THE DEVELOPMENT OF EFFECTIVE HOMEOTHERMY AND ENDOTHERMY BY NESTLING STARLINGS

Transcription

1 USDA National Wildlife Research Center From the SelectedWorks of Larry Clark 1982 THE DEVELOPMENT OF EFFECTIVE HOMEOTHERMY AND ENDOTHERMY BY NESTLING STARLINGS Larry Clark Available at:

2 Comp. Biochem. Phpiol. Vol. 73A. No. 2. pp. 253 to Printed in Great Britain. 03OO-9629/82/ $03.00/ Pergamon Press Ltd THE DEVELOPMENT OF EFFECTIVE HOMEOTHERMY AND ENDOTHERMY BY NESTLING STARLINGS L. CLARK Department of Biology. University of Pennsylvania, Philadelphia. PA 19104, U.S.A. (Received 18 January 1982) Abstract-l. Loss of body temperature and metabolism were measured in nestling European Starlings (Sturnus culyuris L.) for brood sizes of 1, 3, 5 and 7 at four air temperature ranges. Data were collected for age classes l-6, 9, 12, 15 and 18 days. 2. A distinction is made between effective homeothermy and effective endothermy as determined for broods under semi-natural conditions vs homeothermy and endothermy determined for isolated chicks. 3. Endothermy developed slowly. Shivering thermogenesis began at about 6 days; peak metabolism was not reached until about days. Analysis of variance indicated that brood size, air temperature, age, and all two-way interactions were significant. 4. Homeothermy was achieved earlier than endothermy in chicks for all brood sizes under all test conditions. For a given air temperature effective homeothermy was achieved earlier by individuals within larger broods. This was primarily due to inertial effects retarding heat loss. 5. The homeothermic status of broods of various sizes can be predicted from metabolic abilities of individuals within nests. INTRODUCTION Hatchling altricial birds are poikilothermic and ectothermic but develop into homeothermic-endotherms within the first few days of life [see McNab (1978) for definition of terms]. This transition is remarkably rapid. King & Farner (1961) postulated that the ontogeny of homeothermy was principally due to increased insulation, a more favorable heat loss to production ratio (surface area: volume), and maturatlonal effects associated with increasing age of chicks. King and Farner s seminal paper stimulated interest in the development of homeothermy and endothermy on isolated individuals (Dawson & Hudson, 1970; Ricklefs, 1974). However, serious problems arise if laboratory derived data is used in conjunction with field studies, because in the wild, chicks exist in nests with brood-mates. This may influence the insulation, heat loss to production ratio, and the capacity of individuals within broods to maintain or increase metabolism (endothermic response) in the face of a thermal gradient. A growing number of studies have already indicated that the ontogeny of homeothermy as determined in isolated individuals differs from the ontogeny of homeothermy determined on chicks within nests and siblings (Mertens, 1969; Yarbrough, 1970; O Connor, 1975; Dunn, 1976, 1979; Clark & Balda, 1981). Further, Clark & Balda (1981) indicated that large broods are able to maintain an endothermic response earlier than small broods. This ability to exhibit an endothermic and homeothermic response as a function of brood size is called effective endothermy and effective homeothermy (Clark & Balda, 1981). The rational for making the distinction between endothermy and homeothermy as derived in isolated individuals and effective endothermy and effective homeothermy is of importance when considering pat- terns of the ontogeny of thermo- and temperature regulation under natural conditions. I studied the Starling (Sturnus vulgaris) to document the relationships of metabolism and the ability of broods to maintain body temperature (Tb) as a function of age of chicks and environmental air temperature (Ta). Questions concerning the exact partitioning of the mechanisms of these differences (Clark, ms) and of correlating physiological responses of broods to adult behavior (Clark, in prep.) are treated elsewhere. Handling of youny METHODS I obtained Starling nestlings from two source colonies of 90 nest boxes each. Individually marked chicks of known age were removed from nest boxes ca. 45 min prior to dusk and transported to a field laboratory at the Stroud Water Research Center, Avondale, PA in 1980 and Waterloo Mills Research Station, Devon, PA in Transport to the laboratory took between 20 and 45 min. I kept chicks in Styrofoam hamburger containers (MacDonalds) to prevent excessive cooling during their transport to the laboratory. At the laboratory I fed chicks a mixture of canned dog food, whole wheat bread, and cottage cheese until they were replete, after which the chicks were kept under a brooding lamp until used for metabolic trials. For metabolic trials I randomly constructed broods of 1. 3, 5 and 7 from chicks of the same age. Experiments I recorded initial proventricular Tb to the nearest 0.5 C using a YSI thermistor and recorder in 1980 and copperconstantan thermocouples connected to an Omega model 141 mv recorder with cold junction compensator (+0.5 C) in Both temperature probes were calibrated in a water bath against a mercury thermometer with calibration traceable to the National Bureau of Standards. Individuals always began each metabolic trial with a Tb between 38 and 40 C. I placed chicks into natural nests 253

3 254 L. CLARK constructed of dried grass. The nest boxes used were identical to those used by starlings at the field sites. The nest boxes were 12 x 12 x 8 in. in dimension, constructed of S/8 in. exterior grade plywood. Inside edges of the nest boxes were sealed with ceramic cement. The lip of the nest boxes was covered with a layer of silicone. The lid of the nest box was fitted with pliable rubber so as to form a tight air seal when the lip was clamped to the box. Air input through the box was via a l/4 in. diameter hole centered 1 inch above the nest in the wall opposite the entry port. At the onset of metabolic trials sealed boxes with broods were placed into constant temperature environmental cabinets (Hotpoint). Air was drawn from the cabinet atmosphere through ascarite, the nest box. dierite, ascarite. drierite. a How meter (Gilbert), and sample port via a vacuum pump (Manostat). Air samples were taken from the sample port with three 5Occ plastic syringes. After sampling the syringes were sealed. Preliminary tests showed that a seal was maintained for at least 36 hr without appreciable diffusion occurring. This technique was necessary because an 02 analyser was not immediately available at the field site in Broods were allowed to acclimate to experimental conditions for 30 min. Air samples were taken at 30, 45 and 60 min after initiation of the metabolic trial. Oxygen fractions were estimated using a Beckman F-3 oxygen analyzer in 1980 and an Apphed Electrochemistry oxygen analyzer in Oxygen values recorded here represent the average of 3 values per trial. corrected to STP. Variance about this average was small. I placed each box with a brood into the environmental cabinet for 60min at one of four air temperature (Ta) ranges: 1 (6.S C). 2 (13.5-2O.O C), 3 ( O C) and 4 ( C). After each trial I removed the nest boxes from the cabinet and recorded the positions and Tb of the chicks within the nests. The Tb reported in this paper represents the mean Tb of all brood members. At this time I fed the chicks until they were replete. If a chick s Tb was below 38 C it was rewarmed under a brooding lamp until its Tb was between 38 and 4O C. Broods were then tested at the next lowest Ta range (the approximate decrease in Ta for successive trials was 7 C). A given brood was tested once at each of the four Ta ranges. There was no certainty of association among individuals during subsequent trials for older ages. No individual was used more than 3 times out of the IO age classes tested. A 4 x 4 x 6 factorial analysis of variance design was used to test for significant differences in trends in Tb loss (mean initial Tb-mean final Tb) and metabolic rate of broods for the main effects. brood size (BS), age (A), air temperature outside the nest box (Ta), and all interactions. Sums of squares and F ratios were calculated using weighted means and a fixed effects model. Measures of Tb loss and metabolism through each age were independent with respect to broods tested, though each age and stage of development is a function of the previous age. The measure of Tb loss and metabolism per brood at each level of Ta represent repeated measures. The resulting bias in the analysis is towards a fatigue effect, i.e. birds within broods do more poorly as they are tested at progressively lower Ta. This is observed for young birds, but the fact that birds improve with age, that the improvement differs with respect to brood size. and that broods are allowed to recover diminishes the significance of this source of bias, Body tmywraturrs RESULTS The loss of Tb recorded during a trial indicated that all but the BS x A x Ta interaction were highly significant in the anova model. Examination of the profiles for the main effects indicated that larger broods maintained Tb better than smaller broods across all levels of A and Ta tested (Fig. la, F 3,321 = 43.48, P < 0.001). Also, chicks were better able to maintain Tb as they grew older (across all levels of BS and Ta, Fig. lb, F5,321 = 19.26, P < 0.001). Further, chicks maintained Tb more effectively at higher Ta than at lower Ta (across all levels of BS and A, Fig. lc, F,,,,, = 82.84, P < 0.001). These results conformed to results obtained in all previous studies of homeothermy in altrical birds. The interaction terms were of interest in detecting patterns of effective homeothermy. The significant BS x A interaction (across all levels of Ta) indicated that larger broods minimized Tb loss more effectively at each age (Fig. Id, F,5,321 = 5.34, P < 0.001). Rates of improvement were most rapid in BS = 7, about the same in BS = 3 and 5, and slowest in BS = 1, where rate of improvement resembled a step function, changing at 5 days of age. The A x Ta interaction (across all levels of BS) illustrated that temperature loss was minimized at higher air temperatures, particularly for younger chicks. (Fig. le. F = 6.57, P < 0.001). The BS x Ta interaction (across all levels of A) showed that for each level of BS, broods did worse at lower Ta (Fig. lf, F,,,,, = 8.07, P < 0.001). If chicks between the ages l-6 within nests do not effectively thermoregulate (show a metabolic capacity to maintain Tb) one might expect that the Tb loss patterns are some function of brood size. The nature of this function can be determined by assuming broods realise Tb loss l/n times as effectively as a brood of one, where n is the number of chicks within a brood. This quantity is adjusted by the exponent since adding broodmates to the brood increases the surface area of the brood more slowly than the potential metabolizing mass of the brood increases. Thus, treating the brood as a unit, the theoretical Tb loss of a brood should be Tbn = (Tbl)(l/n)exp 0.667, where Tbl is the actual Tb loss for an individual within a nest, and n is the number of chicks within a brood. The theoretically derived values of Tb loss were compared to the actual data in Fig. If. Residual differences were derived by subtracting the observed values for each Ta from the expected. There were no significant differences among the derived values of mean residuals when compared to a theoretical mean of zero (Ta = 1, tj = P < 0.5; Ta = 2, tj = 1.89, P < 0.2; Ta = 3, t, = -2.54, P < 0.1; Ta = 4, t3 = , P < 0.2) This suggests that for the first 6 days of life the mechanism for maintaining Tb for all brood sizes was primarily inertial resistance to heat loss. This approach also demonstrates that useful information about Tb loss in broods can be reasonably approximated from studies based on individuals tested within nests. Mrtuholism One can ask whether the observed patterns of development of homeothermy are due to inertial effects or due to significantly increased thermogenesis during the first 6 days of life. The latter would be indicated if there was an observed increase in metabolism level with age. The anova for metabolism indicates all but

4 The development of elective homeothermy and endothermy 25s AIR TEMPERATURE l BROOD SIZE t s G OL f BROOD SIZE Fig. 1. Profile analysis (Wiener, 1971) of main and all two-way interaction effects for body temperature loss data (mean initial Tbmean final Tb). (A) Air temperature profile. (B) Brood size profile. (C) Age profile. (Df Age x brood size interaction profile, where BS = 1 (solid circles), BS = 3 (open circles), BS = 5 (solid squares) and BS = 7 (open squares). (E) Age x air temperature interaction profile, where T = 1 (solid circles), T = 2 (open circles), T = 3 (solid squares) and T = 4 (open squares). (F) Brood size x air temperature interaction profile, where T = 1 (solid circles), T = 2 (open circles), T = 3 (solid squares and T = 4 (solid squares). The dashed lines are the observed values. The solid lines represent the profile for theoretic values of Tb loss (see Table 2 and results). Each point represents the mean Tb loss for a given level of effect plotted along the abscissa. BS = brood size, A = age (in days) of chicks within a brood and T = air temperature ranges, where T = 1 ( o C), T = 2 ( C). T = 3 ( o C) and T = 4 ( O C). the age x brood size x air temperature interaction ring within this time frame. The same basic increase were statistically significant. in metabolic level as a function of age was seen for There was a strong age component to the develop- each level of Ta (Fig. 2e, F15,269 = 2.05, P < 0.01). ment of metabolic capacity among broods for all Ta The lower ranges of Ta were characterized by lower tested (Fig. 2b, F5,269 = 80.18, P < 0.001). Birds levels of metabolism during the first 4 days of life. showed higher metabolic levels as a function of This trend was reversed at the fifth day post-hatching, increasing age, suggesting some development is occur- suggesting that major developmental changes for

5 256 L. CLARK A? BROOD SIZE d D,,o--- 0 I 8,d I,.. u.,p CT,:y;g: &. I n 30r E I 1 * 1 * I, 12345G F 0 A, * Ta Fig. 2. Profile analysis (Wiener, 1971) of main and all two-way interaction effects for metabolism data. Brood size is the number of chicks within a nest. Age of the chicks is in days with zero being the day of hatching. Air temperatures are ranges of ambient conditions: 1 (6.5-13,O C). 2 ( C), 3 ( o C) and 4 ( ,o C). (A) Brood size profile. (B) Age profile. (C) Air temperature profile. (D) Age x brood size interaction profile, where BS = 1 (solid circles), BS = 3 (open squares), BS = 5 (solid squares) and BS = 7 (open circles). (E) Age x air temperature interaction profile, where T = 1 (solid circles), T = 2 (open circles), T = 3 (solid squares) and T = 4 (open squares). (F) Brood size x air temperature interaction profile, where BS = 1 (solid circles), BS = 3 (open squares), BS = 5 (solid squares) and BS = 7 (solid circles). Each point represents the mean metabolism for a given level of effect plotted on the abscissa. thermogenesis did not occur before this time. Chicks were first observed to visibly shiver at 5 days of age. It might be possible that as chicks increase in age and weight the ratio of heat production to loss becomes greater. Thus, larger broods losing less Tb (Fig. 1) would be better able to sustain metabolism and be more resistant to hypothermia. Despite a significant brood size x age interaction (F = 5.77, P < O.OOl), no discernable pattern emerges that is consistent with this interpretation (Fig. 2d). Also, the main effect, brood size (F3.269 = 52.82, P < O.OOl), showed broods of 1 and 7 possessing higher metabolisms than broods of 3 and 5 across all Ta and ages (Fig. 2a). Further, no clear ranking of metabolism for brood size was observed within each level of Ta (Fig. 2f> F9.269 = 3.55, P < 0.001). These results are probably attributed to a signi~cant but irratic Ta response of chicks during the first 6 days (Fig. 2c, F 3,2h9 = 4.59, P < 0.005). Differences should be more pronounced in the directions indicated if active thermogenesis were a strong component for maintaining homeothermy. &f&t of weight Figure 3 is a plot of Tb loss x weight for ages l-6 and 9, 12, 15 and 18 days of age. Least squares regressions were fit to the data in Fig. 3 (Table 1). Two regression lines were simultaneously fit, one for the range of weights where Tb loss was decreasing and one where Tb loss was relatively constant, with a

6 The developme~l of effective homeothermy and endothermy 257 T= 1 T=2 T=3 JO T-4 JO. 0 Fig. 3. Loss of body tem~rature by broods as a function of brood weight for brood sizes of 1, 3, 5 and 7 (columns) and air temperature ranges 1 ( O Q 2 (13.5-2O.O C). 3 (Z&5-27S C) and 4 (27.s34.O C) (rows). Regressions were fitted by least squares analysis (Table 1). slope close to zero. The intersection of the two regres- definition of effective homeothermy from here. Within sions estimated the lowest brood weight at which a brood size the weight at which effective homeocomplete homeothermy could be maintained by all thermy was attained was Iower for higher Ta. Also, members of a brood. This is used as the operational larger broods realized an effective homeothe~y at a Table 1. Least squares regressions for Tb x wt for each brood size (BS) and air temperature (T,) class Cooling range Homeothermy range BS T, Slope SE Int SE Slope SE fnt SE s o.wo I o.oocl Regressions were derived by minimizing mean square error terms for a pair of regression lines. See Fig. 3 and results for a more complete description of curve fitting techniques.

7 L. CLARK Fig. 4. Metabolism of broods as a function of body weight for brood sizes of 1, 3, 5 and 7 (columns) and air temperature ranges 1 (6.SP13.0 C), 2 ( cC), 3 ( O C) and 4 ( O C) (rows). Regressions were fitted by least squares analysis (Table 2). given Ta range at earlier stages of achieved growth age. Least squares regressions were fit to the data in (i.e. the weight per individual at achieved homeo- the manner described above (Table 2). The intersecthermy is smaller in larger brood sizes). tion of the two regressions defines the stage of effec- Figure 4 is a plot of metabolism as a function tive endothermy. As broods increased in mass the brood weight for ages l-6, 9, 12, 15 and 18 days of metabolic level increased and eventually reached an Table 2. Least squares regressions for metabolism x wt for each brood size (BS) and air temperature (Ta) class BS T, Development range Asymptotic range Slope SE Int SE Slope SE Int SE I I O.ooO Regressions were derived by minimizing mean square error terms for a pair of regression lines. See Fig. 4 and results for a more complete description of curve fitting techniques.

8 The development of effective homeothermy and endothermy 259 Table 3. Weights (g) of broods and individuals within broods at effective homeothermy and peak or asymptotic metabolism for a given Ta range BS Ta l-b Metabolism Brood Individual Brood Individual weight weight weight weight asymptote for a given Ta range for broods of 3 and 5. Broods of 7 reached a metabolic peak at some intermediate weight then decreased their metabolism. Broods of 1 showed a constant increase in metabolism up to maximum achieved weights. Mean weights of individuals within broods, where metabolism peaked for a given Ta, were g (5-6 days) for all brood sizes greater than 1 (Table 3). The trend was for individuals within larger broods to show this metabolic peak at lighter weights (30 g). Comparison of brood weight at metabolic asymptote and weight of the brood at effective homeothermy suggested that effective homeothermy and effective endothermy were closely correlated at low Ta ranges, and that homeothermy preceded endothermy at high Ta ranges (Table 3). DISCUSSION Development of effective homeothermy is a function of the changes of the physical environment surrounding the nestlings. As age and weight of the brood increase there is a smaller decrease in Tb loss for a given BS and Ta. Tb loss is a function of BS and Ta. Thus, the data on Tb loss can be summarized by saying that functional homeothermy develops earlier in larger broods for a given Ta. Similar findings were reported for House Wrens, Troglodytes aedon (Dunn, 1976), Tree Swallows, Iridoprocne hicolor (Dunn, 1979), and Pinon Jays, Gymnorhinus cyanocephalus (Clark & Balda, 1981). Pinon Jays showed an earlier developed capacity to maintain or increase metabolism for a given Ta. The more numerous and detailed data on metabolism presented here, if examined by age, also show these trends. In contrast to metabolism per se (which increases with age and weight), endothermic responses do not occur before age 5 and 6 days (Fig. 2e). Isolated individuals begin to show endothermic responses at about the same age, and are fully homeothermic and endothermic at days of age (Ricklefs, unpub.). At lower Ta, individuals in small broods are more depen- T.H.P. 73.2~ H dent on thermogenesis to maintain effective homeothermy than are individuals in larger broods. Also, differences in timing of effective homeothermy among broods with respect to age and weight suggests that individuals within smaller broods depend on the development of thermogenic capacities to maintain Tb to a greater degree than individuals within larger broods. As chicks mature increased mass allows for a more favorable heat production to loss ratio. By being larger a chick or a brood fills the nest cup more fully, thereby decreasing the amount of exposed surface area which acts as an avenue of heat loss. With age the chicks mature neurologically and increase the density and contractile capacity of muscle tissue. This corresponds to more efficient and higher quantities of thermogenic tissue per unit mass. Feather development during this stage of development (age l-6) is inconsequential. The time parent birds spend brooding their young decreases as the chicks age (e.g. Davis, 1960; Morton & Carey, 1971). Workers have generally assumed that the decrease in time spent brooding young was due to a parental response to the increased homeothermic capacity of individuals within the brood. What has not been shown was whether the observed adult attendance patterns were a facultative response to the increased homeothermic capacity of older chicks, whether adults are responding to the increased energy demands of the maturing chicks, whether it is some combination of the two, or whether adults brooding schedules were behaviorally fixed for a given stage of development of the young. Johnson & Cowan (1974) experimentally showed that adult Starlings and Mynas (S. cristaltellus) were inflexible in their incubation patterns. This pattern does not hold true for brooding behavior in Starlings (Clark, in prep.), in which brooding decreases directly in relation to the effective homeothermic capacity of chicks within broods. As shown here, the ability to maintain Tb is a dynamic process, dependent not only on the maturational status of the young, but at least

9 260 L. CI.ARK initially, also on brood size. Thus data on homeothermy based on isolated individuals could prove misleading for assessments about adult time and energy budgets. Nevertheless, I have shown here that the large body of literature on individual homeothermy still might be of value for ecologists. Estimates of effective homeothermy for broods based on individuals compared favorably to actual data obtained on effective homeothermy. The question of whether chicks within different brood sizes benefit from a facultative response by adults is addressed elsewhere (Clark, in prep.). A~knowledgrmmts~Parts of this study were funded by the Frank M. Chapman Memorial Fund, The Harris Foundation, the Department of Biology of the University of Pennsylvania. and a National Science Foundation grant to R. E. Ricklefs. My thanks go to all of these agencies. R. Vannote and The Academy of Natural Sciences of Philadelphia kindly allowed me use of the Stroud Water Research Station in Avondale. PA. D. Goldstein, T. Webb and N. Lung provided valuable assistance in the field, thus freeing me to do the metabolic studies. I also wish to thank A. E. Dunham. F. B. Gill and J. B. Williams for criticallv reading an earlier draft of this paper. I wish to thank R. E. Ricklefs for allowing me access to unpublished data on metabolism of isolated nestlings. This study was part of a doctoral dissertation from the Department of Biology, University of Pennsylvania. REFERENCES CLARK L. & BALDA R. P. (1981) The development of effective endothermy and homeothermy by nestling Pinon Jays. Auk 98, DAVIS J. (1960) Nesting behavior of the Rufus-sided Towhee in coastal California. Condor 62, DUNN E. H. (1976) The relationship between brood size and age of effective homeothermy in nestling House Wrens. Wilson Bull. 88, DUNN E. H. (1979) Age of effective homeothermy in nestling Tree Swallows according to brood size. Wilson Bull. 91, DAWSON W. R. & HUDSON J. W. (1970) Birds. Comparative Physiology of Thermoregulation, Vol. 1 (Edited by WHIT- TOW G. C.), pp Academic Press, New York. JOHNSON S. R. & COWAN I. M. (1974) Thermal adaptation as a factor affecting colonizing success of introduced Sturnidae (Aves) in North America. Can. J. Zool. 52, KING J. R. & FARNER D. S. (1961) Energy metabolism, thermoregulation, and body temperature. In Biology und Compar&e Physioloyy qf bird.s.~vol. 2 (Edited by MAR- SHALL A. J.), pp Academic Press, New York. MCNAB B. K. (1978) The evolution of endothermy in the phylogeny of mammals. Am. Nut. 112, l-21. MERTENS J. A. L. (1969) The influence of brood size on the energy metabolism and water loss of nestling Great Tits, Parus mcrjor major. Ibis 111, I1 16. MORTON M. L. & CAREY C. (1971) Growth and the development of endothermy in the Mountain White-crowned Sparrow (Zonotrichia lrucophyrs oriunthe). Phq siol. Zool. 44, O CONNOR R. J. (1975) The influence of brood size upon metabolic rate and body temperature in nestling Blue Tits (Parus caeruleus) and House Sparrows (Pusser domesticus). J. Zool., Lond. 175, RICKLEFS R. E. (1974) Energetics of reproduction in birds. In Acian Energeticc Vol. I5 (Edited by PAYNTER R. A.), pp WIENER R. J. (1971) Stutisticul Principles in Experimentul Design. McGraw-Hill, New York. YARBROUGH C. G. (1970) The development of endothermy in nestling Gray-crowned Rosy Finches, Leucosticte tephrocoiis yriseonucha. Comp. Biochem. Physiol. 34,