Article Title: Respiratory gas exchange of sea turtle nests

Transcription

1 ARTICLE INFORMATION Author (s): Ackerman, R.A. Article Title: Respiratory gas exchange of sea turtle nests Journal Title: Respiration Physiology Year: 1977 Volume: 31 Pages: 19-38

2 ~~.~J~irurio~ Physiofog,y ( 1977) 31, EisevieriNorth-Holland Biomedical Press THE RESPIRATORY GAS EXCHANGE OF SEA TURTLE NESTS (CHELONIA, CARETTA) 1 RALPH A. ACKERMAN Department qf Zoolog_v, Unioersity of Floridu, Gaintwillr, Floridu U.S.,4 Abstract. Sea turtles lay about 100 leathery-shelled eggs in a 25 cm diameter chamber carefully excavated about 50 cm deep in a nesting beach, where the eggs exchange gases (at approximately 28 C) during their 60-day incubation period. The sand surrounding the spherical nest chamber restricts the diffusion of gases into and out of the nest so that as embryonic development progresses, PO2 decreases and Pco, increases in the gas inside the nest. Po, falls to torr and Pco rises to torr inside loo-egg manmade C~e~~~~ and Caretra nests. The change in gas tensions &I the nest during development is very similar to that seen in the air cell of the chicken egg. Gas tensions inside the turtle nest and in the sand surrounding the nest can be described by a radial steady-state diffusion model. The rate of diffusion of gases in the sand is 3C-507, of the rate found in the nest and G-127; of the rate found in an equal volume of air. The sand surrounding the turtle nest appears to determine the gas exchange of the eggs in the nest and is functionally analogous to the shell surrounding the chicken embryo. The female sea turtle may construct her nest so as to maximize its gas exchange and minimize gas partial pressure gradients inside the nest. Diffusion Egg incubation Eggshell gas exchange Turtle eggs Turtle nests Female sea turtles deposit about 100 golf-ball-sized, soft-shelled, leathery eggs in urn-shaped nests excavated in marine beaches scattered throughout the tropical and semitropi~l Iatitudes of the world (Carr, 1967). The nest is carefully and stereotypically constructed by the females of all five sea turtle genera. The eggs are then deposited, free of sand, in the nest chamber (Carr, 1967; Carr and Ogren, 1960) and the chamber, which neither loses its shape nor fills with sand, is then buried and the Accepted j&r pab~i~at~o~ 23 February This work was in part supported by a NDEA Title IV Fellowship. Present address: Department of Physiology, State University of New York at Buffalo, Buffalo, New York 14214, U.S.A. 19

3 20 R. A. ACKERMAN nest obscured. Sixty days Iater the eggs hatch simultaneously and the hatchlings collectively dig their way to the beach surface. The individual sea turtle embryo (Cheloniu ~ZJY&ZS) exchanges gases across a shell that is much more permeable to gases (Ackerman and Prange, 1972) and water vapor (Rahn, personal communication) than the bird egg-shell. This observation is correlated with the presence of fully water-saturated gas inside the nest. When the oxygen consumption of a near-term egg (Ido, z 9.7 x IO- ml. set- ) is used together with shell permeability (K g 6.5 x lo- ml. see- STPD-~~-*. mm Hg- ) and shell area (A E 28 cm ) to estimate the oxygen gradient across the shell (dpoz = tio,/k. A), it is apparent that only a very small oxygen gradient (dpoz z 5 mm Hg) and no carbon dioxide gradient should occur (Prange and Ackerman, 1974). This is in marked contrast to the bird egg, where a low shell ~rmeability and relatively high oxygen consumption (and carbon dioxide production) result in substantial 0, and CO, gradients across the shell near hatching (Wangensteen and Rahn, ; Wangensteen, 1972). However, Prange and Ackerman (1974) have measured decreased 0, tensions and increased CO, tensions in air surrounding green turtle eggs of about 45 days incubation. Based on the individual rates of metabolism of hatchlings and eggs near hatching, the metabolic demand for oxygen and production of carbon dioxide by the average loo-egg clutch is about ml. hr-, This metabolic activity, together with the diffusion resistance offered by the sand, should produce substantial gas concentration gradients between the nest and surrounding beach. The magnitude of these gradients should be shaped by a number of biological variables, such as clutch size, nest construction, nest depth and a number of geological or physical variables such as sand grain size or sand air and water content, all of which must interact to produce the respiratory environment in which sea turtle eggs develop. This study describes the changes in oxygen and carbon dioxide tension in the sea turtle nest and in the sand surrounding the nest during incubation. A steady-state, radial diffusion model of nest gas exchange is presented which relates the mass flux of gas in and out of the nest to the gas permeability of the sand surrounding the nest, the nest geometry and the gas concentration gradients between the nest and the beach. It will be shown that the diffusion characteristics of the nesting beach play a fundamental role in the gas exchange of sea turtle eggs, a role that is functionally analogous to the role played by the bird s eggshell. Materials and methods The study of gas exchange by the sea turtle nest was carried out during the summers of 1972 and 1974 on the nesting beach of the green turtle (Cheloniu my&s) at Tortuguero, Costa Rica, and during the summers of 1973 and 1974 on the loggerhead turtle (Cclretta carella) nesting beach on Hutchinson Island, Florida. Freshly-laid sea turtle eggs were collected along a S-mile section of Tortuguero

4 SEA TURTLE NEST GAS EXCHANGE 21 beach and a 15-mile section of Hutchinson Island beach. At Tortuguero it was possible to locate females in the act of laying eggs and mark the nest for collection of eggs early the next morning. This was impossible on the Florida beach due to the distances involved, so the beach was walked at dawn and fresh nests were excavated in an attempt to find eggs. Eggs were removed from the nest cavity and counted. Those from each nest were packed in separate containers and placed in beach nests established at the Green Turtle Station, Tortuguero, and the House of Refuge, Hutchinson Island, respectively. Eggs were gently rinsed with fresh water to remove as much sand as possible and reburied by hand in lots of 100 and 120 eggs per nest. These numbers represent the average clutch size of the species studied (table 1). Care was taken to handle the eggs as little and as gently as possible. The artificial nests were excavated in the beach by shovel, and the nest chambers were spherically shaped by hand. The eggs were placed in the nest and a gas sampling tube leading to the surface was placed in the center of the egg mass. The egg mass was then covered with the moist sand previously excavated. Care was taken to avoid filling the air space between the eggs which is normally present. The depth of all nests (distance from the surface to the top of the egg mass) was about 40 cm (fig. l), and from the top of the egg mass to the bottom about 20 cm. The sand Ii11 was not packed; it was assumed that the prolonged incubation periods (45-70 days) would permit weathering of the original sand to a condition approximating that over natural nests, The gas sampling tube was PE 160 tubing with a piece of cheesecloth or tip of a plastic syringe on the end placed in the egg mass to provide dead space, with about 20 cm of the tubing extending above the surface of the beach. Ex~~mental controls were established by burying sampling tubes 40 cm deep in undisturbed beach. Sampling tubes were also placed in the relatively undisturbed natural nests of turtles. These natural nests were excavated through the shaft dug by the female, and the topmost eggs removed, so that the sampling tube could be placed as close as possible to the center. The eggs were then replaced in their original order and reburied. TABLE 1 Average clutch size of green turtle (Cheloniu) and loggerhead (Carefta) nests collected at Tortuguero, Costa Rica, and Hutchinson Island, Florida, respectively Number Mean of clutch Standard Species Season nests size deviation - _. ~ Chelonia m_vdas k21.9 I f 18.1 Caretta caretta k I f

5 22 R. A. ACKERMAN ~ Gas Sampling Tube 7. II cm Fig. I. A drawing, to scale, of an average green turtle (Chelonicr ~J&LY) nest with 100 eggs. The left panel shows more closely the asymmetrlcal nest chamber at the end of the entry shaft. The adult female would be oriented with her head to the left. The right panel depicts a nest with sampling tubes in position with the scale expanded. Gas samples were collected from artificial nests and natural nests, and from undisturbed beach at the same depth in the Tortuguero hatchery every 3-5 days and in the Hutchinson Island hatchery weekly throughout the incubation period. The tube dead-space volume was aspirated and discarded before collecting a sample of l-2 cm3. Syringes containing the gas samples were placed in a water-filled container and taken to a field laboratory for immediate 0, and CO, analysis with a Scholander 0.5 cm3 gas analyzer (Scholander, 1947). 0, and CO, values were also measured in the sand surrounding three Costa Rican green turtle and three Florida loggerhead turtle nests. Sampling in the sand was accomplished by glueing sampling tubes to a 30-cm metal shaft at 5-cm intervals and then burying the shafts (with tubing) at nest level so that one end of the shaft (and one sampling tube) just touched the nest perimeter and the other end extended away from the nest in one of three configurations : (1) parallel to the beach surface, (2) at a right angle from the top of the egg mass upward, (3) at a right angle from the bottom of the egg mass downward. Sand samples from Tortuguero, Costa Rica, and Hutchinson Island, Flbrida, and Ascension Island (a green turtle nesting colony in the South Atlantic) were analyzed for grain size distribution and sand air and water content with standard techniques (Hillel, 1971). Surface area of sand grains in a known volume of sand was also estimated. This was done by separating the sand sample into size classes based on grain

6 SEA TURTLE NEST GAS EXCHANGE 23 size (by sieving) and presuming that all the grains in a particular size class were spherical and had the same radius. The volume of each class was then measured by water displacement and the surface area of individual grains calculated using the sieve radius for the class. These surface areas were summed grain by grain and class by class to estimate total surface area in the sample. Results 0, and CO, partial pressure in the centers of two experimental and two natural nests of Chelonia and three experimental and one natural Curetta nest are shown in fig. 2. These results are representative of data collected from ten Cheloniu and ten Caretta nests. Gas tensions found in undisturbed beach are shown in table 2. PO2 in the nests of both Chelonia and Caretta begins to decrease about halfway through incubation, falling to the lowest level at hatching. Pco2 in the nests ofboth species rises in the same manner beginning about halfway through incubation and peaking at hatching. In all cases the nests were excavated at the time hatching appeared to occur to insure that this was indeed the case. There appears to be no difference between experimental and natural nests. Po, and Pco2 appear to be constant in beach gas. The Po, and Pco2 in the sand surrounding three Cheloniaand three Caretta nests are shown in figs. 3 &id 4. The eggs of both species were 50 days old at the time. This was immediately prior to hatching for the loggerhead nests and about lo-14 days prior to hatching for the green turtle. (Eggs from all the nests were opened after gas measure- TIME (days) Fig. 2. The Poland Pco, changes with time in the center of loo-egg Cheloniu and Curetra nests. Open circles represent natural nests while closed lsymbols represent artificial nests.

7 24 R. A. ACKERMAN TABLE 2 Oxygen and carbon dioxide tensions (torr) at a depth of about 40 cm in the beach at Tortuguero. Costa Rica, and Hutchinson Island, Florida P 02 (torr) P,.,> (torr) Tortuguero. Costa Rica (n = 33) 149.X Hutchinson Island, Florida (n = 22) 149.0* ? I.3 Gas samples were collected over a l-2 month period and are shown as mean values (dry), with standard deviations. Temperature was considered to be 28 C and samples were presumed to be fully saturated with water vapor. CHELONIA DISTANCE FROM CENTER (cm) jf ),. OOT,, io PARTIAL PRESSURE (torr) Fig. 3. Po, and Pco, in the nest center and sand surrounding 1 O&egg green turtle nests (Cheloniu) at 50 days of incubation. The closed symbols represent the mean (k SD) gas tensions in three nests. The solid line represents the predicted gas tensions using the model described in the text. The drawing is constructed to represent a vertical transect through the nest; the values through the top and bottom half of the transect are mirror images and when actually measured cannot be distinguished.

8 SEA TURTLE NEST GAS EXCHANGE 25 DISTANCE FROM CENTER (cm) EGGS ACHY PARTIAL PRESSURE {tow) Fig. 4. Po2 and Pco2 in the nest center and sand surrounding loo-egg loggerhead turtle nests (Corrttu) at 50 days of incubation. The closed symbols represent the mean ( f SD) gas tensions in three nests. The solid line represents the predicted gas tensions using the model described in the text. The figure is constructed as in fig. 3, ment and the embryos weighed to confirm the appropriate developmental age.) Po, decreased and PcoZ increased from the center to the edge of the nest and from the edge out into the sand. The results of standard sand analysis (Hillel, 1971) from Costa Rica, Florida and Ascension Island are shown in table 3. Sand from the Ascension beaches is clearly the coarsest, having the largest sand grains, while the Costa Rican beach has the finest grain size. On the other hand, the Tortuguero sand contains the largest amount of water in the field and air when dried. The Costa Rican beach also has the greatest surface area per unit volume, a reflection of the greater proportion of small sand grains in this particuiar beach. Discussion NEST GAS EXCHANGE Po, decreases and Pco, increases in the center of the sea turtle nest beginning about halfway through incubation. This pattern is similar to that found in the air cell of the

9 26 R. A. ACKERMAN TABLE 3 Some estrmates of the physical characteristics related to gas diffusion m sand from Tortuguero, Cosia Rica. Hutchinson island, Florida. and Ascension Island Sieve size Tortuguero. Costa Rica Hutchinson Florida Island, Ascenston Island Grain size distribution ( t,, of total dry weight of sand sample) IO I I X x Air volume (I,, dry volume) dry sand Water volume in the field (I,, total volume) Il Air volume (I,, total volume) in the field Estimated surface area (cm*.cm- ) Volume measurements represent oh of total initial sand sample. Estimated surface area is a measure ot surface of sand grains in I cm sand sample. bird egg (Wangensteen and Rahn, 1970/71). These changes are attributable to the increasing metabolic activity of the embryos and the slow diffusion of gases through the sand surrounding the nest. Since the eggs exchange gases by diffusion, substantial gas partial pressure differences must develop between the nest and the beach sand. Thus the respiratory environment in which the sea turtle embryo is developing must be established to a very large extent by the diffusion characteristics of the beach sand in which incubation occurs. The reported oscillations in PO, and Pco, in the center of Chelonia mydas nests at the end of incubation may be related to the hatching and digging activity of the embryos and hatchlings. Prange and Ackerman (1974) have measured a 3-fold increase in oxygen uptake as C~eZon~a eggs hatch with a return to earlier levels after hatching. Carr and Hirth (1961) have also shown that digging out by hatchlings is a cyclic phenomenon with intense periods of digging followed by quiescent periods. At the termination of the experiment (63 days) I reexcavated the nest and discovered that the eggs had indeed hatched and the hatchlings dug up beyond my sampling tubes.

10 SEA TURTLE NEST GAS EXCHANGE 27 NEST MODEL A simple two-compartment model will be used to describe nest gas exchange. It will be assumed that radial, steady-state diffusion of gases into and out of the nest is occurring. One compartment will describe diffusion within the discrete egg mass, i.e. for radii less than the radius ofthe spherical nest, and the second compartment will describe diffusion in thesandoutsideandsurrounding theeggmass, i.e. forradiigreaterthan the nest radius. Using the nomenclature of Piiper et al. (1971) we can make the following definitions : ti OZ.CO2 = mass flux (grams. see- I) of oxygen or carbon dioxide D 02. co* = diffusion coefficient (cm. set- ) of the gas inside the nest D002, co1 = diffusion coefficient (cm2. set- ) of the gas in the sand surrounding the nest r = nest radius (cm) X = radial distance (cm) from the center of nest in any direction Pxo*, co1 = partial pressure (torr) of the gas at radius x P oz, co2 = ambient (in undisturbed sand) partial pressure of the gas (torr) PN 02, co1 = partial pressure of the gas in the center of the nest (torr) R = gas constant (cm3. torr g-. K- ) T = absolute temperature ( K) A = surface area (cm2) We can describe diffusion in the sand compartments more specifically by using oxygen as the diffusing gas and letting the oxygen flux across a given spherical surface at radius x be equal to the diffusion coefficient of oxygen in sand multiplied by the oxygen partial pressure gradient across the surface. The net oxygen flux must equal the oxygen consumption of the nest (per unit time) and we obtain : -njioz = -Do,> (1) Rearranging and integrating yields : PO1 = K- &I,*RT 47cxDcio2 where K is an integration constant that can be evaluated by allowing x to approach infinity. In that case Pxo2 = PIOl - &I,~RT 47~xD0~~ and the constant represents the ambient partial pressure of oxygen in sand with no nests present. The nest compartment (x 4 r) presents a different situation.. As oxygen diffuses (2)

11 28 R. A. ACKERMAN from the nest wall into the center of the egg mass, an amount is consumed as each element of surface, 471x2, is passed. There is a progressive depletion of oxygen as the center is approached due to the oxygen consumption of the eggs. What must be accomplished first is an estimate of the oxygen consumption of any sphere of radius x as x approaches r, the nest radius. The oxygen consumption per unit nest volume is - lvl,,,/$rr and oxygen consumption of any sphere of radius x (where x < r) is some fraction of this : -k Or 4 4 ~. 771x3 = - Mc12x3jr3 T7cr3 This is the oxygen consumption per unit volume times the volume of a sphere of radius x. The negative sign is related to the direction in which diffusion occurs. Then proceeding as before, the quantity of oxygen (- Mo2x3/r3) diffusing across a surface of radius x from the center ofthe nest (but less than r, the nest radius) is: -kolx3/r3 = - Dloz (5) Rearranging and integrating yields : n;l,2rt~2 pxo2 = 8D1-. meti + K' (6) 02 where K is now an integration constant that can be evaluated by letting x go to zero, giving : pxoz kfo2rt~2 = PNO, Dr02nr3 (7) The oxygen partial pressure at the center of the nest can be evaluated further by recalling that at x = r, Pxo2 is defined by eq. (3). Substituting eq. (3) into eq. (7) yields:. + 2% PNOz = Doo2 (8) Equation (8) can be simplified further by solving for &lo2 : * (Plo2 - pq (91 Equations (8) and (9) describe the oxygen tension at the center of the nest in terms of the ambient oxygen tension, the nest radius, the nest oxygen consumption and ap parent diffusion constants for oxygen in the nest and sand. Oxygen tension at any given distance x from the center of the nest may be evaluated by solving eq. (8) and eq. (9) for the appropriate distance. A similar model can be derived describing the carbon dioxide concentration gradient in the nest-beach system. In this case we let &I,, remain positive because

12 SEA TURTLE NEST GAS EXCHANGE 29 the direction is reversed. Given the same assumptions and following the same derivation the following two-compartment models result: for x 2 r Pxcol = Mco,RT 4Doco17rx + P,,J (10) for x 5 r PXcoz = PNCOz- Mco,RTx2 8DicoJrrr3 (11) for x = r PNcO1 = and solving for k,oz: 87cr &I_ = -. ~-22 Do,, Dice RT 2Dk02 + D%02. (P%& - %o*) (12) (13) MODEL PARAMETERS The diffusion model presented here describes the oxygen and carbon dioxide tensions inside the sea turtle nest and in the sand surrounding the nest in terms ofthe metabolic activity of the total egg clutch (&I,, and &I,& the nest shape or geometry (nest radius), and the physical nature of diffusion in the air spaces of the sand and nest (DI and Do). The model can be fitted to the data in several ways. I chose to measure all the parameters except the diffusion coefficients and then fit the model to the data by estimating the diffusion coefficients. All other parameters (Pxo,, Pxco2, h;loz, &I,-oJ except nest radius could be measured more or less directly. ho2 and h;lcoz were measured as described elsewhere (Ackerman, 1975), and these values for eggs at the appropriate incubation times are presented in table 4. Measurement of the nest radius presents a problem since turtles do not dig perfectly spherical nests (Carr and Ogren, 1960) and empty nests collapse when disturbed. Therefore, geometric models of nests were constructed and total volume calculated. The nest radius was TABLE 4 Rates of metabolism used to fit the diffusion model Age (days) Weight of Weight of I&, egg embryo (g) (g) (mlihr) Clutch size (no. of eggs) Moz of nest (mlihr) Chelonia Caretta Rates were measured for individual eggs as described in the text. Models were fitted at day 50 of incubation for both Chelonia and Carerra. This is just prior to hatching for Caretfa and 1615 days prior to hatching for Chelonia.

13 30 R. A. ACKERMAN calculated using this volume. Two models were examined : the first presumed the egg layers to be directly atop one another yielding cubic packages of 4 eggs. This gave a maximum estimate of the nest radius. The second model presumed the egg layers were shifted with respect to each other so that an egg above lay between 3 eggs below. This model yields a tetrahedral package of 4 eggs and a minimum estimate of the nest radius. Table 5 describes the maximum and minimum nest radii of the nests with average egg numbers for both turtle species on the beaches of interest. Po2 and Pcol in the nest and sand were measured as described earlier and are presented in figs. 3 and 4. TABLE 5 Estimates of sea turtle nest geometry calculated from two packing models. Egg dimensions are means measured in the field. Model Eggs in Egg Total nest Nest Nest Gas volume clutch radius volume radius surface in nest area (n) (cm) (cm ) Icm) (cm ) (,, of total vol.) Cheloniu tetrahedral cubic too Currrra tetrahedral cubic MODEL SOLUTION The model equations were then programmed into a Tektronic Model 31 digital computer, and the equations solved for gas partial pressure at various distances from the nest center. Various estimates of the diffusion coefficients for sand, Do, and nest, DI, were tested by substituting each into the model so that the predicted partial pressures coincided (by eye) with the measured partial pressures. The fit of the measured to the theoretical values could be checked by noticing that eq. (3) which describes partial pressures in sand, is an equation for a straight line when the partial pressures at any radius (Px,,) is plotted against the inverse of the radius (I/x) yielding a slope of - I$,J47rDo. Regression of Px,, on l/x gave correlation coefftcients that always exceeded 0.90 and usually exceeded The estimated values of the slope (determined from estimated values of h;lo2 and Do) always fell within the 95 % confidence intervals of the measured slope. The gas tensions predicted by the models are shown in figs. 3 and 4 as solid lines. The solid line inside the nest proper represents the gas tensions that result when the minimum estimate of radius is used. The maxims nest radius results in a flatter curve that does not fit the data as weli. The measured gas tensions are the closed circles and represent means (+ SD) of three nests.

14 SEA TURTLE NEST GAS EXCHANGE 31 TABLE 6 Fitted diffusion coefficients of oxygen and carbon dioxide in the nest and sand surrounding the nest of sea turtles Diffusion coefficient inside nest (cm* I set- ') Diffusion coefficient in beach sand (cm I set- ') 02 CO, 02 CO2 _..._...~~. -~ -~ ~~-._-... _~ _... Ckloniu :, of Dair (Costa Rica) CLfW/lU 7, of DJic (Florida) Temperature of the system was considered to be 30 C and binary diffusion constants (D,J were fitted as described in the text. DIFFUSION COEFFIClENTS IN NEST AND SAND The fitted values of DI (nest) and Do (sand) for oxygen and carbon dioxide are given in table 6. The nest diffusion coefficients (Do) are about cm2. see- r in all the nests studied. The diffusion coefficient of oxygen in air at 30 C by comparison is cm. set-. The diffusion coefficients for 0, and CO, in the sand of the Costa Rican beach are about one-third of those for the nest while in the sand of the Florida beach the estimated diffusion coefficients are two-thirds of the diffusion coefficients found in the nest. The diffusion coefficients ought to be proportional to the amount of air present in the nest and beach sand. Table 3 gives the gas volume (7; of total volume) present in the various beaches. Table 5 shows the gas volume present in the nest. The rate of diffusion in sand is quite clearly much slower than that inside the nest. The nest diffusion coefficient is about 20 % of that for air while the minimum estimate of air in the nest is 26 %. Given the geometric uncertainty the agreement is good, and suggests that the tetrahedral packing model is the more appropriate estimate ofegg packing as one might intuitively expect. The diffusion coefficient of sand, however, appears to be consistently lower than predicted from the volume of air present. Further, the Tortuguero beach ought to be more permeable to gas movement (more gas present) than the Florida beach, yet the Tortuguero diffusion coefficient is only half that of the Hutchinson Island beach. The sand of both beaches cont;,tins about 30 % air by volume, yet the diffusion constants are only 12 y0 and 8 y0 of those in air. A factor called tortuosity (Hillel, 1971) is probably responsible for this deviation. Tortuosity is a measure of the path a molecule must follow diffusing through soils. The finer the grain size distribution, the smaller the pore size and the greater the tortuosity, the inverse of which is used to correct the expected diffusion coefficient

15 32 R. A. ACKERMAN to get the actual coefficient. Unfortunately there appears to be no genera1 theoretical means of predicting tortuosity. The sand grain surface area (per 1 ml sand) has been estimated and is presented in table 2 as an indirect measure of resistance to flow. It is clear that the Tortuguero, Costa Rica, beach sand has much more surface area than the other beaches and therefore ought to retard diffusion to a greater extent as in fact it does. An interesting and perhaps coincidental observation is that the ratio of surface area (Tortuguero) to surface area (Hutchinson) is about the same as the ratio of their fitted diffusion coefficients. Based on this analysis the Ascension Island nesting beach ought to have greatest conductivity to gas; this has not, however, been measured in situ. MODEL ASSUMPTIONS Oxygen and carbon dioxide partial pressures in the nest and sand surrounding the nest appear to be predicted reasonably well by the model equations. Two major assumptions were made in deriving the gas exchange mode1 : (1) that diffusion could be described in one dimension and (2) that steady-state diffusion occurs. The genera1 fit of observed to predicted gas partial pressure supports the first assumption but rnc ;e convincing evidence can be collected by measuring gas concentrations in various directions from the nest and comparing the values. Table 7 shows data collected from three Caretta nests. These data are taken from vertical transects from the beach surface through the center of the nest down to about 30 cm below the lower surface of the nest. Similar data were collected from Chelonia, but the transect is L-shaped with a TABLE 7 0, and CO, values in samples taken at various distances out into the sand surrounding Caretta nests of about 50 days of incubation Nest: I 2 3 P co* (torr) P (tzr) P ct o rf, P (Zrr) P co; (torr) P (todrr) Direction : up down up down up down up down up down up down Nest center Nest edge cm cm cm cm cm Transect direction is given above each column. Regression analysis (as described in the text) cannot dis- tinguish between the transects from an individual nest.

16 SEA TURTLE NEST GAS EXCHANGE 33 vertical transect from the beach surface to nest center, then a horizontal transect at 90 out to about 30 cm into the beach. There are two sets of data for each nest, each representing a different half of each transect. The data can be compared by plotting Pxo2 against l/x as described earlier and doing a regression analysis. Correlation coefficients exceeded 0.90 and-the confidence intervals of the slopes overlapped. The data appear to be equivalent no matter in what direction they are collected. The data from Chelonia nests are similar. Diffusion into both Carerta and Chelonia nests does indeed appear to be spherically symmetrical. The second assumption, that of steady-state diffusion, is more difficult to deal with. Crank (1957) has derived an equation describing a continuous point source (or sink) in an infinite medium which can be used to evaluate the establishment of a steady state. In this case the nest is the point source or sink for carbon dioxide and oxygen respectively, and the beach is the infinite medium. The observation that the distribution ofgas tensions in sand are nearly symmetrical supports the idea ofthe beach behaving as an infinite medium. From Crank s equation, it appears that a steady state is approached during the last week of incubation. This is due in part to the very slow rate of growth (Ackerman, 1975) of the embryos during that period. At least two other parameters have been neglected by the model : temperature and pressure gradients. It has been shown that a temperature gradient exists between the nest and the sand surrounding it (Carr and Hirth, 1961) but it is probably not more than 24 C. The effect of the temperature gradient would be to vary the diffusion coefficients with distance from the nest (Crank, 1957). The importance of such an effect is not clear but is probably not large. The presence ofa total pressure gradient would result in convective currents and therefore bulk flow of gases in the sand; however, Prange and Ackerman (1974) have discounted pressure gradients as a significant source of gas exchange, and that would seem to be confirmed by this paper. ANALOGY BETWEEN BEACH SAND AND BIRD EGGSHELL The sea turtle nesting beach appears to be functionally analogous to the bird s eggshell in the sense that it too presents a fixed resistance to the diffusion of gases. The beach, by permitting the accumulation of carbon dioxide and reduction of oxygen in the nest, presumably allows the adjustment of embryonic acid-base balance and gas transport systems to a status appropriate to lung breathing at birth, as is apparently the case in the chicken embryo (Erasmus et al., 1970/71; Wangensteen, 1972). The analogy can be carried even further if we recall the two model equations describing nest gas exchange for oxygen and carbon dioxide: (13)

17 34 R. A. ACKERMAN and lq=t(rrr. RT!. (PI& - PNOL) (9) Equation (13) can now be divided by eq. (9) which, after cancelling common terms, yields : (14) Equation (1) simplifies to : When we let MCo2fMOz = R, the respiratory exchange ratio and P~co~ = 0. The complex term relating the diffusion constants becomes 0.78 when the appropriate values are substituted in. Equation (15) is interesting because it is the same equation derived by Wangensteen and Rahn (1970/71) to describe gas exchange by the bird embryo where pnco2 and PN~* were the carbon dioxide tensions and oxygen tensions in the air cell of the bird egg. I am using my terms as if they were analogous to the air cell partial pressure and treating a clutch of about 100 sea turtle eggs as if it were functionally a single egg! Equation (15) can be rearranged : PNCO = o.7*. wo2 - PNo2) (16) Since PI,* is constant, PN,, can be plotted against pncoz with slope equal to (R/O.%). The line produced is the diffusion R-line and intercepts the oxygen partial pressure axis at the ambient partial pressure of oxygen corrected for the vapor pressure at the appropriate temperature. 1 have measured nest temperatures at about 2X-30 C. This is in the same range reported by Carr and Hirth (1961) and for beach temperatures by Hirth (1961). The nest temperature wa s set at 30 C, and since air in the nest is fully saturated with water, the ambient partial pressure of oxygen is about 150 mm Hg. The gas tensions measured in the centers of all the Cheloniu and Curetta nests studied are plotted in fig. 5 along with a diffusion R-line (for gas) equal to 0.7. That the data appear to fall along an air R-line of 0.7 is not surprising since a major energy source of eggs are lipids. Wangensteen and Rahn (1970/71) have shown that gas tensions in the air space of chicken eggs fall between an R-line of 0.7 and 0.8, coming closer to 0.7 as hatching nears, and Romanoff (1967) reports that chicken eggs metabolize lipid almost exclusively near hatching. It does appear that a clutch of about 100 eggs exchanges respiratory gases in a manner quite similar to that of the bird egg, with a fixed resistance to diffusion interposed between the embryos and their external environment. It is tempting to suggest that the presence of similar gas partial pressure differences during the embryonic development of two rather dissimilar groups of vertebrates is not fortuitous and serves a physiological function

18 SEA TURTLE NEST GAS EXCHANGE 35 PN o2 (torr) Fig. S. The Pol Pco2 diagrams of gas samples from the four C/w/onic~ and four < urcrru nests shown m lig. 2. The solid line is a diffusion R-line with exchange ratio R = 0.7. important to both reptilian and avian embryos. Not only are similar gradients produced across the bird eggshell and in the sea turtle nest, but the relative time course of gradient development is similar. That is, at proportionally the same time in the incubation period, the gradients are the same size. Among all the birds, despite a very wide range in egg size and incubation time, the metabolic demands ofthe embryo and the diffusive resistance ofthe shell are matched so that gas partial pressure differences, where they have been measured, are very nearly the same (Rahn er al., 1974). Since the sea turtles produce a similar respiratory environment for their eggs that is always saturated with water vapor and since the :esistance of the eggshell to diffusion is very reduced, it would be interesting to speculate that the turtle egg represents the primitive condition and that the bird eggshell represents an evolutionarily controllable resistance that has been fixed so as to reduce water loss while maintaining respiratory gas tensions within certain narrowly defined limits. In this respect it would be extraordinarily interesting to study the mound builders (Meyapodidae), birds that bury their eggs in sand or soil. Raltin (1969) reports gas tensions inside the mound of Alectura that are similar to those found inside the sea turtle nest. What role does the eggshell play here? IMPLICATIONS OF GAS EXCHANGE IN FIXED RESPIRATORY ENVIRONMENTS The bird has at least the theoretical possibility of adjusting the porosity of its shell in response to or to take advantage of environmental conditions. The gas exchange of

19 36 R. A. ACKERMAN PARTIAL PRESSURE (tom) 70 CHELONIA EGGS DISTANCE FROM CENTER (cm) Fig. 6. Gas tensions inside a loo-egg green turtle nest. The line labelled sand represents gas tensions in a nest with normal air space filled with sand. The line labelled air represents gas tensions in a nest with the normal air space between eggs present. the sea turtle egg is more circumscribed since the diffusion characteristics of a particular nesting beach are fixed, and the conductance of the eggshell to gases is already very high and can only be decreased, reducing the 0, tension and increasing the CO, tension presented to the embryo. The female sea turtle can, however, influence the gas exchange of her nest by selecting beaches appropriate to her clutch size or matching clutch size (in a evolutionary sense) to a particular beach. Sea turtles habitually return to the same beach year after year and are capable ofextraordinary feats of homing (Carr, 1967). Perhaps one of the evolutionary forces influencing this behavior has been the need to maintain a constant and stable respiratory environment for the eggs, maximizing hatching success by returning to a beach that provides suitable gas exchange. The female sea turtle can also alter nest gas exchange by the way in which she constructs the nest. The nest is constructed in what appears to be an evolutionarily conservative, highly stereotyped manner (Carr and Ogren, 1960) so that the eggs are free of sand. C&r (1967) has suggested that this behavior must be of considerable evolutionary importance but was unable to provide an explanation. Ackerman (1975 and in preparation) has shown that embryonic growth and survival are markedly affected by the changes in respiratory gas tensions that deviate away from those normally observed in the nest. Given this information, it is interesting to use the nest gas exchange model to examine the effects of filling the space among eggs with sand rather than air. Figure 6 illustrates the effect of packing the eggs in the nest chamber in sand and using the appropriate diffusion coefficient to determine gas

20 SEA TURTLE NEST GAS EXCHANGE 37 tensions in the nest. A greater hypoxia and hypercapnia in the center of the nest and a much larger gas partial pressure difference between the center of the egg mass and the edge is produced. Individual eggs, depending on position, would develop in very different respiratory environments with consequently different embryonic growth rates and survival. The female sea turtle in constructing her nest in a fixed medium in the manner in which she does may well be creating an optimal, reasonably uniform respiratory environment for her entire clutch and thereby enhancing the probability of her eggs surviving to reach the ocean. The sea turtle nesting beach is fundamentally important in shaping the gas exchange of sea turtle eggs and probably influences the nesting behavior of female sea turtles. While similarities between the sea turtle clutch and the bird s egg and between the nesting beach and the bird s eggshell are striking, the nesting beach places fixed constraints on any tissue incubating within and probably plays an important role in explaining why all sea turtles lay about 100 eggs in a very carefully constructed nest. Acknowledgements I would like to thank Dr. Henry Prange and Dr. Archie F. Carr, Jr., of the University of Florida for their encouragement throughout this study, and Dr. Hermann Rahn and Dr. Charles Paganelli of the State University of New York at Buffalo for their invaluable criticism of this paper. The author is also greatly indebted to the Society of Sigma Xi, the Caribbean Conservation Corporation, and the Florida Department of Natural Resources for their support, and to Dr. Marvin Pokrant, Department of Physics, University of Florida, for his help with the mathematical formulations. References Ackerman, R. A. and H. D. Prange (1972). Oxygen diffusion across a sea turtle (Chrlonia m$~s) egg shell. Comp. Biochem. Physiol. 43A: Ackerman, R. A. (1975). Diffusion and the gas exchange of sea turtle eggs. Ph. D. Dissertation, University of Florida, Gainesville, Florida. Baltin, S. (1969). Zur Biologie und Ethologie des Telegalla-Huhns (Alectura lathomi (Gray)) unter besonderer Beriicksichtigung des Verhaltens wahrend der Brutperiode. 2. Tierpsychol. 26: Carr, A. F. and L. Ogren (1960). The ecology and migration of sea turtles. 4. Bull. Am. Mus. Nar. Hist. 121: 749. Carr, A. F. and H. Hirth (1961). Social facilitation in Green Turtle siblings. Anim. Behau. 9: Carr, A. F. (1967). So Excellent a Fishe. New York, Natural History Press, 248 pp. Crank,.I. (1957). The Mathematics of Diffusion. London, Oxford University Press, 347 pp. Erasmus, B. dew., B. J. Howell and H. Rahn (1970/71). Ontogeny of acid-base balance in the bullfrog and chicken. Respir. Physiol. 11: Hillel. D. (1971). Soil and Water. New York, Academic Press, 288 pp. Hirth, H. F. (1961). The ecology of two lizards on a tropical beach. Ecol. Monogr. 33: 83-l 12. Piiper, J., P. Dejours, P. Haab and H. Rahn (1971). Concepts and basic quantities in gas exchange physiology. Respir. Physiol. 13:

21 38 R. A. ACKERMAN Prange, H. D. and R. A. Ackerman (1974). Oxygen consumption and mechanisms of gas exchange of Green Turtle (Cheloniu m~dus) eggs and hatchlings. Cop&, pp. 75X-763. Rahn. H., C. V. Paganelli and A. Ar (1974). The avian egg: air-cell gas tension, metabolism and incubation time. Respir. Phq siol. 22: Romanoff, A. L. (1967). Biochemistry of the Avian Embryo. New York, John Wiley and Sons, 398 pp. Scholander. P. F. (1947). Analyzer for accurate estimation of respiratory gases in one-halfcubic centimeter samples. J. Biol. Chem. 167 : I-l 5. Wangensteen, 0. D. and H. Rahn (1970/71). Respiratory exchange by the avian embryo. Rrspir. Physiol. I I : Wangensteen, 0. D. ( 1972). Gas exchange by a bird s embryo. Respir. Physrol. 14: