A Model for Evaluating Time Constraints on Short-term Reproductive Success in Altricial Birds 1

AMER. ZOOL., 28:853-862 (1988) A Model for Evaluating Time Constraints on Short-term Reproductive Success in Altricial Birds 1 LARRY CLARK Monell Chemical Senses Center, 35 Market Street, Philadelphia, Pennsylvania 1914 AND ROBERT E. RICKLEFS Department of Biology, University of Pennsylvania, Philadelphia, Pennsylvania 1914 SYNOPSIS. We develop a simple single dimension model incorporating the time and energy commitments of breeding altricial birds in an effort to understand the evolutionary constraints on parental care. We chose time as the dimension of preference, because it is a naturally bounded constraint, e.g., length of day or breeding season. The utility of the model was evaluated by comparing simulations of time allocation of various breeding scenarios for the European starling (Sturnus vulgaris) to field data. The structure of the model may prove useful in determining the evolutionary constraints on parental care imposed by the developmental pattern of chicks. INTRODUCTION Animal energetics is an attractive, quantifiable subdiscipline which addresses the proposition that individuals must continuously acquire chemical energy if they are to maintain physiological homeostasis. Further justification for energetic studies stems from the idea that allocation of finitely available resources to maintenance and growth influences survival and reproduction, hence evolutionary fitness (Fisher, 1958; Gadgil and Bossert, 197). Investigations of limits to reproduction via energy acquisition and expenditure have produced a formidable body of empirical studies on avian energetics (Dawson and Hudson, 197; Farner, 1973; Paynter, 1974; Wolf and Hainsworth, 1978; Walsberg, 1983). However, the implicit assumption that energy is the only critically limiting resource has weakened the energetic approach. Empirical studies of energy budgets have not produced a synthesis of the factors constraining the evolution of lifehistory traits. Energetics is one aspect of a life history; multiple limitations must be incorporated into a single model to adequately assess constraints on reproduction. If evolution tends to maximize lifetime fecundity then investigation of tradeoffs among (1) the allocation of time to particular activities, (2) the energetic costs of those activities, and (3) the risk associated with the timeenergy commitment employed should lead to a clearer understanding of limits placed upon the evolution of specific life-history traits. Both time and energy devoted to activities have upper, definable limits of allocation, as well as less definable interactions that affect fecundity and survivorship. Here we use an integrative timeenergy budget approach to study constraints on reproductive biology (sensu Pennycuick and Bartholomew, 1973). We develop a simple model for the allocation of time to parental care for passerine birds. Rate of energy demand and acquisition are expressed as time equivalents. We then examine the consequences of changing the allocation of time and energy resources to different compartments of the model for short-term reproductive success. We also examine mortality patterns of passerines as a means of evaluating the utility of the model. Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218 1 From the Symposium on Energetics and Animal Behavior presented at the Annual Meeting of the American Society of Zoologists, 27-3 December 1986, at Nashville, Tennessee. THE TIME MODEL We express our model in units of time, because total time is fixed and must be allo- 853

854 L. CLARK AND R. E. RICKLEFS cated among discrete, mutually exclusive activities. For altricial passerines, these activities generalize to the following categories: maintenance behavior (preening, drinking), territorial defense, courtship, maintenance of pair bonds, other social interactions, transit, foraging, night or daylight inactivity, and incubation, brooding or shading of young. Energetic costs necessary to support these activities, in addition to the energy invested in eggs and offspring, can be expressed in units of time. Thus, a Foraging Time Equivalent (FTE) is the time a parent must forage to support an energetic expenditure, such that FTE = Ej/F [hr] (1) where E, is the energetic cost of activity i [ J] and F is the net rate at which energy is acquired during foraging [J/hr], after the energy requirement of foraging itself is satisfied. The energetic cost of an activity is calculated as [J] (2) where D ; is the duration of activity i [hr], M : is the rate of energy consumption for activity i [J/hr], and e f is the efficiency of energy assimilation. Assuming a parent or the young being raised do not use stored energy reserves or catabolize other body tissue, the daily allocation of time to all energy-consuming activities plus the foraging time equivalent required to support these activities must sum to 24 hr or less, i.e., FTE: < 24 hr. (3a) Because each activity is a mutually exclusive event, an increase in one component must be balanced by a decrease in the other components. Furthermore, consequences of any tradeoffs made under time and energy constraints can now be evaluated with respect to a utility function that relates time allocation to survivorship probabilities and fecundity. Time, rather than energy, should be the dimension of preference in the constraint function. In some cases, the energetic costs of an activity may be secondary to the importance of that behavior to the probability of short-term reproductive success or survival (e.g., vigilance, thermoregulatory or brooding behaviors; Herbers, 1981; Clark, 1987). The paradigm first proposed by Lack (1954), that parents will raise as many offspring as they can feed, may only be partially correct. Rather, parents may raise as many offspring for which they have time. For a parent raising a brood, eq. 3a takes the form 24-N N + 2 (S + B + FL + M a + M af + M ca + M cg + M c ) < 24 hr (3b) where N is the duration of nocturnal inactivity and is set by latitude and the time of year at which breeding occurs. During daylight hours, time is allocated to social behavior (S), brooding of the young (B), and flight to and from the foraging area (FL). In addition, the time budget is defined by the foraging time required to balance the energetic cost of a parent's maintenance and thermoregulatory costs, exclusive of the energetic cost of flight (M a ), the energetic cost of flight (M af ), the energetic costs of activity for chicks (M ac ), the energetic cost of growth (M cg ), and the energetic cost of maintenance and thermoregulation of the brood (M c ). TIME ALLOCATION OF PARENTAL CARE: A CASE STUDY We evaluate the components of eq. 3b using the European starling (Sturnus vulgaris) as an example. Because temperature, which affects B, M c and M a, changes throughout the day, we base our calculations on 1-hr intervals to which an average temperature is assigned, and run these over a 24-hr period. Brooding behavior Altricial young hatch as ectothermicpoikilotherms and rapidly develop into endothermic-homeotherms (Dawson and Evans, 1957). Until this transition is complete, young chicks must be brooded by the parents if a high, constant T b is to be maintained. Failure of parents to do so may retard development of the young or Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

TIME CONSTRAINTS ON REPRODUCTION 855 increase their risk of death (King and Farner, 1961; Yarbrough, 197). The ability of chicks to maintain a high T b against a thermal gradient is called homeothermic capacity, HC (Clark, 1984), which is determined by (T bf - TJ/^bo - T a ), where T bf is the mean body temperature of chicks after exposure to air temperature T a for one hour, and T^ is the mean initial body temperature of chicks. The mean homeothermic capacity of a brood is a function of the thermogenic capacity of chicks, the insulating properties of the plumage, nest environment and broodmates, and the operative temperature of the environment which drives the potential for heat flow (Mertens, 1977a, b). This complex relationship can be reduced by empirically quantifying the relationship between HC and age, brood size and ambient air temperature (Clark, 1982). Under ordinary environmental conditions, adults brood chicks to maintain their T b 's at least at 9 percent of full homeothermic capacity. Clark (1984) experimentally determined that adults altered brooding time as a function of HC rather than as a consequence of the energetic requirements associated with changes in brood size per se. Thus, on an hourly basis, the time chicks are brooded is B = b - ^HC (4) where b and b, are fitted constants. FTE and time for flight As central-place foragers, parents must travel to and from the foraging site to gather food for their young. This time in transit is FL = (FD/FS)2NF [hr] (5) where FD is the distance from the nest to the foraging site [m], FS is the speed at which a parent flies [m/hr], and NF is the number of flights a parent makes to a foraging area. Specific values of flight speeds have been empirically determined for starlings (Torre-Bueno and La Rochelle, 1978). The number of trips to a foraging area reflects the energy requirement of the brood and the profitability of the foraging environment (Tinbergen, 1976; Wester- terp et al, 1982). Nonetheless, the number of flights usually monotonically increase to an asymptote. This relationship is approximated by a linearly increasing function NF = (n + n,t)bs (6) where n and n, are fitted constants for a parent feeding a single chick, and a constant value representing the asymptotic number of flights required for a single chick. The number of flights at the asymptotic value for larger brood sizes is simply multiplied by the brood size. FTE for rearing chicks The metabolic energy chicks devote to maintenance and thermoregulation varies with environmental conditions, including the extent to which the chicks are brooded, development status of the chicks and the brood size. The time required to support the energetic cost of metabolism can be described approximately by a logistic equation M c = [A mc /(1 + e- k < T " T "]/(e f F) [hr] (7) where A mc is the asymptotic energy of metabolism for a brood [J], k is a rate constant, T is the age of the brood (days posthatching), and Tj is the age of the brood at inflection of the logistic curve. The parameter values A mc, k and T; are empirically determined and vary as a function of T a and brood size (Clark, 1984). The value of M c for the simulation was summed over 24 hr for environmental conditions typically prevailing during the first two weeks of May, and adjusted for the percentage of time the chicks need to be brooded so as to maintain full homeothermic capacity (Clark, 1983). The time required to support the energetic costs of normal daily growth for chicks is M cg = [(W T D T ) - (W-^.D^BS -(e f F) [hr] (8) where W T is the mass of the brood at age T and D T is the energy density of individuals at age T [J/g]. The mass of chicks is estimated from the logistic growth curve of the form Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

856 L. CLARK AND R. E. RICKLEFS W = A w /(1 + [g] (9) where A w and g are the asymptotic mass of nestlings and the growth constant, respectively (Clark, 1983). The density of tissue is estimated from bomb calorimetry studies of nestlings D = d o + d,t [J/g] (1) where d and d, are empirically fitted constants (Westerterp, 1973). The time required to support the energetic cost of activity for chicks is M ac = [BS(a + ait)]/(e f F) [hr] (11) where a and a, are empirically derived constants (Westerterp, 1973). FTE of adult activity and maintenance The time required to support the nonflight related activity, thermoregulation and maintenance metabolism of adults is derived from standard metabolic curves (Hart, 1964; Dmi'el and Tel-Tzur, 1985). When T a exceeds the lower critical temperature M a = [(m + m,t a )/(e f F)]2.5 [hr] (12) where m and m, are fitted constants. The multiplier 2.5 is based on field estimates of activity costs relative to the standard metabolic rate (Mugaas and King, 1981; Walsberg, 1983). Breeding behavior In southeastern Pennsylvania the most productive breeding period of starlings occurs during the first breeding pulse in mid April. Chicks hatch by the first week of May, when daytime temperatures are still generally cool. Subsequent renesting attempts following failure of first breeding efforts or successful reproduction occur in late May, with the chicks hatching by mid June. Temperatures during this period can become quite warm (Clark, 1987). As in other passerines, the reproductive output of starlings decreases as a function of season. The modal clutch and brood size occurring in April and May is five, while the modal clutch size and brood size is four in late May and June. Although males initiatenest-buildingand complete most of the nest, females bear most of the burden of incubating, brooding and feeding. Assistance by males is highly variable, ranging from zero to about one half the effort needed to rear young (Kessel, 1957). Increased parental care by males seems to be limited to periods of high energy demand by the chicks, or to periods of inclement weather (personal observation). SIMULATIONS Simulations of time commitments required for each category of parental care were carried out for one and two parents raising broods of 3, 5, and 7 (Table 1). We assumed that during the course of development neither parents nor chicks used stored energy reserves. This sometimes resulted in a violation of eq. 3b, and was expressed as negative time available during a 24-hr period. Such time deficits may be made up by reallocating time devoted to the models' component activities, e.g., brooding or flight time, by decreasing the distance traveled to a foraging site, or by assuming a negative energy balance. Broods of three The energetic cost of raising a brood of three is small when compared to the investment in raising a brood of five. Indeed, two parents would seem to have no problem in raising a brood of three under any reasonable condition of foraging return, temperature, or distance traveled to a foraging site (Fig. 1). However, unless foraging returns were uncharacteristically high, temperatures unseasonally warm, or the average distances travelled to the foraging sites short, a single parent would not have sufficient time to successfully rear three young (Fig. 1). Regardless of the level of foraging return or distance traveled to the foraging site, the most severe limit to a single parent's allocation of time is brooding. This is true even though the energetic cost of rearing a brood of three for the first few days posthatching is small. For example, by the third day post-hatching and under standard foraging distances and energy return (5 m and 1 J/hr), up to 6 of the 14 available Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

TIME CONSTRAINTS ON REPRODUCTION 857 TABLE 1. An example of the breakdown of the time allocation requirements (hr) necessary for a single adult to raise a healthy brood of five nestlings, assuming a foraging return of 1 kj, an assimilation efficiency of.7, and an average distance to the foraging site of 5 m (Fig. 2, bottom panel).* Age 1 2 3 4 5 6 7 8 9 1 11 12 13 14 15 16 17 18 Chick metab.14.36.82 1.44 2.34 3.47 4.63 5.6 6.29 6.72 6.96 7.1 7.18 7.22 7.24 7.25 7.25 7.25 Forag»ing time equivalents Chick growth Chick activ Adult flight.93 1.28 1.7 2.14 2.55 2.85 2.99 2.96 2.78 2.52 2.24 1.97 1.74 1.55 1.4 1.3 1.22 1.16.8.8.8.8.8.17.39.61.83 1.6 1.28 1.5 1.72 1.94 2.16 2.24 2.24 2.24.25.36.5.65.83 1.3 1.25 Adult metab 5.5 5.5 5.5 Brood 1.4 5.11.12 Flight 2.75 3.31 3.88 4.44 5.1 5.57 6.14 Sum 19.69 16. 12.62 14.25 16.31 18.59 2.9 21.82 22.55 22.95 23.13 23.22 23.29 23.36 23.45 23.44 23.36 23.36 Difference -5.69-2. 1.4 -.25-2.31-4.59-6.9-7.82-8.55-8.95-9.13-9.22-9.29-9.36-9.45-9.44-9.36-9.36 * Age is in days post-hatching. Sum is the sum of all diurnal activity requirements. Difference is the quantity Sum subtracted from the available daylight prevailing in early May (14 hr). daylight hours in May must be spent brooding young. Lowering the mean daily temperature by 8 degrees increases that commitment to 1 hr. It is the combination of brooding requirement with the FTE required to support a parent's energy requirement that results in a conflict over the allocation of time. A solution would be to either brood the young and thereby restrict foraging time, or reduce the brooding commitment in an effort to support FTEs. The former would result in reduced growth for the young or weight loss to the adult. The latter would increase the risk of cold exposure and mortality for the young. Should the young survive to thermal independence, the allocation of a parent's time balances foraging for the chicks and for itself. Parents foraging at a distance of 5 m, and at constant temperatures and foraging return (May, 1 J/hr), require 1 hr of FTE to support metabolic costs and flight time; whereas, the FTE necessary to support the chicks is 6 hr. When the foraging distance is doubled, that ratio increases to 14:6 hr. Increasing the rate of foraging return may provide a solution, but food availability must impose an upper limit to F. In this simulation, at foraging distances of 1 km, adults must gather energy near known maximally achieved values (Tinbergen and Drent, 198). A single parent could successfully raise a brood of three if it could forage closer without a loss of efficiency in foraging. Broods of five Although a brood of five conserves heat better than a brood of three, adults still must brood the young for the first few days post-hatching. Furthermore, the additional energetic cost of raising two extra chicks does not generally seem to create conflicts for a pair's time allocations. Exceptions occur under the most severe conditions of temperature, foraging return and travel distance, where the brooding requirement is in conflict with other behaviors (Fig. 2). Single parents are invariably unable to allocate sufficient time to raise a brood. This failure is initially due to a high brooding commitment for the first two days posthatching (normally 1 and 5 hr, respectively). However, even if a parent could Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

-5 ONE PARENT TWO PARENTS ONE PARENT TWO PARENTS ONE PARENT TWO PARENTS 2 4 6 8 1 2 4 6 8 1 DAYS POST-HATCHING m < o -5 2 4 6 8 1 2 4 6 8 1 DAYS POST-HATCHING -5 ID a O -5 FORAGING 2 \ 1 "N. \ 5 \ A... TEMPERATURE w >^ 'A.. DISTANCE DISTANCE.28 ^\ FORAGING 2o"^ ^"v TEMPERATURE.25 ^^^^ 2 4 6 8 1 2 4 6 8 1 DAYS POST-HATCHING Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

TIME CONSTRAINTS ON REPRODUCTION 859 avoid the time constraint that brooding imposes, it would not be able to avoid conflict for the allocation of time to FTE required to support the chicks, itself, or flight requirements. Of the three, the time spent in flight (6 hr) normally is greatest, with 5.3 hr required for chick FTE and 3.3 hr FTE for adult metabolism at the chicks' asymptotic growth. Broods of seven Brooding behavior is of little concern for large brood sizes. The young are thermally independent under most circumstances by the second day post-hatching. The primary constraint on the parents' ability to rear young is the FTE of foraging and flight time. As can be seen for the simulation in Figure 3, single parents are incapable of rearing seven young, largely because the FTE necessary to support the chicks becomes unmanageable as the chicks grow older and energetic costs increase. For the same reasons, a pair of adults are limited to successful reproduction only under circumstances where time spent in flight is short or adults forage at maximally efficient rates. DISCUSSION Mortality patterns of chicks are indicative of potential selective pressures, which, in turn, can help us understand the factors important to the evolution of parental care. The importance of any specific mortality factor varies widely among studies (Ricklefs, 1969). Starvation resulting from inability of parents to adequately feed large chicks has received much attention (Crossner, 1977; O'Connor, 1979). But the age distribution of chick mortality suggests that factors influencing time allocation of parental care to small chicks may be very important. For example, in many studies over 75 percent of the total mortality of passerine chicks occurs before one-half of the nestling period is completed (Fig. 4; Hudec and Folk, 1961; Bogucki, 1972; Askenmo, 1977; Korpimaki, 1978; Ricklefs and Peters, 1979). Some of this mortality may be attributable to the parents' inability to support the simultaneous energy requirements of growth and thermoregulation of chicks. However, fifty percent of the total mortality occurs prior to chicks developing full homeothermic capacity. This period of development ends before chicks attain their peak energy requirements (Ricklefs, 1974). Mortality patterns of starling chicks are in some degree related to the ontogeny of homeothermic capacity (Fig. 5). Deaths attributable to cold exposure occur significantly earlier in nestling development than deaths attributable to predation or associated with emaciated young (Kolomogorov-Smirnov test, P <.1). To estimate the vulnerability of chicks to exposure we estimated the level of risk associated with prevailing environmental conditions FIG. 1. Simulation of time allocation model for parent(s) raising a brood of three. The ordinate represents the hypothetical daylight hours available to parents for social behaviors. The model assumes that parents are rearing a brood that is synchronously hatched, and that neither the chicks nor the parent(s) experience any weight loss. The top panels represent conditions of varying foraging return (J/hr) for constant distance traveled (5 m) to the foraging site and for average diurnal temperature conditions prevailing during the first two weeks of May. The middle panels represent conditions of varying diurnal temperature (N = first two weeks of May, W = N + 8, C = N 8 ) for constant distance traveled (5 m) to the foraging site and constant foraging return rate (1 J/hr). The lower panels represent varying distances parent(s) travel to the foraging site (m) for average diurnal temperatures during May and fixed level of foraging return (1 J/hr). FIG. 2. Simulation of time allocation model for parent(s) raising a brood of five. The panels (top, middle, and bottom) represent varying conditions of foraging return, mean diurnal temperature and distance traveled to the foraging site. See Figure 1 legend for further description of the simulation conditions. FIG. 3. Simulation of time allocation model for parent(s) raising a brood of seven. The panels (top, middle, and bottom) represent varying conditions of foraging return, mean diurnal temperature and distance traveled to the foraging site. See Figure 1 legend for more detailed description of the simulation conditions. Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

86 L. CLARK AND R. E. RICKLEFS 1 8 ^ ^ u 6 4 2 h 1&i 1 5 7 9 AGE AT DEATH tlji - 11 (DAYS)... 1-197-73 1=1 1979-82 n = 129 n=24 1 n. 13 15 17 FIG. 4. Age distribution of mortality for nestling starlings at the Stroud Water Research Center colony during two studies (Ricklefs and Peters, 1979; Clark, 1983). The line represents cumulative percent mortality for each study. Mortality distributions did not vary between studies (P >.1). and developmental status of the chicks. Risk was denned as the fraction of broods dying at specified sets of air temperatures, ages, and brood sizes. Chicks older than six days post-hatching did not succumb to hypothermia (F,, 7 = 13.92, P <.1), whereas younger chicks were highly susceptible (Fig. 5). Also, chicks exposed to temperatures below 13 C experienced higher risk than those exposed to warmer air temperatures (F li7 = 13.44, P <.1). Smaller brood sizes (BS = 1, 2, 3) were at higher risk than larger brood sizes (BS = 4, 5 and BS = 6, 7)(F Si7 = 6.12, P <.5). Earlier, Clark (1984) showed that parents brooded chicks as a function of the mean homeothermic capacity of the brood, irrespective of the energy demands imposed by brood size. In that study, parents maintained broods at about 9 percent of full homeothermic capacity. Yet the mortality data would suggest that exposure is a significant cause of mortality. This apparent contradiction is explained as follows. The exposure related mortality occurred during long intervals of inclement weather. Adults initially brooded young in the manner outlined by Clark (1984) an example of time allocation to a behavior regardless of the energetic costs. However, prolonged duration of poor foraging conditions and inclement weather eventually resulted in the apparent shift in commitment from brooding to foraging behavior. Reproductive success was apparently a tradeoff between the risk of prolonged brooding absences for the chicks and the risk to the adult of experiencing severe energy deprivation (Clark and Gabaldon, 1979). It is clear from this example that more research is needed to quantitatively assess the effects of energy deprivation on behavioral choices and surviability. Time allocation models are useful in identifying periods of conflicting allocation of time for parental behavior. The mortality data on starlings is consistent with the simulation predictions that time allocation is constrained early during nestling development, primarily by the brooding requirements of small chicks, and later by food requirements of larger chicks. Our model adequately describes constraints on the parents' time, and it is consistent with field data for reproductive success. The model may also be useful in evaluating hypothetical limits on reproduction resulting from intrinsic constraints of an adult's flight physiology and the ontogeny of thermoregulation in chicks. Modeling approaches can have distinct advantages in providing investigators with focused questions which pinpoint areas Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

TIME CONSTRAINTS ON REPRODUCTION 861 1 8 6 u 3 EXPOSURE STARVATION PREDATION OTHER 1 2 3 4 5 6 7 8 9 1-12 13-15 19-2 FIG. 5. Age distribution of mortality for nestling starlings at the Stroud Water Research Center during the period, 1979-82. Mortality was partitioned into categories of apparent cause (Clark, 1983). where empirically derived data are needed to test hypotheses on the evolutionary constraints on the allocation of parental care and reproductive effort. REFERENCES Askenmo, C. 1977. Effects of addition and removal of nestlings on nestling weight, nestling survival, and female weight loss in the Pied Flycatcher, Fidedula hypoeua (Pallus). Ornis. Scand. 8:1-8. Bogucki, Z. 1972. Studies on the activity of Starlings (Sturnus vulgaris) in the breeding season. Acta Zool. Cracoviensia 17:97-121. Clark, L. 1982. The development of effective homeothermy and endothermy by nestling Starlings. Comp. Biochem. Physiol. 73A:253-26. Clark, L. 1983. Constraints on parental care and reproduction in the European Starling, Sturnus vulgaris. Ph.D. Diss., University of Pennsylvania. Clark, L. 1984. Consequences of homeothermic capacity of nestlings on parental care in the european starling. Oecologia (Berlin) 65:387-393. Clark, L. 1987. Thermal constraints on foraging in adult european starlings. Oecologia (Berlin) 71: 233-238. Clark, L. and D.J. Gabaldon. 1979. Nest desertion by the Pinon Jay. Auk 96:796-798. Crossner, K. A. 1977. Natural selection and clutch size in the European starling. Ecology 58:885-892. Dawson, W. R. and F. C. Evans. 1957. Relation of growth and development to temperature regulation in nestling field and chipping sparrows. Physiol. Zool. 3:313-327. Dawson, W. R. and J. W. Hudson. 197. Birds. In G. W. Whittow (ed.), Comparative physiology of thermoregulalion, Vol. l,pp. 223 31. Academic Press, New York. Dmi'el, R. and D. Tel-Tzur. 1985. Heat balance of two starling species (Sturnus vulgaris and Onychognathus tristram) from temperate and desert habitats. J. Comp. Physiol. 155B: 195-22. Farner, D. S. 1973. Breeding biology of birds. National Academy of Sciences, Washington, D.C. Fisher, R. A. 1958. The genetical theory of natural selection. Dover, New York. Gadgil, M. and W. H. Bossert. 197. Life historical consequences of natural selection. Amer. Naturalist 14:1-24. Hart,J. S. 1964. Seasonal acclimation of four species of small birds. Physiol. Zool. 35:224-236. Herbers.J. M. 1981. Time resources and laziness in animals. Oecologia (Berlin) 49:252-262. Hudec, K. and C. Folk. 1961. Postnatal development in the Starling (Sturnus vulgaris L.) under natural conditions. Folia. Zool. 1:35-33. Kessel, B. 1957. A study of the breeding biology of the European Starling (Sturnus vulgaris L.) in N. America. Amer. Midi. Naturalist 58:257-331. King, J. R. and D. S. Farner. 1961. Energy metabolism, thermoregulation and body temperature. In A. J. Marshall (ed.), Biology and comparative physiology of birds, Vol. 2, pp. 215-288. Academic Press, New York. Korpimaki,E. 1978. Breeding biology of the Starling (Sturnus vulgaris) in western Finland. Ornis Fenn. 55:93-14. Lack, D. 1954. The natural regulation of animal numbers. Clarendon Press, Oxford. Mertens, J. A. L. 1977a. Thermal conditions for successful breeding in Great Tits (Parus major major L.). I. Relation of growth and development 4 of temperature regulation in nestling Great Tits. Oecologia (Berlin) 28:1-3. < 3 Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218

862 L. CLARK AND R. E. RICKLEFS Mertens,J. A. L. 19776. Thermal conditions for successful breeding in Great Tits (Parus major major L.). II. Thermal properties of nest and nestboxes and their implication for the range of temperature tolerance of Great Tit broods. Oecologia (Berlin) 28:31-56. Mugaas,J. N. andj. R. King. 1981. Annual variation of daily energy expenditure by the black-billed magpie: A study of thermal and behavioral energetics. Studies in Avian Biology 5:1 78. O'Connor, R. J. 1979. Brood reduction in birds: Selection for fratricide, infanticide and suicide? Anim. Behav. 26:79-96. Paynter, R. A. 1974. Avian energetics. Nuttall Ornithological Club, Cambridge. Pennycuick, C. J. and G. A. Bartholomew. 1973. Energy budget of the lesser flamingo (Phoemnicunaias minor Geoffroy). E. Afr. Wildl. J. 11:199-27. Ricklefs, R. E. 1969. An analysis of nestling mortality in birds. Smithsonian Contr. Zool. 9:1-48. Ricklefs, R. E. 1974. Energetics of reproduction in birds. In R. A. Paynter (ed.), Avian energetics, pp. 152-292. Nuttall Ornithological Club, no. 15, Cambridge. Ricklefs, R. E. and S. Peters. 1979. Intraspecific variation in the growth rates of nestling European Starlings. Bird Banding 5:338-348. Tinbergen, J. M. 1976. How Starlings (Sturnus vulgaris L.) apportion their time in a virtual single prey situation on a meadow. Ardea 64:155-17. Tinbergen, J. M. and R. Drent. 198. The Starling as a successful forager. In D. J. Feare and E. N. Wright (eds.), Understanding bird problems, pp. 83-96. Springer-Verlag, Berlin. Torre-Bueno, J. R. and J. La Rochelle. 1978. The metabolic cost of flight in unrestrained birds. J. Exp. Biol. 75:231-236. Walsberg, G. E. 1983. Avian energetics. In]. R. King, D. S. Farner, andk. E. Parkes(eds.), Avian biology, Vol. VII, pp. 161-22. Academic Press, New York. Westerterp, K. 1973. The energy budget of the nestling Starling (Sturnus vulgaris), afieldstudy. Ardea 61:137-158. Westerterp, K., W. Gortmaker, and H. Wijngaarden. 1982. An energetic optimum in brood-raising in the starling Sturnus vulgaris: An experimental study. Ardea 7:153-162. Wolf, L. and F. R. Hainsworth. 1978. Energy: Expenditures and intake. In M. Florkin and B. T. Scheer (eds.), Chemical zoology, Vol. X, pp. 37-358. Academic Press, New York. Yarbrough, C. G. 197. The development of endothermy in nestling Gray-crowned Rosy Finches (Leucosticte tephrocotis griseonucba). Comp. Biochem. Physiol. 34A:917-925. Downloaded from https://academic.oup.com/icb/article-abstract/28/3/853/99179 by guest on 4 October 218