1 s2.0 0012825289900202 Main
1 s2.0 0012825289900202 Main
1 s2.0 0012825289900202 Main
RUUD WEIJERMARS
ABSTRACT
Weijermars, R., 1989. Global tectonics since the breakup of Pangea 180 million years ago: evolution maps and
lithospheric budget. Earth-Sci. Rev., 26: 113-162.
Pangea, the Earth's youngest single supercontinent, broke up 180 million years ago. Tectonic plates were
subsequently formed by dispersal of the continental fragments and accretion of new oceanic lithosphere. The
configurations of all the major lithospheric plates at 0, 20, 65, 95, 140, 180 Ma BP are reconstructed on six globes of
the Earth, each with a radius of 10 cm. It appears that plate boundaries maintain a remarkably close fit on model
globes of constant radii if the reconstructions include the recovery of subducted spreading patterns. This is illustrated
with maps in equatorial orthographic, oblique orthographic and transverse Hammer-Aitoff projections. The snug fit of
the tectonic plates at every tested time since the breakup of Pangea 180 Ma BP is consistent with the theory of plate
tectonics on a non-expanding Earth and contradicts rapidly expanding Earth models.
The areas of oceanic lithosphere produced and consumed during the past 180 Ma BP are estimated from surface
measurements of the globes reconstructed on the basis of particular assumptions. These measurements suggest a
consistent increase in the production rate of oceanic lithosphere during the past 140 Ma. It was decided to revise the
assumptions and see if alternative reconstructions of the ancient spreading patterns on the floors of the Tethys and
Eo-Pacific oceans could avoid implying an increase of lithospheric production rates with time. This appeared to be
possible. The revised maps suggest that ophiolites older than 180 Ma BP may have been obducted in Cenozoic
collision zones of the Himalayas, Andes, Rockies, and the western part of the Banda Arc (Timor, New Guinea).
Estimates of the ocean floor production and consumption budget appear to be quite similar for both map series,
and only the possible ranges are summarized here, time averaged for the past 180 Ma. World-wide production and
consumption of oceanic lithosphere appears to have varied between 2.6 and 3.5 km2 a -1 at most. The mean of
world-wide spreading velocities ranges between 2.4 and 3.5 cm a - l and world-wide means of subduction velocities lie
between 4.2 and 6.3 cm a-1. More accurate estimates are possible on the basis of particular assumptions and have
been specified for the various time intervals distinguished: i.e., the Neogene (0-20 Ma BP), Palaeogene (20-65 Ma
BP), Upper Cretaceous (65-95 Ma BP), Lower Cretaceous (95-140 Ma BP), and Middle and Upper Jurassic (140-180
Ma BP).
subsequently stated that a neat fit between most accurate means of reconstructing past
past plate boundaries is only possible if the plate configurations. It was first seriously ap-
Earth's radius was 80% of its modern value plied by Bullard et al. (1965) in a search for
about 200 Ma BP (Million years Before Pres- the best fit between the Atlantic outlines of
ent). His maps suggest a linear increase in Africa and South America. However, Bullard
Earth radius with time such that the Earth's et al. did not have access to a detailed age
radius was still only 85% of the modern value pattern of the oceanic lithosphere, which has
at 140 Ma BP. only recently been compiled after twenty years
Recently, it has been argued that the fast of oceanic exploration (Sclater et al., 1981). A
Earth expansion invoked by Owen (1983) is detailed comparison, between the evolution
contradicted by several physical principles maps of Owen (1983) based on a fast Earth
(Weijermars, 1986). It is therefore desirable expansion hypothesis, and those obtained here
that a world-wide reconstruction of past plate assuming an approximately constant Earth
geometries and positions on a reference globe radius for the past 180 Ma, is discussed in
of constant radius becomes available. Until Appendix 1.
now plate reconstructions on a reference Earth The oceanic spreading pattern compiled by
of constant radius were only available for Sclater et al. (1981) is used here not only to
particular parts of the Earth (e.g., Pitman and reconstruct the tectonic evolution of spread-
Talwani, 1972; N o r t o n and Sclater, 1979; ing patterns and geographical positions of
Barron and Harrison, 1980). plate boundaries and continents, but also to
A world-wide reconstruction on a con- estimate the budget of world-wide oceanic
stant-radius Earth is attempted here by refit- plate production and loss. The velocities in-
ting the outlines of ancient lithospheric plates. volved in the subduction and spreading pro-
This method does not involve any external cess can also be inferred. Estimates of mod-
reference frame and therefore provides the ern and some past spreading and subduction
rates have been attempted previously (e.g., position with time and this can be inferred
Deffeyes, 1970; Chase, 1972; Sclater et al., from the globes and maps presented here.
1980; Parsons, 1981). Most of these im- This makes the description of plate motions
portant estimates have been confirmed in the in terms of subsequent Euler pole positions
present study. Additionally, this paper pays and rotation angles very laborious (Cronin,
attention to all the geological data which are 1987), and it seems therefore most adequate
relevant and pertinent to the reconstruction to simply map the position of tectonic plates
of the motion of tectonic plates. at particular time intervals, a strategy adopted
here.
M E T H O D O L O G Y A N D MAPPING METHODS The oblique and equatorial orthographic
projections in Figs. 1 to 4 are obtained from
Five major bands were distinguished in the photographs of the original globes published
ocean floor on the basis of an existing compi- elsewhere (Weijermars, 1986). The projections
lation (Sclater et al., 1981), and plotted on a of Figs. 5 and 6 have been constructed from
globe of 10 cm radius (Fig. 1). Old oceanic unpublished photographs of globes recon-
crust with an age of 180-140 Ma occurs in structed of the Earth at 140 and 180 Ma BP,
three sites: in the Pacific, along the east coast respectively. It should be noted that photo-
of North America and along the west coast of graphic techniques to obtain orthographic
North Africa (Fig. 1). The oldest oceanic maps may introduce minor deviations from
crust of about 180 Ma BP occurs nearest to true orthographic projections (c.f., Robinson
the continental margins, which indicates that and Sale, 1968), but these have been ne-
Pangea broke up in the Middle Jurassic 180 glected here.
Ma BP (cf., Sclater et al., 1981). Such old Additionally, all results have also been
oceanic crust does not occur along the passive transferred to transverse Hammer-Aitoff pro-
margins of Southern Africa, South America, jections. This type of projection has been
Australia, India, and Antarctica. The spread- introduced for the presentation of oceanic
ing pattern record thus implies that the part spreading patterns (Weijermars, 1984), be-
of Pangea termed Gondwanaland by Suess cause it gives an overview of the entire ocean
(1885, p. 768) was still a single continent at floor in a single map. The equatorial
the end of the Jurassic 140 Ma BP. Hammer-Aitoff projection had been indepen-
The spreading pattern on the globe of the dently developed by Hammer from the Lam-
present-day Earth (cf. Fig. 1) was used to bert azimuthal equal area projection in 1892
reconstruct the evolution of the plate config- and earlier, in 1889, by Aitoff who had
uration since the breakup of Pangea 180 Ma adapted the equidistant projection (c.f.,
BP. Details of the construction of the globes Leighly, 1955). The projection applied in Figs.
have been outlined elsewhere (Weijermars, l g - 6 g is termed a transverse Hammer-Aitoff
1986). Figs. 1 to 6 summarise the develop- projection, because the globe and the projec-
ment of the oceanic spreading pattern in six tion net have been rotated 90 ° relative to
stages (i.e., 180, 140, 95, 65, 20 and 0 Ma BP). each other from the usual equatorial orienta-
The motion of the tectonic plates is repre- tion. This equal area projection places the
sented by their successive positions with re- north polar axis at the top and south polar
spect to the geographic grid at particular time axis in the central node of longitude lines of
intervals. No reference is made to their Euler the projected ellipse. The geometric centre of
poles of angular rotation, because the trajec- the ellipse is at 0 ° longitude and 70 ° south-
tory of any point on a plate is generally not a ern latitude, and its axial ratio is 1:2. The
small circle around a single axis of relative south polar latitude circles stay close to true
motion. This is due to the fact that the Euler circles, but neither parallels nor are meridians
pole of any plate may shift its geographic simple curves. Incidentally, this type of pro-
116
Fig. 1. Maps of the oceanic spreading pattern on the modern Earth. a - f : Global views in oblique (a, c-e) and
equatorial (b, f) orthographic projections, g: Overview of the world's modern oceanic spreading patterns, ridges and
transform faults mapped in transverse Hammer-Aitoff projection. Compiled after the data of Sclater et al. (1981).
117
160 ,180
O C E A N I C CRtJ
Neoge
(0-20
Pelael
(20-6
~ Upper
165-9 - CRUST
Lower
(95-1
I Mlddl~ ont
(140-
More detailed subdivisions and precise dating of the ages distinguished can be found in the time tables of Palmer
(1983), Harland et al. (1985), Berggren et al. (1985), and Kent and Gradstein (1985).
118
jection has been used by oceanographers to butted as follows. Computer fits are also sub-
visualize an overview of the global water cir- jective in the sense that in the refit procedure
culation pattern above oceanic lithosphere they require a series of algorithms which pre-
(cf., Dietrich and Ulrich, 1968, fig. 13). scribe the ratio of overlaps and gaps allowed
for, and possibly their preferred location. De-
BASIC ASSUMPTIONS signing of this algorithm in itself would be a
subjective procedure similar to that occurring
The basic assumptions underlying the con- during manual refitting. Such an algorithm
struction of all the maps in Figs. 1-6 will be would also be difficult to explain to the gen-
outlined here. These assumptions, valid for eral reader and therefore makes any com-
the reconstructions of the tectonic evolution puterfit more difficult to reproduce indepen-
of the Earth since the breakup of Pangea 180 dently than those on material globes. The
Ma BP, are that: manual method is pedagogically more power-
(1) manual reconstructions of the move- ful than any computer method because the
ment of tectonic plates on globes of 10 cm results presented here are reproducible
radius are sufficiently accurate; without the large research budget, machinery
(2) any Earth expansion has been negligi- and detailed software needed for computer
ble; simulations. Nonetheless, the manual con-
(3) the surface area of continental litho- struction was used here principally because
sphere has remained approximately constant the development of the software necessary for
(i.e., only the addition and loss of oceanic computer methods is extremely time-con-
lithosphere is significant to the budget, and suming in view of the application and alterna-
any change in continental area is insignifi- tive approach considered here. I look forward
cant); to any future corrections of my work on the
(4) Antarctica has remained fixed relative basis of computer fits.
to both the mantle and the geographic grid; Comparing the areas occupied by oceanic
(5) spreading patterns have been generated and continental lithospheres on the modern
symmetrically about the oceanic ridges (with Earth measured on real globes and digitised
special reference to the Pacific); in a computer gives some idea of the dif-
(6) the Tethyan collision zone may be sim- ference (and perhaps accuracy) of the two
plified into a single suture; and methods. Oceanic and continental litho-
(7) no spreading ridges, other than linear spheres occupy 59.5 and 40.5% of the total
extension of those still visible today (i.e., the surface area, respectively, according to mea-
Pacific Ridge) have been subducted; this par- surements on my reference globe of the mod-
ticular assumption was revised in the con- ern Earth, whereas 59 and 41% is the distribu-
struction of alternative maps in Figs. 9-12 tion determined from computer maps of the
(see later). modern Earth (Cogley, 1984). The difference
All these assumptions are based on careful is no more than 0.5% and the accuracy in my
considerations and these will be discussed in estimates is meaningful, because a non-di-
turn below. mensional surface area of 0.5% on my globes
corresponds to about 6 cm 2. More specifi-
(1) Accuracy cally, the total area of 1256 c m 2 o n any of my
globes corresponds to 5.1. l0 s km 2 on the
Many colleagues have suggested that com- real Earth, because the radius (r) of each
puter reconstructions of past plate configura- globe is 10 cm, that of the real Earth is 6370
tions would have been less subjective and km, and the surface area of any sphere is
more accurate than manual constructions on given by 47rr 2. I assume that my measure-
material globes. This criticism can be re- ments on the globe are accurate within + 1
119
between 1.7 and 3 Ga (All~gre and Jaupart, that part of the overlaps in my maps, in
1985). There exist a variety of smooth con- particular those occurring between the
tinental growth curves estimated on the basis Atlantic margins of the northern continents
of isotopic systems (e.g., Reymer and Schu- (Fig. 6), may be explained by various mecha-
bert, 1984), but even extreme models with nisms of crustal stretching operating during
episodic growth have been suggested (Reymer the formation of passive continental margins
and Schubert, 1986). However, such episodes (cf., Van der Linden, 1979; Lister et al., 1986).
may be only apparent and due to accelerated Continental shortening by compression and
recycling rates of continental material through transpression is occurring along subduction
former subduction zones whilst the chemical and continental collision zones (cf., Helwig,
fractionation of the mantle may still follow a 1976). Crustal shortening due to nappe
smooth curve (Gurnis and Davies, 1986a). thrusting and ductile thickening in the Pe-
Estimates of continental growth rates may ruvian Andes amounts to at least 115 km
be based either on isotopic systems (cf., (M6gard, 1984; cf., Jordan et al., 1983) al-
Kr~Sner, 1985; All~gre and Jaupart, 1985) or though this may have been partly com-
the development of continental freeboard pensated by coeval continental growth both
through geological time (McLennan and due to off-shore accretion of sedimentary
Taylor, 1983; Schubert and Reymer, 1985; wedges (McCourt et al., 1984) and exten-
cf., Taylor and McLennan, 1985). Even if one sional tectonics such as observed in the con-
would adopt estimates of extremely fast net tinental crust above the subducting Cocos
growth rates of the order of 1 km 3 a -1 (Rey- plate (Aubouin et al., 1984). Crustal shorten-
mer and Schubert, 1986) this only adds some ing in the Canadian Rockies has been esti-
(1.8.108 km3/50 km = ) 3.6.106 km 2 to the mated at 160 to 240 km (Bally, 1981; Price,
continental area in 180 Ma. This is less than 1981).
0.01% of the modern surface area of the Earth Crustal shortening in continental collision
and may therefore be neglected in reconstruc- zones seems much larger than that occurring
tions of modern plate tectonics. above oceanic subduction zones. For exam-
Increase of continental areas may also oc- ple, crustal shortening in the Swiss Alps has
cur by lateral extension due to rifting, basin- been estimated at about 600 km (Trumpy,
and-range tectonics and the formation of in- 1973). More spectacularly, shortening in the
tercontinental basins by crustal thinning. For Eurasian continent caused since the collision
example, seismic reflection profiles of the of India and Eurasia about 40 Ma BP, may
continental floor of the North Sea Basin sug- be of the order of 2000 km (Patriat and
gest over 100% crustal stretching by normal Acharche, 1984; Besse et al., 1984). Most of
faulting during the past 160 Ma (Barton and this shortening may be due to "escape tecton-
Wood, 1984). Palinspastic reconstruction of ics" or movement of southeast Asia towards
the Pannonian Basin in eastern Europe sug- the Pacific by strike-slip faulting. The crustal
gests 75-100 km Neogene extension (Royden thickening in the Himalayan Chain seems to
et al., 1982). Palaeomagnetic data from the account for only 600 km or 30% of the total
Basin-and-Range Province in the United shortening (Pelzer, 1989). Crustal thickening
States suggest that 300 km spreading may and denudation by exogenic erosion is still
have occurred since 90 Ma BP (Frei, 1986) by continuing today due to indentation of the
extensional tectonics (Wernicke, 1985; Eurasian continent by the Indian plate at a
Wernicke and Burchfield, 1982). Very Long rate of about 5 cm a -~ (Tapponier et al.,
Baseline Interferometry (VLBI) measure- 1982).
ments suggest that the Basin-and-Range Pro- Present knowledge is insufficient to assess
vince is still extending E - W today at a rate of current estimates of the world-wide extension
0.3-1.5 cm a -~ (Allenby, 1979). Note also and thickening budgets of continental litho-
121
sphere in the Mesozoic and Cenozoic. I there- 70 Ma (Yumi and Wako, 1970; Enslin, 1978).
fore had to assume in my global reconstruc- It should be noted that this true polar
tions and the calculation of production and wandering is in fact, a movement of the whole
subduction budget of oceanic lithosphere (see Earth relative to the axis of rotation, which
later) that the net surface area of continental aside from precession an nutation, is assumed
lithosphere has remained constant during the to remain fixed relative to the Sun (Gold,
past 180 Ma. Even under such assumptions 1955). This reorientation of the Earth in space
care should be taken with estimates of sub- is likely to have its origin in core and mantle
duction rates balanced by spreading, and vice dynamics (Goldreich and Toomre, 1969).
versa, in view of the possibility that oceanic Geodetic observations cannot help much
lithosphere may be shortened by folding or further in unravelling how the Earth's rota-
thrusting. At least the northeast floor of the tion axis has wandered in the remote past,
Indian Ocean may have been buckled elasti- but other geophysical observations can. Plate
cally into l o n g - w a v e l e n g t h ( 1 0 0 - 3 0 0 positions relative to the geographic poles can
km)/small-amplitude (3 km) folds (Weissel et be constrained using palaeomagnetic and hot
al., 1980; McAdoo and Sandwell, 1985) and spot data, in combination with palaeoclimatic
the Gorda plate may be deformed by shear data, as follows. The Earth's magnetic dipole
zones and kink bands (Wilson, 1986). How- field caused by dynamics in the Earth's liquid
ever, neglecting such effects seems still justi- outer core (Busse, 1975) is assumed to be
fied so long as the local shortening rates of aligned with Earth's rotation axis if averaged
ocean floor are small compared to contem- over long periods. The time-averaging re-
poraneous subduction rates. moves the effect of the secular variation of
the magnetic poles relative to the instanta-
(4) Palaeographic grid neous rotation axis (cf., Brown and Mussett,
1981). Hence, the movement of the geomag-
The relative position of tectonic plates in netic dipole relative to the mantle also reveals
the past can be recovered by refitting their that of the instantaneous rotation axis and
ancient outlines, but this approach cannot the associated positions of ancient geographic
recover their absolute or palaeogeographic poles relative to the Earth's mantle.
position. First, accept that a meaningful geo- This is now known in detail. The motion of
graphic grid should be fixed to the Earth's the palaeomagnetic pole, relative to a hot spot
rotation axis at any time; only then will the reference frame fixed to either the mantle or
various (palaeo-) climatic zones be related to core-mantle boundary, depending on the
the geographic grid in a fashion similar to origin of mantle plumes (cf., Loper, 1984,
that seen today. However, locating litho- 1985), has been recently determined for the
spheric plates relative to the geographic spin past 180 Ma (Andrews, 1985). The past rota-
poles rather than just to each other is com- tion pole has slowly spiralled to its present
plicated by the fact that, like the tectonic position. Its furthest offset occurred about
plates, the Earth's rotation axis is also moving 180 Ma BP when it lay 20 ° from the present
relative to the Earth's mantle. The respective pole. It has since come progressively closer,
directions and displacement rates of the poles so that it has never been offset more than 10 o
and plates need not be similar, for they are during the past 100 Ma.
likely to be due to different mechanisms. Consequently, the pole of the geographic
That the Earth's rotation axis moves rela- grid has not travelled far relative to the man-
tive to its interior is now well established. tle. But how far did the indiuidual tectonic
Geodetic measurements in celestial and satel- plates themselves displace ooer the underlying
lite reference frames reveal a recent secular mantle? Africa has been considered as a stable
motion of 10 cm a -~ averaged over the past plate on the basis of stationary hot-spots
122
Fig. 2. The relative distribution of the continents and the oceanic spreading pattern at the end of the Palaeogene 20
Ma BP. Projections as in Fig. 1.
123
160 180
m \ /
-0
O C E A N I C CRU
Palaeq mp in r e f i t
(20-6
~ Upper v e r l a p in r e f i t
165"9
N T A L CRUST
Lower
(95- 1 ~elf
~ Pre-Ci
12140 ontinent
~J
~D
t~
~o
~o
125
m 160\ /180
-0
3p ~n r e f i t
OCEANIC CRL
Upper verlap in r e f i t
165-9~
NTAL CRUST
Lower
(95-1 1elf
l Pre-Ci
(~-140 )ntlnent
Cr~
~D
~D
&.
C~
127
160 180
--0
ap in r e f i t
v e r l a p in r e f i t
OCEANIC CRI
a /
Fig. 5. The face of the Earth at the end of the Jurassic 140 Ma BP, just before the breakup of Gondwanaland.
Projections as in Fig. 1.
129
160 180
\ /
-0
ap in r e f i t
verlap in r e f i t
NTAL CRUST
lelf
OCEA
)ntinent
130
Fig. 6. The Earth at the end of the Lower Jurassic 180 Ma BP, just before the breakup of Pangea. Projections as in Fig.
1.
131
160 ~oo
-0
gap in r e f i t
overlap in refit
4ENTAL CRUST
shelf
0CI
continent
i:-lou md ~r~
132
(Burke and Dewey, 1974), but this is now much larger than the modern Pacific. Diffi-
contradicted by very mobile traces of hot- culties met in reconstruction attempts are due
spots suggested by others (Duncan, 1981; Van to four boundary conditions: ( 1 ) t h e Eo-
Houten, 1983). Absolute plate motions in a Pacific has been progressively shrinking since
hot-spot reference frame assuming no net ro- the breakup of Pangea; (2) the Eo-Pacific has
tation of the lithosphere also suggest that been entirely surrounded by subduction zones,
Africa is moving rapidly northeast, but these rather than passive continental margins (pos-
movements are only reliable for the past 10 sibly, but not necessarily, since the breakup
Ma (Minster and Jordan, 1978). of Pangea); (3) back-arc spreading has oc-
Antarctica seems to lie on a plate which is curred along the West-Pacific margin; and (4)
most stable relative to the mantle. This is the Pacific spreading ridge may not only have
because it has been surrounded by spreading displaced laterally with time, but possibly also
ridges since the breakup of Pangea 180 Ma branched once or twice (see later).
BP (cf., Knopoff and Leeds, 1972), and only Any reconstruction of the Pacific Ocean
minor subduction is likely to have occurred floor needs to take account of all these four
during the Tertiary folding of the Mesozoic problems for complete solutions of its plate
belt that forms the Antarctic Peninsula. Mod- tectonic history. Only partial solutions have
ern absolute displacement velocities also sug- been suggested until now. The displacement
gest that the Antarctic plate is nearly sta- of the margins of the Eo-Pacific was only
tionary with respect to the mantle (Minster recently included in such studies, but is still
and Jordan, 1978). poorly documented and then only back to 74
If Antarctica has not travelled far over the Ma BP (Whitman et al., 1983) or 100 Ma BP
underlying mantle, and if the geographic pole (Rea and Duncan, 1986). The maps in Figs.
has not moved very far either, it seems logical 1-8 and 8-12 account for the convergence of
to keep the geographic grid of palaeogeo- the margins and associated decrease in size of
graphic maps fixed relative to Antarctica. the Eo-Pacific in the last 180 Ma. This was
Such a choice is independently supported by possible because the entire world-wide
palaeoclimatic observations (cf., King, 1961; spreading pattern of Sclater et al. (1981) has
Gill, 1961; Lloyd, 1984), which suggest that been used.
Antarctica has remained at relatively high The Eo-Pacific has been diminishing since
southern latitudes during the past 180 Ma. It the breakup of Pangea, and therefore subduc-
is therefore Antarctica that has been kept tion has been occurring along its margins
fixed relative to the geographic grid in all the throughout the last 180 Ma even without any
evolution maps of plate tectonics in this study contemporaneous spreading of the Eo-Pacific
(Figs. 1-6 and 8-12). Incidentally, polar ocean floor. Pangea has been a single conti-
wandering may have caused internal defor- nent for at least 100 Ma before it was torn
mation of tectonic plates by normal, reverse apart and dispersed (Livermore et al., 1986;
and strike-slip faulting (Weijermars, 1985), Boucot and Gray, 1983), so that the Eo-Pacific
but these effects have been neglected here. must have been in existence throughout at
least the Permo-Triassic Period. Given only
(5) Symmetric spreading patterns in the Pacific one continent (Pangea) and one ocean (the
Eo-Pacific) this implies that the Eo-Pacific
Structural development of the Pacific Oc- must have comprised a constant area
ean floor poses the greatest problems to plate throughout the Permo-Triassic Period. Note
tectonic reconstructions. The Pacific is the that there would have been no need for sub-
remnant of the Eo-Pacific or Panthalassa, the duction to occur if there were no spreading
single palaeo-ocean occupying Pangea before ridges in Permo-Triassic times. However, dat-
its dispersal 180 Ma BP; this Eo-Pacific was ing of orogenic phases in the Andes (Mc-
133
Court et al., 1984) and Rockies (Condie, 1981) times. Note that this assumption is not trivial,
suggests that more or less continuous subduc- in view of Menard's (1984) suggestion that
tion has occurred along the Pacific east coast spreading ridges may evolve asymmetrically.
since the Palaeozoic. This implies that the The current East Pacific Rise appears to
floor of the Eo-Pacific must have been ac- have been the Mid-Pacific Ridge at 65 Ma
tively spreading even before Pangea broke up. BP, if it is assumed to have been as far west
It is not clear whether or not the West- as the logic allows (Fig. 3). The floor of the
Pacific trenches have all been active since 180 Eo-Pacific in Figs. 4 and 5 also shows the
Ma BP. Palaeomagnetic reversal data from Pacific Ridge furthest west as it could have
the Phillipine Basin suggest that back-arc been at 95 and 140 Ma BP, respectively.
spreading has occurred there for the last 65 These tentative interpretations will be revised
Ma (Hilde and Lee, 1984; Seno a n d later (Figs. 8-12) after calculation of the
Maruyama, 1984), so that subduction in the world-wide spreading and subduction budgets
Mariana Trench must have occurred since based on the assumption of a single (western-
that time at least. Other subduction-related most) spreading ridge in the evolving Eo-
back-arc spreading regions in the west-Pacific Pacific (Figs. 1-6).
fall within the same age range (Jurdy and
Stefanick, 1983; McCabe, 1984). (6) Single Tethyan suture
A rigid connection between the floors of
the Pacific and Tethys oceans was suggested The modern subduction pattern in the
by Hilde et al. (1977) and widely accepted. Mediterranean region is one of the most intri-
This assumption results in a spreading cate on Earth and involves many tiny sub-
scenario where the Pacific Ridge migrated plates (e.g., McKenzie, 1972, 1978; Dewey et
from a westernmost position towards its pres- al., 1973; Le Pichon and Angelier, 1979). In
ent location in the east-Pacific. If this was so, fact, the current picture of the Eurasian-Afri-
either a slow oblique subduction zone or a can plate boundary is so conjectural that a
passive margin seems likely in the west-Pacific revision of its western part still seems neces-
until about 65 Ma BP. My preliminary recon- sary on the basis of detailed field work in
struction of the Pacific Ridge positions in southern Spain (Weijermars, 1987b, 1988a).
Figs. 1-6 is largely similar to that of Hilde et The collision zone between the African and
al. (1977), although I neglected the detailed Eurasian plates is therefore tentatively sim-
reorganisation of spreading ridges which may plified here into a convenient single suture of
have occurred during the Tertiary evolution east-west strike after Irving (1977).
of the Eo-Pacific (see later, cf. Appendix 3). Reopening of Tethys appears to be the
This exclusion is accounted for by assump- only way to refit the Atlantic plate boundaries
tion 7 (p. 134) and has been revised in Figs. at 20 Ma BP snugly (Figs. 2, a and b). Re-
8-12. opening of Tethys back in time is seen in the
The location of the Pacific Ridge at 20 Ma reconstructions to be a direct consequence of
BP (Fig. 2) was obtained by preserving a rigid closure of the Atlantic Ocean. Tethys appears
connection with the Antarctic Ridge during to have had her maximum extent when
the refit procedure on a globe (cf., Weijer- Gondwanaland broke up 140 Ma BP (Fig. 5).
mars, 1986). The age pattern of the floor in The final refit of the continents in Pangea
the eastern Pacific was subsequently recon- (Fig. 6) suggests that Tethys did not yet exist
structed by retracting oceanic lithosphere out 180 Ma BP. This is a controversial point of
of the subduction zones (mainly the east- view and will be explained below.
Pacific trenches) making the assumption that The absence of Tethys in my reconstruc-
such ocean floor would haoe been generated tion of Pangea at 180 Ma BP (Figs. 6 and 8)
essentially symmetrically about the ridge at all - - w h e n Panthalassa or the Eo-Pacific was the
134
only ocean in existence--is due to two fac- may have formed in the North Pacific be-
tors: (1) a revised fit of Gondwanaland at 140 tween 82 Ma and 50 Ma BP before merging
Ma BP; and consequently, (2) a preferred fit with the Pacific plate (Rea and Dixon, 1983),
of Pangea at 180 Ma BP. The fit of the would be an ancient example of such branch-
southern continents in Gondwanaland (Fig. ing.
5) differs from all previous reconstructions The other, more widely, recognized mecha-
(including that of Weijermars, 1986, figs. nism of subplate formation which leads to
7el-e6). This new fit is favoured because it is changes in spreading ridge geometry is the
the only one which allows all the southern breakup of larger plates. A well established
continents to have smooth travel paths to- example is the fragmentation of the former
wards lower latitudes (see Appendix 2). After Farallon plate into the Cocos and Nazca
the removal of the Caribbean plate between plates due to changes in the intraplate stress
South and North America (for which there field at 25-30 Ma BP (Wortel and Cloetingh,
are good arguments, e.g. Jordan, 1975, and 1981; Wortel, 1986). Fragmentation of the
Durham, 1985) this new fit of the southern plates and consequent changes in spreading
continents in Gondwanaland also allows a ridge geometry has probably occurred episod-
good fit of the northern continents to the ically throughout the evolution of the Eo-
margins of Gondwanaland in which there is Pacific's floor. However, opinions about the
no space for Tethys (Fig. 6). details of the Eo-Pacific's ancient plate con-
figuration are highly varied (see Appendix 3).
(7) No subducted spreading ridges Spreading ridges may even die before sub-
ducting. One particular clear example, but
The recovery of ancient spreading patterns not always recognized as such (e.g., see the
of oceanic plates along active plate margins is hot spot interpretation of Mahoney et al.,
a complex undertaking. The spreading pat- 1983), is the Ninetyeast Ridge in the Indian
tern of the Pacific Ocean floor has been re- Ocean. Spreading at this ridge during the
constructed by retracting oceanic lithosphere dispersal of G o n d w a n a l a n d separated
out of the Circum-Pacific subduction zones Australia from India until it ceased spreading
assuming that oceanic ridges have always had at about 20 Ma BP (Figs. 3-5). Since then the
symmetric spreading patterns. However, one relative positions of the Indian and Australian
associated problem is that particular spread- continents remained fixed (compare Figs. 1
ing ridges may have disappeared completely. and 2).
Spreading ridges may change their geom- It is very difficult to prove that particular
etry not only by transform faulting (cf., spreading ridges have existed if they are no
Menard, 1984) but also by the formation of longer directly visible. Still, careful plate re-
triple junctions whether this is called plate construction should be based on logical argu-
breakup or ridge branching. One recently dis- ments because only then can the results be
covered spectacular example of ridge branch- reproducible. A preliminary assumption (Figs.
ing is responsible for the fast formation of the 1-6) therefore was that only the linear exten-
Easter microplate, a circular Neogene platelet sion of spreading ridges still visible today has
about 200 km across forming along the East been subducted (i.e., the Pacific Ridge ). This
Pacific Rise between the Pacific and Nazca assumption required that the retracted oc-
plates (Hey et al., 1985; Schilling et al., 1985). eanic lithosphere which appeared during the
Branching is defined here as the local split- retrogressive reopening of Tethys had to be
ting of a spreading ridge into two (not neces- older than at least 140 Ma. This provides a
sarily parallel) spreading centres between firm basis for revising current opinions on the
which a new plate of fresh ocean floor may nature of Tethys (Figs. 8-12) after calcula-
form. The ancient Chinook microplate which tions of the world-wide spreading and sub-
135
duction budgets based on this assumption. It and 20 Ma BP (which appears almost full-
appears useful to revise this assumption and faced on the globe in Fig. 2) and which
propose other former spreading ridges in the subsequently disappeared from the modern
retracted floor (see later). globe (Fig. 1) allows an estimate of how much
65-20 old lithosphere has been subducted
PRODUCTION AND SUBDUCTION BUDGETS since its production (Table 1, section b).
This procedure can be applied to obtain
Areal distributions of oceanic and con- simple loss-estimates for most old oceanic
tinental lithosphere on modern Earth were lithosphere, but reconstruction of the ocean
measured on the globe drawn in Fig 1. Recall floor before 140 Ma is somewhat com-
that the measurements are assumed to be plicated. This is because it is unclear how
accurate within + 1 cm 2, which corresponds much spreading occurred in the period
to +(637) 2 km 2 or 0.1% of the global area. 140-180 Ma BP. Only 10.6.106 km 2 of such
The area of oceanic spreading patterns of the old ocean floor has survived until today, but
various ages distinguished in Fig. 1 are given it is likely that more existed at 140 Ma BP.
in Table I (section a). Note that more than It is obvious that the total surface area
one third (36%) of the present ocean floor is which the modern ocean floor occupies (303.2
20-65 Ma old. In other words, more than • 10 6 km 2) must have been composed entirely
60% of the oceanic lithosphere is of Tertiary of rocks older than 180 Ma when Pangea
age. broke up at 180 Ma BP. However, even though
no ocean floor older than 180 Ma has survived
Area of subducted lithosphere on the modern Earth surface, it can be in-
ferred that at least 303.2-106 km 2 of ocean
The area of lithosphere subducted since floor older than 180 Ma has been subducted.
180 Ma BP can also be estimated from the If the spreading ridges active between 180
globes, as follows. Assume that only oceanic and 140 Ma BP produced more than the
lithosphere has formed or been subducted in 10.6- 10 6 k m 2 still preserved today, then this
the last 180 Ma BP. The area of new litho- extra area of 140-180 Ma old ocean floor
sphere produced between, for e example, 65 should be added to the minimum subduction
TABLE I
Analysis of plate activity since the breakup of Pangea 180 Ma BP, based on the reconstructions in Figs. 1-6
TABLE I (continued)
Legend:
Section (a). Surface areas of the oceanic lithosphere measured on the globe of the modem Earth (Fig. 1).
Section (b). Subducted lithosphere measured from the globes mapped in Figs. 1-6.
Section (c). The area of the lithosphere produced since 140 Ma BP can be inferred by adding the estimates of sections
(a) and (b).
Section (d). Rate of production of the oceanic lithosphere was converted to spreading velocities by applying expression
1 in the text, after measuring the ridge lengths from the globes in Figs. 1-6.
*~ 1 cm2 of area on the model globe corresponds to (637) 2 km 2 of the real Earth.
,2 Including rocks older than 180 Ma BP. Minimum estimate.
,3 Minimum estimate. Possible spreading in excess of 10.6- 106 km 2 between 140 and 180 Ma BP has been excluded.
,4 1 cm length on the model globe corresponds to 637 km on the real Earth.
estimates given for ocean floor older than 140 older than 140 Ma, the production rate of
Ma (Table I, section b, see also footnotes 407.5 • 106 km 2 in 140 Ma corresponds to 2.9
there). km 2 a -1. This agrees remarkably well with
the independent estimate of Parsons (1981)
WorM-wide lithospheric production who suggests a mean rate of 2.99 km 2 a - l
Other earlier estimates include 2.65 km 2 a -~
The minimum estimate for the total area of (Deffeyes, 1970) and 3 km 2 a -1 (Chase, 1972)•
oceanic lithosphere subducted in the last 180 The production rate for the last 5-10 Ma
Ma BP is 418.1 • 10 6 km 2 (Table I, section b). has previously been estimated at 3.1 km 2 a-1
Neglecting continental growth and excluding (Garfunkel, 1975), which lies close to m y in-
Earth expansion, it is obvious that the total dependent estimate of 3.5 km 2 a -~ for the
amount of ocean floor subducted must equal last 20 Ma (Table I, section c). The latter
that which has surfaced during the same 180 estimate is in excellent agreement with Sclater
Ma. Details of the surface production of oc- et al.'s (1980) previous estimate of a modern
eanic lithosphere of a certain age have been production rate of 3.45 km 2 a i
estimated (Table 1, section c) by adding the
particular area of lithosphere subducted (Ta- EVIDENCE FOR SUBDUCTED SPREADING
ble I, section b) to lithosphere of the same age RIDGES
which is still visible on the present-day Earth
(Table I, section a). The mean production rates of ocean floor
A minimum total production rate of 418.1 for all the particular time intervals dis-
• 10 6 km 2 ocean floor in 180 Ma implies a tinguished here (Fig. 1) are specified in Table
mean production rate of 2.3 km 2 a -1. Exclud- I (section c). These suggest that there has
ing the minimum estimates of production been a remarkable increase in the lithosphere
137
~0
laP in r e f i t
) v e r l a p in r e f i t
ENTAL CRUST
;helf
OCE
:ontinent
Fig. 8. The Earth just before the breakup of Pangea 180 Ma BP. Similar to Figure 6g but with a different map
annotation for consistency with the revised maps in Figures 9a-12b. No spreading ridges are indicated, but they are
likely to have occurred and their approximate locations could be estimated retrospectively on the basis of Figure 9b.
The total ridge length might have been 6 × 10 4 krn and a possible representative production rate for oceanic
lithosphere is 3.5 km 2 a-1, corresponding to a worldwide mean spreading rate of 3 cm a-1.
139
(Table II) and the rest of the ocean floor cm a 1 (see later) implies that a representa-
([303-140]. 106 km 2) must have been older. tive ridge length is 6 • 104 km, at any time (cf.
This estimate can now be used, together with eq. 1).
further geological and geophysical constraint (9) Note that as Tethys did not exist be-
(see later), to revise the globe for the Earth of fore 180 Ma BP (Fig. 8) the fragmentation of
140 Ma BP (Fig. 5). This implies that a new Pangea would have created an active spread-
map annotation needs to be introduced to ing ridge on the floor of Tethys at about 180
distinguish between 140-180 Ma and older Ma BP.
lithosphere. It was decided to reserve the an- (10) There is also geological evidence for a
notation used for pre-140 Ma old ocean floor Tethyan spreading ridge that was still active
in Fig. 1 for 140-180 Ma old ocean floor in when Tethys reached her maximum extent at
the revisions (Figs. 9-12). Note that this 140 Ma BP. Obducted ocean floor is now
makes revision of Fig. 1 unnecessary. How- exposed as ophiolites in the Alpine Chain
ever, this also implies, for the maintainance of extending from the western Mediterranean to
uniformity of presentation in what follows, the Indian Ocean (cf. Lagabrielle et al., 1986),
that the map for the Earth at 180 Ma BP (Fig. and has ages ranging from 185-138 Ma BP
6g) needs a new annotation (Fig. 8). (Liguria), 180-150 Ma BP (Dinarides, Hel-
Fig. 9a shows the available data in order to lenides), 110-84 Ma BP (Arabia) to 178 158
clarify how I revised my reconstruction for Ma BP (Pontides-Lesser Caucasus) (see
the Earth of 140 Ma BP. The 140-180 Ma old Knipper et al., 1986).
remnant of the Pacific plate is shown on the (11) The author is not aware of any previ-
floor of the Eo-Pacific in the approximate ous attempts to reconstruct the spreading pat-
location suggested by palaeo-latitude data of tern at the subducted floor of Tethys. How-
Larson and Chase (1972) together with ever, precedents do exist for the evolution of
palaeo-longitude data of Henderson et al. parts of the Pacific ocean floor (see Appendix
(1984). 3). These are invariably based on somewhat
How can these fragments of oceanic litho- restricted philosophies because they consider
sphere on the base map of Fig. 9a, which the tectonic evolution of the Pacific as a
together occupy only 3.3% of the total oceanic problem isolated from the world context. The
area (cf Table I), lead to a complete re- world-wide approach advocated here involves
construction? Five further constraints are sug- more constraints for a revised reconstruction.
gested as replacing assumption (7) made Existing models for evolution of the Pacific
earlier, involving (7) the estimates of the display a wide variety and even where re-
surface area of particular spreading patterns, constructions by different workers agree, the
(8) average ridge lengths, (9) inferences from same tectonic plates may bear different names
the available sequence of snapshots (cf. Figs. (see Appendix 3). Nonetheless, the five
1-6), (10) detailed geological observations, spreading ridges indicated on Fig. 9b as sep-
and (11) previous attempts. These additional arating the Kula-Izanagi, Pacific, Farallon,
assumptions are discussed in turn below. and Phoenix plates have been suggested pre-
(7) Table II suggests that at 140 Ma BP viously (cf., Larson and Chase, 1972; Larson
half Earth's oceanic floor comprised 140-180 and Pitman, 1972; Hilde et al., 1977; Menard,
Ma old lithosphere, whilst the rest was older. 1978; Henderson et al., 1984; and Appendix
Note that these estimates are based on the 3).
adopted production rate of 3.5 km 2 a -1 in-
ferred from the reconstructions in Figs. 1-6. Jurassic ridges in Tethys and Eo-Pacific
(8) The ratio of the assumed steady pro-
duction rate of 3.5 km 2 a -a and representa- Both the 185-84 Ma old Tethyan ophio-
tive world-wide mean spreading velocity of 3 lites (Knipper et al., 1986) and the opening of
140 16o 180
-0
Fig. 9. a: Base map of the Earth 140 Ma BP showing the palaeogeographic positions of the continents and the
remnants of 140-180 Ma old ocean floor still preserved today. The location of the 140-180 Ma old remnant of the
Pacific plate in the Eo-Pacific has been inferred from palaeogeographic data of Henderson et al. (1984). b: Revised
map of the Earth at the end of the Jurassic 140 Ma BP, when Gondwanaland broke up. Modified from the transverse
H a m m e r - A i t o f f maps of Figs. 5g and 9a on the basis of Table II and arguments discussed in the text. N o t e that a
141
160 180
/
_0
lap in r e f i t
~verlap in r e f i t
l ~ Middle
helf
(140-1
Pre-Mi :ontlnent
(2180 [. . . . . .
distinction has been made between oceanic lithosphere 140-180 Ma old and that older than 180 Ma. Two triple
junctions in the Eo-Pacific separate the following plates: Kula-Izanagi plate (bottom of the ellipse), Pacific plate (east
of Australia), Phoenix plate (adjacent to Antarctica), and the Farrallon plate (American west coast). See also
Appendix 3.
142
160 180
-0
Fig. 10. a: Base map of Earth 95 Ma BP showing the palaeogeographic position of the continents and the remnants of
95-180 Ma old ocean floor still preserved today. The location of the 95-180 Ma old remnant of the Pacific plate in the
143
160 180
\ /
-0
ileP in r e f i t
O C E A N I C CRU
c~verlap in r e f i t
Lower
(95-1, ENTAL CRUST
Middle ~helf
(140-'
Pre-Mi ,ontlnent
(>180
Eo-Pacific has been inferred from the palaeogeographic data of Henderson et al. (1984). b: Revised map of the Earth
at the end of the Lower Cretaceous 95 Ma BP. Modified from Figs. 4g and 10a on the basis of Table II and arguments
discussed in the text.
144
160 180
\ /
-0
Fig. 11. a: Base map of Earth 65 Ma BP showing the palaeogeographic position of the continents and the remnants of
65-180 Ma old ocean floor still preserved today. The location of the 65-180 Ma old remnant of the Pacific plate in
145
160
\
~0
-0
OCEANIC CRI.
tap in r e f i t
~ Upper
165-9!
)verlap in r e f i t
Lower
195-1 ENTAL CRUST
Middle ;hell
(140-
~ ] re-M ~ontlnent
(>180 .....
the Eo-Pacific has been inferred from Henderson et al. (1984) and Engebretson et al. (1984). b: Revised map of the
Earth at the end of the Cretaceous 65 Ma BP. Modified from Figs. 3g and 11 on the basis of Table II and the
arguments discussed in the text.
146
160 180
\ /
-0
Fig. 12. a: Base map of Earth 20 Ma BP showing the palaeogeographic position of the continents and the remnants of
20-180 Ma old ocean floor still preserved today. The location of the Pacific spreading pattern is obtained by keeping a
rigid connection with the Antarctic ridge and agrees with that suggested by Whitman et al. (1983). b: Revised map of
147
160 ~Bo
-0
O C E A N I C CRtJ
Pelaeq
(20-6 gap in r e f i t
~ pper
165-9
o v e r l a p In r e f i t
Lower
(95-1 IENTAL CRUST
Middl( shelf
(140-
Pre-M continent
(>180 .o .~,
the Earth at the end of the Palaeogene 20 Ma BP. Modified from Figs. 2g and 12a on the basis of Table II and
arguments discussed in the text. Note that the Ninetyeast Ridge between Australia and India ceased to spread at this
stage so that both India and Australia become part of a single plate (i.e., the Indian plate) as is portrayed on the map
of the modern Earth in Fig. 1.
148
Tethys by rifting of Pangea into Gondwana- The ridges in the Eo-Pacific of 140 Ma BP
land and Laurasia at 180 Ma BP (Fig. 8) (Fig. 9b) have been reconstructed on the basis
imply the former existence of a Tethyan of the estimated 4 . 1 0 4 km ridge length and
spreading ridge (Fig. 9b). The fact that their locations take into account previous
Gondwanaland broke up 140 Ma BP which suggestions of palaeo-ridges. The meridian
later caused separation between India and length across the Eo-Pacific of about 2 . 1 0 4
Australia, suggests that the Tethyan Ridge km, suggests that there was space for two
had a triple junction with the proto-Nine- parallel spreading ridges across the entire
tyeast Ridge extending southward (Fig. 9b). width of the Eo-Pacific. However, as many
Although the area occupied by the 140-180 previous workers have argued (see Appendix
Ma old ocean floor in Tethys is somewhat 3), the detailed spreading pattern of the rem-
arbitrary (Fig. 9b), the majority of such ocean nant of the Pacific plate on Fig. 9a suggests
floor must have occurred somewhere in the the occurrence of three triple junctions. Com-
Eo-Pacific, because about one-fourth of its bination of these two constraints suggests the
total area of 140.10 6 km 2 (Table II) could geometric solution in Fig. 9b as the most
find space in the Tethys. Again, the problem simple. In particular, the ridges suggested by
is to decide where the remaining 108 km z of Hilde et al. (1977) were helpful in the recon-
140-180 Ma old ocean floor was exactly, or struction.
even approximately, located. The Tethyan and The spreading pattern in the Eo-Pacific at
proto-Ninetyeast ridges may only account for 140 Ma BP (Fig. 9b) is constrained by the
2.104 km length at most (30 cm on my fact that about half the Eo-Pacific's ocean
globe) (and possibly contributed less than the floor must have consisted of 140-180 Ma old
mean production rate), whereas the repre- ocean floor with the rest older (see above).
sentative worldwide ridge length has been Most of the transform fault pattern, of course,
estimated at 6 • 10 4 km (assumption 8). This is entirely hypothetical and only introduced
suggests that at least 4- 10 4 km ridge length to give the map of Fig. 9b a realistic ap-
(60 cm on my globes) may have occurred in pearance. However, the large transform fault
the Eo-Pacific. indicated between Australia and the Banda
T A B L E II
Data used to revise the reconstructions in Figs. 2 - 5 as illustrated in Figs. 9 b - 1 2 b
Legend:
Section (a). Lithosphere production during each particular time interval on the basis of a constant production rate of
3.5 km2 a 1.
Section (b). Surface area of oceanic lithosphere missing on the maps of Figs. 2-5.
Section (¢). Same as in section (b), but converted to scale model dimensions (1 cm 2 = (637) 2 km2).
149
Arc is necessary to decouple the floors of the den oceanic lithosphere which is older than
Eo-Pacific and Tethys oceans, in order to 180 Ma (Figs. 9b-12b). Hence this construc-
guide and allow the relatively fast subduction tion predicts the occurrence of ophiolites older
of the Pacific spreading patterns. than 180 Ma in the Himalayas and possibly
Subduction along the margins of Tethys in the tectonic melange and thrust zones of
was initiated by at least 140 Ma BP, because Timor and New Guinea (cf. Hamilton, 1979).
Tethys has been shrinking ever since (Figs. The evolution maps (Figs. 8 and 9b-12b) also
1-5). The polarity of the subduction has been predict the occurrence of ophiolites older than
northward, as can be inferred from geological 180 Ma in the Andes and Rockies. For exam-
field observations in the Alps (Frisch, 1979) ple, the extensive Nicoya ophiolite complex in
and the Hellenic-Cyprean trenches (Lister et the northwest part of the Andes (Bourgois et
al., 1984). This has now been independently al., 1984) might comprise such old ocean floor.
confirmed by large-scale tomographic The constructions also show why ocean floor
analyses of body waves in the upper mantle older than 180 Ma BP has not been obducted
beneath Europe, the Mediterranean and the in the Mediterranean suture of Tethys, be-
Middle-East. These reveal that a northward cause Tethys had not yet opened by 180 Ma
dipping zone of cold (oceanic) lithosphere is BP (Fig. 8).
still present beneath the entire length of the
Alpine orogenic belt (Spakman, 1986; S P R E A D I N G A N D S U B D U C T I O N VELOCITIES
Meulenkamp et al., 1988). The Tethyan Ridge
is therefore being subducted northward under Table III shows how the assumed produc-
Europe on the map of 95 Ma BP (Fig. 10b). tion rate of 3.5 km 2 a-1 (Table II) results in a
This age can be inferred from the fact that revised subduction budget. The areal litho-
the only ophiolites younger than 95 Ma BP spheric production and subduction rates given
found in the Mediterranean suture of Tethys in Table I can be regarded as the slowest
are the 110-84 Ma old Arabian ophiolites possible (Figs. 1-6), whereas Table III gives a
which can be attributed to local spreading budget which may be more realistic, or per-
(Knipper et al., 1986). haps a fast extreme (Figs. 9-12). The produc-
The progressive evolution of the spreading tion and subduction rates in Tables I and III
patterns in both the Tethys and Eo-Pacific can both be used to obtain estimates of
are indicated in Figs. 9b-12b, which are mod- spreading and subduction velocities (see later).
ified versions of Figs. 2-5. The base maps
used are indicated in Figs. 9a-12a. These World-wide spreading velocities
show the relevant spreading patterns still visi-
ble on the Earth today, and the fit of the Spreading velocities (VL) can be obtained
continents along these extended passive an- from the production rate of lithospheric
cient outlines. The palaeogeographic posi- surface area (A) and the instantaneous length
tions of the ancient fragments of the Pacific (L) of the spreading ridge system:
plate have been inferred from Henderson et
al. (1984), and Whitman et al. (1983). Note Vc = A / ( 2 L ) (1)
that the incompleteness of preserved ocean The factor 1/2 is involved since the spreading
floor (Figs. 9a-10a) cast suspicion on in- is in two opposite directions and the spread-
ferences, like those made by Larson and Pit- ing velocity VL is given relative to the ridge.
man (1972), that there was a period of super- The mean world-wide ridge lengths for the
fast spreading, worM-wide, between 110 and Neogene (0-20 Ma BP), Palaeogene (20-65
85 Ma BP. Ma BP), Upper Cretaceous(65-95 Ma BP)
Note also that both the Indian and the and Lower Cretaceous (95-140 Ma BP) have
Australian plates are likely to have overrid- been estimated from the globes in Figs. 1-4,
150
TABLE III
Analysis of plate activity since the breakup of Pangea 180 Ma BP, corresponding to the revised reconstructions of Figs
8-11
Legend:
Section (a). Lithosphere production for each particular time interval on the basis of a constant production rate of 3.5
km 2 a -1
Section (b). Area still present on the modern Earth.
Section (c). Subducted lithosphere inferred from the difference between the estimates in sections (a) and (b).
Section (d). Spreading velocities estimated by applying expression 1 in the text, after measuring the ridge length on the
reconstructions of Figs. 8-11.
respectively. Table I (section d) shows these spreading velocities obtained by applying ex-
ridge lengths together with spreading veloci- pression 1. Comparison of the spreading
ties obtained by applying expression 1. These velocities in Tables I and III (sections d)
spreading velocities are based not only on reveals that these alternative approaches both
minimum estimates of A, but also on con- result in spreading velocities which may have
servative ridge lengths. This may partly ex- varied only little about 3 cm a-1 during the
plain why my world-wide mean for the mod- past 180 Ma.
ern spreading velocity of 3 cm a -1 differs It may be worthwhile to emphasize that to
from the 2.5 cm a -1 estimated by Parsons express plate velocities in cm a -1 is strictly
(1981). Note also that the total ridge length of speaking incorrect, although useful because
58.6- 103 km on the modern Earth estimated its real meaning is easier to envisage than the
here is slightly different from the 59.2.103 physically more correct expression using km
km previously suggested by Forsyth and Ma-1. I prefer to quote imaginary velocities,
Uyeda (1975) before a detailed map of the although plate velocities might not be con-
oceanic spreading pattern became available tinuous on the time scale of years (see Ap-
(Sclater et al., 1981). pendix 4).
The revised maps (Figs. 8 and 9b-12b) and
production rates (Table II) may be used to WorM-wide subduction velocities
make alternative estimates for the spreading
velocities. Table III (section d) shows the Subduction velocities ( ~ ) can be obtained
ridge lengths estimated from the revised re- from the subduction rate of lithospheric area
constructions (Figs. 8 and 9b-12b) and the (A) and the instantaneous length (S) of sub-
151
TABLE IV
Subduction budgets implied by the global reconstructions of Figs. 1-6 and Figs. 8-11
(a) Area of lithosphere (b) Mean areal (c) Mean length (d) Mean
subducted per tectonic subduction of subduction velocity
domain over past 180 Ma rate zones (cm a - l )
model Earth (%) (km 2 a l) model Earth
(cm2) (106 km 2) (cm) (103 km)
Figs. 1-6
(I) Tethys closure 134 54.4 13 0.3 27 17 1.8
(II) Circum-Pacific
Trenches 896 363.6 87 2.0 60 38.2 5.2
(II1) Total 1030 418 100 2.3 87 55.2 4.2
Figs. 8-12
(I) Tethys closure 233 94.6 15 0.5 27 17 2.9
(II) Circum-Pacific
Trenches 1322 536.4 85 3.0 60 38.2 7.9
(III) Total 1555 631 100 3.5 87 55.2 6.3
Legend:
Section (a). (I) The total area of oceanic lithosphere subducted in the Tethys Ocean, measured on the globes of Figs. 5
and 6 (upper half of the table) and Figs. 8-11 (lower half of the table) during the past 180 Ma. (II) The oceanic
lithosphere which disappeared along the Circum-Pacific Trenches can be estimated by subtracting the total amount of
oceanic lithosphere subducted in the Tethys closure from the total area of oceanic lithosphere subducted since 180 Ma
BP (given in Table I, section (b) or Table 3, section (c)).
Section (b). Mean subduction rates during the past 180 Ma, obtained by dividing the areal estimates in section (a) by
180 Ma.
Section (c). Length of subduction zones remained approximately constant during the past 140 Ma and were measured
on the global models of Figs. 1-6 (upper half of the table) and Figs. 8-11 (lower half of the table).
Section (d). The subduction rate of the oceanic surface area can be converted to absolute subduction velocities by
applying expression 2 in the text.
152
given in Table IV (section c). Alternative sub- lowed very little speculation. The snug fit
duction scenarios can be considered. The sub- between past plate boundaries on model
duction in Figs. 1-6 suggests that only 13% of globes of constant radii is in contradiction to
the subducted ocean floor disappeared in the the recent suggestion that this would only be
Tethys Ocean, whereas 87% was subducted in possible on an expanding Earth (Owen, 1983;
the Circum-Pacific Trenches (Table IV, lower see for further discussion Appendix 1 and
part of section a). Figs. 8-11 suggest respec- Weijermars, 1986).
tive percentages of 15% and 85% (Table IV, Calculations of the budget of lithospheric
lower part of section a). The inferred subduc- spreading and subduction from these recon-
tion rate of lithospheric area (Table IV, sec- structions have been specified in Table I and
tion b) has been used to estimate the respec- the upper part of Table IV. The average pro-
tive subduction velocities (Table IV, section duction and subduction rates are estimated to
d). The estimates of world-wide mean sub- have been 2.3 km 2 a 1. World-wide spreading
duction velocities since the breakup of Pan- velocities have varied between 2.4 and 3.5 cm
gea 180 Ma BP vary between 4.2 and 6.3 cm a-1 since the breakup of Pangea at 180 Ma
a 1. The subduction velocities in Tethys may BP. The world-wide subduction velocity is
have been 1.8-2.9 cm a -1 and that in the found to have been 6.3 cm a -1 during the
Circum-Pacific Trenches 5.2-7.9 cm a 1. past 20 Ma, whereas a mean velocity of 4.2
cm a I may be calculated for the past 180
CONCLUSIONS Ma. Subduction velocities for the Tethyan
Trenches (1.8 cm a 1) appear to have been
The oceanic spreading pattern on the mod- much lower than those for the Circum-Pacific
ern Earth (Fig. 1) was used to reconstruct the Trenches (5.2 cm a - l ) .
dispersal of continents and tectonic plates Lithospheric budget calculations based on
since the breakup of Pangea 180 million years the reconstruction of Figs. 1-6 provided a
ago. The completeness of the spreading pat- sound basis for suggesting revisions of the
tern which lies between passive continental spreading pattern history. Fig. 7 shows how
margins makes the refit of such margins far the production (and therefore the subduction)
easier than refitting those interfering with rates of oceanic lithosphere consistently in-
subduction trenches. Large parts of the an- creased during at least the past 140 Ma if the
cient outlines of tectonic plates in the Eo- Figs. 1-6 are entirely correct. It was then
Pacific and Tethys oceans, are subducted with suggested that it may be more realistic, given
the plates and lost to direct study. Several the poor constraints of the past spreading
assumptions are necessary to allow logical patterns in the Tethys end Eo-Pacific oceans,
reconstructions. It was therefore tentatively to assume a constant production rate of 3.5
assumed that spreading patterns have always km 2 a - l , and see if their past spreading pat-
been generated symmetrically about spread- terns could be adjusted to comply with this.
ing ridges and that no ridges other than linear The assumption of a constant production
extensions of those still present today have rate also allowed a distinction between ocean
disappeared since the breakup of Pangea 180 floor of 140-180 Ma and older, on the revised
Ma BP. For the sake of simplicity and in view maps of Figs. 8 and 9b-12b. These maps
of the scale of this study the complex micro- further differ from those in Figs. 2-6 in that
plate structure of the Mediterranean Collision new constraints were used to restore the
Zone (cf. Weijermars, 1987b, 1988a) was ne- former spreading patterns of the floors of the
glected. Tethys and the Eo-Pacific oceans. These con-
The maps of Figs. 1-6 visualize a careful straints are: (1) a particular surface area of
reconstruction of past plate positions and lost oceanic lithosphere implied by the adop-
spreading patterns, and the assumptions al- tion of a constant production rate (Table II);
153
(2) a representative ridge length of 6 - 1 0 4 would imply a shift of Antarctica over the
km; (3) the opening of Tethys when Pangea underlying mantle.
broke up at 180 Ma BP (Fig. 8); (4) the The construction of the original globes used
obduction of 185-84 Ma old Tethyan Ocean in this study has been outlined elsewhere
floor (Knipper et al., 1986) most, if not all of (Weijermars, 1986). Alternatively, the spread-
which has since been subducted northward ing data discussed here can be transferred to
under Eurasia (Spakman, 1986); and (5) pre- flat sheet projections of polyhedral approxi-
vious estimates of ancient ridge locations in mations to a sphere according to a method
the Eo-Pacific (Appendix 3). The direct geo- outlined by Dutch (1985). Such sheets can be
logical implications of Figs. 8 and 9b-12b are folded and quickly assembled to form a pseu-
the possible occurrences of ophiolites older d o g l o b e for assessing the global spreading
than 180 Ma in the Cenozoic collision zones pattern in three dimensions. These could be
of the Himalayas, Andes, Rockies, and the made commercially available at costs much
western part of the Banda Arc (Timor, New lower than that of true globes.
Guinea).
The budget of production and subduction
of lithosphere calculated from the revised
maps of Figs. 8-12 are given in Table III and
ACKNOWLEDGEMENTS
the lower part of Table IV, respectively. The
mean world-wide production and subduction
rates were fixed at 3.5 km 2 a -a, using data in The preparation of this article would not
Table I. World-wide spreading velocities have been possible without the support of
would then have varied between 2.6 and 3.2 Mrs. Christina WernstriSm, the extremely
cm a -1, whilst the mean subduction velocity skillful draftswoman at the Geological In-
would have been 6.3 cm a -1 over the past 180 stitute of Uppsala University. She started
Ma. The mean long-term subduction veloci- drafting in 1983 and transformed the colour
ties for the Tethyan and Circum-Pacific maps of the author into neatly annotated
trenches being 2.9 and 7.9 cm a -a, respec- drawings. She drew all the ornaments by hand
tively. in order to achieve optimum visual contrast
The most complete spreading patterns ob- between the various types of lithosphere dis-
tained here are presented in the H a m m e r - tinguished.
Aitoff maps of Figs. 1, 8 and 9b-12b. It Warm thanks are also due to professor
should be emphasized that all the maps in Hans Annersten, who gave his generous and
this study aim to be palaeogeographically unreserved support when the project started
scaled and not only reconstructions of rela- to yield potentially useful results. Mr. Sandy
tive plate positions. The geographic grid has Cruden of Uppsala University and Earth Sci-
been kept fixed relative to Antarctica for ence Reviews editors professor K. Kobayashi,
reasons discussed in detail. The maps thus K. Tamaki and a third anonymous referee are
obtained provide useful bases to plot thanked for very useful comments on the
palaeoclimatic data, and eventually, if neces- manuscript. Finally, but certainly not least I
sary, readjust the position of the (palaeo-) greatfully acknowledge inspiring discussions
geographic poles. Until now, there are no with professor C.J. Talbot, Uppsala. I thank
conclusive data concerning these positions. him for being so liberal in allowing me this
Nonetheless, I invite palaeoclimatic experts to and other episodic deviations from me Ph.D
find out if Antarctica has moved significantly study on thermal convection (1983-1987). The
relative to the geographic pole over the past author was funded by the Swedish Natural
180 Ma. My chain of arguments (see section Science Research Council (NFR) under grant
on palaeogeographic grid) shows how this no G-DT 1858-101.
154
APPENDIX 1--COMPARISON OF THE PRESENT the older oceanic lithosphere on the eastern side of the
MAPS WITH THOSE OF OWEN (1983) BASED ON island arcs, just as Owen (1983) did in his maps 48, 50,
A FAST EARTH EXPANSION HYPOTHESIS 51, and 53 of the expanding Earth. The gaps on Owen's
maps 46, 49 and 52 near Panama and in the Bering Sea
Owen (1983) claims that a close fit between past occur on my reconstructions as well (Figs. 3g-6g).
plate boundaries is only possible if the Earth's radius The gap which occupies a very large area in Owen's
was 80% of its modern value at about 200 Ma BP. He (1983) maps (i.e., maps 19, 22, 29, 32, 36, 39, 42, 46, 49
presents maps which suggest a linear increase of Earth's and 52) is the Tethys Ocean. For some unstated reason
radius with time such that this radius on plate tectonic Owen (1983) does not appear to accept that the bottom
maps for 140 Ma BP is about 85% of the modern value. of the Tethys Ocean was no gap but consisted of
Reconstructions of Pangea (200 Ma BP, or better 180 oceanic lithosphere. The opening by spreading and sub-
Ma BP) and Gondwanaland (140 Ma BP) on an Earth sequent closure of Tethys has been reconstructed here
of constant radius should therefore reveal gaps which on transverse Hammer-Aitoff maps (Figs. 1, 8 and
occupy 37 and 28% of the Earth's surface, respectively, 9b-12b). These suggest that any Earth expansion is
if Owen's (1983) expansion hypothesis is correct. unlikely during the past 180 Ma (cf. Weijermars, 1986).
My globes and the derived maps (Figs. 2-6) demon-
strate that, in contradiction to Owen (1983), there is a
remarkable close fit between all the plate boundaries of APPENDIX 2--PREFERRED FIT OF GOND-
the particular ages considered here. In my simple recon- WANALAND
struction, gaps of 0.4. 2, 2, 1.5 and 2.6% of the total
surface area of the Earth occurred in the reconstruc- Gondwanaland comprised Antarctica, Africa, South
tions of 20, 65, 95, 140 and 180 Ma BP, respectively. America, India, Australia and Madagascar (cf. Suess,
These gaps (indicated in black) occur between North 1885). Although there are a number of criteria which
and South America (Figs. 2a-6a) and along the could be used to constrain the detailed fit of the conti-
Antarctic Ridge (Figs. 4c-4e), the Indian Ridge (Fig. nents in Gondwanaland, existing models show a great
3d), the North Atlantic Ridge (Fig. 3a), and the variety. This fact has been recently emphasized in re-
Azores-Gibraltar Fault (Figs. 2a-4a). Significant over- views of the various Gondwanalands previously pro-
laps occurred only in the reconstructions of 65, 95, 140 posed (Barron et al., 1978; Powell et al., 1980; Martin
and 180 Ma BP and amounted to 0.5, 0.9, 0.8 and 0.6% and Hartnady, 1986). Fig. 13a and b show the two
of the total surface, respectively. Overlaps (indicated by major fits considered before: the west fit (Fig. 13a: Du
cross-hatching) occur near Greenland (Figs. 3a-6a and Toit, 1937; Smith and Hallam, 1970; Norton and Sclater,
3b-6b) and in the fit of Gondwanaland (Fig. 5, c and 1979) and the east fit (Fig. 13b; Tarling, 1972; Powell et
e). al., 1980; Barron and Harrison, 1980).
Similar but minor gaps which approximately corre- Fig. 13, d and e outline the successive positions of
spond to those I found on my globes (Figs. 2-6) appear the southern continents for 0, 20, 65, 95 and 140 Ma BP
on Owen's (1983) maps 6, 9 and 12 (Arctic region), 29
and 32 (South Atlantic), 36, 39 and 42 (Indian Ocean)
and 46, 49 and 52 (Pacific Ocean). For example, the
gaps in the Arctic region appearing in Figs. 4g and 5g Fig. 13. Fits of the southern continents in Gondwana-
here and those in the North Atlantic near Labrador, the land. a: West fit (e.g., Du Toit, 1937; Smith and Hal-
Azores and Panama (Figs. 2g and 3g) correspond in size lam, 1970; Norton and Sclater, 1979). b: East fit (e.g.,
and location to the gaps on Owen's (1983) maps 16 and Tarling, 1972; Barron and Harrison, 1980). c: New fit
19. The misfit between Africa and South America in of the southern continents 140 Ma BP proposed here on
Owen's (1983) map 32 contradicts the famous recon- the basis of the likelihood of smooth plate travel paths
struction of Bullard et al. (1965), but in either case is so since the breakup of Gondwanaland. The grid used in a
small that it is hardly significant. and b is the central part of a transverse Hammer-Aitoff
However, the large gaps in the Indian Ocean near projection (i.e., the Southern Hemisphere), which en-
the Ninetyeast Ridge (Owen, 1983, map 36) do not ables direct comparison with the other maps in this
appear if instead a relatively small gap is allowed to study. Gaps in the refit are black and overlaps are
remain between the Australian and Antarctic continen- cross-hatched, d-f: Transverse Hammer-Aitoff equal
tal margins (see Fig. 3g). The misfit in Owen's (1983) area projections showing the displacement of the con-
maps 46, 49 and 52 of the Pacific Ocean on an Earth of tinental outlines relative to Antarctica which is here
constant radius is negligible. The back-arc basins along considered to have been stationary since the breakup of
the West-Pacific Continental Margin are indicated as Gondwanaland 140 Ma BP. Centre of the originally
gaps on all three of these particular maps of Owen elliptical maps is at 0 o longitude and 70 o southern
(1983). This could be avoided if the back-arc basins are latitude, but the lower part of the ellipses has been cut
closed and compensated for by increasing the area of away along the equator to preserve space.
155
-20 0 +20
- ~0 60
, ,/' \ /
-8-060 -40 -20 0 +20 +4.0 +60+80 ,. -60-4.0 -20 0 +20 +40 +60+80
I _ _.:.,
, / . / / - i\\ -i\"
I] ~,11
. . . . . .i l. L. .:.-.- - - - - - ~ - " P
:::; ;' ; ~',:,'-..i
i i ~----~ = \\ .... ,;,';:-~----------..-" ;','<:'~"~" -- \
/~ ii-444+4
J - ~ ~ 1 4 0 Ma BP ~ -- "
180 ~ 180
156
was much smaller than today and at 180 Ma BP oc- the Galapagos spreading centre at about 33 Ma BP, but
cupied an area approximately outlined by the present in a complex fashion. The northern part was therefore
remnant of 180-140 Ma old ocean floor in the West termed the Guadelupe plate until about 17-12 Ma BP,
Pacific east of the Mariana Trench (Fig. 1). Hayes and when the northern part of the Guadelupe plate gradu-
Pitman (1970) first suggested the existence of a triple ally disappeared under California and a new triple
ridge junction in the Mesozoic northwest Pacific, sep- junction formed near the Baja California Peninsula (cf.
arating the Pacific plate in the west from the Kula plate Hagstrum et al., 1985). The surviving southern part of
in the north and the Farallon plate in the east as was the Guadelupe plate is the present day Cocos plate
adopted here (Fig. 9b). (Menard, 1978). The southern part of the (South) Faral-
The Izanagi plate has been suggested as a 120-180 Ion (or F1) plate is termed the Nazca plate since the
Ma BP precursor of the Kula plate (Woods and Davies, formation of the Guadelupe plate by propagation of the
1982); Rea and Dixon (1983) renamed the same plate Galapagos Rift about 30 Ma BP (Menard, 1978).
the Bering plate. The Kula plate itself would have
formed by the propagation of a rift through the north- APPENDIX 4 - - E X P R E S S I O N OF PLATE VELOCI-
ern part of the Farallon plate, but not until the Bering TIES
plate was completely subducted (Rea and Dixon, 1983).
A detailed reconstruction of the absolute palaeogeo- The world-wide mean spreading velocity should be
graphic locations of the Pacific-Farallon-Kula (Izanagi) formally expressed as 30 km Ma-~ rather than 3 cm
triple junction has recently been published (Henderson a -1, because the motion may be either periodic or
et al., 1984; Engebretson et al., 1984, 1985). episodic rather than continuous on time scales of years
A second triple junction in the South Pacific (cf. Fig. (which is much shorter than the Maxwell relaxation
9b) could have existed between 180 and 65 Ma BP and time for viscoelastic plates). Geodetic measurements
separated the southwestern Phoenix plate from the across the propagating Red Sea Rift and observations
northern Farallon and Western Pacific plates (Larson along the Atlantic Ridge near Krafla on Iceland, dem-
and Chase, 1972; Larson and Pitman, 1972). Yet another onstrate that the walls of these rifts may suddenly be
triple junction, even further south along the Pacific separated by several metres in connection with fresh
Ridge (Fig. 9b), may have separated the Pacific and dyke intrusions parallel to the rift axes (Goguel, 1983).
Phoenix plates from the Indian-Australian-Antarctic The latter author also suggested that such intermittent
plate (Hilde et al., 1977). On the basis of palaeomagnetic spreading " m a y be explained by elastic strain on the
evidence given earlier (Larson and Chase, 1972), the sides of the crack".
latter authors also show the primitive Pacific plate in its More specifically, intermittent spreading of ridges
135 Ma BP position 4500 km to the south of where the could be understood by an adaption of the stick-slip
180-140 Ma old portion now lies (Fig. 9b). model used for explaining periodic motions along major
The younger evolution of the east Pacific seems very transcurrent faults. The classical stick-slip model as-
complex. The Farallon plate may have broken up, at sumes that elastic simple shear displacement in the
about 55 Ma BP, into a northern Vancouver plate and a walls of transcurrent faults is periodically interrupted
southern part which was still termed the Farallon plate by the release of stored elastic energy as sudden slip
(Menard, 1978). A breakup of the FaraUon plate would motion along the fault plane which causes earthquakes
already have occurred at 100 Ma BP according to Rea (cf. Turcotte and Schubert, 1982). A variant suggested
and Dixon (1983; cf. Rea and Duncan, 1986), who here is that plate-driving forces cause elastic dilation of
termed the equivalent of the Vancouver plate simply the an active, but temporarily seismically quiet, shallow
North Farallon plate and the southern fragment the zone in the ridge until stresses are high enough to cause
South Farallon plate. They apparently missed Menard's brittle failure along the ridge axis, producing earth-
(1978) paper. Similarly, Whitman et al. (1983) simply quakes and allowing intrusion of fresh dykes from the
termed the northern and southern Farallon plates F 2 pockets of partial melt deep under the ridge axis. The
and F 1, respectively, but these would not have been solidification of the dykes may be an efficient crack-
formed until about 50 Ma BP which agrees with sealing mechanism and correspond to the stick-stage of
Menard's (1978) scenario. the strike-slip equivalent of this model. Note that this
The Vancouver (or North Farallon or F2) plate may mechanism is quite different from the steady state ridge
have changed name, but is really the Juan de Fuca plate propagation model suggested by Morgan and Parmen-
from the moment its continuation with the (South) tier (1985).
FaraUon (or F1) plate via the East Pacific Ridge was It may be worthwhile to emphasize that the lack of
subducted at about 25-30 Ma BP (Riddihough, 1984; inertia has nothing to do with discontinuities in plate
cf. Barrash and Venkatakrishnan, 1982). Menard's motions as suggested by Goguel (1983). Inertia is in-
(1978) scenario for the subsequent evolution of the deed negligible in mantle convection and any other
(South) Farallon (or F1) plate suggests the formation of solid state creep of rocks because it is damped by their
158
extremely high viscosity (Weijermars and Schmeling, Barton, P. and Wood, R., 1984. Tectonic evolution of
1986; Weijermars, 1987a, 1988b, c). In other words, the North Sea Basin: Crustal stretching and subsi-
although the kinetic energy of solid rock deformation dence. Geophys. J.R. Astron. Soc., 79: 987-1022.
may involve inertia during fracture and earthquake Berggren, W.A., Kent, D.V., Flynn, J.J. and Van
occurrences, it is always negligibly small in plate mo- Couvering, J.A., 1985. Cenozoic geochronology. Geol.
tions by ductile flow. This was earlier recognised by Soc. Am. Bull., 96: 1407-1418.
Goguel (1983) who therefore noted that the term plate Besse, J., Courtillot, V., Pozzi, J.P., Westphal, M. and
collision is somewhat misleading. In reality, plate diver- Zhou, Y.X., 1984. Palaeomagnetic estimates of crustal
gence or convergence would instantaneously stop shortening in the Himalayan thrusts and Zangbo
without causing any further internal plate deformation suture. Nature, 311: 621-626.
(neglecting elastic effects) if the driving force of plate Boucot, A.J. and Gray, J., 1983. A Paleozoic Pangea.
motions were to cease spontaneously. However, to im- Science, 222: 571-581.
ply that lack of inertia could therefore explain intermit- Bourgois, J., Azemy, J., Baumgartner, P.O., Tournon, J.,
tent plate motions is incorrect because the driving forces Desmet, A. and Auboin, J., 1984. The geologic his-
of such motions themselves cannot change instanta- tory of the Carribean-Cocos plate boundary with
neously. Dynamically scaled convection experiments special reference to the Nicoya ophiolite complex
show that convection patterns driving tectonic plates (Costa Rica) and D.S.D.P. results (Legs 67 and 84
cannot reorganize instantaneously, but may change their off Guatemala): a synthesis. Tectonophysics, 108:
geometry gradually over time scales of at least 50 Ma 1-32.
and only when boundary conditions are changed Brown, G.C. and Mussett, A.E., 198l. The Inaccessible
(Richter and Parsons, 1975). Earth. Allen and Unwin, London, 235 pp.
Bullard, E.C., Everett, J.E. and Smith, A.G., 1965. The
fit of the continents around the Atlantic. Phil. Trans.
REFERENCES R. Soc. London, A 258: 41-51.
Burke, K. and Dewey, J., 1974. Two plates in Africa
Allrgre, C.J. and Jaupart, C., 1985. Continental tecton- during the Cretaceous? Nature, 249: 313-316.
ics and continental kinetics. Earth Planet. Sci. Lett., Busse, F.H., 1975. A model of the geodynamo. Geo-
74 : 171-181. phys. J., 42: 437-459.
Allenby, R.J., 1979. Implications of very long baseline Carey, S.W., 1976. The Expanding Earth. Elsevier,
interferometry measurements on North American in- Amsterdam, 488 pp.
tra-plate crustal deformation. Tectonophysics, 60: Chase, C.G., 1972. The N plate problem of plate tecton-
T27-T35. ics. Geophys. J.R. Astron. Soc., 29: 117-122.
Andrews, J.A., 1985. True polar wander: an analysis of Cogley, K.G., 1984. Continental margins and the extent
Cenozoic and Mesozoic palaeomagnetic poles. J. and number of the continents. Rev. Geophys. Space
Geophys. Res., 90: 7737-7750. Phys., 22: 101-122.
Aubouin, J., Bourgois, J. and Azrma, J., 1984. A new Condie, K., 1981. Plate Tectonics and Crustal Evolu-
type of active margin: the convergent-extensional tion. Pergamon Press, Oxford, 2nd ed., 280 pp.
margin, as exemplified by the Middle America Cronin, V.S., 1987, Cycloid kinematics of relative plate
Trench off Guatemala. Earth Planet. Sci. Lett., 67: motion. Geology, 15: 1006-1009.
211-218. Deffeyes, K.S., 1970. The axial valley. A steady-state
Bally, A.W., 1981. Thoughts on the tectonics of folded feature of the terrain. In: H. Johnson and B.L. Smith
belts. In: K.R. McClay and N.J. Price (Editors), (Editors), The Megatectonics of Continents and Oc-
Thrust and Nappe Tectonics. Spec. Publication Geol. eans. Rutgers University Press, USA, pp. 194-200.
Soc. London, 9: 13-22. Dewey, J.F., Pitman, W.C., Ryan, W.B.F. and Bonnin,
Barrash, W. and Venkatakrishnan, R., 1982. Timing of J., 1973. Plate tectonics and the evolution of the
late Cenozoic volcanic and tectonic events along the Alpine system. Geol. Soc. Am. Bull., 84: 3137-3180.
western margin of the North American plate. Geol. Dietrich, G. and Ulrich, J., 1968. Atlas zur Ozeanogra-
Soc. Am. Bull., 93: 977-989. phie. Bibliographisches Institut, Mannheim, 103 pp.
Barron, E.J. and Harrison, C.G.A., 1980. An analysis of Duncan, R.A., 1981. Hotspots in southern oceans: an
past plate motions: the South Atlantic and Indian absolute frame of reference for motion of the
Oceans. In: Mechanisms of Continental Drift and Gondwana continents. Tectonophysics, 7: 29-42.
Plate Tectonics. P. Davies and S.K. Runcorn (Edi- Durham, J.W., 1985. Movement of the Carribean plate
tors), Academic Press, London, pp. 89-109. and its importance for biogeography in the Car-
Barron, E.J., Harrison, C.G.A. and Hay, W.W., 1978. A ribean. Geology, 13: 123-125.
revised reconstruction of the southern continents. Dutch, S.I., 1985. Pseudoglobes for the earth-science
EOS Trans. AGU, 59: 436-449. classroom. J. Geol. Educ., 33: 128-131.
159
Du Toit, A.L., 1937. Our Wandering Continents. Oliver Hayes, D.E. and Pitman llI, W.C., 1970. Magnetic
and Boyd, London, 366 pp. lineations in the North Pacific. Geol. Soc. Am. Mem.,
Engebretson, D.C., Cox, A. and Gordon, R.G., 1984. 126: 291-314.
Relative motions between oceanic plates of the Helwig, J., 1976. Shortening of continental crust in
Pacific basin. J. Geophys. Res., 89: 10291-10310. orogenic belts and plate tectonics. Nature, 260:
Engebretson, D.C., Cox, A. and Gordon, R.G., 1985. 768-770.
Relative motions between oceanic plates of the Henderson, L.J., Gordon, R.G. and Engebretson, D.C.,
Pacific basin. J. Geol. Soc. Am. Spec. Pap., 206: 1984. Mesozoic aseismic ridges on the Farallon plate
1-59. and southward migration of shallow subduction dur-
Enslin, H., 1978. Determination of the rotation of the ing the Laramide orogeny. Tectonophysics, 3:
Earth. In: P. Brosche and J. Sundermann (Editors), 121-132.
Tidal Friction and the Earth's Rotation. Springer, Hey, R.N., Naar, D.F., Kleinrock, M.C., Phipps Morgan,
Berlin, pp. 28-42. W.J., Morales, E. and Schilling, J.G., 1985. Micro-
Forsyth, D. and Uyeda, S., 1975. On the relative impor- plate tectonics along a superfast sea floor spreading
tance of driving forces of plate motion. Geophys. system near Easter Island. Nature, 317: 320-325.
J.R. Astron. Soc., 43: 163-200. Hilde, T.W.C. and Lee, C.S., 1984. Origin and evolution
Frei, L.S., 1986. Additional palaeomagnetic results from of the west Phillipine Basin: A new interpretation.
the Sierra Nevada: further constraints on basin and Tectonophysics, 102: 85-104.
range extension and northward displacement in the Hilde, T.W.C., Uyeda, S. and Kronenke, L., 1977. Drift
western United States. Geol. Soc. Am. Bull., 97: of the major continental blocks since the Devonian.
840-849. Nature, 270: 304-309.
Frisch, W., 1979. Tectonic progradation and plate Jordan, T.E., Isacks, B.L., Allmendinger, R.W., Brewer,
tectonic evolution of the Alps. Tectonophysics, 60: J.A., Ramos, V.A. and Ando, C.J., 1983. Andean
121-139. tectonics related to geometry of subducted Nazca
Garfunkel, Z., 1975. Growth, shrinking and long-term plate. Geol. Soc. Am. Bull., 94: 341-361.
evolution of plates and their implications for the Jordan, T.H., 1975. The present-day motions of the
flow pattern in the mantle. J. Geophys. Res., 80: Carribean plate. J. Geophys. Res., 80: 4433-4439.
4425 -4432. Jurdy, D.M. and Stefanick, M., 1983. Flow models for
Gill, E.D., 1961. The climates of Gondwanaland in back-arc spreading. Tectonophysics, 99: 191-206.
Kainozoic times. In: A.E.M. Nairn (Editor), De- Kent, D.V. and Gradstein, F.M., 1985. A Cretaceous
scriptive Palaeoclimatology. Interscience, New York, and Jurassic geochronology. Geol. Soc. Am. Bull.,
N.Y., pp. 332-353. 96; 1419-1427.
Goguel, J., 1983. A short note on continuity or discon- King, L.C., 1961. The Palaeoclimatology of Gondwana-
tinuity in the global tectonic plate velocities. land during the Palaeozoic and Mesozoic Eras. In:
Tectonophysics, 100: 1-4. A.E.M. Nairn (Editor), Descriptive Palaeoclimatol-
Gold, T., 1955. Instability of the Earth's axis of rota- ogy. Interscience, New York, N.Y., pp. 307-331.
tion. Nature, 175: 526-529. King, L.C., 1983. Wandering Continents and Spreading
Goldreich, P. and Toomre, A., 1969. Some remarks on Sea Floors on an Expanding Earth. Wiley, New
polar wandering, J. Geophys. Res., 74: 2555-2567. York, N.Y., 232 pp.
Gurnis, M. and Davies, G.F., 1986a. Apparent episodic Knipper, A., Ricou, L.-E. and Dercourt, J., 1986.
crustal growth arising from a smoothly evolving Ophiolites as indicators of the geodynamic evolution
mantle. Geology, 14: 396-399. of the Tethyan Ocean. Tectonophysics, 123: 213-240.
Gurnis, M. and Davies, G., 1986b. Mixing in numerical Knopoff, L. and Leeds, A., 1972. Lithospheric moments
models of mantle convection incorporating plate and the deceleration of the Earth. Nature, 237:
kinematics. J. Geophys. Res., 91: 6375-6395. 93-95.
Hagstrum, J.T., McWilliams, M., Howell, D.G. and KrSner, A., 1985. Evolution of the Archean continental
Grommr, S., 1985. Mesozoic paleomagnetism and crust. Ann. Rev. Earth Planet. Sci., 13: 49-74.
northward translation of the Baja California Lagabrielle, Y., Whitechurch, H., Marcoux, J., Juteau,
Peninsula. Geol. Soc. Am. Bull., 96: 1077-1090. T., Reuber, I., Guillocheau, F. and Gapan, U., 1986.
Hamilton, W., 1979. Tectonics of the Indonesian Re- Obduction-related ophiolitic polymict breccias
gion. U.S. Geol. Surv. Prof. Pap., 1078, 345 pp. covering the ophiolites of Atalya (southwestern
Harland, W.B., Cox, A.V., Llewellyn, P.G., Pickton, Turkey). Geology, 14: 734-737.
C.A.G., Smith, A.G. and Waiters, R., 1985. A Geo- Larson, R.L. and Chase, C.G., 1972. Late Mesozoic
logic Time Scale. Cambridge University Press, Cam- evolution of the Western Pacific Ocean. Geol. Soc.
bridge, 128 pp. Am. Bull., 83: 3627-3644.
160
Larson, R.L. and Pitman III, W.C., 1972. World-wide rounding regions. Geophys. J.R. Astron. Soc., 55:
correlation of Mesozoic magnetic anomalies, and its 217-254.
implications. Geol. Soc. Am. Bull., 83: 3645-3662. McLennan, S.M. and Taylor, S.R., 1983. Continental
Larson, R.L., Pitman III, W.C., Golovchenko, X., freeboard, sedimentation rates and growth of con-
Cande, S.C., Dewey, J.F., Haxby, W.F. and tinental crust. Nature, 306: 169-172.
LaBrecque, J.L., 1985. The Bed Rocky Geology of M~gard, F., 1984. The Andean orogenic period and its
the World, (a map with explanation). Freeman, New major structures in central and northern Peru. J.
York, N.Y. Geol. Soc. London, 141: 893-900.
Leighly, J.B., 1955. Aitoff and Hammer, Geograph. Menard, H.W., 1978. Fragmentation of the Farallon
Rev., 45: 246-249. plate by pivoting subduction. J. Geol., 86: 99-110.
Le Pichon, X. and Angelier, J., 1979. The Hellenic arc Menard, H.W., 1984. Evolution of ridges by asymmetri-
and trench system: a key to the neotectonic evolu- cal spreading. Geology, 12: 177-180.
tion of the eastern Mediterranean area. Tectonophy- Meulenkamp, J.E., Wortel, M.J.R., van Wamel, W.A.,
sics, 60: 1-42. Spakman, W. and Hoogerduyn, E., 1988. On the
Lister, G.S., Banga, G. and Feenstra, A., 1984. Meta- Hellenic subduction zone and the geodynamic evolu-
morphic core complexes of Cordilleran type in the tion of Crete since the late Middle Miocene.
Cyclades, Aegean Sea, Greece. Geology, 12: 221-225. Tectonophysics, 146: 203-215.
Lister, G.S., Etheridge, M.A. and Symonds, P.A., 1986. Minster, J.B. and Jordan, T.H., 1978. Present-day plate
Detachment faulting and the evolution of passive motions. J. Geophys. Res., 83: 5331-5354.
continental margins. Geology, 14: 246-250. Morgan, J.P. and Parmentier, E.M., 1985. Causes and
Livermore, R.A., Smith, A.G. and Vine, F.J., 1986. Late rate-limiting mechanisms of ridge propagation: a
Palaeozoic to early Mesozoic evolution of Pangea. fracture mechanism model. J. Geophys. Res., 90:
Nature, 322: 162-165. 8603-8612.
Lloyd, C.R., 1984. Pre-Pleistocene palaeoclimates: the Norton, I.O. and Sclater, J.G., 1979. A model for the
geological and palaeontological evidence; modeling evolution of the Indian Ocean and the breakup of
strategies, boundary conditions, and some pre- Gondwanaland. J. Geophys. Res., 84: 6803-6830.
liminary results. Adv. Geophys., 26: 35-140. Owen, H.G., 1983. Atlas of Continental Displacement,
Loper, D.E., 1984. Structure of the core and lower 200 Million Years to Present. Cambridge, 159 pp.
mantle. Adv. Geophys., 26: 1-34. Palmer, A.R., 1983. The decade of North American
Loper, D.E., 1985. A simple model of whole-mantle geology 1983 geologic time scale. Geology, 11:
convection. J. Geophys. Res., 90: 1809-1836. 503-504.
Mahoney, J.J., Macdougall, J.D., Lugmair, G.W. and Parsons, B., 1981. The rates of plate consumption and
Gopalan, K., 1983. Kerguelen hotspot source for creation. J. Geophys. R. Astron. Soc., 67: 437-448.
Rajmahal Traps and Ninetyeast Ridge? Nature, 303: Patriat, P. and Acharche, J., 1984. India-Eurasia colli-
385-389. sion chronology has implications for crustal shorten-
Martin, A.K. and Hartnady, C.J.H., 1986. Plate tectonic ing and driving mechanisms of plates. Nature, 311:
development of the South West Indian Ocean: a 615-621.
revised reconstruction of East Antarctica and Africa. Pelzer, G., 1989. Bull. Geol. Inst. Univ. Uppsala. In
J. Geophys. Res., 91: 4767-4786. press.
Mayer, A.E.S. and Burdett, M., 1986. The ocean drilling Pitman, W.C. and Talwani, M., 1972. Sea-floor spread-
program: organization and progress. Episodes, 9: ing in the North Atlantic. Geol. Soc. Am. Bull., 83:
71-77. 619-646.
McAdoo, D.C. and Sandwell, D.T., 1985. Folding of Powell, C.M.A., Johnson, B.D. and Veevers, J.J., 1980.
oceanic lithosphere. J. Geophys. Res., 90: 8563-8569. A revised fit of east and west Gondwanaland.
McCabe, R., 1984. Implications of palaeomagnetic data Tectonophysics, 63: 13-29.
on the collision related bending of island arcs. Price, R.A., 1981. The cordilleran foreland thrust and
Tectonophysics, 3: 409-428. fold belt in the Canadian Rocky Mountains. In:
McCourt, W.J., Aspden, J.A. and Brook, M., 1984. New K.R. McClay and N.J. Price (Editors), Thrust and
geological and geochronological data from the Col- Nappe Tectonics. Spec. Publ. Geol. Soc. London, pp.
umbian Andes: continental growth by multiple 427-448.
accretion. J. Geol. Soc. London, 141: 831-845. Rea, D.K. and Dixon, J.M., 1983. Late Cretaceous and
McKenzie, D., 1972. Active tectonics of the Mediter- Paleogene tectonic evolution of the North Pacific
ranean region. Geophys. J.R. Astron. Soc., 30: Ocean. Earth Planet. Sci. Lett., 65: 145-166.
109-185. Rea, D.K. and Duncan, R.A., 1986. North Pacific plate
McKenzie, D., 1978. Active tectonics of the convergence: a quantitative record of the past 140
Alpine-Himalayan belt: the Aegean Sea and sur- m.y. Geology, 14: 373-376.
161
Reymer, A. and Schubert, G., 1984. Phanerozoic ad- Van der Linden, W.J.M., 1979. The Atlantic margin of
dition rates to the continental crust and crustal Iberia and Morocco, a reinterpretation. Tectono-
growth. Tectonophysics, 3: 63-78. physics, 59: 185-199.
Reymer, A. and Schubert, G., 1986. Rapid growth of Van Houten, F.B., 1983. Sirte Basin, north-central
some major segments of continental crust. Geology, Libya: Cretaceous rifting above a fixed mantle
14: 299-302. hotspot? Geology, 11: 115-118.
Richter, F.M. and Parsons, B., 1975. On the interaction Weijermars, R., 1984. New structural theorems on new
of two scales of convection in the mantle. J. Geo- global tectonics. Terra Cognita, 4: 175-183.
phys. Res., 80: 2529-2541. Weijermars, R., 1985. Comment on " O n the breakup of
Riddihough, R., 1984. Recent movements of the Juan tectonic plates by polar wandering" by Han-Shou
de Fuca plate system. J. Geophys. Res., 89: Liu. J. Geophys. Res., 90: 11559-11564.
6980-6994. Weijermars, R., 1986. Slow but not fast global expan-
Robinson, A.H. and Sale, R.D., 1968. Elements of sion may explain the surface dichotomy of Earth.
Cartography. Wiley, New York, N.Y., 3rd ed., 310 Phys. Earth Planet. Inter., 43: 67-89.
pp. Weijermars, R., 1987a. Convection Experiments in High
Royden, L.H., Horvath, F. and Burchfiel, B.C., 1982. Prandtl Number Silicones. Ph.D.Thesis. Uppsala
Transform faulting, extension, and subduction in the University, Uppsala, 172 pp.
Carpathian Pannonian region. Geol. Soc. Am. Bull., Weijermars, R., 1987b. A revision of the Eurasian-Af-
93: 717-725. rican plate boundary in the western Mediterranean.
Schilling, J.-G., Sigurdson, H., Davis, A.N. and Hey, Geol. Rundsch. 75: 667-676.
R.N., 1985. Easter microplate evolution. Nature, 317: Weijermars, R., 1988a. Where does Africa meet Europe?
325 -331. Nature, 332: 118.
Schubert, G. and Reymer, A.P.S., 1985. In: Nature, Weijermars, R., 1988b. Convection experiments in high
316: 336-339. Prandtl number silicones, Part 1. Rheology,
Sclater, J.G., Jaupart, C. and Galson, D., 1980. The heat equipment, nomograms and dynamic scaling of
flow through oceanic and continental crust and the stress- and temperature-dependent convection in a
heat loss of the Earth. Rev. Geophys. Space Phys., centrifuge. Tectonophysics, 154: 71-96.
18: 269-311. Weijermars, R., 1988c. Convection experiments in high
Sclater, J.G., Parsons, B. and Jaupart, C., 1981. Oceans Prandtl number silicones, Part 2. Deformation, dis-
and continents: similarities and differences in the placement and mixing in the Earth's mantle.
mechanisms of heat-loss. J. Geophys. Res., 86: Tectonophysics, 154: 97-123.
11535-11552. Weijermars, R., in prep. New fit of Gondwanaland
Seno, T. and Maruyama, S., 1984. Palaeogeographic predicts continuation of South African ore province
reconstruction and origin of the Phillipine Sea. in Precambrian basement of India. Submitted for
Tectonophysics, 102: 53-84. review.
Smith, A.G. and Hallam, A., 1970. The fit of the Weijermars, R. and Schmeling, H., 1986. Scaling of
southern continents. Nature, 225: 139-145. Newtonian and non-Newtonian fluid dynamics
Spakman, W., 1986. Subduction beneath Eurasia in without inertia for quantitative modelling of rock
connection with the Mesozoic Tethys. Geol. flow due to gravity (including the concept of rheo-
Mijnbouw, 65: 145-153. logical similarity). Phys. Earth Planet. Inter., 43:
Suess, E., 1885. Das Anlitz der Erde. Tempsky, Vienna, 316-330.
Vol. 1,768 pp. Weissel, J.K., Anderson, R.N. and Geller, C.A., 1980.
Tapponier, P., Peltzer, G., Le Dian, A.Y., Armijo, R. Deformation of the Indo-Australian plate. Nature,
and Cobbold, P., 1982. Propagating extrusion tecton- 287: 284-291.
ics in Asia: new insights from simple experiments Wernicke, B., 1985. Uniform-sense normal simple shear
with plasticine. Geology, 10: 611-616. of the continental lithosphere. Can. J. Earth Sci., 22:
Tarling, D.H., 1972. Another Gondwanaland. Nature, 108-125.
238: 92-93. Wernicke, B. and Burchfield, B.C., 1982. Modes of
Taylor, S.R. and Mclennan, S.M., 1985. The Continen- extensional tectonics. J. Struct. Geol., 4: 105-115.
tal Crust: Its Composition and Evolution. Blackwell, Wezel, F.-C., 1988. Earth structural patterns and
London, 312 pp. rhythmic tectonism. Tectonophysics, 99: 241-249.
Trumpy, R., 1973. The timing of orogenic events in the Whitman, J.M., Harrison, C.G.A. and Brass, G.W.,
Central Alps. In: K.A. de Jong and R. Scholten 1983. Tectonic evolution of the Pacific Ocean since
(Editors), Gravity and Tectonics. Wiley, London, pp. 74 Ma. Tectonophysics, 99: 241-249.
229-251. Wilson, F.D., 1986. A kinematic model for the Gorda
Turcotte, D.L. and Schubert, G., 1982. Geodynamics. Deformation Zone as a diffuse southern boundary of
Wiley, New York, N.Y., 450 pp.
162
the Juan de Fuca plate, J. Geophys. Res., 91: Cocos-Nazca spreading center. Geology, 9: 425-430.
10259-10269. Yumi, S. and Wako, Y., 1970. Secular motion of the
Windley, B.F., 1984. The Evolving Continents. Wiley, pole. In: L. Mansinha, D.E. Smylie and A.E. Beck
Chichester, 2nd ed., 399 pp. (Editors), Earthquake Displacement Fields and the
Woods, M.T. and Davies, G.F., 1982. Late Cretaceous Rotation of the Earth. Reidel, New York, N.Y., pp.
genesis of the Kula plate. Earth Planet. Sci. Lett., 58: 82-87.
161-166.
Wortel, M.J.R., 1986. Dynamical aspects of active con-
tinental margins. Geol. Mijnbouw, 65: 119-132. [Received December 22, 1987; accepted after revision
Wortel, R. and Cloetingh D., 1981. On the origin of the July 25, 1988]