THERMAL AND WATER RELATIONS OF EMU EGGS DURING NATURAL INCUBATION'

THERMAL AND WATER RELATIONS OF EMU EGGS DURING NATURAL INCUBATION' WILLIAM A. BUTTEMER,2 LEE B. ASTHEIMER,2 AND TERENCE J. DAWSON School of Zoology, University of New South Wales, Kensington, Australia 2033 (Accepted 4/12/88) Naturally incubated emu eggs lost only 10% of their initial mass over their 54.5- day incubation period. This low rate of water loss was largely due to emu eggs having a conductance to water vapor that was 65% of that predicted for eggs of identical mass and incubation period. Temperatures of naturally incubated artificial eggs rose steadily from 32 to 34 C over the first 10 days of incubation, remained at 34 C for the next 15 days, then rose gradually to 36 C by day 35 and remained at that temperature until hatching time. The thermal sensitivity of embryonic oxygen consumption was inversely related to both embryo age and egg temperature. Because eggs laid after the start of incubation experience warmer initial incubation temperatures, late additions should grow more rapidly than other embryos in the clutch. We have estimated that eggs added 3, 6, and 9 days after the initiation of incubation will require 1, 2, and 3 days less, respectively, to reach the plateau of the ratite growth curve. Consequently, the pattern of rising incubation temperatureshould facilitate hatching synchronization for all members of the clutch. INTRODUCTION Normal development of avian embryos requires maintenance of egg temperature and nest humidity within specific limits during incubation (Landauer 1967). Such conditions are usually provided by one or both parents through nest construction and incubatory behavior. As adults engaged in the latter activities sacrifice their time available for feeding, they are confronted with the conflict of satisfying the maintenance requirements of their developing progeny at the expense of their own energetic needs. Partitioning of time and energy for reproduction by adults is especially evident in emus (Dromaius novaehollandiae). Male emus, like all other ratites (sensu Cracroft 1974) except ostriches, assume sole responsibility for incubation (Handford and Mares 1985). The incubation pe- ' We thank Ray Williams, Jeff Vaughan, and Shane Maloney for facilitating all aspects of our field study. Bruce Young generously furnished a datalogger for the initial part of the study. Financial support was provided by the Australian Research Grants Scheme and by the University of New South Wales. 2 Present address: Department of Zoology, University of Washington, Seattle, Washington 98195. Physiol. Zool. 61(6):483-494. 1988. c 1988 by The University of Chicago. All rights reserved. 0031-935X/88/6106-87102$02.00 483 riod of emus occurs during winter and lasts about 8 wk, during which time the attending males are presumed neither to eat nor to drink (Davies 1976). Although it has been suggested that incubating males conserve energy through reduced thermoregulatory expenditures (Davies 1976), the extent to which this occurs is limited by the thermal requirements for successful embryo development. This latter concern includes the need to minimize hatching asynchrony of eggs incubated for disparate periods. A female emu may continue adding eggs to a clutch up to 13 days after the male begins incubation (Fleay 1936). As emu chicks are precocious and become ambulatory within 2 days, the attending male must shift duties from egg incubation to brooding and shepherding the mobile chicks. Consequently, fledging success would improve if late-laid eggs required less incubation time than those laid earlier. Insight into the emu's incubation process thus requires evaluation of both the physical conditions attending this period and the response of embryos to their variation. Accordingly, we present information concerning the thermal and hygric environment of emu eggs undergoing natural incubation and include an assessment of the effects of thermal variation on embryo metabolism.

484 W. A. BUTTEMER, L. B. ASTHEIMER, AND T. J. DAWSON MATERIAL AND METHODS EGGS AND STUDY SITE We conducted studies on a colony of 10 adult emus (six males, four females) maintained 50 km north of Sydney at the Cowan Field Station of the School of Zoology, University of New South Wales. The emus ranged freely in a large (100 X 200 m) section of dry sclerophyll forest. During the 1985 breeding season, the enclosure was surveyed usually daily for eggs which, when found, were labeled with an indelible marker, weighed to the nearest 0.01 g in an adjacent laboratory, and then returned to their original site. These eggs were subsequently weighed at least once each week to gauge their rate of water loss both before and during incubation. EGG WATER VAPOR CONDUCTANCE AND HYGROMETRY We evaluated water vapor conductance (GH20) of six eggs collected from two clutches using methods similar to those of Ar et al. (1974). Our measurements of GH20 are based on daily weighings of eggs held in desiccators at 26 C for 5 days. Subsequently, these eggs were emptied of their contents and fashioned into egg-diffusion hygrometers following the procedures of Rahn, Ackerman, and Paganelli (1977a). This involved filling the eggs with silica gel and sealing them with detachable nylon screw caps. We calibrated these hygrometers, both before and following their placement in nests, by determining their average daily rate of mass gain when exposed to atmospheres with vapor pressures of 23.8, 17.9, or 7.7 Torr. These conditions were achieved by placing the hygrometers for 6 days each in desiccators maintained at 25 C and containing distilled water, or a saturated NaCl solution, or a saturated solution of MgCl2, respectively (Winston and Bates 1960). Nest humidity was measured by placing a previously weighed egg-diffusion hygrometer among a given clutch of eggs. Hygrometers were left in the nests for at least 3 days before being removed, reweighed, and filled with fresh silica gel. As the rate of mass gained by the hygrometers is affected by temperature and barometric pressure (Paganelli, Ackerman, and Rahn 1978), differences between incubation temperature (see below) and that of their calibration were accounted for when determining nest vapor pressure for a given interval. Although we did not monitor barometric pressure during the field study, we adjusted the hygrometer calibrations for pressure differences expected for the field station's altitude (217 m) using the appropriate table in List (1971). Ambient vapor pressure was evaluated with an egg hygrometer placed at ground level beneath a caravan adjacent to the incubating birds. TEMPERATURE MEASUREMENTS We prepared three emu eggs as temperature probes by first evacuating their contents, sawing them in two, and lining their inner surfaces with a layer of fiberglass matting coated with resin. Each egg then received a miniature temperature transmitter (J. Stewart Enterprises), which was first rendered buoyant by surrounding its borders with Styrofoam before encapsulating it with a polyvinyl wax. After the transmitters had been inserted, the egg was joined with resin and filled with mineral oil. This permitted the transmitters to remain near each egg's upper surface, as early embryos do, despite egg turning by the incubating male. The transmitter signals were converted to a voltage output through a Telonics receiver and digital processor. We calibrated this telemetry system by measuring the voltage associated with the immersion of these eggs in water ranging from 10 to 45 C as detected by a certified mercury in glass thermometer. These eggs were recalibrated using the same procedures at the end of the study. For the one egg which displayed slight divergence in these calibrations, all data were interpolated linearly over time between the two calibration dates. Once incubation was under way, each transmitter egg was placed among a clutch of eggs and its output recorded once every 10 min on a Datataker 100F datalogger. Air temperature was measured using a shaded 36-gauge copper-constantan thermocouple positioned 0.5 m above the ground at a site 25 m from the incubating emus.

THERMAL AND WATER RELATIONS OF EMU EGGS 485 EMBRYO THERMAL SENSITIVITY During the course of our studies, two nests were abandoned and the eggs were transferred to an incubator held at 35 C. Humidity was adjusted such that eggs lost mass at a similar rate (approximately 1 g per day) to those undergoing natural brooding. The rate of oxygen consumption (Vo2) of these eggs and many of those being naturally incubated was periodically measured using methods described by Hoyt, Vleck, and Vleck (1978). We selected three eggs on a given day, placed them individually in open 4-liter paint cans, and maintained them at 30, 34, or 38 C for at least 2.5 h within a constanttemperature cabinet. We had established earlier that this time was sufficient for emu eggs to reach thermal equilibrium as determined by monitoring egg surface temperatures (Hoyt et al.). Following this equilibration period, we sealed the cans and used a syringe to inject 120 cm3 of cabinet air into each through three-way valves. The syringe plunger was pulled back and forth several times to ensure gas mixing before a 100-cm3 air sample was withdrawn for determination of initial oxygen concentration (F1). Air pressure within the cans was then equilibrated to atmospheric pressure by venting the extra 20 cm3 before closing the valves. A second 100-cm3 sample was withdrawn after sufficientime had elapsed to reduce the oxygen concentration in the chamber to between 20.0% and 20.5%. Fractional oxygen concentration of initial and final gas samples were measured with a Servomex oxygen analyzer (Model OA-272) after absorbing water vapor and CO2 with Drierite (anhydrous CaSO4) and Ascarite (sodium asbestos anhydride), respectively. We calculated oxygen consumption using the formulation of Vleck (1987): V02 = V (FI - FE)/(1.O - FE)t, (1) where V is the dry gas volume (STPD) in the chamber, F, and FE are the initial and final fractional oxygen concentrations in the chamber, and t is the time interval between initial and final gas sampling. During these determinations, an empty fourth chamber was treated similarly and served as a blank. We measured each egg's V02 at 30, 34, and 38 C on the same day and evaluated their thermal sensitivity by computing their temperature coefficient (Qio) for the temperature intervals 30-34 and 34-38 C through Q1o = (R2/R, ),O-oi2 ), (2) where T2 and T, are the higher and lower temperatures, respectively, at which the corresponding rates of oxygen consumption (R2 and R1) were measured. Eggs that were being naturally incubated were subjected to metabolic measurements no more than once per week. Unless stated otherwise, values given represent means plus or minus their standard error. RESULTS EGG WATER LOSSES Females laid eggs at approximately 3- day intervals (2.9 + 0.1 days; n = 27). Three clutches were incubated, one for 14 days and two until all eggs hatched. Prior to incubation, eggs lost 174.8 + 11.2 mg of mass daily (n = 21). As the chick's respiratory quotient during development results in the masses of oxygen consumed and carbon dioxide produced balancing one another, all mass losses represent water losses from the egg (MH20)(Drent 1975). Mean MH20 ofemu eggs rises substantially during the course of incubation (fig. 1). Measurements taken during the first quarter of incubation (<13 days) differ significantly (P <.001) from those obtained from the last quarter (>39 days), averaging 906.9 + 14.5 (n = 39) and 1,242.9 + 21.3 (n = 36) mg day-', respectively. Assuming that our last measurement of MH20 for each egg was representative of its rate of water loss through hatching, we estimated egg masses at the time of hatching and total water lost by eggs of both clutches. These values did not differ significantly between nests (P >.1). The total mass lost averaged 9.9% + 0.3% of initial egg mass for eggs of both clutches, with individual estimates ranging from 8.1% to 10.7%. Although incubation of these nests was initiated 2 wk apart, the mean daily air temperature, ambient va-

486 W. A. BUTTEMER, L. B. ASTHEIMER, AND T. J. DAWSON 1400 1300 - o "o 1200 0-0 o oo oc E 1100 " 1000 o _ 900 oo - o - o 800 0-700 I I I 0 10 20 30 40 50 60 Days Incubated FIG. 1.-Daily rate of emu egg water loss over the course of natural incubation. Mean values (circles) of eggs from up to three different clutches are bounded by their 95% confidence limits (horizontalines). por pressure, and nest vapor pressure did not differ significantly between them (table 1). Our measurements of water vapor conductance (GH20) averaged 45.7 + 1.6 mg day-' Torr-' for the six eggs we sampled (mean egg mass = 673 + 18.2 g). This value, which is corrected to 25 C and 760 Torr as described by Paganelli et al. (1978), is indistinguishable from that measured by Vleck, Vleck, and Hoyt (1980b). INCUBATION PERIOD Incubation was initiated before clutches were completed and appeared to be uninterrupted from then until approximately 8 wk later. It began after the sixth egg was laid in each of the three nests, with two eggs added subsequently to the clutch that was eventually abandoned and three eggs each to the other clutches. The latter additions occurred 1, 4, and 7 days after the start of incubation in both nests. These TABLE 1 DATA SUMMARY FOR TWO EMU NESTS THAT SUCCESSFULLY FLEDGED CHICKS Variable Nest 1 Nest 2 Date of incubation initiation July 31 August 14 Clutch size 9 9 Initial egg mass (g) 656.7 629.1 (+7.7; N= 9) (+3.4; N= 9) Initial mass lost during incubation (%) 9.4 10.3 (+.5; N = 8) (+.2; N= 8) Nest vapor pressure (torr) 15.8 16.3 (+.8; N= 9) (+.4; N= 8) Ambient vapor pressure (torr) 9.1 9.4 (+.7; N = 11) (+.6; N = 10) Daily air temperature (C) 12.5 13.1 (+.4; N = 31) (+.4; N = 33) Incubation period (days) 54.7 54.4 (+.3; N= 8) (+.6; N= 8) NOTE.-Values for variables that are accompanied by parenthetical values representheir mean, standard error, and sample size, respectively.

THERMAL AND WATER RELATIONS OF EMU EGGS 487 clutches had similar mean incubation periods for eggs that hatched (table 1), but, in each clutch, eggs laid after the start of incubation required less time to hatch. For example, in nest 2, all eggs hatched within 42 h of one another, but incubation times extended from 50.5 to 56.5 days, with the egg laid last requiring the least incubation. INCUBATION TEMPERATURES During the first 10 days of incubation, mean egg temperatures (Tegg) rose steadily from about 32 to 34 C in all three nests (fig. 2). From days 10 through 25, Tegg's remained at about 34 C followed by an ascension to a plateau of approximately 36 C from day 35 onward (fig. 2). Part of the variability in mean Tegg values likely represents daily differences in the location of the transmitter egg within the clutch. We evaluated this by placing three transmitter eggs in nest 2 on day 44 of its incubation. The average Ta on this day was 10.7 +- 0.3 C and the Tegg's averaged 35.6, 35.7, and 36.0 C for the three eggs. The intraclutch Tegg differences of these three eggs are even greater when evaluated on an hourly basis (fig. 3), perhaps reflecting periodic egg shifting between central and peripheral nest locations by the incubating bird. EMBRYO THERMAL SENSITIVITY The effect of temperature on embryo Vo2 varied markedly over the course of incubation. Their thermal sensitivity, based on evaluation of Qio from Vo2 measurements, was greater at lower temperatures (30.0-34.0 C) than at warmer ones (34.0-38.0 C), and sensitivity declined significantly as incubation proceeded (figs. 4, 5). Between 30.0 and 34.0 C, the relation between embryonic Qio and time to hatching is described by the equation Qo = 1.03 + 0.053 (DBH) (r =.81;Sb = 0.006; Sy.x = 0.362; n = 39), (3) 38.0-0 0 0 4- E 0- + + 0o o *o0 + 0 Emu 11 32.0 --oo * Emu 5 C0 + Emu 7 30.0 0 10 20 30 40 50 60 Days Incubated FIG. 2.-Mean daily internal temperature of oil-filled emu eggs over the course of natural incubation in three emu nests (see text for details).

488 W. A. BUTTEMER, L. B. ASTHEIMER, AND T. J. DAWSON where DBH is days before hatching. Similarly, for the 34.0-38.0 C temperature interval Qio = 1.00 + 0.032 (DBH) (r =.71;Sb = 0.005; SY.x = 0.303; n = 39). (4) From analysis of covariance, both the slopes and elevations of equations (3) and (4) differ significantly from one another (P <.05 and P <.01, respectively). Thus, the thermal sensitivity of embryonic metabolism within the Tegg range of 30-38 C is inversely related to both temperature and embryo age. DISCUSSION WATER RELATIONS Without exception, avian eggs lose water throughout their incubation. Based on a compilation of egg water losses for 81 species, Ar and Rahn (1980) found that eggs of these birds lost, on average, 15% of their initial mass despite considerable species differences in both incubation period and fresh egg mass. Our measurements on emu eggs are at variance with the concepts of Ar and Rahn (1980) in two important aspects. First, the percentage of initial egg mass lost as water by emu eggs is lower than that of all but one of the 81 species that Ar and Rahn (1980) reported, and, second, contrary to the assumed constant rate of egg water loss, the MH20 of emu eggs increased significantly during incubation. In view of these differences, an examination of the physical factors affecting egg vapor exchange is pertinent. As a diffusive process, egg water loss can be described by the relation (Ar et al. 1974) MH20 = GH20 (Pegg - Pnest), (5) where the rate of egg water loss (MH20 in mg day-') is the product of the eggshell's conductance to water vapor (GH20 in mg day-' Torr-') and the vapor pressure difference between the egg contents and the nest environment (Pgg and Pnest, both in Torr, respectively). From the above relation it is apparent that, for MH20 to remain invariant throughout incubation, GH20 and the vapor partial pressure difference (AP = Pgg - P,,t) must either remain constant or vary in compensating directions. The sub- 38.0-37.0 do00 e 0 34.0 1500 2100 0300 0900 1500 Time of Day,, h FIG. 3.-Daily variation of internal egg temperature for three oil-filled eggs placed in an emu nest on day 44 of incubation.

4.0 - THERMAL AND WATER RELATIONS OF EMU EGGS 489 3.0 + 30-34 C 0 2.0 + + 0.0 35 30 25 20 15 10 5 0 Days Before Hatching FIG. 4.-Thermal sensitivity of emu embryonic oxygen consumption as a function of embryo age for the temperature interval 30-34 C. Solid line represents the least-squares regression for individual values (crosses). stantial increase in MH20 of emu eggs during their incubation (fig. 1)1 suggests that GH20 and/or AP rises over this period. GH20 increases during early incubation in several avian species (Carey 1983), and it is possible that similar variations might 4.0 34-38 C 3.0 0 + + Q 2.0 ++ 1.0 ++ 0.0 35 30 25 20 15 10 5 0 Days Before Hatching FIG. 5.-Same as fig. 4, but for temperature interval 34-38 C

490 W. A. BUTTEMER, L. B. ASTHEIMER, AND T. J. DAWSON T 55- E 45- o 30 o 0 NY 0 M 20 -. 0 o 0 0 10 20 30 40 50 60 Days Incubated FIG. 6.-Estimated differences of external from internal egg vapor pressures (AP) and eggshell water vapor conductances (GH20) for emu eggs over the course of their incubation. The latter values were calculated from the relation GH20 = MH20/AP using field measurements of MH20 and estimates of AP(see text for details). underlie the changes in MH20 that we observed. Although we did not evaluate GH20 for the eggs undergoing incubation, substitution of our measurements of MH20, Pnest (from egg diffusion hygrometers), and Pgg (through measurement of Ten from our telemetered eggs) into a rearrangement of equation (5) permits indirect assessment of GH20 for this interval. Comparison of these estimates to those of AP throughout incubation (fig. 6) indicates that only the latter parameter varies in a manner consistent with our measurements for MH20. Because all but two of the estimates of GH20 depicted (fig. 6) lie within the 95% confidence interval for our direct measurements of GH20 presented earlier, we conclude that the increased MH20 displayed by these eggs results from variation in AP. Furthermore, as Pnest did not vary by more than 2 Torr from the mean 16.1 Torr for the two nests throughout incubation, the rise in AP (fig. 6) directly results from variation in Tegg (fig. 2). Similar proportionate increases in MH20 have been reported for Laysan and blackfooted albatrosses (Diomedea immutabilis and D. nigripes, respectively) over the course of their incubation (Grant et al. 1982). As for the emus, the variation in MH20 of the latter species' eggs reflected increased AP because of rising T,, and were not a consequence of changes in GH20 (Grant et al. 1982). There is general agreement between the indirect estimation of GH20 (fig. 6) and our laboratory values, after correction for temperature and altitude differences between sites. The field-estimated GH2O averages 9% lower than our direct measurements but is consistent With our use of telemetered egg temperatures in estimating Pe,. Since these temperatures derive from transmitters floating near the upper surface of the egg which'is in contact with the incubating bird, their values tend to be higher than the average Tegg that drives Pegg (Drent 1970). Although inference of egg temperature from estimates of Pe is unreliable (Grant 1982; Walsberg 1985), our field estimates of GH20 would be explained by average egg temperatures being 1.5 C lower than our telemetered values. Given the relatively low percentage of water lost by emu eggs (table 1) compared to values reported by Ar and Rahn (1980), GH20 and/or AP should be lower for emus than for other species. Based on our measurements of GH20 and loss of egg mass throughout incubation, AP averaged 24.2 Torr for eggs in the two nests that we studied. This value is statistically indistinguishable (t-test; P>.05) from the 27-Torr average AP based on direct measurement of Pnest and Pegg for 18 avian species (Walsberg 1980). Although the latter birds had a higher average Pegg than emus in our study (46.5 vs. 40.2 Torr, respectively), this is mostly offset by the emus' Pnest averaging 4 Torr lower than that of the other birds. The water vapor conductance of avian eggs increases directly as a function of egg mass (Drent 1970; Ar and Rahn 1980) but is inversely related to incubation length (Rahn and Ar 1974). Our measurements of GH20 averaged 37% lower than values predicted from an equation incorporating considerations for both egg mass and incubation period (Hoyt 1980). Divergence of GH20 from expected values has been interpreted as an adaptive response by birds nesting in atypical gaseous environments. For example, lower than predicted GH20 values have been found in birds breeding

THERMAL AND WATER RELATIONS OF EMU EGGS 491 under conditons of low ambient vapor pressure such as altitude (e.g, Rahn et al. 1977b; Carey 1980; Carey et al. 1983) or in arid environments (Grant 1982), whereas some birds nesting in environments with higher than normal vapor pressure have correspondingly high values of GH20 (e.g, Lomholt 1976; Ackerman and Platter- Rieger 1979; Howell 1979; Seymour and Ackerman 1980). In view of these adaptive responses, it is not surprising that the lower than predicted GH20 of emu eggs has been interpreted as an adaptation for nesting in arid environments (Vleck et al. 1980b). Nevertheless, our measurements indicate that a Pnest of 3 Torr would be necessary to produce egg water losses totaling 15% of initial egg mass for emus nesting at sea level. Such a low nest humidity, which is 85% lower than the average Pnest of desert-nesting birds (Walsberg 1980), is unlikely to attend incubating emus, especially given their widespread breeding range throughout Australia (Blakers, Davies, and Reilly 1984). Consequently, instead of representing an adaptation for arid nesting, the divergent GH20 of emu eggs may reflect a setpoint for total egg water loss similar to that associated with optimal hatchability in chickens ( 11% of fresh egg mass; Lundy 1969). Interestingly, emus breeding in Western Australia have egg water losses similar to those that we recorded (10.4% of fresh egg mass; Curry 1979). However, in view of recent evidence that avian embryos can tolerate much greater excursions in egg water loss than previously expected (Carey 1986), functional interpretation of water loss by emu eggs must await evaluation under controlled conditions. THERMAL RELATIONS Rising egg temperatures following the onset of incubation have been variously attributed to increased parental attentiveness (e.g, Drent 1970), increased heating potential of the developing brood patch (e.g., Farner 1958), and increased thermogenesis by developing embryos (e.g., Grant et al. 1982). In emus, we can eliminate embryo thermogenesis as contributing to such rises during the first 2 wk of incubation. Vleck et al. (1980b) found that emu 38-0.7 S36 0.60.0 L 34 -""0 -.5 o 0.4 3032-0.3 no op 3 0-0 g E,0.2 28-0.1 26 " 0.0 0 10 20 30 40 50 60 Days Incubated FIG. 7.-Internal temperature of oil-filled eggs as a function of elapsed incubation time. Mean daily values from eggs placed in three different clutches (open circles) are accompanied by their daily thermal ranges (vertical lines). The dashed line represents estimated net embryonic heat production as a function of time incubated, based on metabolic data of Vleck et al. ( 1980b) and measurements of egg water loss from this study (see text for details)..

492 W. A. BUTTEMER, L. B. ASTHEIMER, AND T. J. DAWSON 9 u 6 S) 5 7- ")7 (D 3. S 2 "'........... 0 10 20 30 40 50 60 Days Incubated FIG. 8.-Estimated differences in embryo age of eggs added 3, 6, and 9 days after the start of incubation from that of embryos incubated from its onset. See text for methods and underlying assumptions. embryo oxygen consumption was negligible for the first 15 days of artificial incubation. After adjusting their Vo2 measurements for the longer incubation period of our naturally incubated eggs (51 days- Vleck et al. 1980b; 54.5 days-this study) and assuming a thermal equivalent of 20.1 J/cm3 02, we can compare the time course of embryo heat production to that of our telemetered egg temperatures (fig. 7). This estimate of net heat production, which incorporates evaporative heat losses associated with measurements of MH20 over the same interval (assuming 2.42 J/mg H20), reveals a marked rise between 25 and 35 days of incubation (fig. 7). This timing is coincident with the increase in telemetered Teg and suggests that embryonic heat production in adjacent eggs is responsible for the second phase of egg temperature rise during natural incubation (fig. 7). Although we did not measure temperatures of fertile eggs, we expect that embryo thermogenesis would elevate their values above those of telemetered Tegg's from day 25 onward. The rise in Tem over the first week of incubation (fig. 7) may be due to incubating birds sitting gradually more tightly on their eggs. At this time, variations in Tem were larger and peak daily values lower (fig. 7) than seen later in incubation. Because we neither maintained surveillance of incubating emus nor monitored egg surface temperatures, we can not assess the roles of parental behavioral and/or physiological variation in the initial increase in Tegg. However, variation in daily mean Tgg over the course of incubation is not attributable to changes in adult body temperature. Daily telemetered intraperitoneal temperatures of the two males successfully hatching chicks were very stable and averaged 37.8 C (Buttemer and Dawson 1988). How does such variation in egg temperature during incubation affect the rate of embryonic development, particularly in the eggs laid after the onset of incubation? Because of the rise in egg temperature over the first 10 days of incubation (fig. 7), eggs laid after the start of incubation experience higher initial incubation temperatures. Therefore, these later additions may develop more rapidly than their nestmates, reducing differences in intraclutch hatching times. The extent to which this occurs is dependent on both the thermal sensitivity of embryonic growth and the difference in egg temperature at the time that later eggs were added. Although we did not measure directly the thermal sensitivity of growth, simultaneous measurements of growth and oxyen consumption of altricial and precocial avian embryos by Vleck et al. (1980a) demonstrated that embryonic Vo2 related directly to growth. Consequently, we assume that the inverse relation between emu embryo age and the thermal sensitivity of their aerobic metabolism (figs. 4, 5) is representative of temperature effects on their growth rate. Lacking measurements for embryos younger than 20 days, however, we assume that growth processes have a Qio of 2.5 on the first day of incubation and decline to a value of 1.0 at the time of hatching. We can evaluate our thermal model of emu growth by compar- ing the length of natural incubation to that for artificial incubation at 36.0 C (Vleck et al. 1980b). For example, on the first day of incubation there is a temperature difference of 4.0 C between the natural and artificial conditions. With a Qio of2.5, the latter embryo will grow 1.44 times more rapidly than the naturally incubated embryo. Thereafter, the difference in predicted daily growth rates declines because of the rapid rise in Teg during natural incubation (fig. 7) and decreasing metabolic

THERMAL AND WATER RELATIONS OF EMU EGGS 493 sensitivity of embryos (figs. 4, 5). Over the incubation cycle, the model predicts that artificially incubated eggs will require only 50.9 days and naturally incubated eggs 55.6 days to attain the same state of development. This predicted hatching time for the former eggs is nearly identical to the 50.7-day incubation period observed by Vleck et al. (1980b). This model can be applied to the estimated thermal differences of eggs added to a clutch 3, 6, and 9 days after incubation. The higher relative incubation temperatures of the "late" eggs reduces their extra incubation time by 0.9, 1.8, and 2.7 days, respectively (fig. 8). Although this growth advantage for late additions does not account fully for our observation that all emu nestmates hatched within 48 h of one another, such increases in growth rate would help facilitate hatching synchrony. The accelerated hatching observed in precocial birds (e.g., Vince 1966) requires that younger chicks are developed sufficiently to respond to acoustic stimuli elicited from older chicks prior to their hatching (Woolf, Bixby, and Capranica 1976). Ostrich, emu, and common rhea (Hoyt et al. 1978; Vleck et al. 1980b) have unique growth curves that also help these birds attain hatching synchrony. Embryonic growth rates for these ratites decline before reaching hatching age and this descent is ACKERMAN, R. A., and M. PLATTER-RIEGER. 1979. Water loss by pied-billed grebe (Podilymbus podiceps) eggs. Am. Zool. 19:921. AR, A., C. V. PAGANELLI, R. B. REEVES, D. G. GREENE, and H. RAHN. 1974. The avian egg: water vapor conductance, shell thickness, and functional pore area. Condor 76:153-158. AR, A., and H. RAHN. 1978. Interdependence of gas conductance, incubation length, and weight of the avian egg. Pages 227-236 in J. PIIPER, ed. Respiratory function in birds, adult and embryonic. Springer-Verlag, New York.. 1980. Water in the avian egg: overall budget of incubation. Am. Zool. 20:373-384. BLAKERS, M., S. J. J. F. DAVIES, and P. N. REILLY. 1984. The atlas of Australian birds. Royal Australasian Ornithologist Union. Melbourne University Press, Carlton, Victoria. BUTTEMER, W. A., and T. J. DAWSON. 1988. Emu winter incubation: thermal, water, and energy relations. In C. BECH and R. E. REINERTSEN, eds. Physiology of cold adaptation in birds. NATO Advanced Research Workshop. Plenum, New York (in press). more pronounced and occurs relatively earlier in their incubation period than is seen in other birds (Vleck et al. 1980b). These data indicate that embryo growth rate peaks at 75% of the way through incubation and declines thereafter until just before hatching (fig. 7). Furthermore, ostrich and emu embryos that died about 80% of the way through their typical incubation period had yolk-free body masses equal to those of siblings that hatched on schedule (Vleck 1978). Despite similarity in mass, however, these younger embryos had not fully reabsorbed their yolk sacs and probably had not attained the same extent of neuromuscular development as that of embryos incubated full term (Vleck et al. 1980b). Gains in hatching synchrony, and, therefore, reproductive output, would follow promotion of more rapid growth by embryos in later-laid eggs. Differential growth rates were considered unimportant for hatching synchrony in another ratite, Darwin's rhea (Pterocnemia pennata; Cannon, Carpenter, and Ackerman 1986). However, this conclusion was based on comparison of embryonic growth rates LITERATURE CITED from eggs held artificially at constant temperature. We emphasize that more rapid growth of late clutch additions will follow from combinations of variable, but stead- ily rising, incubation temperatures and a thermally labile embryonic growth rate. CANNON, M. E., R. E. CARPENTER, and R. A. ACK- ERMAN. 1986. Synchronous hatching and oxygen consumption of Darwin's rhea eggs (Pterocnemia pennata). Physiol. Zool. 59:95-108. CAREY, C. 1980. Adaptation of the avian egg to high altitude. Am. Zool. 20:449-459.. 1983. Structure and function of avian eggs. Pages 69-103 in R. F. JOHNSTON, ed. Current ornithology. Vol. 1. Plenum, New York.. 1986. Tolerance of variation in eggshell conductance, water loss, and water content by redwinged blackbird embryos. Physiol. Zool. 59: 109-122. CAREY, C., S. D. GARBER, E. L. THOMPSON, and F. C. JAMES. 1983. Avian reproduction over an altitudinal gradient. II. Physical characteristics and water loss of eggs. Physiol. Zool. 56:340-352. CRACROFT, J. 1974. Phylogeny and evolution of the ratite birds. Ibis 116:494-521. CURRY, P. J. 1979. The young emu and its family life in captivity. M.S. thesis. University of Melbourne, Melbourne, Victoria. DAVIES, S. J. J. F. 1976. The natural history of the emu in comparison with that of the other ratites.

494 W. A. BUTTEMER, L. B. ASTHEIMER, AND T. J. DAWSON Pp. 109-120 in Proceedings of the 16th International Ornithological Conference, Canberra, Australia, 1974. DRENT, R. 1970. Functional aspects of incubation in the herring gull. Behaviour, Suppl. 17:1-132.. 1975. Incubation. Pages 333-420 in D. S. FARNER and J. R. KING, eds. Avian biology. Vol. 5. Academic Press, New York. FARNER, D. S. 1958. Incubation and body temperatures in the yellow-eyed penguin. Auk 75:249-262. FLEAY, D. 1936. Nesting of the emu. Emu 35:202-210. GRANT, G. S. 1982. Avian incubation: egg temperature, nest humidity, and behavioral thermoregulation in a hot environment. Ornithological Monogr. 30:1-75. GRANT, G. S., T. N. PETTIT, H. RAHN, G. C. WHIT- TOW, and C. V. PAGANELLI. 1982. Water loss from Laysan and black-footed albatross eggs. Physiol. Zool. 55:405-414. HANDFORD, P., and M. A. MARES. 1985. The mating systems of ratites and tinamous: an evolutionary perspective. Biol. J. Linnean Soc. 25:77-104. HOWELL, T. R. 1979. Breeding biology of the Egyptian plover, Pluvianus aegyptus (Aves:Glareolidae). Univ. Calif. Publ. Zool. 113:1-76. HOYT, D. F. 1980. Adaptation of avian eggs to incubation period: variability around allometric regressions is corrected with time. Am. Zool. 20: 417-425. HOYT, D. F., D. VLECK, and C. M. VLECK. 1978. Metabolism of avian embryos: ontogeny and temperature effects in the ostrich. Condor 80:265-271. LANDAUER, W. 1967. The hatchability of chicken eggs as influenced by environment and heredity. Conn. Agric. Exp. Stat. Monogr. 1:1-315 (revised). LIST, R. J. 1971. Smithsonian meteorological tables. 6th rev.ed. Smithsonian Institution Press, Washington, D.C. LOMHOLT, J. P. 1976. Relationship of weight loss to ambient humidity of bird eggs during incubation. J. Comp. Physiol. 105:189-196. LUNDY, H. 1969. A review of the effects of temperature, turning, and gaseous environment on the hatchability of the hen's egg. Pages 143-176 in T. C. CARTER and B. M. FREEMAN, eds. The fer- tility and hatchability of the hen's egg. Oliver & Boyd, Edinburgh. PAGANELLI, C. V., R. A. ACKERMAN, and H. RAHN. 1978. The avian egg: in vivo conductances to oxygen, carbon dioxide, and water vapor in late development. Pages 212-218 in J. PIIPER, ed. Respiratory function in birds, adult and embryonic. Springer-Verlag, New York. RAHN, H., R. A. ACKERMAN, and C. V. PAGANELLI. 1977a. Humidity in the avian nest and egg water loss during incubation. Physiol. Zool. 50:269-283. RAHN, H., and A. AR. 1974. The avian egg: incubation time and water loss. Condor 76:147-152. RAHN, H., C. CAREY, K. BALMAS, B. BHATIA, and C. V. PAGANELLI. 1977b. Reduction of pore area of the avian eggshell as an adaptation to altitude. Proc. Natl. Acad. Sci. 74:3095-3098. SEYMOUR, R. S., and R. A. ACKERMAN. 1980. Adaptations to underground nesting in birds and reptiles. Am. Zool. 20:437-447. VINCE, M. A. 1966. Artificial acceleration of hatching in quail embryos. Anim. Behav. 14:389-394. VLECK, C. M., D. VLECK, and D. F. HOYT. 1980a. Patterns of metabolism and growth in avian embryos. Am. Zool. 20:405-416. VLECK, D. 1978. The energetics of activity and growth. Ph.D. diss. University of California, Los Angeles.. 1987. Measurement of 02 consumption, CO2 production, and water vapor production in a closed system. J. Appl. Physiol. 62:2103-2106. VLECK, D., C. M. VLECK, and D. F. HOYT. 1980b. Metabolism of avian embryos: ontogeny of oxygen consumption in the rhea and emu. Physiol. Zool. 53:125-135. WALSBERG, G. E. 1980. The gaseous microclimate of the avian nest during incubation. Am. Zool. 20: 363-372.. 1985. A test for regulation of egg dehydration by control of shell conductance in mourning doves. Physiol. Zool. 58:473-477. WINSTON, P. W., and D. H. BATES. 1960. Saturated solutions for the control of humidity in biological research. Ecology 41:232-237. WOOLF, N. K., J. L. BIXBY, and R. R. CAPRANICA. 1976. Prenatal experience and avian develop- ment: brief auditory stimulation accelerates the hatching of Japanese quail. Science 194:959-960.