Hyatt, Heart weight gravity JAP 85

Influence of the heart on the vertical gradient of transpulmonary pressure in dogs ROBERT E. HYATT, EPHRAIM BAR-YISHAY, AND MARTIN D. ABEL Division of Thoracic Diseases.and Internal Medicine, Departments of Physiology and Biophysics and Anesthesiology, Mayo Clinic and Mayo Foundation, Rochester, Minnesota 55905 erted very little effect on the lung in this position. Of interest is the fact that the distribution of regional lung volumes differs in supine and prone dogs. This was suggested indirectly by the study of Lai et al. (ll), who found that the slope of phase III of the single-breath O2 maneuver was significantly lower in prone than in supine dogs. More recently, Hubmayr et al. (9), using intraparenchymal metallic markers, demonstrated a vertical gradient of regional lung volumes in supine dogs but not in prone dogs. These findings would be compatible with the suggestion that the lungs in part support the heart in the supine but not in the prone posture. We therefore tested more directly the sugges tion that in the dog the lungs help to support the heart and that thi s support is one of the determinants of the VGTP. We studied dogs in the head-up posture and used esophageal pressure to estimate pleural surface pressure and to obtain a measure of the VGTP. After the induction of a total bilateral pneumothorax, the VGTP disappeared-a phenomenon initially reported by Thurlbeck and Marshall (16). Roentgenograms showed that heart-lung interaction; head-upposture; pneumothorax; esoph- the heart moved in a caudal-dorsal direction. Increasing agealpressure;pleural pressuregradient; lung inflation the heart weight by adding Hg to the heart increased the VGTP. These data are consistent with the suggestion that in the dog the lungs help to support the heart and A NUMBER OF INVESTIGATORS have found a vertical that the VGTP is in part due to the pressure distribution gradient of transpulmonary pressure (VGTP) in the up- required for the support of the heart by the lungs. right dog when pressure was mea sured eith .er in the pleural ca.vity or in the esophagus (Tab1 .e 1). Some (8 9 METHODS 17) but not all (2) have found that the VGTP does not Dogs weighing 16.4 t 2.1 kg (mean t SD) were used. change with passive lung inflation. Agostoni ( 1) has reviewed the various mechanisms that In the first phase of the experiment, dogs were anesthetized with intravenously administered pentobarbital .sohave been proposed to explain the origin of the VGTP. It has been suggested that VGTP results from gravity dium (30 mg/kg). In most of the dogs, catheters were placed in one or both ventricles. A tracheal cannula was acting directly on the lung or on the chest wall, thus deforming the chest wall so that the natural shapes of tied in place. The dog was first placed supine in a special the lung and chest wall do not match. cast ‘, similar to that used by Banchero et al. (4). The dog One possible mechanism that has received relatively was secured by attaching the incisors to the top of the little attention is that the heart may be supported by the cast and by strapping all 1imbs to the cast. In addition, lungs. Banchero et al. (4) suggested that the density of an e lastic bandage around the abdome n helped to. mainthe heart was important in determining the VGTP in tain a constant posture and prevented any sizeable inthe lower thorax. Similarly, McMahon et al. (12) ex- crease in functional residual capacity (FRC) when the plained their measures of regional pleural pressure on dog was put in the head-up position. The relative constancy of FRC was estimated by using a large calibrated the basis of the weight of the mediastinal structures, especially the heart. However, Miserocchi et al. (14) syringe and measuring expiratory reserve volume (airway pressure = -30 cmH20) and inspiratory capacity (airway suggested that in supine dogs the heart was supported through the negative pressure of the pleural liquid in the pressure = +30 cmH20) when the dog was both supine lung-free sternal region and that the heart’s weight ex- and head up. HYATT, ROBERT E., EPHRAIM BAR-YISHAY, AND MARTIN D. ABEL. ~~f~u~~~~ of the heart on the vertical gradient of trans~ul~ona~ pressure in dogs. J. Appl. Physiol. 58(l): 52-57, 1985-Estimations were madeof the vertical gradient of transpulmonary pressure(VGTP) from measurementsof esophageal pressurein nine head-up dogsat functional residual capacity (FRC) when alive, when dead, and after total bilateral pneumothorax. The VGTP of 0.4 cmH@/cm height in the alive state was abolished by pneumothorax, and roentgenograms showedthat the heart movedin a caudal-dorsaldirection. There wasa smallbut significant increasein the VGTP on going from FRC to near total lung capacity (TLC) in alive head-up dogs. In eight dead head-up dogs heart weight was increased by replacingvarious amounts of heart blood with Hg. The VGTP wassignificantly increasedfrom 0.28to 0.51cmHzO/cm height. The fractional increasein the VGTP was similar to the fractional increase in heart weight. In five dogs extrapolation to zero heart weight gave an average VGTP of 0.14 cmHzO/cm height. We conclude that the lungs help support the heart in the head-updog and that the VGTP is in part determined by the pressuredistribution required for this support. 52 0161-7567/85 $1.50 Copyright0 1985 the AmericanPhysiological Society VERTICAL GRADIENT OF TRANSPULMONARY 1. Summary of measurements of VGTP in head-up dogs at FRC TABLE Site of Measurement Pleural cavity and esophagus Pleural cavity Pleural cavity Pleural cavity Pleural cavity Pleural cavity Esophagus Pleural cavity Esophagus Ref. No. 4 10 15 53 PRESSURE VGTP, cmH20/cm 0.72 0.21 0.3 0.3 0.16-o. 0.4 0.42 0.60 0.40 Ht 8 12 2 16 7 Present study VGTP is vertical gradient of transpulmonary pressure. FRC is functional residual capacity. the expiratory reserve volume from the lungs to the pleural cavities. Equal amounts were pl .aced i n each cavity, and the cavities were connected to allow equilibration. Thus, overall thoracic volume was maintained constant before and after pneumothorax. The VGTP was again measured. Next, the chest was widely opened and the VGTP remeasured. In all 14 dogs, the VGTP was measured at FRC and TLC in the alive anesthetized state. At some stage in the experiment after death, the VGTP could not be measured in five dogs because of the accumulation in the esophagus of air or fluid (or both), which could not be completely removed. Thus, comparisons of the VGTP were obtained at FRC in nine dogs when alive, when dead, and after pneumothorax. The effect of altering heart weight was evaluated in eight head-up dogs. Double-lumen catheters with thin rubber balloons covering the distal lumen were placed in the right ventricle in all eight dogs and in the left ventricle in four of the dogs. Heparin was administered, and the dogs were killed by anesthetic overdose. The VGTP was measured at FRC. Next, 12-50 ml of blood were removed from the heart via the proximal catheter lumen, and equal volumes of Hg were added to the balloons via the distal lumen (see Table 2 for volumes added). The VGTP was again measured. In dogs 25 and 26, Hg was added in two steps bracketed by measures of the VGTP. In five of eight dogs, the heart was removed, drained free of blood, and weighed. In three dogs, the balloon broke during heart removal, spilling Hg into the heart. Since Hg could not be completely removed from the ventricular trabeculae, an accurate estimate of heart weight was not obtained in these three dogs. Anteroposterior and lateral roentgenograms were obtained each time that the VGTP was measured. A tube with a 2-mm focal spot was used. Exposure time was 0.3 s at 80-90 kV and 15 mA. Focal spot to film distance was 125 cm, and the film was in contact with the chest wall. 1. The VGTP was measured by plotting the static transpulmonary pressure (esophageal pressure minus airway pressure at zero airflow) against the vertical height. Since the relationship was fairly linear, the slope of the relationship (the VGTP) was estimated by linear regression analysis. In all but four dogs, static transpulmonary pressure (Ptp) was obtained by moving a 4-cm-long balloon up the esophagus in steps of 2-4 cm. The balloon was 3.5 cm in circumference and contained 0.7 ml of air, which placed the balloon on the flat portion of its pressure-volume curve. In the spontaneously breathing dog, the cardia and the region of the tracheal artifact (13) were identified by occluding the airway and requiring a constant Ptp during inspiratory efforts, except for slight rarefaction effects. In the dead dog, validity of pressure recordings was checked by occluding the airway, manually compressing the chest, and requiring similar constancy of Ptp. In four dogs, the VGTP was measured using three 6-cm-long balloons arranged in series. This arrangement provided three simultaneous estimates of Ptp over a distance of 12 cm, on the assumption that pressure was estimated at the top of each balloon. Similar precautions regarding balloon volumes and pressure checks were taken with this system. RESULTS The VGTP was measured in 14 alive head-up anesthetized dogs. In all instances throughout the study, three Effect of pneumothorax. The mean data measured at or four technically acceptable estimates of the VGTP FRC in nine dogs when alive, when dead, and after were obtained and the results were averaged. In these 14 bilateral pneumothorax are shown in Fig. 1. The distance dogs, the VGTP was measured at FRC and at a volume 100 ml below total lung capacity (TLC). TLC was taken TABLE 2. Effects of increased heart weight as the lung volume obtained when airway pressure was on VGTP at FRC +30 cmHz0. TLC minus 100 ml was designated “TLC” and was chosen to avoid unduly compromising venous VGTP, Volume of Dog Wt, cmH20/cm Ht return. Hg Added, Dog No. kg ml The dog was then heparinized and rapidly killed by an Control Hg overdose of anesthetic; the airway was occluded prior to 30* 0.06 0.27 5 16 the overdose in a n attempt to main .tain lung vo lume 0.07 0.08 6 17 30* con .stant at FRC. C ontrast material (Renografin) was 0.25 0.58 11 19 50* 0.48 0.76 12 19 25t then introduced into one or both ventricles, and the 0.37 0.65 23 16 17-r VGTP was measured again. Preliminary studies indi17t 0.47 0.83 24 12 cated that replacing the heart blood with an equal volume 16 12f0.27 0.55 25 of contrast material had no effect on the VGTP. A total 0.28 0.35 26 20 12-f bilateral pneumothorax was produced by inserting a VGTP is vertical gradient of transpulmonary pressure. FRC is blunt needle connected to the tracheal ca nnula into each functional residual capacity. *Divided evenly between right and left pleural cavity and transferrin .g a volume of gas equal to ventricle. -FAdded to right ventricle only. HYATT, 54 BAR-YISHAY, AND ABEL first or second thoracic vertebrae in 10 dogs. The distance from the landmark to the most caudal portion of the heart was measured on both anteroposterior and lateral roentgenograms. In all dogs, the heart moved upward with lung inflation; the mean displacement was 1.2 cm (range 0.2-2.5). Effect of increasing heart weight. In all but dog 6, there -16 -12 -6 -4 0 Transpulmonary pressure (cm H20) FIG. 1. Static transpulmonary pressure plotted as a function of height (distance from incisor teeth) in 9 head-up dogs in 3 states. Measurements were made at same end-expiratory thoracic gas volume. PNX refers to data after total bilateral pneumothorax. Horizontal bars, 1 SE. along the esophagus that yielded acceptable measurements of pressure was nearly the same in all dogs alive and dead. Therefore, Ptp was plotted as a function of absolute distance from the incisor teeth. Some difference was seen in the pattern of Ptp vs. height between the alive and the dead dog. The mean overall VGTP was 0.4 and 0.2 cmHzO/cm height in the alive and dead state, respectively. However, the VGTP in the alive and dead dogs was not statistically significant by paired t test. There was no gradient after the pneumothorax. Mean Ptp was -4.2 cmHpO alive, -4.3 dead, and +0.5 after pneumothorax. There were no differences in gradients or mean Ptp between the pneumothorax state and the open thorax state; the latter data are not shown. With induction of the bilateral pneumothorax, the heart moved slightly in the caudal direction and very strikingly in the dorsal direction (Fig. 2). Metal markers were placed on the ribs and sternum of some dogs to facilitate comparison of roentgenograms. In these dogs, acceptable measurements of pressure were obtained over a distance of 58-40 cm from the incisors. On the lateral roentgenogram (upper right panel), this distance corresponded to a length from 4 cm below the top of the diaphragm to 2 cm above the base of the heart. Effect of lung inflation. In 10 alive dogs, the VGTP, as measured from the esophagus, did not disappear with lung inflation (Fig. 3). The mean VGTP in these 10 dogs plus the 4 dogs studied with the three-balloon system increased from 0.43 to 0.58 cmHzO/cm height at FRC and “TLC,” respectively. This difference was significant by paired t test (P < 0.05). Mean Ptp at “TLC” was -14.6 cm HzO. Roentgenograms taken at FRC and “TLC” showed that with lung inflation the heart was no longer in contact with the diaphragm (Fig. 4); this was true in all 14 dogs. In 11 of the 14 dogs, the heart was no longer in contact with the sternum. These changes in heart position occurred with or without the abdominal binder. In an attempt to ascertain if the heart was displaced upward with lung inflation, landmarks were identified on the FIG. 2. Anteroposterior (left column) and lateral (right column) roentgenograms before (top row) and after (bottom row) induction of total bilateral pneumothorax in a dead dog. Catheters are seen entering ventricles that contain contrast material, and there is a radiopaque catheter in esophagus. Screws and metal markers identify bony landmarks. In bottom row are seen S needles used to induce pneumothorax. After bilateral pneumothorax, heart moves in a caudal-dorsal direction. F :: $f 44 - .; 46 - .E 48 - E e 50 - 52 - : s ;; 54 - 56 - 8 50 - - I -20 I I -16 I I -12 I I -5 I I -4 I I II 0 Transpulmonary pressure (cm H20) FIG. 3. Static transpulmonary pressure plotted as function of height (distance from incisor teeth) in 14 alive head-up dogs at functional residual capacity (FRC) and total lung capacity minus 100 ml (“TLC”). Horizontal bars, 1 SE. VERTICAL GRADIENT OF TRANSPULMONARY 55 PRESSURE dogs; this was accomplished by linear regression analysis in dogs 25 and 26. The mean VGTP at zero heart weight was 0.14 cmH20/cm height, with a range from 0.07 to 0.26. Since the data on the ratio of heart blood volume to body weight were few, the results were evaluated on the assumption that the ratio was either overestimated or underestimated by 50%. An overestimate of 50% would, at zero heart weight, give a VGTP of 0.22 cmHPO/ cm height, whereas an underestimate of 50% would give a figure of 0.11 cmHzO/cm height. Hence, the extrapolated VGTP is not too sensitive to the heart blood/body weight ratio. The mean fractional changes in the VGTP with the addition of Hg and the mean fractional increase in heart weight with the Hg addition were also calculated. The mean fractional change in VGTP was 0.68, and the mean fractional change in heart weight was 0.85; these differences were not significant (P > 0.3). DISCUSSION FIG. 4. Anteroposterior (left column) and lateral (right column) roentgenograms at functional residual capacity (top row) and total lung capacity minus 100 ml (bottom row) in alive dog. Radiopaque catheter with l-cm gradations is seen in esophagus. was an increase in the VGTP after the addition of Hg to the heart of the dead head-up dog (Fig. 5). By paired t test, the increase in gradient from 0.28 to 0.51 cmHaO/ cm height was significant (P < 0.005). In several dogs, the heart was viewed fluoroscopically as Hg was added, and it was noted that the heart moved in a caudal-dorsal direction. There was no obvious association among dogs between the magnitude of the change in the VGTP and either the amount of Hg added or the size of the dog (Table 2). However, in a given dog there appeared to be a correlation between the change in heart weight and the change in the VGTP (see below). There are little data on the blood volume of the canine heart as a function of body weight. Based on preliminary data (E. Hoffman, personal communication), a value of 7.5 ml blood/kg body wt was used as the heart blood volume. In five dogs, we knew the weight of the heart as well as the volume of Hg added. Therefore, using this figure for heart blood volume, we could estimate the heart weight in both the control and the Hg-weighted states. In Fig. 6, heart weight was plotted against the VGTP. In dogs 25 and 26, Hg was added in two steps. The data were extrapolated to zero heart weight in all Esophageal pressure was used to estimate the VGTP. Previous work from our laboratory indicated that esophageal pressure was a valid reflection of regional pleural pressure in supine and prone dogs (6). Thurlbeck and Marshall (16) reached a similar conclusion and also presented data suggesting that the esophagus provides a valid estimate of the topography of regional pleural pressure. Like the study of Thurlbeck and Marshall (16), the present study gave an estimate of the VGTP that was similar to the values measured directly from the pleural cavity (Table 1). In this study, care was taken to be sure that the esophagus was free of air and fluid and that the length of esophagus that provided valid pressures was identified by appropriate checks during airway occlusion (13). Therefore, we believe our estimates of the VGTP are valid, although there may have been some errors in our estimates of height because we did not use roentgenograms to quantify vertical height (5). We assumed that any errors in estimates of height were constant in a given dog. As noted previously, little attention has been given to the possibility that support of the heart by the lung has played a role in determining the VGTP in dogs. Our roentgenograms (Figs. 2 and 4) and direct observations at autopsy (see below) indicate that the heart does not receive significant support from the sternopericardiac ligament, in contrast to a previous suggestion (3). Furthermore, the lack of tension in the pericardial membrane in intact dogs (3) indicates that the heart is not supported by the pericardium. From the present study, we conclude that the lungs do help to support the heart in the head-up dog. The dramatic dorsal movement of the heart after pneumothorax (Fig. 2) and the small upward movement with lung inflation (Fig. 4) support this conclusion. In several dogs we carefully cut away the intercostal muscles, permitting inspection of the interior of the chest cavity. With lungs inflated in the supine position, the heart could be seen to be lifted ventrally, coming into close proximity to the sternum. The present results also indicate that the pressure distribution required to support the heart helps to deter- 56 HYATT, BAR-YISHAY, AND ABEL o OH9 #5 32 36 Control #6 Ml #12 40 44 48 52 56 60 4cl 5 I -4 -8 , ??t -8 -8 -4 ,:, 0 -16 -12 -8 -4 #25 #26 R l l l -8 -4 0 - 12 -8 Transpulmonary -4 0 pressure -8 FIG. 5. Static transpulmonary pressure plotted as a function of height (distance from incisor teeth) in 8 dead headup dogs in control state (W) and after introduction of Hg into heart (W). In dogs 25 and 26, Hg was added in 2 steps; interrupted line represents second step. A 3-balloon system was utilized in dogs 23, 24, 25, and 26; pressures could not be recorded from middle balloon in dog 23 due to a balloon leak. See Table 2 for further details. -4 l ‘is, l \ 0 -8 -4 0 (cm H *O) mine the VGTP. In seven of eight dogs, the VGTP increased after the heart’s weight increased with the addition of Hg (Fig. 5 and Table 2). Thus, any increase in heart weight above the normal state appeared to be borne by the lungs and resulted in an increase in the VGTP. Extrapolation of the VGTP to the “weightless” heart state (Fig. 6) gave a gradient close to that expected from the action of gravity on the lungs alone, namely 0.22 cmH20/cm height (10). Miserocchi et al. (14) argued that the heart in the supine dog is supported by the negativity of the pleural liquid pressure in the region of apposition of the heart to the sternum. We find it difficult to visualize how a sufficiently negative pressure could be maintained without the accumulation of fluid. Furthermore, in the headup dog this mechanism would require that there be no shear stress in this region. This seems unlikely. Further evidence against this mechanism (14) was the fact that at FRC the fractional change in the VGTP was similar to the fractional change in heart weight, an unlikely result if the heart in the control state was supported by the relatively noncompliant sternum. Furthermore, the roentgenograms taken near TLC (Fig. 4) show that the heart is no longer in contact with the sternum. This was evident in 11 of 14 dogs. Separation of the sternum and the heart was also noted in three of six dogs at a volume midway between FRC and “TLC.” Under these circumstances, the sternal support mechanism could not be operative. The fact that the VGTP increased only slightly between FRC and “TLC” is consistent with the notion that the lungs are almost totally supporting the heart at FRC. Our finding that the VGTP did not disappear near TLC agrees with the findings of Hoppin et al. (8). However, Agostoni and Miserocchi (2), using a different technique, found in the head-up dog that the VGTP decreased progressively with passive lung inflation so $’ 5 $ g 1.4 - 1.2 - 5: .c3, 1.0 - 12 0 23 @ 24 @ 25 l $j z k 5 ‘a8 E 5 0.6 - A 26 Heart weight (g) 6. Vertical gradient of transpulmonary pressure is plotted against heart weight in 5 dead head-up dogs. Heart weight was increased by introducing Hg into ventricles. See text for further details. FIG. that it became zero at TLC. We cannot explain these discrepant results. It should be emphasized that our findings apply only to the dog, a species in which there is very loose coupling of the pericardium to sternum or diaphragm. In the rabbit, for example, the heart is firmly bound to the sternum by ligaments (14). The reasons for, and implications of, these species differences in support of the heart warrant further study. Our study does not permit speculation regarding other mechanisms proposed as determinants of the VGTP, such as gravitational forces acting on the lung or chest wall. The extrapolation data (Fig. 6) do suggest, however, that gradients greater than 0.11-0.22 cmHzO/cm height in the head-up dog are due to the lungs’ support of the heart. It has been observed that the substantial vertical gradient of regional lung volume in the supine dog disappears in the prone dog (9). This raises the additional VERTICAL GRADIENT OF TRANSPULMONARY 57 PRESSURE possibility that the weight and support of the he art have a role in determining the vertical distribution of regional lung volumes, if the lungs partial .ly support the heart in the supine bu t not in the prone posture. However, the present data do not address this question directly. Our findings could, for example, reflect pleural effects localized to the region traversed by the esophagus. In addition, the cardiac (mediastinal) lobe in the dog might provide most of the support for the heart. If so, the effect of the heart in determining regional lung volumes could be primarily localized to this lobe. Additional studies are required before we can evaluate with confidence the possible role the heart has in determining lung volumes. overall regional The authors thank Mark A. Schroeder and Darrell L. Loeffler for their excellent technical assistance. We also thank Drs. Theodore A. Wilson, Joseph R. Rodarte, and Kai Rehder for stimulating discussions and critical review of this work. This investigation was supported in part by National Heart, Lung, and Blood Institute Grant HL-21584. E. Bar-Yishay was supported by a Parker B. Francis Foundation Fellowship. Address for reprint requests: R. E. Hyatt, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. Received 30 March 1984; accepted in final form 30 July 1984. REFERENCES 1. AGOSTONI, 128,1972. 2, AGOSTONI, E. Mechanics of the pleural space. Physiol. Rev. 52: 57- E,, AND G. MISEROCCHI. Vertical gradient of transpulmonary pressure with active and artificial lung expansion. J. Appl. 3. 4. 5. 6. 7. Physiol. 29: 705-712, 1970. BANCHERO, N., W. J. RUTISHAUSER, A. G. TSAKIRIS, AND E. H. WOOD. Pericardial pressure during transverse acceleration in dogs without thoracotomy. Circ. Res. 20: 65-77, 1967. BANCHERO, N., P. E. SCHWARTZ, A. G. TSAKIRIS, AND E. H. WOOD. Pleural and esophageal pressures in the upright body position. J. Appl. Physiol. 23: 228-234, 1967. BANCHERO, N., P. E. SCHWARTZ, AND E. H. WOOD. Intraesophageal pressure gradient in man. J. Appl. Physiol. 22: 1066-1074, 1967. GILLESPIE, D. J., Y.-L. LAI, AND R. E. HYATT. Comparison of esophageal and pleural pressures in the anesthetized dog. J. Appl. Physiol. 35: 709-713, 1973. GROPPER, M. A., J. P. WIENER-KRONISH, AND S. J. LAI-FOOK. Pleural liquid pressure (Ppl) in upright dogs measured through implanted rib capsules (Abstract). Federation Proc. 42: 1270, 1983. 8. HOPPIN, F. G., JR., I. D. GREEN, AND J. MEAD. Distribution of pleural surface pressure in dogs. J. Appl. Physiol. 27: 863-873,1969. 9. HUBMAYR, R. D., B. J. WALTERS, P. A. CHEVALIER, J. R. Ro- DARTE, AND L. E. OLSON. Topographical distribution of regional lung volume in anesthetized dogs. J. Appl. Physiol. 54: 1048-1056, 1983. 10. KRUEGER, J. J., T. BAIN, gradient of intrathoracic AND J. L. PATTERSON, JR. Elevation pressure. J. Appl. Physiol. 16: 465-468, 1961. 11. LAI, Y.-L., J. R. RODARTE, AND R. E. HYATT. Effect of body position on lung emptying in recumbent anesthetized dogs. J. Appl. Physiol. 43: 983-987, 12. MCMAHON, S. M., 1977. D. F. PROCTOR, AND S. PERMUTT. Pleural surface pressure in dogs. J. Appl. Physiol. 27: 881-885, 1969. 13. MILIC-EMILI, J., J. MEAD, AND J. M. TURNER. Topography of esophageal pressure as a function of posture in man. J. Appl. Physiol. 19: 212-216, 1964. 14. MISEROCCHI, G., E. D’ANGELO, AND E. AGOSTONI. Topography of pleural surface pressure after pneumo- or hydrothorax. J. Appl. Physiol. 32: 296-303, 1972. 15. PROCTOR, D. F., P. CALDINI, AND S. PERMUTT. The pressure surrounding the lungs. Respir. Physiol. 5: 130-144, 1968. 16. THURLBECK, W. M., AND R. M. MARSHALL. Topography of esophageal pressure in the dog. J. Appl. Physiol. 34: 590-596, 1973. 17. TURNER, J. M. Distribution of lung surface pressure as a function of posture in dogs (Abstract). Physiologist 5: 223, 1962.