Blind Landings

Blind Landings
This page intentionally left blank

Blind Landings
Low-Visibility Operations in American
Aviation, 1918–1958
erik m. conway
The Johns Hopkins University Press

Baltimore
© 2006 The Johns Hopkins University Press
All rights reserved. Published 2006
Printed in the United States of America on acid-free paper
246897531
The Johns Hopkins University Press

2715 North Charles Street
Baltimore, Maryland 21218-4363
www.press.jhu.edu
Library of Congress Cataloging-in-Publication Data

Conway, Erik M., 1965–
Blind landings : low-visibility operations in American aviation, 1918–1958 /
Erik M. Conway.
p. cm.
Includes bibliographical references and index.
isbn 0-8018-8449-7 (hardcover : alk. paper)
1. Instrument flying—United States—History—20th Century. 2. Airplanes—
Landing—United States—History—20th Century. 3. Landing aids (Aeronautics)—
United States—History—20th Century. I. Title.
tl711.b6c66 2006
629.132′521409730904—dc22 2006003795
A catalog record for this book is available from the British Library.
To my parents, Richard and Barrie-Lynn Conway
Contents
Acknowledgments ix
List of Abbreviations xiii
Introduction 1
1 Instrumental Faith 12
2 Places to Land Blind 35
3 Radio Blind Flying 57
4 The Promise of Microwaves 80
5 Instrument Landing Goes to War 104
6 The Intrusion of Newcomers 123
7 The Politics of Blind Landing 137
8 Transformations 163
Conclusion 177
Notes 187
Index 211
Acknowledgments
A dozen years ago, I was afloat in the vast tracklessness of the Pacific Ocean
when I received an e-mail from John Beatty welcoming me to the University of
Minnesota’s Program in the History of Science and Technology. “Welcoming”
proved an understatement, and I want to thank John, Sally Gregory Kohlstedt,
Arthur Norberg, Robert Seidel, Alan Shapiro, Roger Stuewer, and John Euler for
fostering an atmosphere of collegiality amid the horribly cold Minnesota winters.
The friendship and camaraderie o¤ered there made a sometimes-diªcult tran-
sition from navy oªcer to (hopeful) scholar the most pleasurable years of my life.
The idea for this project came from a conversation with another of Minnesota’s
graduates, Robert Ferguson. I am certain that this is far from what he expected it
to be, but I hope that it is still satisfactory. My friends Eileen and Andrew Wein-
gram provided a comfortable home away from home on too rare vacations. And
I cannot thank Barton C. Hacker and Margaret Vining enough for the many re-
search visits they have hosted at their Clio house in Washington, DC, over the
years.
As all historians do, I have accumulated debts to librarians and archivists at
various institutions. At the New England branch of the National Archives, Helen
Engle aided this project enormously through her knowledge of the MIT Radia-
tion Lab papers and her work-in-progress finding aid, more than 600 pages long.
Patricia White at the Stanford University Archives guided me to important files
on klystron development and the Sperry Gyroscope Company’s involvement in
blind landing research. Mary Pavlovich, librarian at the National Air and Space
Museum, was unfailing in her attempts to procure obscure materials via inter-
library loan. Marjorie McNinch at the Hagley Museum and Library provided ac-
cess to the Aircraft Owners and Pilots Association records and additional Sperry
Gyroscope Corporation records, and she arranged permission for use of Hagley
images in this book. Kate Igoe at the National Air and Space Museum provided
access to many images used in this book, and I am grateful for her e¤orts to iden-
tify and secure permission for several rare photographs.
x acknowledgments
Completing this project required access to records still held by operating agen-
cies. Mr. V. Pattanyak, at the International Civil Aviation Organization’s head-
quarters in Montreal, kindly provided access to an extensive collection of docu-
ments related to international technical negotiations. The records oªcer at the
U.S. Federal Aviation Administration, Lisa McGlasson, provided access to many
documents still in the agency’s possession, while FAA’s now retired historian,
Ned Preston, provided key documents from its few remaining historical files. Fi-
nally, Hal Moses at the Radio Technical Commission on Aeronautics allowed me
to use the organization’s collection of 1940’s documents. Without the assistance
of these agencies and their representatives, this history would be poorer.
The assistance of survivors of the World War II–era U.S. Eighth Air Force was
also invaluable. More than thirty veterans responded to an advertisement I placed
in the Eighth Air Force Historical Society’s newsletter to tell me about their ex-
periences flying and landing in the European theater. Collectively, they provided
an insight into wartime problems that I could not have gotten from the published
literature. I cannot thank them all here, but Dr. John A. Clark, Col. Bruce Ed-
wards, Col. Francis L. Grable, Charles W. Halpen, J. W. Howland, Joseph E.
Manos, William D. Reeder, and Bill Rose took the time to write detailed letters
about their experiences. Dan Raymond and George Reynolds sent me articles that
had appeared in various other veterans’ newsletters that I would never have
found, and Chuck Derr sent me a self-published book on his wartime experi-
ences. Most extensive of all has been the correspondence of Col. Robert W. Vin-
cent. Thank you all for your sacrifices.
The National Air and Space Museum supported this project through a pre-
doctoral Guggenheim Fellowship, freeing me from the usual demands of gradu-
ate student teaching and allowing access to the vast archival collections there. The
U.S. Army Signal Corps alone collected more than 13,000 pages on blind land-
ing work between 1933 and 1939, and the Air Corps, Navy Bureau of Aeronautics,
and Civil Aeronautics Authority each accumulated similar volumes of material.
Residence in the Washington area made this project feasible. The National Sci-
ence Foundation, via dissertation improvement grant NSF/SBR-9729517, sup-
ported research conducted outside Washington. I largely completed the revisions
to this manuscript while a contract historian at the National Aeronautics and
Space Administration’s Langley Research Center in Hampton, Virginia.
I am indebted to my advisors, Arthur Norberg at Minnesota and Michael
Neufeld at the National Air and Space Museum, and to Sally Gregory Kohlstedt
at Minnesota, for their careful readings of and comments on earlier versions of
this manuscript. Conversations with my friends and colleagues Deborah G.
acknowledgments xi
Douglas and Stuart G. McCook were important to my thinking about airfields and
landing aids. Their thoughtful comments have improved this book immeasur-
ably, yet all errors, whether of fact or interpretation, are mine alone.
Finally, I thank the National Air and Space Museum, the Hagley Museum and
Library, Unisys Corporation, and the Aircraft Owners and Pilots Association for
permission to reproduce images contained in this work, and to the Johns Hop-
kins University Press for permission to reproduce much of chapter 7 from an ear-
lier article.
Abbreviations
A-1 a U.S. Army Air Corps landing aid, also known as the “Hegen-
berger system”
AAF Army Air Forces
ALPA Air Line Pilots Association
AN/MPN-1 a World War II–era GCA system
AOPA Aircraft Owners and Pilots Association
ARINC Aeronautical Radio, Inc.
ARL Aircraft Radio Laboratory (Wright Field, Dayton, Ohio)
ATA Air Transport Association
BABS blind approach beacon system
CAA Civil Aeronautics Authority (until 1940); Civil Aeronautics Ad-
ministration (after 1940)
FCC Federal Communications Commission
FIDO a thermal fog dispersion device
GCA ground-controlled approach
IATCB Interagency Air Traªc Control Board
ILS instrument landing system
ITRM/ITD International Telephone and Radio Manufacture; International
Telephone Development Company (same company as the Federal
Telephone Development Corporation)
NAS National Academy of Sciences
NBS U.S. National Bureau of Standards
PGP pulsed glide path
PPI planned position indicator
RAE Royal Aircraft Establishment
RTCA Radio Technical Commission on Aeronautics
SBA standard beam approach system
SCS-51 U.S. Army version of the instrument landing system
UAL United Air Lines
xiv abbreviations
UHF ultra-high frequency (after 1944; frequencies between 3 MHz and

30 MHz)
USSTAF U.S. Strategic and Tactical Air Forces
VHF very high frequency (after 1944; frequencies between 30 MHz and
3 GHz)
YB a U.S. Navy version of ILS
Blind Landings
Introduction
Pilot and writer Antoine de Saint-Exupéry, in his classic novel Night Flight, told
of a French mail pilot and his radio operator who, surrounded by storms while
flying between Patagonia and Buenos Aires, were unable to find a place to land.
The pilot’s supervisor, driven by the need to conquer the distance between iso-
lated cities, had ignored the worsening weather and insisted that the mail go
through on time. Saint-Exupéry did not recount their deaths; instead, the two men
simply vanished from the narrative.1 Their fictional deaths, like the real deaths
that occurred all too frequently in the first thirty years of the aerial enterprise,
were counted as merely two more casualties in the struggle to bring about avia-
tion’s great future.
Saint-Exupéry’s novel reflected Charles Lindbergh’s own experience. Lind-
bergh was flying for a mail contractor one night on the St. Louis–Chicago run in
1926 and became lost in a dense fog. When he was nearly out of fuel, he de-
scended as close to the ground as he dared given the unknown terrain and
dropped a parachute flare that he carried for such occasions. He hoped that the
flare would illuminate the ground, permitting him to land in a convenient field,
but it did not. Using the last of his fuel to climb as high as he could, he jumped
out, relying on his parachute to save his life. The next day, he rescued the gasoline-
soaked mail from the plane’s wreckage, which a farmer had found a couple of
hundred yards from his house.2 This happened to Lindbergh twice that year.
His experience was not unusual. Virtually all mail pilots in the 1920s had to
make forced landings, using holes in the clouds to try to find any field big enough
to land their aircraft in to circumvent a crash. A report generated by the Post
2 blind landings
Oªce’s Air Mail Service in 1925 detailed the magnitude of the problem. Between
July 1924 and July 1925, the mail service’s pilots made 750 forced landings, 554
of which (77 percent) were caused by the weather.3 That report’s recommenda-
tions set the next decade’s research agenda for the Army Air Service, the airlines
created by privatization of the air mail in 1925, the regulatory authority created
the following year, and the National Bureau of Standards’ radio laboratory.
Before 1940, researchers called it the “blind” or “fog” landing problem. In-
ventors, engineers, and scientists of diverse backgrounds proposed literally hun-
dreds of technological solutions. These ranged from the suspension of lit balloons
as markers above airfields to the use of high-powered X-rays as guidance beams.
The balloon method was popular with British researchers in the 1920s; happily,
the U.S. Army Signal Corps checked with radiation specialists before seriously
exploring the X-ray proposal. Also popular with inventors were systems based on
the perceived ability of infrared light to penetrate clouds and fog, systems using
underground cables (and some with the cables held above the surface on poles),
acoustic systems, and radio beam systems.
The sheer diversity of proposed solutions to the blind landing problem sug-
gests how large it loomed. Most proposals came from private individuals who sub-
mitted them to the Army Air or Signal Corps in the hope that one of the agencies
would find their idea worthy and develop it. Several proposals received significant
attention from the Corps’ investigators, but few were actually developed into func-
tional prototypes. This book will explore some of these prototype systems. I have
not been able to locate documents relating to every prototype, including one that
gave important service during World War II. These historical orphans, whose ex-
istence is known but whose development I cannot document, will deserve men-
tion in later chapters, but this book is built around documented development
e¤orts. That means, in essence, that it is a history of government-funded devel-
opment projects, primarily in the United States but also, to a lesser extent, in
Western Europe. The majority of these projects followed directly from the rec-
ommendations made in the Post Oªce’s reports, and hence we can trace much
of the infrastructural development of commercial and military aviation to the Post
Oªce’s airmail experiment in the 1920s.
Not all blind landing work stemmed from the airmail experiment. World War
II produced important techniques and technologies as well. The vast air war the
Royal Air Force and the Army Air Forces pursued in Europe relied heavily on the
technologies devised to solve the airmail carriers’ problems while inspiring many
more, some of which were made available after the war to civil aviation interests.
The infrastructure necessary to support all-weather flying in the United States is
introduction 3
thus a product of what a more conspiracy-minded scholar might call a “military-

postal-industrial complex” that was joined in the late 1930s by two universities,
Stanford and MIT.4
The goal that the Air Corps, the Post Oªce, and, after airmail privatization,
the airlines shared was all-weather operations. They, the privately financed Gug-
genheim Foundation for the promotion of Aeronautics, and the National Bureau
of Standards all exchanged knowledge and equipment to reach that goal. While
the United States Navy also pursued blind landings, it kept its research secret and
hence had little impact. The same is true of French work. During the 1920s,
France pursued a much di¤erent set of technologies in secret, permitting only
the barest of descriptions to be published and thus having no influence at all in
the United States. Germany, in contrast, openly adapted American research after
1932, when it was largely freed from Versailles Treaty constraints. After World
War II, U.S. technologies were adopted as the world standard, eliminating the di-
versity of techniques that marked prewar aviation.
Histories of aviation are often written as quest narratives, and the extant his-
tories of blind landing research are of this type. James Hansen has criticized his-
torians of “going native” when studying aviation and reinforcing the master nar-
rative of linear development and untrammeled success. Failures are thus typically
forgotten. The blind landing story has been told in this way in the past.5 But such
an approach disguises a far more interesting story. While blind landings have
been achieved, they have not become routine. Instead, the tremendous e¤ort that
went into solving the blind landing problem produced an infrastructure capable
of supporting almost all-weather operations. Yet the goal, completely blind land-
ings, remained elusive at the end of aviation’s first century despite substantial
advances in precision operations. This book thus will investigate a seeming para-
dox: while scientists and engineers developed systems increasingly able to pro-
vide reliable blind landings, pilots and regulators became less and less convinced
that routine blind landings were either feasible or even a good idea. The result
was that equipment intended to produce blind landings was installed worldwide
but was used for a lesser purpose.
This book is both a history of the technologies of blind landing and a history
of the idea of blind landings. It will trace the technologies of blind landing—at
least some of them—through the cycle of invention, testing, improvement, ne-
gotiation, and adoption, while also following the gradual collapse of the concept
of “blind landings.” Through following these two threads, this book seeks to ex-
plain why the quest for blind landings failed.
When I first began looking into the history of aircraft landing aids, to use the
4 blind landings
post–World War II term, I was struck by the use of the terms “blind flying” and
especially “blind landing” in the popular aviation magazines and newspaper ar-
ticles of the interwar period. Living at the tag end of the twentieth century, and
jaded with the knowledge that blind landing was simply not done, the idea of at-
tempting it struck me as both dangerous and silly. The safe thing to do if bad
weather is predicted at your destination or along your chosen route is stay home,
as generations of prospective pilots have been taught in ground school. To facili-
tate that obvious bit of wisdom, the first piece of national aviation infrastructure
built in the United States was an aviation weather network, constructed during
the middle 1920s.6 Yet something drove fliers like Saint-Exupéry and Lindbergh
to ignore wisdom and challenge the weather anyway. Understanding why they did
so is our first task.
Aviation was the first form of transportation to seek perfect all-weather oper-
ations, at least in any organized way. Automotive pioneers made no attempt to
make cars and trucks capable of blind driving, no sane sea captain sailed into port
in a fog—that’s what anchors are for—and trains quite literally stopped in their
tracks when weather became so bad that the engineer could no longer see. These
other forms of transportation had one advantage over airplanes in dealing with
the weather in that they could stop and wait it out. Once airborne, an airplane
must always be in forward motion. Circling to wait out the weather was an option
for aircraft, but only within the limits imposed by fuel, as Lindbergh’s story sug-
gests. Hence one could argue that safety motivated the development of blind
landing systems, but that argument runs afoul of my earlier safety point: if one’s
primary concern is flight safety, one stays firmly on the ground. Since safety is-
sues could be resolved by not flying, safety was not the sole motivation behind the
development of blind landing systems.
Saint-Exupéry’s novel suggests both of the primary reasons aviation pioneers
sought to achieve blind landings: profit and progress. The profit motive is fairly
obvious. Aircraft that sit on the ground produce losses, not profits. One subtext
of Saint-Exupéry’s story is the pressure to perform the airmail supervisor placed
on his men, both pilots and mechanics. In the United States the Post Oªce was
the original source of the drive to beat the weather. While the Post Oªce itself
was not a profit-seeking enterprise, its leaders intended the airmail service to
become profitable so that it could be commercialized, thus forming an eco-
nomic basis for commercial air service on a national scale. Bad weather interfered
with the potential for profitability. Successful commercialization meant that the
weather’s e¤ects on aviation had to be mitigated through some technological
means. While the weather also interfered with trains, which the airmail pioneers
introduction 5
considered their primary competition, the interference was much less. This was
because visibility at ground level is almost never completely absent, while solid
overcast at a thousand feet or so is quite common.
Perhaps more important to the profit motive, however, was aviation’s “dirty
little secret”: the airplane was not, in a practical sense, faster than trains were
when airmail service began. While airplanes might cruise at 100 mph and the
iron horse steamed along at only 40 mph, the train still won because aircraft had
to stop at night and for bad weather. Trains could travel round the clock, while a
wintertime airplane flight had only eight hours of daylight and spent about an
hour of that on the ground getting fuel. Worse, mail pilots often found that they
might be able to land in worsening weather but not take o¤ again, introducing
another sort of delay. Realization that the airplane was actually slower than the
railroad was quick in coming. The Post Oªce tried several remedies beginning
in 1920, but the most successful was a string of beacons, in reality extremely
bright searchlights on rotating platforms, that it erected during 1923–24 to con-
struct a transcontinental airway able to support round the clock flying. Mail pi-
lots flew from searchlight to searchlight, and most of the original group of mail
pilots died doing it. Yet the searchlights fixed the “night flying” problem well
enough so that the airmail could beat the trains in good weather conditions.7
Once the Post Oªce had proven night flying possible and profitable, it priva-
tized the airmail, contracting with groups of entrepreneurs to provide airmail ser-
vice over the Post Oªce’s old routes. These entrepreneurs founded the airlines
that dominated U.S. commercial aviation through deregulation in 1978. The Post
Oªce provided a deliberately generous subsidy to the airmail companies to en-
sure the service thrived. The Post Oªce’s leaders also chose and assigned routes
to expand service and prevent excessive competition. Finally, it structured the
mail contracts to encourage airmail companies to carry passengers. Passenger
service, in turn, demanded that the airlines achieve a much greater consistency
of operation than the Post Oªce did. The mail cared very little if it arrived a few
hours late due to bad weather or wound up in a di¤erent city than expected.
Passengers accustomed to scheduled rail service were much less forgiving. Be-
cause the Post Oªce’s, and airlines’, goal was the creation of economically self-
supporting scheduled passenger service, routine blind landing assumed sub-
stantial economic potential.
But the dependence of the air carriers on government subsidies (which con-
tinued in the United States into the 1950s) raises two other questions. What mo-
tivated governments to create these businesses in the first place and, in the sec-
ond place, to keep them afloat financially? There are several parts to that question,
6 blind landings
and they all revolve around the western notion of “progress,” especially the tech-
nological kind.
Three historians have discussed the modernist notion of “progress” in the con-
text of aviation, from very di¤erent approaches. The first of these is Robert Wohl,
who has examined western society’s response to the airplane before 1918. He
shows that Americans and Europeans responded to the airplane with enormous
emotion once it had, beyond a doubt, been proven true by the Wright brothers’
highly publicized demonstrations in 1908. Wohl calls the enthusiasm that the
Wrights’ e¤orts generated a “passion for wings.” That passion inspired artists, ar-
chitects, poets, and writers to present the airplane as a cultural symbol, while
other young men, and a few young women, devoted their lives to promoting the
new technology through air shows, races, and attempts to set speed, distance, or
altitude records. Some people gave up otherwise successful careers to pursue the
airplane. Briton T. E. Lawrence, who gained fame as the leader of an Arab upris-
ing against the Ottoman Empire during World War I, gave up his oªcers’ com-
mission and the life of celebrity his exploits had earned him to become a private
in the Royal Air Force, believing that “the air is the only first-class thing our gen-
eration has to do.”8
The airplane’s ability to break free from the earth, if only briefly, enabled it to
become the vessel for long-held hopes for human betterment. Historian Joseph
Corn calls the belief people like T. E. Lawrence held the “winged gospel”: the air-
plane would improve the human condition when intercontinental flight permit-
ted the free admixture of peoples and ideas. This, these true believers fervently
hoped, would eliminate the misunderstandings they perceived as the causes of
war. Corn writes that Americans “viewed mechanical flight as portending a won-
drous era of peace and harmony, of culture and prosperity.”9 The airplane’s abil-
ity to evoke such utopian imagery, he argues, descended in part from Christianity.
Christian religious symbols, from the Star of David through the angelic hosts to
the ascension of Christ, linked Heaven directly to flight. The flying machine, by
its very nature, drew on these ancient symbols to stimulate a utopian vision of the
future. That utopian aerial future would come about simply through the opera-
tion of unimpeded progress.
The religious symbolism of flight also reinforced another traditional thread in
American culture, technological utopianism.10 Corn traces this thread partly to
the rise of evangelical Christianity and partly to the sheer pace of technological
change in the nineteenth century. In a single lifetime, the development of large-
scale production techniques and the deployment of mass transportation tech-
introduction 7
nologies of steamboat, steamship, and railroad had transformed American life.

One could not live through such an age without developing an implicit recogni-
tion of the power of these new technologies to alter old patterns of life. Those
transformations, in turn, fed upon the Enlightenment notion of progress and its
seeming promise to deliver social improvement concomitant with the advance of
technology.
Corn shows that this vision of “aerial progress” had strong democratic over-
tones. Belief that the airplane was a force for the spread of democracy ran deep
enough that, before 1940, the federal government attempted to develop an air-
plane for the “everyman.” There was wide public support for the goal of an “air-
plane in every garage.”11 In part, this was a reaction to the hated railroad trusts,
since personal airplanes could allow the public to simply bypass them. It was also
an extension of the public’s response to Henry Ford’s notion of personal auto-
mobility. Ford’s own 1925 entry into airplane manufacturing was greeted with
enormous enthusiasm by a public that believed he could make airplanes available
to the masses the same way he had with the car. He failed, but others kept trying
through the late 1940s, when the public’s enthusiasm for the winged gospel
finally faded. The dream was kept alive, however, by private fliers’ clubs and lobby
organizations, which will be important later in this book.
Yet there was more than one possible “aerial future” implicit in the winged
gospel’s vision. While Corn focuses on the mainstream version, he briefly ac-
knowledges a minority thread within the vision: the airplane as mass transit. This
was the vision held by Otto Praeger, the assistant postmaster general who built
the initial airmail service; by Herbert Hoover, who as secretary of commerce
helped produce the initial air carriers; by entrepreneurs like Great War ace Eddie
Rickenbacker, who headed Eastern Airlines; by the near-legendary William Boe-
ing, whose two air-related startups, Boeing Air Transport and Boeing Aircraft
Company, eventually became the two largest businesses of their type in the world
(United Airlines and The Boeing Company, respectively); and by investors like
Henry Ford, William A. Rockefeller, and Cornelius Vanderbilt Whitney.12 The
airplane-in-every-garage was not at all what they perceived in the airplane. In-
stead, they imagined a ticket counter in every drugstore. This commercial vision
of the airplane-as-mass-transit was much more important to the development of
blind landing systems than was the more populist democratic vision. The drive
to maintain reliable, scheduled service provided the direct impetus for many of
the innovations this book documents. It was the commercial version of aerial
progress that motivated government support and regulation of aviation in the
8 blind landings
early years, too, most clearly reflected in the Post Oªce’s involvement, in the
placement of aviation regulation in the Department of Commerce, and in the
foundational legislation’s very title: the Air Commerce Act.
Missing from Corn’s book entirely is a third thread that was an inherent part
of the aerial future: the airplane-as-ultimate-weapon. This was the dark side of
aviation, recognized much more strongly in Europe than in America. Frenchman
Jules Verne recognized the potential well before the Wrights actually flew, and
British and Italian strategic theorists generated the doctrine known as “strategic
bombing” during and after World War I. Yet plenty of Americans recognized,
and approved of, this version of “aerial progress.” Most famous among them
was Brigadier General William Mitchell, whom the U.S. Army eventually court-
martialed for publicly criticizing army aviation policy. Many historians of military
aviation have documented Mitchell’s public promotion of the airplane-as-weapon
and his quest for a military air arm equal in stature (if not superior) to the army
and navy. Seldom explicitly acknowledged was that Mitchell’s proposed Depart-
ment of Aviation would have placed all aviation under military control, a very
di¤erent vision of aviation’s future and one that General Henry “Hap” Arnold
tried again to achieve in 1942 by working to militarize the Civil Aeronautics Au-
thority.13 Civil pilots and the airlines were vehemently opposed to military rule,
although it was the navy’s refusal to release the many naval reservists within Civil
Aeronautics Authority to the army that directly blocked Arnold’s attempt at dom-
ination. Mitchell’s and Arnold’s military vision of the aerial future was thus dia-
metrically opposed to the ideal of democratic aviation because it envisioned cen-
tralized control over all aviation activities in the United States.
These three visions of aviation’s future mostly manifested themselves in
conflict over infrastructure. While one could design individual aircraft for specific
purposes, the enthusiasts of the air all agreed that only one infrastructure to sup-
port aviation should be built. The selection of technologies for that infrastructure
became the most hotly contested part of aviation development because advocates
promoted technologies that “fit” their vision of the future. The plethora of blind
landing system designs created during the interwar period provided wide latitude
for this politicized process of selection to operate. On some occasions, the process
of resolving disputes over blind landing system selection took the form of nego-
tiations. In others, disputants resorted to congressional politics to get their way.
Di¤ering visions of aerial progress thus directly a¤ected the development and de-
ployment of blind landing systems.
In a recent book, historian Eric Schatzberg has investigated the role played by
belief in what he called “the progress ideology of metal.” He argues that belief in
introduction 9
this ideology, which promoted metal as the “modern material” in contrast to old-
fashioned wood, caused aircraft manufacturers, regulators, and purchasers to
demand all-metal aircraft after the Great War. They did this despite what he calls
the technological indeterminacy of the situation: researchers were never able to
demonstrate convincingly that either material was superior in terms of strength,
maintainability, or ease of manufacture. Schatzberg contends that this inability
to prove objective superiority left a “social space” in which ideology could oper-
ate to influence the research agenda.14 A similar analysis can be applied to the
case of blind landing systems: several systems worked about equally well. None
could be proven superior, and that lack of objectively demonstrated superiority al-
lowed the three competing visions of aviation’s future to influence the selection
of blind landing systems.
Besides ideological motivations, there were several other major influences on
the development of blind landing systems. “Nature” gave inventors, engineers,
scientists, and entrepreneurs no end of trouble, and early hopes for a quick solu-
tion to the blind landing problem collapsed as the magnitude of nature’s impact
on the operation of these systems became clear. Both what William Cronon has
called “first nature,” the natural environment, and “second nature,” the human-
built environment, caused problems in roughly equal measure.15 It would be too
much to claim that these inventors negotiated with nature to achieve success, but
they gradually learned that a rather intimate understanding of the e¤ects of
both regional and local operating environments, and a close tailoring of both the
equipment design and specific, local installations to those conditions were nec-
essary to achieve even a possibility of success. Systems that could not be tailored
to local conditions were early casualties of the selection process. Even the mature,
more flexible landing system adopted worldwide after World War II could not be
tailored to all existing localities, and a great many airfields worldwide therefore
cannot use it. While progress ideology was important as a motivator for techno-
logical change, “nature” influenced the technologies at a very detailed level.
So too did that amorphous, hotly debated thing called “human nature.” I ar-
gue that the blind landing quest’s failure was not an engineering failure but a hu-
man one. Scientists and engineers achieved technological systems capable of
blind landings, a reality evident to anyone familiar with “smart weapons” and ro-
bot aircraft, both of which were in use by the waning days of World War II. If one
can make a bomb find its own way to a target the size of a ship, one can certainly
make an airplane find its way to a much larger runway. The first transatlantic
flight of a fully robotized aircraft duly occurred in 1948.16 Yet by this time the very
idea of blind landings had vanished from the literature, replaced on the orders of
10 blind landings
the armed services and the Civil Aeronautics Authority by a new term, “instru-
ment approach.” Flying and landing an aircraft “blind,” using only the instru-
ments provided in the aircraft, was ultimately an act of technological faith. The
evangelists of blind landing discovered to their dismay that average pilots and reg-
ulators had come to believe that no machine was trustworthy enough to actually
land an airplane without visual verification by the human crew. Hence the final
thread in this history is a social one. Pilots were the arbiters of technological
progress in aviation: technologies that did not earn their faith vanished; those
technologies that succeeded are still with us. Faith was neither automatic nor ab-
solute. Pilots had to be convinced to trust their equipment enough to fly blind.
Most never achieved that level of trust, however. Faith had its limits, and we must
therefore examine how pilots, not just the famous great pilots but ordinary fliers
as well, responded to these technologies to get at those limits. This also means
we must examine the development of procedures for the use of blind flying and
landing technologies, because building reliable procedures and training pilots to
use them were key elements in the production of pilot faith.
By the end of World War II, the majority of the technologies of the modern air
traªc system had been developed in the course of e¤orts to produce all-weather
operations. Yet individual technologies were insuªcient to solve the weather
problem. What no one foresaw before the war was the need for integration.
Wartime operations made that clear to aviation leaders, who began to make
progress toward integrated airport approach and landing control in the late
1940s. The e¤ort stalled out as the nation returned to prewar levels of budget
stringency, but the immediate postwar e¤ort included all of the prewar develop-
ments, radar derived from wartime programs, and, in a complete rejection of
blind landing, runway lighting. With the exception of digital computers, this suite
of technologies was the basis of the modern air traªc system, deployed nation-
wide after a series of catastrophic accidents created renewed political pressure for
reform.17 While the construction of modern air traªc control is beyond the scope
of this work, its roots are in this immediate postwar drive for integration.
The book begins with an examination the development of blind, later called
“instrument,” flying, to argue that the development of technique was as impor-
tant as the development of technologies for flying blind. In Chapter 2, I explore
the relationship between airfield design and blind landing systems, to argue that
the conversion of paved runways in the United States during the late 1920s ren-
dered one possible blind landing system unusable. The early history of the cur-
rent “instrument landing system” is the subject of Chapter 3, which argues that
the instability of early versions helped delay the system’s adoption while forcing
introduction 11
engineers to innovate to stabilize them. Chapter 4 examines one potential solu-

tion to the instability problem, microwaves, pursued jointly by Stanford, MIT, the
Army Signal Corps, the Civil Aeronautics Authority, and the Sperry Gyroscope
Company, in the context of disputes and negotiations between government agen-
cies and airlines over whether to adopt improved versions of the original National
Bureau of Standards system as the U.S. standard or the microwave system.
World War II serves as a major division in the book. The European air war
marked the first attempt to use air power in all weather conditions, and the Al-
lied forces quickly found that they needed a blind landing system—or several.
Chapters 5 and 6 each examine the development and deployment of one such sys-
tem, one using ultra-high frequencies (UHF), which eliminated the microwave
system that had appeared so promising in 1939, and a radar based system that
became famous as “ground controlled approach.” Chapter 7 traces the postwar
fight between supporters of these two systems over which to adopt as the national
standard, to argue that the struggle represented a conflict between two very di¤er-
ent visions of aviation’s future. Chapter 8 examines the transformation of the two
systems into an almost-all-weather system of landing aids, marking the final ac-
ceptance that routine blind landings were simply not humanly feasible.
chapter one
Instrumental Faith
In the 1920s and 1930s, poor visibility was much more likely at the altitudes
at which airplanes flew than it was at ground level due to the frequent occurrence
of low clouds in the most populous regions of the United States. Early aircraft
were restricted to altitudes of only a few thousand feet (at best), and thus they
could not fly above the weather. They had no choice but to fly through it. For this
reason, the blind flying problem was a more immediate economic concern than
was the blind landing problem; fortunately, it proved a more tractable issue. By
the mid-1930s, it had largely been solved.
There were two aspects to the blind flight problem: control of the aircraft in
the absence of outside visual references, and navigation of the aircraft from one
point to another without external vision. Innovation in point-to-point aerial nav-
igation has been well documented elsewhere.1 This chapter focuses primarily but
not exclusively on the first aspect of blind flying, control of the aircraft. Before the
late 1920s, the distinction had not yet been made. Many early pilots believed that
their own motion senses provided all the information necessary to maintain con-
trol of the aircraft, and hence point-to-point navigation was considered the only
problem. The real danger in blind flying, however, was not in getting lost. It was
in inadvertently losing control of the aircraft. Before 1926, no one knew why this
happened, and most fliers simply blamed it on insuªcient skill among the un-
fortunate. There was a physiological reason for loss of control during blind flight,
however, and it required a solution.
The solution was not primarily mechanical. The instrumentation necessary to
resolve the blind flying problem already existed at the end of World War I, but for
instrumental faith 13
a decade after, no one used it for that purpose. Instead, pilots continued to fly as
they always had—by the seats of their pants. In the late 1920s, a pair of U.S. Army
oªcers, William Ocker and Carl Crane, began a campaign to convince their peers
that this was no longer an acceptable method for professional pilots. Human mo-
tion senses did not function accurately in the three-dimensional world of flight, they
claimed. Instead, these senses misled pilots and caused them, ultimately, to kill
themselves. Based upon research they had done on other army pilots and on air-
mail service pilots, they contended that aviators had to ignore their motion senses
and follow a set of rules built around their instruments in order to fly blind safely.
Their mechanistic version of flying ran directly counter to the prevailing
mythology of flight that good pilots were born, not made. They met a great deal
of resistance at first, but a series of events beginning with Charles Lindbergh’s fa-
mous transatlantic flight gradually overcame professional pilots’ resistance to the
new rule-based method of blind flying. The most important such event was the
airmail crisis of 1934, which caused the adoption of simulators to ensure that
the new rules of flying were instilled as habit—without leaving the ground. Ha-
bituation was necessary because the central problem in blind flying was one of
faith. Pilots had to learn to believe in their instruments rather than themselves.
This was not an easy task.
instruments for blind flying
The Great War provided the impetus for the development of a wide range of
aircraft instruments and aerial navigation techniques. Prewar aviation had been
exclusively a fair-weather a¤air, and most flying was local. Aircraft took o¤ from
and landed on the same airfield, having gone no more than a few miles away. Pi-
lots could use familiar landmarks for what little navigation they did. They could
pick a landmark and fly toward it, then choose a new one for their next waypoint.
They also made use of roads, railroads, and rivers as navigation aids by simply
following them. This kind of flying became known as contact flying because the
pilot remained in visual contact with the ground.
Contact flying proved to be unwise during the war for a number of reasons.
Ground fire was one obvious problem. Even small arms could hit an aircraft flying
close enough to the ground to see landmarks, and by the end of the war antiair-
craft guns capable of reaching above ten thousand feet had been produced. The
desire to stay alive drove pilots to fly too high to rely on landmark navigation. Even
discounting the e¤ect of gunfire, however, contact flying was not very useful. Con-
tact flying worked best when a pilot was familiar with the route over which he was
14 blind landings
flying, but pilots in the Great War (especially American pilots in the war’s final
eighteen months) were expected to operate over hostile and unfamiliar terrain.
There were no aeronautical charts of enemy territory, and charts of friendly ter-
ritory were new and not yet widely available. Railroad maps became precious to
pilots because the iron horses’ tracks were permanent, visible from high altitudes
even at night (their polished surfaces reflected moonlight), and always led some-
where. But following the railroads imposed limitations and dangers. They did not
go everywhere along the front; hence relying on them limited potential targets.
And if planes routinely followed the tracks, the enemy would quickly line the
tracks with guns to shoot them down. The “iron compass” of the railroads pro-
vided a decent emergency navigation aid, but turning the airplane into a really
useful weapon meant replacing contact flying with navigation less dependent on
detailed knowledge of enemy terrain.
The obvious solution was the same kind of chart, compass, and stopwatch nav-
igation that navies of the world had used since the fifteenth century—except that
the now-ancient technology of the magnetic compass did not seem to work in the
air.2 More precisely, it worked in tethered balloons but not in airplanes, where it
behaved erratically. Sometimes the compass seemed to “stick” and not respond
at all to a turn, while at other times it responded too quickly and overshot the
plane’s actual heading. It might start to turn one direction when the plane was
turning the other direction. The compass provided reliable indication only when
the plane was in straight, level flight. This capricious behavior caused fliers to
mistrust the magnetic compass and caused researchers to seek an explanation
and a means of fixing the problem.
Keith Lucas, a researcher at the Royal Aircraft Factory, explained the compass’s
capriciousness in late 1914. The curvature of the earth’s magnetic field was the
culprit. Because the north-seeking end of the needle is pulled “down” toward
the north pole (in the Northern Hemisphere), compass manufacturers placed the
pivot point o¤ center slightly to keep the needle from actually pointing down. As
long as the compass was level (and high-quality compasses included bubble lev-
els for precise reading), the needle stayed flat. But since airplanes turned by bank-
ing, thus raising one wing above the other, the compass was not level with respect
to the earth’s magnetic field during turns. As the plane rotated around the nee-
dle, the o¤-center pivot combined with the field’s downward pull caused erratic
movement. With Lucas’s explanation at hand, the Royal Aircraft Factory was able
to ameliorate, but not fix, the magnetic compass’s unreliability.3 Yet recognition
that the magnetic compass could never be used as a turn indicator led investiga-
tors to try other techniques.
One approach was to detect the degree of an aircraft’s bank. That proved sim-
ple to do, and by the end of the war a bank indicator consisting of a fluid level
with a small metal sphere in it to mark the angle had been developed. While fine
for detecting a bank, it proved of little use as an instrument to supplement the
magnetic compass, partly because one can turn a plane without banking (prop-
erly called skidding), but mostly because it was not very precise. One could not
use it to make small course changes because it did not detect small angles of bank
reliably and because it was slow to respond in any case. Hence it was not a good
turn indicator.
Because of its well-known ability to maintain a constant orientation in space,
the gyroscope was the immediate favorite of inventors interested in making a turn
indicator. The idea of a gyro turn indicator actually preceded the airplane, ac-
cording to an internal memo from the Pioneer Instrument Company, a spin-o¤
from the more famous Sperry Gyroscope Company. British patent #3587, dated
February 17, 1899, covered the basic idea of a turn indicator and included draw-
ings that functionally replicated the equipment that Pioneer Instrument began to
produce in 1919. An air-jet powered gyroscope rotor was suspended in a frame
that was free to rotate about an axis perpendicular to the axis of rotation. As the
aircraft turned, it literally rolled around the gyro, and the geared frame moved an
indicator needle that showed the divergence between the gyro’s orientation and
the aircraft’s. Elmer Sperry’s son Lawrence test flew the device in 1917 but found
that the instrument was not responsive enough. Two years of work by Elmer and
Charles H. Colvin produced one that Lawrence found acceptable. Elmer filed his
patent for it in 1920, but at least a half-dozen other gyroscope-based turn indica-
tors appeared by 1919. The Royal Aircraft Factory, for example, produced a de-
tachable gyro turn indicator that was simply clamped somewhere convenient on
the airframe, with the open gyro wheel powered directly by the air flowing past
the aircraft, while the German Drexler “Steering Gauge” used an electrically pow-
ered gyro to do the same thing.4
Sperry Gyroscope spun o¤ Pioneer Instrument Corporation in February 1919,
when several of Sperry’s employees decided they wanted to keep working in aero-
nautical instrumentation after Elmer Sperry proclaimed that there was no future
in them. They were able to license the patent rights from Sperry to design the
turn indicator shown in Figure 1.1. Pioneer received a substantial order for the
turn indicator from the French l’Armée de l’Air in 1919, but the U.S. Army Air
Service, a unit of the Army Signal Corps, did not follow the French lead.5
Yet the army’s failure to adopt the instrument immediately did not matter be-
cause fliers had not yet recognized the turn indicator’s most important use. It had
16 blind landings
been invented as a means to overcome the instability of the magnetic compass

during a deliberate turn, which was how those pilots who had it used it. The turn
instrument was a rate instrument, which meant that it displayed, in e¤ect, how
fast the aircraft was turning. One could therefore make turns accurately by plac-
ing the needle at a particular spot on the gauge face and then timing the turn.
With practice, a pilot could learn to accurately make a turn of any size by this
method. The use of the turn indicator was therefore based on intention—pilots
used it when they intended to make a turn. Otherwise, they concentrated on the
magnetic compass.
But fliers with the turn indicator kept dying, including some accomplished he-
roes. Capt. John Alcock, who made the first nonstop transatlantic flight in June
15, 1919, died in a fog-related crash that December, leading Elmer Sperry to write
to Lawrence, “We have absolutely got to solve the problem; if we die in the attempt
and could have registered a single notch in advance, it seems to me that it would
be well worth while.” Tragically, Lawrence did not live to see the problem solved,
drowning in the English Channel on December 13, 1923, after a daylight crash.6
But Elmer’s attitude was hardly unusual. In the same letter, he quoted Charles
Kettering: “We do not yet know this airplane business. One need only fly for a few
months to realize that the airplane is developed to about the same point as if we
had ships with absolutely no paraphernalia of navigation. . . . The only way we
will bring [navigation paraphernalia] into existence is to try to do something; to
find out what we can do and then do it. And that is why it is right to continue the
airmails. That is the reason I believe in them.”7
The airmails, as Kettering prophesied, were the laboratory in which blind
flying was perfected. The Post Oªce did not work alone at this. Most of the hard-
ware the Post Oªce tested was developed by the Army Air Service at Wright Field
or by private companies. But the Post Oªce did run a small experimental facility
at the Air Mail Service’s Development Division at Monmouth Field, Illinois,
where it orchestrated testing of these innovations.
At war’s end, blind flying research work largely ceased in the U.S. Army until
the mid-1920s, while the U.S. Post Oªce, which then prided itself on its tech-
nologically progressive nature, took up the mantle of aviation progress. The Post
Oªce had begun canvassing Congress to get funding for an airmail experiment
in 1910, without success. In early 1918, the Army Air Service had approached sec-
ond assistant postmaster general Otto Praeger, who was lobbying hard to get air-
mail funding from the U.S. Congress, with an o¤er to fly the mail. The Air Ser-
vice needed the experience with cross-country flying, and its oªcers believed that
flying the mail was the way to get it. Praeger could hardly turn down the army’s
Figure 1.1. Pioneer Instruments turn indicator. National Air and Space Museum,
Smithsonian Institution (SI 83-7649).
o¤er, given the patriotic fervor of the day, and the army began flying the mail in
May. The army’s leadership quickly abandoned the project however, and by Au-
gust had turned over pilots and aircraft to the Post Oªce.8
Praeger believed that the army had failed because it was not devoted enough
to the airmail cause. Although not a pilot, he was a true believer in the airplane’s
great potential. Like Kettering, he believed that only the government could de-
velop the airplane to a point at which it could be commercially viable. He expected
that the airmail would lose money and pilots for years, but like the Sperrys, he
believed the sacrifice was necessary to bring about the aerial age.
Praeger’s belief in the airplane’s potential, coupled with his own lack of per-
sonal knowledge of the technology’s current limitations, led him to place enor-
mous pressure on his pilots to fly in all conditions. Praeger had suspected that
the army’s pilots routinely had canceled flights for bad weather when the trip was
actually flyable. He forced the mail pilots to sign contracts that specified their re-
sponsibility to fly in fair weather or foul, and acted on them. When four pilots re-
18 blind landings
fused to fly the New York to Washington, DC, run on July 22, 1919, because visi-
bility was less than two hundred feet, he ordered the two senior pilots fired and
the two junior pilots reprimanded.9 This action came at the end of a month in
which two aircraft had been demolished in weather-related accidents from which
the pilots had walked away and three days after the death of a pilot who crashed
after becoming disoriented in cloud cover. The New York pilots initiated a brief
strike over the firings which was resolved when Praeger reinstated one of the two
pilots and granted airfield managers the authority to determine flyability (thereby
removing the authority from the pilots and from himself ). The tone was set, how-
ever, and the pilots’ contracts continued to state that refusal to fly constituted their
resignation.
Praeger’s successors were somewhat more safety minded, but the pressure to
fly in all weather was endemic to the enterprise. Businesses depended on the mail
more for reliability than speed, and unless the airplane could be made to perform
on schedule, it could never compete with the older, better established, and more
reliable forms of mail transportation. Its sole advantage over those forms of trans-
portation was its superior speed, but that counted for nothing if the plane had to
stop at night or for bad weather. Express trains, the primary means of carrying
mail in the first half of the century, were slower but stopped for almost nothing.
During the long winter nights especially, the trains could overtake and pass the
faster, but grounded, mail planes. Hence the mail planes did not start to beat the
express trains routinely on the New York to San Francisco route until they started
to fly the segment from Chicago to Cheyenne, Wyoming, at night in 1923.10
The airmail service did not become profitable for the Post Oªce until it began
flying the New York to Chicago route at night in 1925. A group of Chicago bankers
had carefully explained to Praeger’s successor, Paul Henderson, that because the
airmail left both cities in the morning and arrived in the evening, it still arrived
in the banks the next morning along with the mail that went by train. There was
no speed advantage from the banks’ point of view and thus no point in paying the
higher airmail postage. If the mail left at night, however, and arrived the next day,
it would essentially eliminate the transit time for bank drafts (since the banks
were closed at night), saving the banks millions of dollars in “float” costs per
year.11 This savings would make airmail well worth the cost. The Post Oªce lis-
tened, and its airmail volume between New York and Chicago exploded during
the last two years it flew the mail before privatization.
Hence the airplane’s miles-per-hour speed was irrelevant if it did not translate
into faster service than that provided by the railroads. That, in turn, meant all-
weather, round-the-clock operation. Kettering, Sperry, Praeger, Henderson, and a
few other entrepreneurs intended to use the airmail to develop the technologies
and techniques to permit all-weather flying. In essence, the airmail was a real-
world laboratory for flying. Most significant about the airmail experience was not
new technologies; in fact, the Post Oªce developed and deployed no instruments
or other navigation technologies that had not existed during World War I. What
it did was build an infrastructure to support flying, while its pilots began to build
a new kind of flying around the instruments wartime development had be-
queathed them.
That a new kind of flying was necessary was not immediately obvious to the
Post Oªce’s pilots, and most of them continued to fly the way they always had.
Quite a number of them died doing it, and more crashed but lived. There were
essentially two kinds of weather-related accidents during the airmail years. The
first type, flight into terrain, were caused when pilots in poor weather descended
to very low altitudes to keep the ground and its handy navigation aids in sight. On
October 14, 1919, for example, pilot Lyman Doty was flying the Washington to
New York route when he descended below treetop level in an attempt to keep the
ground in sight. He hit a tree, the plane caught fire, and he burned to death. On
March 10, 1920, Clayton Stoner was the victim of another in-flight collision with
a tree on the Chicago to Cleveland route, and later that month Harry Sherlock hit
a smokestack while approaching Newark’s Heller Field. While flying in four-
hundred-foot visibility from Chicago to Cleveland in April 1921, J. Titus Chris-
tensen was following the Cuyahoga River when he entered a cell of clear air over
the river, which was surrounded by tall buildings and bridges on all sides. He tried
to climb out in a tight circle, staying inside the clear cell, but stalled the plane and
crashed into railroad tracks.12 Trying to contact fly in bad visibility was extremely
hazardous.
Yet the obvious alternative, staying up high in the clouds where trees and
buildings did not reach, was perhaps even more frightening. Many of the mail pi-
lots who tried this alternative died too. While trying to fly over the Snowshoe
range of the Allegheny Mountains in Ohio in 1919, Charles Lamborn’s plane en-
tered the clouds only to reappear a few minutes later, nose down, and plummet
into the ground. John Charlton died the same way that year, apparently after spin-
ning into a mountaintop. On January 4, 1923, Henry Boonstra tried to fly over a
storm at eighteen thousand feet when his engine died. Trying to glide down
through the clouds, he lost control of the airplane and went into spins five times
before finally hitting the ground.13 Fortunately, he lived to tell the story. A lot of
other pilots experienced these spins in clouds and recovered from them when the
spin dropped them to lower altitudes with a visible horizon.
20 blind landings
The mail pilots found these spins to be deeply frustrating because they could
not explain the phenomenon. Their aircraft were aerodynamically stable, mean-
ing that the airplane would return without pilot action to straight, level flight af-
ter a small disturbance. Thus the pilots were becoming disoriented and causing
the spins themselves, but they did not know why. Most early mail pilots chose to
take the collision risk of flying low rather than to face the spin risk of flying blind.
The pilots who chose to face the spin threat, however, were the ones who ulti-
mately produced the new kind of flying.
The disorientation problem, they learned from anatomists, was a result of
how humans sensed motion. Beginning in 1926, two army oªcers, pilot Capt.
William C. Ocker and flight surgeon Maj. David Meyers began a campaign to con-
vince pilots that the error lay in believing that people could fly using their sense
of balance. People maintained their equilibrium through a combination of the
fluid levels in the canals of the inner ear, through muscle balance, and through
vision. Deprived of vision by clouds, the other two senses became misleading,
causing pilots to make inappropriate control actions. The two men had employed
a Jones-Barany chair, essentially a barber’s chair, to demonstrate that “after ac-
celerations, either angular or linear, a man cannot competently interpret the
movements of his body unless he can see some plane of reference such as the
earth or objects stationary in relation to it.”14 By spinning pilots in the chair a few
moments at a constant speed, then slowing it, Ocker and Meyers demonstrated
that their subjects believed that they were now spinning in the opposite direction.
This e¤ect, to which they applied the old term for dizziness, vertigo, was what
caused pilots to kill themselves. If a plane started into a turn, a pilot could sense
it accurately. If he then applied the proper control corrections to stop the nascent
turn, he would sense that the plane was now turning in the opposite direction
when it was not actually doing so. Correcting for that apparent motion led to fur-
ther inaccurate sensations of motion and further inappropriate control move-
ments. After only a few minutes, the di¤erence between what a pilot thought the
plane was doing and what it was actually doing became so great that the pilot lost
control. In reality, as the gap between apparent and actual motion grew, pilots
began to overcontrol, eventually making too tight a turn and stalling the plane,
which in turn caused a spin.
Between 1927 and 1932, when Ocker and army pilot Carl Crane published
their full study in a highly influential book, they tested a large number of mail,
airline, and army pilots to find that only 3 percent could fly blind for more than
twenty minutes without entering a spin.15 Even when equipped with the turn in-
dicator, pilots would revert to believing their sense of balance in flight and mis-
trust the instruments that claimed something counter to their own senses. Ocker,
Meyers, and Crane therefore promoted a method of blind flying that was based
on complete dependence upon the instruments.
The Ocker-Meyers method of blind flight was based on the idea that pilots
should create in their minds a virtual artificial horizon, based on the information
presented by their instruments.16 As no mechanical artificial horizon existed in
the United States before 1929, they believed that a virtual horizon was necessary
to produce safe blind flight. Constructing such a horizon mentally based upon
the turn-and-bank indicator, the climb indicator, and the compass took great con-
centration. The magnitude of concentration required to create and maintain that
mental image of the horizon induced fatigue fairly quickly in most pilots, and
only a very few (like Charles Lindbergh, an early convert to the method) could
maintain it for hours at a time. This diªculty helped generate a great deal of con-
troversy over their method in the aeronautical journals, and Ocker and his sup-
porters used their Jones-Barany chair to prove their case.
Ocker built and patented what was literally a black box positioned over the sub-
ject’s eyes while seated in the Jones-Barany chair.17 The box contained the stan-
dard combined turn-and-bank indicator with an air pump to keep the gyro spin-
ning and a magnetic compass. For the first test run with a subject, the box’s lights
were turned o¤ so that the subject was literally blind, and the subject was spun
in the chair to demonstrate the vertigo e¤ect that prevented him from accurately
sensing motion. The second test run was done with the box’s lights on and the
turn indicator’s gyro spun up, to demonstrate that the turn indicator often op-
posed the subject’s own sense of motion. The turn indicator was always correct—
the demonstration depended upon pilots recognizing that a gyroscope could not
be wrong.
Acceptance of the Ocker-Meyers method of blind flight did not come quickly.
Their earliest articles in 1926 produced controversy in the flying community be-
cause most (nonmail) pilots believed completely in their ability to fly blind with-
out the luxury of turn-and-bank indicators or the o¤ensive idea that pilots should
focus completely on instruments. Most pilots flew for the joy of seeing the earth
from the air and had no desire to place their attention inside the cockpit. Many
of these amateur pilots had encountered blind conditions briefly and survived
them, after all, a reality that Ocker admitted readily. The large majority of pilots
placed themselves in a class that Ocker and his advocates called “natural pilots,”
who flew by sense. They did not wish to become, and did not believe it necessary
to become, “mechanical pilots” who flew by memorized rules built around in-
struments.18
22 blind landings
Even pilots who had accepted the Ocker-Meyers method did not particularly
care for it. The sheer diªculty in constructing and maintaining a virtual artificial
horizon caused professional pilots to return to the idea of developing a real in-
strument to do this for them. Ocker and Crane promoted a “flight integrator” that
was supposed to solve this problem, but they never described it in much detail.
In any case, a mechanical artificial horizon was not a new idea in the late 1920s.
In 1916, the U.S. Navy’s Oªce of Naval Aeronautics had asked Sperry for a gyro-
scopic compass that established directional and horizontal planes of reference.
Elmer Sperry had responded that they needed two instruments to accomplish
this, a gyrocompass and a gyroscopic artificial horizon.19 Sperry failed to produce
either device during the war, and the navy dropped the contract at war’s end.
The appearance of a German artificial horizon instrument called the Gyrorec-
tor in 1926 helped stimulate new U.S. interest. The army obtained one and tested
it at Wright Field for several days before declaring its inadequacy for army use.20
Its chief problems, according to the army’s report, were that it was too slow to re-
spond and much too heavy (over thirty pounds). Weight in 1920s aircraft was a
major issue. Adding a thirty-pound instrument to the mail planes of the day re-
duced the amount of carriable mail by 10 percent, and the mail planes were sim-
ply modified army planes. The army wanted something a good deal lighter. The
slow response of the gyrorector was also a substantial operational drawback, as
pilots ideally wanted an instrument that responded instantly. Ultimately, this was
the easier problem to fix. All instruments have to be damped in some fashion to
prevent unwanted oscillations, a reality known since the nineteenth century. Too
much damping and the instrument responds slowly; too little and it will appear
wild to the user. Striking the proper balance between speed of response and sta-
bility is less a technical problem than a cultural one. The instrument designer has
to understand where the user’s preference lies and build that into the instrument.
That was a simpler issue than shaving twenty pounds o¤ the device without mak-
ing it too fragile to survive flying.
Fragility had doomed Elmer Sperry’s earlier work on the artificial horizon. In
his wartime foray into aeronautical instruments, he had discovered that the ac-
celeration forces endemic to flight were much greater than those encountered by
the seagoing gyroscopes in which he had previously specialized.21 Worse, be-
cause airplanes were extremely weight-limited, the instruments had to be much
lighter. That closed o¤ the obvious solution of making the instruments more
rugged by making them more massive. They had to be less massive and more
rugged, a combination that Sperry did not master during the war.
Sperry received a chance to redeem his failure in 1927. Early in the year, Daniel
Guggenheim established a fund, run by his son Harry, devoted to the improve-
ment of aeronautics. This Guggenheim Fund for the Promotion of Aeronautics
financed schools of aeronautics at the University of Michigan, the Massachusetts
Institute of Technology (MIT), the California Institute of Technology (CalTech),
Stanford, the University of Washington, and the Georgia School of Technology,
while undertaking a real-world project to perfect blind flying. The fund arranged
to borrow aircraft from the army, hired army reserve pilot James “Jimmy” Doolit-
tle, and set up an experimental “Full Flight Laboratory” at Mitchell Field, New
York, to carry out its blind flying research using radio equipment borrowed from
the National Bureau of Standards, Bell Labs, and Western Electric. The goal was
to produce a complete blind flight, including a blind landing. Doolittle ran the
blind flying e¤ort at Mitchell Field, assisted by MIT professor William G. Brown.
Doolittle’s recognition of the limitations of the turn-and-bank indicator and his
belief in the rightness of the Ocker-Meyers method led him back to Elmer Sperry.
Sperry decided what Doolittle needed was a horizon indicator and turned the task
over to his son Elmer Jr.22
The instrument Elmer Jr. devised used an air-spun gyro designed to maintain
a vertical orientation. The instrument face was designed to replicate the natural
horizon, with a fixed set of wings representing the aircraft and a “horizon” line
that appeared to move in relation to the wings. The upper half of the instrument
face was light blue, to indicate skyward, and the lower half was black to represent
the earth (Figure 1.2). For the Guggenheim blind flying tests, the new device was
mounted in a Consolidated NY2 borrowed from the army.23
Doolittle made his first complete blind flight on September 29, 1929, taking
o¤ from and landing at Mitchell Field. Using the horizon, a directional gyroscope,
a specially designed altimeter, and a set of borrowed radio equipment, Doolittle
took o¤ under the hood like that in Figure 1.3 and flew a well-practiced track away
from and then back toward the field. He landed using an experimental blind land-
ing system devised by the National Bureau of Standards (discussed in Chapter
3).24 For his e¤ort toward making blind flying possible, Doolittle received the
1929 Collier Trophy.
The Sperry artificial horizon was rapidly adopted by the airlines, whose pilots
found that it made keeping the plane straight and level much easier. This was be-
cause the artificial horizon sensed the absolute amount of deviation of the wings
from true horizontal, while the turn indicator displayed the plane’s rate of turn.
A slight bank showed up on the artificial horizon instantly, while the same bank
might not translate into a turn rate substantial enough to show on the turn indi-
cator for several minutes. Although many promoters of the artificial horizon ad-
Figure 1.2. Sperry artificial horizon. Reprinted with the permission of Unisys
Corporation, courtesy National Air and Space Museum, Smithsonian Institution
(SI 99-41447).
Figure 1.3. Boeing School of Aeronautics blind flight trainer. Note the leather hood over
the front cockpit. Blind flight meant blind flight. National Air and Space Museum,
vocated it as a replacement for the turn-and-bank indicator, most airlines put both
on their aircraft once the horizon became available. Pilots found that while the
horizon made maintaining level flight easier, the turn-and-bank indicator pro-
moted simpler recovery when the aircraft departed from level flight.
Simply telling pilots how to use the instruments and sending them up did not
solve the problem. Pilots tended to believe that the turn indicator and artificial
horizon worked fine as long as they could see, but once in the clouds, they thought
the instruments went haywire and reported turns that the pilots were certain the
airplane was not making.25 The turns did happen, of course; that was the point.
Pilots had to be broken of their dependence on their sense of balance.
The first step toward freeing pilots from this dangerous dependence upon
themselves came from the army, which began screening its new volunteers for
flying duty for instrument compatibility. Once a potential recruit had been taught
the basics of flying, he was given two hours of classroom instruction on how the
instruments worked and why they were necessary. An instructor then took him
aloft in a training aircraft with a hood fitted over the student’s cockpit, and once
high enough for safety turned the plane over to the student. If the student tried
to maintain control via the instruments, he was acceptable. If he did not, he was
given two more chances after additional classroom instruction. If on the third
training hop he continued to rely on his own sense of balance, he was washed out
of the flight program. Henceforth, willingness to rely completely on instrumen-
tation was a fundamental requirement for army fliers.
The airlines went further in checking out all of their existing pilots, motivated
in part by a Department of Commerce ruling that all airline pilots had to pass
a blind flying test by January 1933.26 Eastern Air Transport, for example, sent
Howard Stark around their system to retrain all of the company’s pilots. By 1930,
Stark was a recognized expert in blind flying because he had improved upon the
Ocker and Meyers method. While they had advocated that pilots learn to fly by in-
struments alone, Stark had devised an explicit methodology for using the instru-
ments in blind flying. Called the 1-2-3 method by Stark (and the A-B-C method by
the army), it was rapidly adopted in both army and airline flying.
Stark had begun his flying career as a barnstormer and joined one of the new
airlines in 1927. He quickly learned that he did not really know how to fly; when
he asked his fellow pilots how to use the instruments he was told to “keep them
centered.” Keeping them centered proved not to be easy. After entering a spin at
four thousand feet and plunging to less than seven hundred feet before recover-
ing, he realized that “he really could not keep control of the plane if [he lost his]
26 blind landings
balance, and this might happen to anyone in rough air such as thunderstorms or
line squalls.”27
After his personal epiphany, Stark developed his 1-2-3 method during 1928 and
1929. The steps of the method were: (1) center the turn indicator with rudder only;
(2) center bank indicator with ailerons only; and (3) center climb indicator with
elevators only. The order in which pilots were to take these steps mattered, too,
he explained in his 1931 textbook. If the turn indicator was not centered first, the
bank indicator would “give a false index for level flight.”28 The method relied
upon the realization that in a properly banked turn, which by definition is a turn
intended to exactly mimic the acceleration of gravity, the bank indicator would in-
dicate no bank at all and the pilot would not feel the turn through the inner ear’s
balance function, because neither of these sensors can discriminate between
gravitational acceleration and centrifugal acceleration. By using the rudder to cen-
ter the turn indicator first, a pilot would automatically create an improperly
banked turn that would permit the bank indicator to deflect. Only then could the
pilot level the aircraft’s wings using the bank indicator as a reference. In essence,
the bank indicator only accurately represented the position of the wings when the
plane was not turning. This is why the order of the instruments and the steps re-
lated to them mattered. The turn had to be stopped before the bank indicator
could be relied upon.
The turn-and-bank indicator had assumed a new function in Stark’s method.
While it had been designed to permit pilots to make accurate turns, supple-
menting the unreliable magnetic compass, in this method it became an instru-
ment used to prevent the plane from making unwanted turns. Originally, the turn
indicator was used only occasionally by pilots when they wanted to turn, but now
pilots had to pay near-constant attention to it. Stark was not the first to perceive
this use; Lawrence Sperry seems to have understood this use of the instrument,
as did the old hands from whom Stark had asked advice when he took the mail
job. Realization that the turn indicator could be used this way was probably wide-
spread among professional pilots by the mid-1920s, but their success at using it
varied greatly. It seems likely that many individuals had worked out their own way
to use the turn-and-bank indicator, and Stark may well have codified something
that was already common knowledge among professional fliers. By a year after
his book’s publication, however, his method was explicitly adopted at flying
schools, by airlines, and by the army.29
Stark’s method also separated the aircraft’s controls in pilots’ minds. That sep-
aration was a false one, and a later, more complicated method was proposed to
restore a more realistic perception of control actions. By telling pilots to use the
rudder to manage the turn indicator and the ailerons to manage the bank indi-
cator, Stark was implicitly arguing that the rudder was for turning and the
ailerons for banking, but actually the rudder causes banking in addition to turn-
ing, while the ailerons cause turning in addition to banking. In fact, only raw be-
ginners use the rudder to make a turn; the ailerons are far more e¤ective. By the
late 1930s, designers were making airplanes without rudders, although the trend
did not catch on among pilots, in part because of their training and in part be-
cause the rudder was important in emergencies like spins and loss of an engine
in multiengine aircraft. Stark’s rules removed the confusion that two controls
with similar e¤ects could create in pilots’ minds. They were, in e¤ect, a heuristic
approach designed to produce consistent results, even if the process of applying
it was inelegant and suboptimal.
Because of the need for pilots to continuously refer to the turn indicator to pre-
vent turns, much of Stark’s short book was devoted to the subject of proper in-
strument placement. For most e¤ective use of his method, he believed, the “turn
indicator group” of instruments should be placed in the following order: airspeed
indicator, turn indicator (which included the bank indicator), climb indicator, and
altimeter. Pilots could scan this set of instruments left to right, centering each one
in its proper turn.
The introduction of the Sperry artificial horizon two years later led him to re-
vise his recommendations only slightly. While pilots were to use the turn indica-
tor group to recover straight and level flight after a disturbance caused the plane
to turn, a new “Sperry group” that included the horizon and the directional gyro
or gyrocompass was to be used to maintain straight and level flight. The Sperry
group was to be in a specific layout, too. If only a Sperry group was to be installed,
Stark advocated placing the horizon atop an inverted T, with the gyrocompass di-
rectly below it, and airspeed indicator and altimeter to the gyrocompass’s left and
right, respectively. The horizon’s position atop the layout (where it would be di-
rectly under the magnetic compass, which was always placed above the instru-
ment panel to limit electromagnetic interference), reflected its new status as the
most used instrument.
Just as Ocker and Meyers had found advocates willing to promote their view
of the incapacity of humans to fly blind unassisted by instruments, Stark found
willing champions of his rule-based method. Ernest A. Cutrell, who had flown for
the Department of Commerce in the early 1930s and was hired away to be an in-
strument instructor for American Airlines, advocated Stark’s method to Avia-
tion’s readership. At New York University, E. B. Schaefer from the Guggenheim
School of Aeronautics wrote a detailed exegesis of it for Aviation Engineering.
28 blind landings
Ocker himself gave the biggest impetus to Stark’s method by republishing it (un-
der a new name) in his co-authored 1932 book, which became the most influen-
tial text on instrument flying before World War II. By 1938, a former mail pilot
and instrument instructor for American Airlines could claim that “instrument
flying is based on cold logic; there is nothing of the ‘born pilot’ about it” before
explaining how to use Stark’s 1-2-3 method.30 The mechanistic method of flying
that Stark had derived during his airmail experience had come to dominate com-
mercial and military aviation.
Stark’s new kind of flying was diªcult enough that the airlines had to set up
training programs to teach it. Pan American, for example, set up a school in
Brownsville, Texas. The school employed the same equipment that Ocker and
Meyers had used to demonstrate the disorientation e¤ect on the ground, the
Jones-Barany chair and Ocker’s black box, to familiarize its pilots with the prob-
lem. Pan Am also equipped an airplane (a Fairchild FC-2) with a second, blacked-
out cockpit behind the original pilot’s position and used that for initial in-flight
training. Once student pilots could fly blind in this smaller aircraft, they were ad-
vanced to training in larger multiengine aircraft. Pan Am required a minimum
of ten hours of instruction. It also required what would eventually be called re-
current training, making pilots redemonstrate their abilities at regular intervals.31
Like United Air Lines, which instituted a similar program, Pan Am required an
hour of practice monthly after students graduated from its course. The monthly
practice was necessary to keep pilots who did not routinely encounter blind con-
ditions from backsliding after the training course and returning to dependence
on their sense of balance.
Stark’s rule-based flying led directly to a radical idea, perhaps best stated by
American Airlines instructor Karl Day in a 1938 textbook, that pilots should be
taught to fly on instruments first, preferably while sitting on the ground.32 Con-
tact flying taught habits that might be acceptable for amateur pilots, but profes-
sional pilots needed to learn to reject the misleading siren calls of their motion
senses and accept the truth as revealed by the instruments. He believed that in-
stead of permitting pilots to gain hundreds of hours of experience in contact
flying before being trained on instruments, they should be taught instrument
flight before learning contact flying or, second best, immediately after they suc-
cessfully completed basic flight training. Day took the teachings of Ocker, Mey-
ers, Crane, and Stark to their logical conclusion. If instrument flying was to be
based on rules, those rules should be ingrained in pilots as habits, and training
programs should be arranged so that proper habits were formed early.
Training nonpilots to fly on instruments first, however, was an idea whose time
never came, in part because of the danger to instructors imposed by a completely

untrained student at controls enclosed in a small, dark space in which he might
panic. The idea of training new graduates from flight school on instrument flying
was adopted by the Army Air Corps, which appended instrument flight training
to its basic flight school curriculum. It did not catch on among private fliers, how-
ever, because instrument training was very expensive compared to basic flight
training. The ten hours of instrument training that the airlines and the army con-
sidered necessary to produce a proficient instrument pilot was more than the to-
tal amount of flying time required for many students to receive their private pi-
lots’ license during the 1930s. Physicist Luis Alvarez, for example, received three
hours of flight training before soloing.33 The Bureau of Air Commerce, which
was founded in 1928 to regulate civil aviation, was not about to require a sudden
doubling or tripling of the training requirements for all pilots. The organization
rightly perceived that doing so would be political suicide. Only commercial and
military pilots saw their training requirements increase during the 1930s to re-
flect the increased importance of “mechanical flying.”
The cost of that increased training requirement led indirectly to the fulfillment
of Day’s claim that instrument training was best done on the ground. Day was
not the first to believe so, and by the time he wrote his textbook, ground-based
training was old news, if not yet famous. The idea of ground-based flight train-
ing was around even before the Great War, embodied in, for example, the 1910
“Sanders Teacher.”34 But ground-based flight training did not catch on until the
ideology of instrument flight was accepted in the army and the airlines. It made
little sense to adopt ground training for pilots when most pilots flew for the joy
of the act of flying. Most people in the 1920s learned to fly from itinerant barn-
stormers after a handful of hours of instruction, and therefore organized flight
schools were rare before the early 1930s, when the Bureau of Air Commerce be-
gan to formalize, if not tighten, licensing requirements. There was thus no mar-
ket for ground trainers, regardless of their functionality. There were, nonetheless,
a number of such systems invented in the early years of aviation, but only one
happened to coincide with the rise of instrument flying.
The lucky inventor was Edwin A. Link Jr., the son of a manufacturer of player
pianos. Born in Huntington, Indiana, in 1904, Link fell in love with the airplane
during his youth. He took his first flight in 1920 with Charlie Chaplin’s brother
Sidney, and abandoned prep school in favor of a more hands-on education, which
he got at his father’s factory, then in Binghamton, New York. He devoted his spare
time to tinkering with electronics and learning to fly, and he finally soloed in
1926. He purchased one of the new Cessna Aircraft Company’s first planes the
30 blind landings
Figure 1.4. Ed Link posing with his “blue box,” an early blind flight trainer. National Air
and Space Museum, Smithsonian Institution (SI 99-4151).
next year. But like many previous inventors, Link thought that flight training was
too expensive. If the first few hours of “flight time” could be done on the ground,
he thought training could be made less expensive and safer.35
He was encouraged in this belief when he discovered that the French had
trained their pilots during the Great War by teaching them how to control the
plane while taxiing around an airfield. This “penguin” system used planes with
worn-out engines and clipped wings that would not allow takeo¤. It also seemed
to reflect his own experience, and he began to construct a trainer that would feel
like an aircraft bouncing around an airfield. A year and a half of Link’s spare time
went into constructing his device which, when completed in 1929, included no
blind flying instruments (Figure 1.4). It was merely a device intended to help pro-
spective pilots get the feel of an aircraft’s controls.36
As Link’s biographer reports, almost no one in aviation wanted his trainer
when it was introduced. Pioneer Instruments bought one to demonstrate its in-
struments to potential customers on the ground, and the U.S. Navy bought one
to try out at its flight training school at Pensacola, Florida. Amusement parks and
fairs showed far more enthusiasm. Link sold fifty trainers to the amusements in-
dustry before 1933.37 It took a disaster to introduce the professional flying com-
munity to Link’s device, and that disaster descended from the airmail.
In February 1925, the U.S. Congress had passed the Air Mail Act, which di-
rected the Post Oªce to let contracts to private companies to carry the airmail.
There were no airlines at the time; instead, during the year groups of investors
proposed airlines on paper in order to participate in the bidding. The postmaster
general awarded contracts that October to five new airlines—Colonial Air Trans-
port, Robertson Aircraft Corporation, National Air Transport, Western Air Ex-
press, and Varney Air Lines—to serve routes that the Post Oªce’s mail service
did not cover. It continued to fly the major trunk route (New York to San Fran-
cisco via Chicago) with its own aircraft. Once the new airlines had proven their
ability to maintain scheduled service, the Post Oªce opened bidding for this final
route in 1927, which it divided into two contracts. One contract covered the New
York to Chicago segment, which was awarded to National Air Transport; the other,
which went to Boeing Air Transport, covered the long Chicago–San Francisco
run. The Post Oªce also added new contract routes that year, mostly to new start-
ups.38
The result of the bidding process, however, was a crazy quilt of routes. For a
piece of mail to get from Boston to Seattle in 1928, for example, it flew from
Boston to New York on Colonial Air Transport, then had to be put on a National
Air Transport plane to fly to Chicago. Once in Chicago, the letter was transferred
to a Boeing Air Transport Plane, was flown to San Francisco, and then went on a
Pacific Air Transport plane up to Seattle. This made no business sense at all. Pres-
ident Hoover’s postmaster general, Walter Folger Brown, decided to rationalize
the mail service through a set of “shotgun marriages.” Out of Brown’s weddings
came the airlines that dominated American commercial aviation through dereg-
ulation in the late 1970s: American Air Line, Eastern Air Transport, Northwest
Air Lines, Transcontinental and Western Air (TWA), and United Air Line. Much
of this manipulation occurred at a single conference that began May 19, 1930, in
Washington, known ever after as the “Spoils Conference.”39 The conference pro-
duced a relatively rational, eªcient route system based on a few large carriers,
while e¤ectively locking smaller airlines, especially new startups, out of the lu-
crative mail business.
Franklin Delano Roosevelt held rather di¤erent ideas about what constituted
competition, however, and after his election he chose to make an example of what
critics had begun to call the “airline trusts.” Roosevelt was driven to act against
the airlines in February 1934 by an investigation started by a political opponent,
32 blind landings
Senator Hugo Black, who had received a distorted version of Brown’s work from
a reporter. The reporter had been approached by several airline hopefuls foiled by
Brown’s focus on constructing a few large carriers. He had prepared an exposé,
only to find it suppressed by his editors. Black’s investigation made the spoils con-
ference public, prompting Roosevelt to act precipitously. On February 9, Roo-
sevelt asked Maj. Gen. Benjamin Foulois, chief of the Army Air Corps, if his or-
ganization was prepared to fly the mail. Foulois had answered that it was, and
Roosevelt directed Postmaster General James Farley to cancel all the airmail con-
tracts.40
The result was a disaster but proved to be Ed Link’s opportunity. The airlines
(except United) received the majority of their revenue from the mail contracts,
and with the exception of United and American, they shut down immediately in
hopes of fending o¤ bankruptcy until the political situation could be repaired. All
the airlines launched an immediate media o¤ensive against Roosevelt, and they
were aided immeasurably, although certainly not deliberately, by the Army Air
Corps.
The Air Corps failed Roosevelt catastrophically. The Chief of the Air Corps had
assured Roosevelt that they “had a great deal of experience in flying at night, and
in flying in fogs and bad weather, in blind flying, and in flying under all other con-
ditions.” Yet over the thirty days following the February 19 cancellation, ten army
fliers died in the Air Corps attempt to imitate the airlines. The newspapers turned
their deaths into a vicious political attack on Roosevelt, and the incident became
a major political crisis. On March 10, the Air Corps called a safety halt and re-
sumed flying the mail, only in fair weather and by daylight, on the nineteenth.
The airlines got the mail back in mid-May, after being forced to change their
names slightly (United Air Line became United Air Lines) in order to bid for new
contracts. The rebidding did result in contracts going to a few new competitors,
primarily Brani¤ and Delta, so the cancellation had not been an entirely futile
act.41
The Air Corps failure at the airmail came not from a sudden increase in deaths;
in fact, its attempt to fly the mail produced the same accident rate its normal op-
erations in previous years had. But that rate was far worse than the airlines’, and
many of the accidents came specifically from failing at blind flying. The Air Corps
used the same instruments the airlines did and had the same procedures de-
scribed by Ocker and Crane’s text. What its pilots lacked was proficiency. In its
normal operations, the corps had no reason to fly blind. As numerous historians
of air power have noted, Air Corps war-fighting doctrine was predicated on des-
ertlike weather, clear and dry. One could not bomb a target e¤ectively if one could
not see it, and a di¤erent version of the same incontrovertible rule held for
fighters (then called pursuits). Air combat required being able to see the enemy at
a distance of miles until the development of radar, still years away in the United
States. The corps had no reason to practice blind flying routinely, and it did not.
Flight time was expensive, after all, and the corps was the poorest of the military
branches by far.
Although Foulois had assured Roosevelt of the corps’s competence, either he
or some of his subordinates had been considerably less certain of the Air Corps’s
capabilities. One of Link’s marketing partners was invited to demonstrate Link’s
trainer at the Casey Jones flying school in Newark on the eleventh, a scant two
days after the cancellation announcement but several days before the corps
started mail service. The Air Corps oªcers left impressed with the trainer’s po-
tential to provide inexpensive proficiency training, but they had no money to buy
it. Link’s partners canvassed Congress to rouse political support for an emergency
appropriation to enable an Air Corps purchase of ten and, aided by the Corps’ fail-
ure in the following days, succeeded. Roosevelt signed the appropriation bill in
late March, and the first six trainers were delivered in June.42 The Air Corps pub-
lished new training procedures built around the Link trainer the following year.
The Air Corps’s purchase served as validation of Link’s trainer for other avia-
tion groups. Japan was the first to buy (another ten), followed by the Soviet
Union’s ubiquitous Amtorg Trading Company. Orders from the United King-
dom, France, and various airlines followed, and Link was soon able to begin re-
selling to the same customers by introducing new models, often tailored to rep-
resent specific aircraft. Thirty-five countries used Link’s trainers to train their
pilots and to maintain their instrument proficiency by the eve of World War II.
conclusion
By 1940, professional pilots had accepted the once-heretical notion that blind
flying required complete faith in their instruments. Their organizations had in-
stitutionalized training programs to induce and maintain that faith, based around
a machine whose own foundational concept—that pilots should be taught to fly
on the ground—was not popular in the community’s early years. Finally, profes-
sional pilots had adapted to a new kind of flying that replaced intuition and feel
with instruments and rigid rules about how to use them.
By the late 1930s, therefore, blind flight was possible for pilots possessing the
proper training and skills, although it remained dangerous for those without
them—or for those who were simply out of practice. Success, however, had re-
34 blind landings
quired instilling in pilots an extreme form of technological faith that required

them to reject the sensory motion data upon which they had previously relied.
Maintaining that faith meant ongoing retraining in mechanical flying.
Because blind flying depended upon pilots’ faith in their instruments, achieve-
ment of blind landing necessitated development of a highly reliable system. A
blind landing system would have to work in all weather (to support training in
good weather as well as bad). It also would have to earn pilots’ trust. And it would
have to function within the existing aviation infrastructure.
chapter two
Places to Land Blind
While the goal of all-weather operations forced pilots to learn to fly blind, it

also caused promoters of air commerce to seek landing fields capable of sup-
porting all-weather flying. Airfields during the Great War had simply been grass
fields, usually circular, and aircraft took o¤ and landed in whichever direction
happened to be into the wind. But during the 1920s, airfield operators wishing
to attract the mail business began laying surfaced runways, usually in the two or
three prevailing wind directions at their individual sites. Why they did this is a
matter of some dispute in aviation history, and we must examine the shift to run-
ways in the context of the problems they created for blind landing technologies.
The transition from grass field to paved runways is usually linked anecdotally
to the introduction in the early 1930s of heavy, all-metal airliners like the Boeing
247 and the Douglas DC-3. The chief diªculty with this argument is that airports
began adopting surfaced runways well before those aircraft were developed. A
more recent interpretation places the responsibility for the adoption of surfaced
runways on architects, whose professional desires drove them to promote the per-
manent construction of geographic forms in the ground.1 Unfortunately, this ex-
planation reflects the popular sin of social determinism—architects’ professional
desires no doubt a¤ected runway layouts, but they were not the reason airfield
managers and airlines wanted grass replaced in the first place. This argument
cannot explain why airport managers wrote many letters to journals of the day ex-
plicating detailed physical problems at their airports. Runways emerged as a so-
lution to specific problems that various airports had in meeting demands for reg-
ular service. Surfaced runways were invented as a solution to particular local
36 blind landings
environmental problems that prevented the achievement of regularity.2 They are

thus part of the same drive toward “all-weather operations” that encompassed
blind flying and landing e¤orts.
The deployment of surfaced runways, however, led to a substantial increase in
the diªculty of designing an acceptable blind landing system. Making a blind
landing system capable of guiding a plane into a large, open grass field was one
thing, while designing one to guide it onto a narrow strip of pavement proved to
be quite another. Runways required greater precision than did the grass fields of
the Great War. Hence, an early blind landing system developed to serve grass
fields proved unacceptable to pilots and airport operators.
why runways?
The earliest airfields in the United States were unprepared in any way save
clearing of trees and shrubs. The U.S. Army used parade grounds as takeo¤ and
landing areas with its first aircraft but quickly found that marching soldiers and
landing airplanes were not a good mix. The First Aero Squadron’s initial opera-
tional deployment, during the U.S. campaign to track down Francisco “Pancho”
Villa, cast the problem of landing fields into stark relief. As Benjamin Foulois re-
ported in his autobiography, the soldiers sent to find landing fields for his
squadron often chose poorly. Although naturally flat fields of suªcient size
abounded in the campaign area, they were not smooth enough for safe opera-
tions. Rocks, holes, and even large clumps of grass caused crackups.3 Although
few serious injuries resulted, the damage to aircraft was expensive, in terms of
repair costs and in reduced availability of the aircraft. The unprepared field was
therefore obsolescent by the time of U.S. entry into World War I, used only for
emergency landings. Proper airfields had to be leveled and graded.
The Army Air Service’s Specifications for Municipal Landing Fields, published
in 1919, was the earliest basis for airfield design in the United States.4 The army
identified the square as the best shape for an airfield based on its experience
during the Great War.5 A first class field, in the army’s opinion, was a square
1,800 feet by 1,800 feet, with a central 150-foot concrete “runway” in the shape
of a cross. The runway was for takeo¤, not landing; indeed few army fields were
built with this feature. Instead, most army airfields remained what were called
“all-over” fields (with no specific runways) through the mid-1930s. They were
groomed and drained turf fields that allowed aircraft to land in any direction.
During the early 1920s, pilots and airport managers considered the use of turf
places to land blind 37
as an airport surfacing material beneficial. Because aircraft did not have brakes,
grass provided friction that reduced stopping distance. Airplanes had tail skids to
slow them, which made the self-repairing feature of good turf vital. Grass fields
also kept down dust. Aviators believed that blowing dust and dirt would keep
the public away from airfields, making aviation commercially unappealing. Dust
also entered aircraft engines through carburetors, causing rapid wearing of cylin-
der surfaces and bearings, reducing aircraft reliability, safety, and profitability. Fi-
nally, pilots also believed that turf was softer than hard surfaces and produced less
stress on their aircraft’s landing gear, another seeming advantage of the prepared
grass field.
There was, therefore, an entire industry devoted to the maintenance of grass
airfields. One can find numerous articles and advertisements in the aviation mag-
azines of the 1920s and 1930s extolling these virtues and instructing airport own-
ers and managers in the proper methods of draining, grading, disking, seeding,
and rolling grass fields. The image of the eªcacy of turf as a field covering was
reinforced with references to the great “model” aerodromes of Europe, Croydon
in London and Templehof serving Berlin.6 Both retained grass landing and take-
o¤ areas (and, in a departure from U.S. practice, both were roughly circular) but
had surfaced taxi and parking areas.
Yet the all-over field was obsolete in American aviation by the end of the 1920s.
Although military airfields and those airfields not on the major commercial
routes remained turf all-over fields into the late 1930s, airfields serving major
U.S. cities adopted various forms of surfaced runway in the late 1920s. What
brought about this divergence of commercial and military practice?
Three intertwined factors influenced the departure of civil practice from mil-
itary. The first was aircraft design and, in particular, the tire pressures chosen for
aircraft use. The second was the nature of civil aviation as a business, particularly
in its demand for regularity. Finally, the geographic diversity of the nascent com-
mercial industry imposed soil and climatological conditions that, when combined
with the above factors, drove airport managers to adopt surfaced runways rapidly
during the late 1920s.
Aircraft designers have to consider a number of external factors in the mak-
ing of their creations. One of the least sexy, little regarded, yet very important such
factors is the surface from which the craft is expected to operate. The most obvi-
ous di¤erence results from a choice of water rather than land as the operating
surface. Planes designed for water used floats, boatlike hulls, or both, while land
planes had wheels (for ground) or skis (for snow and ice). The Wrights had ini-
38 blind landings
tially used skids, but builders abandoned them rapidly as they began to design
aircraft too heavy to be simply picked up and carried o¤ the field, as the Wright
Flyers were.
The apparent lack of options for landing on the ground (wheels or wheels)
might suggest that designers had little choice, but in fact there is wide diversity
in wheels, particularly in tires. Bicycles, for example, rely upon di¤erent tire pres-
sures for di¤erent surfaces: mountain bike tires are typically thirty-five pounds
per square inch (psi) for mud and soft dirt to sixty psi (for hard ground), while
common road bike tires run as high as 110 psi. The tire pressure is, to within a
few psi, the pressure a bicycle (or plane or car) exerts on the surface, and each sort
of surface can bear a di¤erent load. A 1944 textbook on airport design reported
that most soils could withstand between fifty and seventy psi once prepared by
rolling and when dry, but moisture reduced that bearing capacity to as little as ten
psi.7 Tires with higher pressures than a particular soil can support dig into it,
which is why road bikes tend to damage lawns (especially wet ones), and why dirt
bike rallies inevitably trash the soft soils they are held in.
This is also one of the factors in the failure of turf airfields in the United States.
A U.S. Army Air Service circular from 1922 lists the standard aircraft tire pres-
sures in use at the time as fifty to sixty psi, and its authors demonstrated that those
pressures translated into pressures of fifty-two to seventy-one psi exerted on the
ground.8 The 1944 text shows these pressures were acceptable, if marginal, for
smooth, dry turf fields, but moisture meant a day (or more) of no flying lest the
field be damaged.
There were two options available to engineers interested in solving this land-
ing problem. One was to use lower pressure tires on aircraft, as the bicycle ex-
ample suggests. Private aircraft owners often did just that, adopting low-pressure
“balloon” tires (ten to twenty-five psi, in many cases) because such tires allowed
the convenience of operating from relatively unprepared fields. That made the
airfields they used less expensive to build and maintain.
For larger aircraft, however, the low-pressure tire approach rapidly became in-
feasible. In order to support an aircraft of the same weight, a low-pressure tire
had to be much larger than a high-pressure one because the tire had to distribute
the same load (the aircraft’s weight) over a larger area. Increasing the plane’s
weight would result in even greater increases in size. Larger tires meant more
drag on the aircraft in flight, while again increasing the aircraft’s weight. Neither
of these e¤ects was desirable to commercial and military aircraft designers, for
di¤erent reasons. Commercial operators were initially paid by weight for carry-
ing the mail, and therefore minimizing aircraft weight in favor of greater mail ca-
pacity was of economically valuable to the airlines, as was reducing drag in favor
of fuel economy. Military aircraft have always been designed for high perfor-
mance, and increased drag and weight imposed performance penalties that were
unacceptable to military customers. Retractable landing gear, introduced to com-
mercial aircraft in the early 1930s, was one consequence of the desire to improve
fuel economy and performance by reducing drag.9 Once tires had to be stored in-
side the aircraft in flight, designers were under further pressure to reduce the size
of tires, which again meant increasing tire pressure.
Nevertheless, aircraft designers were able to keep aircraft tire pressures within
the fifty to seventy psi range through the mid-1930s (the Douglas DC-3 used sixty-
three psi tires) in order to maintain aircraft compatibility with existing fields. Af-
ter that, however, the massive increase in aircraft size made this impossible. The
1935 Boeing B-17, for example, weighed twice as much as the DC-3, so its tires
would have been twice as wide as the sixteen-inch tires on the DC-3 at the same
pressure. Storing a thirty-two inch wide tire was a great waste of space, and tire
pressures therefore reached one hundred psi by the onset of World War II. The
escalation stopped there only when designers adopted multitired landing gear.10
The commercial and military demand for higher performance aircraft of ever
greater weight meant that the low-pressure tire solution to the landing field
problem was no solution at all. At best, it was a stopgap measure until the other
possible solution could be implemented. That measure was the adoption of arti-
ficially surfaced runways. Artificial surfaces promised to overcome the weather
sensitivity of turf fields just as they had helped overcome the same problems for
automobiles and bicycles. The need to overcome weather sensitivity, in turn, de-
scended directly from the commercial aspects of aviation.
Airports are designed to deal primarily with aircraft traveling from place to
place in the pursuit of commerce, just as the primary function of seaports is to
deal with commercial ships hauling goods from place to place. City planners im-
mediately grasped the commercial consequences of airports and deliberately
chose sites based upon relative proximity to surface transportation. Since people
and goods carried by aircraft had to get from the airport to their actual destina-
tion, architects and city planners focused on airports as eªcient nodes of trans-
fer.11 Selection of airports as eªcient and safe landing sites was distinctly sec-
ondary, as was the consideration of site suitability for proper turf growth.
The primacy of eªcient location of the airport with respect to surface trans-
portation in urban areas was enhanced by a second commercial influence on avi-
ation. That is due to the simple reality that people (especially business people)
only want to go where other people are. The United States was an urban nation
40 blind landings
already by 1930, and its population was rapidly becoming more centralized. In
addition, the largest cities were within the snow-belt states. With those demo-
graphic and geographic realities, the Post Oªce established the transcontinental
mail route across the northern states in the 1920s: Washington, New York City,
Cleveland, Chicago, Salt Lake City, and San Francisco were the major stops. Of
these, only Washington and San Francisco were relatively free of snow, and the
drained swamp upon which Washington sat proved to be at least as troublesome
for airport building. Rainfall along most of the route was substantial, and the soil
tended to retain water, which was good for growing but bad for building load-
bearing structures. The demographics of the United States demanded, in sum,
that commercial operators, building upon the Post Oªce mail routes, operate in
precisely those places that Benjamin Foulois had argued should be avoided by the
army in establishing its airfields: the cold, wet, foggy, snowy North.12
Finally, commerce dictated regularity. Although the greater speed of aircraft
made air travel appealing to the first brave passengers (almost always business-
people), more important for the issue of runway development was business trav-
elers’ demand for scheduled service. Airplanes are not faster than trains if they
have substantial weather-induced delays, such as an airfield turned to mud by a
sudden storm or the spring thaw. Diversion of incoming planes to other fields
from similar causes was equally unwelcome to business travelers. This, too, was
reinforced by the Post Oªce, which insisted its airmail contractors deliver the
mail daily, regardless of weather or airport conditions. Private, and indeed mili-
tary, fliers could and did wait for airfields to dry out enough after a storm to re-
sume flying without damaging the turf surfaces, a practice common even during
World War I.13 The mail could not wait. Airfields whose owners wished them to
be airports had to ensure continuous availability of their fields to attract a mail
contract, and therefore an airline.
These three demands of commerce conspired to place major commercial air-
ports in what were relatively poor sites in terms of their suitability as landing
fields. The meteorological conditions of these areas, combined with soil condi-
tions at specific sites, brought about the rapid demise of grass as a commercial
airport surface technology in the United States.
As early as 1923, the problem of maintaining turf fields for aviation was rec-
ognized by construction engineer Archibald Black, who stated that the need for
special runways had grown out of the breakdown of turf under regular use. This
breakdown was accelerated by the “softening” of the turf that occurred during
“certain periods of the year.” Softening, caused by higher soil moisture content,
meant that aircraft tires did more damage. This was especially true where the soil
froze over the winter, which prevented drainage of surface water for several weeks
during the spring. Freezing areas with soils of high organic content were also sub-
ject to “lensing,” which created zones of weakness in the soil.14 Black noted that
there was a trend toward specially prepared runways, and he predicted that the
tendency would increase. Black did not refer to runways in our modern concep-
tion, however.
Black specifically identified several possible runway surfaces that, while not
hard in the sense that concrete is hard, were certainly considerably harder than
turf. Based upon standard road construction practices, Black reviewed the possi-
bilities of using gravel, crushed stone, cinder, and asphalted, or oiled, runways.
Black’s articles were reinforced by a 1929 textbook he published on airport engi-
neering.15 In it he focused on the similarities between the needs of automobiles
for a good road surface and the needs of aircraft for take-o¤ and landing surfaces,
finding that the two machines had essentially the same requirements. He pre-
scribed working crushed stone or gravel into the ground through repeated wa-
tering and rolling, and crowning this with oiled cinders, sand, or smaller gravel,
depending on the local availability of these materials. These recommendations
were based upon existing road construction practice, as he readily admitted.
Black’s review of these surfacing methods suggests that he expected improved
drainage to be insuªcient to keep turf fields in reliable service at many sites. His
analysis was supported in many other articles published during the later 1920s.
Quincy Campbell, an Air Service reserve pilot, stated that sod was initially the best
choice for fields, but increased traªc had made them unmaintainable. Tail skids
cut up the sod in “soft” weather, and ruts developed from wheels. Frequent re-
rolling to eliminate ruts further damaged the sod, resulting in dust and mud.16
Campbell did not discuss the cost involved in turf maintenance, but it was
clearly substantial. Removing fields from use for rerolling and reseeding meant
less revenue for airport operators as well as increased direct costs. The cost to air-
craft owners of poorly maintained fields is diªcult to assess quantitatively, as no
statistics breaking out surface-caused accidents from other types exist. Qualita-
tive examination is easy, however. Airports contained an “Airports in Pictures” sec-
tion often devoted to photographs of aircraft wrecked by poor field conditions.
Although these accidents rarely appear to have been fatal, they represented an
unacceptable financial risk to the infant airlines created in the wake of the Post
Oªce’s 1925 decision to privatize the airmail. Maintenance of turf was clearly a
substantial burden that at least some airport operators were either unwilling, or
unable, to shoulder.
The airlines, the Post Oªce, and the Department of Commerce drafted an air-
42 blind landings
port rating standard in 1926 that substantially influenced a wave of airport build-
ing that began after privatization of the airmail. The standard was not mandatory,
as the newly formed Aeronautics Branch of the Commerce Department had no
legal authority over airports. Nonetheless, the rating standards provided a frame-
work for comparison and certification of airports.17
Cities often went well beyond the standard and certainly did so with runways.
The standard only required year-round serviceability and a minimum width and
length. It did not require surfacing of any kind, stipulating vaguely that fields be
smoothed to the point that a passenger car could drive over it at thirty mph with-
out “undue discomfort” to the occupants.18 Yet cities across the country built sur-
faced runways. A few examples will suªce to show the reasons behind cities’ de-
cisions to adopt runways over turf.
In 1929, the airport superintendent for Columbus, Ohio, reported to Airports
magazine that he had chosen paved runways over turf due to airline demands for
regularity of service and the high volume of traªc his city hoped to attract. Ac-
cordingly, his runways were built with a five-inch concrete base with a 1.5 inch as-
phalt crown. This decision drew an attack from a soil engineer named Wendell
Miller, who argued that competent soil engineers could have built an “all over”
turf field suitable for year-round use for half the cost. “Soil technology,” he wrote,
“as a definite engineering science, capable of lessening the cost of airports, has
not yet been recognized by the oªcials of air transport companies who have had
the direction of airport construction policies.”19
At issue here was the problem of drainage. Since aircraft certainly could use a
turf field in dry weather, airline demands for regularity required some means of
maintaining a dry field in wet weather. Miller believed that a properly designed
drainage system could achieve this. Unfortunately, neither article addresses the
system installed at Columbus (all airports, turf or not, had drainage systems of
one sort or another), but details on other wet climate fields are available, and us-
ing these we can at least qualitatively explore the limits of drainage.
The supervisor of public works for Rochester, New York, described his city’s
choice of flying field and the work that had to be done to make it useable. Stating
his preference for turf fields, he wrote that the field selected by the city, located in
a bend of the Genesee River so that it could serve seaplanes and float planes too,
tended to be waterlogged near the center, and the heavy, clay-rich soil exacerbated
the problem. He had laid 7,500 lineal feet of drain tiling to remove the excess wa-
ter from where the runways were to be built. It was not enough piping to provide
a firm surface for an all-over field, however. It was simply enough to protect the
runways from being undermined. At Rochester, the pipes were placed under the
runway centers, further minimizing construction costs. The runways themselves

were built of 17,000 tons of stone brought in by team and wagon. The stone was
worked into the soil, and crowned with several inches of cinders bound with as-
phaltic oil to make the runways waterproof and provide a resilient surface capa-
ble of withstanding tail skids.20 Later, the cinders were replaced with asphalt
because cinders that worked loose tended to damage propellers. Rochester’s super-
intendent chose what he thought was the least expensive approach to building an
airport, with limited drainage intended specifically to protect the runways.
Similar reports of airport construction abound. Iowa City, Iowa, laid 150 by
2,800 foot runways consisting of an eight-inch crushed limestone base, packed
hard and rolled. Drain tiling was again installed to protect the runways, due to an
“excess of moisture during winter months, freezing and thawing of the ground,
[and] the heavy black soil.” St. Paul, Minnesota, one of the first additions to the
transcontinental mail route, spent $295,000 in building its municipal airport in
1926 and floated an additional half-million dollars worth of bonds by 1929. Like
Rochester’s original airport, St. Paul’s was built on low land next to a river (the
Mississippi). Here, the problem was drainage and snow removal. Photographs
show that plowing the runways for the mail planes coming from Chicago (which
had snow far less often, and therefore mail planes could not fly in and out with
skis) severely damaged the surface. This meant an expensive reconstruction job
each spring, until the airport replaced turf runways with hard surfacing.21
Airport managers chose a combination of drainage and runways to reduce the
construction costs of their fields. Runways, like roads, required a firm, dry sub-
surface to prevent cracking of the pavement, and airplanes needed a firm, dry sur-
face to prevent tire rutting and bogging down. Managers could not avoid drainage
entirely, but the use of runways reduced dramatically the areas that had to be
drained. Drain piping had to be buried between four and twelve feet down, with
wetter areas requiring shallower piping. Shallower piping, in turn, drained a
smaller surface area, which meant fields in wet areas required a great deal more
piping. An all-over field in a wet area, such as Rochester, meant an extensive
subsurface drainage system. Having to drain only runway areas seemed a cost-
e¤ective solution, especially in the light of the army’s well-known drainage prob-
lems with two of its major fields during the 1920s.
Perhaps the best evidence of the inability of drainage alone to solve the prob-
lem of boggy airfields at a reasonable cost comes from the army’s struggles at
Selfridge Field near Detroit and the original Bolling Field near Washington, DC.
Both had been chosen for their proximity to important cities during World War I.
Selfridge’s location was chosen to ease the testing of aircraft that the automobile
44 blind landings
industry was expected to produce for the war e¤ort; Bolling’s location was cho-
sen for easy access to the army’s bureaucratic center. Both had terrible drainage.
Selfridge was the second army field to receive a surfaced runway after McCook
Field in Ohio. Despite the army’s continued adherence to all-over fields, it was
forced by conditions at low-lying Selfridge to install runways after sixty-three
miles of drainage tiling failed to produce a useable year-round field.22 The e¤orts
of Rochester, in laying a mere 1.5 miles of tiling, pale in comparison.
The army’s experience with Bolling Field was worse still. Originally, it was lo-
cated on the Anacostia River, on hydraulic fill dredged from the river. The Inte-
rior Department had intended to make the area a park until the army conducted
a campaign in Congress to get possession of the field. Because the field was only
a few inches above the river, it was prone to repeated flooding. Consisting of the
extremely fine stu¤ that typically makes up river bottoms, the soil was largely im-
permeable and did not drain. Drainage piping proved useless, as the ground a
few inches down was fully saturated. After an embarrassing cancellation of a ma-
jor exercise in 1927, the army began a new campaign to move the entire field to
a nearby site, which Congress approved in 1929.23 In the meantime, the army
built a runway on the old site, in order to get some use out of the field. These well-
publicized failures no doubt impressed upon airline and airport managers that
even massive drainage systems might be insuªcient when faced with low-lying
fields in wet climates and no doubt influenced their decisions to reject soil engi-
neers’ claims for all-over turf fields in favor of runways.
Bolling and Selfridge were poor choices from the standpoint of flyability, and
the army paid a high price for its choices. But just as the army chose Selfridge
and Bolling Fields without much attention to geography, cities and city planners
were more concerned with placing airports near other forms of transportation
than with choosing high-quality flying environments. The scarcity of large, un-
developed tracts of land near surface transportation in urban areas severely lim-
ited the choices of sites available to cities. Some cities chose to make new land,
as the number of airports built on fill attest: San Diego’s Lindbergh Field, Boston’s
Municipal Airport, New York’s La Guardia, and San Francisco’s Municipal were
all built during this period at least partly on fill. Fill, especially hydraulic fill as the
Bolling example suggests, seldom resulted in high-quality, naturally drained land,
thus leading to the use of runways. The claims of soil engineers notwithstanding,
it seems clear that the eªcacy of drainage had its limits, and many municipal air-
ports exceeded them. This problem, combined with the unwillingness of military
and commercial organizations to accept the performance limitations imposed by
low-pressure tires, meant that runways were the only realistic solution. Con-
fronted with snow and ice removal in the north, water nearly everywhere, and the
demands placed by airlines on regularity, safety, and eªcient transshipment of
passengers and mail, the technology of the all-over turf field became obsolete in
the United States during the 1920s.
making a technology fit:

evolution of the leader cable system
The obsolescence of grass fields in turn impacted blind landing technologies

intended for them. The United States, the United Kingdom, and France had all
experimented with balloons, acoustic location, and “leader cables” in their at-
tempts to solve the blind landing problem during the Great War. After the war,
primary interest in France and Great Britain focused on leader cables. In France,
the system was known as the “Loth system” after its inventor William Loth; in
Great Britain and the United States, it was simply called the leader cable system.
A number of variations were developed, and I use the term leader cable system to
refer to the class of systems in general and designate specific incarnations by the
name of the inventor.
Charles Stevenson of the Royal Society of Edinburgh first proposed the leader
cable system’s basic idea in 1893. He described the testing of a single cable laid
along the sea floor to see if induction through sea water could happen without
the use of parallel wires. Stevenson undertook this experiment after he had the
idea that such a cable could be of great navigational assistance to ships. He laid a
single cable on the floor of a lake and fed it with eighty volts to produce an elec-
tromagnetic field. The combination of an uninsulated copper wire coil hung at
each end of the boat, connected with a wire hooked to a telephone receiver, acted
as a detector. Using the telephone, an operator could hear a tone that increased
in intensity as the boat neared the cable, as long as the wire in the boat remained
roughly perpendicular to the cable. Once the boat was centered over the cable, the
tone stopped. The author presented no scientific analysis of the device’s function,
and how, exactly, the leader cable system worked remained subject to debate in
France through the 1930s. Lack of scientific understanding of a useful new de-
vice rarely stands in the way of its adoption, however, and the German navy put
a version of this idea into service before the Great War. The British Royal Navy be-
came interested in leader cables after finding out about them from captured Ger-
man sailors and installed an eighteen-mile cable run into Portsmouth. It aban-
doned the idea due to the high cost of maintaining under water cable, particularly
since ships’ anchors tended to hook and cut it.24 The Royal Navy’s rapid aban-
46 blind landings
donment of the nautical leader cable anticipated some of the reasons for the aero-
nautical version’s failure.
France took the leader cable idea the furthest after the war, developing a guid-
ance version for point-to-point aerial navigation and a landing-aid version. The
French government voted to build guide cables from Paris to London, Paris to
Brussels, Paris to Strasbourg, and from Paris to Versailles in 1922, and consid-
ered building one across the Sahara desert. The British Parliament approved
funding for the section of the Paris to London route on British soil in 1923. The
records available in the United States suggest that little was actually built, but they
are hardly definitive. The one published work on air navigation before World War
II does not discuss leader cables at all and indicates that aerial navigation in Eu-
rope was carried out through the use of radio direction-finding stations. However,
a series of articles in the French aeronautical press during the 1930s debates the
merits of the leader cable system, suggesting but not confirming that at least parts
of the planned system were built.25 The records related to the landing-aid version
are fortunately somewhat better.
The first published reference to any of these aircraft systems is Loth’s “On the
Problem of Guiding Aircraft in a Fog or by Night When There Is No Visibility.” It
was presented to l’Academie des Sciences in December 1921, and it described the
system for overland navigation and blind landing of aircraft based on the gener-
ation of an artificial electromagnetic field. He had been asked to develop the sys-
tem by the undersecretary of state for aeronautics and had built his prototype at
Villacoublay. To provide guidance for aircraft trying to land, his system passed an
alternating current of six hundred cycles per second through a loop of cable sur-
rounding the landing field. This produced a varying electromagnetic field that a
properly equipped aircraft could detect.
An aircraft’s equipment consisted of three sets of two frames, or loops, ori-
ented along the aircraft’s three axes. Passing any wire loop through a magnetic
field perpendicular to the field’s lines of force induces a current in that wire, and
to detect the current induced in the wire frames, the system employed an audio
amplifier. Two sets of the frames were vertical, with one aligned with the longi-
tudinal axis of the aircraft, and the other perpendicular to it. The third frame was
horizontal. The longitudinal frame received best when parallel to the guiding
wire, while the transverse frame received best when perpendicular to the wire.
They could be connected to a goniometer to display the aircraft’s inclination to
the guide wire, while connecting the longitudinal and transverse frames indicated
whether the aircraft was right or left of the guide. Finally, connecting the trans-
verse and horizontal frames showed vertical position with respect to the cable. It
thus provided three-dimensional navigation information. Early versions of this

system relied upon a pilot’s hearing to interpret the system’s audio signals. Later
versions employed visual indicators on the pilot’s instrument panel.26
French engineers published a series of articles in the journal l’Aéronautique
during the 1930s which explained the operation of the system and described the
results of tests conducted at the aerodromes of Villacoublay and le Bourget. Un-
fortunately, those articles argue the scientific rationale of the system and give no
insight as to how e¤ective it was or how widespread its use became in France.
Loth also founded a company to develop and manufacture his system, the Société
Industrielle des Procédés Loth, and even an American subsidiary, the American Loth
Company.27
Loth’s system does not appear to have been placed in either commercial or
military service. Although no source provides a reason, Christienne’s history of
French military aviation provides one clue. The French aeronautics establishment
was fiscally poor and organized to create prototypes but not actual procurement
and production. The result was the rapid obsolescence of French aircraft and avi-
ation infrastructure. The death of the Count de la Vaux, who was killed along with
his pilot when his plane collided with a landing cable installed at Vaux sur Seine,
did not help the system’s chances.28
Across the Channel, British investigators working at the Electrical Research
Section of the Royal Aircraft Establishment (RAE) had begun exploring a similar
application of electromagnetic principles to solving the blind landing problem
some time before 1920. Unfortunately, security concerns prevented publication
of their work before 1926. Led first by one Major McAlpine, and after 1923 by
Flight Lt. H. Cooch, the RAE’s team designed a blind landing system similar in
principle, but significantly di¤erent in detail, to the Loth leader cable system.
Three key di¤erences were the use of a racetrack shape for the cable, with one of
the straight sides marking the landing area; the use of thirty-four Hz current in
the ground cable; and the use of a frequency reverser in the plane to convert the
received thirty-four Hz energy to direct current.29
The use of a mechanical frequency reverser was the most significant di¤er-
ence in the British system. The French had relied upon the generation of a weak
electromagnetic field by the cable and used a vacuum-tube-based audio frequency
amplifier in the aircraft to amplify the signal. The British chose to generate a
much stronger electromagnetic field on the ground, and because direct-current
reading instruments were more sensitive than alternating-current instruments,
convert the received alternating current energy in the plane to direct current via
a mechanical-relay-based device. Because mechanical relays could not operate
48 blind landings
very rapidly, the frequency of the generated electromagnetic field had to be

kept low. This created two problems for the British. First, as was becoming well
known, low-frequency signals had unpredictable propagation characteristics.
Worse, the Electricity Commissioners had recently chosen fifty Hz as the stan-
dard frequency for England’s electric train system.30 The nearness of the two fre-
quencies, coupled with the much higher-intensity electromagnetic fields gener-
ated by the very high current use of the trains, resulted in significant interference.
Although the system tested successfully in a small-scale laboratory setting, when
fully installed at Farnborough, it did not do so well.
Cooch presented his findings to the Royal Aeronautical Society on February
19, 1926, and the audience’s response makes clear that it was well aware of the
frequency problem. They were also aware of the French Loth System and of the
theoretical benefits that they could gain through its use of higher frequencies.
Higher frequencies, however, required the use of higher speed switches of some
sort or other, and only the vacuum tube, or as the British called them, thermionic
valves, were suitable for use either as high-speed switches, or as amplifiers. Nei-
ther Cooch nor his audience that evening believed that thermionic valves were re-
liable enough or, apparently, that they could be made reliable enough for use in
a safety-related system.31
Therefore, by 1930, the RAE had entirely abandoned the leader cable system.
Instead, it began pursuing an idea that air forces had tried unsuccessfully during
World War I: the tethering of a balloon above the fog where pilots could see it. A
pilot seeing the balloon maneuvered as close to it as he could and then simply de-
scended at a constant rate along the proper compass bearing until the plane hit
the ground. The RAE realized that such a landing would be at best frightening to
passengers, and so it devised a “ground proximeter,” consisting of a weight hang-
ing from thirteen feet of linen thread below the wheels. When the weight hit the
ground, the tension on the thread was released, causing a light to flash on the in-
strument panel. This was to give the pilot enough time to pull the nose of the
plane up to reduce the shock of the landing. A strengthened undercarriage was
still required.32
The ground proximeter did nothing to resolve the problems that World War I
fliers had encountered with the balloon method. The altitude of fog banks
changed over time, often obscuring the balloon, and the inaccuracies of baro-
metric altimeters meant that a constant rate descent was often not constant at all,
which occasionally led to reaching the ground outside the boundaries of the aero-
drome. Pilots tended to be uncomfortable about such possibilities and in the
United States rejected another system that also relied upon the barometric al-
timeter.33
Finally, as was pointed out by F. Tymms of the Directorate of Civil Aviation
during a discussion of the balloon test results, “However valuable this result
might be, it could not be considered a satisfactory solution for the needs of air
transport, whose successful development demands regularity. Further, if pilots
were merely to use the system when accidentally caught in fog, they would never
attain confidence in it.”34
The emergency nature of the balloon system ensured that it did not appeal to
commercial aviation interests. Because the balloons had to be deployed whenever
poor conditions occurred, they were not always available. Pilots would not be able
to practice using them during good weather, and practice was the only way to de-
velop pilots’ confidence in their ability to use the system safely. An emergency
use only system was not what commercial airlines sought, and it is surprising
that the RAE seriously considered the balloon landing idea at all. The British
abandoned the leader cable system, which could have been a part of everyday
flying, for an emergency system that could never be routine. Rejecting thermionic
valves thus led them to a technological dead end that is all the more surprising
given that wireless, which relied upon tubes, was already in widespread use for
navigation in Europe and the Empire.
In Europe, therefore, no single cause for the nonadoption of leader-cable-
based systems for landing aircraft exists. Instead, each country had a unique ex-
perience. In France, an idiosyncratic state policy is the likely cause. In Great
Britain, a conscious choice against vacuum tubes provoked abandonment and re-
placement by what was an even less suitable procedure. Germany, whose aviation
activities were severely limited by treaty restrictions, did not conduct blind land-
ing research until 1931, when Telefunken, Lorenz A.G., and Deutschen Versuch-
sanstalt für Luftfahrt, the state-supported aeronautics research laboratory, em-
barked upon a radio-based development program based loosely on a U.S. Bureau
of Standards’ idea.
Although the leader-cable-based systems failed in Europe, they nonetheless fit
within the surrounding technologies of airplane and landing field. They did
not produce great precision, but none was necessary to place an aircraft some-
where within the confines of a large aerodrome. They were thus suited to open-
field landings by aircraft nimble enough to negotiate a relatively short approach,
which were all that existed in Europe throughout the mid-1930s. They did not fit
within the technological context of American aviation, however. The U.S. need
Figure 2.1. The American Loth Company’s version of the Loth leader cable system. The upper portion of the diagram shows the approach path as viewed
from the ground, with the Morse code signals displayed in the different approach zones. The “leveling surface” is simply the magnetic field intended to
indicate to the pilot that it was safe to level out and land. The lower portion of the diagram shows the bird’s-eye view of an approach, keyed to the visual
indications that the system provided to a pilot. Notice the lack of anything resembling runways on this figure. Aviation (December 1931): 199.
for greater precision was due to the adoption of runways in North America more
than a decade before their widespread use in Europe. To provide enough preci-
sion for runway landings, the leader cable idea had to be modified.
During the war, the United States had also explored a blind landing system
based upon the leader cable principle. On Armistice Day 1918, the U.S. Bureau
of Standards tested a prototype consisting of a “transmitter” composed of a 160-
foot coil of copper wire, which the bureau’s researchers wrapped around the roof
of the Radio Building, powered by a 500-watt generator. Wire coils wrapped
around the lower wings of a JN-4 “Jenny” biplane connected to an audio fre-
quency amplifier served as a detector.35 Although the pilot was able to hear the
signal the system produced, he did not attempt to land on the building’s roof us-
ing it, likely because the roof was too small. Several more months of testing at the
College Park, Maryland, airfield followed, but postwar austerity measures put an
end to the project. The Post Oªce, which assumed airmail responsibilities from
the Army Air Service after the war, considered navigation between airfields a
more important problem and put its limited funding into the construction of
lighted airways. No further organized research into blind landings was done in
the United States until 1927, when the Bureau of Standards was given the task by
then secretary of commerce Herbert Hoover. In Britain and France, however,
blind landing e¤orts continued under government auspices and focused on the
leader cable approach.
In the United States, proponents of leader-cable-type systems developed three
di¤erent variants in an attempt to make the idea fit into a di¤erent context. The
first was a product of the American Loth Company and built upon Loth’s work in
France. Beginning in 1930, the Company modified his original formulation to re-
place the single, aboveground loop with a series of concentric underground ca-
bles, which may have been a response to Count de la Vaux’s death (Figure 2.1).
As a plane flew toward this installation, its pilot first heard a Morse code D sig-
nal until the craft crossed the field boundary, where D was replaced with a con-
tinuous tone. Inside the field, the system provided a Morse code U until the air-
craft was twenty feet above the surface, when the small inner cable’s I signal
overwhelmed the U. That was the signal to land.36
This ingenious arrangement met with no greater success than its parent sys-
tem. No one ever adopted it. By 1932, all major U.S. airports had adopted run-
ways, and as the geometry of the system makes very clear, it could not serve a
runway-based airfield. Like its European forbears, it relied upon the provision of
an open landing area. The circular form allowed landing from any direction,
which a runway-based airport could not permit. Although it might have served
52 blind landings
the aerodrome-like U.S. Army fields, the army emphasized the need for porta-
bility in any blind landing equipment throughout the 1930s, which the Loth sys-
tem certainly was not. The American Loth Company’s system remained too much
like its European forebear to fit the U.S. technological context. Instead, inventors
had to modify the leader cable arrangement to suit U.S. airfields.
While the American Loth Company worked on its circular system, Earl C.
Hanson, who had installed a leader cable for ships in New York harbor, designed
a system that could serve runway-based fields. This was installed and tested at the
Lansing, Illinois, Ford Airfield, which had opened in 1928 with the first concrete
runways in the United States (Figure 2.2). It relied upon two single-turn cables
extending nearly four thousand feet beyond the airfield boundaries and aligned
along the approach path. By exciting both cables with a 1,000 Hz frequency and
using switches to alternately open and close the circuits, a “dot” pattern was ra-
diated in one cable and a “dash” pattern in the other. A pilot flying precisely
halfway between the two cables heard a continuous tone. Further, a “gun coil,”
which produced a loud tone when the plane passed directly over it, informed the
pilot when the plane had arrived inside the airfield boundaries. An inductive al-
timeter provided altitude indication. Hanson’s successful demonstration in 1930
drew congratulatory telegrams from Lee De Forest and Capt. Stanford C. Hooper
of the Navy Department’s radio section, who took interest in the basic design of
the system for its applicability to airships.37
Hanson’s work seems to have inspired Hooper to support further develop-
ment of the basic idea, which was carried out by a navy radio engineer working
for Hooper’s Radio Section of the Bureau of Engineering. Edward Dingley de-
vised the system shown in Figure 2.3 in the mid-1930s. The system fed alternat-
ing current through the cable arrangement shown in the diagram, from the gen-
erator labeled “3.” The cable set defined by points 36-37-3 was the main loop,
which generated a powerful electromagnetic field. Each of the smaller loops (for
example, 38-39-3) was fed with current of the opposite phase to that in the main
loop, establishing an opposing electromagnetic field. This resulted in destructive
interference between the two fields. By carefully choosing the currents in each ca-
ble, engineers could establish an electromagnetic field whose intensity decreased
as an aircraft approached the runway. Using a radio receiver designed to detect a
line of constant intensity, the pilot could follow what would appear as a straight
glide path to the runway. The aircraft installation consisted of the same wire
frames, in the same orientation, as those originally proposed by Loth and were
connected to a cross-pointer instrument much like that proposed by the Bureau
of Standards in 1931.
Figure 2.2. Hanson’s modified leader cable system. The diagram is clearly not to scale,
but it demonstrates how this system could serve runway-based airfields. The center “on
course zone” is 200 ft wide just before crossing the airport boundary, which was a
common runway width. The approach length is a bit short even for 1930 aircraft. Other
systems designed between 1929 and 1933 generally supported a one- to two-mile
approach. Since commercial aircraft of the day landed at about 60 mph, Hanson’s
approach provided about half a minute of guidance before landing. Note that the system
provided no altitude indication, a problem that caused pilots to reject other landing aids
during the 1930s. Aviation (22 February 1930): 402.
Dingley advertised its chief advantages as its independence of radio frequen-

cies and the consequent atmospheric and ground distortions. It did not employ
vacuum tubes, and Dingley argued that this made its maintenance costs negligi-
ble.38 The electromagnetic field provided lateral and vertical guidance, so a sepa-
rate localizer was unnecessary, as were marker beacons. Finally, since the system
did not require antennae on the ground, or perhaps more properly, the antennae
were underground, it presented no obstructions on the field.
To test Dingley’s vision of a landing system, the navy built a prototype within
the confines of its airship base at Lakehurst, New Jersey. That forced a short ap-
proach and an accordingly steep descent, which was fine for lighter-than-air craft
but impossible for airplanes. The navy considered lengthening the approach path
enough for airplane use during 1940, until the Lakehurst base commander
Figure 2.3. The Dingley leader cable system. Like Hanson’s it was supposed to guide
planes down a narrowing funnel of electromagnetic energy. Unlike Hanson’s, it
provided a “glide path” through the use of successive additions of oppositely
polarized electromagnetic fields. Destructive interference between the main field
and the succession of smaller fields resulted in a decreasing field intensity which
appeared to the pilot as a glide path. By 1938, the army was demanding approaches of
at least ten miles, and preferably fifteen, for its aircraft. For this system, the cable array
would have had to be extended the entire approach distance. Edward Dingley Jr., “An
Instrument Landing System,” Communications (June 1938): 7.
pointed out that lengthening the approach meant having to secure easements on
land well outside the base. In Lakehurst’s case, that meant having to get ease-
ments on cranberry bogs, which he expected to be expensive and an installation
nightmare. It was briefly examined by Vannevar Bush’s National Academy of Sci-
ences committee on instrument landing aids in 1939 and by the CAA at the same
time, but both considered radio a better solution, most likely due to the high costs
involved in building and maintaining what would have been ten-mile-long cables
in order to guide the much larger, faster aircraft of the late 1930s to safety. Ad-
miral Stark, the Chief of Naval Operations, finally ordered the project’s cancella-
tion in September 1941.39
conclusion
One clear factor that emerges in discussion of landing aids and landing areas
is that of the environment in which the technologies of aviation had to work. Turf
fields and leader cables were not the only technologies a¤ected by the physical en-
vironment of aviation, and later chapters show similar problems with other tech-
nologies. Technologies designed without explicit recognition of these environ-
mental constraints had severe problems, and the leader cable designers were not
the only innovators who failed to overcome them.
The environmental problems encountered by aviation technologies cannot be
separated from economic factors, however. Cities could have, and in retrospect
perhaps should have, put their airports inland, away from water problems for sur-
faces and landing aids, as well as navigation hazards imposed by urban struc-
tures. They did not do so, due to city planners’ emphasis on airports as nodes of
transfer, which reflected their primary function as spaces for commerce.
Leader cables were therefore made obsolete in the United States by the adop-
tion of runways driven by the intertwined demands of commerce and geography.
Although various inventors attempted to make the system fit the new technolog-
ical context, the result was systems that fit just as poorly as their predecessors.
The intertwined demands of commerce and geography thus e¤ectively doomed
the leader cable to the same fate they had spelled out for the turf field, relegating
both to the technological scrap heap. But the same airline demand for regularity
which had played such an important role in creating runways, and which still reg-
ularly ran afoul of bad weather in the form of poor visibility, remained unsatisfied.
While leader cable inventors were tinkering with their systems, therefore, the
airlines put pressure on the Post Oªce, which paid them to carry the mail, and
the Department of Commerce, which maintained the federal airways for them,
56 blind landings
to find a solution to the blind landing problem. The government’s proposal had
nothing to do with leader cables. Instead, the commerce department turned to
the only technical development group available to the civil half of the U.S. gov-
ernment: the National Bureau of Standards, which possessed a radio research
laboratory.
chapter three
Radio Blind Flying
Two government organizations played key roles in the late 1920s and early
1930s blind landing research, the U.S. National Bureau of Standards (NBS) and
the Aeronautics Branch of the Department of Commerce. The U.S. Congress had
established the NBS in 1901 to devise and maintain a uniform system of weights
and measures for the nation. That mission included maintenance of frequency
standards for radio transmission, and the bureau established a radio research lab-
oratory in 1913 to carry out that responsibility.1 Researchers at that laboratory cre-
ated the basic model on which the current instrument landing system is based,
the now familiar marker beacons, localizer, and glide path.
While the Bureau of Standards’ authority was well established, the Aeronau-
tics Branch was a new agency when blind landing work began. Civil aviation in
the United States had been unregulated before 1926, with the exception of the
Post Oªce’s mail service. The postmaster general’s decision to privatize airmail
carriage in 1925 resulted in a new commercial airline industry and its almost in-
stantaneous demands for federal regulation. Airline pioneers feared that each
state would impose its own regulations in the absence of a federal oversight, re-
sulting in chaotic operating conditions and consequently little chance of prof-
itability. The airlines prevailed upon Herbert Hoover, the secretary of commerce,
who helped draft and push through Congress legislation creating the Aeronau-
tics Branch in 1926. That organization, in turn, was reorganized in 1934 into the
Bureau of Air Commerce, still within the Commerce Department.2
Hoover had also arranged for a joint program between the new Aeronautics
Branch, the Bureau of Standards’ radio lab, and the Bureau of Lighthouses to re-
58 blind landings
place the giant searchlights that the Post Oªce had used to light the federal air-
ways with radio ranges that would better facilitate blind flying. The radio ranges
were an outgrowth of an army project at Wright Field. They broadcast signals that
pilots could hear in their radio headphones. Using the signals, pilots would be
able to fly blind from range to range across the country once all the ranges were
in place. Beginning in June 1926, NBS engineers had designed these ranges with
the Aeronautics Branch’s support, and to them also fell the task of producing a
blind landing system. Unsurprisingly, the Radio Section chose to base the blind
landing system on the range design.
The ranges’ instability made the choice problematic. Pilots could not depend
upon them to maintain a consistent course, and they dubbed the ranges, not en-
tirely humorously, “rotating ranges.” The tendency of these ranges to vary was
dangerous but rarely fatal. Most of the time, the variations amounted to a few de-
grees and were not enough to lead pilots into danger. Variation of a few degrees
in a blind landing system, however, proved intolerably dangerous. To land safely
under blind conditions on a runway, aircraft had to be aligned within fractions of
a degree to the runway centerline, and stability was therefore vital. The Bureau of
Standards sentenced engineers to decades of work on this instability problem by
choosing radio as the basis of a blind landing system.
The instability descended from the use of low-frequency radio signals, in the
100 to 300 kilohertz (KHz) band. This was “best practice” in the 1920s. Higher
frequencies were known but could only be broadcast at relatively low power. The
lower-frequency AM stations were much more likely to su¤er from environ-
mental interference than were the higher-frequency FM stations. In this discus-
sion, “environmental” conditions refers both to those interferences deriving from
the “natural” environment, such as storms, and those extending from the “built”
environment, like bridges, power lines, and buildings. Interference e¤ects get
worse as one moves down the radio frequency spectrum, and the radio ranges in
use during the late 1920s and 1930s broadcast on frequencies far below those
used in modern AM radio broadcasting.
At root, the problem blind landing pioneers faced was that radio waves prop-
agate through the environment and are a¤ected by it. Because “first” nature, the
nonhuman environment, and “second” nature, the built environment, are in-
consistent, radio propagation through them is as well. Fixing that instability
problem meant finding a way around nature’s inconsistency. Pavement had been
the answer to the inconsistency of the Earth’s physical surface, giving planes a
smooth, flat, hard place to land; unfortunately, consistency in radio-based blind
landing equipment proved much harder to achieve.
radio blind flying 59
flying the (unstable) beam: the national bureau

of standards system
The National Bureau of Standards had been asked by Secretary of Commerce

Herbert Hoover in June 1926 to begin developing radio navigation beacons to re-
place the high-powered searchlights that had provided guidance along the Post
Oªce’s airmail routes. These beacons, which were called ranges even though
they did not provide a pilot with a distance indication, provided four courses along
which pilots could fly either toward, or away from, the ranges. The navigational
“beams” were formed by overlapping two signals, one coded with a Morse code
A and the other with a Morse code N, as Figure 3.1 shows. The airplane’s receiver
simply transmitted Morse code signals to the pilot’s headphones, and pilots flew
these ranges by trying to make the two signals merge into a single, continuous
tone. When a pilot heard a continuous tone, the aircraft was in one of the “on
course” zones shown in the figure.3 This type of arrangement came to be called
an equal signal, or equi-signal, transmitter.
Flying the radio ranges during the late 1920s was not a simple task. Staying
on course meant pilots had to listen intently to the tones in the radio headphones
for hours on end and use them to orient the aircraft in aural space. Because the
ranges operated at low frequencies, they su¤ered from static interference that
could overwhelm the real signals, and they also su¤ered from fading due to
changes in the upper atmosphere. Further, the ranges had more than one task.
Their signals were interrupted every thirty minutes or so for weather and traªc
advisories, which were transmitted by Morse code. All of these issues made the
ranges challenging to fly, and in bad weather their demands on a pilot’s attention
were exhausting. There are numerous anecdotal accounts of accidents caused by
simple fatigue. Nonetheless, the ranges worked well enough to warrant deploy-
ment throughout the United States, and during World War II the army stripped
many lesser-used air routes of their ranges so the equipment could be shipped
overseas.
In internal NBS memoranda in 1928, Harry Diamond, an engineer in the ra-
dio section, proposed the use of a modified low-frequency four-course range (330
KHz) as a landing aid. This was to allow “blind” landings to be made by provid-
ing a pilot with information about the aircraft’s lateral position with respect to the
runway. The addition of two low-powered, high-frequency (97.3 MHz) beacons in
line with the approach path, one at the airport boundary and one two miles away,
would provide a pilot with two positive position “fixes.”4 A pilot following the di-
60 blind landings
Figure 3.1. Radiation pattern of a four-course range, 1929. Pilots heard the Morse code
A-N signals in their headphones, and maneuvered their aircraft so that the two signals
merged to form a continuous tone, effectively orienting themselves in aural space.
Ranges of this type were called aural ranges. Replacing the Morse code signals with
two different modulation frequencies produced a signal that could be detected by an
instrument. A range of that type was called a “visual range,” and initially utilized the
tuned reed indicator (see Fig. 3.2). Reprinted with permission from Nick Komons,
Bonfires to Beacons: Federal Civil Aviation Policy under the Air Commerce Act, 1926 –1938
(Washington, DC: Smithsonian Institution, 1989), 157.
rectional beam (called a localizer) would pass over the outer beacon and begin de-
scending at a constant rate. The inner marker, once heard, indicated that the plane
was inside the airport boundaries and could safely land.
Doolittle’s 1929 blind flying experiments at Mitchell Field were the first to em-
ploy the localizer idea. He borrowed from the army a Consolidated NY2 trainer
and equipped it with special shock absorbers, a Sperry artificial horizon, a Sperry
directional gyroscope, a specially designed Kollsman precision altimeter, and stan-
dard instruments. He also borrowed a low-frequency localizer from the NBS.
With this equipment, Doolittle was able to make the first deliberate blind landing
on September 29, 1929.5 He reported that over the several months of the exper-
iments, several hundred blind landings were made with the NY2.
The localizer he used was a visual type, which was designed to operate a tuned
reed indicator in the cockpit (Figure 3.2). The visual indicator was an innovation
intended to replace the aural signals of the standard four-course range. It had a
pair of metal reeds, each of which was mechanically tuned to resonate at the mod-
ulation frequency of one side of the equi-signal localizer. When the aircraft was
on course, the two reeds appeared to be the same length. If the plane was to the
right or left of course, the reed on that side would appear longer than the other,
and the pilot would then turn toward the shorter of the two reeds.
Doolittle’s landing procedure in the Guggenheim experiments was to fly to-
ward the beacon maintaining one thousand feet altitude. When the aircraft
passed directly over the beacon, the reeds stopped vibrating, signaling that the
plane was over it. The box containing the reeds then had to be taken out of the re-
ceiver, turned over, and reinserted, while the aircraft flew about four miles. The
pilot then made a 180-degree turn, turned over the reed box again, and descended
to four hundred feet. Once the aircraft had passed over the beacon again, Doolit-
Figure 3.2. Tuned Reed indicator. This is a later version that fixed the problem of having
to remove the reed container from its housing, flip it, and reinsert it during an approach.
The “To” side was on top while flying toward a radio range, and the “From” side was
rotated up when flying away from one. The two reeds are the white lines at the center
of the indicator. From H. Diamond and F. Dunmore, “A Radio Beacon and Receiving
System for Blind Landing of Aircraft,” Bureau of Standards Journal of Research RP 238
(1930): 905.
62 blind landings
tle idled the engine, kept the wings level, and maintained a sixty-mile-per-hour
airspeed as the aircraft descended to fifty feet. Once there, Doolittle opened the
throttle again slightly to reduce his descent rate from one thousand feet per
minute to six hundred, which was the most he could comfortably tolerate.6 The
aircraft’s course was maintained with the directional gyroscope after passing over
the beacon house. Doolittle then flew the plane into the ground. Because Mitchell
Field was still a grass all-over field, it did not matter where on the airport surface
he landed.
Doolittle readily admitted that this was an experimental system, and it never
became more than that. The reed indicator, which was also used with the radio
ranges along the federal airways system, was diªcult to use, and the need to flip
it over during an approach made it dangerous. Doolittle’s precision altimeter,
though novel in that it was synchronized by radio with a barometer on the ground,
also gave pilots pause as a landing instrument because as a barometer, it merely
read pressure above sea level and not above the ground. Perhaps more impor-
tantly, it did not show distance above obstacles on the ground, which were a sig-
nificant hazard to fliers approaching an airfield. Barometric altimeters did not in-
dicate what pilots called absolute altitude, which meant an aircraft’s altitude “above
terrain,” including the ground and objects on the ground.
In addition, because they were mechanical devices, barometric altimeters
su¤ered significant lag, and they were not particularly accurate. Although Doolit-
tle’s specially made altimeter was reputedly capable of an accuracy of plus or mi-
nus ten feet, most pilots believed that plus or minus fifty feet was the best one
could expect from service equipment.7 A fifty-foot error was quite enough to
cause accidents, either by causing a pilot to come in too low and land short of the
field with the attendant risk of collision, or by approaching too high and over-
shooting the field, with the same result. Barometric altimeters, then, provided
neither the kind of information pilots wanted, such as absolute altitude, nor
suªcient precision to insure landing within the safe confines of an airfield. Pi-
lots were unwilling to trust them when flying close to the ground.
The lack of absolute altitude indication in the 1929 experiments proved to be
one of the most diªcult problems to solve in the development of landing aids.
Possible solutions using acoustic height finders (used by Germany’s Zeppelins)
and radio altimeters (demonstrated by E. F. W. Alexanderson) met with a luke-
warm reception from the aviation community.8 This disappointment was largely
a reflection of the state of the art. These were bulky, heavy devices that required
a dedicated operator. As aircraft of the period were small, and every pound con-
sumed by operators and equipment was a pound of capacity unavailable for pay-
ing cargo, these devices also did not suit the interests of the air carriers. Alexan-
derson’s device, for example, required the operator to transmit two signals in
close proximity and then analyze the resulting graphical trace to determine the
altitude. Then the operator could pass the information on to the pilot. The delay
was enough to make the device too slow to assist with a landing. The Zeppelin
acoustic height finder, although fine for airships that approached their mooring
towers at only a few miles per hour, was also much too slow for aircraft. Hence,
these ideas did not grab the attention of the community, and radio altimeters were
developed for bombing use during World War II, while acoustic altimeters went
with the Zeppelin into technological oblivion.
A di¤erent idea did catch the community’s interest. Another scientist at the
NBS radio section had conceived a means to provide another radio beam, angled
up from the ground, which would essentially provide a safe inclined path to the
landing field. Since the beam would be angled to clear all obstacles, an aircraft
following it would also clear them, thereby eliminating the obstacle problem.
Francis Dunmore’s “landing beam,” later called a “glide path,” combined with
Harry Diamond’s localizer and marker beacons, was intended to provide a three-
dimensional path in the air (see Figure 3.3). Two separate instruments were orig-
inally designed to display this path on the instrument panel, but after test pilots
remarked on the diªculty of trying to coordinate the use of two special landing
instruments and the Sperry artificial horizon, the bureau’s researchers combined
the landing instruments into one: the famous cross pointer (Figure 3.4). With the
circle at the instrument’s center representing the center of the two beams, and
the intersection of the two pointers representing the plane’s position with respect
to the beams, the pilot simply had to keep the pointers crossed over the circle to
e¤ect a safe landing. The NBS first published the full description of its system in
1930 and tested it with the aid of the Aeronautics Branch at College Park, Mary-
land, and Newark, New Jersey, through 1934. The proposal drew wide acclaim
from the aeronautical and engineering press. The German journal Zeitscrift für
das Weltflugwesen, which published a lengthy abstract of the NBS publication,
commented that this was the best solution to the blind landing problem yet de-
vised, while the U.S. secretary of the navy wrote to the secretary of commerce
to congratulate him for the “wonderful progress made” in blind landing after
demonstrations during the summer of 1933. Various inventors continued to in-
vestigate other solutions, but none drew the attention that this idea did.9
Diamond and Dunmore intended the landing beam to work as follows. Figure
3.5 shows what the two engineers believed was the result of propagating a radio
beam at 93.7 MHz from a set of horizontal dipole antennas arranged at a slight
64 blind landings
Figure 3.3. The beam pattern for the Bureau of Standards System. Note the curvature of
the glide path. This diagram presents the approach path for the Indianapolis Experiment
Station, and the image was probably made in 1936, when the Bureau of Air Commerce
began to work with that system again. The sensing of the instrument displayed in the
lower right is opposite that of the instrument in Fig. 3.4. The same image appeared in
numerous publications between 1937 and 1941. National Air and Space Museum,
angle to the earth’s surface. On the figure, the glide path is a surface of equal field
intensity, leading to this type of glide path being called a constant-intensity glide
path. Because the airplane’s motion toward the transmitter increases the received
field intensity, maintaining a constant received signal strength required the pilot
to drop away continuously from the center of the beam. This descent would, ac-
cording to the authors, inscribe a curve that was to touch the ground at a point
about two thousand feet from the antenna.
Despite the widespread press attention, however, the NBS system remained
an experimental apparatus. Test pilot Marshall Boggs’s ability to land occasion-
ally using the installation at College Park did not translate into an ability to do it
regularly. He reported in a 1931 letter to the director of aeronautical development
of the Aeronautics Branch, “I am beginning to doubt that the present approach
Figure 3.4. The cross-pointer indicator in an early form. The two zones at the bottom of
the instrument face were colored green and red, left to right, respectively. These were
later changed to yellow and blue because green and red could not be read under red
night-time cockpit lighting. The sensing of this instrument is the inverse of that in Fig.
3.3. The aircraft is represented by the intersection of the two needles, with the center of
the beam indicated by the large circle at the center. Author’s collection.
Figure 3.5. The constant-intensity glide path proposed in 1929. From Harry Diamond and
Francis Dunmore, “A Radiobeacon and Receiving System for the Blind Landing of
Aircraft,” Bureau of Standards Journal of Research RP 238 (October 1930): 924.
66 blind landings
immediately adjacent to the field on the localizer and landing beam can be made
consistently . . . It can be made occasionally, as I have already proved by my two
‘blind landings’ to date. It takes almost perfect conditions, however, to accomplish
it,—i.e. no wind or a steady south wind, freedom from rough air or bumps, free-
dom from static and the perfect functioning of the radio transmitting and re-
ceiving apparatus.”10
Further, he remarked upon the unsuitability of the College Park runway, which
at 250 feet wide was too narrow. With a forty-three-foot-wide plane, only 103 feet
to either side of the runway centerline remained, and he found that the aircraft
(a Ford Trimotor, designated NS1) could drift through such a width before he
could respond. He stated that at least a five-hundred-foot width was required to
ensure reliable, safe landings. These conditions, and the desire of the Aeronau-
tics Branch to test the system under commercial conditions, led it to arrange an
installation at Newark during 1933, a field that had been built with four-hundred-
foot-wide runways at the demand of Transcontinental Air Transport, which held
the airmail contract for New York City. The airport was expanded during Febru-
ary 1934 to make “virtually the whole area” available for landings.11 The increased
area available made the landing problem much simpler.
Boggs’s statement also points directly to one of the system’s environmental
problems. Bumpy air made flying the beams diªcult but did not alter the beams
themselves. Static conditions, the electricity buildup that occurs in storm condi-
tions, on the other hand, could substantially alter the localizer course and the
glide path angle, endangering the aircraft. The buildup of static electricity in the
atmosphere “pulled” the localizer course toward the storm. An aircraft following
the beam in would therefore not be aligned with the runway course, making a
safe landing impossible.
In order to determine how well the system would perform at other locations
and to begin training pilots in its use, the Aeronautics Branch arranged to install
two copies of the system at Cleveland, Ohio, and Oakland, California, during
1933.12 The two cities had the large, relatively unobstructed airfields that the sys-
tem’s long, low approach path required. Continued problems with the Newark in-
stallation along the lines Boggs had reported two years before, however, caused
branch oªcials to reconsider that decision, and only the Oakland installation was
made. The branch’s oªcials were concerned that they would not have enough
funds to improve the equipment if they made too many installations. The Oak-
land equipment had already been completed when that decision was made, how-
ever, and Oakland had agreed to pay the installation cost. The Aeronautics Branch
could not back out of that deal, and Oakland got its equipment.
In January 1934, Eastern Air Transport sent pilot E. A. Cutrell to conduct fur-
ther tests of the NBS’s Newark experimental installation. Cutrell found that the
localizer had a serious bend, probably caused by a nearby radio range, which,
combined with the high speed of the Branch’s Ford Trimotor airplane, made the
beam diªcult to fly. Cutrell accumulated nearly twenty-nine hours flying the sys-
tem, but he did not make blind landings because he was unfamiliar with the
plane’s landing qualities. He also had “no first hand knowledge of how much
abuse the airplane would stand.”13 He clearly expected blind landings using the
system to be less than smooth.
The methodology used by Doolittle, Boggs, and Cutrell was essentially the
same in all of these tests. They did not fly in bad weather. Doolittle and Boggs
quite literally flew under a hood that was buttoned over the cockpit, while a safety
pilot, who could see, monitored the approach. The safety pilot could take the con-
trols in the event of a bad approach and fly the plane back to a safe altitude. This
provided a significant psychological advantage that pilots flying under real con-
ditions did not have. Pilots “under the hood” knew that their safety pilot would
not let them crash. Without the reassurance provided by the safety pilot, pilots
had to rely entirely on their instruments, which therefore had to represent a sta-
ble, reliable system.
The Bureau of Air Commerce’s records, primarily because the agency dropped
the NBS work in 1934 under pressure from the army and the airmail crisis, do
not reveal the full magnitude of the system’s problems. The U.S. Navy picked
up the development work in 1933, however, via the Oªce of Naval Intelligence
(ONI).14 Seeking a landing aid for aircraft carriers and seaplane bases, ONI had
a group called the Washington Institute of Technology, based at College Park,
clone the NBS system. The Washington Institute was composed primarily of
laid-o¤ NBS researchers who had been collected by the director of naval intelli-
gence and given a contract to complete and “navalize” the NBS system. The navy’s
records on it are the most extensive of all the agencies involved in investigating
the NBS landing system, and they reflect the substantial technological diªculties
the aeronautical community had with it.
The navy’s clone was portable, with localizer and glide path transmitters in a
trailer and the marker beacons in motorcycle sidecars. It was oªcially designated
the YB system and was released in sealed containers in 1936 for commercial use
under the name Air-Track system, named for its College Park manufacturer, the
Air-Track Corporation. The Air-Track system has the distinction of providing the
first blind landing to a passenger-carrying flight in Pittsburgh on January 28,
1938.15 But the sealed nature of Air-Track, done to assuage navy national security
68 blind landings
demands, made the system impossible to maintain. It never became a serious

competitor to other variations on the NBS’s theme.
The navy’s tests on the YB system began in 1935, with tests aboard the USS
Langley. The tests were inconclusive but convinced the navy that although the sys-
tem was not yet suitable for aircraft carriers it could be useful for seaplane bases.
The navy continued developing the system and invited the chief signal oªcer of
the army to attend a demonstration the following year. The report the two oªcers
sent to the Signal Corps from the Aircraft Radio Laboratory at Wright Field pro-
vides an excellent review of problems encountered by the navy. They found that
the glide path beam tended to vary its angle with the ground depending on the
weather. On one particular day at the manufacturer’s College Park testing facil-
ity, the pilots reported that as the ground dried in the sun after an early morning
rain, the glide path angle slowly elevated so that by midafternoon, the path was
unflyably steep. They also found that the localizer beam was pulled o¤ course by
the presence of static conditions near the airfield.16 In other words, the system
did not function reliably in bad weather—which was when it was most needed.
Nevertheless, the navy was satisfied enough with the system to establish an
engineering group called Project Baker within the Patrol Utility Wing at San
Diego in 1937. The navy’s lack of concern with the above problems probably was
due to the di¤erent landing requirements of seaplanes. Seaplanes did not have to
reliably touch down at a well-defined point on the surface, as did land planes us-
ing runways. Course variations were therefore much less important. The San
Diego location imposed an unexpected diªculty, however. The development
team found that splits had developed in the localizer beam between the College
Park tests and those done in San Diego the following year. Instead of producing
a single navigable course along the approach path, the system was producing two
courses. After months of fruitless modifications to the equipment, the group at-
tached a graphic recorder to the receiver outputs in the test plane to find that the
split course was caused by an interference pattern that surrounded the transmit-
ter. Several more months of letter writing to various specialists eventually caused
Project Baker’s sta¤ to decide that the interference was most likely caused by
mountains more than one hundred miles away, in Mexico. They labeled the prob-
lem “uniformly recurrent multiple course phenomena,” a suitably frightening-
sounding label for a design problem experienced by all systems based upon low-
frequency localizers. Because low frequencies could not be made directional,
interference from any direction distorted the courses. The localizer splits delayed
the navy’s production decision into 1940, despite the advocacy of the comman-
der of the Aircraft Scouting Force, Pacific Fleet, whose aircraft and pilots were be-
ing used for the tests throughout 1938.17
The navy finally accepted the equipment for its seaplane bases after receiving
testimonials from several fleet pilots whom Project Baker had trained. Although
the testimonials were written in support of the system, they also foreshadow one
of the reasons for the navy’s 1943 abandonment of it. Ensigns H. P. Gerdon and
P. H. Craig thought that the long, low approach provided by the glide path beam’s
curvature was less safe than a straight one would be, due to the possibility of ob-
stacles. Gerdon called the low approach a “mental hazard.”18 Mental hazard is an
excellent term for what pilots experienced while using these variants of the NBS’s
system. Knowing that the beam led them at very low altitude over the surface, and
knowing that the beam might not be stable at a given time, place, or weather con-
dition, gave pilots good reason to be less than trusting. A minor change in the
beam angle at such a low altitude could easily prove fatal to a crew. That danger
imposed a mental hazard on pilots which caused them to mistrust the NBS sys-
tem and its YB clone and drove radio engineers to look for ways of stabilizing the
beams.
The instability problem doomed YB from the outset. The navy had not con-
ducted site surveys of its intended installations because its leaders did not un-
derstand the new radio technology well enough to grasp the importance of site
selection. Of its four highest priorities—Reykjavik, Iceland, and Sitka, Kodiak,
and Dutch Harbor, Alaska—only Reykjavik possessed an approach path clear of
nearby mountains. Mountains and mountainous islands in the approach paths
of the Alaska facilities produced far more severe localizer distortions than had
been present in San Diego, and which had not occurred at all in College Park. YB
proved unusable at every Alaska facility where the navy tried to use it after the sys-
tem’s deployment began in 1941. By late 1943, the navy had abandoned the sys-
tem completely.19
Both YB and its twin, the NBS system, su¤ered the same problem. Changing
conditions on the earth’s surface, from soil moisture to topographic features,
caused changes in the glide path. Hills and even trees caused “bumps” in it that
made flying the beam diªcult, particularly for inexperienced pilots. Nearby
mountains caused reflections that made the beams unflyable, while even distant
ones could cause enough interference to be noticeable to pilots.
The instability of the localizer and glide path caused pilots to mistrust the sys-
tem, and they would not use it as a blind landing system. Although I have focused
on the system’s instability, the major component of the NBS/YB system’s failure
70 blind landings
was lack of trust. Landing an aircraft blind required an act of faith from pilots,
and they would not place that faith, and their lives, in a system they knew was un-
stable. Getting pilots to trust the system meant finding a way to evade nature’s in-
consistency.
That engineers would eventually fix the instability problem was not fore-
ordained, however, because the problem had nearly resulted in the NBS system’s
complete abandonment by the Bureau of Air Commerce in 1934. The poor results
achieved by the system during the 1932 and 1933 trials had caused Eugene Vidal,
Director of Air Commerce, to examine an army invention, hoping that it would
be a more acceptable blind landing system. Capt. Albert F. Hegenberger at Wright
Field had developed the army competitor between 1931 and 1933. Hegenberger
was a veteran of the first nonstop flight from the West Coast to Hawaii and be-
cause of that feat was regarded as an expert in radio navigation.20 In late 1933, Vi-
dal asked the Air Corps to send Hegenberger and his system to Newark for test-
ing by the Bureau of Air Commerce.
Hegenberger’s system, formally named the A-1 system by the army but more
commonly called the Hegenberger system, relied upon a radio compass to pro-
vide directional information to the pilot. Radio compasses, often called radio di-
rection finders and now referred to as automatic direction finders, were a deriva-
tive of World War I radio research. Wartime demands for improved navigation
had inspired radio engineers and radio physicists to investigate methods of lo-
cating radio transmissions, and one result was the deployment of a European net-
work of radio direction finding stations that served as the primary means of
navigation for aircraft.21 These ground stations detected an aircraft’s radio trans-
missions, triangulated with other stations, and then reported the plane’s position
to its pilot. What the U.S. Army Air Corps wanted, however, was a system that did
the inverse. It wanted a radio station on the ground to activate an instrument in
the airplane that the pilot could follow.
In 1931, the Air Corps hired Geo¤rey Kreusi from Western Air Express. While
at the airline, he had begun working on an airborne radio direction finder that
the Air Corps leaders thought showed promise. It relied primarily upon a crossed-
loop antenna on the aircraft’s nose or belly to receive a radio transmission. Any
radio transmission would suªce, so pilots could easily tune the device to com-
mercial radio stations and use them as navigation aids if they wished.22 The di-
rection was accurate to within a few degrees—quite excellent by contemporary
air navigation standards.
Kreusi’s e¤orts at the Air Corps’s Wright Field development center had suc-
ceeded by 1933, and Hegenberger decided to base a blind landing system on one.
He used two low-frequency, omnidirectional radio transmitters that he called

compass locators to provide navigational references for the compass. Pilots were to
fly between them using a specific pattern, shown in Figure 3.6. The two locator
stations, which were truck-mounted, were to be placed just outside the airfield
boundary (station A on the diagram) and 1.5 miles out (station B). Pilots set their
radio receiver for station A’s frequency and used the radio compass to fly toward
it. Once the plane had passed over the inner station (indicated by the compass
needle’s sudden swing to point in the opposite direction), a pilot was to make a
180-degree turn, reset the receiver to station B’s frequency, and fly to B. After con-
tinuing past B for about three more miles, pilots were to execute a standard rate
turn and line up on B again, which they were to fly over at five hundred feet. The
radio receiver was then retuned to A’s frequency. Descending at four hundred feet
per minute, pilots were supposed to fly over A at 150 feet and maintain the four
hundred feet per minute descent until they touched down. The descent was mea-
sured with the Kollsman sensitive altimeter.
The Bureau of Air Commerce’s testing of the Hegenberger system took place
in early 1934, in the midst of what aviation historians call the airmail crisis. Eu-
gene Vidal asked Foulois to send Hegenberger and his system to Newark to be
tested alongside the NBS system installed there in February. In April, Eastern Air
Transport’s Cutrell flew the system, as he had the NBS system tests that January.
He and his safety pilot completed thirty-seven landings in ninety-one attempts, a
much better showing than they had achieved on the NBS system. On the strength
of those results, and given the political situation, Vidal chose the army system. In
June, he wrote to Foulois that the Bureau of Air Commerce wanted to adopt it as
the civil standard. “Our investigation indicated clearly the superiority of the U.S.
Army Air Corps system over the so called Bureau of Standards system for gen-
eral adaptation to commercial aircraft,” he wrote. His decision was based upon a
report submitted by Major Snow, head of the blind landing program. Snow con-
sidered the tests conclusive and recommended that “any further tests of any blind
landing system utilizing radio beams . . . be discontinued as wrong in principle
and practice and of unwarranted expense in view of their past failure and slight
future chance of success.”23 The army’s Hegenberger, or A-1, system thus became
the first oªcially sanctioned blind landing system in the United States.
Snow was certainly biased toward his own system, but his analysis was not
wrong. The NBS system had not worked well in its oªcial tests. He was correct
that blind landings were impracticable using radio beams, and decades later that
has not changed. Radio beams in that frequency range were and are not stable
enough to permit safe blind landings. But the commercial pilots who were invited
Figure 3.6. The Hegenberger, or A-1, approach procedure. This pattern took 1930s aircraft twenty minutes to fly, and because the aircraft was
required to overfly the airfield during the approach, only one aircraft could be on approach at a time. This produced a very low landing rate.
National Air and Space Museum, Smithsonian Institution (SI 98-15788).
to fly the A-1 system during 1934 believed that it was, in fact, inferior to the NBS
system on which they had been practicing unoªcially for the preceding year.
They and their superiors were quick to take exception to Vidal’s decision, and they
inundated the bureau with letters explaining in great detail why they felt this way.
J. R. Cunningham, Superintendent of Communications of United Air Lines,
wrote in October that his pilots did not “consider the sensitive altimeter depend-
able to the limits required for landing.” In November, C. C. Shangraw of Ameri-
can Air Lines wrote to Major Snow that A-1 left a pilot “approximately 200⬘ in the
air over the inner guiding station and headed toward the field or runway . . . [hav-
ing] no definite assurance of his height above ground, which obviously means
that he has to feel his way down which required a longer runway or larger field
than is normally available in commercial transport operation.” Cunningham fol-
lowed his October letter the following month with a detailed list of critiques that
appear to summarize several of his pilots’ opinions. The Kruesi compass per-
mitted the plane to drift from the proper approach path after passing over beacon
B, and in a blind landing the pilot had no way to detect that drift. And because
the sensitive altimeter was typically o¤ by twenty feet or so (and he reported de-
viations of up to one hundred feet in some aircraft), it was possible to land short
of the field, or to land too far down field to stop before running out of space. Lack
of a stable vertical indication, then, meant that the Hegenberger system did not
define an adequate “point of contact” on the field. Several other pilots wrote the
bureau with similar misgivings about the system’s ability to provide safe blind
landings at commercial fields.24
Clearly, the Air Corps and the commercial airlines had very di¤erent ideas
about what worked as a landing system. That disagreement derives at least in part
from the di¤erences in their operating environments. The Air Corps in 1934 still
relied overwhelmingly on all-over turf fields for its flying, for which an accurate
point of contact was unnecessary. The corps also used smaller, lighter aircraft,
mostly biplanes. The airlines had already junked such planes for faster, heavier
all-metal monoplanes—Ford Trimotors, Boeing 247s, and DC-3s—and had de-
manded the adoption of surfaced runways to support all-weather operation of
these relative monsters. One consequence of the airlines’ adoption of surfaced
runways is that landing on one required more precise guidance than the A-1 could
give. What worked for the army’s all-over fields thus did not work for the airlines
and their runways.
The airlines were successful at getting Vidal’s decision reversed, and the NBS
abandoned the A-1 before installing it. Their e¤orts were helped by the Air Corps’
excessive zeal to get the system into production. The corps delivered eighteen sets
74 blind landings
of ground equipment to the bureau, and all of them proved faulty. The bureau’s
oªcials vacillated for much of 1935 over whether to ask for money to fix them or
to drop the whole idea but in the end decided to do what the airlines wanted. They
let the program die. The Air Corps leaders saw that decision as a betrayal, com-
plicating the relations between the two organizations right through World War II.
The airlines’ reaction to the A-1 decision, in retrospect, was excessively nega-
tive. They were certainly correct that the A-1 was not an acceptable blind landing
system, and during World War II the army air forces supported the development
of several more systems in hopes of producing a more acceptable one. Yet the sys-
tem’s radio compass was an excellent navigational aid for finding an airport. The
Bureau of Air Commerce’s four-course ranges were e¤ective en route navigation
tools, but because their courses were not aligned with airport runways, they pro-
vided no help in making an approach and landing. Once o¤ the range, pilots had
to navigate to airfields by landmark and magnetic compass. Magnetic compasses
were unreliable for that because of the aircraft’s vertical motion during a turn.
Landmarks, of course, were diªcult to see through clouds, and in bad weather
pilots had to risk flying close to the ground to find their way to a field. Some of
the pilots who wrote to protest the A-1 system’s adoption of a landing aid made a
distinction that became very important later on: the radio compass was, they be-
lieved, an excellent approach aid because it made getting onto an approach course
easy.25 After World War II, the Air Line Pilots Association insisted that the com-
pass locator stations be added to the NBS system’s marker beacons for precisely
this reason. Excessive devotion to blind landing in 1934 and 1935 thus drove the
airlines and their pilots to reject a good approach aid because it was a poor land-
ing aid.
The idea that Diamond and Dunmore had presented, then, a virtual path from
sky to ground was so compelling that the commercial airlines and their pilots im-
mediately adopted it as the best blind landing solution despite the minor prob-
lem that it did not work—yet. That it could be made to work, however, was an un-
questioned belief. On that basis, the airlines managed to suppress the army’s A-1
system, and United Air Lines took over the Oakland NBS installation and began
working to fix the system’s instability problems.
Recognizing that the NBS possessed most of the technical expertise on the sys-
tem, United Air Lines’ Cunningham approached the National Bureau of Stan-
dards and the Bureau of Air Commerce about obtaining their assistance in over-
coming the system’s problems in May 1935. He had made an arrangement with
Eclipse Aviation, a subsidiary of Bendix Aviation, and he wanted to borrow some
of the organizations’ engineers to aid the two companies’ development work. He
agreed in advance to pay their usual salaries, making the deal no-cost to the gov-
ernment. The navy initially opposed the plan, with several oªcers concerned that
the project would compromise the security of the YB system. Indeed, the secre-
tary of the navy as early as 1933 had voiced national security concerns in his con-
gratulatory letter to the secretary of commerce, and rightly so. The Soviet Union,
Japan, France, Germany, and the United Kingdom all copied the NBS system, and
Germany used the localizer as the basis for a “blind bombing” aid during World
War II. But Adm. Ernest J. King, then chief of the Bureau of Aeronautics, con-
vinced Adm. William H. Standley, the chief of naval operations, that not sup-
porting the proposal would place the navy “in a position of encouraging the
establishment of a monopoly” by making the Air-Track Corporation the sole man-
ufacturer of a system that had, after all, originally been developed by a civilian
agency for commercial purposes.26 After nearly a year of delay, therefore, the al-
liance went ahead.
When United took over the NBS work in 1934, it immediately adopted a UHF
localizer, primarily to reduce the system’s cost. If localizer and glide path oper-
ated at the same frequency, only one transmitter and one receiver was necessary.
That would reduce the ground installation’s cost and the amount of weight that
had to be carried by aircraft equipped to use the system. The choice of UHF for
the localizer proved to be a good one, as we have already seen through the Lorenz
case. It did nothing to help the already UHF glide path, however, and with its part-
ner, Bendix Radio, United conducted an extensive investigation into the glide path
stability problem. Primarily, the two companies focused on experimenting with
antenna designs and with polarization. They found that a glide path using hori-
zontal polarization was less a¤ected by ground conditions than one relying upon
vertical polarization, but simply changing the polarity was not enough to com-
pletely resolve the instability problem. They also abandoned the equal intensity
glide path for an equi-signal type, in which the strengths of two di¤erently mod-
ulated signals were compared. United Air Lines’ engineers did this to help pre-
vent changes in receiver sensitivity from changing the indicated glide path, thus
eliminating another source of unreliability.27 Yet neither they nor anyone else
managed to completely eliminate glide path instability without shifting to mi-
crowave frequencies.
If the glide path could not be made completely stable, it could not be safely
used for blind landings. But United’s managers realized that truly blind condi-
tions were very rare. They occurred less than 1 percent of the time, while ceilings
below five hundred feet, the existing legal minimum, occurred between 5 and 10
percent of the time.28 If one could produce a system that would permit safely cut-
76 blind landings
ting the minimum in half, well over half of all visibility-related delays could be
eliminated. So United reduced its goals. It did not use the system for blind land-
ings. Instead, it was satisfied to use the equipment as an approach aid, with land-
ing to be done visually. United expected that the system would pay for itself in a
few years by allowing completion of flights that would otherwise have to be can-
celed or diverted. Careful site surveys and equipment installation, combined with
extensive pilot training, made this goal both reasonable and achievable.
Essentially, United redefined the problem to be solved as one of establishing
an instrument approach system rather than a landing system. Without the de-
mand for blind landing, one could reduce the twin needs of stability and preci-
sion. Precise electronic navigation was unnecessary if pilots were going to land
visually. A system that could direct pilots to within a few degrees of the runway
course was far better than nothing, and the remaining instability in the glide path
was not severe enough to jeopardize aircraft that did not rely on it below two hun-
dred feet. Safety would still be preserved, while most of the cost of weather delays
would be avoided. This was, in fact, the ultimate “resolution” of the instability
problem, but the Bureau of Air Commerce, the navy, and, to a lesser extent, the
army required several more years of experience with failure to accept United’s
rather pragmatic attitude.
The United-Bendix system, therefore, was the most successful of the first gen-
eration of descendants of the NBS system, not because it was technically supe-
rior but because the organization using it recognized its limitations. It was in-
stalled at eight United fields in the United States, and between 1936 and 1937
more than three thousand landings were made with it “under the hood.” It did
not become more widespread because the equipment was expensive (around
$10,000 per installation) and because there was no nationwide oªcial standard.
Airport managers waited for federal money to pay for these installations, and that,
in turn, meant waiting for federal agencies to agree on which blind landing sys-
tem to buy.
The attention given to the NBS research had not been exclusively American.
While Japan and the Soviet Union hired U.S. engineers to duplicate the NBS sys-
tem for them, in Germany the localizer–glide path–marker beacon idea, broad-
cast there by a translation of Diamond and Dunmore’s 1930 article, had inspired
radio engineer Ernst Kramar to develop his own version. The result, known as
the Lorenz system after the eventual manufacturer, Carl Lorenz A. G. of Berlin,
emerged from a joint project conducted by Lorenz, Telefunken, and the German
aeronautical research laboratory DVL in 1932. The Lorenz system operated on
33.3 MHz, a frequency not available in the United States, where it was used by po-
lice and forest-fire patrols. It was markedly di¤erent than the NBS system in that
it combined localizer and glide path into a single unit by alternately exciting two
reflectors on either side of a vertical antenna with dots on one side, dashes on the
other. This produced two elliptical patterns in space, with the major axes parallel
to the on-course heading. The result was an equi-signal zone much like that pro-
duced by the NBS localizer. The glide path was in some ways a parasite: instead
of producing one separately, a constant intensity line was simply selected from
the existing “localizer” field pattern by the receiver. The United-Bendix team al-
most certainly got this idea from the Lorenz system. The system also employed
two marker beacons operating on 38 MHz, with modulations of 700 Hz and
1,700 Hz to distinguish them.29
The company began testing the system in late 1932 at Templehof and began
marketing it in 1935. The U.S. Army Signal Corps found that by 1936 installa-
tions were being made at Berlin, Frankfurt/Main, Hanover, Cologne, Hamburg,
Konigsberg, Leipzig, Stuttgart, Danzig, Breslau, Munich, Nurnberg and Stettin
in Germany, at Zurich, Switzerland, and Vienna, Austria, at Le Bourget in Paris,
and at London’s Heston aerodrome. Plans were being made for installations at
Prague and Moscow, and one unit was shipped to Japan.30
Despite its rapid adoption in Europe, the Lorenz system shared the faults of
its ancestor. In response to questioning by the U.S. assistant naval attaché for air
in Paris, the director of Le Bourget aerodrome stated that “the curved beam did
not maintain a definite course and shape due to meteorological conditions and
building and hangar interference.” The Lorenz company had such diªculty with
its glide path that in a visit to the chief signal oªcer in February 1936, Carl Lorenz
stated that his system did not include a glide path because it was still “experi-
mental,” despite his own sales literature and the testimony of Le Bourget’s di-
rector.31 As a result of the beam instability, pilots flew the Lorenz installations
made during the mid-1930s with their glide path receivers disconnected, allow-
ing them to use the localizer and marker beacons as a landing aid, but subject to
the same lack of altitude indication that inspired commercial dislike of the A-1
system in the United States.
Even without the glide path, European pilots benefited from the Lorenz sys-
tem, which simplified the basic task of finding the airport. It was especially use-
ful given European air navigation practice. Instead of radio ranges for point-to-
point navigation, European pilots relied on radio direction finding stations on the
ground for position information. These stations triangulated on the plane’s radio
transmission, and then sent the resulting fix to the pilot, who then had to plot it
on a map.32 This was fine for en route navigation and in some ways was superior
78 blind landings
to the U.S. method, which provided no guidance for pilots flying outside the es-
tablished airways, but it was much too slow for an approach. The procedure could
put aircraft within the Lorenz system’s beam, however, making the two an e¤ec-
tive combination for making “instrument low approaches,” although it was still
not a safe blind landing system.
From the standpoint of the stability problem, the Lorenz system, in a sense,
simultaneously represented a step forward and a step backward. Its UHF local-
izer was much less susceptible to atmospheric and long-range interference phe-
nomena than the NBS system’s low-frequency localizer had been. At the same
time, it was more susceptible to local reflections (inelegantly called multipath
e¤ects). Local interferences, however, proved more manageable in the long run.
Perhaps based on the success of the Lorenz system, UHF was quickly selected by
U.S. radio engineers as the right choice for localizers because one could, by care-
ful siting, prevent localizer refiections. Often that involved physical modification
of airport surfaces, via movement of hangars, landscaping, and rerouting of traf-
fic. But this was a far easier problem to solve than the “multiple recurrent course
phenomena” the navy had run into in San Diego with its low-frequency localizer.
That could only have been fixed by removing the o¤ending mountains in Mex-
ico, a far more substantial undertaking than merely moving a hangar or two.
Hence, the UHF localizer required careful siting and perhaps some rearrange-
ment of the local landscape, but nothing drastic. Protecting a UHF localizer from
local interference was simply easier than protecting a low-frequency one from
long-range interference.
By 1937, then, radio engineers and aviators in the United States and Europe had
largely accepted that higher frequencies were better for landing aids. Environment-
induced instabilities were more tractable with higher frequencies because they
were local, and Chapter 4 examines engineers’ attempts to move blind landing
into the microwave realm. In that frequency range, they hoped to detach the sig-
nals from the earth’s surface completely, making the environment irrelevant. Yet
there was another way out of the stability trap, as the foregoing suggests: one
could give up the notion of blind landings entirely. United already had, as had the
European airlines. The United-Bendix, Lorenz, and A-1 systems were capable of
producing reliable instrument approaches had U.S. authorities been willing to
take a more pragmatic approach. The record shows, however, that neither the
Army Air Corps nor the Bureau of Air Commerce were ready to give up. They still
wanted blind landings.
conclusion
The blind landing system pioneered by the National Bureau of Standards’ Ra-
dio Section proved to have severe stability problems that led, in part, to its can-
cellation in favor of the army’s A-1 system. Yet the mismatch between the capa-
bilities of the A-1 system and the operating environment the airlines functioned
within caused the airlines to resist the A-1. It simply was not and, as far as airline
pilots could see, could not be made accurate enough to permit safe blind land-
ings at most airports. The airlines’ advocacy of the NBS system earned it a re-
prieve. The system’s problems still had to be fixed, however, and that need drove
organizations to seek more solutions.
That there was even a choice to be made descended from the NBS system’s in-
teraction with nature. If the original NBS system had been stable, there is little
doubt that it would have been deployed beginning in 1933, and the army’s chal-
lenger would not have been adopted. Nature thus posed a significant constraint
on the evolution of blind landing systems. The need for some sort of system was
a product of the collision between commercial aviation’s demand for regularity
and the earth’s refusal to be consistent. That very inconsistency prevented the sys-
tem and its UHF descendants from achieving the goal established for them by
their inventors and institutional supporters: regular blind landings. That con-
straint, however, also motivated researchers to try new technologies and tech-
niques in pursuit of the original goal. Nature’s inconsistency constrained and
drove the evolution of blind landing systems.
During the 1930s, therefore, engineers pursued several possible solutions to
the beam instability problem, while policy makers sought to define an acceptable
beam pattern. These two e¤orts a¤ected one another. As the ability to generate
straighter, more stable beams evolved, achieving agreement on a technical stan-
dard became increasingly diªcult.
chapter four
The Promise of Microwaves
Commercial pilots’ rejection of the Hegenberger system left a schism in the

aviation community over how best to achieve blind landings. The Air Corps was
satisfied with its radio compass approach, while commercial pilots wanted some
version of the National Bureau of Standards’ (NBS) tripartite system. In 1934, the
community had not agreed on how a blind landing system should work.
Between 1934 and 1939, however, the Air Corps moved closer to the airlines’
position, agreeing by 1939 that the three-part NBS model was, indeed, the best
route to a solution. The remaining dissension was over details: how long should
the glide path be? Must it be straight? What frequencies should it use? With con-
sensus around a particular model of the solution achieved, the various organiza-
tions in the aviation community had common ground to negotiate over. Never-
theless, determining the system’s details proved diªcult.
The most vexing detail was the shape of the glide path beam. The Air Corps
insisted on a straight glide slope, while airlines were satisfied with the “bent
beam” idea and pushed for rapid deployment of such a system. The Air Corps, in
turn, demanded a new system that relied on microwave radio, which was being
designed in a unique collaboration between Civil Aeronautics Authority (CAA),
the Massachusetts Institute of Technology (MIT), the Air Corps, Stanford Uni-
versity, and Sperry Gyroscope. The microwave system promised to provide the
long, straight glide path that Air Corps oªcers believed was necessary, and they
sought a way of delaying the airlines’ drive for the bent-beam system. Gen. Henry
“Hap” Arnold thus deployed the classic Washington delaying tactic—he re-
quested an independent review of the various blind landing programs. In e¤ect,
the promise of microwaves 81
this placed the decision in the hands of President Roosevelt, who turned it over
to the National Academy of Sciences. Their recommendations would become pol-
icy until U.S. entry into World War II changed everything.
In the only existing history of this subject, William Leary has argued that
Arnold’s review was an “unnecessary complication” to the adoption process.1 But
Arnold had good reason, in the form of a new microwave-based system being de-
veloped for the Army Air Corps. As we have seen, the descendants of Diamond
and Dunmore’s system had serious stability problems, and Arnold believed the
new microwave solution, not the troubled CAA system, was the correct choice.
There is a tinge of bitterness in Leary’s argument because Arnold’s review had
the e¤ect of preventing the adoption of any commercial blind landing system un-
til well after the war, costing the airlines money—and costing pilots’ and pas-
sengers’ lives. Arnold held a vision of a microwave future free of stability prob-
lems. The National Academy’s review panel agreed with him.
defining a “standard” system:

the radio technical committee for aeronautics
In 1936, the Bureau of Air Commerce formed a committee to help it deal with
problems it had found in the use of radio technologies in aviation. Called the Ra-
dio Technical Committee on Aeronautics (RTCA), the group was composed of se-
nior people from the technical departments of aviation-related organizations. The
Army Air Corps and Navy Bureau of Aeronautics each had two representatives,
as did the Bureau of Air Commerce. The Federal Communications Commission
and State Department each provided one member. From outside the federal gov-
ernment came representatives from the Air Transport Association, the trade or-
ganization representing the scheduled airlines, and the Air Line Pilots Associa-
tion, the labor union representing commercial pilots. The chair of the committee
was elected and tended to rotate among the group. Private, or general, aviation
was not represented until after World War II.
One important note is the RTCA’s legal basis of authority. It had none. The
Bureau of Air Commerce had not sought, or received, blessing from Congress in
the form of a law, or from the president in the shape of an executive order. The
bureau’s hope had been that by bringing together the full range of aviation in-
terests from inside and outside government, whatever agreement was reached
would have what Osmun has called the “authority of agreement.”2 The reason-
ing went that if everyone agreed to a standard, that standard would hold moral
authority, even in the absence of legal authority. And absence of legal authority is
82 blind landings
precisely what existed: the Bureau of Air Commerce could recommend standards
for airport equipment, but it had no enforcement power, and it was legally barred
from funding airport improvements. It could not buy equipment to install at air-
ports, blocking that possible route to establishing a de facto standard.
In addition to the RTCA’s lack of legal authority, no other mechanism existed
to coordinate aviation standards among Army, Navy, and Commerce Depart-
ments. The RTCA provided a forum for discussion of radio issues and for the ne-
gotiation of proposed standards, but for a standard to succeed, each member had
to convince his (there were no women on the committee in the first twenty years
of the RTCA’s existence) parent organization to adopt the proposed standard.
That proved diªcult. Nevertheless, by providing a forum for discussion of needs
and possibilities among various groups, the RTCA served an important role in
the development of blind landing systems by establishing minimum criteria and,
more importantly, by defining problems that still needed to be solved.
The Bureau of Air Commerce began to revisit the blind landing system ques-
tion in 1936, after the airlines refused to use the Hegenberger system. It rein-
stalled the Newark equipment at Indianapolis and began once again to investi-
gate the Bureau of Standards system as a possible solution to the blind landing
problem. The rapid advancements made in the state of the art by other organi-
zations, particularly the United-Bendix partnership, however, meant that the
Newark installation was obsolete. This fact and the growing plethora of derivative
systems led the bureau to ask the RTCA to examine all of the existing blind land-
ing systems and, based on that, recommend a suitable standard. In this way, the
bureau intended to capitalize on the knowledge gained and innovations made
over the preceding few years by other organizations, without committing itself to
any one company’s design.
The RTCA’s Subcommittee on Instrument Landing Devices, therefore, visited
College Park, Maryland, to examine the Washington Institute of Technology’s
Air-Track system (the civil name for the navy’s YB system), Kansas City, where
United demonstrated the UAL-Bendix system, and finally Indianapolis, where on
May 14 and 15, 1937, it tested the Bureau of Air Commerce’s slightly modified sys-
tem. More importantly, the Indianapolis visit also resulted in an extended testing
of the Lorenz system, imported from Germany by International Telephone and
Telegraph (IT&T), the parent company of Lorenz A. G. This first hands-on review
of the only European instrument landing system brought extensive press atten-
tion, as well as detailed scrutiny by the Army Signal Corps.
The Signal Corps’ representatives, Capt. George V. Holloman and Maj. F. S.
Borum, argued that the curved, constant intensity flight path produced by the sys-
tem required pilots to constantly change the aircraft’s attitude and throttle setting
during landing, which was in opposition to army training.3 The army trained its
pilots to land by maintaining the aircraft in a level attitude and slowly closing the
throttles to maintain a powered glide into the airfield. This reduced the pilot’s
workload and was relatively easy to train pilots to do.
Pilot training was not the only problem the army foresaw in the curved glide
path. The curvature placed aircraft at low altitudes (around fifty feet) at a half-mile
from airport boundaries. The army believed that few of its airfields, and even
fewer commercial airports, were suªciently clear of obstacles to make this a safe
approach altitude. The army’s standard for obstacle clearance was to ensure a
clear 20:1 slope for aircraft approaches, which was approximately a 2.8-degree
glide. The Lorenz system provided a 1-degree glide from a half mile out, with the
curvature increasing rapidly as distance from the airport increased. Beyond two
miles out, pilots considered the path unflyably steep. They therefore had to in-
tercept the glide path at roughly 700 feet at the outer marker 1.9 miles from the
airfield, descend rapidly to fifty feet, and then level out into the final portion of
the approach.
The two issues of pilot training and obstacle clearance mentioned in the two
pilots’ report was likely informed by a third, unstated, concern. Although the
16,000- to 25,000-pound aircraft commonly used by the air transport industry
could negotiate the curved glide path in the hands of a skilled pilot, the army had
much larger planes on its drawing boards. In fact, the 55,000-pound Boeing
B-17 prototype had begun flying in 1936, and by 1937 Air Corps leaders were be-
ginning to realize that aircraft of such size were not maneuverable enough to fol-
low either this curved glide path, or the very short overall approach that the
Lorenz—and to be fair, all other existing systems—provided. As a result, the di-
rector of the Aircraft Radio Laboratory, Col. John O. Mauborgne, added an en-
dorsement to Holloman’s report suggesting that the army needed to design, or
have designed for it, a straight glide path.4
The two oªcers’ criticisms were certainly valid. The training issue is diªcult
to assess, but the “army way” to land a plane became the “right way” after World
War II, despite the army’s complete replacement of its outdated instrument flying
training program by a commercial one in 1941.5 The army could have trained its
pilots to use a curved glide path, but it rightly perceived no benefit in doing so.
The more significant issues were obstacle clearance and aircraft size. Obsta-
cles in airport approaches were one of the pilots’ union’s biggest heartaches, and
its president agitated in Congress to get that body to grant the Bureau of Air Com-
merce and its successor, the Civil Aeronautics Authority, the authority to regulate
84 blind landings
airport approaches. In 1937, in fact, pilots for Eastern Airlines and American Air-
lines refused to fly DC-2 and DC-3 aircraft into Washington-Hoover airport, cit-
ing the dangers posed by smokestacks, power lines, radio towers, and a highway
that ran through the middle of the field. They were joined by Eleanor Roosevelt,
who said she feared for her friends’ lives in using the field, and a photo essay in
the Washington Times showing the o¤ending structures. The union also listed
Fort Worth, Wichita, Chicago, San Francisco, Newark, Fresno, and Bakersfield as
cities with seriously obstructed airfields. The Lorenz “bent beam” system would
certainly not have been safe at any of these cities, until the municipalities agreed
to remove the obstructions or move the airport.6 The army’s oªcers were correct
in their belief that the system would not serve many airports without substantial
modification of the airports’ surroundings.
It is important to note that the Lorenz system was not the only curved glide
path system. All systems that had been based on the Bureau of Standards work
generated such a path and at roughly the same angle of inclination. Commercial
operators had considered this desirable at first because following it reduced the
descent rate of the aircraft as the plane approached touchdown, thereby reducing
impact. When it had tested them in 1935 and 1936, the army had criticized this
aspect of all these systems in its internal reports (which were apparently released
to the airlines), but because the army was still clearly wedded to the Hegenberger
system, which included no glide path at all, no one seems to have taken the army’s
criticism seriously. Colonel Mauborgne’s suggestion that the army campaign for
a straight glide path is the first indication that someone with significant author-
ity in the army at least recognized the unsuitability of the army’s own A-1 system.
It also presaged a long-running argument between CAA, which wanted to deploy
a bent-beam-type system, and the army, which demanded a straight glide path.
Following the demonstrations at Indianapolis, United Airlines hosted a meet-
ing between representatives of the other major airlines, the FCC, the Bureau of
Air Commerce, and Bendix Radio at its Chicago Oªce on June 23, 1937. At that
meeting, a standard was agreed upon that was validated by the RTCA that Sep-
tember. Although it did not require a straight glide path, it stated that “study
should be made of the possibility for obtaining a straight line constant rate of de-
scent glide path.” The airlines had also recognized the eventual need for a straight
path, which might well have been based on their own expectations of much larger
aircraft appearing in the near future. Paul Goldsborough, president of Aeronau-
tical Radio, Inc. (ARINC), a company established by the airlines to make radio
equipment especially for airline use, forwarded a copy of this proposed standard
to the Air Corps headquarters. The proposed standard was published in the Au-
gust 1937 newsletter of the Air Line Pilots Association, accompanied by a state-
ment by R. T. Freng that “personally, I feel that this is the thing that we have been
looking for a long time.”7 It was approved by the RTCA and published in the new
Civil Aeronautics Authority’s first technical report that October, along with a rec-
ommendation that CAA pursue development of the items listed under “Projected
Developments,” which included the straight-line glide path.
By late 1937, then, the Air Corps and commercial aviation groups had recog-
nized that a straight glide path o¤ered advantages to the curved one that all pre-
vious versions of the Bureau of Standards’ system had generated. These advan-
tages included a simpler landing procedure, greater length and consequently a
longer approach from a higher altitude, and better obstacle clearance. But the abil-
ity to produce a straight path did not exist, and airlines believed that it was in their
interests to adopt a curved path while a straight one was developed. It was over
this point that the army disagreed, believing that waiting was better.
The response to the proposed standard within the Air Corps was mixed. The
advantages of the proposed equipment were well understood, but criticism fo-
cused on the cost of the system. In a letter to the secretary of commerce, the
secretary of war (very politely) pointed out that the shift to UHF frequencies that
the standard imposed left Air Corps aircraft unable to use the system. Internal
memos were much less polite, pointing out that the Air Corps, with more than
3,000 aircraft, faced a very expensive procurement program to outfit its planes in
order to keep up with the Bureau of Air Commerce’s improvements to the Fed-
eral Airways System. Not lost on the army men was the reality that the bureau
was considering the needs of only about 300 aircraft operated by the commercial
airlines, whose equipment the bureau did not have to pay for.8 Hoping that in two
or three years a straight path would be developed that might not be UHF, the Air
Corps considered a temporary system unpalatable.
The major airlines were not only willing to pay for the receivers (which cost
about $1,500 per aircraft) but were practically demanding the opportunity to do
so. Despite believing that the equipment they wanted to buy in 1938 would be ob-
solete in at most four years due to the projected movement to a microwave sys-
tem, the airlines were willing to install it. They wanted an instrument landing sys-
tem to improve their profitability because delayed and canceled flights cost them
a great deal of money. The airline demand for regularity that had led to surfaced
runways also drove them to push for an instrument landing system. Airlines ex-
pected to make back the cost of receivers within a year of operations, as the addi-
86 blind landings
tional income they expected to derive from use of the system o¤set the initial cost.
Hence, the airlines’ economic model indicated that rapid adoption of a “good
enough” system was a better strategy than waiting for perfection.
The airlines therefore pushed the Bureau of Air Commerce to let a contract to
someone to design a system that conformed to the published standard. In early
1938, the bureau awarded a contract to International Telephone Development
Corporation (ITD), a subsidiary of International Telephone and Telegraph, to
build a new system intended to meet the RTCA standard. The company’s system
was based upon two key patents held by one of its engineers. It consisted of a 110
MHz equi-signal localizer, a 93 MHz constant-intensity (curved) glide path, and
two marker beacons. The installation at Indianapolis included two localizer trans-
mitters and four glide path transmitters to allow approaches to be made from any
of the four runway directions available at the field without having to move equip-
ment. It employed a specialized array of five horizontal loop antennae for the lo-
calizer, which allowed variation of the course width by changing the amount of
energy radiated by each of the loops. This invention was an important step in over-
coming the problem of bent or split courses, because careful tuning of the sys-
tem, combined with careful siting on the airfield surface, could provide a narrow
enough beam pattern to avoid obstacles. The other significant patent was a trans-
mitter bridge circuit that automatically compensated for aging of system com-
ponents.9 This allowed the system to maintain a constant output automatically,
eliminating another source of system instability. Neither innovation dealt with
the issue of glide-path curvature, however. As initially demonstrated in 1939, the
Indianapolis system possessed the same two to three mile constant intensity
curved glide path as its predecessor.
In September, the Radio Technical Commission for Aeronautics’ Committee
on Instrument Landing Devices met to formally evaluate the International Tele-
phone Development Corporation’s system at Indianapolis and the MIT system.
Forty-six representatives of various organizations, including six from CAA, two
from the Army Air Corps, and three from the Army Signal Corps met on Sep-
tember 13, 1939, at the CAA Experiment Station in Indianapolis. Six airlines were
represented, as were MIT, Sperry Gyroscope, Bendix Radio, Bell Labs, Aeronau-
tical Radio Inc., RCA, the International Telephone Development Corporation, the
Washington Institute of Technology, and the Federal Communications Commis-
sion. Although there was no formally appointed member of the Air Line Pilots
Association listed among the representatives, five commercial pilots were pres-
ent, flew the system, and submitted an addendum to the report o¤ering specific
recommendations for modifications. E. A. Cutrell, who had participated in the
1934 tests of the A-1 and NBS systems, was one of these, as was the chief pilot for
United Air Lines, R. T. Freng. The senior Air Corps representative was Maj. A. W.
Marriner, of the Air Corps Communications Department in the Materials Divi-
sion; the senior Signal Corps representative was Colonel Mitchell, director of the
Aircraft Radio Laboratory. The organizer was Richard Gazely, head of the Tech-
nical Development Division of the CAA, and the chair was J. R. Cunningham, di-
rector of communications for United.10
The report of this committee indicates that Colonel Mitchell was active in sup-
porting the army’s interests, and in fact the chair requested his recommendation
on the form in which the proposed standard was to be cast. Mitchell recom-
mended that the standard be “left suªciently broad to admit of many approaches
to the problem, that is, that a performance specification only should be written.”11
The group unanimously agreed that a three-element system be chosen (marker
beacons, glide path, and localizer, the elements of the Bureau of Standards’ con-
cept), and Mitchell suggested that the standard be written so as not to rule out use
of a “heading device,” by which he obviously meant a radio compass. The group
again agreed and placed that device in the “projected developments” category as
desirable, but not necessary.
The vital question of glide path shape, however, proved controversial. Preston
Bassett of Sperry Gyroscope Company reported in an internal memo that “the
meeting devolved into a wide variance of opinions, one section of which insisted
on the adoption of the IT&T system for the next step, another group insisting that
the system was impossible, and no blind landings could ever be made with it.”
Bassett also reported that there was a “general feeling among many of the pilots
and operators that the bent beam was not the final solution.”12 Ultimately, the
group chose to defer the decision to the pilots who were then out flying the sys-
tem. The pilots’ addendum to the report, accepted by the group as the commit-
tee’s recommendation, specified that “a glide path intersection shall be obtained
at 1500⬘ altitude at a distance of 6 miles from the transmitter end of the runway
. . . At the point of contact the glide path shall have an angle with the runway not
less than 1 degree and not more than 2 degrees. The glide path shall pass through
the following points—not less than 500⬘ nor more than 700⬘ altitude at a dis-
tance of 3 miles from far end of runway, and 1500⬘ altitude at 6 miles.”13
The commercial pilots had adopted an “almost straight” glide path of six miles
length. This was well in excess of the two to three mile glide path that the Indi-
anapolis system had achieved. Once again, the group placed the straight glide
path in the “desirable developments” section of the standard.
Although the army members of the committee did not oppose the adoption of
88 blind landings
this standard, Mitchell was apparently not satisfied with it. He commented, in yet
another attached addendum, that the Indianapolis system had merit as an ex-
perimental apparatus, especially since its ability to produce a straight glide path
and a curved one would finally allow determination of which was safer. He
thought that the straight glide path was too short and that the system should be
extensively flown before final adoption. The army needed a ten to fifteen mile ap-
proach, instead of six miles. He obviously had in mind the Air Corps’ new bombers,
which were more than twice as heavy as the biggest commercial transport in
service. He emphasized the need for a common CAA/army/navy standard and
pointed out that the weight of the equipment made it unsuited to the army’s
smaller aircraft (which were, of course, the majority of army aircraft).14
It is important to note that the CAA does not seem to have intended the Indi-
anapolis system to be the ultimate system. The organizer and head of the CAA’s
research division, Richard Gazely, pointed out, “The only really worthwhile opin-
ion we believe is one based on extensive pilot experience. Only after 500 or more
pilots have had a chance to observe the behaviour of the equipment under every
condition encountered in routine operations and only after these pilots tell us
that they would trust the lives of their passengers and their companies’ equip-
ment to the guidance which a system gives them will we permit ourselves to be-
lieve that the system is good. But to get this kind of pilot opinion requires some-
thing more than a laboratory installation. A rather costly large-scale installation
is necessary.”15
He concluded with the hope that the experience gained would allow the sys-
tem to be “improved out of existence.”16 The CAA intended to buy ten of these
systems in 1940 so as to enable this extended testing to take place in a variety of
locations. It also planned an additional purchase of fifteen for 1941.
The CAA was able to commit to a standard and buy equipment for airports in
support of it because Congress had given it new legal authority in August 1938 in
a major reorganization.17 The CAA’s legal authority to adopt a standard for civil
aviation did not supplant military prerogatives, however, and it had to negotiate
with the War and Navy Departments to achieve military/civil standardization.
Widespread belief that a single standard for military and civil aviation was nec-
essary, combined with a technical desire to eliminate the ground-induced stabil-
ity problem, led it and the RTCA to view the Indianapolis system as transitional.
Any adoption of it was to be temporary, to produce the sort of information re-
garding performance at many locations which all previous testing had not been
able to provide, while continuing to investigate methods of producing a straight
glide path. Following the RTCA’s recommendation, the CAA continued its re-
search program to develop the straight, microwave path that was clearly the so-
lution preferred by many of its members. The RTCA’s standard, then, was a
means of establishing a coherent development program while providing limited
service to the airlines.
the microwave future
The root of the glide path stability problem was that Diamond and Dunmore’s
“landing beam” was not a beam at all. Instead, the glide path antenna’s radiation
pattern relied upon the earth’s surface to act as a reflector. That meant surface
conditions a¤ected the radiation pattern. Changes in the soil moisture content of
the surface altered the soil’s conductivity and thus the glide path’s transmission
pattern. That, in turn, changed the glide path angle. This was why the army’s 1936
tests found that the glide path angle changed as the ground dried out after a rain.
It also meant that changes in surface conductivity caused by large pipes, under-
ground electrical conduits, and similar artificial structures caused “bumps” in the
glide path. Fixing the glide path, then, meant finding a way to detach the glide
path’s radio propagation from the earth’s surface.
Wilmer L. Barrow, an MIT engineer, had discovered in 1938 that radio waves
could be propagated directionally, without reflection by the ground, by using
metal horns. In other words, the use of these “horn radiators,” as Barrow called
them, could generate radio beams independent of the earth’s surface. This was
clearly a potential solution to the instability problem, but it brought with it a sub-
stantial diªculty: the horns were very large. The size depended upon frequency
in an inverse relationship.18 Higher frequencies needed smaller horns.
In 1936, CAA engineer Irving Metcalf had proposed a blind landing system
based on three beams of energy, which would appear as three spots on a cathode
ray tube display. Initially, he had believed that infrared energy would be the best
choice, but after he consulted with Edward Bowles at MIT he decided microwaves
were a better choice. Working at the Round Hill experiment station, Bowles’s re-
search group had demonstrated that infrared would not penetrate all types of fog
and mist.19 Microwave radio seemed a better solution.
Metcalf was able to convince his superiors to fund a twelve-month contract
with MIT to build a prototype blind landing system based upon Barrow’s horns
and microwaves at fifty-centimeter wavelength.20 The horns were still large but
much more manageable than they would have been at lower frequencies. Ulti-
mately, the Round Hill group devised a system that produced four beams instead
of three. By overlapping two vertical fan-shaped beams along their flat faces, the
90 blind landings
system established a plane that provided an indication of proper course. Similarly,

it overlapped two horizontally fanned beams to produce a plane for vertical navi-
gation reference. Although all four beams were transmitted at fifty centimeters,
each was modulated with a separate audio frequency, which was then separated
by filters in the receiver.
The resulting system had two great advantages over the Bureau of Standards
system’s UHF descendants: the beams were perfectly straight, and they were
nearly immune from changes in ground conditions. It therefore answered pilots’
demands for a more stable system. A successful demonstration of the system in
early 1938 led to a six-month extension of the bureau’s contract, so that Bowles
could finish the report he was required to submit. The system had one major
drawback, however: the cluster of triodes which generated the microwave signal
could only produce about three watts of power, resulting in a useful range of only
about a mile. More power was clearly necessary for a useful blind landing system.
Fortunately, a potential solution already existed, a new vacuum tube called a kly-
stron.
Physicist William Hansen and two other researchers, Russell and Sigurd Var-
ian, developed the klystron in the Stanford University physics department. Stan-
ford’s physics program was oriented toward research into high-powered devices.
In 1936, Hansen had developed a microwave tube called a rhumbatron, but this
did not produce much power. Department chair William Webster had unsuc-
cessfully sought money for a “million-volt X-ray” device in 1937, and after this fail-
ure Hansen turned toward research into techniques that might be less expensive.
Russell Varian, who had been a student of Hansen’s before going into private
enterprise, was also interested in this research area, and he kept in touch with
Hansen. According to Russell, Hansen wrote to him in 1937 that hollow res-
onators seemed to be very eªcient at generating high-frequency radio energy.21
This was the basic idea behind the klystron.
Russell and his brother Sigurd, a Pan Am pilot, had also corresponded about
possible uses of microwave energy, including aircraft detection and blind land-
ing. Sigurd had expressed concern over the concentration of power—particularly
air power—in hands of European dictators. The Spanish Civil War had made
clear the future importance of air power, and Sigurd believed that aircraft could
easily bomb targets via “blind flying” techniques, while the targets could not de-
fend themselves against the invisible aircraft. Microwave-based radio locators
might resolve this problem. In September 1937, the two brothers went to Stan-
ford and argued for a project to develop a resonator-based tube that might pro-
duce the high-energy microwave radiation an aircraft detector would require.
Sigurd was particularly adamant about it, according to Russell, and they suc-
ceeded in persuading Hansen. Hansen got Stanford president Ray Lyman Wilbur
to support the e¤ort with a $100 grant. The university’s board of directors ap-
proved the arrangements in early October.22
The Varians quickly succeeded at assembling an apparatus that produced mi-
crowave energy and, nearing the end of their finances, tried to produce interest
in the device by visiting the local navy and CAA oªces. Irving Metcalf happened
to be in town, and he and Hugh Willis of Sperry Gyroscope were immediately in-
terested in seeing it in action. Metcalf realized that combining the klystron with
Bowles’s construct at MIT was the obvious route to making a stable, useable glide
path. Willis perceived a new business opportunity, and he quickly arranged a deal
with the Varians and Stanford to take over financial support of the klystron
e¤ort.23
The microwave work at MIT and Stanford had also attracted the attention of
the U.S. Army Signal Corps. Col. John Mauborgne, who had been director of the
Aircraft Radio Laboratory at Wright Field, had been promoted to brigadier gen-
eral and assigned chief signal oªcer of the army in early 1938. In April of 1938
General Mauborgne visited MIT, where he met with Karl Compton and Edward
Bowles and witnessed a demonstration of the CAA-MIT microwave landing sys-
tem.24 He came away from that meeting convinced that it was the solution to the
army’s need for a straight glide path, and he arranged to send people from the
Signal Corp’s Radio Laboratory to MIT to learn more about it. Bowles’s assistance,
in turn, led the Signal Corps to establish an in-house research program into mi-
crowave landing systems at Wright Field. At first, the army was only interested in
the glide path part of the system and arranged to borrow a localizer from the CAA
to use in their tests.
Mauborgne’s replacement as director of the Aircraft Radio Lab, Lt. Col. Hugh
Mitchell, kept him informed of new research through correspondence with one
of his sta¤ members, Col. Luis Bender. Mitchell heard of Stanford’s klystron from
Metcalf and had corresponded with Webster about it. Mitchell had then described
the device to Bender in a handwritten note, emphasizing the importance of this
device for blind landing of aircraft, as he knew that his boss was interested in
the subject.25 He believed it was the best possibility for extending the range of
Bowles’s glide path from a mile to the minimum of ten that the army wanted.
Mauborgne arranged to borrow one of the prototype klystrons from Stanford
for use in the army’s copy of Bowles’s glide path transmitter, believing that join-
ing the two would provide a useable range. The army first flew the klystron to
MIT, however, where Bowles was able to use it for a few weeks before the army’s
92 blind landings
equipment was ready. Initial experiments proved very exciting, but flight tests
originally scheduled for May 1939 had to be postponed into June because the kly-
stron had to be sent back to Stanford for repairs. The research projects at MIT
and Wright Field su¤ered from frequent problems with the prototype klystrons,
forcing them to send the devices back and forth between MIT, Wright Field, and
Stanford, which considerably delayed their work.
The shortage of klystrons in early 1939 led the CAA, MIT, and the army’s Air-
craft Radio Lab to cooperate at MIT. MIT, which expected to get a new klystron
before the army’s was repaired, approved an army request to fly MIT’s system
once the klystron arrived. The CAA had renewed its research contract with MIT,
although its e¤orts in the microwave area were waning as the UHF system it had
contracted for with ITD neared completion at Indianapolis. The Sperry Gyro-
scope Company, which had established a development laboratory for the klystron
in San Carlos, California, was also informally involved in the undertaking at this
point.26
The cooperation was shaky, however. General Arnold had written to Clinton
Hester, administrator of Civil Aeronautics, proposing a cooperative development
e¤ort in the microwave field. Hester happily agreed to participate and even to pro-
vide the equipment.27 Arnold then demanded establishment of an overall policy
of cooperation before agreeing to a cooperative program in this specific case, how-
ever, which Hester appears to have been unwilling to commit to. Relations be-
tween the two organizations deteriorated, while Bowles’s group continued its
work at MIT and the Signal Corps prepared its own version of MIT’s system at
Wright Field. But the research at MIT and Stanford had produced a new set of
technologies that would satisfy the army’s requirements for a straight glide path.
The remaining issue for the army was how to get everyone else to adopt what it
wanted.
the national academy of sciences:

validating a research program
The completion and relatively successful testing of the International Tele-

phone Development Corporation’s design at Indianapolis during the summer of
1939, while the army’s enthusiasm for microwaves was still growing, set the stage
for what other historians have presented as an army attempt to block the RTCA’s
standard. Wilson and Leary have argued, based upon CAA records, that General
Arnold sought to prevent the CAA from carrying out the RTCA’s plan by calling
for a National Academy of Sciences (NAS) review, apparently in the hope that
NAS would recommend something else. Army records suggest, however, that
Arnold’s memos were misunderstood in the White House and that Arnold had
no intention to block limited commercial deployment, which was, after all, the
RTCA’s own plan. Instead, Arnold sought validation of his belief that microwaves
were the best solution to the glide path problem and therefore of the Signal Corps
research program. Once he had that validation, he intended to insure that the
CAA conformed to the limited deployment of the RTCA’s system, as he believed
it had agreed to do.
Arnold, after a conversation with the assistant secretary of war for air in Au-
gust 1939, had written for the secretary of war’s signature a memo to President
Roosevelt. In it, Arnold requested that the NAS be asked to evaluate the existing
blind landing systems. They were, he argued, a “disinterested group of distin-
guished scientists who would serve much as a court of justice serves in our sys-
tem of jurisprudence.” The Air Corps had gotten good advice from them in the
past, Arnold believed, and he thought that their national prestige would end all
“quibbling or contraversy [sic]” once they had announced their decision. Arnold
had begun working with members of the National Research Council in 1936, as
Gen. Oscar Westover’s assistant. That experience had convinced him that scien-
tists were willing and able to advise the Air Corps on new technologies in which
it had no expertise.28 The NAS was therefore a reasonable place for him to turn
to resolve what he perceived as a divergence between the needs of the Air Corps
and the plans of the CAA.
Arnold’s letter to President Roosevelt got an immediate response. In a letter
dated the same day, August 30, Roosevelt asked Paul Brockett, the executive sec-
retary of the National Academy of Sciences, to look into the problem of stan-
dardizing upon a common system. Brockett turned the letter over to Frank Jew-
ett, then president of the NAS, who responded on September 1. Jewett agreed to
appoint a committee to look into the problem, with the caveat that it might not
be possible to attempt complete standardization.29 Jewett’s initial sense, that stan-
dardization might not have been possible given the state of the art, is certainly
borne out by the RTCA’s construction of an interim standard.
Jewett chose Vannevar Bush to chair that committee and assembled a number
of prominent researchers to serve on it. The conference committee consisted of
Oliver Buckley, executive vice president of Bell Labs; electrical engineer Dano
Gunn; mechanical engineer W. F. Durand from Stanford; physiologist Joseph
Erlanger of Washington University; telephone engineer Bancroft Gherardi, who
was retired from AT&T; biological chemist Lawrence Joseph Henderson of Har-
vard; aeronautical engineer Jerome Hunsaker at MIT; and physicist Max Mason
94 blind landings
of the California Institute of Technology.30 Bush also wrote to the various aviation
interests, including the CAA, the Air Corps, the navy, the Air Line Pilots Associ-
ation, and the Air Transport Association, seeking their positions on the issue.
In October, Charles Stanton, acting administrator for Civil Aeronautics, re-
sponded to Bush’s request with a long explanation of the CAA’s current policy.
Stanton explained that the CAA intended to install ten sets of equipment at im-
portant airports, while continuing to improve the equipment. The CAA recog-
nized the need to familiarize pilots with instrument approach procedures well in
advance of attempting blind landings in routine operations. He asserted that the
CAA intended to continue funding MIT’s research into a microwave-based sys-
tem. A reliable, manufacturable microwave system, he believed, was three years
away, and because airlines expected aircraft radio equipment to become obsolete
every three to four years, he felt no concern that deploying a UHF system would
prevent adoption of a microwave one once one was available. Air carriers, he
pointed out, had “shown eagerness to accept the use of new equipment which
tend to increase the regularity with safety of scheduled operations.”31 Stanton ar-
gued for the approval of ten ground stations.
Bush received similar advice from Edgar S. Gorrell, president of the Air Trans-
port Association. Gorrell had written to each of the major airlines to get their sup-
port for Bush’s investigation, and expressed their “sincerest cooperation with the
National Academy of Sciences in the subject of one instrument landing system
for common adoption.” Gorrell also pointed out that the CAA had a project under
way to produce a standard landing system through the RTCA and attached a copy
of the September 14, 1939, proposed standard to his letter for Bush’s edification.
The air carriers, he said, were convinced by their many years experience in test-
ing various systems that the CAA’s new system was the best that they had seen,
and he urged that “nothing be done to retard the program.”32 The airlines, he
said, were unanimous in their wish that the CAA proceed with the initial ten in-
stallations. The airlines were clearly satisfied that even as an interim system, the
RTCA’s proposal would provide suªcient economic benefit to justify their costs.
Bush’s committee had no significant funding, and it met only once, on Octo-
ber 14 and 15, 1939. The group saw presentations by the army, the CAA, and the
navy. It drew up a draft report on the fifteenth, and the committee decided that
additional information should be gathered from the various involved agencies. It
also decided that Bush should personally fly each of the systems in question, par-
ticularly the CAA’s Indianapolis installation and the MIT microwave experiment.
He was to keep the committee informed by letter.
At the meeting, the committee focused primarily on CAA’s system and MIT’s
microwave project. The report it drafted on the fifteenth proposed continuing
the CAA’s plans to install ten copies of the Indianapolis system, with the caveat
that it be modified to meet the six-mile, 1,500-foot approach that the pilots had
specified the month before but that ITD’s engineers had not yet managed to
achieve. The draft also supported the army’s desire for a straight path and advo-
cated continuing microwave research as the best means to produce it. In short,
the NAS’s draft report echoed the RTCA program almost exactly, with somewhat
more emphasis placed on microwave development than in the RTCA docu-
ment.33
As the largest unresolved issue was the straight versus curved glide path prob-
lem, Bush sought more information from various parties. He received a letter
from Charles Stanton on November 3 notifying him that the Indianapolis glide
path was now in conformance with the RTCA specification. Stanton also for-
warded a letter from Colonel Mitchell to Richard Gazely, which stated that the
army found the CAA’s specification acceptable for test purposes, but for army use
a completely straight path, inclined at 3 degrees, was necessary. The colonel reit-
erated that he supported the installation of ten sets of the CAA’s system for pilot
training and familiarization, but deployment should not go further until the sys-
tem’s real world performance was well understood.34
Bush also received a long letter from Bowles that discussed the Indianapolis
system and the MIT microwave experiments. Bowles told Bush he had spoken
with William Jackson from CAA’s Technical Development Section and had been
told that the straight six-mile glide path that had been produced had not yet been
thoroughly tested for straightness and “freedom from irregularities.” Further,
Bowles had been left with the impression that the straight portion could not be
substantially lengthened with the existing equipment because this was done by
moving the transmitter equipment further to one side of the runway. The further
away from the runway the equipment was moved, the greater the probability of
interference with the radiation pattern from reflections and other surface condi-
tions. Although Bowles did not discuss the issue in any detail, he stated that ITD’s
engineers believed that an equal signal glide path was possible at UHF wave-
lengths.35
Bowles argued explicitly for an equal signal glide path in his much more de-
tailed discussion of the microwave system. “I believe that the groups interested
in the instrument landing of airplanes subscribe to the idea that the ultimate sys-
tem will be a microwave system and that the value of a longer wave system lies
96 blind landings
principally in the fact that these systems are now in such form that at least they
o¤er immediate means for the training of commercial pilots in the technique of
instrument approach and instrument landing.”36
He added that the constant-intensity glide path at Indianapolis “must sooner
or later degenerate into a curved path” if the CAA made further attempts to
lengthen it. He had demonstrated already that microwave radiators could produce
a straight equal signal glide path by means of a sharp ultra-high-frequency
beam.37 The army, he thought, agreed with his point of view and were working
on such a system at Wright Field. Finally, he asserted that an equal signal glide
path, whether UHF or microwave, was most likely to resolve the straightness
problem and the instability problem because it was an “actual path in space” that
could be made independent of transmitter output and receiver sensitivity.
In a prescient paragraph, he also argued that the “flare” at the bottom, which
had been specified by the pilots to reduce impact upon landing, would be un-
necessary and should be dropped because the system would only be used as an
approach system, not a landing system.38 If the system was only flown as an ap-
proach system, with the pilot making a visual landing, any desired flare out could
be done visually. Bowles’s estimation of the unfeasibility of a true “blind” land-
ing system was no di¤erent than that of United Air Lines, which had already rec-
ognized that blind landings belonged to the remote future, if they were possible
at all. Within a year, the army also picked up on the improbability of truly blind
landings and tried to change the name given to this class of technologies from
“instrument landing systems” to “instrument approach systems.”
Bowles argued, finally, that the fastest path to a microwave system was a
UHF/microwave hybrid, utilizing the UHF localizer of the Indianapolis system
and a microwave glide path, put together by a commercial company. He pointed
out that RCA, General Electric, and Sperry Gyroscope were working on equip-
ment suitable for this task. He also contended that the “ultimate” system should
be all microwave, to resolve the remaining problems with bends in UHF localiz-
ers, and should employ a ten centimeter wavelength, instead of the fifty cen-
timeters that he and the Aircraft Radio Laboratory had been working on. He be-
lieved that this could reasonably be done in two years.39 The Aircraft Radio Lab
had already taken this approach at Wright Field, probably on Bowles’s recom-
mendation.
Having digested all of the arguments presented by his interested parties, Bush
flew o¤ to try out the various systems. At the Aircraft Radio Lab, he was able to
try a Bendix system, the old Army Hegenberger system, and at nearby Patterson
Field, the radio lab’s prototype microwave system. He wrote the committee on
November 20 that when flying the Bendix system, he was “further impressed
with the desirability of a long, substantially straight path, although the system in
general operated successfully and satisfactorily.” On the old army A-1 system, his
pilot, whom he called “one of the most skillful and experienced,” made two ap-
proaches and missed both. Bush was “completely convinced” that this system
made “too severe demands upon the pilot,” and this was reason enough for its
abandonment. At Patterson Field, he found the microwave equipment to be “in
experimental condition only.” He was unable to fly the system because both of
the available microwave receivers broke during his visit. He was convinced of the
microwave system’s “great future promise” but was clearly less optimistic than
Bowles had been about its nearness to commercial utility.40
At Indianapolis, Bush found that William Jackson’s group had done an excel-
lent job on the equipment. His pilot made satisfactory instrument landings from
all four directions. He judged it “completely satisfactory for commercial use.” He
agreed that the substantially straight part of the glide path was six miles long and
admitted that while the army wanted an even longer glide path, he was “inclined
to believe that much more is not really necessary.”41
The issue of the length of the glide path is diªcult to evaluate. The length to
be chosen depended on how air traªc was managed and what sort of aircraft were
using a facility. In commercial operations, especially during poor weather, aircraft
were typically stacked above a radio marker beacon at some distance from the air-
port, with the most recent arrival at the top of the stack, and the next aircraft to
be cleared to land at the bottom. Since commercial aircraft flew at a maximum of
10,000 feet (they were unpressurized), and aircraft were separated by at least
1,000 feet within a stack, having the base of the stack at 1,500 feet allowed a nine-
layer stack. The width of a stack, and therefore its overall volume, was defined by
aircraft performance. Faster aircraft, and heavier aircraft with poorer turning abil-
ities, meant a bigger stack that then had to be further away from an airfield in or-
der to prevent congestion close to the field. Because Bush was not a pilot and flew
these systems in commercial aircraft with pilots familiar with commercial needs
and in good weather, the six-mile glide path seemed perfectly reasonable to him.
It meshed with the current suite of technologies in use by the airlines.
By 1939, however, the Air Corps was getting aircraft that flew higher, weighed
more, and were faster than anything in commercial use and therefore demanded
larger, more distant stacks. That in turn meant a longer approach. The need to
stack those planes was reflected in the Air Corps’ demand for a longer glide path.
It is true that the Air Corps could probably have lived with a localizer-only ap-
proach for the first few miles of a descent (and, in fact, it did not have a glide path
98 blind landings
until 1944), but that largely defeated its purpose in adopting a glide path, which
was to reduce accidents through improved vertical guidance.
The army wanted its pilots to pick up the glide path fifteen miles out and fly
the twin beams in, instead of picking up one beam at fifteen miles and the other
at six miles. Since the cross-pointer instrument displayed vertical and horizontal
position, it provided a vertical indication even if it was not receiving a glide path
signal—the elevation needle did not disappear in the absence of a signal. Instead,
no signal resulted in a “fly down” indication, which a pilot would have had to ig-
nore until approximately six miles out, with no certain indication of when he
should start receiving the signal and therefore start paying attention to the eleva-
tion needle. Hence, the Air Corps’ demand was driven by the performance char-
acteristics it had designed into its new bombers, combined with the need to pro-
vide its pilots with a positive instrument indication. Bush could not have realized
this from his own experience with these systems, but he deferred to the army’s
demands in his final report.
After returning to Washington, Bush made some minor changes to the draft
report, the most significant of which were a specification that the army and the
CAA needed to agree on a common sensing of the cross-pointer instrument
through the RTCA, and a statement that the committee did not feel that a proper
solution lay in the low-frequency direction. He requested that the committee tele-
graph their approvals by November 24 so that the report could be submitted
promptly.
Bush apparently received the approval quickly, for the report he sent to Jewett
bears the date November 21. Jewett, in turn, transmitted it directly to President
Roosevelt. Roosevelt had his personal secretary, Gen. Edwin Watson, send copies
of it to the army, the navy, and the CAA on December 5. The CAA responded with
a highly laudatory letter to Bush, thanking the committee for its e¤orts in inves-
tigation of this “highly important and complex subject.”42
Although the report was eventually approved unchanged, the process of ap-
proving it took longer than drafting the report itself. The first stumbling block
was the Civil Aeronautics Authority, which believed it had gotten exactly what it
had wanted and pushed the National Academy of Sciences to make the report
public immediately. The academy’s executive secretary wisely checked with the
White House and found that Roosevelt considered it a confidential report until
he, personally, chose to release it. The CAA had apparently not been quite so cir-
cumspect, and details found their way into the press anyway. In Brockett’s words,
this had left the White House “just a little annoyed” with the CAA.43
The secretary of war, Harry Woodring, responded on December 28 to General
Watson’s memo. The War Department, he said, concurred with the NAS report,
and he emphasized that the department considered the plan for future develop-
ment it proposed “to be of material assistance in guiding the development to-
wards standardization of an instrument landing system available to all aviation
services.”44 The records show that Arnold himself drafted Woodring’s reply.
Arnold, therefore, was primarily interested in the report for its validation of the
army’s research plans, which were clearly present in the report’s call for mi-
crowaves.
On the day Arnold wrote Woodring’s response to the NAS report, he also acted
upon a memo initiated by the chief signal oªcer in October. General Mauborgne
had requested Arnold’s opinion on whether to pursue the army’s microwave re-
search by letting a contract to Sperry Gyroscope Company for a complete micro-
wave system, based upon the klystron. Arnold had clearly held onto the memo
while waiting for the NAS report. In his answer to Mauborgne, Arnold requested
that the Signal Corps “expedite” its research program and then quoted the NAS
report’s microwave section verbatim.45 His holding of the chief signal oªcer’s
request and his reliance on the Bush committee’s findings suggest that Arnold
had already come to rely heavily on scientists’ recommendations in subjects out-
side his expertise. He sought validation of the RTCA’s program through the NAS
and wound up causing a great deal more trouble than he intended.
The secretary of war’s concurrence was not enough to establish the academy’s
report as a formal policy. That there was also no policy to guide its adoption led
to a great deal of confusion. Because War and Navy were Cabinet departments
and the CAA was an independent agency, no authority existed to approve and en-
force the report’s recommendations other than Roosevelt himself. The White
House did not immediately recognize this, and after Roosevelt received the sec-
retary of war’s statement of approval, he asked his secretary, “What do we do
next?” Watson was equally in the dark, and he wrote to the secretary of war ask-
ing his opinion, who then told the chiefs of the Air Corps and Signal Corps to
submit their ideas on how to proceed.46
With much of oªcial Washington still on the traditional Christmas hiatus, nei-
ther the chief signal oªcer nor the chief of the Air Corps responded in person.
Instead, their executive oªcers responded, most likely after consulting with their
bosses. In any case, the executive oªcer of the Signal Corps, Col. Clyde Eastman,
suggested that the president approve the report and “furnish copies of it to the
interested agencies for their guidance.” The executive oªcer of the Air Corps,
Maj. C. E. Duncan, concurred. The secretary of war accepted their recommenda-
tion and so recommended to General Watson.47
100 blind landings
By putting the decision back in the White House, the Air and Signal Corps
placed a technical decision in the hands of nontechnical people. Roosevelt relied
heavily upon Watson’s advice in army matters, and Watson, a retired infantry oª-
cer, did not understand the issues. A series of letters between Watson, Woodring,
and Arnold ensued as Arnold attempted to explain to the others what the Air
Corps did and did not want. A particular set of personal memos between Arnold
and Watson had convinced the confused Watson that Arnold disliked the CAA’s
plans as embodied in the NAS report, and that perception was transmitted to
Clinton Hester at the CAA. The CAA’s records therefore document army resis-
tance that did not exist.
The confusion arose from a combination of Watson’s lack of comprehension
of the basic issues and an imprecise memo from Arnold. Watson had asked
Arnold for his opinion of the NAS’s report in a memo dated February 6. Arnold
had responded:
Dear “Pa:”
The Air Corps has no fault to find with Dr. Bush’s proposed solution. We still, of
course, believe that the system which we developed is better suited to meet Air
Corps needs, but we realize fully that it is an impossible situation to have di¤erent
aeronautical agencies each develop their own systems, no two of which any one
airplane could use. It is absolutely essential that one common system be in use by
the whole aviation industry and we are perfectly willing to give and take and com-
promise in order to arrive at that universal system.48
Watson seems to have understood this memo as supporting the army’s old
Hegenberger system over the academy’s recommendations. That interpretation
would certainly have indicated hostility on Arnold’s part towards the report’s rec-
ommendations, as Bush and his committee had clearly consigned the Hegen-
berger system to the scrap heap.
The context suggests that Arnold had meant to support the army’s microwave
system with this memo. Arnold was well aware of the Aircraft Radio Lab’s work
with Bowles on that prototype. He had sent Air Corps pilots to MIT to test the mi-
crowave system that the CAA had funded there. He had been satisfied enough
with the Aircraft Radio Laboratory’s progress with the system to ask the chief sig-
nal oªcer to expedite a procurement contract with Sperry Gyroscope for a man-
ufacturing prototype less than two months before. Finally, he had personally
signed the Air Corps’ approval of the NAS report and drafted a paragraph espe-
cially supportive of its “future developments” section, which included the mi-
crowave straight-line glide path system being experimented on by the army and
MIT.49 If we interpret Arnold’s memo as referring to the microwave system, then

there appears to be little ground for a charge of army resistance to the report.
Bush’s committee had, after all, validated the army’s need for a straight glide path
and had strongly preferred a microwave one.
The tone of Arnold’s memo suggests that he had some unstated concerns,
however, which may well have helped confuse General Watson. What those con-
cerns may have been are suggested in the minutes of a conference held March 11
in Arnold’s oªce. The administrator of Civil Aeronautics had requested the con-
ference in order to find out why the Air Corps was resisting the report’s recom-
mendations. At that conference, an anonymous oªcer recorded, it was carefully
explained to Hester that neither the Air Corps nor the Bureau of Aeronautics
(Navy Department) had any objections to “ten purely experimental instrument
landing systems, but felt that with the microwave system so far along, it would
be undesirable to invest funds in radio equipment which would soon become ob-
solete.”50 The drafter reported that Hester had agreed that these were purely ex-
perimental and that the CAA would support the installation of a better system as
soon as one was developed and satisfactory for general use.
Clearly, the Air Corps did not intend to adopt the CAA’s system at all, as it pre-
ferred to wait the expected two years before a procurable microwave system was
available. Arnold’s concern was probably over the possibility that Hester was not
negotiating in good faith. Arnold remembered clearly the Bureau of Air Com-
merce’s adoption and near-simultaneous abandonment of the Hegenberger sys-
tem. Many oªcers in the Air Corps had seen that as an act of betrayal. In this
case, Arnold was likely concerned that the CAA might take the NAS’s approval
of the Indianapolis system as an excuse to deploy the system widely, making get-
ting rid of it nearly impossible. Widespread commercial deployment of the sys-
tem would almost inevitably force the Air Corps to adopt it, whether or not it
satisfied their perceived needs.
The CAA did have larger plans for the Indianapolis system than the ten in-
stallations that NAS had approved, and it had budget authority for twenty-five in
the 1940 and 1941 fiscal years. One CAA oªcial had also suggested to the RTCA
(with the army members present) that up to fifty might be deployed before a mi-
crowave system became available. That amounted to an investment of $1.25 mil-
lion in 1940, if CAA’s 1939 estimate of $25,000 per installation had been accu-
rate, all of which would have to be abandoned when a microwave system was
deployed in two or three years. Because Arnold complained bitterly in his auto-
biography about the shoestring budget of his Air Corps, it seems likely that he
did not consider such a sum easily abandonable. Congress may well have balked
102 blind landings
at any CAA plan to replace the UHF system, too, as it was already unhappy at the
cost of maintaining the Federal Airways System.51 Arnold’s concern with being
stuck with an unsuitable system was probably well founded.
The March meeting adjourned with a promise by Arnold’s oªce to draft a
memo to General Watson to correct his misunderstanding of the Air Corps’ po-
sition. Arnold submitted such a memo, which unfortunately no longer exists, but
it was not enough to satisfy Watson. On April 3, the War Department chief of sta¤
requested that Brigadier General Yount, Arnold’s assistant chief of the Air Corps,
go to the White House to explain the matter to General Watson in person. Gen-
eral Yount’s conference with General Watson resulted in a request for yet another
memo, this time signed by the secretary of war, which was duly generated and
sent to Watson on April 13, 1940. Watson dispatched that memo to President Roo-
sevelt, and Roosevelt attached his “OK. Execute” on May 2nd.52 The RTCA’s rec-
ommendation, changed only in emphasis by the NAS committee, had finally
achieved the status of an oªcial standard.
conclusion
The process of constructing a standard landing system was made painful by

several interrelated factors. Each organization involved in the negotiations had a
particular set of technical demands, which mostly, but not completely, over-
lapped. Their technical demands, in turn, were based upon the types of aircraft
that they intended to operate. For example, the navy’s prewar focus on seaplanes
made it in many ways the “odd man out” in these talks, playing a distinctly sec-
ondary role to the Air Corps and the CAA. The curved glide path was well suited
to its seaplanes, and because no one else used flying boats, it was allowed to go
its own way.53 The navy’s unusual technological suite imposed unique require-
ments on its choice of landing system. Since it did not need to share airfields with
the army or airlines, it had no reason to conform with any standard, and no one
else chose to take issue with its nonconformity.
The Air Corps and commercial airlines, on the other hand, recognized that
their aircraft had to be able to use each other’s airfields in case of emergency or
war and agreed that a standard was necessary, but they disagreed on the timing
of that standard. An “almost straight” glide path was temporarily acceptable to
the airlines, until such time as the “ultimate” straight path was available. The air-
lines, and the government agency created to serve them, wanted an interim sys-
tem sooner rather than a final system later. This was a matter of economics, as
suggested earlier. A landing aid system, even one at only ten airports, was ex-
pected to pay for itself quickly though increasing the number of completed
flights. Hence, the technical issue of glide path curvature was secondary to the
airlines’ economic interests. The CAA’s goal was happy constituents, and the
airlines had powerful friends in Congress to whom unhappiness could be ex-
pressed—and it was the following year. The CAA’s support for a temporary so-
lution, even one costing large sums, was therefore perfectly reasonable, especially
since the CAA’s investment in ground transmitters amounted to less than the air-
line investment in receivers.
For the Air Corps, the technical and economic issues suggested a di¤erent
timetable. The Air Corps believed that its new big bombers needed a longer ap-
proach than that provided by the Indianapolis system, and it therefore wanted to
proceed directly to the ultimate system, which it expected to be available at the
same time that these aircraft reached full production. Without a war to provide
the operational necessity to justify the expense of a temporary solution, the Air
Corps’ technological needs were reinforced by its own financial analysis. With ten
times as many aircraft to equip as the scheduled air carriers possessed, the Air
Corps faced an enormous investment in receivers with no visible return, as well
as the additional burden of having to explain to a Congress still suspicious of mil-
itary spending why such an investment would have to be scrapped in a mere three
years. Focusing on the ultimate system was the most reasonable approach for the
Air Corps, technically and economically. The di¤ering needs of these organiza-
tions was merely exacerbated by the lack of a policy to guide approval of inter-
departmental standards.
Finally, the agreement achieved closure within the community on the proper
form that the solution to the blind landing problem should take. It had agreed on
the tripartite NBS system, consisting of marker beacons, localizer, and a straight
glide path. This represented less an exemplary artifact than an exemplary model
upon which equipment and training could be based. That model, which places
the information needed to land a plane directly in the cockpit, I call the pilot-
control model. Only two years after Roosevelt approved the NAS report, that
model was challenged by a new one based on a radically di¤erent principle of op-
eration. The result was a political crisis.
chapter five
Instrument Landing Goes to War
President Roosevelt’s May 1940 approval of the National Academy of Science’s

report on blind landing systems had established a “standard” interim system and
a research program to develop a final system. That approval allowed the army and
CAA to continue supporting Edward Bowles’s microwave project while CAA be-
gan installation of the allowed ten copies of the Indianapolis system. But U.S. en-
try into World War II caused the aviation community to abandon its carefully ne-
gotiated agreement. The army, suddenly needing a system sooner than later,
embarked on three di¤erent development programs to produce a better glide
path, while also supporting an MIT Radiation Lab project to produce a radar-
based system. That radar project, called ground-controlled approach, is the sub-
ject of Chapter 6.
The three glide path projects the Army Air Forces embarked on were a straight
very high frequency (VHF) glide path, for which it contracted with International
Telephone Development Corporation, CAA’s contractor; a ten-centimeter con-
tinuous wave (CW) microwave system based on Bowles’s work, carried out un-
der a Signal Corps contract to Sperry Gyroscope Company; and a ten-centimeter
pulsed glide path (PGP), which the MIT Radiation Lab began on its own initia-
tive. The VHF system, called SCS-51 during the war, was used by the U.S. Army
Air Forces in Europe, North America, and in the Pacific theater by transport air-
craft and bombers. It was functionally identical to CAA’s postwar instrument
landing system (ILS), and di¤ered from the prewar Indianapolis system only in
being portable and in using a straight, equi-signal, VHF glide path.1 The tempo-
rary system thus became a permanent fixture of the aviation infrastructure, while
instrument landing goes to war 105
both microwave systems—the wave of the future in 1940—have vanished into

the past.
There is no other explanation for the entrenchment of VHF at the expense of
microwaves other than the war itself. Entry into the war caused the United States
to provide the AAF far greater resources, both financial and scientific, with which
to pursue technological development. At the same time, those projects were con-
strained by immediate needs. The AAF could no longer wait for the future. It
needed a landing aid immediately to avoid the huge weather-related losses that
its new ally, the Royal Air Force, was already experiencing. The VHF system was
the obvious choice if its glide path could be straightened quickly. At the same
time, Sperry Gyroscope’s microwave project ran afoul of the MIT Radiation Lab’s
belief in the superiority of its own work. In an odd paradox, the lab’s devotion to
magnetrons prevented the production of a microwave glide path during the war,
leaving the field to the “low-tech” VHF system.
entrenching the low-tech solution, part 1:

scs-51 and world war ii
In the two years following Franklin D. Roosevelt’s blessing of the National

Academy of Sciences report, CAA’s Indianapolis system reached its nadir and
was then suddenly, and permanently, rescued. A mistaken cost estimate nearly
undid CAA’s program, while the pressing demands of a world war provided a
golden opportunity for CAA to salvage its shipwrecked plan and, in a stunning
reversal, forced the Army Air Forces to adopt it too. The war thus put the allegedly
temporary Indianapolis system firmly on the road to permanency.
In 1940, CAA found that it had greatly underestimated the Indianapolis sys-
tem’s cost. CAA had based its budget request for funding of the ten systems based
on estimates made by its own experiment station personnel. They, in turn, had
assumed that production units could be had at the same cost as the prototype con-
tract, $25,000. CAA had therefore made a $250,000 request in its 1940 budget
estimate, submitted to the Bureau of the Budget in late 1938. It received the full
amount and found it was not nearly enough when it opened the three bids sub-
mitted by Air-Track, Bendix, and the International Telephone Development Com-
pany (ITD). Instead of ten units, it had only suªcient funds for four. Although
CAA did let the contract to ITD, which was the low bidder, realization that only
four sets would be installed during 1940 threw its plans into disarray. The air-
lines decided that four installations were too few to justify the cost of installing
receivers in their planes and suddenly balked at CAA’s plans. At least six instal-
106 blind landings
lations were necessary to make back the cost of receivers before the whole system
would be discarded in favor of a microwave one in three years or so, the airlines
informed one of their Army Air Forces contacts.2
CAA’s contractor also ran into engineering diªculties on top of the financial
problems. The extensive modifications CAA had made to the original prototype
meant that it had to be reverse engineered in order to make accurate drawings
that could be used as a basis for manufacturing. The glide path transmitter had
undergone the most extensive changes and required the most engineering work.
That, in turn, consumed time, further delaying production and increasing costs.
CAA did not get any equipment before 1941 and still had not managed to con-
vince the airlines to buy receivers.3
CAA’s mistaken estimate threatened to kill the UHF system before it reached
deployment. With the temporary Indianapolis system delayed for at least a year
while the microwave system work forged ahead, the airlines believed it made lit-
tle sense to invest in UHF receivers. The airlines therefore abandoned the system
they had been pushing for years, causing no little feeling of betrayal at CAA. Nev-
ertheless, CAA went ahead with its purchase of the four installations, while try-
ing to convince the airlines to return to the fold.
While CAA was having problems with its system, events in the larger world
convinced the Army Air Corps that it could not a¤ord to wait for a microwave sys-
tem to emerge from Sperry Gyroscope’s production facilities in a couple of years.
The contacts forged between the Royal Air Force and Army Air Forces during
1941 brought home just how severe a problem lack of a blind landing system
could be. Although Bomber Command’s statistics do not break out landing acci-
dents, it lost hundreds of aircraft to noncombat causes during 1939–1941. Shortly
before the war, the RAF had arranged a license to produce the Lorenz system in
England, under the name standard beam approach (SBA) system. Like its prede-
cessor, the SBA system was unstable, and in 1942 physicist David Langmuir re-
ported to Lee DuBridge at the MIT Radiation Lab that RAF pilots had stopped us-
ing it. Lack of an e¤ective landing aid coupled with the extreme density of aircraft
involved in landing several hundred bombers in an area of about a hundred
square miles within forty-five minutes, it seems reasonable to infer that many of
Bomber Command’s noncombat losses were approach and landing accidents.
Hap Arnold, who traveled to Britain in April 1941 for conferences with his RAF
counterparts and an audience with the king, could not have been unaware of
RAF’s problems.4
The army thus continued to fund Sperry’s continuous wave glide path project,
while dropping the microwave localizer and marker beacons. The army intended
to graft a microwave glide path onto the UHF CAA system, just as Bowles had
recommended to Bush as the fastest approach to a useable system. With this “mix
and match” approach to a landing aid, the ten-unit restriction was no longer nec-
essary. With the approval of the army, therefore, RTCA recommended that the
ten-unit restriction in the agreed-upon plan be dropped. This was intended to al-
low civil and military aviation to install and use individual elements of the ap-
proved system (i.e., localizer and marker beacons) while work continued on a va-
riety of approaches to solving the glide path problem.5 One could make a low
approach using only the localizer and marker beacons, after all, and although this
was not a perfect solution to the blind landing problem, it was very much better
than nothing. A later RTCA decision added the Army Air Corps’ compass loca-
tor stations as optional equipment; after the war, commercial pilots insisted that
they be required because they made interception of the very narrow localizer
beam much easier. The army’s A-1 system hardware and CAA’s system thus be-
came completely integrated, on paper at least, by the end of 1941.
The Japanese attack on Pearl Harbor precipitated a massive installation pro-
gram. With active prosecution of a strategic air war targeted for late 1942, the
newly renamed Army Air Forces had to construct a ferry route up the East Coast
to Newfoundland and thence across Iceland and into Scotland to get its new
bombers to England. With CAA’s system the only one immediately available, the
AAF oªcially adopted all but the glide path part in December 1941. Recognizing
that CAA was the expert on its own system, the secretary of war transferred more
than $1 million to CAA to procure and install the marker beacons and localizer
portions of its system throughout the United States, based on a prioritized list
provided by AAF.6 Glide path transmitters were to be provided by the army once
it had developed one to its liking.
With only the glide path standing in the way of a complete system, and flush
with suddenly vast development resources, the army pursued three di¤erent glide
path projects in the hope that at least one might work out, and do so quickly. It
continued supporting Sperry Gyroscope’s continuous wave glide path project,
based on Bowles’s work, it entered into a contract with ITD for a 330 MHz equal
signal glide path, and it approved a Radiation Laboratory proposal to build a
pulsed glide path (PGP) based on the cavity magnetron. It also established a con-
tract with ITD to build a completely portable and militarized version of CAA’s sys-
tem (minus the glide path, of course), which the company had begun to do with
its own resources earlier in the year.7
The Signal Corps had taken over CAA’s Indianapolis Experiment Station, its
personnel, and its equipment, in early 1942 in order to more rapidly bring about
108 blind landings
an army version of the CAA’s system. It brought aboard ITD, which had changed
its name to International Telephone and Radio Manufacturing Company (ITRM),
to do the production engineering work on a portable system that was to be de-
signed around standard tube sets. The army intended the use of standardized,
easily available tubes to prevent having the system’s production delayed by parts
shortages, which were already becoming common. It also improved the system’s
maintainability by eliminating the need for specialty tubes, which would not
normally be stocked in the supply system. These were important considerations
in a wartime system.
The army still insisted on a straight glide path, however, and while ITRM’s en-
gineers reworked the localizer and marker beacons for portability and standard
parts, they also had to deal with the army’s demand for a long, straight glide path.
They did this by raising the frequency, changing the antenna, and adapting the
equal signal method used in CAA’s localizer for use in the glide path.
CAA’s glide path had operated on 93.7 MHz, while the localizer worked on 110
MHz. The two di¤erent frequencies required two transmitters, consuming more
tubes and extra space. ITRM’s engineers replaced the glide path transmitter with
a frequency multiplier, which raised the localizer transmitter frequency to 330
MHz (a VHF frequency) for broadcast through the glide path antenna. This saved
space and tubes, but more importantly, the higher frequency reduced the envi-
ronmental sensitivity of the glide path while improving the predictability of its
propagation. Predictability was important because the adoption of the equal
signal method allowed the glide path to be controlled to some extent. (Figure 5.1
shows the propagation pattern of ITRM’s glide path.) By broadcasting a 150-Hz-
modulated signal from the upper antenna of the new glide path antenna system,
and a 90-Hz-modulated signal from the lower antenna, the system created an
overlapping series of lobes that could be adjusted relative to the ground by alter-
ing the relative intensity of the signals transmitted from the two antennas. A tech-
nician could thereby adjust the glide path to compensate for changing ground
conditions.
The major drawback of this method is that it produced several possible glide
paths, most of which were unflyable. Each of the 150-Hz lobes could appear as
the correct glide path to a pilot. The possibility of intercepting the wrong glide
path was reduced by procedure: the approach chart for each field directed pilots
to be at a particular altitude when they reached the outer marker beacon, which
corresponded to the altitude of the correct glide path at that point. (Figure 5.2 is
the 1946 approach chart for Newark.) The chart’s lower section shows an eleva-
tion view of the approach, specifying that an approaching aircraft be at 800 feet
Figure 5.1. The SCS-51 glide path antenna radiation pattern. Only the lowest 150-cycle
lobe (light gray) is a useable approach angle for most aircraft. Because the two antennae
were independently excited, by adjusting the amount of energy supplied to them, an
operator could adjust the angle that the 150-Hz lobes made with the ground to a limited
extent. This helped compensate for changes in moisture. A monitoring device sounded
an alarm if the lobe patterns deviated more than a preestablished amount from the
desired pattern. M. E. Montgomery, “Latest Type AAF Blind Landing Equipment,”
Electronic Industries (January 1945): 101.
when passing through the outer marker beacon at Metuchen, thirteen miles from
the field. An aircraft at that altitude, and on the proper course, would intercept
the correct glide path.
The use of systematic procedures to reduce the probability of accidents was al-
ready common practice in aviation, although its application to airfield approaches
had not been standardized and codified, even for individual airfields, before
World War II. Chaotic conditions due to the vast increase in air traªc during 1942
forced the AAF and CAA to establish a joint board to standardize and publish the
approach procedure for each major airfield in the United States so that all pilots
near an airport would be following the same procedure, reducing the probability
of collisions while also increasing airfield handling capacity.8 That standardiza-
tion was also necessary for successful use of the new glide path; however, the
army’s newly completed landing system did not drive the standardization proc-
ess. Standardization began as a response to vastly increased traªc.
International Telephone and Radio’s glide path was placed into a competitive
flyo¤ in late 1942 against the two microwave systems, where it performed well
enough to win a Signal Corps order for 350 units, which, combined with the lo-
calizer and marker beacons, were to bear the designation SCS-51 (see Figure 5.3).
110 blind landings
Figure 5.2. The standardized 1946 approach chart for Newark airport. The top section
shows the approach as it appears from above, while the smaller middle section shows
the approach’s vertical aspect. The bottom section provides rate of descent for several
landing speeds, and the length of time a descent will last at that speed and rate of
descent from the inner marker (Elizabeth, 0.9 miles from the field). The minimum
ceiling at this airport was 500 feet, as the chart shows. The approach manual required
that pilots unable to see the ground from 500 feet immediately climb back to 2,000 feet
and request to “go around” for another try. Army Air Forces Instrument Letdown
Procedures, 1 September 1946, p. 91.
Early production models were subjected to extensive testing at various locations

in the United States during 1943, including locations in Alaska. There, the joint
air forces/navy command found that it worked well enough to recommend per-
manent adoption, and largely on the strength of the navy commander’s recom-
mendation, the navy agreed to accept SCS-51 as its own new standard later that
year.9 The Army Air Forces established SCS-51 as its standard instrument ap-
proach in 1943 as well, despite the system’s unsuitability for small aircraft.
The portable SCS-51 was deployed domestically first, because a low priority rat-
ing hamstrung CAA’s program to make permanent installations of similar equip-
ment. Equipment for domestic use was automatically placed low on the priority
list, while equipment intended for use in the combat theaters was given higher
Figure 5.3. An illustration of the SCS-51 localizer path. Top of the image shows the
localizer truck. The correct “path” is the surface of equivalent signal—when the signal
strengths of the 90-cycle and 150-cycle lobes are equal. National Air and Space Museum
(NASM A-4974-A), Smithsonian Institution.
112 blind landings
priority. Because SCS-51 was designated vital tactical equipment, it received one
of the highest priorities. By November 1943, less than a year after its selection,
SCS-51 had been installed at fields along the northeast all-weather ferry and air
cargo route to Europe, from New York (Mitchell Field and Newark), through West-
over Field in Massachusetts; Presque Isle, Maine; and Harmon Field, Newfound-
land. The vast number of receivers needed to equip all of the AAF aircraft to use
the system, however, took many months despite the high priority. High volume
production of receivers took until mid-1944 to achieve, with new production air-
craft, particularly the B-29s, getting most of the first batches.10 Hap Arnold in-
tended to deploy SCS-51 to the Pacific theater with the early Twentieth Air Force
B-29 units, although I have found no records attesting to specific locations. Ad-
ministrative records dealing with the Pacific theater are in general far less avail-
able than those for the European theater, and to describe SCS-51’s operational
record, we will have to rely on Eighth Air Force experiences with it.
All U.S. air bases in Britain had been built with RAF equipment, and they
were equipped with SBA systems. Accordingly, Eighth Bomber Command had
adopted RAF’s system as its landing aid when it first arrived in Britain. Bringing
in a new system made little sense, and in any case, the AAF did not have one avail-
able. Eighth Bomber Command ordered several thousand sets of SBA equipment
in 1942, which were to be shipped back to the United States and installed in the
aircraft at the factories. These were never built. Although it did receive small
numbers of SBA receivers, the Eighth never obtained enough to equip most of
its aircraft, and therefore had no blind landing system at all until late 1944.
Therefore, like the RAF, which by 1943 had begun abandoning SBA due to
lack of receivers, Eighth Bomber Command did not fly missions if the ceiling for
the return flight was expected to be less than 500 feet, and if forecasts turned out
to be wrong, bomb groups broke up into squadrons, which descended to one hun-
dred feet or so above the English Channel and then flew by visual landmark back
to their fields. Various bomb groups worked out their own approach procedures
based on their home field and aircraft equipment. Aircraft equipped with Gee, a
navigation system designed primarily for bombing, could make approaches us-
ing that system, with the plane’s navigator feeding the pilot directions.11 Aircraft
without Gee sometimes could use compass locator stations called “slashers,” and
intended as raid-forming beacons to shoot radio compass approaches to their
home fields, much like the procedure for the army’s Hegenberger system. In truly
blind conditions, airfields also stationed men with flare guns at either end of the
landing field. The soldier at the approach end fired a green flare when he heard
a plane passing overhead; if the man at the departure end heard a plane flying
overhead, he released a red flare. The gun crews in the planes were expected to
watch for the flares and tell the pilot whether to land or not. There were, there-
fore, several field solutions to the blind approach problem, although all were dan-
gerous for novice pilots.
With the establishment of SCS-51s along the ferry route in late 1943, the equip-
ment started to earn a reputation for itself, and Gen. Carl Spaatz, commander of
U.S. air forces in Europe (USSTAF), requested that a unit be sent to the United
Kingdom for trial there. The equipment arrived in late January 1944 and in early
February a two-week long series of tests took place at De¤ord, attended by senior
commanders from Eighth and Ninth Air Forces, RAF’s various commands, and
the joint chiefs. Based on these tests, the joint chiefs adopted SCS-51 for use by
both RAF and USAAF in the European theater immediately. Initially, thirty sets
of ground equipment were required, with enough receiver sets to entirely equip
the Eighth and Ninth Bomber Commands’ aircraft.12
The major selling point for European theater commanders was SCS-51’s easy
adaptability to automatic control. At the tests at De¤ord, Signal Corps Lt. Col.
Francis L. Moseley, formerly an engineer for Sperry Gyroscope, had demonstrated
a device that converted the system’s localizer receiver output into a signal that the
aircraft’s autopilot could use. Although the device was not intended to land the
plane, it significantly reduced the pilot’s workload during the approach by doing
most of the flying. Essentially, the pilot’s job during an approach became moni-
toring the system’s performance, dealing with the throttles, and taking over when
the field was in sight. Spaatz liked the idea enough to demand that the experi-
mental device be procured immediately as an integral part of the system. The de-
mand for this “automatic coupler” was driven, in part, by the legacy of the failed
SBA system, which had left pilots distrustful of the very similar SCS-51.13 The au-
tomated approach removed much of the mental stress that blind landings en-
tailed, while removing many of the skill requirements (and therefore training re-
quirements) as well. In short, it fit the Army Air Force’s wartime need for a system
that its relatively inexperienced pilots could use.
Spaatz’s demand for the automatic landing coupler caused the Signal Corps
to embark upon a crash program to develop versions of it compatible with the va-
riety of automatic pilots in use, including the Honeywell C-1 and Sperry A-5,
which ran the duration of the war. The device never made it to combat theaters,
however, because it could not be produced in time. Similarly, the Eighth Bomber
Command did not receive enough glide path receivers for all of its aircraft until
114 blind landings
January 1945, despite the receiver’s design having been completed in late 1943.14
Even high priority equipment could not be manufactured so quickly in such large
volumes.
USSTAF headquarters kept up the pressure, however, complaining in Sep-
tember 1944 that lack of glide path receivers was preventing C-47s of Air Trans-
port Command from keeping supplies moving to the front, and that the auto-
matic coupler was vitally needed.15 By early 1945, enough glide path receivers had
reached Europe to allow equipping all bombardment and transport aircraft with
them. This left the AAF with a training problem, since its pilots had not been
trained to use the system.
The AAF had attempted to establish an instrument landing training program
in 1942 using sets of YB equipment borrowed from the U.S. Navy.16 Because the
YB system was very similar in operation to the SBA system pilots deploying to
Europe were expected to use, pilots would be able to gain familiarity with the
SBA’s landing procedures while still at the army’s training schools. The idea
failed in application, however, for the same reason the navy abandoned the sys-
tem the following year: it was simply too unstable, especially for training use. The
training schools quickly dropped the idea, and pilots deployed to Europe with no
training in instrument landing.
Instrument landing training therefore had to be done in the combat theater.
To accomplish that, USSTAF pressed into service the Link trainer, which could
be equipped to simulate instrument landing the same way it already simulated
instrument flying.17 A cross-pointer instrument in the trainer’s cockpit was ma-
nipulated remotely by the trainer operator. Pilots in the trainers then used the in-
dications to fly the trainer with, as a pilot would if the trainer were a real aircraft.
Some pilots reported that they received no training at all, while others report hav-
ing been able to practice on the Link equipment between missions. Hence the op-
portunity to train on the Link equipment was clearly not available to all pilots.
To help ameliorate the training problem, the AAF’s training manual suggested
that pilots practice on the real system as much as possible, by making their ap-
proaches “under the hood” even in good weather, with the copilot completing the
landing visually. With only thirty SCS-51s in the European theater, however, most
fields did not have one to practice on.18 Instead, the sites for SCS-51 installation
were usually chosen so that the equipped field could serve as an emergency field
for several nearby bases. That was necessary due to the dense spacing of fields in
England, combined with the limited number of channels available to SCS-51. Al-
though more equipment could have been set up, the mutual interference would
have rendered the sites useless. Hence training for SCS-51 proved a severe prob-
lem for the AAF, which had thousands of crews to train in Britain.
Even without the automatic approach coupler, however, and with a limited
training program, pilots who had access to SCS-51 seem to have appreciated it.
One pilot remembered the glide path was the most satisfying part of the system
because it eliminated the altitude uncertainty remaining in the barometric al-
timeter. It greatly reduced the number of “go arounds” because aircraft broke
through the clouds in consistently better position for landing. The AAF had long
ago given up on the idea of routine blind landings, and this pilot’s recollection is
exactly what the AAF’s leaders had hoped to achieve.
World War II, therefore, rescued the Indianapolis system from an oblivion vir-
tually guaranteed by CAA’s mistaken cost estimate and the resulting defection of
the airlines. The war also established the system’s reputation by demonstrating
that a di¤erent VHF glide path would work. Finally, by instigating widespread de-
ployment of this hybrid system, the war had made it easy to justify adopting the
system permanently. Vannevar Bush had warned against the dangers of deploy-
ing a system too early, thus making far more diªcult its replacement by a supe-
rior system, and as we will see, his foresight had been correct.
a brief interlude: demise of the high-tech solution
The army did not stop supporting Sperry Gyroscope’s work to construct a mi-
crowave glide path when it decided to purchase International Telephone’s VHF
one. Instead, it pursued both projects. Success of the VHF glide path was there-
fore not solely responsible for the failure of the microwave one. The MIT Radia-
tion Lab undid Sperry’s work by challenging Sperry’s glide path with one of its
own.
Sperry Gyroscope, which had long been a contractor for both the U.S. Army
and U.S. Navy for navigation equipment, entered microwave work when it con-
tracted with Stanford’s physics department to fund the development of the kly-
stron into a commercial product. Chapter 4 details Edward Bowles’s work on a
prototype forty-centimeter blind landing system at MIT, and the Signal Corps’
award of a contract to Sperry to build a complete blind landing system based on
Bowles’s work. That system proved to have one major problem: size. The radia-
tor horns needed to broadcast eªciently at forty centimeters were too big to be
portable, and were also a significant collision hazard for aircraft. The obvious so-
lution was to use a shorter wavelength, which would require smaller radiators.
116 blind landings
Figure 5.4. The Sperry Gyroscope microwave glide path transmitter. Note that the original
horn radiators have been replaced by what is essentially a vertical “slice” through them.
With microwave transmission, the width of the beam is inversely related to the transmitting
antenna’s dimensions. In this case, the antenna produced a pattern that was wide
horizontally but very narrow vertically. Courtesy of Hagley Museum and Library.
The Signal Corps, CAA, and Sperry chose to stop work on a forty-centimeter sys-
tem, and put their development e¤ort into a ten-centimeter system.19 They be-
gan testing that system in 1942 (see Figure 5.4).
Stanford’s researchers and Sperry’s engineers had always intended to use kly-
strons for continuous wave transmission, and the Sperry Gyroscope ten-centimeter
glide path was built in that tradition. In two works examining the development
of the two early forms of radio transmission, radiotelegraphy (based on inter-
rupted wave signaling) and radiotelephony (based on continuous wave transmis-
sion), Hugh Aitken has argued that the two di¤erent forms of transmission rep-
resented two di¤erent intellectual traditions.20 Practitioners of one could not
easily adapt to the other, because the mental tools needed to work in one tradition
were inappropriate for the other. Although Aitken relied on radiotelephony nar-
rowly defined—we would call it radio broadcasting—continuous wave transmis-
sion with, or without, voice modulation, utilizes the same techniques. Sperry’s
engineers thus worked within a tradition, or technological frame, of continuous
wave transmission. When faced with pulsed transmissions, they adapted poorly.
The pulsed system, which was supposed to supersede Sperry’s work, was a
product of a new organization, the MIT Radiation Laboratory. It had been founded
in 1940 by Vannevar Bush’s National Defense Research Committee to pursue the
development of microwave techniques for use in the war e¤ort.21 The lab’s foun-
dation had been provided by the arrival of British physicist Henry Tizard in the
United States with a new device for generating microwaves: the cavity magnetron.
The magnetron, like the klystron, relied upon electrically resonant cavities to pro-
duce microwaves. The internal structures of the two tubes, however, was com-
pletely di¤erent, with the result that the magnetron was better suited to provid-
ing high power output if operated intermittently. Because higher power translated
directly into longer range, clearly a benefit for weapons systems, the Radiation
Lab focused exclusively on designing equipment for pulsed use, and the lab’s en-
trant into the glide path competition was no di¤erent. Instead of broadcasting
continuously, the Pulsed Glide Path (PGP) transmitted a train of pulses that air-
craft could receive.
The lab decided to produce a glide path after Vannevar Bush assembled a com-
mittee in November 1941 to revisit the blind landing system progress obtained
since his investigation two years earlier. Alfred Loomis, a former industrialist
turned amateur physicist, chaired the committee. The group reinvestigated the
blind landing work being done by CAA, army, navy, and private companies
around the country. It reviewed nine systems in all: Hegenberger, Air Track, the
CAA’s VHF system, the CAA-MIT microwave system, Sperry’s microwave sys-
tem, PGP, GCA, the new 330-MHz glide path being made by ITRM for the army,
and the Lorenz system. The committee recommended that the lab develop the
PGP for production due to the rapid accumulation of pulsed techniques and the
increasing availability of equipment designed for pulsed operations.22 Perhaps
unsurprisingly, it thought little of the possibilities for any of the non–Radiation
118 blind landings
Lab projects, however, and other documents suggest that it thought Sperry’s sys-
tem too complex to work. Hence, the lab decided to make a full-scale push toward
getting the PGP adopted for production by the Signal Corps.
The PGP project had begun informally earlier in 1941, administered under the
lab’s Group 73, the “Landing Group.” J. S. Buck was the project engineer. His goal
was to produce a glide path operated on three centimeters to minimize the radi-
ators’ size. The PGP, which the lab developed in partnership with General Mo-
tors’ Delco Electronics subsidiary, operated by sending a train of pulses through
a horn radiator, producing a narrow beam. The receiver employed an averaging
circuit so that it provided a continuous indication on a cross-pointer type instru-
ment, making the pulses invisible to the pilot. Unfortunately, there is practically
no other information available on the PGP project. I have been unable to locate
either photographs or a detailed description of its operation. Fortunately, there
are documents relating to flight testing and the acceptance of the two microwave
systems still available, despite the lack of technical documentation.
A series of tests held in late 1942 between PGP, the Sperry continuous wave
system, and the SCS-51 system from the ITRM held at Pittsburgh, Cincinnati, In-
dianapolis, and Wright Field, resulted in an army decision to buy both PGP and
the ITRM system. The Sperry continuous wave system was to be abandoned. The
reports do not make clear why PGP was chosen over Sperry’s system, as each ap-
peared to perform equally well (or poorly, as neither put in particularly encour-
aging performances.) The evidence suggests, however, that the AAF had already
come to rely heavily on the advice of the Radiation Lab’s physicists in its elec-
tronics procurement, and it is very clear that the Radiation Lab supported its own
program over Sperry’s.23 The Signal Corps chose to assign production of PGP to
Sperry Gyroscope, a decision apparently based on an analysis that suggested
Delco Electronics did not have suªcient production resources to manufacture
PGP in addition to its other obligations. Sperry, in turn, never put PGP into pro-
duction.
PGP did not reach production for two reasons, related directly to the com-
pany’s previous work with the klystron and Bowles’s continuous wave glide path.
Sperry’s engineers were not experienced in “pulse techniques,” as the Radiation
Lab’s project supervisor put it.24 Sperry’s engineers were unable, or at least very
unwilling, to adopt pulse techniques over the continuous wave design that they
had spent several years developing. The company’s managers also preferred to
leave their engineers assigned to Sperry’s own continuous wave system. The Ra-
diation Lab’s oªcial historian, Henry Guerlac, attributes the nonproduction of
PGP to Sperry’s having lost interest in the project.25 It seems fair to say, however,
that Sperry was never very interested in the first place. Sperry had spent a great
deal of time, energy, and money developing the klystron, upon which the contin-
uous wave system was based and to which it also owned the rights. It could not
recoup that investment if the klystron were supplanted by the cavity magnetron.
The PGP was not in Sperry’s long-range financial interests, any more than it was
within the realm of Sperry’s existing technical experience. Hence, Sperry’s man-
agement chose to keep its engineers at work on the klystron-based continuous
wave glide path throughout the war. This combination of lack of expertise and
lack of management interest spelled the end of PGP, and ultimately of microwave
glide paths.
The Radiation Lab’s intrusion into microwave glide paths thus prevented the
production of a microwave glide path during the war, but Sperry continued to
work its microwave system, hoping that it could convince the AAF and CAA to
replace the temporary VHF system. In 1946, it launched a substantial marketing
e¤ort to get its microwave system adopted as the U.S. and international standard.
Yet the substantial number of SCS-51 and fixed CAA ILS systems installed dur-
ing the war, and the tens of thousands of receivers for it produced during the war,
proved to be an insurmountable obstacle. To suggest how fully the war had biased
the aviation community against a rapid replacement of the “temporary” system,
we examine the decision of the infant Provisional International Aviation Organi-
zation (PICAO) to rely upon the SCS-51 as the basis for an international landing
aids standard. That decision served as the next nail in the microwave glide path’s
coªn.
entrenching the low-tech solution, part ii: picao
The Provisional International Civil Aviation Organization had been formed as

the result of a series of conferences between Britain, the United States, Canada,
and a host of newly restored governments, governments in exile, and govern-
ments of the few noncombatant states in late 1944 and early 1945.26 PICAO’s
function was to establish the framework for a permanent organization to regu-
late international civil aviation. One of the powers granted to the organization by
the member states was the ability to set technical standards for navigation and
communication equipment, so that aircraft flying between nations did not have
to carry di¤erent equipment for each of its destinations. The delegations all rec-
ognized the financial absurdity of that situation.
The problem of landing aids was only one of a number of technical issues that
PICAO needed to resolve and probably the least controversial. The real “battle
120 blind landings
between systems” at the international level took place over short- and medium-
range air navigation systems and was between the British delegation, led by the
famous radar physicist Robert Watson-Watt, and the American group, led by ca-
reer bureaucrat Charles Stanton, who had been administrator of civil aeronautics
until fired by President Roosevelt in 1944.27 This fight suggests some of the
reasoning that likely underlay the less controversial, and therefore less well-
documented, landing aids decision.
Watson-Watt demanded the adoption of the wartime Gee system as the future
international short- and medium-range air navigation system. Gee operated via a
series of ground stations laid out in a network. These stations generated time-
based signals that an airborne receiver detected and decoded as a navigational
grid, which could then be used by a navigator to determine the receiving aircraft’s
position. It was operationally similar to the long-range Loran system but more
precise. Its major advantage, according to Watson-Watt, was that it allowed air-
craft to operate anywhere within the area of broadcast coverage, meaning that air-
craft were not confined to specific, narrow “airways” like those provided by the
four-course ranges in the United States. This, he contended, would allow higher
traªc densities than an airway-based system, which he believed would be neces-
sary to serve Europe’s dense population.28
Stanton, in contrast, promoted the U.S. airways model. Instead of producing
a network of lines from which to derive a position, the U.S. system of radio ranges
simply produced a line of bearing. Pilots then simply flew from range to range
along well-defined airways. These airways were essentially highways in the sky.
Fliers could leave the airways, but outside them pilots had no e¤ective naviga-
tional references other than landmarks on the ground. Initially, the ranges were
the four-course type discussed in Chapter 3, but during World War II a new sort
of omnidirectional range had been devised that provided a reasonably accurate
course in all directions. This was called a visual omni-range (VOR), the visual
meaning that the information was displayed to the pilot on an instrument instead
of aurally. Distance to the range was to be provided by another wartime innova-
tion, called Distance Measuring Equipment (DME). The great advantage of this
VOR-based system was that it was relatively simple and inexpensive, and Stanton
openly ridiculed Watson-Watt over the cost of installing and operating the hun-
dreds of short-range Gee transmitters necessary to cover North America.29 VOR’s
biggest disadvantage, Watson-Watt correctly noted, was that it confined fliers to
specific routes, greatly reducing the volume of airspace useable by instrument
fliers.
Neither man presented a rigorous argument or substantial data to back up his
claims, however, and the delegates chose the U.S. system. Most nations simply
took over the airways stations that the U.S. Military Air Transport Service had in-
stalled to facilitate its worldwide wartime operations, and later expanded upon
them.30 It made little sense immediately after the war for nations to spend a lot
of money on navigational aids when a system suªcient for their immediate, ba-
sic needs had already been installed.
A similar sort of reasoning made the adoption of the USAAF’s SCS-51 as the
international standard completely uncontroversial. SCS-51 had also already been
installed, again by the U.S. Military Air Transport Service, at a number of major
European airfields. It made sense to adopt a system that was available immedi-
ately and relatively cheap. New installations, at $70,000 apiece, also seemed a less
budget-busting solution than the other technologies presented during demon-
strations held in late 1946 for PICAO’s benefit, which included the Sperry mi-
crowave system and the MIT Radiation Laboratory’s Ground Controlled Ap-
proach system, expected to cost $200,000.
Further, as the Belgian delegate pointed out, other nations did not want to be
forced to buy all of their equipment from a single source, and many countries
wanted to be able to manufacture it for themselves.31 He raised this as a condi-
tion for the acceptance of all PICAO standard equipments, and both Watson-Watt
and Stanton quickly agreed. The chosen standard thus could not be proprietary
or contain military secrets. It also had to be fairly easy to manufacture, which was
manifestly not true for microwave equipment.
The most widely reported reason for PICAO’s selection of ILS as the interna-
tional standard, however, was its ease of adaptability to automatic landing. The
automatic approach coupler that General Spaatz had demanded for Eighth Air
Force in 1944 was very important to the plans of PICAO’s technical committees,
which sought greater automation of flying overall. Why they sought to automate
landings was not made explicit, but it seems reasonable to assume that their goal
was the reduction of missed approaches and landing accidents by elimination of
“pilot error.” No one seems to have expected the elimination of pilots themselves,
since the control of a plane’s attitude during landing, which was all the approach
coupler did, was only one part of pilots’ jobs. Landing was the highest workload
phase of a flight, and the delegates no doubt believed that automating part of that
workload would result in safer flights and improve the all-important regularity of
service. When the first president of ICAO, American aeronautical engineer Ed-
ward P. Warner, discussed the technical work being done by the organization, he
focused on the need for automation to reduce workload and errors.32
PICAO’s delegates thus had several good reasons to adopt the low-tech SCS-
122 blind landings
51 over its competition. They had the evidence of its wartime performance, they
were able to fly the systems during demonstrations held in England and at Indi-
anapolis, and they had specific goals of economy, national self-interest, and au-
tomaticity to help sway their decisions. The war, however, had provided the con-
ditions that informed the choices of PICAO’s national delegations.
conclusion
From a scenic overview of the technoscape of aviation dated 1940, the future
of Bowles’s microwave work seemed assured. Microwaves were clearly the future.
World War II altered that technoscape of aviation dramatically, however, by driv-
ing the Army Air Forces to adopt the CAA’s previously “inadequate” system as a
“good enough” expedient to support its wartime operations. At the same time, by
causing Winston Churchill to dispatch Henry Tizard to the United States with
one of the United Kingdom’s most valuable possessions, the cavity magnetron,
the war established the basis of the MIT Radiation Lab, and thus of the Pulsed
Glide Path system. Without the pressing needs generated by the war, the AAF
would not have contracted with International Telephone and Radio for the VHF
glide path, which the Civil Aeronautics Administration incorporated into its
postwar ILS, and the Radiation Lab would never have been founded to produce
its challenger to Bowles’s work. In sum, neither challenger to the Bowles-MIT-
Sperry Gyroscope continuous wave system would have existed. Hap Arnold would
have adopted Sperry’s glide path when it was ready, and the airlines, as their sud-
den reluctance to adopt the CAA’s VHF system in 1940 suggests, would happily
have followed the AAF’s lead. Without the war, finally, the conditions under which
PICAO was established and made its technological choices would also not have
existed. The pressing demands of World War II, then, radically altered the out-
come of the landing aids development process.
The exigencies of war did not merely reverse the fortunes of the two 1940 glide
path projects, however. The MIT Radiation Lab produced yet another challenger
to the CAA’s system, and this one was not merely a di¤erent way to produce a
glide path. Instead, this radar-based system replaced the entire model pioneered
by the National Bureau of Standards system. Instead of using radio beams to ac-
tivate a cockpit instrument, the ground-controlled approach system used radar to
inform operators on the ground of a plane’s position. This system thus chal-
lenged the community’s very conception of how a landing system should work,
shattering the consensus that the National Bureau of Standard’s pilot-centered
model was the one best way for a landing aid to function.
chapter six
The Intrusion of Newcomers
The blind landing systems detailed thus far worked in accordance with a sin-
gle model, for lack of a better word. All of them, whether based upon cables or ra-
dio transmissions, provided a signal or set of signals that directly operated one or
more instruments in a plane’s cockpit. They were a primitive automated remote
control. Pilots translated the instruments’ readings into information meaningful
to them in accordance with their training and experience and then maneuvered
their aircraft. Because pilots retained control of the information translation
process, they also retained their own autonomy. They needed no help from peo-
ple on the ground, although, of course, they required a great deal of assistance
from devices on the ground.
The preceding several chapters have shown a process of building a consensus
within the aviation community toward what was less an exemplary artifact than
a preferred model of a solution. By 1940, every group involved in the process of
building a blind landing system had accepted that a complete system had to in-
clude marker beacons, a localizer, a straight glide path, and a set of receivers in
the plane. No other possible solution was under serious consideration. This was
the National Bureau of Standards’ model, and it belonged to a category of systems
that functioned under what I called in Chapter 4 the pilot control model.
It was not the only way to build a blind landing system, however. Three years
after the National Academy of Science’s (NAS) report was so painfully hammered
out, a radically di¤erent way of directing aircraft to a safe landing came into ex-
istence. The MIT Radiation Laboratory, in particular, future Nobel laureate Luis
Alvarez, was responsible for challenging the pilot-control model. As a new orga-
124 blind landings
nization, the Radiation Lab certainly was not a member of the dominant techno-
logical frame, and its leadership reinforced its outsider status by choosing to re-
cruit from the ranks of nuclear physicists, as opposed to radio physicists and en-
gineers. The Radiation Lab was thus able to create and introduce a radically new
model solution to the blind landing problem. This had far-reaching e¤ects. The
old model left control of the aircraft in pilots’ hands, where it had always been;
the new system placed control on the ground. Not everyone proved happy with
that sort of progress.
inventing a new model:

the cyclotroneers invade aviation
A month after President Franklin Roosevelt had signed the NAS report on the
status of blind landing systems, Vannevar Bush was back in his oªce, proposing
a scientific research organization devoted to developing new weapons, devices,
procedures, and materials for the war that both men believed the United States
could not avoid much longer. The result of that June 12, 1940, meeting with Roo-
sevelt was the National Defense Research Committee, chaired by Bush, which
was to oversee research and development activities directed primarily at military
needs. The committee’s original structure was poorly designed for the operation
of a large bureaucracy in oªcial Washington, however, and in June 1941, Bush
obtained another executive order from Roosevelt submerging the committee un-
der the Oªce of Scientific Research and Development (OSRD), once again, un-
der Bush. The committee continued to exist but was restructured into numbered
divisions. Bush initially assigned all things microwave to Karl Compton’s Divi-
sion D, but after the reorganization, it fell under Division 14, and more particu-
larly, the microwave committee chaired by Alfred Loomis.1
The spark that precipitated the MIT Radiation Laboratory was the arrival on
U.S. shores of Henry Tizard, a British physicist sent to secure American assis-
tance in making use of several key British inventions. Tizard brought with him
the resonant cavity magnetron, which was capable of producing microwaves at
several orders of magnitude higher output than was the klystron that American
radars, and Bowles’s microwave landing system, were based on. Demonstrated
to Alfred Loomis and Karl Compton on September 19, the magnetron’s potential
convinced the two men that a central laboratory to develop applications for it was
necessary. Another meeting and demonstration held at Loomis’s Tuxedo Park re-
treat in New York on October 12 and 13 cemented in his mind the need for an in-
dependent lab, and he convinced Bush. He and Bush chose MIT as the best site
the intrusion of newcomers 125
for the lab, largely due to its existing microwave experience. Loomis and Bush
badgered Frank Jewett into accepting their decision, and the three then lobbied
Karl Compton, president of MIT, to accept it. He did so on October 17.2
With a site for the lab approved, the three men needed a sta¤. They decided to
recruit from the ranks of nuclear physicists, since members of that discipline had
developed extensive experience in dealing with high-frequency energy in the
course of their cyclotron experiments. As historians John Heilbron and Robert
Seidel put it, “cyclotroneers and their fellow travelers did not fear big projects, did
not disdain to scrounge when necessary, did not insist on perfection or protocol.
They were ideal people for crash programs.”3 With that understanding, Bush,
Loomis, and Jewett turned to Ernest Lawrence to head the lab, but he chose to re-
main with his own work in Berkeley. Lawrence instead recruited Lee DuBridge,
a former student of his who was then chair of the physics department at the Uni-
versity of Rochester, to fill the position. I. I. Rabi, of Columbia University, and F.
Wheeler Loomis from the University of Illinois also signed on. For this story, how-
ever, the most important recruit Lawrence sent was another of his protégées, Luis
Alvarez, who arrived in mid-November. Alvarez recruited a friend from his grad-
uate student days at Chicago, George Comstock, who had taken up a teaching po-
sition, and slightly later, Lawrence Johnston, whom Alvarez had left in charge of
his lab at Berkeley when he moved to MIT.4
Alvarez had joined Lawrence’s sta¤ in 1936 after completing his thesis under
Arthur Compton at the University of Chicago, where he had learned to build
things and to navigate a library, although he considered his education in physics
there to have been poor. At Lawrence’s lab, though, he learned the ropes of his
chosen field. He finally read the physics he was supposed to have learned, and,
more importantly, he learned to work in a team, which Chicago had frowned
upon. Because Lawrence’s cyclotron required a great deal of labor to maintain and
operate, teamwork was necessary to keep the machine running. The machine re-
quired a six-man crew to run it each day, in shifts of two.5 Alvarez blamed the reg-
imen imposed by the machine for the lab’s failure to make some of the funda-
mental discoveries in physics made elsewhere, but it served him in good stead at
MIT, where his education in teamwork and hands-on operation of complicated
equipment proved essential.
Physics was only one of his two careers, however. The other was aviation.6 Both
of his careers are important to this story. Alvarez had earned a private pilot’s li-
cense in 1933, after three hours of instruction. He did not learn instrument flying
until early 1942, when Aviation Chief Machinist’s Mate Bruce Griªn taught him
at the Squantum Naval Air Station in Massachusetts. He was therefore only a
126 blind landings
“contact rules” pilot when he conceived of his famous ground-controlled ap-

proach system, or GCA, in 1941. He was qualified to fly only within sight of the
ground. His lack of instrument training meant that he had no mental commit-
ments to the model upon which the other blind landing systems we have seen op-
erated, although he knew their general principles. He was not set in the way it
had always been done, nor was he a member of the dominant technological frame
based upon the ideal of pilot control. He was free to imagine something com-
pletely di¤erent.
According to Johnston, Alvarez became interested in the blind landing prob-
lem while recuperating from gallbladder surgery in Rochester, Minnesota, in the
summer of 1941. When he returned to MIT in August, he saw a demonstration
of a gun-laying radar that gave him the basic idea for the GCA. The gun-laying
radar, called XT-1 by the lab, was designed to fulfill one of the three primary tasks
the microwave committee had assigned to the lab: an airborne intercept radar, the
gun-laying radar, and a long-range air navigation system.7 The XT-1, later given
the oªcial name SCR-584 (the SCR designates the equipment as Signal Corps
Radio gear), was a truck-mounted, conical scan radar that automatically tracked
an aircraft, and fed range, azimuth and elevation to an analog fire control com-
puter that predicted the aircraft’s future position, which could then be used to aim
gunfire accurately.
While watching a test of the rooftop prototype XT-1, Alvarez realized that if
radar could track an aircraft accurately enough to hit it with an artillery shell, it
certainly could track one accurately enough to guide it to a runway. The basic idea,
then, was to employ radar information to facilitate blind landings, instead of us-
ing a set of radio guide beams, as the various incarnations of the NBS system had.
As it turned out, given the existing technologies, radar could be used in at least
two ways to land aircraft, and both were used during World War II. One was Al-
varez’s ground-controlled approach; the other, a field innovation, was the blind
approach beacon system (BABS).8
Alvarez’s conception was to use the XT-1 to track an incoming plane and to use
some modification of the analog gun director to provide an operator on the
ground with range, bearing, and altitude information that he could then pass on
to the plane’s pilot via voice radio, which every military aircraft carried. The radar
operator on the ground could therefore talk planes down. This was not a new idea.
The navy already used a talk-down system aboard its aircraft carriers, with the
“talking” done not by voice but by light wands in the hands of a crewman sta-
tioned on the ship’s deck (the Landing Safety Enlisted, LSE). There was, of course,
no radar involved in that process, but because the navy was already familiar with
the basic principle of talking down, and in fact never did find a better way to fa-
cilitate carrier landings, it became interested in Alvarez’s modification of the ba-
sic idea early on and after the war was its biggest proponent.9
Alvarez’s idea got its first important boost from the battle against fog that the
Royal Air Force (RAF) was fighting in Britain that fall. In some raids, Bomber
Command was losing as many aircraft in landing accidents as it was over enemy
territory due to the poor performance of its standard beam approach (SBA) sys-
tem. Vannevar Bush therefore assembled another committee in November 1941
to revisit the blind landing system progress obtained since his investigation of
two years before. Alfred Loomis chaired this committee, and Alvarez, who by then
had been put in charge of Lab Division 7, the beacons group, was invited to be a
member along with several representatives of the armed services. One outcome
of this committee was pulsed glide path (PGP), as we saw in the previous chap-
ter.10
The committee recommended that Alvarez pursue his ground-controlled ap-
proach idea quickly. Overlooked in all of the previous work in blind landing of air-
craft had been the problem of small aircraft, which could not carry the receivers
and instruments that the various versions of the NBS system required. This was
particularly important for fighter aircraft (or pursuits, as they were called), which
could a¤ord neither the weight of the receivers nor the drag imposed by the two
antennae without sacrificing performance. This had simply not been an issue be-
fore 1941, since without radar, fighters could not find targets in bad weather and
therefore had no reason to be flying. With the development of radar, however,
fighters could function in poor weather but only as long as visibility was still good
enough to land. Because all of the equipment for GCA was on the ground, it im-
posed no weight or drag penalties, making it ideal for fighters and other small
aircraft. GCA promised to complete a technological suite capable of allowing
fighters all-weather operations.
The blessing of Loomis’s committee permitted Alvarez to pursue GCA full
time. He appointed Lawrence Johnston project engineer, and with George Com-
stock and David Griggs, a Harvard geophysicist and amateur pilot, set to work
building an optical version of his GCA idea as a proof of method experiment. Two
modified theodolites and a range-only radar were combined through a director
device to produce a distance-to-landing indication, and two readouts of departure
from an ideal glide path were established for the aircraft. During March 1942, the
team put the optical equipment to work, directing Chief Machinists’ Mate Griªn,
flying a Grumman J2F “Duck,” to safe landings via voice radio.11
The prototype XT-1 finally became available to the group during May 1942, and
128 blind landings
the navy invited Alvarez’s group to Oceana Naval Air Station to try their landing
idea out with the antiaircraft radar that had inspired Alvarez’s idea in the first
place. There, however, the project also nearly ended. The same reflection phe-
nomenon that had caused so much grief for the NBS system nearly undid GCA.
The system tried to track the aircraft’s underground mirror image as well as the
plane itself. Figure 6.1 shows the two types of mirror-image phenomena that
XT-1 was susceptible to at low angles. As the figure indicates, the XT-1’s antenna
did not have a narrow enough reception pattern to keep the energy reflected from
the ground from returning to the receiver at the low angle (3 degrees) above the
ground that landing the aircraft required. The system had no way to distinguish
the real airplane from its image, and its automatic tracking system hunted be-
tween the real and virtual planes, making it useless.12 No simple modification to
the XT-1 would turn it into a blind landing system. Alvarez had to completely re-
design it.
Depressed over his misfortune, Alvarez met with Loomis in Boston. Alvarez
later reminisced that during dinner at the Ritz Carlton, Loomis told him that “we
both know that GCA is the only way planes will be blind-landed in this war, so we
have to find some way to make it work.”13 Loomis forbade him from leaving the
table until he had figured out a solution. The two agreed that the best approach
was to replace the single XT antenna with two antennae, one vertical and one hor-
izontal, which were to be mechanically scanned through narrow arcs. By making
the antennae beaver-tail shaped, they narrowed the beam patterns, reducing the
probability of reflections from anything but the target. These two antennae pro-
vided the short-range, highly precise indications needed to e¤ect a landing, while
a third search radar enabled the GCA’s crew to direct aircraft into the narrow pre-
cision beams.
With Loomis’s pledge of support to buy ten of the resulting systems with
OSRD funds, Alvarez’s group got to work assembling the Mark I GCA. Alvarez
thought it would help speed production if the contractor sent engineers to the lab
to learn as it was assembled, so OSRD asked the Signal Corps to suggest possi-
ble contractors. The Signal Corps originally selected Paramount Pictures to man-
ufacture it, based apparently on the presence of a radio engineer named Homer
Tasker on its sta¤. The movie company decided that it did not wish to enter the
electronics business and agreed to loan Tasker to a Los Angeles company that
made household radio receivers, Gilfillan Brothers.14 Gilfillan thereby won the
contract for the first ten units. Gilfillan dispatched Tasker and three others to the
Radiation Lab in late 1942 to begin learning the system from its experts.
Figure 6.1. Two mirror effects prevented Alvarez’s original idea for GCA from working
effectively. In the “primary mirror effect,” the aircraft reflected both radio energy directly
from the antenna and energy which first bounced off the ground. The receiving antenna
thus “saw” two aircraft, the actual plane, and its virtual reflection underground. In
Lloyd’s mirror effect, radio energy reflected off the ground interfered with radio energy
following a direct path to the aircraft. The resultant of that interference is the dotted
radar beam emanating from the “antenna’s image.” To the operator in the truck, the
aircraft appeared displaced from its actual position. Lawrence Johnston, “Radiation
Laboratory Report 438,” file RB 334.8, box 1428, Chief Signal Officer Central Files, RG
111, National Archives.
The Mark I consisted of a gasoline-driven generator powering the radars

mounted in one truck, with the antennae on its roof. A second truck contained
the radar screens and the voice radio sets that controllers used to communicate
directions to pilots. Both trucks were parked fifty feet to the left side of the run-
way in use, from the pilots’ point of view. For production, Gilfillan’s engineers
replaced the second truck with a large panel trailer. The first production unit,
named Mark II, is shown in Figure 6.2.15 The engineers moved the antennae to
the trailer, with the search antenna on upper right, elevation on the left corner,
130 blind landings
Figure 6.2. AN/MPN-1 #1, at the Gilfillan plant in Los Angeles. The trailer contained the
search radar (upper left), the precision azimuth antenna (lower center), and the precision
elevation antenna (upper right), in addition to the operators’ position (internal). The
truck contained the power generators and air conditioning units. The individuals are
not identified. Author’s collection.
and azimuth in the front center of the trailer. The Mark II prime mover, the cov-
ered truck, contained a generator and air conditioning systems to keep the hun-
dreds of vacuum tubes in the trailer from overheating.
Alvarez’s group completed the Mark I in November 1942, and it was demon-
strated to observers from the U.S. Army and the British Royal Air Force from mid-
January to early February at East Boston airfield. A few days later, Gen. Harold
McClelland, the director of Army Air Forces’ (AAF) technical services, invited the
group to demonstrate the system to more senior oªcers on Valentine’s Day at
National Airport. After several days of fixing wires and replacing tubes damaged
by the long drive, Alvarez and his group successfully demonstrated the system to
the AAF’s satisfaction on the seventeenth.16 Hap Arnold had anticipated success
and had already requested that McClelland buy ten for further tests; by the end
of March, the AAF had fifty-seven on order.
Much of the GCA testing before this oªcial demonstration had been done at
Quonset Point Naval Air Station in Rhode Island, using navy planes and pilots,
and was therefore already well known to the navy. The base commander had un-
stintingly praised the system, particularly after a January 1, 1943, incident in
which a group of 3 PBYs was caught in a snowstorm and became lost. He had
called Alvarez’s group from the control tower and asked if they could bring the
planes in. Alvarez talked them in, after gaining the trust of the pilots, who did
not, according to Larry Johnston, even know radar existed. Alvarez was able to
convince the pilots he could see them by talking them through various maneu-
vers before trying to land them. Once he had their trust, he was able to bring them
in. The base commander reported the feat to the commander of Fleet Air at
Quonset Point, and he soon requested that copies be procured by the navy.17
The initial orders from the AAF and the navy amounted to fifty-seven and
twenty, respectively, in addition to the ten that OSRD purchased. The Army Sig-
nal Corps chose to let the AAF contracts to Gilfillan Brothers, while the navy chose
to rely upon one of its traditional contractors, Bendix Radio. Both army and navy
expected to receive their orders by the end of 1943; in fact, neither received any
equipment before mid-1944, and neither service was able to provide them to com-
bat theaters until 1945. GCA proved extremely diªcult to get built because Al-
varez was not as satisfied with it as the armed forces had been. Alvarez’s group
therefore spent the spring of 1943 converting the Mark I to the Mark II, which re-
lied upon linear array antennae like those employed on one of Alvarez’s other cre-
ations, the Eagle blind bombing radar. The advantage gained with this change was
elimination of the mechanical scanning of the two precision antennae, which was
a severe maintenance problem. The inertia of the large antennae destroyed the
drive gearing quickly. The Mark II design also incorporated several changes to the
radar displays, the most important of which was replacement of the B scan scopes
with plan position indicator (PPI) displays, which showed the straight approach
as a straight line, instead of a curve. With the improved presentation, the com-
plex mechanical error readers that the Mark I had employed to mathematically
transform the curved display into useful information about the actual straight ap-
proach path were no longer needed, reducing the number of operators and the
complexity of the equipment.
By making these changes, Alvarez’s group delayed achievement of a key mile-
stone in production engineering, design lock-in. Without a fixed design to work
with, the contractors could not make any headway toward establishing a manu-
facturing process for the system or obtaining parts. Although Tasker moved to
the MIT Radiation Lab in late 1942 and became essentially part of Alvarez’s team,
he was not able to provide final specifics and had to frequently send changes back
132 blind landings
to Gilfillan. The company then had to alter the orders it had placed with its sub-
contractors, delaying the process further. This was one consequence of Alvarez’s
decision to change the design.
Another was that the contractors did not have the extensive diagrams they
needed to engineer system components because the Radiation Lab group had not
yet made them. Throughout 1943, with the design in flux, they did not finalize
the diagrams. The Signal Corps’ Monmouth procurement district oªce, which
was responsible for overseeing the purchasing and production of GCA, criticized
the lab for refusing to turn over the prototype so that it could be reverse engi-
neered to make the necessary drawings. Without drawings or the Mark I, the con-
tractors had nothing to work with. Complete diagrams did not reach the con-
tractors until October 1943, and Tasker had to have many of them redone. He
borrowed underemployed electrical engineers from the aircraft manufacturing
plants in the Los Angeles area to do that, but this delayed completion of the first
unit into February 1944.18 Gilfillan did not complete the first ten OSRD units un-
til May, after which full-scale production began.
The Mark I was also unavailable to the contractors because Alvarez still seems
to have thought that he needed to sell the system to the military brass, despite the
rapid placement of orders for GCA in March. He arranged to pack up the Mark I
in June and ship it to Britain for demonstration there, while Larry Johnston joined
Tasker at Gilfillan to engineer the Mark II. In Britain, Alvarez’s team set up the
Mark I at Elsham Wolds, a bomber command base, for more demonstrations. The
British appointed an oªcer named Arthur C. Clarke to learn the equipment. Al-
varez’s group trained Clarke and his women (the RAF chose to use members of
its Women’s Auxiliary Air Force to serve as crew under the supervision of male
regular oªcers) in maintaining and operating Mark I, and they in turn trained
other crews once Alvarez and his group left in August. The Mark I moved to
Davidstowe Moor in Cornwall, a coastal command base, after the Elsham Wolds
tests. It remained there as the centerpiece of the RAF’s GCA school until mid-
1944, when its antenna drive gearing failed. The Mark I was cannibalized for
spares to use in the first three preproduction GCAs, which were given by the Ra-
diation Lab to the British, as they had operators trained to use them.19 The RAF
thus adopted GCA as quickly as had the AAF and the U.S. Navy.
In December 1944, George Comstock, who replaced Johnston as project en-
gineer when Johnston left for Los Alamos in late 1943, reported that Gillfillan had
manufactured forty-seven GCA sets for the army, with fifty-five more on order.
Due to the extensive delays in production, the army had also placed an order for
one hundred more sets from Federal Telephone and Radio, only a few of which
had been produced by that date. Bendix had also produced only a few of the eighty
sets for which the navy had contracted (sixty of which were intended for the
British). Therefore, a total of something like sixty Mark II sets existed at the end
of 1944, but none reached combat theaters until February 1945.
The bottleneck after 1944 was not equipment availability, then, but operator
training. The Mark II GCA required a much larger crew than the original XT-1 in-
spired idea. Figure 6.3 shows the internal seating of the trailer. A traªc director
operated the search set, while a plane selector designated a specific aircraft to
land. The azimuth and elevation trackers operated the azimuth and elevation pre-
cision radars, respectively, and passed corrections to the controller, who directed
the incoming pilot. Stu¤ed into a trailer, the five-person GCA crew required a
good deal of teamwork among themselves to properly operate the system, in ad-
dition to the teamwork necessary between the controllers and pilots whom they
were directing. The radar operators had to feed the controller the proper infor-
mation quickly and accurately; the controllers had to reissue the information in
the form of steering commands (“right 3 degrees,” etc.) to the pilots; the pilots
had to respond immediately to the controllers’ commands, especially during the
last few moments of a “blind” landing. Delay there could be fatal. GCA required
a great deal of human coordination to work, and training with the actual equip-
ment was the only way to forge such a team.
Because the Mark I had stayed in England, there was no equipment on which
to train operators in the United States until mid-1944, when Gilfillan finally de-
livered the ten preproduction units. Nine of those went to serve as the basis of
training schools, three each to the AAF, the RAF, and the U.S. Navy, while the
tenth went to the Radiation Lab for further research. Training programs took time
to establish, and more time to produce their first class. The AAF’s training pro-
gram, for example, was a twelve-week course and initially graduated thirty-three
teams in early 1945.20 By the German surrender in May, only eight GCAs with
crews had reached AAF units in the European theater (all at fighter/bomber
fields), and I can document only two in the Pacific theater by the Japanese sur-
render in August, although Radiation Lab documents indicate that ten crews had
been allocated there.21 A total of between ten and twenty GCA units, therefore,
were in operation by the end of the war. This was a tiny fraction of all the airfields
in use during the war, and GCA’s contribution to the war e¤ort was small. It was,
however, an instant sensation among pilots.
The two great selling points for GCA’s talk-down method were that it did not
need special equipment in the aircraft and, more importantly, that it required no
pilot training. Whereas all variations of the NBS system required extensive pilot
134 blind landings
Figure 6.3. AN/MPN-1 GCA internal trailer seating. The traffic director on the right
operated the search radar, while the plane selector assigned an aircraft to the controller
for final approach. The traffic director maneuvered the selected plane (via voice radio
to the pilot) into the precision radars’ beams, whereupon the controller took over the
approach. The azimuth and elevation trackers fed the controller information from the
precision scopes to allow completion of the “blind” landing. Author’s collection.
training to read and properly interpret its instruments and to gain the pilots’ faith
in the system itself and their ability to use it, the GCA only required the ability to
respond to voice commands. Military pilots, at least, got that training beginning
with their first day in flight school. Particularly important during the war was the
human contact that GCA provided. Since crews often returned from missions
with damaged aircraft, exhausted and sometimes with injuries to make their land-
ings even more hazardous, being talked in by fellow soldiers was reassuring in a
way that the NBS-type systems could never be. From the pilots’ point of view, GCA
was nearly miraculous. It could see them when they could see nothing and tell
them where they were when they had no idea. The contrast with the NBS system,
with its training requirements, its equipment needs, and its environmental va-
garies, could not have been more striking.
Military pilots, and especially navy pilots, took to GCA immediately. In part,
the system’s basic requirements placed no demands on them other than follow-
ing orders. Thinking was done by someone else; as RAF veteran Frank Griªths
put it, “any fool could use it.”22 Perhaps more importantly, however, GCA fit the
culture of wartime military flying extremely well. Teamwork was required among
bomber crews because no one could manage all of the plane’s functions alone,
any more than Lawrence’s cyclotroneers had been able to individually manage
their massive research instrument. Fighter and fighter-bomber pilots were also
trained to work as teams with their squadron mates and wingmen, both as a mat-
ter of collective defense and to increase combat e¤ectiveness. Hence, GCA’s de-
mand for teamwork between air and ground fit a model of behavior for which
military pilots were already prepared.
This was the same model of behavior that Alvarez claims for Lawrence’s lab-
oratory, and to that esteemed place we should return to find GCA’s genesis. While
all of the inventors and innovators working under the NBS model attempted to
reduce a blind landing system to one-man operation by a pilot from a lofty,
bumpy, and moving perch—thus placing yet another burden on the pilot—Al-
varez spread the work of a blind landing out among a group specially trained for
it. Just as teamwork allowed Lawrence’s cyclotron to keep going, teamwork pow-
ered GCA. Alvarez certainly did not learn the benefits of teamwork at Chicago,
which he described as a solitary experience. He learned it from his experiences at
Berkeley and built it into his invention at MIT. We must therefore place some of
the praise for GCA—and some blame, for not everyone loved it—squarely at the
feet of the cycolotroneers.
conclusion
The MIT Radiation Lab closed down shortly after the war, and the intruders
from nuclear physics mostly, but not entirely, returned to their labs. Some who
did not became very influential in aviation’s technical community. George Com-
stock, for example, who completed the GCA project after Alvarez and Johnston
left for Los Alamos, was one of the few who stayed in aviation for the rest of their
lives. He joined the Airborne Instruments Lab, became its president in 1960, and
served on President Kennedy’s Project Beacon commission, formed to recom-
mend a new air traªc control system.
More important than the individuals that the lab gave to aviation, and far more
controversial, was its bequest of the ground-control model. Before World War II,
pilots’ authority and autonomy in the air was unquestioned because no one but
136 blind landings
the pilots knew where they were and what the status of their aircraft was. The
rudimentary air traªc control system that the Bureau of Air Commerce had be-
gun piecing together in 1936 actually had no communication with aircraft at all,
let alone any means to track them. It exerted no real control over pilots. Instead,
it served as a means of exchanging information between airports as to expected
arrivals and departures. It was, in its best days, a flight-following system, not re-
ally a control system. The wartime development of radar, however, promised—or
threatened, depending on one’s point of view—to place pilots under continual
surveillance and control from the ground. The intrusion of Lawrence’s cyclo-
troneers thus dramatically altered the technoscape of American aviation during
World War II.
The war itself, however, was the driving force behind these changes. Without
World War II, there would have been no Radiation Lab to challenge the dominant
model of control. The war caused politicians to make available far greater re-
sources for technological development than had previously been contemplated,
and that in turn enabled investigation into a variety of di¤erent possible landing
aids that would never have been pursued otherwise. The war established condi-
tions within which unparalleled creativity was possible and rewarded, and in
which any landing aid could get a trial by fire. The politics of blind landing, in
essence, were suspended during the war because the need was so great that the
AAF was willing to try virtually anything that promised to work better than what
little it had, and it had the money to spend on multiple avenues of research. It
also had no competition for control over aviation during the war and therefore
did not have to negotiate with the airlines or with the CAA. The army could dic-
tate the type of infrastructure created to serve the needs of wartime aviation with-
out consideration of other organizations’ needs, and equally importantly, AAF
leaders came to rely upon the advice and expertise of academic scientists to help
them make technological decisions. The war thus radically altered the environ-
ment in which landing aids research was done.
Wartime suspension of the politics of blind landing did not last longer than it
took the ink to dry on the Japanese surrender document, however. A vicious po-
litical battle took place after 1945 over the selection of a common landing aid for
the United States. Chapter 4 examined the negotiation of a standard system be-
fore the war to suggest how diªcult reaching agreement could be even when all
parties agreed to the basic form of the desired solution. After the war, no such
agreement existed, with some organizations supporting GCA, and others sup-
porting SCS-51. With the appropriate model of the solution itself open to ques-
tion, the choice of a landing aid became deeply political.
chapter seven
The Politics of Blind Landing
In May 1947, Charles V. Murray told readers of Life magazine that the Civil
Aeronautics Administration (CAA) had “declined into a mental outlook more
befitting the guild of harness-makers.” At issue was CAA’s continued insistence
on deploying its “old-fashioned” instrument landing system (ILS), which di¤ered
from the army’s wartime SCS-51 only in that it was housed in buildings instead
of in a truck, and not the army/navy AN/MPN-1 ground-controlled approach
(GCA) system. GCA’s glories had been trumpeted throughout the U.S. aviation
press as the best solution—and by Aviation News and AOPA Pilot as the only so-
lution—to that now-ancient bugaboo of fliers, the blind landing.1 Criticism of
CAA’s “anti-radar” policy began in late 1945, almost immediately after cessation
of hostilities, and became increasingly vitriolic throughout 1946. The August
1946 decision of the Collier Trophy Committee, custodians of the most presti-
gious aviation honor, to present the 1945 award for “the greatest achievement
in aviation in America” to Luis Alvarez was merely one more blow to CAA’s
credibility. Over the next few months, Congress joined the media criticism and
launched an investigation into air safety.2
Part of the controversy descended from the landing aids decisions made by the
U.S. Army Air Forces (AAF) and the U.S. Navy. The navy had decided by late 1945
to abandon SCS-51/ILS completely in favor of AN/MPN-1/GCA, while the AAF
publicly promoted GCA.3 Nonetheless, the AAF continued to install both mobile
SCS-51s and fixed ILS stations throughout the United States and along overseas
routes. The majority of its GCA sets, moreover, stayed in storage despite its own
rhetoric. Manpower shortages kept the army from utilizing the GCA sets it pos-
138 blind landings
sessed, regardless of policy decisions. The AAF had chosen not to choose, in
essence.
The aeronautical press bore a bias at least as deep as the one it attributed to
CAA. It took at face value the words of GCA enthusiasts, even when their “facts”
were somewhat questionable. The historian will search in vain through the pages
of Aviation News, for example, for a single positive mention of ILS, despite its solid
war record. Instead, one will find page after page devoted to reporting on the num-
ber of airmen “saved” by GCA. Only by referring to the National Archives can one
discover that there were more than twice as many ILS systems in successful op-
eration at military airfields in the United States in 1946 as GCA sets. Biased re-
porting created much heat but shed little light on significant issues.
The CAA’s administrator in the first post–World War II years was Theodore P.
Wright, a well-respected aeronautical engineer. Appointed in August 1944, he re-
placed Charles Stanton, who then became assistant administrator. Wright had
come to Roosevelt’s attention through his service as a member of the Aircraft Pro-
duction Board, where he had earned a reputation for working well with the mili-
tary. He also had foreign contacts and was a member of the Royal Aeronautical
Society, which President Roosevelt had hoped would better enable him to work
with the new Provisional International Civil Aviation Organization. Wright was
also a believer in what Joseph Corn has called the “winged gospel,” and he wres-
tled with the inconsistency between his interest in private aviation and his duty
to the primary users of the airways, whom he felt were commercial and military
fliers.4 As a result, he spent most of his term being attacked by the Aircraft Own-
ers and Pilots Association for not doing enough to help the “little man” gain ac-
cess to the airways. The ILS/GCA debate is one place where Wright most spec-
tacularly failed private fliers.
The issue of access to the skies is a deeply political one. Both GCA and ILS
worked as e¤ective landing aids, but how they worked—not how well—caused
private and some military fliers to line up in the GCA camp, while the Air Line
Pilots Association and most airlines supported the CAA and its ILS. In short,
GCA was easier and cheaper for an “average” pilot to use, which is precisely why
airline pilots were opposed to it. GCA enthusiasts saw it as a way to promote a
democratic vision of aviation, one in which every adult could own and operate an
airplane. The Air Line Pilots Association, by contrast, saw it as a threat to pilot au-
tonomy and a means of deskilling the profession. Ultimately, however, the out-
come of the controversy was determined by how the two systems allocated costs.
the politics of blind landing 139
the controversy
The great debate over ILS and GCA took place in the aeronautical and main-
stream media between December 1945 and the end of 1947, and in a congres-
sional investigation into aviation safety which ran from January to March 1947.5
The media universally took the pro-GCA position, granting no quarter to ILS sup-
porters, all of whom were painted as “conservatives” refusing to adopt or even se-
riously consider new technologies, particularly radar. The leading GCA support-
ers were J. B. Hartranft Jr., head of the private-aviation-oriented Aircraft Owners
and Pilots Association (AOPA) and editor of its newsletter, AOPA Pilot; and an
aviation writer named William Kroger, who wrote primarily for the magazine Avi-
ation News. The defenders of ILS were CAA’s “old guard,” led by Charles Stanton
and the president of the Air Line Pilots Association, David Behncke.
A cartoon appearing in AOPA Pilot in June 1946 illustrates the participants
and their positions in this rather complex process, at least from Hartranft’s per-
spective (Figure 7.1). The man in the flight suit represents private fliers (and by
implication, all fliers, as no ALPA character appears), while the “fat cat” business-
man next to him represents the CAA. The absence of a figure for the Air Trans-
port Association (ATA) implies that the CAA figure represented both the agency
and its preferred client, ATA. CAA’s oªcials were hardly well paid; Wright’s
salary went up when he left the administrator’s oªce for Cornell University.
Clearly, AOPA considered CAA biased toward business and combined CAA and
ATA into a single figure.
The cartoon also accurately represents the positions of the armed services. A
navy oªcer seeks to help convince the reluctant agency that GCA was the right
choice. The navy had adopted GCA for several reasons. First, the ILS was unus-
able aboard aircraft carriers, as the beam required a stable reflecting surface. Sec-
ond, the navy had always used a form of “ground-controlled approach” aboard air-
craft carriers. A safety petty oªcer guided aircraft aboard using light wands and
hand signals, so that Alvarez’s GCA was merely an extension of its existing op-
erating procedures. Third, as the navy’s head of aircraft research told an inter-
viewer, GCA was cheaper from the navy’s point of view. Admiral Luis de Florez
explained that while the GCA ground installation cost twice as much as ILS, the
cost of the onboard ILS receivers greatly outweighed the ILS ground installation’s
lower cost. But the real savings from GCA came from reduced training costs. ILS
required training thousands of pilots and, since skills atrophied when they went
Figure 7.1. “Only GCA can save us!” The Aircraft Owners and Pilots Association’s view of the ILS/GCA conflict. AOPA Pilot, June 1946,
52a. Courtesy of Hagley Museum and Library.
unused, necessitated recurrent training every few months. GCA had no pilot
training costs, only operator costs. Hundreds of pilots might use a GCA installa-
tion in the course of a day, but a new two-man GCA being developed would need
only eight operators and one maintenance technician to serve them round the
clock.6 From the navy’s point of view, GCA was an economical choice.
While the navy’s position in the conflict was unambiguous, the cartoon makes
clear that the Air Force’s was not. The agency’s leadership refused to take a pub-
lic position on the issue, arguing instead that the two systems should be seen as
complementary, not as rivals. This reflected internal arguments over the two sys-
tems’ relative merits. In 1945, the AAF’s chief instrument instructor pilot, Lt. Col.
J. B. Duckworth, had argued that neither system was completely suitable and rec-
ommended further development. The agency then ran a series of tests during
1946 on the two systems at the Civil Aeronautics Administration’s Indianapolis
center which resulted in a similar set of recommendations. In its review of those
tests, the Air University, which was charged with professional education of Air
Force oªcers, argued that GCA, while it had put in a better showing during the
test series, was problematic because it violated the principle of “air leadership”
by placing control of aircraft in ground controllers’ hands—precisely the pilots
union’s argument. It also recognized that ILS required greater levels of training,
was more diªcult to fly, and depended on special equipment in the aircraft. The
Air Force’s leadership ultimately decided that it wanted a fully automated system
that allowed pilots to decide whether to use it or not; until it developed that sys-
tem, it supported installing both ILS and GCA.7
Two major players are missing from the cartoon. The lack of representation
for ALPA, the commercial pilots union, suggests that AOPA did not wish to re-
veal division among pilots over the two systems. It may also indicate that AOPA
considered ALPA illegitimate. The two organizations rarely saw eye to eye, with
the union consistently fighting for tighter regulation of the industry, while AOPA
preferred a laissez-faire approach. AOPA may have felt that commercial pilots did
not deserve separate representation. The other missing party, of course, is the
U.S. Congress. That omission is forgivable. No one could have foreseen the po-
litical attention that the great ILS/GCA controversy eventually drew, but perhaps
someone should have. Ultimately, the choice between the two systems was a po-
litical one, and thus well within that body’s traditional rights.
If I have accurately represented the positions of the groups involved in the
blind landing controversy, no single issue dominated the debate. Commercial pi-
lots perceived in GCA a threat to their professional autonomy, and private pilots
saw it as a means to safer (and less expensive) personal flying.8 Autonomy was
142 blind landings
never an issue for navy pilots.9 Cost was that agency’s major concern. And au-
tonomy was just one of several issues for the Air Force, while cost appears to have
been irrelevant. Di¤erent issues mattered to di¤erent organizations.
The great controversy began inauspiciously, in routine meetings of the Inter-
agency Air Traªc Control Board, which had been formed to standardize approach
procedures for airfields throughout the country. In meeting no. 359, held on Oc-
tober 25, 1944, the members voted to allow the installation of what the AAF re-
ferred to as the “instrument low approach” system without restriction.10 (The
Army Air Forces had decided there was no such thing as an “instrument landing
system,” and this was simply the army name for CAA’s fixed ILS installations.)
Before that action, each installation had to be individually reviewed and approved,
slowing installation work. Production problems that had slowed the CAA-AAF
program to install fixed ILS stations during the war had been overcome, and the
army had finally begun providing CAA with complete sets of ground equipment,
including the AAF’s straight-line glide path. Because many of the installations
were being made at commercial airfields that had been taken over by the AAF at
the onset of the war, CAA expected both fields and equipment to revert to com-
mercial use very shortly after the end of the war. That would allow almost imme-
diate availability of a standardized landing aid to its airline customers. The end
of CAA’s decade-long quest to improve airline regularity appeared, finally, to be
in sight.
A few months later, in May 1945, CAA’s bubble burst. The Interagency Air
Traªc Control Board, which was dominated by the army and navy, reversed itself
and proposed immediate cessation of all installations and commissionings of the
instrument low approach systems, “until a standardization policy applicable to all
systems could be established.” On the record the reason given for this change in
policy was to resolve some conflicts over frequency allocations and flight pattern
interferences between adjacent stations. The internal CAA memo reports, how-
ever, that the real reason for the sudden reversal of the earlier decision was
“that an important segment of military opinion favors discontinuance of the In-
strument Approach Program by the Army and proceeding with installations of
Ground Controlled Approach Systems of the radar type.”11 CAA’s representative
arrived at this conclusion after an informal conversation with Lt. Col. Clarence B.
Sproul, the senior Army Air Force member. CAA’s representatives strongly op-
posed the recommendation, of course. They pointed out that CAA’s system was
standardized, according to previous agreements established years earlier. It was
congressionally approved and funded, which, they insisted, made carrying out the
program mandatory. CAA was able to block final approval of the new policy, but
its representatives recognized their organization suddenly faced a real problem.
The ground-controlled approach system was not unknown to CAA, although
it was certainly unfamiliar to many of its personnel. Thomas Bourne, the direc-
tor of Federal Airways, had accidentally been invited to the oªcial tests that the
MIT Radiation Lab had held at National Airport in February 1943. He had been
required to sign papers barring him from discussing it, even within CAA, but by
1944, GCA was an open secret within the organization. Except for that early test
series, however, no one else in the organization had seen it or operated it, let alone
learned its details of operation and maintenance procedures. Regardless of the
system’s performance, CAA was not prepared to operate it in February 1945.
During the controversy, accusations flew regarding who knew what, and when,
with ground controlled approach partisans attempting to depict CAA as incom-
petently slow even to try the system. Thomas Bourne was attacked because he had
known about GCA since early 1943 but had not done anything about it, despite
the wraps of secrecy around the system through late 1944.12 Bourne did not help
matters any by claiming that he was unaware of GCA before February 1945, which
was actually the date CAA was first allowed to send sta¤ to train on and evaluate
a GCA installation. Aviation News also reported Maj. Gen. Harold McClelland’s
claim that he had o¤ered Bourne a GCA set during the fall of 1944 but had been
refused, to further paint Bourne into the “old guard” anti-GCA camp. The record,
however, shows that McClelland had o¤ered to let CAA evaluate a GCA set in No-
vember and that Bourne had accepted it the following month. Those tests were
scheduled for February at Bryan; that same February Bourne also formally re-
quested that the Army Air Forces loan a GCA set to CAA for more extensive tests
and to serve as the basis for drafting specifications for a civil version.13 Bourne
also met with the president of Gilfillan Brothers, the army’s ground controlled
approach manufacturer, and Homer Tasker, Gilfillan’s GCA project engineer,
in February to arrange technical training and maintenance assistance. The re-
quested unit arrived in May at the Indianapolis Experiment Station, still nomi-
nally an AAF-run facility. Kroger, the leading GCA partisan among the aviation
writers of the day, later summarized this exchange in an “objective” history of the
controversy, which consistently portrayed McClelland’s claims as the true version
of events.14
The record seems to show that CAA moved as quickly as one can expect of a
large bureaucracy, especially given its almost complete isolation from the vast ar-
144 blind landings
ray of new aviation technologies developed during the war. As agency historian
John Wilson has pointed out, CAA had started the war ten years ahead of the AAF
in navigation technologies and ended it ten years behind. That was hardly the
agency’s fault. All of the new technologies were secret and, once unveiled, took
time to learn. They also required money, and CAA’s budget for radio and radar
research was a mere $87,000 for FY1946, and $180,000 for FY1947, which
Ben Stern, the assistant administrator for public a¤airs, told the editor of the
Chicago Times was one-fifth what the agency had requested.15 The amount was
not enough to buy a set of any type of radar, and CAA was therefore entirely de-
pendent on handouts from the Pentagon for its investigations. It obtained its first
radar gear from the navy, which sent ten truckloads of various sorts of equipment
in late 1944 to the Indianapolis Experiment Station. That delivery is what had
prompted Bourne to try to obtain a GCA set, which the navy had not included in
its bequest.
Glen Gilbert, chief of CAA’s Air Traªc Control Division, conducted the
agency’s first evaluation of GCA during February 1945 at the Army Air Force’s
Instrument Instructor School at Bryan, Texas. Gilbert had been one of the first
group of controllers hired by CAA to run its first experimental “en route” air
traªc control center, which went into operation in 1936.16 His analysis was cer-
tainly biased by his job assignment, and he reported primarily on GCA’s utility
as a traªc control aid.
Gilbert described discussions he had held with various sta¤ members at the
Instrument Instructor School. According to his report, they believed that because
the final controller could handle only one aircraft at a time, the maximum land-
ing rate for GCA was one plane every two minutes, based on the two-mile range
of the precision radars. Additional precision scopes and an extra controller could
raise this capacity to one per minute, which “was just about the capacity of a sin-
gle runway under any weather conditions.” By contrast, the instructors thought
that SCS-51 could only handle one aircraft every three minutes, due to beam
deflections caused by other aircraft. Although this appeared to give advantage to
GCA, both CAA and AAF found that the two-mile approach the AN/MPN-1
o¤ered was too short for the largest aircraft and increased the precision radars’
range to six miles, and then to ten miles, reversing the capacity advantage.17 In
actuality, the capacity issue was never subject to public controversy, and through-
out 1945 and 1946 both CAA and AAF sought to find ways to increase the safe
capacity of both systems.
More important was Gilbert’s evaluation of the ease of use of the two systems:
“Based on discussions with oªcers who have had considerable experience in

both systems, and considering my personal reaction in making consecutive ap-
proaches under the hood (SCS-51 in an AT-6 and GCA in a B-24), it is the writer’s
personal opinion that making instrument approaches under GCA will be easier
for the average pilot than when using SCS-51.”18
Gilbert was seconded by A. H. Hadfield, the assistant chief of the Airways En-
gineering Division.19 This had, of course, been one of the system’s major selling
points for the generals and admirals who had witnessed the oªcial Radiation Lab
demonstrations in February 1943. Alvarez had intended it to be easy for pilots,
and Gilbert’s statement reiterates that original goal. It was also a prophetic state-
ment, and the first warning to CAA’s old guard that they might face a public re-
volt.
By the end of the year, the two men found their opinions being echoed by
members of the aeronautical press. William Kroger reported in December 1945
that “private flyers and segments of the industry [were] arranged against CAA,
criticizing its instrument approach system as too complicated.” He blamed CAA’s
conservative faction for continued promotion of ILS, and in subsequent articles
various other Aviation News authors named Stanton and Bourne as the leaders of
that faction. The Aircraft Owners and Pilots Association was the segment of the
industry in question, as the Air Transport Association supported CAA’s ILS. The
two exceptions were Pan Am (not represented by the Air Transport Association,
since it served overseas routes exclusively), and Jack Frye’s TWA. Pan Am, along
with a number of transatlantic airlines, rented a GCA set from the army and
operated it at Gander, Newfoundland, beginning in January 1947. TWA did the
same at Reading, Pennsylvania, as an emergency field for its major hub at Phil-
adelphia. For that, the two received great accolades from Aviation News, Popular
Science, and even the New York Times.20 Adm. Emory S. Land, president of the Air
Transport Association, pushed CAA to make experimental GCA installations, yet
Land and his technical vice-president, retired general Milton Arnold, emphasized
GCA use as traªc control aid, not primarily as a landing aid. Hence, “segments
of the industry” is at best a misleading phrase, and many aviation writers seem
to have followed Kroger in misconstruing support for GCA as a control aid as
support for it as a landing aid.
Although the aeronautical press castigated CAA for its “anti-radar” faction, it
was clearly an unfair characterization. Administrator Wright wrote to James John-
son, president of the Springfield (Missouri) Flying Service, that the CAA program
included “the adoption of ground radar search installations for use in connection
146 blind landings
with airport traªc control.”21 He did not, however, mention its use as a landing
aid. That distinction was never made clear within the aeronautical press, and a
large portion of the acrimony over radar was probably derived from sheer confu-
sion. CAA’s policy was not anti-radar, but its oªcials by and large were opposed
to its use as a landing aid.
Leaving aside the innate bias Bourne, Stanton, and others had toward ILS, the
use of radar as a landing aid appeared to be a very bad deal to CAA. One problem
it perceived was liability. If a pilot using ILS made a mistake and crashed, it was
legally “pilot error” and therefore an airline responsibility. If the ILS were broken,
aircraft could be sent elsewhere. If a GCA final controller made a mistake, then
CAA clearly was responsible. Since pilots had no other source of information on
their approach position in a GCA-only approach, they could not evaluate the con-
trollers’ orders, and hence CAA could not pin the blame on them.
A larger problem for CAA was the sheer number of operators that a full-scale
GCA deployment would have entailed. The wartime sets, as described in Chap-
ter 6, required five operators per shift, with at least one additional maintenance
person. CAA worked with Gilfillan Brothers during 1946 to reduce that to two
operators per shift, but for twenty-four-hour operation that still required eight op-
erators and two maintenance crew per installation. The agency planned to install
ILS at 180 airports throughout the country by 1950, which would mean an addi-
tional 1,800 new employees if GCA replaced ILS in its plans. That was a 10 per-
cent increase over CAA’s 1946 payroll. CAA was under no illusion that Congress
would happily provide it with so many new payroll lines; indeed, the agency found
Congress unwilling to provide suªcient payroll funds to bring it to its authorized
manpower. Worse, Congress e¤ectively cut CAA’s payroll in 1946, prompting the
closure of fifty-five airways communications stations.22 Given its tight budget,
it is unsurprising that CAA’s policy was to use radar sparingly. Since ILS was
automated, and required only one maintenance person whose duties could be
split between the ILS and other airport electronics gear, it made far more sense
to oªcials.
Then there was the issue of maintenance costs. Although CAA had received
and tested a GCA set during 1945, it did not undertake anything like the rigorous
testing necessary to really evaluate the system’s performance or determine its true
operating costs in civil use until 1946. Military cost estimates were essentially
useless because military operations did not suªciently resemble civil ones. With
the exception of Military Air Transport Service operations, military aircraft took
o¤ and returned nearly simultaneously. That meant the system had to handle
high volume, but only for short periods of time. The system could then be put in
standby for several hours before the next wave of aircraft operations. Military op-
erations did not require continuous availability, and maintenance could be sched-
uled between operating periods. Commercial use, by contrast, required continu-
ous operation of the system for at least twelve and even, at the largest airfields,
twenty-four hours a day. Maintenance time came only at the expense of service to
aircraft, which meant commercial use required a vastly more reliable piece of
equipment than the military needed.
CAA found during its investigations in 1946 that the Gilfillan-built AN/MPN-1
it had been loaned required 900 hours of maintenance for 1,600 hours of oper-
ation, although some of that maintenance could be conducted while the system
was in operation. Gilfillan Brothers hotly contested that figure, but based on the
Signal Corps’ own maintenance records for the twenty-eight GCA sets it operated
during 1945 and 1946, it may have been accurate. There was extreme variability
in the number of maintenance hours needed to keep the equipment running:
some units spent more time being maintained than being operated, while others
appear to have been very reliable.23 One likely explanation is poor quality control
at the factory. With 600 vacuum tubes stu¤ed into a trailer, minor variations
in placement, quality of soldered connections, and quality of components could
cause large di¤erences in reliability between units. CAA might have gotten a
lemon. Another explanation is quality of maintenance personnel. Inexperienced
or poorly trained maintenance people require much more time to accomplish a
given task than someone who knows the equipment well, and CAA’s own lack of
radar acumen no doubt drove up its tally of maintenance hours. Finally, the ba-
sic design of Gilfillan’s equipment seems to have a¤ected its reliability. During
the war, Bendix engineers sent to Gilfillan Brothers by the navy to learn how to
build GCA had substantial disagreements over design with Gilfillan’s chief engi-
neer; after the war, CAA found that Bendix-made GCA sets, known oªcially as
AN/MPN-1(B), were easier to maintain and much more reliable.
Regardless of cause, CAA’s oªcials were convinced that the maintenance re-
quirements of the wartime GCA sets were too great for civil use. They worked
with Gilfillan to try to reduce those costs. The agency reported its e¤orts in CAA
Journal throughout 1946 and 1947. Re-engineering GCA for commercial service
meant, of course, several years of delay before any feasible implementation of ei-
ther full GCA or even just the search radar. Engineering work required time and
money, neither of which were granted to the besieged CAA.
Cost, however, depends on one’s point of view. As noted above, from the navy’s
perspective GCA was less expensive. The only highly trained people required by
the system were the ground controllers, and even a large-scale GCA installation
148 blind landings
would require fewer operators than there were pilots who, in an ILS world, would
have to be extensively trained, and constantly retrained, in order to maintain their
skills. “Under such circumstances,” Adm. Luis de Florez had stated, “GCA is in-
comparably cheaper . . . The training you give the ground operators of GCA cov-
ers literally the hundreds of pilots a day he [sic] can talk in safely.”24
The admiral was most certainly correct, from a certain point of view. Elimi-
nating a great mass of pilot training to train a much smaller group of controllers
meant significant savings to the navy. For the CAA, however, which did not have
to pay pilot training, the addition of controllers and increased training was a ma-
jor budget blow. ILS minimized cost to the agency, at the price of raising the air-
lines’ costs. The airlines did not object, since they expected additional revenue
from whatever landing aid was selected, and most were not thrilled with various
cost-sharing proposals that were bandied about in Congress to fund GCA. Hence
the di¤erent economics of commercial and military aviation influenced techno-
logical choice in a way that was never made clear in the media. GCA probably was
cheaper from a global economic standpoint, but such a consideration was irrele-
vant to the CAA, the Budget Bureau, and to the Congress.
Therefore, CAA did indeed object to radar landing but not solely because of
the conservatism of Stanton’s old guard. The agency believed it had legitimate
concerns based on its funding history balanced against the costs it expected radar
landing to impose over and above a much more limited traªc control radar de-
ployment. The media’s own bias, however, prevented it from explaining the real
costs of radar landing. In fact, the media often misrepresented the issue. AOPA
Pilot, for example, told its readers that one GCA set could serve Washington Na-
tional, Bolling Field, and Anacostia Naval Air Station.25 That, of course, was an
absurd claim, but one the average person could not evaluate. Nor could the me-
dia advocates accept that re-engineering the wartime GCA was necessary to make
it commercially useful, and they insisted that CAA adopt the trailer-mounted sets
immediately.
Pressure thus built on Administrator Wright to do something about GCA dur-
ing 1946, and he decided to try to obtain some additional GCA units to install at
commercial fields in order to determine the true costs of operating them as com-
mercial installations. He was probably persuaded to try this by Milton Arnold, an
ex-brigadier general who had commanded a bomber group in Europe during the
war and had left the army to become head of the Air Transport Association’s en-
gineering section. Arnold had broached the issue of the Army Air Forces loaning
three sets of AN/MPN-1 equipment to CAA as well, and the AAF’s internal re-
sponse is worth repeating verbatim: “The primary purpose is to demonstrate con-
clusively to CAA the advantages of search radar in simplifying airport traªc con-
trol and at the same time improve airline schedules and operations during the
coming winter. Early adoption of airport radar by the CAA will be advantageous
to the AAF as well as the airlines, consequently this Headquarters [Air Materiel
Command] concurs in the desirability of aiding in the project even to the extent
of loaning developmental equipment.”26
Air Materiel Command’s letter suggests that neither the Air Transport Asso-
ciation nor the Army Air Forces was dedicated to forcing CAA to adopt GCA’s
landing function. It clearly is aimed at GCA’s traªc control function, in the same
way that Glen Gilbert of CAA’s Air Traªc Control Division had emphasized the
search radar’s utility in his 1945 report. That traªc control was the important is-
sue to the Air Transport Association’s leaders is unsurprising, despite the media’s
devotion to radar landing. Its oªcials were well aware that if CAA suddenly
dropped ILS for radar landing, widespread deployment would be several years
away due to the vagaries of federal budgeting and the need to re-engineer the
equipment. ILS was already installed at thirty-nine major commercial fields, and
CAA had funding for a large number of smaller fields as well. The only reason it
was not already operational was lack of receivers for the aircraft, and those too
were expected in a few months. The ATA and its member airlines had been
burned before by waiting for something “better” to be perfected when a “good
enough” system already existed, and it was not about to make that mistake again.
From the airlines’ point of view, radar landing was not worth waiting for.
However, the letter also implies that Air Materiel Command agreed with the
aeronautical press that some elements within CAA were not giving GCA’s search
function its due. Both Wright and Glen Gilbert, however, were already enthusi-
astic about GCA’s possibilities as a traªc control device. Agency poverty pre-
vented them from doing much about it, although Wright did seek to have CAA’s
radar research budget increased. The Budget Bureau also denied this request. So
borrowing was all he really could do.
Wright happily accepted the Army Air Forces o¤er to loan three GCA sets for
installation at Chicago, La Guardia, and Washington National after getting the Air
Transport Association to agree to pay the installation cost (a not-inexpensive
$20,000 per site) and arranging for Gilfillan Brothers to install three experi-
mental remoting kits to place the radar screens in the airport control towers in-
stead of in the trailer parked on the field. Those kits were also part of the AAF’s
loan. Where the money for their installation came from is not clear in the records,
but CAA had to pay the twenty-one operators necessary to run the three stations,
borrowing people from other parts of its airways operation service to crew the
150 blind landings
units.27 The GCA sets did not come back from Gilfillan until late in the year and
were not ready for service before January 1947.
Despite Wright’s e¤orts to overcome whatever bias existed within Stanton’s
old guard, it was never enough to satisfy the GCA enthusiasm Kroger and other
aviation writers promoted. He could not have satisfied it because he worked
within constraints imposed on him by CAA’s position as a poorly funded subunit
of the Commerce Department. His primary constituency, the commercial air-
lines, wanted a system of landing aids immediately, not years in the future, and
ILS was all that was available. Radar landing, even if he had been able to recon-
cile its high cost with his comparatively small budget through some means of cre-
ative cost-sharing with the airlines, was years away from service if for no other
reason than the time needed to re-engineer the equipment to reduce its mainte-
nance and operator costs. During the war emergency, when cost was no object,
GCA had taken a year and a half to get into production, and the aviation writers’
presumption that a civilianized, rigorously tested radar system could be had in
mid-1946 was absurd. Ultimately, the GCA enthusiasts could not see that “com-
bat proven” was not a good enough appellation for civil use. Civil aviation had
di¤erent needs.
Yet CAA’s concerns about radar landing were not entirely legitimate, either.
The budget problem CAA’s oªcials perceived was not really CAA’s problem at
all. It belonged to Congress, which had the authority to fund whatever it chose to
approve. CAA’s budgetary concerns, although perhaps honorable, were thus
somewhat misplaced. As one member of the House Interstate Commerce Com-
mittee soon pointed out, Congress was ultimately responsible to the public for
the air transportation system, not the unelected CAA. Part of the agency’s annual
responsibilities was to provide a report on what it needed to safely maintain the
nation’s airways so that Congress could then make informed decisions in fulfill-
ment of its duty. In this view, CAA had prevented Congress from exercising its
responsibility to the flying public by proposing only ILS and trying to ignore radar
landing. The agency’s devotion to ILS thus seemed to many members of Con-
gress to be protection of a pet program, not pursuit of the best technologies. By
late 1946, the media had convinced Congress of GCA’s superiority, and Congress
joined the anti-CAA alliance.
congress enters the fray
The midterm election of 1946 returned the first Republican majority to Con-
gress since Herbert Hoover left oªce in 1932 and thus provided the disenfran-
chised Republicans their first opportunity to roll back the “creeping socialism” of
the New Deal and attack Truman and the leftover New Dealers in his adminis-
tration. The Republican leadership probably perceived that Harry Truman was
particularly vulnerable to an investigation into his administration’s handling of
aviation matters, since as a U.S. senator he had been a leading advocate of the
Civil Aeronautics Act that had founded CAA. He had also chaired a long-running
investigation into defense procurement during World War II, which focused on
the aircraft industry.28
The Republicans had two pretexts for mounting the investigation. The major
newspapers and the aeronautical press had noticed a “rash” of fatal air crashes
during 1946, which the airlines claimed had resulted in a sudden drop in air
travel. A number of articles claimed radar or GCA could have prevented many of
these. The acrimonious debate over the two landing systems, which of course in-
volved the much-criticized lack of radar along the airways, provided the other ex-
cuse for a full-blown congressional investigation.
The rash of accidents, in reality, was a statistical artifact. The commercial air-
line accident rate for 1946 was actually only about half that of 1945, but because
the amount of traªc had nearly tripled, the number of accidents went up, and ac-
cordingly, so did media coverage (see Table 7.1). The crashes, according to United
and Eastern Airlines spokesmen, had resulted in a “marked decrease” in pas-
senger traªc from the previous year. Either the airline spokesmen were unaware
of the statistics that their superiors gave CAA or they chose to project an image
of reduced safety in order to “help” CAA get more money from Congress. Re-
gardless, the airlines’ claims inspired great leaps of congressional rhetoric. Rep.
L. Mendel Rivers (D-SC) declared, “The American people are horrified and scared
to death.”29
The perception that CAA was not doing its job was further reinforced by
claims that radar (in the New York Times) or GCA (in Aviation News) could have
prevented some of the crashes.30 The airlines had contributed to the perception
table 7. 1
Fatalities per 100 million passenger miles
Year NATS ATC US Airlines
1944 18.1 5.5 2.1

1945 19.7 6.0 2.1
1946 10.8 8.2 1.2
Source: CAA Statistical Handbook, 1950.
Notes: NATS is the Naval Air Transport Service, while ATC was the Air Trans-
port Command.
152 blind landings
of danger by trumpeting the merits of wartime radio and radar devices as safety
advances, without making clear that these were years from being ready for actual
commercial service. The public expected the number of accidents to drop as a re-
sult of these advances, but instead the number went up. Nevertheless, the avia-
tion system was safer. It simply did not appear that way to a public misled by a
number of sensational crash stories and equally ostentatious tales of radar’s “all-
seeing eye.” Congress chose to act on the appearance rather than the reality.
The second pretext for Congress’s investigation into CAA’s management of
the aviation system was the ILS/GCA controversy, which had been rekindled in
late November 1946 when the New York Herald Tribune and Chicago Tribune, as
well as Aviation News, published parts of some allegedly “secret Army Air Forces’
tests.”31 These reports cited Lt. Col. Clarence B. Sproul’s claims for GCA’s supe-
riority over ILS and provided much scientific-appearing data to back him. The ap-
pearance of these pro-GCA articles marked the beginning of a full-blown public
controversy. Combined with the rash of air crashes, they virtually ensured con-
gressional attention to the landing aids issue.
The House Committee on Interstate and Foreign Commerce, which oversaw
the Commerce Department, convened hearings in January 1947 to investigate the
air safety crisis and CAA’s management of the federal airways. The full commit-
tee was chaired by Rep. Charles Wolverton (R-NJ), while Carl Hinshaw (R-CA)
chaired the subcommittee established to run the air safety hearings. In general,
the transcript provides little evidence of political grandstanding, and the members
seem to have been genuinely concerned with the issues in question, although the
majority clearly did not understand the technologies involved. Hinshaw, whose
Pasadena-based district contained substantial portions of the aviation industry
and had long been interested in aviation matters, was one exception. Also, M. Har-
ris Ellsworth (R-OR) clearly recognized the political nature of the controversy.
The air safety subcommittee heard a veritable parade of witnesses from the
aviation industry over the next two months. Generals Ira Eaker and Harold Mc-
Clellan were among the earliest witnesses, while the navy’s primary witness was
Rear Adm. J. W. Reeves, head of the Naval Air Transport Service (NATS). Wright,
and his superior in the Commerce Department, William Burden, were CAA’s ma-
jor representatives, while Milton Arnold represented Air Transport Association
and David Behncke spoke for the commercial pilot’s union, ALPA. Last to speak
to the committee was AOPA’s Hartranft, who was asked no questions.
While the army’s General McClelland carefully avoided taking a position on
the great controversy despite several attempts to pin him down, Admiral Reeves
was happy to discuss the navy’s decision to use GCA exclusively. It had, he re-
ported, enabled NATS to reduce its weather delays to an average of 2 percent of

all flights. He also told that committee that in January 1947, in a period during
which 45.7 percent of total commercial scheduled operations in the San Francisco
area were canceled, only 1 percent of NATS operations at GCA-equipped Mo¤ett
Field were canceled.32 Reeves also referred the committee to Colonel Sproul’s re-
port, which was entered into the record as evidence. Publication of Sproul’s test
reports in the hearing transcript marked the first public appearance of the full re-
port and no doubt helped convince the committee of CAA’s backwardness on the
issue.
Clarence Sproul, now an ex-lieutenant colonel and the same man who had in-
formed CAA’s Interagency Air Traªc Control Board (IATCB) members of the
AAF’s sudden change of policy toward GCA in May 1945, had characterized his
tests for the media as “proving” the superiority of GCA’s radar landing function
over ILS. Pilots using GCA, Sproul reported, “achieved 350 per cent better run-
way alignment than those following the glide path-runway localizer beams of the
I.L.S.” The most detailed publication of Sproul’s findings was in Aviation News.
William Kroger reported that airline captains who participated had an average of
6,560 hours flying time, 40 percent had previously flown GCA, while 20 percent
had flown ILS. Ninety-one percent were able to touch down with GCA, while only
54 percent succeeded with ILS. A group of fighter pilots, Kroger reported, flying
the C-54 test aircraft for the first time, and with an average of only 805 flying
hours, “were far more accurate in landings on GCA than with ILS.”33
Before trying to glean anything meaningful out of Sproul’s data, some dis-
cussion of the test methodology is necessary. Unsurprisingly, neither of the writ-
ers bothered to ask about how the tests were conducted, since “scientific tests”
were supposed to be objective and report the truth. Sproul’s tests, however, con-
tain a serious methodological flaw that renders the details suspect. The AN/MPN-
1 set’s precision radars were used to evaluate the performance of both themselves
and the ILS.34 There was therefore no independent verification of either system
and thus no adequate control despite the selection of “control groups” of pilots.
If GCA happened to be in error, for whatever reason, the error could not be de-
tected.
In response to publication of Sproul’s tests, Wright informed the aviation ed-
itor of the Boston Traveler that CAA had yet to find “a single case where any pilot
using the instrument landing system missed the runway at its end by as much as
326 feet, which Mr. Sproul reports as the average error for pilots experienced in
ILS.”35
CAA’s director of Air Navigation Facilities Service contended that an error that
154 blind landings
great at touchdown represented a 3.3 degree deviation from the landing course,
which would result in a full-scale deflection of the cross-pointer instrument. No
certificated pilot, he claimed, could be so consistently poor.36 Yet airline pilots did
even worse than the control group average, according to Sproul’s data, achieving
a dismal average error of 726 feet.
Despite the bias of CAA’s people toward their own gear, they were certainly
correct in their belief that no pilot who had survived 6,000 or more flight hours
could be so bad. The poor showing of ILS was probably due to a misalignment
between the two systems, which would cause GCA’s performance to appear far
better than SCS-51’s. Sproul certainly had no means to ensure such precise align-
ment, and in actual operation, very precise alignment is unnecessary if landing
aircraft is to be done visually. Only automatic systems require the kind of preci-
sion alignment necessary to eliminate the possibility of beam divergence. Hence
Sproul’s detailed data is at best unreliable, and it certainly misled both the media
and congressional investigators as to the relative performance of the systems.
Nevertheless, some of Sproul’s data remains useful. One piece of data in each
series, the “percentage of Approaches from which a Landing Could Have Been
Made” is based in part on input from the check pilots, who were able to see the
airfield. That eliminates some, but not all, of the pro-GCA bias in the numbers.
Since only a few diehards within CAA still considered blind landings feasible—
Sproul’s report rejects it explicitly—pilots would make visual landings after break-
ing through an overcast. Once CAA placed ILS in commercial service, it imposed
a 300-foot ceiling, reducing it to 200 feet for each airline as its pilots gained ex-
perience. Hence, the 200-foot data permits some useful conclusions (Table 7.2).
table 7.2
Percent approaches from which a landing could have been made
Test Group GCA SCS-51
Control Group #1 100% 83%

Control Group #2 100 92
Air Transport Command 97 88
Air Material Command 94 61
HQ, Army Air Forces 93 64
Bolling Field 91 75
Fifteenth Air Force 87 79
CAA 100 91
Air Line Captains 100 72
Training Command 100 100
All Pilots 95 80
Source: House Committee on Interstate and Foreign Commerce, Safety in Air
Navigation, 80th Cong., 1st sess., Jan. 1947, 531–37.
All pilot groups achieved better results on GCA than ILS, with the exception
of the AAF’s Training Command instructors, who achieved 100 percent on both
systems. The worst GCA performance was 87 percent, by the fighter pilots of Fif-
teenth Air Force, who were flying a C-54 transport, an unfamiliar aircraft of much
greater weight than the fighters to which they were accustomed. That certainly
made flying both systems more diªcult, as they would not have been able to ac-
curately judge the aircraft’s handling. All other pilot groups achieved above 90
percent. GCA direction was thus very consistent in its performance regardless of
the experience of the pilots flying it.
The SCS-51 results, in contrast, vary from a high of 100 percent for the Train-
ing Command pilots to a low of 61 percent for the Air Materiel Command pilots.
The worst-scoring AAF pilots were the least experienced in instrument flying. Es-
pecially telling is the 64 percent showing for the AAF headquarters pilots. Al-
though they had a relatively high flight hours average, they were in nonflying bil-
lets and had not flown much in the ninety days preceding the tests. The poor
showing indicates that their instrument skills had atrophied. Their much better
performance under GCA in turn suggests that Glen Gilbert’s 1945 assessment of
GCA’s greater ease of use was correct. The SCS-51 required more training to use
successfully, and it required that pilots use it more frequently in order to main-
tain their skill level.
GCA was precisely what AOPA and its general manager, J. B. Hartranft Jr,
wanted in a landing aid. Hartranft, a former AAF oªcer, had been the other AAF
member of the Interdepartmental Air Traªc Control Board when Colonel Sproul
had pushed for the cessation of ILS installations in early 1945. He declaimed
against ILS and CAA’s old guard from the pages of AOPA Pilot, emphasizing that
GCA would cost fliers nothing if adopted. With all equipment on the ground, pro-
vided by CAA, private fliers would have nothing to buy and no training to pay
for.37 To the House Committee, he emphasized that the receivers were too large
and heavy for private aircraft to carry. They were also too expensive at an estimated
$750 apiece, in an era when an airplane cost about $1,500. They were not even
available yet, and would not be for private fliers for several years. ILS was useless
to the members of his organization.
Hartranft then argued that ILS would not serve the majority of airways users.
According to his numbers there were 80,000 private aircraft in the United States,
and only 800 aircraft operated by scheduled air carriers. The airlines thus did not
constitute the vast majority of the flying public. Since ILS could only be used by
the 800 commercial aircraft, taxpayer money should not be spent on it, especially
since GCA was the “proven superior system.”38 He therefore demanded that Con-
156 blind landings
gress act to stop CAA’s ILS installation program, force it to dismantle the instal-
lations already in existence, and buy GCA.
Certainly GCA was the better choice for private fliers, for both economic and
safety reasons. Implicit to Hartranft’s argument, however, was an assumption
about who, exactly, counted as an airways user. He had pointed out that “the
amount of cross-country navigational flying accomplished by personal aircraft
owners and private pilots is much in excess of that accomplished by commercial
carriers.”39 Pilots, he assumed, were the only airways user, and therefore private
pilots were the dominant user, with military pilots second. The mere 7,000 com-
mercial pilots were a distant third. The air navigation system should thus be de-
signed to suit private pilots, and he railed against CAA’s “reverse thinking” in
favoring commercial airlines and for choosing to solve only the problems of
commercial users.
Wright refuted Hartranft’s claim that private fliers were the majority of air-
ways users, explaining that the “person-hour” was the best measure of airways
use and that “the individual passenger in a civil or military transport aircraft is
just as much a user of the [air navigation] system as the pilot of a military aircraft
or personal aircraft.”40 Therefore, of the 173,000,000 person-hours flown in
1946, 163,000,000 were flown in commercial and military aircraft, which could
be equipped for ILS (see Table 7.3). In Wright’s formulation, the majority of users
were passengers on commercial aircraft. The air navigation system should there-
fore be built primarily to serve that user group most eªciently, despite the im-
position of higher costs on the private fliers.
AOPA and CAA could not even agree on who airways users were. AOPA’s bias,
of course, is obvious. But CAA was supposed to represent all users, and its cho-
sen statistic, the person-hour, did show a clear commercial bias. A large com-
mercial aircraft, with many more seats, longer range, and higher speed, could
rack up “person hours” a great deal faster than any private aircraft could hope to.
table 7.3
CAA Airways use statistics
Group Millions of Person-Hours Flown in 1946
Scheduled Air Carrier 99

Non-Scheduled Air Carrier 27
Army 20
Navy 17
Personal 10
Source: House Committee on Interstate and Foreign Commerce, Safety in Air
Navigation, 80th Cong., 1st sess., January 1947, p. 145.
The person hour was not the only statistic CAA had to choose from. Admiral
Reeves, for example, had taken other CAA numbers and cleverly demonstrated
that his Naval Air Transport Service was the biggest airways user, before explain-
ing to the committee that anyone could make the numbers prove any particular
case.41 Wright could have chosen some other statistic and thereby made a di¤er-
ent argument. The case Wright chose to prove, however, was the commercial one.
More than any other act Wright could have performed, his choice of the person
hour reveals his agency’s bias.
Wright’s defense of commercial aviation certainly satisfied ILS’s most out-
spoken defender, the Air Line Pilots Association. ALPA had been founded se-
cretly in 1931 by David Behncke, who was still its president, after the failure of an
earlier attempt to unionize commercial pilots.42 Behncke was a labor activist, but
he refused to involve ALPA with other labor organizations. He wanted pilots to
be seen as skilled professionals, not as common laborers. Behncke’s protection
of the image of pilots as skilled professionals helped the union immeasurably in
its aviation safety campaign, which Behncke waged personally in ALPA’s monthly
newsletter, through friends in Congress, and occasionally in more mainstream
publications.
Behncke chose to attack GCA on the issue of safety. In the March 1946 issue
of The Air Line Pilot, he had written to his fellow pilots: “‘GCA’ requires complete
and precise coordination of five operators on the ground and the pilot in the air-
plane in order to e¤ect a safe landing. Think this over for a minute and realize
the possibility of human error involved plus the possibility of malfunctioning
equipment.”43 Behncke believed that the additional humans in what we would
call the “information loop” was bound to increase the probability of human error
and thus cause more accidents than the mostly automated ILS. Because of this,
he argued, ILS was the safer system. He moderated somewhat in his congres-
sional testimony (where he advocated proper runway lighting as the most im-
portant landing aid), and in a later article in Air Transport, the oªcial organ of the
Air Transport Association (and therefore a management publication). There, he
apparently decided to cleave more closely to the airlines’ oªcial position that GCA
should be adopted for traªc control and ILS should be the landing aid.44
In his own newsletter, however, he published the results of a poll he had con-
ducted of the union’s fifty-three regional councils. Forty-six of the councils had
voted for ILS as the primary landing aid, while only one had supported GCA. The
remaining six had abstained, because their members felt they did not have
suªcient experience to properly evaluate the systems. Behncke also noted that
seven councils had written to support the use of GCA as a traªc control aid, al-
158 blind landings
though they had not been asked their opinion on that subject.45 Behncke thus
had overwhelming support from his members for his ILS campaign.
ALPA’s public posture promoted ILS as the safer system, but Behncke also had
other motivations. Historian Nick Komons has argued that ALPA is a schizo-
phrenic organization, which uses a public rhetoric supporting air safety to mask
its real concern, the economic well-being of its members. ALPA is, after all, a la-
bor union, and unions exist primarily to advance their members’ economic in-
terests. Yet unions also have tended to try to protect their members’ skills from
being eliminated by technological change. ALPA’s defense of ILS was also a de-
fense of pilots’ skills. As Sproul’s results suggest, ILS was indeed harder to use
and required more training and practice than did GCA. It therefore maintained,
if not actually increased, the skills needed to operate a plane, while GCA trans-
ferred some of those skills to a set of ground operators. Gilfillan Brothers and the
AAF were already working on an automatic GCA, which was first demonstrated
in 1947.46 The union’s pilots could thus look forward to a complete loss of land-
ing skills within a few years if GCA were adopted. Automation of ILS, because
the automation system resided entirely within the plane and was therefore under
the pilot’s control, was a far more acceptable outcome.
AOPA and ALPA, therefore, had fundamentally di¤erent interpretations of
Luis Alvarez’s GCA. One group perceived it as a means of improving their access
to the airways system, while the other considered it a threat to their profession.
The key issue Congress faced, then, was choosing which of these constituencies
to favor. One member of the House committee recognized that this was, ulti-
mately, a political issue. Rep. Harris Ellsworth tried to make that point clear while
questioning Administrator Wright: “My point is the old question of how to spend
money to do the most good for the most people. We are spending $6,000,000
[on ILS] for the benefit of 900 airplanes in the air today, whereas there are 50,000
planes which could use this other system of approach [GCA] . . . Why not ask
more money and put in this other system which seems to be admittedly good?”47
Wright’s only defense was to repeat that CAA believed ILS was the better sys-
tem, especially if it could be supplemented by GCA at the busiest airports for
traªc control purposes. The issue of better for whom did not, apparently, enter
Wright’s thinking, while Ellsworth clearly understood that was the root of the
problem. Ellsworth rightly believed that GCA was better for the majority of pilots
in the United States, while Wright followed the bias of his organization and its
favored constituents, the commercial airlines, toward believing ILS was a better
choice. Ultimately, Ellsworth sought to persuade his colleagues that they faced a
political choice between di¤erent interest groups, not a technological choice over
competing claims of technical superiority. The two models of operation, ground-

controlled versus pilot-controlled, proved to be inherently political.
The committee chose to ignore Ellsworth, however, and focused on a di¤erent
issue in its final report. The debate in the media had convinced the members that
CAA’s plans di¤ered substantially from the Pentagon’s. The members of the
House Commerce Committee were very attentive to the needs of the AAF and
navy in their 1947 investigation. In part, the members’ bias toward national de-
fense issues derived from the almost complete unpreparedness of the AAF for
war in 1941 and an understandable reluctance to see that situation reoccur. And
although there is no mention of a Soviet threat in the 1947 transcripts, the near-
disintegration of U.S. armed forces during 1946 had not been matched by simi-
lar disarmament in the only other potential world power. The members were at
the very least convinced that civil aviation had to be built to be easily available for
military use in a future emergency, regardless of the source of that emergency,
and they focused on CAA’s well-publicized “disagreement” with the military over
landing aids.
There was, in fact, very little disagreement between the Army Air Forces and
CAA in terms of their landing aids policy. Indeed, Major General McClelland told
the committee that the controversy was “unfortunate,” and had diverted attention
from their joint integration program, while former brigadier general Milton
Arnold, of the Air Transport Association, said he thought that CAA’s GCA work
could not have been done better.48 Mr. Ellsworth ruefully admitted that “there is
a tendency on the part of some of us who are laymen on this subject to grab at
these two combinations of initials and consider that there is a battle on as between
the two of them.”49 But because the majority of witnesses the committee had
heard took sides, and the press clearly had as well, the air safety subcommittee
was already biased toward an either/or choice. Ultimately, it did consider the two
systems competitors, basing its final report on that assumption.
The final report therefore criticized CAA for its consistent refusal to accept the
decisions of the military-dominated Air Coordinating Committee (ACC), which
had called for cessation of the ILS program in 1945. The committee stated that
“any area of disagreement between civil and military aviation as to navigational
and airport approach aids should be immediately resolved in the interest of na-
tional defense.” The Air Coordinating Committee, the House committee mem-
bers believed, was being held hostage by CAA’s refusal to accede to Army Air
Forces and navy demands for GCA. It could do this because the Executive Order
establishing the Air Coordinating Committee required unanimity in decision-
making. The members demanded that CAA stop using its veto power to block
160 blind landings
committee decisions and recommended that Congress establish a new body

whose structure would not allow CAA to dominate the military. Further, it rec-
ommended that the House not fund any new civil landing aids, including the
twenty GCA units the army and CAA had negotiated a loan of, until such time as
the Army Air Forces, navy, and CAA reached agreement on a common air navi-
gation system. The House agreed, as did the Senate, and CAA found its entire air
navigation plan suddenly in ruins.50 Considerations of national defense thus
drove Congress to censure the Civil Aeronautics Administration, which the mem-
bers were convinced was interfering with the Pentagon’s plans.
Yet despite CAA’s loss of FY1948 funding, and the harsh criticism it received
from the House, Congress’s decision actually represented a substantial victory to
the beleaguered agency. It had not been stopped from installing and operating
the hundred or so ILS systems that had been funded in earlier budgets, nor had
it been directed to install GCA in any form. By late 1947, therefore, the major air-
lines had begun using ILS at thirty-nine airports throughout the nation and con-
tinued to do so on more than 1,500 installations in 1998. Despite being castigated
in every media outlet from Aviation News to Life magazine, and being subject of
a congressional censure, CAA got what it wanted: an operational ILS system to
serve its preferred customers, the commercial airlines.
Congress’s nondecision also represented a victory for the commercial airlines,
which got a “good enough” system for immediate operation. They were not forced
to wait another four or five years for radar landing, which would have happened
had Congress forced CAA to conform to the navy’s decision. Not incidentally, the
airline pilots won, preserving their skills and autonomy for at least a few more
years. Their support was probably crucial, since pilots’ refusal to use the Hegen-
berger system in 1934 had doomed that system to failure. Congress certainly lis-
tened when Behncke insisted that landing lights were the most important land-
ing aids, for the House final report elevated lights to its own number one priority.
Finally, without ALPA’s support, CAA would have stood alone. The Air Transport
Association, although supportive of ILS, ultimately wanted both systems installed
universally, an outcome CAA did not support.
The two losers in this controversy were the U.S. Navy and private aviation. The
navy did not lose much. Although its GCA-only policy did not get a congressional
mandate, that hardly mattered. Just as the navy’s unique requirements had kept
it out of the discussion in the RTCA’s landing aids e¤ort during the late 1930s,
its needs clearly made it nearly irrelevant here. No one expected naval considera-
tions to have great impact on civil aviation. Congress was more interested in the
AAF’s opinion, which, as the AOPA Pilot cartoon showed, was mixed, and there-
fore open to interpretation. In the light of widespread media support for GCA,
the members chose to believe that the army, like the navy, preferred GCA.
The real loser, then, was private aviation. Without money, regardless of its in-
tentions, CAA could not provide any kind of radar landing service to private fliers
for at least several more years, and small plane owners had to wait until the tran-
sistor made receivers small enough and light enough to use ILS. Administrator
Wright, who was personally a champion of private aviation and drove the devel-
opment of a simplified plane designed for the “average” private flier, resigned a
few months after the congressional investigation ended, perhaps frustrated that
he could not find a way to please both commercial and private aviation. Private
fliers were thus locked out of the air navigation system during poor weather, giv-
ing the lie to Wright’s own dreams of “commuting in the modern manner” via a
plane in every garage.
conclusion
GCA and ILS had politics built in to their very operation. The pilot-control
model, on which ILS operated, favored skilled, experienced pilots who flew fre-
quently, making it the obvious choice of professional pilots. The ground control
model, on the other hand, favored the occasional pilot, leading to unswerving de-
votion from advocates of “democratic” flying. The two models sprung not from
political di¤erences between the inventors but from their design environment.
Diamond and Dunmore, originators of the pilot-control model, worked closely
with professional pilots in a project driven largely by the demands of airlines for
regularity of service. That their system continued to be the choice of commercial
aviation groups is unsurprising. The ground-control model evolved from Luis Al-
varez’s experiences as a private pilot and a cyclotroneer, while he was working to
overcome a military problem. That a military system was the more “democratic”
may seem surprising, but the drafted and hastily trained armed forces of World
War II had di¤erent technical demands than the post-Vietnam professional mil-
itary. Alvarez’s system had fit the needs of a draftee military nicely, and the very
qualities that made it useful in that circumstance made it the ideal solution for
private aviation.
Just as the way the two systems worked influenced how they were perceived
in the political realm, the way they allocated costs impacted how they were seen
in Washington’s bureaucratic realm. GCA, while less expensive from the per-
spectives of the navy and private fliers, threatened very substantial additional
costs to the Civil Aeronautics Administration. Its leaders had no reason to believe
162 blind landings
the agency would be given the additional funds to acquire and sta¤ the new equip-
ment, and the agency’s budget history suggested that it would be forced to ter-
minate some of its existing functions to provide operators for GCA sets. That is,
after all, what had happened when it tried to obtain a small amount of money to
man the three test sets it had borrowed from the army. The agency quite rightly
believed that GCA, whatever its benefits to aviation, would not receive the finan-
cial support necessary to operate it successfully; as political attention began to
shift away from landing aids to traªc control over the next several years, the
agency’s essential poverty became very clear. For CAA, if not for aviation, ILS was
the correct choice.
The House investigation, finally, proved only a brief embarrassment to the
administration. Congress’s decision to zero-fund landing aids handed Harry Tru-
man plenty of ammunition to shoot back with. Congress’s nondecision also in-
spired widespread derision and disgust in the media. Truman’s own investigative
committee, chaired by John Landis, chairman of the Civil Aeronautics Board, was
thus able to turn the tables on his opponents rather easily. That group, picking
up on the Army Air Forces’ program to integrate the two systems, answered Con-
gress’s censure over the next few months.
chapter eight
Transformations
Congress’s smacking the Civil Aeronautics Administration (CAA) with the

budget axe after the air safety hearings was not quite the final act in the bitter de-
bate between advocates of instrument landing system (ILS) and enthusiasts for
ground control approach (GCA). After a new series of airline accidents in June
1947, the administration turned the tables on Congress. Truman convened a Pres-
idential Special Board of Inquiry on Air Safety, chaired by John Landis, chairman
of the Civil Aeronautics Board. It published an “interim” report in July 1947 that
claimed 35 percent of all fatal accidents in 1946 occurred during approach and
landing. Congress’ deletion of landing aids from the 1948 budget, the group
wrote, guaranteed that “the tragic pattern of the past will inevitably repeat itself.”1
Unless funding for CAA’s landing aids program was restored, the report argued,
several more years of unnecessary accidents would occur.
Landis’s report also attacked Congress for its continued belief that ILS and
GCA were mutually exclusive competitors, despite the testimony of respected
Army Air Force experts that they were, in fact, complementary. The report fur-
ther undermined Congress’ position by explaining work being done jointly by the
AAF, CAA, and the navy at a former navy airfield in Arcata, California, to improve
and integrate ILS, GCA, and two other landing aids into a coordinated system to
resolve, finally, the blind landing problem. This was, it pointed out “in full accord
with the program recommended by the Radio Technical Commission for Aero-
nautics.”2 The president’s board, in short, implied that the great landing aids de-
bate had entirely missed the point. There was no conflict between AAF and CAA
policy. Both sought integration of existing landing aids, not a choice of one over
164 blind landings
the other, and were doing so through a known, respected, and, above all, politi-
cally neutral organization.
Recognition of the need for integration of ILS and GCA derived from serious
operational problems that the Army Air Forces, and to a lesser extent CAA, had
encountered during World War II. Although both systems were successful as
landing aids, that success was circumscribed. They did not solve the blind land-
ing problem. Worse, both proved to have limited capacity. They could be trusted
to land aircraft safely as long as the landing rate was kept relatively low, but nei-
ther could safely handle high-density traªc.
These two operational problems, then, drove the two agencies to reconsider
fundamentally the entire issue of landing aids. They faced two political problems,
however, that made it diªcult for them to promote an integrated, common sys-
tem. The partisan conflict between Republican Congress and Democratic ad-
ministration and the equally fractious conflict between the ILS advocates and the
GCA enthusiasts both had to be overcome to get funding restored. To resolve
these problems, CAA, AAF, and the navy decided via the Air Coordinating Com-
mittee to turn the issue over to a revived Radio Technical Commission on Aero-
nautics (RTCA), which produced a template for not only a landing aids solution
but for a comprehensive, nationwide system of air traªc control. Solving the
blind landing problem, ultimately, led to the recognition of a new national-scale
problem: airport traªc control.
complementarity: operational imperatives

for integration
During World War II, Army Air Forces leaders had believed that solving the
blind landing problem, or even the lesser diªculty of the instrument low ap-
proach, would lower their high accident rate and reduce the number of days dur-
ing which air power was grounded. The Royal Air Force’s Bomber Command did
not fly missions if their meteorologists predicted the ceiling over England to be
less than 1,000 feet during the return flight, and the Eighth Bomber Command
did not fly if the return ceiling was expected to be less than 500 feet.3 This meant
that during the war, about half of the missions that had to be scrubbed due to
weather were canceled not by weather conditions over the target but over En-
gland. That was why physicist David Langmuir had told Lee DuBridge of the MIT
Radiation Lab in 1943 that a blind bombing radar could only solve half the weather
problem.4 Some sort of landing system had also been necessary to resolve the
other half of the bombing halts.
transformations 165
But the 1930s conception of blind landing was insuªcient. Even achievement
of truly blind landings would not have resolved Eighth Air Force’s weather-related
accident problem because of midair collisions during bad-weather operations.
Eighth Bomber Command lost hundreds of bombers to accidents during the war,
many due to midair collisions.5
The Army Air Forces did not get around to trying to resolve this new problem
until early 1945, when it began putting pressure on the MIT Radiation Lab to help
it develop a traªc control radar. After brief resistance by the acting director of the
Radiation Lab, F. Wheeler Loomis, to becoming involved in a new project so close
to the Lab’s expected termination, he agreed to help the Army Air Forces’ Watson
Laboratories design a new air traªc control radar based on a heavily modified
GCA search radar. This was the CPN-18 project, and George Comstock, who had
taken over the GCA project after Luis Alvarez and Lawrence Johnston had fled
Cambridge for Los Alamos, was put in charge until the Lab’s closure. The AAF
later selected General Electric as the contractor.6
The idea that radar could be used to control aircraft, of course, was not new.
The principle was used throughout the war to direct fighters to their targets in a
procedure the United States learned from the British, ground controlled inter-
cept. It was also the basis of Alvarez’s GCA system, whose search radar was
modified into the new traªc control set. The need for a purpose-designed traªc
control radar, in turn, derived from a number of characteristics GCA’s search
radar had that substantially reduced its e¤ectiveness for traªc control.
Alvarez’s search radar su¤ered from several drawbacks. The first was its beam
pattern, which topped out at 4,000 feet.7 The radar could not detect aircraft at
higher altitudes reliably, and a traªc control radar certainly needed to detect air-
craft at normal operating altitudes of more than 10,000 feet. The war had also
seen the development of aircraft capable of operating at up to 30,000 feet, and
those aircraft were expected to spawn commercial variations immediately after
the war. A traªc control radar had to be able to detect all the traªc in an airport’s
area if it were to help prevent collisions, and hence it had to scan much higher al-
titudes e¤ectively.
GCA’s search radar also did not provide its operator with altitude information.
Height finding had been left out of the search radar to simplify the equipment
and because it was unnecessary to the search radar’s function of directing aircraft
into the precision radar’s beams. The accuracy of the barometric altimeter all air-
craft had was good enough to get a plane into the precision beams, as long as it
could be directed accurately in azimuth and distance. That meant, however, that
aircraft at two di¤erent altitudes, but the same bearing and range, appeared to the
166 blind landings
search radar operator as a single aircraft. Further, without altitude information,

the radar could not be used for stacking aircraft in holding patterns, which was
how aircraft were managed during peak traªc periods. GCA’s search radar could
thus be used to separate aircraft by distance but not by altitude, a serious draw-
back in managing a three-dimensional traªc structure.
The CPN-18 project sought to overcome these major issues, while Glen Gil-
bert, of CAA’s Air Traªc Control Division, reported on the presence of other prob-
lems that perhaps were more important to CAA. After the tests Gilbert ran at
Bryan, Texas, in February 1945, he arranged another set of tests at Fort Dix, New
Jersey, that May. The objectives were to evaluate the use of radar as a traªc con-
trol aid in separating aircraft, and as an aid in providing simultaneous navigation
and let-down guidance. Gilbert found that the system had several flaws that had
to be fixed to make it truly e¤ective as an air traªc control radar, including elim-
ination of ground clutter and providing aircraft identification. Identification was
particularly important to CAA, so that operators could quickly identify aircraft in
danger of collisions and issue warnings. Gilbert’s assessment otherwise mirrored
the AAF’s, with only the need for aircraft identification on the radar display
significantly di¤erent.8
CAA administrator Theodore Wright realized that at least for high-traªc air-
fields, an air traªc control radar might be financially justifiable, and he took steps
both to get suitable radars from surplus military stocks and to get a specification
written and circulated for a radar designed specifically for the job, based upon
Gilbert’s evaluation of GCA. Specification CAA-743, for a commercial air traªc
control radar, was the result. The specification was sent to various electronics com-
panies in March 1946, and Gilfillan Brothers, manufacturer of GCA, responded
in May. In his letter, the vice president of the company stated that he could not
in good conscience submit a bid on a project so similar to the Army Air Forces’
CPN-18. He instead proposed modifying some of the wartime GCA sets to serve
until the CPN-18 project reached production in a projected 2.5 years. In his pro-
posed modification the number of radar displays would be reduced to two and
placed in a tower instead of in a trailer.9 This was the same modification eventu-
ally made to the three AN/MPN-1 sets the AAF loaned to CAA for use at Chicago,
La Guardia, and National Airport.
This was the most positive response CAA received to its specification. Other
companies submitted bids ranging from $400,000 to $600,000. By compari-
son, General Electric’s bid on the CPN-18 contract had been $1.1 million. Because
its FY1947 research and development budget was much less than these figures,
CAA had no money with which to begin a radar program. To ensure that his
transformations 167
agency was fiscally positioned to deploy an air traªc control radar once one was
developed, Wright requested funds for fifty installations in his FY1948 budget
submission to the Bureau of the Budget, but that was cut to twenty-five.10 Radar
in any form was not going to reach CAA except as handouts from the wealthier
armed services until at least 1949, and the best Wright could do was to take steps
to ensure appropriate liaison between the Army Air Forces and his organization
in hopes that CAA could influence CPN-18’s development.
The CPN-18 project represented a first attempt to replace the AN/MPN-1 GCA
search radar with a purpose-built air traªc control radar, capable of handling the
high traªc density of major commercial and military fields even in bad weather.
The 1930s version of air traªc control, towers equipped with people wielding
binoculars, had been enough to deal with high traªc demands during good
weather, but faced with all-weather operations demands by both the AAF and the
commercial airlines, the older technique proved thoroughly inadequate. Yet just
any old sort of radar would not necessarily function adequately, either. Instead,
the AAF and CAA determined that a new radar, based on GCA’s search radar, was
required.
While the air traªc control radar was under development, the House of Rep-
resentatives’ air safety investigation heard from Army Air Forces witnesses on the
closely related subject of landing aids. Other witnesses, and most of the ques-
tioning, focused on arguments about which of the two landing aids was better,
but as the president’s Air Safety Board later pointed out, other testimony argued
that question was beside the point. Most of that testimony came from Maj. Gen.
Harold McClelland, still the Air Forces Communications Oªcer, and former
brigadier general Milton Arnold, head of the Air Transport Association’s engi-
neering section.
General McClelland, the oªcer who had requested the Valentine’s Day GCA
test in 1943, told the House committee that integration of all useful equipment
into a single system was the proper solution to the blind landing problem. Mc-
Clelland meant that air navigation and control aids had to be considered as parts
of a system in order to resolve the operational problems the AAF had encoun-
tered during the war. Each individual aid had to be considered in light of its con-
tribution to the whole. Radar was most important for its traªc control uses, while
ILS was useful for its easy adaptability to automatic approach and its higher ca-
pacity. Since neither ILS nor GCA was a complete approach and landing system
in and of itself, the AAF, he told them, was developing two additional parts of an
integrated system: runway lights and a British invention called FIDO. McClel-
land’s testimony was seconded by Milton Arnold, who related his experiences in
168 blind landings
Europe, before laying out an expansive plan for installing a comprehensive set of
approach and landing aids, including the equipment the AAF had under devel-
opment in its All Weather Flying Program. His proposal ran to $36.65 million for
the 160 busiest airports in the United States.11
In late 1944, the Army Air Forces had established the All Weather Flying Pro-
gram to overcome what was rapidly becoming a greater danger to the AAF than
its human enemies, the weather. Originally stationed at Clinton County airfield,
near Wilmington, Ohio, the program moved to Lockbourne Army Air Base and
then back to Wilmington. The moves hindered it, and it did not achieve a great
deal before 1946, when it finally resettled in Wilmington. In early February 1946,
Maj. Gen. Curtis Lemay, then head of the Army Air Forces Research and Devel-
opment Board, called a weeklong conference of both civil and military oªcials to
discuss all available blind flying and landing equipment. At that meeting, the
army, navy, and CAA agreed to establish a landing aids experiment station to thor-
oughly evaluate an integrated set of approach and landing aids under real world
conditions. Because the AAF’s All Weather Flying Program operated out of a field
with insuªciently poor weather, the Landing Aids Experiment Station was to be
established at the former Naval Auxiliary Air Station at Arcata, California, known
as the airfield with the worst weather in the forty-eight states. Initially, the navy
administered the station while looking for a suitable contractor to run it. A brief
contract with the University of Southern California failed in mid-1946, and the
navy re-let the contract to United Airlines the same year.12
The Arcata Landing Aids Experiment Station (LAES) was an attempt to over-
come the inadequacies of both ILS and GCA. Each was e¤ective for the relatively
limited application of dealing with single or very few aircraft. For large numbers,
however, neither system was really satisfactory. ILS had no means of coordinat-
ing the movement of aircraft around an airfield, and that major deficiency re-
sulted in midair collisions among bomber formations during poor weather. ILS
needed a traªc control radar to be e¤ective. Further, with no way to provide pos-
itive distance separation between approaching aircraft, planes had to be separated
in time. That meant one plane had to land before the next one could be allowed
to leave its holding pattern and begin an approach. That reduced the theoretical
ILS handling capacity of one aircraft per minute to an actual rate of one every ten
to fifteen minutes, depending on the airport.13
GCA had its own problems. As already discussed, the search radar’s capabili-
ties were not adequate to serve as a good traªc control radar. The precision radars,
in turn, provided a glide path that was too short for heavy aircraft, and lengthen-
ing it meant substantially reducing its handling capacity because the final con-
transformations 169
troller could handle only one aircraft at a time. Lengthening the approach from
two miles to ten miles meant the controller had to work with each approaching
aircraft for five times as long. Adding operators could increase the capacity, but it
also raised the cost. The ILS had no such problem since once on the beam, pilots
could fly themselves in. One goal of the Arcata project was, therefore, to develop
procedures for using the two systems as complements to each other.
The other major goal was to evaluate two other, nonelectronic systems in-
tended to address the reality that neither ILS nor GCA was an e¤ective blind land-
ing system. Each was an e¤ective approach system, capable of bringing aircraft
down to altitudes of 100 feet or so, but although both systems had achieved com-
plete “zero-zero” landings, they could not achieve that safely with every landing.
The problem was runway alignment. Both systems could accurately put planes
directly above the runway, but could not guarantee that the plane was aligned to
go straight down the runway once the wheels touched. Pilots had to verify aircraft
alignment visually, which meant operationally that they had to see something out-
side the cockpit from an altitude of 100–200 feet, depending on the size of the
aircraft, in order to achieve alignment before touchdown or to abort the landing.
The two existing systems were FIDO and lights. FIDO was a British develop-
ment, first put into the RAF Coastal Command base at Davidstowe Moor, the
same place Alvarez’s prototype GCA was located under Arthur C. Clarke’s care.
It was simply two lines of large oil burners, one down each side of a runway. The
idea was to burn o¤ fog, creating an artificial ceiling of 200 feet or so above the
runway. FIDO burned approximately 100,000 gallons of fuel oil per hour, mak-
ing it a very expensive proposition, but it did remove fog from the field. The U.S.
Navy installed one at Amchitka, Alaska, during 1944 for tests but found the base
too remote for e¤ective testing after the war. At Arcata, FIDO’s fuel consumption
was reduced by half by improving the combustion eªciency of the burners, but
that still-high expense was compounded by the violent turbulence that the rising
heated air generated. Clarke reported that the turbulence was responsible for at
least one accident during the wartime British tests. Although large commercial
aircraft would be less a¤ected by turbulence than smaller planes, the combina-
tion of cost and turbulence made FIDO a commercial failure. One commercial
installation was made at Los Angeles in 1949, but no more followed it.14
FIDO would have been an e¤ective emergency landing aid, but commercial
aircraft could always go somewhere else less fogbound in the 2.5 percent of the
time that ceilings were below 200 feet at a particular field.15 Maintaining an ex-
pensive emergency system thus made little sense to the airlines. There was little
danger of an expensive crash without FIDO, given the prevalence of alternative
170 blind landings
airfields in the United States. Passenger discomfort at landing between two walls
of flame was no doubt another reason the airlines eventually rejected FIDO.
More controversial than FIDO was the issue of approach and landing lights.
Before World War II, both AAF and CAA believed strongly enough in “blind”
landings assisted by radio that they simply did not bother investigating the pos-
sibility of using lights to improve landing safety. In the late 1930s, the pilots’
union (ALPA) had begun pushing for approach lights of some sort, and CAA had
grudgingly begun looking at neon “approach light lanes” to provide some visual
guidance, but that program got lost in the war fever. Actual runway lights were
not explored at all. During the war, the AAF used an RAF lighting scheme that
was nothing more than relatively dim lights around the airfield edges called
perimeter lights; these were simply there to inform a pilot that he was over the
field. That was the total extent of lighting investigations until wartime experience
made it brutally obvious that an appropriate system of lights could make a big
di¤erence in helping pilots get that critical bit of visual alignment information
before landing.
The president of the pilot’s union fired the first public round in the lighting
controversy, which eventually managed to surpass the ILS/GCA controversy in
bitterness.16 In a February 1946 editorial, he argued that approach and runway
lights should be the first priority of the postwar CAA but was soundly ignored by
CAA through 1947. He brought it up again before the House Commerce Com-
mittee in his testimony during the air safety hearings and was rewarded with
placement of lights at the top of the committee’s own priority list, although lack
of agreement over what kind of lights kept Congress from providing funds for
large-scale procurement. The basic conflict was over CAA’s intention to install
1,500-foot strips of neon lights in airport approach zones, while the air force, navy,
the Air Transport Association and the pilot’s union wanted 3,000-foot strips of
high-intensity lights. However, the lights issue did not reach the level of true pub-
lic controversy until the Arcata project results began to appear in late 1948, per-
haps because the aeronautical press could not find a way to make lights as com-
pelling an issue as radar had been.
FAA historian John Wilson deals the lighting controversy in his oªcial history
of CAA, and it bears mention here only to report the denouement of the Arcata
project. Arcata tests demonstrated the superiority of a runway lighting arrange-
ment called “slope-line lighting” during 1948, after which the air force, navy,
CAA, Air Transport Association, and the British and French governments all
agreed to it. The pilots union refused to accept it, however, and in a true tour de
force, managed to defeat slope-line and get adopted its own center-line lighting
transformations 171
system in 1950.17 The pilots union thus humiliated nearly everyone, and in a fit
of pique, Congress ordered the Arcata project shut down. The members saw no
point in funding experimentation to determine the best methods if the results
were simply going to be ignored. Nonetheless, the union got what it wanted over
the next several years: the addition of a standardized lighting arrangement at
every runway served by an ILS system.
Ultimately, then, the Arcata project had mixed results. Although it was able to
reduce FIDO’s voracious fuel appetite and to demonstrate the best lighting
configuration, neither of those technical successes translated into widespread
use. Nor did the project add to the aviation community’s knowledge of air traªc
control issues or of the performance of ILS and GCA under real-world conditions.
Both systems had already been thoroughly lab tested, and the Landing Aids Ex-
periment Station was certainly a less realistic operating environment than that
available at the many ILS and GCA installations already in operation at commer-
cial and military airfields. It did, however, represent the first integration of the
now-common parts of our airport traªc control systems: surveillance radar, ILS,
and approach and landing lights.
The aviation agencies that had established Arcata station and its integrated ap-
proach to solving the blind landing problem thus had operational reasons for it
and CPN-18 traªc control radar projects. The need to speed traªc near congested
airports and prevent midair collisions in poor weather was suªcient to justify
adoption of expensive installations of the integrated system, at least in high-traªc
areas. But the operational considerations that had persuaded the experts had not
been persuasive enough to overcome Congress’ predisposition to consider ILS
and GCA as competitors, as the outcome of the air safety hearings demonstrates.
The aviation community, therefore, had to find a way to overcome the political
fallout from the controversy.
political imperatives for integration
The press attention given to the ILS/GCA controversy during late 1946 and
early 1947 had convinced House investigators that CAA and Army Air Forces
were pursuing di¤erent landing aids, which would impair national defense by
making civil airfields unusable by military aircraft. That the two agencies’ plans
were not actually incompatible was irrelevant. As long as members of Congress
believed that the two agencies disagreed, there existed a political issue that the ad-
ministration could use against the Republican Congress (as Truman’s Presiden-
tial Air Safety Board had done) or Congress could use against the administration
172 blind landings
(as the House Interstate Commerce Committee had done.) The landing aids con-
troversy thus reached the highest level of politics in the United States. Resolving
the 1947 controversy, in turn, meant that CAA and the Army Air Forces, two ex-
ecutive branch agencies, had to convince Congress that they no longer disagreed
and that their proposed solution was not merely a political ploy by the adminis-
tration intended to embarrass Republican congressional leaders. The controversy
made integration more diªcult to achieve, then, because even if the two agencies
agreed on a plan, it was still subject to partisan politics unless some means could
be found to present it in a nonpartisan format.
Even if CAA and the army agreed on landing aids, moreover, it was very clear
that one group was still strongly opposed to CAA’s plans. Despite the pious claims
made by Administrator Wright and the chairman of the Civil Aeronautics Board,
John Landis, during their congressional testimony that “all aviation interests” had
agreed to CAA’s plans, the controversy revealed that one aviation interest had not
been consulted: the Aircraft Owners and Pilots Association. Since AOPA repre-
sented private aviation, and private pilots made up the vast majority of all pilots
in the United States, it had considerable political suasion among Congressmen
interested in ensuring that whatever air navigation system was purchased served
the greatest number of people. Hence resolving the ILS/GCA controversy also
meant finding some way to appease AOPA and its private pilots, thereby satisfy-
ing demands by various members of Congress concerned with pursuing the
“greatest good” with taxpayer dollars, and, of course, keeping the votes of tens of
thousands of private pilots.
Integration was the obvious solution to both of these political imperatives.
Crafting a program that would provide both ILS and GCA services, as well as ap-
proach and landing lights, at virtually all airports and many smaller airfields
would appease AOPA and its congressional supporters, keep the Pentagon happy,
and thereby mollify Congress. Since integration did, in fact, have valid opera-
tional imperatives, a plan to deploy air traªc control radars (Airport Surveillance
Radars, or ASR, beginning in 1946), ground-controlled approach radars (Preci-
sion Approach Radars, or PAR), ILS, and landing lights was defensible from the
standpoint of operational need.18 The plan could not be assailed as purely politi-
cal. The two major potential roadblocks to integration were the cost of the pro-
gram, which would surely upset fiscal conservatives in both Congress and the par-
simonious Bureau of the Budget, and the reality that Republicans in Congress
might perceive a plan coming from the Truman administration as an attack on
them.
The cost of the program, as one CAA oªcial put it, would have to stand on its
transformations 173
own.19 Since public pressure had forced CAA to alter its own program, Wright
came to believe that Congress had to be the final arbiter of whether to fund a more
expansive, and expensive, infrastructure. One result of the controversy was that
after 1947 CAA felt more able to propose an expensive program than its tradi-
tional poverty would have allowed.
CAA was then essentially freed of budgetary restraints in its planning by a de-
cision of the Air Coordinating Committee to allow the Radio Technical Commis-
sion on Aeronautics to devise the program.20 That decision meant the plan would
no longer be the exclusive prerogative of the agency, as the AAF and navy were
also represented on the commission, but because this was a nongovernmental
committee sta¤ed largely by civilians, it had the freedom to design a system with-
out great concern about cost. This is not to say that cost was no object—the Air
Transport Association’s representatives could be counted on to help rein in costs
lest airlines be forced to help pay them—but that CAA’s own budget woes could
be subordinated to operational imperatives within the commission’s work.
More importantly, however, the Air Coordinating Committee’s decision to rely
on the nongovernmental RTCA e¤ectively removed the entire issue from parti-
san politics. Since RTCA was widely perceived to be a group of experts in an age
in which expertise was still considered objective and politically neutral, whatever
their product turned out to be would be acceptable to both the Congress and the
administration. Indeed it was. Fully published in January 1948, the report of
RTCA’s Special Committee 31 was hailed by everyone, immediately adopted by
the Congressional Air Policy Committee as its own, and then blessed by the ad-
ministration, whose own Air Policy Commission published its recommendations
too early to jump on the bandwagon.21 To get around the minor problem that
RTCA had no authority, the administration then established the Air Navigation
Development Board with a mandate to turn the committee’s recommendations
into reality.
RTCA’s grand plan did not, however, emerge directly from the Air Coordinat-
ing Committee’s decision. Although the agency had been moribund during the
war, it had begun meeting in April 1944 to consider how best to make use of all
the new radio-based technologies fostered during the war. The original constitu-
tion of RTCA’s executive committee, which initially had representatives from
AAF, navy, CAA, State Department, ATA, ALPA, Aeronautical Radio, Inc. (AR-
INC), and the Federal Communications Commission, was altered in early 1946
to add the one major “out” group, AOPA, at that organization’s request. This
helped resolve the private flier problem. In December 1945, Delos Rentzel, head
of ARINC, submitted an agenda item to the executive committee requesting that
174 blind landings
RTCA undertake the development of a comprehensive plan to overcome traªc

control and air navigation problems.22 His request, in turn, was based on a letter
from the Air Transport Association. RTCA’s examination of the problem thus be-
gan before the House investigation and the Air Coordinating decision to turn the
decision over to them.
Because much of RTCA’s attention during early 1946 was devoted to matters
concerning coordination with the new Provisional International Civil Aviation
Organization, it did not begin meeting to develop an air navigation plan until Au-
gust 1946. It produced a brief document that reflected an attempt to grasp the
complexities of the new technologies but lacked specificity. It did not detail the
amount or cost of equipment needed, and it merely suggested that more research
and development needed to be done. Nonetheless, this early report included all
of the landing aids that were then being tested at Arcata in its recommendations
for “approach zone facilities.” During 1947, RTCA refined its recommendations
carefully, to ensure that the program e¤ectively served private, military, and com-
mercial fliers. Largely mirroring the Air Transport Association’s program in its
landing aids recommendation, RTCA ultimately recommended an interim target
for 1953 of 320 ILS, 150 airport surveillance radar, and 82 precision approach
radar installations. It also stipulated that some of the ASR and PAR sets be in-
stalled at noncommercial fields so that service was provided to private aviation.
In this way, RTCA intended to heal the rift between CAA and the private fliers’
organization. The widespread deployment of precision approach radar that the
group proposed also ensured that the radar enthusiasm that the ILS/GCA debate
had revealed among the public was addressed, while ILS, already widely deployed,
remained to satisfy the pilots union and the airlines’ desire for a system now, not
in some Buck Rogers future.23 Finally, this resolved the second political impera-
tive that drove integration. By giving each group the particular technology it fa-
vored, RTCA e¤ectively bought their agreement. If no dissenting voices echoed
in Congress, the controversy would simply vanish.
RTCA’s proposal, however, was far more than simply a duplication of the Ar-
cata project at every commercial airport and many private flying fields in the na-
tion. In the strong political winds, the commission had taken a long reach and
proposed nothing less than the creation of an integrated, nationwide, air naviga-
tion and traªc control system. By interlinking (networking, in modern terms) au-
tomatic long- and short-range radar, the group envisioned positive radar control
over every mile of U.S. airspace. Further, it foresaw the use of automation to elim-
inate labor-intensive flight progress strips and flight clearance management, al-
lowing automatic tracking of aircraft from takeo¤ to touchdown, and eventually
transformations 175
from gate to gate. It was an ambitious plan expected to take twenty years to com-
plete. It was also a prescription for a very expensive command research and de-
velopment program, and on the Scylla of cost and the Charybdis of technocracy
the plan eventually ran aground.
conclusion
Solving the “blind landing problem” required overcoming unforeseen diªcul-

ties that derived from both operational and political imperatives. No one in the
1930s had envisioned that massive increases in traªc would necessitate not only
bad-weather landing aids but also robust traªc control aids. The AAF’s experi-
ences during World War II demonstrated that need and further revealed the ut-
ter inadequacy of its earlier attempts to beat the weather. By the end of the war,
both AAF and CAA realized that more research and development was needed to
truly defeat weather constraints on aviation. The operational inadequacies of
SCS-51 and AN/MPN-1 meant both systems needed to be supplemented and
improved, an imperative the aeronautical press entirely missed. Their di¤erent
models of operation, pilot control and ground control, proved highly comple-
mentary, and their integration into an alliance, albeit an uneasy one, was one
foundation of the modern air traªc control system.
The creation of an integrated system of approach and landing aids was not
solely driven by operational demands, however. The public frenzy over air safety
and radar took the issue out of bureaucrats’ oªces and put it before a Congress
only too willing to grab it, attack the administration with it, and demand a solu-
tion. The dizzying political heights to which the ILS/GCA issue ascended in the
United States ensured that the radar versus instrument landing question was at
least addressed, even if the answer, SC-31, eventually failed of its promise.
Despite congressional and administration approval of SC-31 and allocation of
funds for it beginning in FY1949, the next four years saw the program stall. The
Korean War caused the diversion of increasing portions of CAA funding to the
Pentagon, weakening its ability to influence Air Navigation Development Board
research and essentially eliminating CAA’s airways equipment program, while
the election of Dwight Eisenhower installed a president who, as Walter McDou-
gall has argued, resisted the siren call of technocratic management and command
research and development. Without strong leadership from the chief executive,
the Air Navigation Development Board’s cooperative e¤ort quickly disintegrated
into interagency squabbling, and it achieved little. As a result, CAA had a total
of ten precision approach radar systems in 1950, and the number never grew
176 blind landings
larger.24 The ten airport surveillance radar sets it had installed by 1950 also re-
mained unchanged through 1953, and the number of long-range surveillance
radars remained fixed at two—both World War II sets on permanent “loan” from
the air force—until another air safety crisis caused Congress to dismantle the
ine¤ectual CAA and replace it in 1958. The resulting new agency, the Federal Avi-
ation Agency, was responsible for finally pulling together the postwar technolo-
gies of aviation into the modern air traªc system.25
Conclusion
On August 6, 1997, Korean Air flight 801 crashed near the top of Nimitz Hill,
on the U.S. territory of Guam, near the transmitting antenna of the visual omni-
range (VOR) that aircraft approaching the island’s primary airport outside the city
of Agana use to maneuver onto the ILS approach path. Immediately, the press
turned its attention to the “ancient” ILS installation, which had both of its glide
path transmitters out of service. The plane’s crew was aware of the outage and
had been briefed on the approach procedure for use when the glide path trans-
mission is unavailable, but they did not properly follow it. Ultimately, the National
Transportation Safety Board ruled the accident a product of “pilot error.” There
were, of course, contributing factors.1 But as one diehard supporter of Luis Al-
varez’s ground control approach system wrote to the industry magazine Aviation
Week and Space Technology, this accident would not have happened if GCA (or its
civil equivalent, precision approach radar) had been the landing aid chosen forty
years before.2
So much is certainly true. The final controller of a GCA installation would have
been able to warn the plane’s crew that they were descending toward Nimitz Hill
and not the runway three miles further along their flight path. But a GCA world
would have had accidents of its own. Perhaps the most famous GCA failure came
amid its most important success, the Berlin airlift of 1948 and 1949. On August
13, 1948, forever after known as Black Friday, the airfield in the U.S. sector of
Berlin, Templehof, su¤ered a series of accidents. One C-54 missed the runway
and crashed, while the aircraft behind it blew its tires trying to stop before hitting
the wreck. A third C-54 crashed on the auxiliary runway, catching fire and burn-
178 blind landings
ing up plane and cargo.3 The relatively inexperienced final controller in Tempel-
hof ’s GCA unit, who would normally have been able to prevent these sorts of ac-
cident by proper instructions to the approaching pilots, lost his “nerve,” as the eu-
phemism of the day went. As a result of Black Friday, the airlift commander
established much stricter traªc control procedures and put in place one of the
central rules of modern commercial aviation: all aircraft had to operate under in-
strument flight rules, regardless of weather conditions. That e¤ectively placed the
aircraft under radar control, while ensuring that all pilots followed the same ex-
acting flight procedures.
These two bad days forty-nine years apart indicate that neither ILS nor GCA,
ultimately, was perfect. If GCA could have prevented the KAL accident in Guam,
some other catastrophe related to GCA’s imperfections would have taken its place
in the long years since its 1943 debut. What makes the ILS/GCA decision still
contentious forty years later is not competing claims for safety, although the dis-
pute is often expressed in those terms, but the question first raised by Rep. M.
Harris Ellsworth during the 1947 hearings: how to do the most good for the most
people? In other words, whom should the airways be built to serve? That ques-
tion places the landing aids issue into the “democratic technology” arena in the
historiography of technology. What, exactly, a democratic technology might look
like is as problematic as the issue of what a democracy is, but that has not stopped
scholars from trying to define them.
Among the most prominent advocate of rules to define and regulate tech-
nologies is Langdon Winner. Winner proposes a general maxim, that “technolo-
gies be given a scale and structure of the sort that would be immediately intelligible to
non-experts.”4 Acceptable technologies would be limited to those whose operation
is obvious to the proverbial “casual observer.” That would, unfortunately, exclude
both ILS and GCA, as even some commercial pilots cannot explain how ILS
works. They only know how to use it. Indeed, Winner’s rule would strike out all
aviation technologies. Most people cannot explain how aircraft fly and perhaps it
bears mentioning here that no one since the Wrights has been able to get a re-
production of the famed 1903 Flyer aloft. If its flight had not been recorded on
film, historians would probably have had to reject their claim to the first powered
flight because their experiment has proven unrepeatable. If even the Wright
Flyer’s operation is nonobvious, then there are no simple technologies of flight.
But are there no democratic ones either?
Perhaps a more useful way to approach the issue of democratic technologies
can be drawn from feminist historiography of technology. Although much of that
conclusion 179
literature focuses on various aspects of gender and technology, one underex-

plored aspect deals with issues of access. “Access” means many things to many
people, from representation in the design process to physical possession of a tech-
nology, so perhaps substitution of a more precise term is in order. By “accessibil-
ity” I mean what we might call usability. Alvarez’s GCA was clearly more useable
to an average pilot than was ILS, as most pilots, congressmen, and aviation writ-
ers recognized. One implication of GCA’s usability by a broader range of fliers
would have been more opportunity to fly for owners of small aircraft. Though po-
litical philosophers may balk at the crudity of this analysis, it seems reasonable
to assert that a technology whose characteristics make it useable by a larger num-
ber of people than its competitors is the more accessible, and the more demo-
cratic.
Although Joseph Corn locates the death of the dream of democratic aviation
in the late 1940s, the persistence of GCA advocates suggests that the dream is
alive and well in AOPA, still the most contrarian organization in American avia-
tion. J. B. “Doc” Hartranft’s organization has continued to be a persistent critic
of CAA successor, the Federal Aviation Administration. In particular, the organi-
zation has vehemently resisted the adoption and maintenance of radar control of
traªc around the busiest U.S. cities. Although fans of radar landing, AOPA’s
members chafe at the complex traªc control procedures that FAA has built
around its surveillance radars to manage traªc, and blame those procedures for
diminished interest in personal flying among the public. Radar has been, at best,
a mixed blessing for its original champion.
As one might expect, FAA was driven to expand the scope and complexity of
its radar-based traªc control procedures by the rapid growth of commercial avi-
ation. Because the nation’s busiest airports face much greater traªc than did
Berlin during the airlift—one landing every three minutes in Berlin meant one
landing every nine minutes at each of Berlin’s three airports, a trivial rate by mod-
ern Chicago-O’Hare standards—the traªc control procedures are relatively more
intricate and rigorous. The e¤ect of those procedures, as AOPA has complained
for two decades, has been to drive general aviation out of the major airports, re-
ducing its members’ access to the skyways. In some sense, that result is a denial
of the dream of many of the “air-minded” pioneers of aviation that Corn docu-
mented so well, but increased commercial traªc has also meant low enough air-
fares that nearly everyone can fly, even if the piloting is done by someone else. If
the democracy granted by low air fares is not quite what the true believers in dem-
ocratic aviation intended, perhaps they can be forgiven for not quite grasping the
180 blind landings
realities of twenty-first-century America: in an increasingly urban society, not

everyone can own a plane. The parking issues alone make the idea almost laugh-
able.
Along with more complex rules under which flying is done, aviation has seen
a proliferation of interest groups representing specific subsets of the industry.
While CAA had to deal only with the Air Transport Association, ALPA, and AOPA
in addition to the army and navy during the development of ILS and the ILS/GCA
controversy, FAA must now deal with a plethora of what it calls the “alphabet
groups.” These include, for example, the Experimental Aircraft Association (EAA),
representing owners of home-built aircraft; the Allied Pilots Association (APA);
the National Air Traªc Controller’s Organization (NATCO); and the National
Business Aircraft Association.
The proliferation of these groups has created a situation similar to that de-
scribed by Brian Balogh in his study of the decline of the nuclear power industry
in the United States.5 He refers to the construction of “issue networks” sur-
rounding the industry composed of activist groups focused on various aspects
of nuclear plant safety, but in aviation the situation is a bit di¤erent. While
the groups Balogh studied were oriented around a single issue, and were often
ephemeral, the alphabet groups are permanent organizations active in a wide ar-
ray of issues within aviation. Sociologists, of course, also perceive such networks,
redefined as “actor networks,” in their studies of the activities of scientists and
engineers and the managerial professionals that support them. Actor networks,
like the often-ephemeral issue networks Balogh finds, seem too impermanent to
explain the longevity of technologies. We should describe the network formed by
the alphabet groups, FAA, and the Pentagon’s four di¤erent air arms as an insti-
tutional network. Although sociologists understandably balk at allowing faceless
institutions agency, the character of the major aviation organizations has changed
very little over time, and their positions on various aviation issues have also been
changeless. Individuals are shaped, in part, by the institutions they represent, and
the conception of an institutional network allows for personalities to act in that
context.6
The negotiations between CAA, the airlines, and the Army Air Corps during
the 1930s stretched over years due to a combination of di¤ering technical and
training needs and misunderstandings. Commercial adoption of a standardized
landing aid was retarded by the need to negotiate a standard acceptable to all par-
ties in order to gain their support. FAA is often criticized for its inability to adopt
new technologies quickly, and one reason is simply its need to satisfy the conflict-
ing demands of the growing number of alphabet groups. While CAA’s ILS ne-
conclusion 181
gotiations in 1938 and 1939 involved only itself, the air corps, airline representa-
tives, and the representatives of a few radio companies, FAA now must deal with
a much wider range of interest groups. It no longer has the assistance of RTCA’s
extragovernmental committee structure to aid the negotiation process, however.
The Kennedy administration removed the government representatives from
RTCA, believing that allowing industry and government representatives on the
same committee violated conflict-of-interest laws. RTCA still exists as a private
foundation dedicated to aiding the adoption of new technologies into aviation by
fostering communication among the various groups, but it no longer holds the
informal authority it once did.
The Federal Aviation Administration thus faces a more diªcult task in devel-
oping acceptable standards than did its predecessor. Although FAA’s legal au-
thority is much greater than that of its predecessors, it has limits. It cannot sim-
ply ignore the alphabet groups because those groups have proven perfectly willing
to wage media campaigns and lobby congressional supporters over disagree-
ments. And the alphabet groups are often successful in their challenges. The de-
velopment of a complex institutional network has ensured that the alphabet
groups scrutinize every technological decision FAA makes, and aspects of the
technology found wanting by a particular group are subjected to lengthy negoti-
ation.
The composition of the institutional network I have described changed over
time and not simply due to the proliferation of alphabet groups. The relative
power of various groups to a¤ect the industry and its technologies has also
changed. During the 1930s, the demands made by commercial airlines on CAA
and its predecessor, the Bureau of Air Commerce, to deploy networks of weather
stations, radio navigation ranges, and communications stations made commer-
cial considerations the dominant influence in infrastructure development, while
World War II put the Army Air Forces in the pilot’s seat. The cold war and the
related national security state kept the air force strong enough to prevent the
reemergence of a dominant CAA during the 1950s, until Congress, via the Fed-
eral Aviation Act of 1958, rather surprisingly made the new FAA entirely re-
sponsible for navigational infrastructure and air traªc control. That, in turn, en-
sured continuation of the commercial bias in aviation policy, while the research
and development resources available to the air force enabled that organization to
dominate in new technologies of flight. Like CAA at the end of World War II, FAA
is generally left to modify air force technologies for civil use, or more commonly,
wait for a private company to commercialize the technology, then approve it.
The diversification of the institutional structure of U.S. aviation that began in
182 blind landings
the 1930s also permanently linked the industry to academic organizations. While
the Guggenheim Foundation had funded aeronautical research at several uni-
versities during the 1920s, that mode of support for research did not give the re-
searchers a policy voice. Further, aeronautics research, though very important,
was also circumscribed. The laboratories founded by the Guggenheim performed
extensive research into aerodynamics, but not in electronics, propulsion, radio,
or physiology.7 Edward Bowles had to scrape to get funding for his Round Hill
group before CAA took interest in 1937, although MIT possessed one of the
Guggenheim labs. The Army Air Corps’ wartime needs cemented the university
research–aviation linkage, and expanded the scope of government-supported avi-
ation research in universities to include all aspects of flight. Although the MIT
Radiation Lab closed after World War II, the air force had grown so used to hav-
ing its own microwave scientists that it got MIT to establish a new lab to assist
it, the Lincoln Laboratory. That organization, in turn, was crucial to FAA’s micro-
wave landing system (MLS) development program during the 1970s. MLS was
supposed to replace ILS during the 1980s and 1990s, but its financial costs
proved more than either administration or Congress were willing to bear. Simi-
larly, the Pentagon’s JASON group maintained a policy voice for scientists in civil
aviation, perhaps most prominently in advising President Kennedy to build a sep-
arate air traªc control system during the 1960s, instead of using the air force’s
SAGE air defense system.8
There are, therefore, tight links between military and commercial aviation, at
di¤erent levels. They increasingly share the same technologies, while the same
laboratories and scientists serve to advise military and civil aviation policy oª-
cials. Yet this was not always the case, as Chapters 1–3 suggest. During the 1920s
and early 1930s, military and commercial technologies diverged, with commer-
cial airlines demanding runways and greater precision in navigation and landing
aids, while the Army Air Corps was satisfied with grass fields and the first gen-
eration of radio aids—low-frequency communications, ranges, and compass lo-
cators. The cause of the divergence was commercial demands for regularity of
service, an issue that did not matter to the interwar Air Corps. Why the recon-
vergence, then? World War I was fought without runways and radio navigation
aids, as Eddie Rickenbacker’s description of “dud days” attests, and no one seems
to have minded the sometimes days-long absence of aircraft over the battlefields.
What made World War II airmen suddenly adopt commercial technologies de-
signed to allow all-weather flying?
Veterans of the Eighth Air Force generally consider the film Twelve O’Clock
High to be the truest depiction of their war against the Luftwa¤e, and perhaps
conclusion 183
from there we can begin to draw an answer: maximum e¤ort.9 Hap Arnold placed
enormous pressure on Ira Eaker to show results for the vast expenditures made
by the United States to equip Eighth Air Force with its aerial armada. Without a
doubt, that reflected pressure Roosevelt and the joint chiefs placed on Arnold in
order to justify expending those resources on bombers and not on ships, landing
craft, and tanks. In part, this pressure was a result of Arnold and Eaker’s own ar-
gument, that strategic bombing could achieve decisive victories early in the war.
As Eighth failed to produce those victories, the AAF’s field commanders, with
their RAF counterparts, tried instead to exert constant pressure on the enemy.
Eaker thus sought to fly in conditions that would have seen Rickenbacker’s Fly-
ing Circus safely grounded in order to produce that maximum e¤ort, and his re-
placement, James Doolittle, pushed the weather limits still harder. To that end,
the Army Air Forces, leaning heavily on commercial technologies, sought to be-
come more like the airlines. All-weather flying (and bombing and air combat)
would allow near-continual pressure on the enemy. Radar, once integrated with
the commercially inspired radio navigation equipment designed during the
1930s, enabled completion of the “all-weather flying system” the airlines and
CAA had begun through blind landing research. The AAF thus adopted a mili-
tary version of regularity of service during the war as its goal. The motivation of
the two groups was di¤erent: whereas the airlines sought to become profitable al-
ternatives to the railroads, the AAF sought to defeat the Luftwa¤e, justify its bud-
get, and achieve its bid for independence from the army’s hated ground forces.
For di¤erent reasons, then, the AAF and the airlines came to require all-weather
flying.
Recent scholarship in the history of science has shown that the Defense De-
partment launched substantial research e¤orts into atmospheric sciences after
World War II, based on use of the German V-2 missile. Understanding of the
weapon’s operating environment, researchers believed, was necessary to im-
proving its performance.10 Yet that was, in essence, no more than an extension
of the AAF’s belated recognition of the importance of environmental considera-
tions in aviation. Weather-related crashes were not the only evidence of the AAF’s
prewar lack of respect for nature’s inconsistency, after all. The unheated, un-
pressurized B-17 proved murderous to early crews, who could be permanently dis-
abled by frostbite if the Luftwa¤e failed to kill them. Gun crew members also
sometimes froze to death, while others su¤ocated on vomit that froze in their oxy-
gen masks.11 The commercial airlines had recognized the weather as their pri-
mary enemy in the late 1920s; the AAF, by not adopting the dream of all-weather
flight as quickly as the airlines had, found itself nearly as vulnerable to nature as
184 blind landings
it initially was to the Luftwa¤e. The Pentagon’s deployment of vast resources af-
ter World War II to understand what its oªcers perceived as the “operating envi-
ronment” makes perfect sense, given the experiences its leaders had in fighting
the European weather.
By the beginning of the Berlin airlift, then, a remarkable convergence in com-
mercial and military aviation had occurred, centered on the goal of regularity.
Knowledge of the operating environment was merely a means to that end. Shar-
ing the same goal did not necessarily mean that these two subsets of aviation
shared precisely the same technologies. Although Curtis LeMay adopted ILS as
the air force’s primary landing aid in 1950 at its permanent installations, the air
force has kept more modern versions of the ground controlled approach system
in service for deployment to forward airfields because it is much less sensitive to
site conditions. In a practical sense, it is simply more portable than the site-
sensitive ILS. Hence the di¤erences between military and commercial aviation’s
operating circumstances continue to influence their technological choices.
Finally, there is the evolution of the concept of blind landing. The concept had
considerable suasion and staying power. The technological frame constructed
around the technologies of radio blind landing caused its believers to overlook an
“old technology” partial solution, lighting. It should not have taken the experi-
ences of World War II to convince aviation oªcials that some judiciously chosen
application of Thomas Edison’s invention could help fliers land. But belief in the
possibility of blind landings guided solely by radio was compelling enough that
even after the war, Charles Stanton and others insisted on retaining the name in-
strument landing system, still convinced that it could bring about routine blind
landings.
Fortunately for the flying public, retention of the name never led to implemen-
tation of blind landings using ILS, although experimentation along those lines
continued for many years. Much of that experimentation moved into microwave-
based landing aids like Sperry Gyroscope’s in the quest for greater stability. One
system developed during the 1960s for the Lockheed C-141 military transport, the
all-weather landing system (AWLS), was supposed to achieve widespread com-
mercial deployment, but it did not. The only true blind landing systems in use to-
day are the microwave landing system designed in the late 1960s that the NASA
space shuttle uses, and the navy’s carrier controlled approach system, a fully au-
tomated version of GCA. MLS was approved for worldwide deployment in 1977
but was canceled in 1995 due, in part, to commercial development of the Penta-
gon’s global positioning system, which promises to cost FAA an order of magni-
tude less than MLS would have. And although the carrier controlled approach sys-
conclusion 185
tem is installed on all of the U.S. Navy’s dozen aircraft carriers, it is never used.
The navy’s carrier pilots, who are willing to be fired o¤ the deck by big steam-
powered pistons, be shot at by enemies of whatever description, and let their re-
turning aircraft get caught by thick cables and sometimes nets stretched across
the deck, cannot countenance a blind landing.
In this study, we have seen the devolution of “blind landing” into “instrument
approach,” and the reason is simply this: although engineers and scientists have
produced equipment capable of blind landings, they have not been able to make
pilots capable of it. The human element of the system has proven to be the most
diªcult to engineer properly. The problem is not one of training or skills but one
of psychology. Although military and commercial pilots make their living in one
of the most technically oriented professions humanity has ever created, they ap-
pear to be almost genetically incapable of turning the most dangerous part of a
flight over to a machine that, in blind conditions, they cannot monitor. This is the
result of that most basic animal instinct, survival. A minor glitch—all too com-
mon even in the most modern aircraft—in any other part of a flight is generally
an annoyance. A minor glitch during a blind landing is almost certain to be fatal
to pilots, passengers, and quite likely people on the ground, too. Hence RAF vet-
eran Frank Griªths recounts that after several weeks of making blind landings
“under the hood” in 1944 using Major Moseley’s approach coupler, he and his
crew went out to try it for real by making a night approach with field lights and
the plane’s landing lights turned o¤. He found the pressure too much, though,
and faced with dying if the black box was wrong, he cheated and turned on the
plane’s landing lights.12
The devolution of the concept of blind landing occurred as the technology im-
proved, but more importantly, as knowledge of pilots’ reactions to the technolo-
gies of blind landing was developed. That knowledge was not formal. “Human
factors” research in aviation began under AAF funding only during World War II
and remains a hotly contested form of knowledge. Yet there was clearly pilot feed-
back in the process of design, testing, and negotiation, with the pilots’ standard
for the “almost-straight” glide path of 1939 the clearest example. The airlines were
the first to adopt the stand that blind landings were impractical, likely motivated
in part by the repeated refusal of pilots to use the various descendants of the NBS
system as landing systems, as opposed to approach aids. The goal of all-weather
flying drove the evolution of blind landing systems, but was not enough to over-
come the mental hazard pilots were subjected to in a blind landing. Blind land-
ing systems were redefined as instrument approach aids and air traªc control
systems in the face of human psychological limitations.
186 blind landings
Despite its ultimate failure, the dream of blind landing was thus important
to aviation because it stimulated development of an array of technologies that
proved important to related needs. Those technologies, in turn, allowed com-
mercial aviation to achieve greater safety and greater regularity than any other
transportation system while permitting round-the-clock, almost-all-weather aer-
ial combat and bombardment. But the dream asked too much of the human part
of the system. Although Harry Guggenheim announced the achievement of blind
landing in 1929, the twentieth century did not see its routinization.
Notes
abbreviations
AFHRA Air Force Historical Research Agency, Maxwell AFB, AL

NA National Archives, Washington, DC
NAS Archives National Academy of Sciences Archives, Washington, DC
RG record group
USGPO U.S. Government Printing Oªce
introduction
1. Antoine de Saint-Exupéry, Night Flight (New York: Century, 1932).

2. Charles Lindbergh, The Spirit of St. Louis (New York: Charles Scribner’s Sons, 1953),
6–7.
3. “A Classification of the Causes of Forced Landings, from January 1924 to July 1925,
with Recommendations,” 7 November 1925, file A14.48: Mgr. Monmouth Ill. Development
Division, box 47, Records of the Post Oªce Department, General Correspondence of the
Air Mail Service, record group (RG) 28, National Archives, Washington, DC (NA).
4. I’m grateful to my colleague Michael Buckley for the phrase “military-postal-indus-
trial.”
5. William M. Leary, “The Search for an Instrument Landing System, 1918 –1948,” in
Innovation and the Development of Flight, ed. Roger Launius (College Station: Texas A&M
University Press, 1999), 80–99; James R. Hansen, “Aviation History in the Wider View,”
Technology and Culture (1989), 643–56.
6. Nick A. Komons, Bonfires to Beacons: Federal Civil Aviation Policy under the Air Com-
merce Act, 1926–1938 (1978; reprint, Washington, DC: Smithsonian Institution Press,
1989), 151.
7. Ibid., 125–46. The best history of the airmail service is William M. Leary, Aerial Pi-
oneers: The U.S. Airmail Service, 1918–1927 (Smithsonian Institution Press, 1985). Only
nine of the Post Oªce’s original forty mail pilots were still alive in 1925; Roger E. Bilstein,
Flight in America: From the Wrights to the Astronauts, revised ed. (Baltimore: Johns Hopkins
University Press, 1994), 52. For comparison of railroad and airline times during the early
1930s, see Edward W. Constant II, Origins of the Turbojet Revolution (Baltimore: Johns Hop-
kins University Press, 1980), 166–67.
188 notes to pages 6–15
8. Robert Wohl, A Passion for Wings: Aviation and the Western Imagination, 1908 –1918
(New Haven: Yale University Pess, 1994), quote from p. 1.
9. Joseph Corn, The Winged Gospel: America’s Romance with Aviation, 1900 –1950 (New
York: Oxford University Press, 1983), vii.
10. Ibid., 46–50.
11. See Tom D. Crouch, “An Airplane for Everyman: The Department of Commerce and
the Light Airplane Industry, 1933–1937,” in Launius, Innovation and the Development of
Flight, 166–87.
12. Corn, Winged Gospel, 139–40; Komons, Bonfires to Beacons, 68. Boeing Air Trans-
port merged with National Air Transport, Varney Air Lines, and Pacific Air Transport to
form the core of United Air Lines. See Frank J. Taylor, High Horizons (New York: McGraw-
Hill, 1964).
13. The best synthesis of strategic bombing fantasies is Michael S. Sherry, The Rise of
American Airpower: The Creation of Armageddon (New Haven: Yale University Press, 1987),
1–9. The classic histories of strategic bombing are Raymond H. Fredette, The Sky on Fire:
The First Battle of Britain, 1917–1918 (New York: Harcourt, Brace, Jovanovich, 1976) and Lee
Kennett, A History of Strategic Bombing (New York: Scribner, 1982). On Mitchell’s struggle
to militarize aviation, see Sherry, Rise of American Airpower, 29 –31, 34 –37, and Bilstein,
Flight in America, 43. A full length treatment is Alfred F. Hurley, Billy Mitchell: Crusader for
Air Power (Bloomington: Indiana University Press, 1975). John R. M. Wilson, Turbulence
Aloft: The Civil Aeronautics Administration Amid Wars and Rumors of Wars, 1938 –1953 (Wash-
ington, DC: Federal Aviation Administration, 1979), 63–83.
14. Eric Schatzberg, Wings of Wood, Wings of Metal (Princeton: Princeton University
Press, 1999), 5, 15–19.
15. William Cronon, Nature’s Metropolis: Chicago and the Great West (New York: W. W.
Norton, 1991), xix.
16. An Army Air Forces C-54 cargo plane rigged with a special autopilot system flew
nonstop from Stephensville, Newfoundland, to Ireland in 1947. See “Automatic Controls
for Pilotless Ocean Flight,” Electronics (December 1947): 88–92.
17. Erik M. Conway, “Echoes in the Grand Canyon: Public Catastrophes and Technolo-
gies of Control in American Aviation,” History and Technology 20, no. 2 ( June 2004): 115 –
34.
chapter 1: instrumental faith
1. Monte Dwayne Wright, Most Probable Position: A History of Aerial Navigation to 1941
(Lawrence: University Press of Kansas, 1972).
2. See ibid., 70, for a detailed description of the magnetic compass’s airborne misbe-
havior.
3. Annual Report of the Advisory Committee for Aeronautics, 1915 –1916, Flight (24 Au-
gust 1916): 720–22.
4. See Elmer Sperry Jr., ts. dated December 1969, Elmer Sperry Papers, National
Air and Space Museum Archives, Washington, DC; description based on U.S. Patent
1,407,491, Elmer A. Sperry, “Turn Indicator,” 21 February 1922. Pioneer Instrument Com-
notes to pages 15–22 189
pany memos on the patent issues involved make clear that the idea for the Sperry Turn In-
dicator came from Luis de Florez of the Navy Bureau of Construction and Repair (during
World War II, de Florez became head of Navy flight instrument research), while Charles
H. Colvin, who left Sperry Gyroscope to help found Pioneer Instruments in 1919, actually
perfected the turn indicator. Elmer Sperry did not initially believe the turn indicator would
work. See M. M. Titterington to Lt. E. W. Rounds (Aviation Section, Bureau of Construc-
tion and Repair), 9 August 1921, Elmer Sperry Papers. Historian Thomas Hughes gives a
di¤erent account of the turn indicator in Elmer Sperry: Inventor and Engineer (Baltimore:
Johns Hopkins University Press, 1971), 233–36. “Le Gyrorector,” L’Aérophile (1–15 Janvier
1930): 16.
5. Charles Colvin, interview, 27 May 1971, p. 22, Lawrence B. Sperry Papers, National
Air and Space Museum Archives; Elmer Sperry Jr., ts. dated December 1969, Elmer Sperry
Papers.
6. Quote from Elmer Sperry to Lawrence Sperry, August 10, 1920, Lawrence Sperry Pa-
pers. On Alcock’s flight, see Sir John Alcock and Sir Arthur Whitten Brown, Our Transat-
lantic Flight (London: William Kimber, 1969). William Davenport, Gyro! The Life and Times
of Lawrence Sperry (New York: Charles Scribner’s Sons, 1978), 253. Apparently, the engine
on his Sperry Messenger failed and he glided in. A coastguardsman saw the airplane go
down and a search found the plane two hours after the crash about two miles from shore,
but Lawrence was never found.
7. Elmer Sperry to Lawrence Sperry, 10 August 1920, Lawrence Sperry Papers.
8. The best history of the U.S. Post Oªce’s airmail service is William Leary, Aerial Pi-
oneers: The U.S. Air Mail Service, 1918–1927 (Washington, DC: Smithsonian Institution
Press, 1985).
9. Ibid., 87–89.
10. Ibid., 171–85.
11. Ibid., 205.
12. Eighteen pilots died in the first thirty-four months of Post Oªce airmail service,
fully half the original complement of pilots. Ibid., 95–112, 147.
13. Ibid., 86–87, 109, 188–90.
14. Ray H. Boudreaux, “The Ocker-Meyers Method of Blind Flying,” Aero Digest (July
1928): 48, 183–85; William C. Ocker, “Economic Value of Flying by Instruments,” Aero Di-
gest (October 1930): 62–63.
15. William C. Ocker and Carl J. Crane, Blind Fight in Theory and Practice (San Antonio,
TX: Naylor Printing Company, 1932), 7.
16. A German instrument called the “Gyrorector” existed that was a workable artificial
horizon, but it appears not to have been marketed in the United States. The army obtained
one via a military attaché in Europe and tested it at Wright Field in 1926 but did not find
it suitable for military aircraft. Its weight also discouraged airline use.
17. William C. Ocker, “Under the Hood,” Southwestern Aviation (May 1934), 4 –6, 28.
18. Ocker and Crane, Blind Fight, 12; I use the definitions of “natural” and “mechani-
cal” pilots described by Ocker supporter Boudreaux in “The Ocker-Meyers Method,” in this
analysis (see n. 14).
19. Hughes, Elmer Sperry, 234.
20. “Rosenbaum Gyrorector,” War Department report, 28 May 1926, ts., Turn and
Bank Instruments Technical File, National Air and Space Museum Archives.
21. Hughes, Elmer Sperry, 235.
22. Richard Hallion, Legacy of Flight: The Guggenheim Contribution to American Aviation
(Seattle: University of Washington Press, 1977), 101–27; Solving the Problem of Fog Flying
(New York: Daniel Guggenheim Fund for the Promotion of Aeronautics, 1929); Hallion,
Legacy, 117.
23. Elmer A. Sperry Jr., “The Artificial Horizon,” Aeronautical Engineering (October–
December 1930): 289–91.
24. Hallion, Legacy, chap. 6; James H. Doolittle, “Flying an Airplane in Fog,” SAE Jour-
nal (March 1930): 318–20, 345.
25. Wolfgang Langewiesche, “Flying Blind,” Harper’s Magazine (April 1947): 328.
26. “Blind Flying and the Airlines,” Aviation (August 1932): 349. See also Harold C.
Stark, Instrument Flying, rev. ed. (Pawling, NY: Harold C. Stark, 1934), iv.
27. Stark, Instrument Flying, 6, 8.
28. Ibid., 15. The same system was published by two army fliers the following year un-
der the name “A-B-C system.” The army fliers’ book is William C. Ocker and Carl J. Crane,
Blind Flight in Theory and Practice (San Antonio, TX: Naylor Printing, 1932).
29. The Air Corps introduced instruction in June 1930. See “Blind Flying and the Air-
lines,” 350.
30. E. A. Cutrell, “Instrument and Radio Flying,” Aviation (June 1935): 11–14; E. B.
Schaefer, “The Stark System of Instrument Flying,” Aviation Engineering (March 1932):
21–22; Karl S. Day, Instrument and Radio Flying (Garden City, NY: Air Associates, Inc.,
1938), vi.
31. “Blind Flying and the Airlines,” 349–52.
32. Day, Instrument and Radio Flying, 1.
33. Private pilots had no minimum number of solo hours to receive a private license at
this time. Komons, Bonfires to Beacons, 97; Luis W. Alvarez, Adventures of a Physicist (New
York: Basic Books, 1987), 30.
34. Robert B. Parke and Lloyd L. Kelly, The Pilot Maker (New York: Grosset & Dunlap,
1970): 3.
35. Ibid., 18–20, 26.
36. Ibid., 36.
37. Ibid., 38.
38. Leary, Aerial Pioneers, 224–25.
39. The best account of Brown’s activities is F. Robert Van der Linden, “Progressives
and the Post Oªce” (Ph.D. diss., George Washington University, 1996). See also Komons,
Bonfires to Beacons; Leary, Aerial Pioneers, 222–37; Frank J. Taylor, High Horizons (New York:
McGraw-Hill, 1964); and Robert J. Serling, Eagle: The Story of American Airlines (New York:
St. Martin’s, 1985), 63–76. F. Robert van der Linden, The Boeing 247: The First Modern Air-
liner (Seattle: University of Washington Press, 1991), 20–21; Taylor, High Horizons, 73 –94;
Komons, Bonfires to Beacons, 197–210.
40. This account is a very brief synthesis of sources that do not all agree on how, ex-
actly, Roosevelt came to cancel the contracts, although they do agree on why. See Komons,
Bonfires to Beacons, 219–99; Norman E. Borden, Air Mail Emergency 1934 (Freeport, ME:
Bond Wheelwright, 1985); and Benjamin Foulois with Col. C. V. Glines, From the Wright
Brothers to the Astronauts (New York: McGraw-Hill, 1960): 235 –59.
41. Komons, Bonfires to Beacons, 261–62, 272–73.
42. Parke and Kelly, Pilot Maker, 52–53.
chapter 2: places to land blind
1. Deborah G. Douglas, “The Invention of Airports” (Ph.D. diss., University of Penn-

sylvania, 1996), 322–23.
2. John Law has argued for the importance of understanding “natural” factors in the
design and success of technological systems. See his “Technology and Heterogeneous En-
gineering: The Case of Portuguese Expansion,” in The Social Construction of Technological
Systems, ed. Wiebe E Bijker, Thomas P. Hughes, and Trevor Pinch (Cambridge, MA: MIT
Press, 1987), 111–34.
3. Maj. Gen. Benjamin D. Foulois, with Col. C. V. Glines, From the Wright Brothers to
the Astronauts (New York: McGraw-Hill, 1968), chap. 7.
4. Douglas, “Invention of Airports,” 275–77.
5. Capt. Eddie V Rickenbacker, Fighting the Flying Circus (Garden City, NY: Doubleday,
1965), 9–10, describes World War I French aerodromes.
6. A. C. Blackall, “Why Britain Uses Grass for Runways,” Airports (September 1928):
38; Hans-Joachim Braun, “The Airport as Symbol: Air Transport and Politics at Berlin-
Tempelhof, 1923–1948,” in From Airships to Airbus: The History of Civil and Commercial Avi-
ation, Vol. 1: Infrastructure and Environment, ed. William Leary (Washington, DC: Smith-
sonian Institution Press, 1995), 45–54.
7. H. Oakley Sharp, G. Reed Shaw, and John A. Dunlop, Airport Engineering (New York:
John Wiley & Sons, 1944).
8. “Addendum to Air Service Information Circular, vol. IV, no. 303—Discussion of Air-
plane Tires and Wheels” (Washington, DC: USGPO, 15 September 1922), 4.
9. Pressure is, by definition, the force, or load, applied to each square inch of surface
area. Mathematically, that is, Pressure ⫽ Force/Area. With pneumatic tires, increasing the
force, or load, on the tire also increases the tire’s contact area with the ground. The pres-
sure exerted on the ground is therefore not a direct linear relationship to the load applied.
Ronald Miller and David Sawers, The Technical Development of Modern Aviation (London:
Routledge, 1968) is the classic work focusing on the achievement of greater eªciency,
power, and performance in aircraft.
10. Bendix Landing Gear Service Bulletin, 1 March 1945, located in “Landing Gear, Gen-
eral,” Technical File Y4000200, National Air and Space Museum Archives, Washington,
DC.
11. Deborah Douglas, “Airports as Systems and Systems of Airports: Airports and Ur-
ban Development in America Before World War II,” in Leary, Airships to Airbus, 55 –86.
12. Nick A. Komons, From Bonfires to Beacons: Federal Civil Aviation Policy under the Air
Commerce Act, 1926–1938 (1978; reprint, Washington, DC: Smithsonian Institution Press,
1989), 130–31. This was the source of one of Foulois’s early conflicts with Billy Mitchell,
who wanted airfields placed in the northeast, nearer to most of the army’s infrastructure.
See Foulois, Wright Brothers to the Astronauts, 139.
13. Eddie Rickenbaker referred to such days as “dud days” in his memoir of World War
I flying (35).
14. Archibald Black, “Landing Field Roads and Runways,” Aeronautical Digest (April
1923): 253–54, 292; (May 1923): 330–32.
15. Archibald Black, Civil Airports and Airways (New York: Simmons-Boardman Pub-
lishing Company, 1928).
16. Quincy Campbell, “Runways of Brick,” Airports (September 1928): 8 –9.
17. Douglas, “The Invention of Airports,” 289–330.
18. C. N. Connor, “Airports and Transportation Engineering: E¤ect of Airplane Impact
on Airport Surfaces,” Aviation Engineering (December 1932): 18 –21.
19. William F. Centner, letter to the editor, Airports (August 1929): 27, 28, 49; Wendell
Miller, “The Drainage Factor in Airport Site Selection,” Airports (March 1929): 9 –10, 44.
20. The tendency of airport designers to minimize costs using centerline drains ended
when a CAA study in the mid-1930s found that many runways had cracked after the pipes
under them had collapsed. Airport engineers then adopted the current practice of putting
the drainage piping down each edge of the runways, which at least doubled the amount of
piping required; Edwin A. Miller, “Good Runways and Drainage at Rochester Airport,” Air-
ports (August 1929): 42–44, 58.
21. D. W. Crum, “Iowa City,” Airports (March 1931): 19–20; “Airports in Pictures,” Air-
ports (July 1929): 32.
22. Jerold Brown, Where Eagles Land (New York: Greenwood Press, 1990, 40.
23. Ibid., 82.
24. Charles Stevenson, “Induction through Air and Water at Great Distances without
the Use of Parallel Wires,” Proceeding of the Royal Society of Edinburgh, 20 (1892–93): 25 –
27; H. Cooch, “Landing Aircraft in Fog,” Journal of the Royal Aeronautical Society (June
1926): 365–93.
25. John S. Gray, “The Loth Leader Cable System for Electrical Steering of Aeroplanes,”
Proceedings of the Institution of Aeronautical Engineers 9 (1923): 7–30; Monte Duane Wright,
Most Probable Position: A History of Aerial Navigation to 1941 (Lawrence: University Press of
Kansas, 1972); A. Verdurand and J. Blancard “Utilisation des procédés Loth pour le guidage
des avions par ondes hertziennes,” L’Aérotechnique (October 1930): 364 –76; Paul Larivière,
“Le câble de guidage-son emploi pour l’atterrissage sans visiblité,” L’Aéronautique (March
1935): 33–39; Capitain P. Franck and Capitain A. Vomerange, “Le guidage des avions par
câbles électriques,” L’Aéronautique (1924): 39–47.
26. William Loth, “On the Problem of Guiding Aircraft in a Fog or by Night When
There Is No Visibility,” National Advisory Committee for Aeronautics Technical Memo 57,
January 1932. Translated from Comptes Rendus des Seances de l’Academie des Sciences, no. 23,
5 December 1921.
27. A. Verdurand and J. Blancard, “Utilisation des procédés Loth pour le guidage des
avions par ondes hertziennes,” L’Aérotechnique (October 1930): 364 –76; Paul Larivière, “Le
câble de guidage-son emploi pour l’atterrissage sans visiblité,” L’Aéronautique (March 1935):
n o t e s t o p a g e s 4 7 –5 9 193
33 –39. Captain P. Franck and Captain A. Vomerange, “Le guidage des avions par câbles
électriques,” L’Aéronautique (1924): 39–47.
28. Charles Çhristienne and Pierre Lissarague, A History of French Military Aviation,
trans. Francis Kianka (Washington, DC: Smithsonian Institution Press, 1986), 210 –43,
310; Verdurand and Blancard, “Utilisation,” 375.
29. Cooch, “Landing Aircraft in Fog,” 365–93.
30. Ernest T. Williams criticized Cooch’s paper, ibid., 388. However, Cooch discounted
the comments (390–91) as did Mr. Handley Page (383).
31. Ibid., 391.
32. Ibid., 391; F. W. Meredith, “Air Transport in Fog,” Journal of the Royal Aeronautical
Society (February 1931): 75–85.
33. For a detailed discussion, see Chapter 3.
34. Tymms in Meredith, “Air Transport in Fog,” 90. Emphasis added.
Innovation and the Development of Flight, ed. Roger D. Launius (College Station: Texas A&M
University Press, 1999), 80–99.
36. Frederic Celler, “Landing Blind: The Loth System of Energized Cables for Fog Land-
ing Guidance,” Aviation (December 1931): 699–700; M. Heinrich Gloeckener, “Methods
for Facilitating the Blind Landing of Airplanes,” NACA Technical Memo 687, trans. by
Dwight M. Miner from “Verfahren zur Erleichterung von Blindlandungen,” Zeitschrift für
Flugtechnik und Motorluftschi¤ahrt 23 (24 June 1932): 12.
37. “Hanson Announces Fog-Landing Device,” Aviation (22 February 1930): 402.
38. Edward Nelson Dingley Jr., “An Instrument Landing System,” Communications 18,
no. 6 ( June 1938): 7–9.
39. G. H. Mills to Chief of the Bureau of Ships, 14 February 1941; Chief of the Bureau
of Ships to Chief of Naval Operations, 1 August 1941, both in file: C-F42-1/88, box 18, Bu-
reau of Ships, General Correspondence 1940–1945, RG 19, NA. Stark’s order is contained
in the fourth indorsement to the basic letter.
chapter 3: radio blind flying
1. A. Hunter Dupree, Science in the Federal Government (Baltimore: Johns Hopkins Uni-
versity Press, 1986), 271–77. The name National Bureau of Standards was in use from 1901
to 1903, when the organization became Bureau of Standards. That name lasted until 1934,
when the organization was renamed National Bureau of Standards again. It is now the Na-
tional Institute of Standards and Technology. For clarity, I will use Bureau of Standards
throughout and abbreviate it NBS. William F. Snyder and Charles L. Bragaw, Achievement
in Radio (Washington, DC: USGPO, 1986), 43.
2. The oªcial history is Nick Komons, Bonfires to Beacons: Federal Civil Aviation Policy
under the Air Commerce Act, 1926–1938 (Washington, DC: Smithsonian Institution Press,
1989).
3. Ibid., 147–63.
4. H. Diamond, “Use of the Radio Beacon System for Landing in Fog,” 19 November
194 n o t e s t o p a g e s 6 1 –6 6
1928, unpublished ms. in box 10, Papers of J. H. Dellinger, RG 167, NA. Diamond was
born in Minsk, Russia, in 1900 and was naturalized a citizen on 4 June 1923. He had pre-
viously served as an instructor in electrical engineering at Lehigh University until joining
the NBS in July 1927; the term “hertz” was not in use at the time, but for purposes of clar-
ity, I will use it exclusively in this chapter. The historically correct term is “cycles per sec-
ond,” with the usual metric system prefixes “kilo-,” “mega-,” etc. One hertz is by definition
one cycle per second, and they are therefore identical mathematically, if not historically.
5. Richard Hallion, Legacy of Flight (Seattle: University of Washington Press, 1977).
6. James H. Doolittle, “Flying an Airplane in Fog,” SAE Journal (March 1930): 318 –20,
345. Doolittle reported that the aircraft could withstand landing at 1,000 fps but he could
not.
7. Doolittle’s equipment is described in Hallion, Legacy of Flight, and The Daniel
Guggenheim Fund for the Promotion of Aeronautics, Equipment Used in Experiments to
Solve the Problem of Fog Flying (New York: The Guggenheim Fund, 1930). The airline esti-
mate is United’s. See J. R. Cunningham to Major Chester Snow, 16 November 1934, in file
827.1 vol. 1: July 1934–July 1937, folder 1 of 2, box 359, CAA Central Files, RG 237, NA.
8. E. F. W. Alexanderson, “Height of Airplane above Ground by Radio Echo,” Aviation
Engineering (December 1928): 12–13; A. P. Rowe, “Flying and Landing in Fog,” Aircraft En-
gineering (July 1939): 169–72.
9. H. Diamond and F. W. Dunmore, “A Radio Beacon and Receiving System for Blind
Landing of Aircraft,” Bureau of Standards Journal of Research, RP 238 (1930): 897–931; R.
Schultz, “‘Nebellanden’ mit hilfe von Richtfunkbaken” [Fog Landing with Help from Di-
rective Radio Beacons], Zeitschrift für das Weltflugwesen (March 1931): 129 –34; Secretary of
the Navy to Secretary of Commerce, 9 August 1933, file: F42-1/88 vol. 1, box 789, Bureau
of Aeronautics Confidential Correspondence, RG 72, NA; an example is K. Baumann and
A. Ettinger, “A New System for Blind Landing of Airplanes,” Proceedings of the Institute of
Radio Engineers 24, no. 5 (May 1936): 751–54; a long-running program to develop a system
using very low frequency electromagnetic fields is detailed in the previous chapter.
10. Marshall Boggs to Col. H. H. Blee, 14 September 1931, file 827.1 vol. 1: June
1930–June 1934, folder 2 of 2, box 359, CAA Central Files, RG 237, NA.
11. Deborah G. Douglas, “The Invention of Airports” (Ph.D. diss., University of Penn-
sylvania, 1996), 248–50; W. S. Hinman Jr., “Preliminary Report on Tests of the Blind Land-
ing Systems at Newark Airport,” 23 May 1934, file: Blind Landing Newark 1932–1934, box
14, Papers of J. H. Dellinger, RG 167, NA.
12. Some controversy exists over the status of the Newark and Oakland installations. A
1939 Aero Digest article states that United got custody of the discarded Newark installation
from the Bureau of Air Commerce in 1934 and transported it to Oakland. See Henry W.
Roberts, “United Air Lines Radio Laboratory,” Aero Digest (November 1939): 60, 62. Leary
and Douglas have followed this article. Internal documents of the Bureau of Air Com-
merce, file 827.1 vol. 1: Inst. Landing Systems, box 359, CAA Central Files, RG 237, NA,
and W. E. Jackson, “The Status of Instrument Landing Systems,” CAA Technical Develop-
ment Report no. 1 (Washington, DC: Civil Aeronautics Authority, October 1937) make clear
that these were two separate installations. The Newark equipment was moved to Indi-
anapolis in 1936, where the Bureau of Air Commerce began working on it again.
13. W. S. Hinsman Jr., “Preliminary Report on Tests of the Blind Landing Systems at
Newark Airport,” 23 May 1934, file: Blind Landing Newark 1932–1934, box 14, Papers of
J. H. Dellinger, RG 167, NA.
14. Memo from G. Fulton to Head of Material Branch, Bureau of Aeronautics, 16 De-
cember 1935, file: F42-1/88 vol. 1, box 789, Bureau of Aeronautics Confidential Corre-
spondence, RG 72, NA. The Washington Institute of Technology consisted of a number of
laid-o¤ Bureau of Standards radio researchers who were interested in continuing the blind
landing work.
15. William Leary, “The Quest for an Instrument Landing System,” Innovation and the
Technology of Flight, ed. Roger D. Launius (State College, TX, 1999), 93.
16. Ibid., 80–99; Capt. George V. Holloman, “Inspection of Instrument Landing Sys-
tems,” 24 June 1936, file 413.44, July–December 1936, box 1256, Oªce of the Chief Signal
Oªcer Correspondence, RG 111, NA.
17. S. H. Ingersoll to Bureau of Engineering, 31 March 1939; Commander Aircraft,
Scouting Force to Chief of the Bureau of Engineering, 13 December 1937; Capt. Mark
Mitscher to Chief of the Bureau of Engineering, 11 August 1939, all in file: F42-1/88 vol. 1,
box 789, Bureau of Aeronautics Confidential Correspondence, RG 72, NA.
18. Roy Jackson to Chief of the Bureau of Engineering, 31 May 1940, file “F-42-1/88”
vol. 2, box 789, Bureau of Aeronautics Confidential Correspondence, RG 72, NA.
19. Commander, Patrol Wing Seven to Chief of the Bureau of Aeronautics, 14 Febru-
ary 1942, file “F-42-1/88” vol. 3, box 789; Captain C. A. Nicholson to the Chief of the
Bureau of Aeronautics, 10 November 1943, and Chief of BuAer to Chief of BuShips, 10 Oc-
tober 1943, file “F42-9,” box 892, all in Bureau of Aeronautics Confidential Correspon-
dence, RG 72, NA.
20. Foulois, From the Wright Brothers to the Astronauts (New York: McGraw-Hill, 1968),
207, 224.
21. Monte Duane Wright, Most Probable Position: A History of Aerial Navigation to 1941
(Lawrence: University Press of Kansas, 1972), 86–87, 90–94.
22. Ibid., 127.
23. Foulois claims in his memoirs that the army’s work of 1934 was the beginning of
ILS and GCA; Wright Brothers to the Astronauts, 258; Chester Snow to Rex Martin, 6 June
1934, file 827.1 vol. 1: June 1930–June 1934, folder 2 of 2, box 359, CAA Central Files, RG
237, NA.
24. Cunningham to Eugene Vidal, 31 October 1934; C. C. Shangraw, 28 November
1934; Walter J. Addems, 5 November 1934; H. M. Hucke to Cunningham, 7 November
1934; H. B. Sneed to Jack Frye, 4 January 1934, all extracted in Lloyd L. Juleson to Rex Mar-
tin, 5 April 1935, file 827.1: July 1934–July 1937, folder 1 of 2, box 359, CAA Central Files,
RG 237, NA.
25. C. C. Shangraw, 28 November 1934, ibid.
26. J. R. Cunningham to J. Lyman Briggs, 1 May 1935; Chief of the Bureau of Aero-
nautics to Chief of Naval Operations, 25 November 1935, both in file “F42-1/88,” box 7, Bu-
reau of Engineering Confidential Correspondence 1940–1945, RG 19, NA.
27. W. E. Jackson, “The Status of Instrument Landing Systems,” Technical Development
Report no. 1 (Washington, DC: Civil Aeronautics Authority, October 1937).
28. I have not been able to find reliable ceiling statistics for the 1930s. A 1949 docu-
ment puts the percent of time below a 500-foot minimum for La Guardia at 5.5 percent,
and for Los Angeles, 5.2 percent. See Edgar A. Post, “Airline Operating Experiences with
the Instrument Landing System,” Institute of Aeronautical Sciences, 22 July 1949.
29. Diamond and Dunmore, “Radio Beacon and Receiving System”; R. Stüssel, “The
Problem of Landing Commercial Aircraft in Fog,” Journal of the Royal Aeronautical Society
(1934): 807–36. The Lorenz system was produced in conjunction with the German aero-
nautical research institute, Deutsche Versuchsanstalt für Luftfahrt; Jackson, “Status of In-
strument Landing Systems,” 7–9.
30. Carl Lorentz A. G. to Major Gardner, 18 March 1936; First indorsement to Lt. Col.
C. K. Nulsen to Military Attache, Berlin, Germany, 27 January 1936, both in folder 413.44:
Inst. Landing no. 1, box 1256, Oªce of the Chief Signal Oªcer Correspondence, RG 111,
NA.
31. First indorsement to Lt. Col. H. H. Fuller to Assistant Chief of Sta¤ (G-2), 19 Feb-
ruary 1936; third indorsement to Lt. Col. H. H. Fuller, ibid.
32. Wright, Most Probable Position, 103–29.
chapter 4: the promise of microwaves
Innovation and the Development of Flight, ed. Roger Launius (College Station: Texas A&M
University Press, 1999), 94–95.
2. William Osmun, The Authority of Agreement: A History of RTCA (Washington, DC:
RTCA, 1985).
3. Captain G. V. Holloman and Major F. S. Borum, “Lorenz Instrument Landing Sys-
tem,” file 413.44: Inst. Landing no. 4, box 1256, Oªce of the Chief Signal Oªcer Corre-
spondence, RG 111, NA.
4. John O. Mauborgne to Chief Signal Oªcer, 19 May 1937, ibid. The Chief Signal
Oªcer of the Army was responsible for radio development within the Army, including air-
craft radio; therefore the Director of the Aircraft Radio Laboratory at Wright Field worked
for him. This required considerable coordination with the Army Air Corps.
5. Noise abatement requirements ended the “straight in approach” during the 1970s,
except in bad weather. The ILS, obviously, permits only straight approaches.
6. “Pilots Ultimatum Condemns Washington Airport”; “Fears for Friends Flying to
Capital, Mrs. Roosevelt Says,” both in The Air Line Pilot (July 1937): 1; “An Editorial in Pic-
tures,” The Air Line Pilot (November 1937): 1. The editorial was reprinted from the Wash-
ington Times; “Ask Removal of Hazardous Obstructions,” The Air Line Pilot (November
1937): 4; FDR’s decision was to move the airport to Gravelly Point on the Virginia bank of
the Potomac River; it later became National Airport.
7. W. E. Jackson to Gazely, 28 June 1937, file 827.1: July 1934–July 1937, folder 1 of 2,
box 359, CAA Central Files, RG 237, NA; David Behncke, “New Developments on Instru-
ment Landing System,” The Air Line Pilot 6 (August 1937): 1, 5, 7, quote p. 5. Freng was Su-
perintendent of Flying for United at this time; Paul Goldsborough to Major A. W. Marriner,
25 June 1937, file 413.44: Inst. Landing no. 4, box 1257, Oªce of the Chief Signal Oªcer
Correspondence, RG 111, NA. Marriner communicated extensively with the Aircraft Radio
Laboratory, which probably explains why a copy of this letter is in the Signal Oªce’s files.
Donald MacKenzie discusses the influence of ARINC in the context of the laser ring gyro-
scope in “From the Luminiferous Ether to the Boeing 757: A History of the Laser Ring Gy-
roscope,” Technology and Culture (1993): 475–515.
8. Secretary of War to Secretary of Commerce, 21 April 1938; Col. Frank M. Kennedy to
the Chief of the Air Corps, 13 April 1938, both in file 413.44: Inst. Landing no. 5, box 1257,
Oªce of the Chief Signal Oªcer Correspondence, RG 111, NA.
9. H. H. Buttner and A. G. Kandorian, “Development of Aircraft Instrument Landing
Systems,” Electrical Communications 22, no. 2 (1945): 179–92.
10. Report of Meeting of Subcommittee #4, Instrument Landing Devices, Radio Tech-
nical Committee for Aeronautics, file 827.1 vol. 2: November 1938 –December 1939, folder
1 of 2, box 359, CAA Central Files, RG 237, NA. For the record, the pilots were Otis Bryan,
TWA; Fred Davis, Eastern; R. T. Freng, United; E. A. Cuttrell and B. O. Howard, American.
Northwest Air Lines and National Air Line were also represented, but not, apparently, by a
pilot.
11. Ibid., 2.
12. Preston Bassett, 27 October 1939, file: Instrument Landing, box 35, Sperry Gyro-
scope Company Collection, Hagley Museum and Library, Wilmington, Delaware.
13. Report of Meeting of Subcommittee #4.
14. Ibid., 14. Arnold later concurred with Mitchell’s assessment in an independent
memo to Clinton Hester, 2 November 1939, file 413.44, box 673, AAF Central Decimal
Files, RG 18, NA.
15. Report of Meeting of Subcommittee #4, 8.
16. Ibid.
17. John M. R. Wilson, Turbulence Aloft: The Civil Aeronautics Administration in War and
Rumors of War, 1938–1958 (Washington: FAA, 1980), 1–42; CAA had no authority over mil-
itary aviation, unlike the current FAA. That problem was not rectified until 1958, when
CAA (which was organizationally revised in 1941 into an Administration, operating under
the Commerce Department) was replaced with the FAA.
18. See W. L. Barrow and L. J. Chu, “Theory of the Electromagnetic Horn,” Journal of
the Institute of Radio Engineers 27, no. 1 (January 1939): 51.
19. Donald G. Fink, “3 Spots and a Horn,” Aviation 37 (September 1938): 28, 29, 73, 74;
Alex Soojung-Kim Pang, “Edward Bowles and Radio Engineering at MIT, 1920 –1940,”
Historical Studies of the Physical Sciences 20, vol. 2 (1990): 313–37.
20. E. L. Bowles, W. L. Barrow, W. M. Hall, F. D. Lewis, and D. E. Kerr, “The CAA-MIT
Microwave Instrument Landing System,” preprint of AIEE paper dated 2 December 1939,
in file 413.33: Inst. Landing no. 6, box 1458, Oªce of the Chief Signal Oªcer Central Files,
RG 111, NA.
21. Arthur L. Norberg and Robert W. Seidel, “The Contexts for the Development of
Radar: A Comparison of E¤orts in the United States and the United Kingdom in the
1930s,” Tracking the History of Radar, ed. Oskar Blumtritt, Hartmut Petzold, and William
Aspray (Piscataway, NJ: IEEE, 1994), 199–216; undated ts., folder 8, box 4, Russell Varian
series, Varian Papers, SC 345, Stanford University Archives.
198 n o t e s t o p a g e s 9 1 –9 6
22. Undated ts., Russell Varian series; copy of agreement, 26 April 1938, folder 11, box
5, ibid. This later agreement references the original dated 6 October 1937.
23. Undated ts.; agreement between Leland Stanford Junior University and Sperry Gy-
roscope Company, 27 April 1938, folder 11, box 5, Russell Varian series.
24. The Chief Signal Oªcer was a Major General billet, and upon promotion to Brig-
adier General and assignment as CSO, he was breveted to Major General. The visit appears
to have been arranged by the Oªce of the President, in the White House. Compton trans-
mitted his thanks for the visit through a memo addressed to the Oªce of the President,
which then wrote to Mauborgne. See Allen W. Horton to Maj. Gen. John Mauborgne, 30
April 1938, file 413.44, box 1256, Oªce of the Chief Signal Oªcer Correspondence, RG 111,
NA.
25. Col. Hugh Mitchell to Col. Bender (the addressee is “My Dear Bender”), 26 March
1938, ibid.
26. Sperry Gyroscope Company, “A Proposal for an Instrument Landing System,” n.d.
[late 1939], ts., file: Instrument Landing #21, box 35, Sperry Gyroscope Company Collec-
tion, Hagley Museum and Library.
27. Clinton Hester to Maj. Gen. H. H. Arnold, file 413.44, box 673, AAF Central Deci-
mal Files, RG 18, NA.
28. Maj. Gen. H. H. Arnold to Assistant Secretary of War, 20 August 1939, ibid.; H. H.
Arnold, Global Mission (New York: Arno Press, 1949), 165.
29. Franklin D. Roosevelt to Dr. Paul Brockett, 30 August 1939; Frank B. Jewett to Pres-
ident of the United States, 1 September 1939, both in file 413.44, box 673, AAF Central Dec-
imal Files, RG 18, NA.
30. Frank Jewett to President, 24 November 1939, ibid.
31. Charles Stanton to Vannevar Bush, 14 October 1939, file 827.1 vol. 2: November
1938 –December 1939, folder 1 of 2, box 359, CAA Central Files, RG 237, NA.
32. Edgar S. Gorrell to Vannevar Bush, 30 September 1939, Committee on Airplane In-
strument Landing Equipment Collection, National Academy of Sciences Archives, Wash-
ington, DC (NAS Archives).
33. Bush, head of the Carnegie Institute of Washington, was also a member of MIT’s
Electrical Engineering faculty and therefore not quite a disinterested party.
34. Lt. Col. Hugh Mitchell to Richard Gazely, 25 October 1939, General file, Commit-
tee on Airplane Instrument Landing Equipment Collection, NAS Archives. Mitchell’s state-
ment is a summation of one sent by Maj. Gen. Arnold to Clinton Hester, 2 November 1939,
which also stated that the straight glide path should extend to at least ten miles. Letter in
file 413.44, box 673, AAF Central Decimal Files, RG 18, NA.
35. Edward L. Bowles to Vannevar Bush, 2 November 1939, Committee on Airplane In-
strument Landing Equipment Collection, NAS Archives.
36. Ibid.
37. He referred to ultra-high frequency, but he clearly meant what we call (and others
at the time called) microwave. At this time, the frequency spectrum did not yet have its cur-
rent structure, and di¤erent terms were in use for the same frequency band. Col. Mitchell,
for example, referred to Bowles’ work with super-high frequencies, a term Bowles never
used.
38. Bowles to Bush, 2 November 1939.

39. Ibid.
40. Vannevar Bush to Members of the Conference Committee on Instrument Landing
Systems, 20 November 1939, General file, Committee on Airplane Instrument Landing
Equipment Collection, NAS Archives.
41. Ibid.
42. Vannevar Bush to Frank Jewett, 21 November 1939; Frank Jewett to President Roo-
sevelt, 24 November 1939, file 413.44B; Brig. Gen. Edwin Watson to Hon. Harry H.
Woodring, 5 December 1939, cited in Harry H. Woodring to Brig. Gen. Edwin Watson, 28
December 1939, file 413.44, all in box 673, AAF Central Decimal Files, RG 18, NA. Clin-
ton M. Hester to Vannevar Bush, 18 December 1939, General file, Committee on Airplane
Instrument Landing Equipment Collection, NAS Archives.
43. Paul Brockett to Frank Jewett, 21 December 21; Paul Brockett to Frank Jewett, 26
December 1939, both in General file, Committee on Airplane Instrument Landing Equip-
ment Collection, NAS Archives. It isn’t clear whether the White House was irritated with
the Authority, a five-member board that directed the agency, or the Administrator, who was
supposed to actually run it. The confusing structure of the CAA was resolved in 1940, when
the Authority was separated out as the Civil Aeronautics Board, and the administrative
functions were placed in the Civil Aeronautics Administration, a unit of the Commerce De-
partment.
44. Woodring to Watson, 28 December 1939. Woodring’s reply was taken verbatim
from Maj. Gen. Arnold’s response to the report. See 2nd indorsement to Adjutant General
to Chief of the Air Corps, 20 December 1939, file 334.8: National Academy of Sciences, box
1773, Oªce of the Adjutant General Central Files, RG 407, NA.
45. Maj. Gen. H. H. Arnold to Chief Signal Oªcer, 20 December 1939, file 413.44, box
673, AAF Central Decimal Files, RG 18, NA.
46. Franklin Roosevelt to Brig. Gen. Edwin Watson, 5 January 1940, ibid.; Adjutant
General to Chief Signal Oªcer and Chief of the Air Corps, 9 January 1940, file 334.8: Na-
tional Academy of Sciences, box 1773, Oªce of the Adjutant General Central Files, RG 407,
NA.
47. Second indorsement to Adjutant General to Chief Signal Oªcer and Chief of the
Air Corps, 11 January 1940; Harry Woodring to Brig. Gen. Watson, 16 January 1940, file
334.8: National Academy of Sciences, box 1773, Oªce of the Adjutant General Central Files,
RG 407, NA.
48. Maj. Gen. Arnold to Brig. Gen. Watson, 12 February 1940, folder 413.44, box 673,
AAF Central Decimal Files, RG 18.
49. Clinton Hester to Maj. Gen. H. H. Arnold, 12 April 1939; Sixth indorsement to Air-
craft Radio Laboratory to Chief Signal Oªcer, 10 October 1939, ibid. The Sixth indorse-
ment is signed by Arnold, under date 20 December 1939. Second indorsement to Adjutant
General to Chief of the Air Corps, 20 December 1939, file 334.8: National Academy of Sci-
ences, box 1773, Oªce of the Adjutant General Central Files, RG 407, NA. Note that the
date on this indorsement is the same as the date under which Arnold asked the Chief Sig-
nal Oªcer to contract with Sperry for the microwave system. Given that the Chief Signal
Oªcer’s memo had begun its trek through the bureaucracy in early October, it seems very
200 notes to pages 101 –107
likely that Arnold was waiting for the NAS recommendations before making his own rec-
ommendation to the CSO. In his response, Arnold quoted the committee’s microwave rec-
ommendations verbatim.
50. Anonymous to Maj. General Arnold, 12 March 1940, file 413.44, box 673, AAF Cen-
tral Decimal Files, RG 18, NA.
51. Arnold, Global Mission, 165; Wilson, Turbulence Aloft, 28.
52. Lt. Col. Orlando Ward to the Chief of the Air Corps, 3 April 1940; Brig. Gen. R. K
Yount to the Secretary of War, 13 April 1940; draft of memo (with routing info) in Oªce of
the Adjutant General, all in file 334.8: National Academy of Sciences, box 1773, Oªce of
the Adjutant General, RG 407, NA. Final and “EMW” (Brig. Gen. Edwin Watson) to The
President, 2 May 1940, in file 413.4B, box 673, AAF Central Files, RG 18, NA.
53. Pan American, of course, used seaplanes extensively, but that organization had its
own system. Since it was not technically a domestic airline, it also does not seem to have
been invited to these conferences. It certainly never attended.
chapter 5: instrument landing goes to war
1. The glide path was 330 MHz. This was a UHF frequency until 1944 when it was re-
designated VHF.
2. Major E. M. Powers, 20 June 1940, file 413.44, box 673, AAF Central Decimal Files,
RG 18, NA.
3. According to the Army, seven sets built with 1940 and 1941 money and installed dur-
ing 1941 were Chicago, Cleveland, Fort Worth, Kansas, Los Angeles, New York/La Guardia,
and Washington, DC. Oªce of the Chief of the Army Air Forces, “Air Force Communica-
tions Policy: Instrument Approach and Landing Systems,” 31 October 1941, in unmarked
folder, box 404, RD 158, Records of Wright Field (Sarah Clarke Collection), RG 342, NA.
4. Sir Charles Webster and Noble Frankland, The Strategic Air O¤ensive Against Ger-
many, 1939–1945 (London: Her Majesty’s Stationary Oªce, 1961); H. H. Arnold, Global
Mission (New York: Arno Press, 1949), 215–40; David B. Langmuir to Lee DuBridge, 14
January 1943, file: Project 102, box 404, RD 158, Records of Wright Field (Sarah Clarke Col-
lection), RG 342, NA.
5. See “Air Force Communications Policy,” n. 186; David Little, 21 November 1941, file
413.44: Inst. Landing no. 7, box 1459, Oªce of the Chief Signal Oªcer Central Files, RG
111, NA.
6. Henry L. Stimson to Secretary of Commerce, 19 May 1942, file 413.44: Inst. Land-
ing no. 8, box 1459, Oªce of the Chief Signal Oªcer Central Files, RG 111, NA. Although
the date of the oªcial letter is May, the arrangements for this program began in January.
7. International Telephone Development Corp., a subsidiary of International Tele-
phone and Telegraph, changed its name to International Telephone and Radio Manufac-
turing Company (ITRM), in 1941, and then to Federal Telephone and Telegraph Company,
apparently in 1943. The name changes were likely due to its connection to ITT, which was
attacked in the press, and investigated by the military, over its subsidiaries in Germany (of
which Lorenz A.G. was one).
8. John M. R. Wilson, Turbulence Aloft: The Civil Aeronautics Administration during Wars
and Rumors of War (Washington, DC: FAA, 1979). The agency was the Interdepartmental
Air Traªc Control Board.
9. See, for example, CNO to CG, AAF, 21 June 1944, file: F42-9 vol. 1, box 892, Bureau
of Aeronautics Confidential Correspondence, RG 72, NA.
10. Major Lyman D. Swendson to Chief, Maintenance Division, Air Service Command,
11 May 1944, file: Project 102, box 404, RD 158, Records of Wright Field (Sarah Clarke Col-
lection), RG 342, NA, contains the allocation and priority lists for the various theaters. Re-
ceivers were held up for months at shipping depots: see Deputy Directory, ATSC, to CG
ETO, 14 September 1944, file: Project 102 (UK); Major F. L. Moseley, “Visit Report to Air
Transport Command and Oªce of the Chief Signal Oªcer,” 10 November 1943, file: Proj.
102 (CAA Correspondence), both in ibid. By May 1944, there were twelve in operation in
the United States.
11. William Rose to author, 24 March 1998, personal communication. Rose was a pilot
assigned to 92nd BG (H). The ceiling data is from Strategic Bombing Survey, Weather Fac-
tors in Combat Bombardment Operations in the European Theater, 2nd ed. ( January 1947):
9A; J. W. Howland to author, 23 March 1998, personal communication; see also Rose to
author, 24 March 1998. Apparently the Gee method also only worked at certain airfields.
Howland served with 381st BG (H) and 91st BG (H).
12. Col. Daniel C. Doubleday to Chief Signal Oªcer, 8 November 1943; European The-
ater of Operations (Eisenhower) to Wright Field, telegram dated 8 February 1944; War (Sec-
retary of War) to ASCPFO, 13 March 1944, all in file: Project 102 (UK), box 404, RD 158,
Records of Wright Field (Sarah Clarke Collection), RG 342, NA.
13. Frank Griªths, Angel Visits: From Biplane to Jet (London: Thomas Harnsworth,
1986): 101–103. Arnold to McClellan (signed by Spaatz), telegram, 14 March 1944, file:
Project 102 (UK), box 404, RD 158; Carl Spaatz to [recipient unreadable, but probably Gen-
eral McClelland], 13 May 1944, file: Project 102D, box 405, RD 159; 1st Lt. Donald Hansen,
29 January 1945, file: Project 102 (UK), box 404, RD 158, all in Records of Wright Field
(Sarah Clarke Collection), RG 342, NA.
14. Wilbert Swank, 8 May 1944, file: Project 102; 1st Lt. Donald Hansen, 29 January
1945, file: Project 102 (UK), both in box 404, RD 158, Records of Wright Field (Sarah Clarke
Collection), RG 342, NA. This author noted that SBA’s failure continued to cause hesita-
tion to use SCS-51.
15. HQ USSTAF, London, to ATSCWFO, telegram dated 12 September 1944, file: Proj-
ect 102 (UK), box 404, RD 158, Records of Wright Field (Sarah Clarke Collection), RG 342,
NA.
16. See Director of Bombardment to Director of Communications, 8 January 1943, box
1107, AAF General Correspondence, RG 18, NA.
17. Hansen, 29 January 1945 (see n. 14). See also Robert B. Parke, The Pilot Maker (New
York: Grosset and Dunlap, 1970), on the Link Trainer.
18. Only twenty-five were installed. The remaining five were dispatched to the conti-
nent after the Normandy invasion, where they were used to mark the front lines so that pi-
lots would know when they were over friendly territory. This was done to reduce the prob-
202 notes to pages 116 –124
ability of aircraft mistakenly attacking friendly ground forces. With the localizer aligned
parallel to the front, the 90 Hz signal zone marked friendly territory, and the 150 Hz zone
marked enemy territory. Telephone conversation with Timothy Bland, 23 March 1998.
19. Preston R. Basset to T. A. Morgan, 18 July 1941, file: Instrument Landing, box 35,
Sperry Gyroscope Collection, Hagley Museum and Library.
20. Hugh G. J. Aitken, Continuous Wave: Technology and American Radio, 1900 –1932
(Princeton: Princeton University Press, 1985), 1–27. The term “intellectual tradition” is my
own formulation of his much more complex argument. See also his earlier work on ra-
diotelegraphy: Hugh G. J. Aitken, Syntony and Spark: The Origins of Radio (Princeton:
Princeton University Press, 1976).
21. More detail on the Lab’s founding can be found in Chapter 6.
22. A. L. Loomis, “Report of the Ad Hoc Committee on Instrument Landing,” 16 Feb-
ruary 1942, box 53A, Records of the MIT Radiation Lab, RG 227, NA.
23. J. H. Buck to Lee DuBridge, 2 December 1942, file: PGP [1941–1943], box 54A, MIT
Radiation Laboratory Records, RG 227, NA; for comments suggesting this interpretation,
see Buck to DuBridge; Robert Davies to J. H. Buck, 13 April 1943, ibid.; Loomis, “Report”;
MIT Radiation Laboratory Records, RG 227, NA.
24. Buck to DuBridge, 2 December 1942.
25. Ibid.; Henry Guerlac, History of Radar in World War II (New York: Tomash Pub-
lishers, 1987).
26. The best work on ICAO’s foundation is Alan Dobson, Peaceful Air Warfare: The
United States, Britain, and the Politics of International Aviation (Cambridge: Oxford Univer-
sity Press, 1991).
27. On Watson-Watt, see his autobiography: Robert Watson-Watt, Three Steps to Victory
(London: Odhams, 1957); on Stanton, see Wilson, Turbulence Aloft, 137–40.
28. Robert Watson-Watt, “Record of an Informal Fourth CERCA meeting held on Sun-
day, 10 November 1946,” COT/15, PICAO documents, International Civil Aviation Orga-
nization (ICAO) Library and Archives, Montreal, Canada. CERCA stood for “Common-
wealth and Empire Radio Committee Assembly.”
29. Special Radio Technical Division, “Minutes of the Tenth Meeting,” 11 November
1946, COT/21, PICAO documents, ICAO. Stanton’s criticism of Watson-Watt is in COT/22,
same date.
30. MATS had stripped a number of lesser-used airways in the United States of their
ranges during the war and shipped them overseas, where they were installed to facilitate
cargo and aircraft movement. See Wilson, Turbulence Aloft, 119.
31. Special Radio Technical Division, “Minutes of First Meeting,” 30 October 1946, PI-
CAO documents, vol. 23, ICAO Library and Archives, Montreal.
32. Edward P. Warner, “Work of the Interim Council of the PICAO,” Aero Digest (March
1946): 24–25, 148.
chapter 6: the intrusion of newcomers
1. Robert Buderi, The Invention That Changed the World: How a Small Group of Radar
Pioneers Won the Second World War and Launched a Technological Revolution (New York: Si-
mon & Schuster, 1996), 38–51; James Phinney Baxter, Scientists against Time (Cambridge,
MA: MIT Press, 1946), 3–25.
2. Buderi, Invention That Changed the World, 38–51.
3. J. L. Heilbron and Robert W. Seidel, Lawrence and His Laboratory: A History of the
Lawrence Berkeley Laboratory, Vol. 1 (Berkeley: University of California Press, 1989), 493.
4. Ibid.; Buderi, Invention That Changed the World, 46–49; Luis W. Alvarez, Adventures
of a Physicist (New York: Basic Books, 1987), 88; Lawrence Johnston, interview with Fred-
erik Nebeker, 13 June 1991, transcript, MIT Radiation Laboratory Oral History Collection,
IEEE Center for Electrical History, Rutgers, NJ.
5. Alvarez, Adventures, 32–49; Heilbron and Seidel, Lawrence and His Laboratory, 236–37.
6. Alvarez, Adventures, 31.
7. Johnston interview; Buderi, Invention That Changed the World, 43.
8. The name was originally Ground Controlled Landing (GCL), but the Army insisted
that blind landings were impossible, drawing Alvarez’s scorn. The name was changed
nonetheless, and of course the Army was proven correct. Although GCA could land air-
craft completely blind sometimes, it could not do so with perfect reliability. That has been
true of every blind landing system. At least two other approach systems based on the
ground control model were invented during the war, both by technicians in combat the-
aters before Alvarez’s GCA was deployed. In both cases, radars designed for airborne use
(H2S/APS-15 in one case, and SCR-717 in the other) were modified and set up at an airfield
to scan the runway approach sector. An operator then gave pilots landing directions via
voice radio. George Reynolds to author, 23 March 1998, personal communication; “Baby
GCA [1944–1945],” box 35B, MIT Radiation Laboratory Records, RG 227, NA.
9. The LSE still has a job, and modern carrier pilots also have a stabilized visual refer-
ence system on the ship’s deck to assist in the approach. Certain aircraft can utilize a fully
automatic radar-based “carrier controlled approach” system developed between the mid-
1950s and the late 1960s.
10. A. L. Loomis, “Report of the Ad Hoc Committee on Instrument Landing,” 16 Feb-
ruary 1942, file: Report of Ad Hoc Committee on Instrument Landing, box 53A, Records
of the MIT Radiation Lab, RG 227, NA.
11. Alvarez, Adventures, 86–110, and Buderi, Invention That Changed the World, 137–38.
12. Lawrence Johnston, “GCA: Ground Controlled Approach, Radiation Laboratory Re-
port 438” (Cambridge: MIT Radiation Laboratory, 1 October 1943, photocopy of copy #87),
file: RB 334.8 NDRC RL Report ET-2047, box 1428, Oªce of the Chief Signal Oªcer Cen-
tral Files, RG 111, NA.
13. Alvarez, Adventures, 98. See also Jennet Conant, Tuxedo Park: A Wall Street Tycoon
and the Secret Palace of Science That Changed the Course of World War II (New York: Simon
& Schuster, 2002), 260.
14. Homer Tasker, “Technical Report of Radio Set AN/MPN-1 (XE-1),” Report no. 103,
15 June 1945 (Los Angeles: Gilfillan Bros., Inc., photocopy) file: GCA[GCL] GCA-Mk II
[AN/MPN-1] GCA Mk III [1941–1945], box 53B, MIT Radiation Laboratory Records, RG
227, NA.
15. Alvarez, Adventures, 99–100; the military designation for the Gilfillan sets was
AN/MPN-1(A).
204 notes to pages 130– 137
16. Luis Alvarez, “Minutes of the Coordinating Committee,” 20 January 1943, file:
“Group 73, Landing [1942–1943],” box 41B, MIT Radiation Laboratory Records, RG 227,
NA; Alvarez, Adventures, 100; Johnston, interview.
17. Johnston, interview. Date from Johnston, “GCA,” 46 (see n. 12); Commander Fleet
Air, Quonset Point, to Chief of the Bureau of Aeronautics, 26 April 1943, file: F42-9, box
892, Bureau of Aeronautics Records, RG 72, NA.
18. Tasker, “Technical Report,” 8–9. The Mk I required 4,400 drawings totaling 12,800
square feet.
19. A fascinating incident happened during this test series when two RAF Mosquito
bombers, one painted silver, the other black, at the same altitude, gave two very di¤erent
readings. The black one disappeared from the scope at a distance of ten miles while broad-
side to the antenna; the silver one stayed visible until twenty miles out. The AAF took con-
siderable notice: Lt. Lawrence McFadden to Chief Signal Oªcer, 25 August 1943, file: GCA
Correspondence 1943, box 53, MIT Radiation Laboratory Records, RG 227, NA; the infor-
mation in this paragraph has been synthesized from several sources: Johnston interview;
Alvarez, Adventures; Clarke, Glide Path; George Comstock, transcript of interview with
Henry Guerlac, 7 December 1944, in file: GCA/GCL Mk II [AN/MPN-1] Mk III [1941–
1945], box 53B, MIT Radiation Laboratory Records, RG 227, NA. Although Clarke’s work
is a novel, enough of the events it reports match with those reported in the other sources
to consider it to be a largely true account, even if it di¤ers in its chronology.
20. Anonymous memo to the Coordinating Committee, 22 May 1945, file: GCA Mem-
oranda, box 65A, MIT Radiation Laboratory Records, RG 227, NA.
21. Saipan and Iwo Jima. On Iwo Jima, P-61s with experienced pilots could be brought
in under ceilings down to 25 feet, but the base implemented a policy of ordering crews to
bail out over the island (directed by the GCA operator) for ceilings of under 100 feet be-
cause pilots who could not see below that point, even if landed safely, tended to collide with
ground obstacles such as parked aircraft. Anonymous, undated document [c. June 1945],
file: Baby GCA, box 35A, MIT Radiation Laboratory Records, RG 227, NA.
22. Frank Griªths, Angel Visits: From Biplane to Jet (London: Thomas Harnsworth Pub-
lishing, 1986), 99.
chapter 7: the politics of blind landing
1. AOPA Pilot was a members-only insert to Flying magazine, provided by AOPA’s sta¤.
Because Flying had no editorial control over the insert, I have documented it as an inde-
pendent publication. It is essentially a newsletter, and library editions of Flying did not con-
tain it. It is diªcult for scholars to access as a result, but the Hagley Museum and Library
contains a complete set. AOPA Pilot became an independent publication in 1958.
2. Quote from Charles V. Murphy,”The Last 500 Feet,” Life (12 May 1947): 93; Minutes
of the Robert J. Collier Trophy Committee, 8 August 1946, Collier Trophy Collection, Na-
tional Air and Space Museum Archives, Washington, DC; report prepared by Carl Hin-
shaw, Aids to Air Navigation and Landing, House Committee on Interstate and Foreign
Commerce, Subcommittee on Air Safety, 80th Cong., 1st sess., 11 July 1947, 17–18. This
chapter appeared in a shortened form as Erik M. Conway, “The Politics of Blind Landing,”
Technology and Culture 42, no. 1 (January 2001): 81–106.
3. For clarity’s sake, I will refer to both SCS-51 and ILS as ILS and both AN/MPN-1 and
GCA as GCA. Although GCA is less a piece of equipment than a procedure, the distinc-
tion was not made at the time, and all references to GCA in the literature are to Alvarez’s
invention.
4. John M. R. Wilson, Turbulence Aloft: The Civil Aeronautics Administration amid Wars
and Rumors of Wars, 1938–1953 (Washington, DC: FAA, 1979), 164 –68; Joseph Corn, The
Winged Gospel (New York: Oxford University Press, 1983).
5. A parallel Senate investigation also took place. The House record is more detailed
and complete, and I have therefore relied on it.
6. John Stuart, “Aviation Accidents: Ground Control Approach System, Backed by
Many, Runs into Cost Objections,” New York Times (2 February 1947): 14.
7. AAF Instrument Flying Standardization Board, “Recommendations of the Sta¤
Study of the Relative Merits of SCS-51 and GCA,” 26 May 1945, 202.2-54, attachment 225,
Air Force Historical Research Agency (AFHRA), Maxwell AFB, AL; Air Transport Com-
mand, “Report on the Pilot and Instrument Approach System Evaluation Tests Conducted
at Indianapolis,” 11 July 1946, K239.0249-4609AU, AFHRA. The Air University’s evalu-
ation of the Indianapolis tests was requested by Gen. Carl Spaatz by memo dated 1 August
1946 and was forwarded to AAF Headquarters on 18 March 1947. Both are attachments to
K239.0249-4609AU, AFHRA.
8. And, as if in a mirror, the pilots’ union demanded radar control of the airways (but
not radar-directed landing) during the 1950s, after a series of spectacular midair collisions
demonstrated that the skies had become too crowded to permit the absolute freedom that
pilots had traditionally enjoyed. They also fought to ensure that ground controllers’ au-
thority was carefully circumscribed, however. See Stuart I. Rochester, Takeo¤ at Mid-
Century: Federal Civil Aviation Policy in the Eisenhower Years, 1953 –1961 (Washington, DC,
Federal Aviation Authority, 1976), 57–78.
9. Navy pilots retained the right to ignore flight deck orders because the person issu-
ing them was an enlisted rating, and by law enlisted ranks cannot give orders to oªcers—
and almost all Navy pilots were oªcers. Hence, legally, the Navy had to allow pilots the right
to ignore directions from the LSE, as the position is known. Further, the Navy’s tradition
of independence of command applied to aircraft as well as ships. The Navy therefore ex-
pected its pilots to accept directions from the flight deck, but it could not force them to. Navy
pilots thus retained legal autonomy while still typically submitting to deck-based directions.
10. Commerce Member, IATCB to Director of Federal Airways, 10 February 1945, file:
504.0 vol. 6, box 83, CAA Central Files, RG 237, NA.
11. Chief, Ultra-High Frequency Unit to Acting Chief, Radio Engineering Section, 22
May 1945, ibid.
12. “Washington Observer—CAA Touchy on GCA,” Aviation News (3 March 1947): 3.
13. Chief, Air Traªc Control Division, to Director of Federal Airways, 17 February 1945,
file 504.0 vol. 6, box 83, CAA Central Files, RG 237, NA.
14. Ibid.; Thomas Bourne to Brig. Gen. H. M. McClelland, 14 February 1945, file 818.1
vol. 1, box 285, CAA Central Files, RG 237, NA. Unfortunately, Kroger’s version is the one
that CAA’s oªcial historian, Wilson, relied on. See William Kroger, “CAA and Airlines
against GCA,” Air Transport (March 1947): 30–33, 77–78; and Wilson, Turbulence Aloft, 323.
I used the FAA History Oªce’s annotated copy of Kroger’s article in this analysis.
15. Wilson, Turbulence Aloft, 217; Ben Stern to Maurice Roddy, 12 November 1946, file
818.1: ASR vol. 1, box 825, CAA Central Files, RG 237, NA.
16. Gilbert later became one of the leading authorities on Air Traªc Control matters.
See, for example, Glenn A. Gilbert, Air Traªc Control: The Uncrowded Sky (Washington,
DC: Smithsonian Institute Press, 1973); Nick Komons, “Federal Government Helped Forge
New Enterprise, New Profession,” Aviation’s Indispensable Partner Turns 50, Department of
Transportation, n.d.
17. Chief, Air Traªc Control Division, to Director of Federal Airways, 17 February 1945,
file 504.0 vol. 6, box 83, CAA Central Files, RG 237, NA. In combat theaters SCS-51s and
runways were utilized at landing rates of up to one plane every thirty seconds. AN/MPN-
1s were occasionally used at equivalent rates, but that seems to have been accomplished by
landing several aircraft simultaneously, in formation. That was obviously only possible with
small, fighter-sized aircraft. Such rates were clearly too risky for commercial aviation.
18. Ibid.
19. Asst. Chief, Airways Engineering Division, to Director of Federal Airways, 27 Feb-
ruary 1945; Chief, Air Traªc Control Division, to Director of Federal Airways, 17 February
1945, both in file 504.0 vol. 6, box 84, CAA Central Files, RG 237, NA.
20. William Kroger, “Standardized System Expected in Instrument Landing Dispute,”
Aviation News (10 December 1945): 9–10; Robert B. Hotz, “Two Congressional Probes
tackle Air Safety Problem,” Aviation News (27 January 1947): 7–8; “TWA Ready to Use
GCA-ILS,” Aviation News (27 January 1947): 28; Austin Stevens, “Gander Radar Test Guides
Planes In,” New York Times (10 January 1947): 7.
21. T. P. Wright to James C. Johnson, 19 November 1945, file 818.1: ASR vol. 1, box 285,
CAA Central Files, RG 237, NA.
22. CAA’s personnel are estimated in Committee on Interstate and Foreign Com-
merce, Safety in Air Navigation, 80th Cong., 1st sess., January 1947, 244. The cut was not
CAA specific. Congress had passed the FY 1946 budget and then passed the Ramspeck
Act, raising the pay of federal civil service employees 14 percent percent. It then refused to
go back and fix the payroll budgets, leading to widespread furloughs in all federal civilian
agencies. See ibid., 237; and Wilson, Turbulence Aloft, chap. 6.
23. Assistant Administrator for Federal Airways to Administrator, 21 February 1947,
file 818.1: ASR vol. 2, box 286, CAA Central Files, RG 237, NA. Unit maintenance files in
RD 1276–1280, Records of Wright Field (Sarah Clarke Collection), RG 342, NA.
24. John Stuart, “Aviation Accidents: Ground Control Approach System, Backed by
Many, Runs into Cost Objections,” New York Times (2 February 1947): B-14. De Florez had
submitted Alvarez’s name to the Collier committee, and, as a member, had spoken in sup-
port of his nomination.
25. AOPA Pilot (November 1945): 42b–42c.
26. Brig. Gen. Alden Crawford to Commanding General, Air Material Command, 24
July 1946, file 413.44: AN/MPN-1, box 1870, AAF Unclassified Central Decimal Files, RG
18, NA.
27. The three sets used highly modified experimental displays that required only two
operators, one for the search set and one for the two precision sets.
28. Alonzo Hamby, Man of the People: A Life of Harry S. Truman (Cambridge: Oxford
University Press, 1995): 218, 248–60.
29. Anthony Leviero, “Congress Aroused over Air Crashes,” New York Times (19 Janu-
ary 1947): B-16.
30. Ibid.; “Pressure Builds Up for GCA,” Aviation News (6 January 1945): 3.
31. C. B. Allen, “Radar-’ILS’ Rift Studied by CAA Chief,” New York Herald Tribune (24
November 1946): 45; William Kroger, “Collier Trophy Award to Alvarez Spotlights GCA
Development,” Aviation News (16 December 1946): 7–8.
32. Committee on Interstate and Foreign Commerce, Safety in Air Navigation, 572–73
(see n. 22).
33. Allen, “Radar-’ILS’ Rift,” 45; Kroger, “Collier Trophy,” 8.
34. T. P. Wright to Bob Sibley (Aviation Editor, Boston Traveler), 24 January 1947, file
818.1: ASR vol. 2, box 286, CAA Subject File, RG 237, NA. This letter is based on one pro-
vided to Wright by the Asst. Administrator for Federal Airways, 10 January 1947, file 504.0:
vol. 11, box 84, ibid. Stanton held the oªce at that time, and the memo was actually writ-
ten by C. M. Lample, Director of the Air Navigation Facilities Service.
35. Wright to Sibley, 24 January 1947. Emphasis in original.
36. Director, Air Navigation Facilities Service to Asst. Administrator for Federal Air-
ways, 10 January 1947, file 504.0, vol. 11, box 286, CAA Subject File, RG 237, NA.
37. “AOPA Cites Needs for GCA Units,” AOPA Pilot (January 1946): 42a.
38. Committee on Interstate and Foreign Commerce, Safety in Air Navigation, 1431,
1436, 1434.
39. Ibid., 1431.
40. Ibid., 1451.
41. Ibid., 570–71. He used “route miles” for his argument.
42. George E. Hopkins, The Airline Pilots: A Study in Elite Unionization (Cambridge,
MA: Harvard University Press, 1971); George E. Hopkins, Flying the Line: The First Half Cen-
tury of the Air Line Pilots Association (Washington, DC: ALPA, 1982). Behncke was a United
Airlines pilot until his union presidency became a full-time job.
43. David Behncke, “Technically Speaking,” The Air Line Pilot (March 1946): 5.
44. David Behncke, “Airlines Pilots Stake in Safety,” Air Transport (June 1947): 27–30.
45. ALPA does not have locals. Instead, its regional councils have members from each
represented airline that operates in that region; David Behncke, “Pilots Prefer ILS by 46–
1 Margin,” The Air Line Pilot (November 1947): 3, 5.
46. Nick A. Komons, The Third Man: A History of the Airline Crew Complement Contro-
versy, 1947–1981 (Washington, DC: Federal Aviation Administration, 1987); “New Radar
Tests by Army Air Force,” New York Times (15 February 1947): 8.
47. Committee on Interstate and Foreign Commerce, Safety in Air Navigation, 351.
48. Ibid., 489, 804.
49. Ibid., 805.

50. Ibid., 21; “Text of Air Safety Report,” New York Times (3 July 1947): 8; House Commit-
tee on Interstate and Foreign Commerce, Subcommittee on Air Safety, report prepared by Carl
Hinshaw, Aids to Air Navigation and Landing, 80th Cong., 1st sess., 11 July 1947, 17–18.
chapter 8: transformations
1. “Text of Air Safety Report, New York Times (3 July 1947): 8.

2. Ibid.
3. U.S. Strategic Bombing Survey, Military Analysis Division, Weather Factors in Com-
bat Bombardment Operations in the European Theater, 2nd ed. (January 1947), 9A. The di¤er-
ence in the two organizations’ minimum ceiling requirements was due to Bomber Com-
mand’s policy of night bombing, which demanded a higher ceiling in the absence of an
e¤ective low approach aid. The Eighth Air Force bombed, and landed, by day and hence
could a¤ord a lower ceiling within the same technological constraints.
4. David B. Langmuir to Lee DuBridge, 14 January 1943, file: Project 102, box 404, RD
158, Records of Wright Field (Sarah Clarke Collection), RG 342, NA.
5. Slightly more than a hundred collisions between bombers occurred in the thousand
or so flying days the Eighth Bomber Command existed. About half occurred over England.
See Roger A. Freeman, with Alan Crouchman and Vic Maslen, Mighty Eighth War Diary
(New York: Jane’s, 1981), 19.
6. George Comstock, 2 May 1945, file 1022 GCA, box 19, MIT Radiation Laboratory
Records, RG 227, NA. See various correspondence in file: GCA correspondence [1945],
ibid.; Acting Technical Assistant to Director, ANF Operations Service, 17 June 1946, both
in file 818.1: vol. 1, box 285, CAA Subject File, RG 237, NA.
7. Glen Gilbert to Chief, Technical Development Division, 14 December 1945, file 818.1:
ASR vol. 1, box 285, CAA Central Files, RG 237, NA.
8. Chief, Air Traªc Control Division to Chief, Technical Development Division, 14 Sep-
tember 1945; Air Navigation Facility Operation Service, “Operational Tests in the Use of
Ground Radar Aids to Instrument Landing and Air Traªc Control,” 1 June 1945, both in
file 818.1: ASR vol. 1, box 285, CAA Central Files, RG 237, NA.
9. T. P. Wright to Mr. [Donald M.] Stuart, 30 November 1945; J. H. Miles to Donald M.
Stuart, 30 April 1946, both in ibid.
10. Acting Technical Assistant to Director Air Navigation Facilities Operations Service,
17 June 1946; T. P. Wright to Asst. Secretary of Commerce, 26 December 1946, both in
ibid.
11. Committee on Interstate and Foreign Commerce, Safety in Air Navigation, 80th
Cong., 1st sess., January 1947, 768.
12. Edwin E Aldrin, “Fair to Foul Weather Flying,” Aeronautical Engineering Review 5,
no. 7 (1946): 26–29; “Civilian Application of Radar Moves Forward on Two Fronts,” Avia-
tion News (11 February 1946): 12; Deputy and Assistant Chief of Bureau of Aeronautics to
Assistant Secretary of the Navy for Air, 24 January 1947, file: Equipment, Materials, Sup-
plies, 1947, box 19, Bureau of Aeronautics Arcata Landing Aids Experimental Station
Records, RG 72, NA.
13. The one-a-minute rate was a CAA estimate. The AAF believed that only one aircraft
every three minutes could be accommodated, due to deflection of the ILS beams by the
leading aircraft.
14. Arthur C. Clarke, Glide Path (New York: Harcourt, Brace & World, 1963), chaps. 23,
24. Milton Arnold claimed that there had been four FIDO installations in England by the
end of the war, and that FIDO did not stand for anything—it was just a code word; “FIDO
for Los Angeles,” Aero Digest (April 1949): 60.
15. The average figure for Los Angeles was 2.5 percent, using a 200-foot ceiling. I have
found no data presenting an average for all U.S. airports, or even the ten busiest. For com-
parison, La Guardia’s average was 1.7 percent, again using a 200-foot ceiling. From Edgar
A. Post, “Airline Operating Experiences with the Instrument Landing System,” Institute of
Aeronautical Sciences, 22 July 1949.
16. David L. Behncke, “Invitation to Progress,” The Air Line Pilot (February 1946): 2.
See also “Technically Speaking,” The Air Line Pilot (March 1946): 5.
17. John M. R. Wilson, Turbulence Aloft: The Civil Aeronautics Administration amid Wars
and Rumors of Wars, 1938–1953 (Washington, DC: FAA, 1979), 237–41.
18. CAA coined these terms to eliminate confusion over what “GCA” meant. “GCA”
was used in the media to variously represent all radar, the actual AN/MPN-1 equipment,
all landing procedures controlled from the ground regardless of equipment, and just the
short-range precision radars. By splitting the AN/MPN-1’s three radars into two categories,
ASR and PAR, CAA hoped to clarify its policies and improve aviation writers’ under-
standing of its policies.
19. E. M. Sturhahn to Administrator [Wright], 30 January 1947, file 818.1: ASR vol. 2,
box 286, CAA Central Files, RG 237, NA.
20. Wilson, Turbulence Aloft, 225–28.
21. Ibid.
22. Com. Paul Goldsborough, RTCA Executive Committee Meeting minutes, 6 April
1944, binder 1944-1945-1946, courtesy RTCA, Washington, DC. ARINC had been estab-
lished by the airlines during the early 1930s to produce aircraft radios, since the airlines
felt they were not being well-served by existing radio companies. Rentzel replaced Wright
as Administrator of Civil Aeronautics in 1948. L. M. Sherer, RTCA Executive Committee
Meeting minutes, 7 December 1945, binder 1944-1945-1946, courtesy RTCA.
23. RTCA, “Recommended United States Policy,” Air Navigation-Communication-
Traªc Control, 28 August 1946; Wilson, Turbulence Aloft, 233; The “Buck Rogers future”
comment is from Ben Stern to the Editor, Business Week (2 October 1945), file 818.1: ASR
vol. 1, box 285, CAA Central Files, RG 237, NA.
24. Walter McDougall, . . . the Heavens and the Earth: A Political History of the Space Age
(New York: Basic Books, 1985), 229–31; Federal Aviation Agency, FAA Statistical Handbook
of Aviation (Washington, DC, 1959), 22.
25. Erik M. Conway, “Echoes in the Grand Canyon: Public Catastrophes and Tech-
nologies of Control in American Aviation,” History and Technology 20, no. 2 ( June 2004):
115 –34.
conclusion
1. This sketch is based on the opening statement of Gregory A. Feith, the Investigator-
in-Charge, to the public hearing on the accident held March 24, 1998. See also the final in-
vestigation report: National Transportation Safety Board, “Aircraft Accident Report: Con-
trolled Flight into Terrain, Korean Air Flight 801, 6 August 1997,” NTSB/AAR 0001, 13
January 2000, ix.
2. Karl F. Kettler to the editor, Aviation Week and Space Technology (May 4, 1998): 6.
3. Roger G. Miller, To Save a City: The Berlin Airlift, 1948–1949 (Washington, DC: Air
Force History and Museums Program, 1998).
4. Langdon Winner, Autonomous Technology: Technics-out-of-Control as a Theme in Polit-
ical Thought (Cambridge: MIT Press, 1977), 326. Emphasis added.
5. Brian Balogh, Chain Reaction: Expert Debate and Public Participation in American
Commercial Nuclear Power, 1945–1975 (New York: Cambridge University Press, 1991.)
6. Susanne K. Schmidt and Raymond Werle, Coordinating Technology: Studies in the In-
ternational Standardization of Telecommunications (Cambridge, MA: MIT Press, 1998), 20 –
21.
7. See Richard Hallion, Legacy of Flight (Seattle: University of Washington Press, 1976).
8. On SAGE, see Kent C. Redmond and Thomas M. Smith, From Whirlwind to MITRE:
The R&D Story of the SAGE Air Defense Computer (Cambridge, MA: MIT Press, 2000). For
Project Beacon’s interest in SAGE as an air traªc control system, see Richard J. Kent, Safe,
Separated, and Soaring: A History of Federal Civil Aviation Policy, 1961–1972 (Washington,
DC: FAA, 1980), 29–34; and Erik M. Conway, “Echoes in the Grand Canyon: Public Catas-
trophes and Technologies of Control in American Aviation,” History and Technology 20,
no. 2 ( June 2004): 115–34.
9. Twelve O’Clock High, 20th Century Fox, 1949. Based on a book by Lay Beirne, Twelve
O’Clock High (New York: Harper, 1948).
10. See David DeVorkin, Science with a Vengeance (New York: Springer-Verlag, 1992), 3.
11. Dale O. Smith, Screaming Eagle: Memoirs of a B-17 Group Commander (Chapel Hill:
Algonquin Books, 1990), 64.
12. Frank Griªths, Angel Visits: From Biplane to Jet (London: Thomas Harnsworth Pub-
lishing, 1986), 98–115.
Index
A-B-C method, 25 Aircraft Radio Laboratory, 68, 83, 87, 91, 96,
A-I system, 70–74, 77–79, 84, 87, 97, 107 100
acoustic altimeters, 63 airfields: approach procedures for, 109, 110;
acoustic height finders, 62 drainage systems for, 42–44; grass, 36, 37,
Aeronautical Radio, Inc. (ARINC), 84, 86, 173 45; historical background of, 36–40; leader
Aeronautics Branch (Department of Com- cable system and, 45–55; overview of, 35–36;
merce): function of, 57–58; radio blind flying surfaces for, 38, 40–42
and, 63–64, 66, 68, 69 airlines: accident rate among, 151–152; early
Air Commerce Act, 8 training programs by, 27–28; mail contracts
Air Coordinating Committee (ACC), 159, 164, with, 5, 31–32; negotiations between CAA,
173 Army Air Corps and, 180–181
The Air Line Pilot, 157 airmail crisis of 1934, 13, 67, 71
Air Line Pilots Association (ALPA), 74, 85–86, airmail service: economics of, 18; historical
138, 139, 141, 152, 157–160, 170, 180 background of, 1–2, 16–19; weather-related
Air Mail Act of 1925, 31 accidents in, 19–20
Air Mail Service (U.S. Post Oªce), 1–2 airport surveillance radars (ASR), 172, 174
Air Navigation Development Board, 173, 175 airports: construction of, 42–45; design of, 39;
Air Policy Commission, 173 location of, 39–40
Air-Track Corporation, 67, 75, 105 Aitken, Hugh, 117
Air-Track system (Washington Institute of Alcock, John, 16
Technology), 67–68, 82 Alexanderson, E. F. W., 62, 63
Air Traªc Control Division (Civil Aeronautics All Weather Flying Program (Army Air Forces),
Authority), 144, 149, 166 168
air traªc control radar, 165–167, 172 all-weather landing system (AWLS), 184
air traªc system, 10 altimeters, 23, 48, 62–63
Air Transport, 157 Alvarez, Luis: background of, 29, 123, 125–126;
Air Transport Association (ATA), 94, 139, 145, blind landing research and, 126–132, 135,
147, 152, 157–160, 167, 170, 173–174 145; Collier Trophy and, 137; search radar
Air University, 141 and, 165
aircraft control, 12–13 American Air Line, 31
aircraft design: airfield considerations and, 35, American Airlines, 84
37–39; glide path length and, 97 American Loth Company, 47, 50–52
Aircraft Owners and Pilots Association (AOPA), Amtorg Trading Company, 33
138–141, 145, 155–159, 172, 179–180 AN/MPN-I, 137, 144, 147, 148, 153, 166, 167,
Aircraft Production Board, 138 175
212 index
AOPA Pilot, 137, 139, 148, 155, 160 Bell Labs, 23, 86
Arcata project, 168–171, 174 Bender, Luis, 91
Army Air Corps: academic organizations and, Bendix Aviation, 74
182; glide paths and, 85, 97–98; instrument Bendix Radio, 75–77, 84, 86, 105, 131, 133
training by, 29; landing system requirements Berlin airlift of 1948, 177–178
and, 73–74; mail flights by, 32–33; micro- Black, Hugo, 32
wave research and, 81, 86, 93, 103; negotia- black box (Ocker), 21, 28
tions between CAA, airlines and, 180–181; Black Friday, 177–178
radio direction finders and, 70, 80 blind approach beacon system (BABS), 126
Army Air Forces (AAF): air traªc control radar blind flight trainer: Boeing School of Aeronau-
and, 165–167; automatic GCA and, 158; con- tics, 24; Link’s, 30–31, 33
tacts between Royal Air Force and, 106; glide blind flying: on airmail routes, 16, 32–33; Army
path projects of, 104, 107; ILS/GCA contro- Air Corps and, 32–33; instrument depen-
versy and, 141, 142, 153, 159–162, 171; Instru- dence and, 33–34; interwar-period research
ment Instructor School, 144; landing aids de- on, 16–17; Ocker-Meyer method of, 21–22;
cisions of, 137; landing training and, 114; radio, 59–79; use of term, 4
policies of, 183; SCS-51 and, 110–115; VHF blind landing: altitude and, 12; evolution of
glide path and, 122 concept of, 184–185; failure of, 186; govern-
Army Air Service: airfield design and, 36; early ment organizations involved in research on,
airmail flights by, 2, 16–17; turn indicator 57–58; requirements for, 34; use of term, 4;
and, 15 varying aspects of, 12; World War I and, 45,
Army Signal Corps: blind landing research 182; World War II and, 165
and, 77, 82, 99; World War II developments blind landing systems: based on three beams of
and, 107, 113, 118; x-rays and, 2 energy, 89; Bureau of Air Commerce and,
Arnold, Henry “Hap,” 8, 80–81, 92, 93, 99– 82; components of, 123; failure of quest for,
102, 106, 118,122, 130, 183 9–10; historical background of, 1–3, 5, 23;
Arnold, Milton, 145, 148, 152, 159, 167 ILS/GCA controversy and, 138–162, 171;
artificial horizon, 21–25, 27 leader cable system and, 45–55; microwave-
automatic direction finders, 70 based, 89–103; profit motive for, 4–5;
aviation: academic organizations and, 182; ac- progress motive for, 6–9; recommendations
cess to skies and, 138; environmental issues for, 117–118
related to, 6–8, 55; interest groups related to, Boeing, William, 7
180; links between commercial and military, Boeing Aircraft Company, 7
182; notion of progress in, 6–8; religious Boeing Air Transport, 7, 31
symbolism and, 6–7 The Boeing Company, 7
Aviation News, 137, 139, 143, 145, 151–153, 160 Boggs, Marshall, 64, 66, 67
aviation technology: pre–World War I, 13; World Bolling Field, 43, 44
War II, 10, 11, 75, 104, 105, 107, 113, 114, 118, bombers: blind flying and, 75, 90; GCA and,
120, 122. See also specific types of technology 135; World War II, 75, 164
Boonstra, Henry, 19
balloon system, 2, 48–49 Borum, F. S., 82–83
Balogh, Brian, 180 Bourne, Thomas, 143, 144–146
bank indicator, 15, 26 Bowles, Edward, 89–91, 95–97, 100, 104, 107,
banking, 14–15 115, 122, 182
barometric altimeters, 62 Brockett, Paul, 93
Barrow, Wilmer L., 89 Brown, Walter Folger, 31
Bassett, Preston, 87 Brown, William G., 23
Behncke, David, 139, 152, 157–158, 160 Buck, J. S., 118
index 213
Buckley, Oliver, 93 Compton, Karl, 91, 124, 125

Burden, William, 152 Comstock, George, 125, 127, 132, 135, 165
Bureau of Air Commerce: air traªc control sys- Congress, U.S.: aviation policy and, 159–163,
tem and, 136; blind landing development 170; Civil Aeronautics Authority and, 146,
and, 29, 67, 70, 71, 74, 76; commercial air- 148, 150–152, 155–156, 160, 163; ILS/GCA
lines and, 181; Radio Technical Committee controversy and, 141, 158; landing aids con-
on Aeronautics and, 81–83, 86 troversy and, 171–174; 1946 elections and,
Bureau of Lighthouses, 57–58 146, 148, 150–152
Bush, Vannevar, 93–101, 107, 117, 124–125, 127 Congressional Air Policy Committee, 173
constant-intensity glide path, 64, 65
CAA. See Civil Aeronautics Authority (CAA) contact flying: explanation of, 13–14; instru-
California Institute of Technology (CalTech), 23 ment flying vs., 28
Carl Lorenz A. G., 76, 82 continuous wave system, 104–107, 117–119
center-line lighting, 170–171 Cooch, H., 47, 48
Chaplin, Sidney, 29 Corn, Joseph, 6–8, 138, 179
Charlton, John, 19 CPN-18 project, 165–167, 171
Chicago Tribune, 152 Craig, P. H., 69
Christensen, J. Titus, 19 Crane, Carl, 13, 20–22, 28
Churchill, Winston, 122 Cronon, William, 9
Civil Aeronautics Act, 151 cross-pointer indicator, 63, 65, 98
Civil Aeronautics Authority (CAA): Air Traªc Cunningham, J. R., 73, 74, 87
Control Division, 166; air traªc control Cutrell, Ernest A., 27, 67, 71, 86–87
radar and, 166–167; airways use statistics of, cyclotron, 125, 135
154–156; Congress and, 146, 148, 150–152,
155–156, 160, 163; demise of, 176; GCA sys- Day, Karl, 28, 29
tem and, 144–150; glide path and, 108, 122; de la Vaux, Count, 47, 51
ILS and, 10, 137–139, 141, 142, 154, 158, 160– Delco Electronics, 118
162; Interagency Air Traªc Control Board, Department of Commerce: Aeronautics
153; landing aids and, 170–172; microwave Branch, 57, 58, 63–64, 66; aviation regula-
research and, 80, 81, 83, 84, 88, 94, 98, 101– tion by, 8, 55
103, 106; negotiations between airlines, design lock-in, 131
Army Air Corps and, 180, 181; policy stan- Diamond, Harry, 59, 63, 74, 89, 161
dardization and, 142, 143; radar issues and, Dingley, Edward, 52–54
137, 145, 146, 148–151, 175–176; system disorientation: demonstration of, 28; spins re-
problems and, 105, 106 sulting from, 19–20
Civil Aeronautics Board, 163, 172 distance measuring equipment (DME), 120
Clarke, Arthur C., 132, 169 Doolittle, James “Jimmy,” 23, 60–62, 67, 183
clouds, spins in, 19–20 Doty, Lyman, 19
Collier Trophy, 23, 137 drainage systems, airfield, 42–44
Colonial Air Transport, 31 DuBridge, Lee, 106, 125, 164
Columbus, Ohio, 42 Duckworth, J. B., 141
Colvin, Charles H., 15 Duncan, C. E., 99
Committee on Instrument Landing Devices Dunmore, Francis, 63, 74, 89, 161
(Radio Technical Commission for Aeronau- Durand, W. F., 93
tics), 86 DVL, 76
compass: gyroscopic, 22; magnetic, 13, 14, 16;
radio, 70–71, 74, 80 Eagle blind bombing radar, 131
compass locators, 71, 107, 112, 182 Eaker, Ira, 152, 183
214 index
Eastern Airlines, 7, 84, 151 Gerdon, H. P., 69

Eastern Air Transport, 25, 31, 67, 71 Germany: blind bombing and, 75; blind land-
Eastman, Clyde, 99 ing technology in, 3, 22, 49; radio blind land-
Eclipse Aviation, 74 ings and, 76–77
Eighth Air Force, 112–113, 165, 182–183 Gherardi, Bancroft, 93
Eisenhower, Dwight D., 175 Gilbert, Glen, 144–145, 149, 155, 166
Ellsworth, M. Harris, 152, 158–159, 178 Gilfillan Brothers, 128, 129, 131–133, 143, 146,
environmental issues, 6–8, 55 147, 149–150, 158, 166
equal signal transmitter, 59 glide path: Army and, 104–105; CAA, 108;
Erlanger, Joseph, 93 curved, 83, 84, 86, 102–103; development of,
Experimental Aircraft Association (EAA), 63, 64, 122; Indianapolis system and, 88,
180 101; ITRM, 108, 109; length of, 80, 97–98;
microwave, 93, 95, 96, 119; pilot training
Farley, James, 32 and, 83; shape of, 87–89, 96, 97; using hori-
Federal Aviation Act of 1958, 181 zontal polarization, 75
Federal Aviation Administration: function of, Goldsborough, Paul, 84
179, 180; outlook for, 181 Gorrell, Edgar S., 94
Federal Aviation Agency, 176. See also Federal grass airfields, 36, 37, 45
Aviation Administration Griªn, Bruce, 125
Federal Communications Commission, 81, 86, Griªths, Frank, 135, 185
173 Griggs, David, 127
Feminist historiography, 178–179 ground-based flight training, 29–31
FIDO, 167, 169–171 ground-controlled approach (GCA) system: au-
flight instruments: artificial horizon and, 21– tomatic, 158; controversy between ILS and,
25, 27; dependency on, 25; developments in, 138 –163, 171, 174; costs of using, 138, 139,
13–33; placement of, 27; training programs 161–162; data on, 154, 155; development and
in, 25, 27–28; during World War I, 13–14. testing of, 127–133; explanation of, 122, 126;
See also specific instruments function of, 133 –134; instrument approaches
flight integrator, 22 under, 145; integration of ILS and, 163, 164;
flight into terrain, 19 performance data on, 154, 155; politics and,
Florez, Luis de, 139, 148 150 –162; precision radars and, 168 –169;
fog landings. See blind landing; blind landing promotion of, 137, 138; safety issues and,
systems 156, 157, 177, 178; search radar and, 165, 168;
Ford, Henry, 7 as traªc control aid, 157–158; training for
Forest, Lee De, 52 use of, 139, 141; use of, 135, 137–138, 177–
Foulois, Benjamin, 32, 33, 36, 40, 71 179
France: blind landing technology in, 3; leader ground proximeter, 48
cables and, 45–47 Guerlac, Henry, 118
Freng, R. T., 85, 87 Guggenheim, Daniel, 22–23
Frye, Jack, 145 Guggenheim, Harry, 23, 61, 186
Guggenheim Foundation for the Promotion of
Gazely, Richard, 87, 88, 95 Aeronautics, 3, 23, 182
GCA. See ground-controlled approach (GCA) Gunn, Dano, 93
system gyrocompass, 27
Gee system, 112, 120 gyrorector, 22, 23
General Electric, 96, 165 gyroscope, 15
Georgia School of Technology, 23 gyroscopic compass, 22
index 215
Hadfield, A. H., 145 Kettering, Charles, 16–18

Hansen, James, 3 King, Ernest J., 75
Hansen, William, 90, 91 klystrons, 90–92, 116
Hanson, Earl C., 52–54 Komons, Nick, 158
Hartranft, J. B., Jr., 139, 155, 156, 179 Korean War, 175
Hegenberger, Albert F., 70, 71 Kreusi, Geo¤rey, 70
Hegenberger system, 70, 72–73, 80, 96, 100, Kroger, William, 139, 143, 145, 150, 153
101, 112, 160
Heilbron, John, 125 Lamborn, Charles, 19
Henderson, Lawrence Joseph, 93 Land, Emory S., 145
Henderson, Paul, 18–19 Landing Aids Experiment Station (LAES), 168–
Hester, Clinton, 92, 101 171
Hinshaw, Carl, 152 landing beams, 63–64
Holloman, George V., 82–83 landing fields. See airfields
Hooper, Stanford C., 52, 57 Landis, John, 162, 172
Hoover, Herbert, 7, 51, 57, 59 Langmuir, David, 106, 164
Hunsaker, Jerome, 93 L’Armee de l’Air (France), 15
Lawrence, Ernest, 125
ILS. See instrument landing system (ILS) Lawrence, T. E., 6
Indianapolis system, 87–88, 95–97, 101–106, leader cable system: Dingley, 53, 54; Hanson,
115 52–54
instrument flying, 28 leader cable system: applications for, 49, 51;
Instrument Instructor School (Army Air background of, 45–47; Loth, 45, 47, 50; vari-
Forces), 144 ations of, 51–55
instrument landing system (ILS): Air Coordi- Leary, William, 81
nating Committee and, 159; Army Air Forces LeMay, Curtis, 168, 183
use of, 137; CAA and, 10, 137–139, 141, 142, lighting: approach and landing, 169, 170; center-
154, 158, 160 –162; controversy between line, 170–171; slope-line, 170
GCA and, 138 –163, 171, 174; integration of Lindbergh, Charles, 1, 4, 13, 21
GCA and, 163, 164; Korean flight crash and, Link, Edwin A., Jr., 29–33
177; politics and, 152–162; views of, 138, 177, Loomis, Alfred, 117, 124, 125, 127, 128
178 Loomis, F. Wheeler, 125, 165
instrument landing training, during World War Loran system, 120
II, 114 Lorenz, Carl, 77
Interagency Air Traªc Control Board (IATCB) Lorenz system: development, 76–78, 83, 84;
(Civil Aeronautics Authority), 142, 153 production of, 106
International Telephone and Radio Manufac- Loth, William, 45–47, 52
turing Company (ITRM), 108, 109, 118 Loth leader cable system, 45, 47, 48, 50. See also
International Telephone and Telegraph (IT&T), leader cable system
82, 86 Lucas, Keith, 14
International Telephone Development Corpora-
tion (ITD), 86, 92, 104, 105, 108 magnetic compass, 13, 14, 16
magnetron: development of, 117, 122; resonant
Jackson, William, 95, 97
cavity, 124
Jewett, Frank, 93, 98, 125
Mark I GCA, 128–133
Johnson, James, 145–146
Mark II GCA, 129–132
Johnston, Lawrence, 125, 126, 129, 131
Marriner, A. W., 85, 87
Jones-Barany chair, 20, 21, 28
216 index
Mason, Max, 93 Night Flight (Saint-Exupéry), 1, 4

Massachusetts Institute of Technology (MIT), Northwest Air Lines, 31
3, 23, 115; Lincoln Laboratory and, 182; micro-
wave technology and, 80, 86, 89, 91, 92, Ocker, William C., 13, 20–22, 28
94; Radiation Laboratory, 104, 105, 117–119, Ocker-Meyer method, 21–23
122–125, 135, 165 Oªce of Naval Aeronautics (U.S. Navy), 22
Mauborgne, John O., 83, 84, 91, 99 Oªce of Naval Intelligence (ONI), 67
McAlpine, Major, 47 Oªce of Scientific Research and Development
McClelland, Harold, 130, 143, 152, 159, 167 (OSRD), National Defense Research Com-
McDougall, Walter, 175 mittee, 124
mental hazard, 69 1-2-3 method (Stark), 25, 26, 28
Metcalf, Irving, 89, 91
Meyers, David, 20, 21, 28 Pan American Airlines, 28, 145
microwave research: Army support for, 115; pilot-control model, 103, 123
early events in, 89–92; klystron and, 90– pilot training. See training programs
92, 116; magnetron and, 117; National Acad- Pioneer Instrument Company, 15
emy of Sciences and, 92–102; outlook for, Pioneer Instruments, 30
122 plan position indicator (PPI) displays, 131
Miller, Wendell, 42 Popular Science, 145
Mitchell, Hugh, 87, 88, 91, 95 Post Oªce, U.S.: air routes of, 31, 40; early air
Morse code, 59, 60 mail service of, 1–2, 4, 5, 16–19; lighted
Moseley, Francis L., 113 airfields and, 51
multipath e¤ects, 78 Praeger, Otto, 7, 16–19
Murray, Charles V., 137 precision approach radars (PARS), 172
Presidential Special Board of Inquiry on Air
National Academy of Sciences (NAS): blind Safety, 163
landing programs and, 81; RTCA system Project Baker, 68
and, 92–102 Provisional International Civil Aviation Organi-
National Air Traªc Controller’s Organization zation (PICAO), function of, 119–122, 174
(NATCO), 180 pulsed glide path (PGP): development of, 104,
National Air Transport, 31 117, 127; reasons for nonproduction of, 118–
National Bureau of Standards (NBS): blind 119; tests of, 117, 118
landing research and, 51, 57, 67–69, 103; es-
tablishment of, 57; goal of, 3; radio research Rabi, I. I., 125
and, 2, 23, 56, 57–79 radar: air safety and, 151; air traªc control,
National Business Aircraft Association, 180 165–167; airport surveillance, 172; Civil
National Defense Research Committee (Oªce Aeronautics Authority and, 137, 145–146,
of Scientific Research and Development), 148–151, 175–176; eagle blind bombing, 131;
124 networking automatic long- and short-range,
National Research Council, 93 174; precision approach, 172
National Transportation Safety Board, 177 Radiation Laboratory (Massachusetts Institute
Naval Air Transport Service (NATS), 152, 153 of Technology), 117–119, 122–125, 135, 165
Navy, U.S.: blind landing research by, 3; GCA radio altimeters, 62, 63
and, 137, 139, 141, 142, 152–153; NBS landing radio blind flying: A-I system and, 70–74, 78,
system and, 67–69 79, 97; commercial airlines and, 73–75;
NBS. See National Bureau of Standards (NBS) cross pointer and, 63, 65; experimental de-
New York Herald Tribune, 152 velopments in, 66–73; landing beams and,
New York Times, 145, 151 63–64, 66, 71; Lorenz system and, 76–78;
index 217
organizations involved in, 57–58; radio navi- Seidel, Robert, 125

gation beacons and, 59–62; stability issues Selfridge Field, 43, 44
and, 58, 69–70, 78, 79; United Air Lines Shangraw, C. C., 73
and, 74–76 Sherlock, Harry, 19
Radio Technical Commission on Aeronautics slope-line lighting, 170
(RTCA): blind landing research and, 107, Snow, Chester, 71, 73
160; function of, 81, 163, 173–174; Interna- Société Industrielle des Procédés Loth, 47
tional Telephone Development Corporation’s Spaatz, Carl, 113, 121
system and, 86; legal authority issues and, Specifications for Municipal Landing Fields
81–82; recommendations of, 174–175; stan- (Army Air Service), 36
dards system and, 83–89 Sperry, Elmer, 15, 16, 22, 23
railroads, 14 Sperry, Elmer, Jr., 23
RCA, 86, 96 Sperry, Lawrence, 15, 16
recurrent training, 28 Sperry artificial horizon, 23, 24
Reeves, J. W., 152, 153 Sperry Gyroscope Company, 15, 80, 86, 91, 92,
religious symbolism, 6–7 96, 99, 100, 104, 105, 115, 118–119
Rentzel, Delos, 173–174 Sperry Gyroscope microwave glide path trans-
resonant cavity magnetron, 124 mitter, 116–117
Rickenbacker, Eddie, 7, 182, 183 spins: causes of, 19–20; research on, 20–21
Rivers, L. Mendel, 151 Spoils Conference (1930), 31
Robertson Aircraft Corporation, 31 Sproul, Clarence B., 142, 152–155, 158
Rochester, New York, 42–43 St. Paul, Minnesota, 43
Rockefeller, William A., 7 standard beam approach (SBA) system: expla-
Roosevelt, Eleanor, 84 nation of, 106, 112, 114; performance of, 127
Roosevelt, Franklin Delano, 31–33, 81, 93, 98– Standley, William H., 75
100, 102, 104, 120, 183 Stanford University, 3, 23, 80, 90–92
Royal Aeronautical Society, 138 Stanton, Charles: CAA policy and, 94, 95; ILS
Royal Air Force (RAF): blind landing technolo- and, 138, 139, 146; landing aids decisions
gies and, 2, 106, 112; World War II and, 113, and, 120
127, 164 Stark, Howard, 25–28
Royal Aircraft Establishment (RAE), 47–49 Stern, Ben, 144
Royal Aircraft Factory, 14, 15 Stevenson, Charles, 45
Royal Navy, 45–46 Stoner, Clayton, 19
RTCA. See Radio Technical Committee on Subcommittee on Instrument Landing Devices
Aeronautics (RTCA) (Radio Technical Committee on Aeronau-
rudder, 26, 27 tics), 82
Saint-Exupéry, Antoine de, 1, 4 Tasker, Homer, 128, 131, 132, 143

Sanders Teacher, 29 Telfunken, 76
SBA. See standard beam approach (SBA) system Tizard, Henry, 117, 122, 124
SC-31, 175 training programs: curved glide path and, 83;
Schaefer, E. B., 27 GCA, 139, 141; ground-based, 29–31; instru-
Schatzberg, Eric, 8–9 ment flying, 25, 28–29; recurrent, 28; in
SCR-584, 126 Stark’s rule-based flying, 27–28; during
SCS-51: data on, 118, 154, 155; development of, World War II, 114
104, 110–115; instrument approaches under, Transcontinental Air Transport, 66
145; use of, 119, 121–122, 137, 144, 175. See Transcontinental and Western Air (TWA), 31, 145
also very high frequency (VHF) glide path Truman, Harry, 151, 163
218 index
tuned read indicator, 61, 62 Watson, Edwin, 98–102

turf airfields, 38, 40–42 Watson-Watt, Robert, 120
turn-and-bank indicator, 26 weather: accidents related to, 19 –20; forced
turn indicator, 15–16, 17f, 25 –27 landings caused by, 1–2; research on,
Twelve O’Clock High, 182–183 183
Tymms, F., 49 Webster, William, 90
Western Air Express, 31
UHF frequencies, 85 Western Electric, 23
UHF localizers, 67, 75, 78, 96 Westover, Oscar, 93
UHF/microwave hybrid, 96 Whitney, Cornelius Vanderbilt, 7
UHF system, 94, 106, 107 Wilbur, Ray Lyman, 91
United Air Lines: blind landing developments Willis, Hugh, 91
and, 7, 28, 31, 32; radio blind landing sys- Wilson, John, 144, 170
tems and, 74–76; view of blind landings, 96 Winner, Langdon, 178
United-Bendix system, 75–78, 82 Wohl, Robert, 6
United Kingdom, 45, 47–49 Wolverton, Charles, 152
University of Michigan, 23 Woodring, Harry, 98–100
University of Washington, 23 World War I: airfields during, 35, 36, 40; blind
U.S. Air Forces in Europe (USSTAF), 113, 114 landing technology during, 45, 182; pilot
U.S. Military Air Transport Service, 121 training during, 30
World War II: blind bombing in, 75, 164; blind
Varian, Russell, 90, 91 landing technologies during, 10, 11, 107–119;
Varian, Sigurd, 90, 91 defense procurement during, 151; glide path
Varney Air Lines, 31 projects and, 104, 105, 107; visual omni-
Verne, Jules, 8 range and, 120
vertigo, 20, 21 Wright, Theodore P., 138, 145, 148–150, 152,
very high frequency (VHF) glide path, 104– 153, 156, 157, 161, 166, 172
105, 115. See also SCS-51
Vidal, Eugene, 70, 71, 73 XT-I, 126, 127, 133. See also SCR-584
visibility, 12
visual omni-range (VOR) system, 120 YB system, 67–69, 75
Yount, Brigadier General, 102
Warner, Edward P., 121
Washington Institute of Technology, 67, 82, 86 Zeppelin acoustic height finder, 62, 63

Blind Landings - Low-Visibility Operations in American Aviation, 1918 - 1958 (PDFDrive)

Uploaded by

Copyright:

Available Formats

Blind Landings - Low-Visibility Operations in American Aviation, 1918 - 1958 (PDFDrive)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Blind Landings - Low-Visibility Operations in American Aviation, 1918 - 1958 (PDFDrive)

Uploaded by

Copyright:

Available Formats

This page intentionally left blank

The Johns Hopkins University Press

The Johns Hopkins University Press

Library of Congress Cataloging-in-Publication Data

2 Places to Land Blind 35

3 Radio Blind Flying 57

4 The Promise of Microwaves 80

5 Instrument Landing Goes to War 104

6 The Intrusion of Newcomers 123

7 The Politics of Blind Landing 137

UHF ultra-high frequency (after 1944; frequencies between 3 MHz and

thus a product of what a more conspiracy-minded scholar might call a “military-

nologies of steamboat, steamship, and railroad had transformed American life.

engineers to innovate to stabilize them. Chapter 4 examines one potential solu-

instruments for blind flying

been invented as a means to overcome the instability of the magnetic compass

never came, in part because of the danger to instructors imposed by a completely

quired instilling in pilots an extreme form of technological faith that required

Places to Land Blind

While the goal of all-weather operations forced pilots to learn to ﬂy blind, it

environmental problems that prevented the achievement of regularity.2 They are

runway centers, further minimizing construction costs. The runways themselves

making a technology fit:

The obsolescence of grass ﬁelds in turn impacted blind landing technologies

thus provided three-dimensional navigation information. Early versions of this

very rapidly, the frequency of the generated electromagnetic ﬁeld had to be

Dingley advertised its chief advantages as its independence of radio frequen-

Radio Blind Flying

flying the (unstable) beam: the national bureau

The National Bureau of Standards had been asked by Secretary of Commerce

demands, made the system impossible to maintain. It never became a serious

He used two low-frequency, omnidirectional radio transmitters that he called

The Promise of Microwaves

Commercial pilots’ rejection of the Hegenberger system left a schism in the

defining a “standard” system:

the microwave future

system established a plane that provided an indication of proper course. Similarly,

the national academy of sciences:

The completion and relatively successful testing of the International Tele-

MIT.49 If we interpret Arnold’s memo as referring to the microwave system, then

The process of constructing a standard landing system was made painful by

Instrument Landing Goes to War

President Roosevelt’s May 1940 approval of the National Academy of Science’s

both microwave systems—the wave of the future in 1940—have vanished into

entrenching the low-tech solution, part 1:

In the two years following Franklin D. Roosevelt’s blessing of the National

Early production models were subjected to extensive testing at various locations

a brief interlude: demise of the high-tech solution

entrenching the low-tech solution, part ii: picao

The Provisional International Civil Aviation Organization had been formed as

The Intrusion of Newcomers

inventing a new model:

“contact rules” pilot when he conceived of his famous ground-controlled ap-

The Mark I consisted of a gasoline-driven generator powering the radars

The Politics of Blind Landing

“Based on discussions with oªcers who have had considerable experience in

congress enters the fray

1944 18.1 5.5 2.1