The influence of parental behavior on vulnerability to nest predation in tropical thrushes of an Andean cloud forest

The Skutch hypothesis predicts that parental activity around the nest may attract the attention of predators and thus, in the tropics where predation pressure may be high, selection favors reduced parental activity. This hypothesis has been questioned by studies demonstrating that parents can decrease the risk of nest predation through nest defense. The link between parental activity and predation risk may be further confounded by nest site characteristics. We examined the effects of parental behavior and nest site on clutch survival in two sympatric tropical thrushes (Myadestes ralloides and Turdus leucops). We compared survival rates of clutches in three treatments: 1) natural nests at the incubation stage, 2) unattended nests (un‐manipulated nests of the same species, with clutches unattended by parents), and 3) exposed clutches (eggs exposed in unconcealed positions, unprotected by the nest). Parental activity had a positive effect on clutch survival, which was revealed by significantly higher survival rate of clutches in attended nests compared to unattended nests. The effect of nest site was less clear: clutches in unattended natural nests survived better than clutches in exposed sites selected by humans, but results were insignificant. We propose that parent birds can exclude a group of opportunistic predators, that are able to destroy unattended clutches. Nest site characteristics may be less important in determining clutch survival in the tropics, where predator guilds are more diverse, making completely safe sites difficult to find. Our results challenge Skutch's hypothesis and point to the need for more data from tropical latitudes.     

Nest predation of grassland bird species increases with parental activity at the nest

Alexander Skutch predicted that nest predation will increase with activity at nests, and that predation should be greatest during the nestling stage when parents are feeding young. We tested this hypothesis using three ecologically similar grassland bird species nesting on the high altitude grasslands of Wakkerstroom, South Africa. Parental activity, measured as adult arrival and departure frequency from the nest, was greater during the nestling than incubation stage. Nest predation, however, did not increase with parental activity between these stages in all three study species. However, nest site effects could have confounded this result. We therefore conducted an experiment that controlled for parental activity (by reusing natural nests of the study species with artificial clutches) in order to test for nest site effects. Nests that had a high rate of predation when used by active parents had a correspondingly high rate of depredation when the same nests were reused with artificial clutches (i.e. after controlling for parental activity). This result supports the notion that variation in nest site quality is the primary factor affecting nest predation rate. We also tested whether high predation during incubation is related to nest site effects. Nest predation rates of experimental clutches placed in reused nests that had been originally (when active parents were present) depredated during incubation stage were significantly higher than those that were originally depredated during the nestling stage. Finally, once nest site effects were accounted for, nest predation showed a positive increase with parental activity during the nestling stage across species.     

Clutch size limitation in waders: experimental test in redshank Tringa totanus

Several hypotheses have been raised to explain the upper limit of clutch size at four eggs in waders (suborder Charadrii), which may play an important role in the evolution of the variety of mating and parental care systems in this group. Experimental tests of the hypotheses have produced conflicting results. It was recently suggested that the combined effects of several incubation costs of a larger clutch suffice to limit its size to four eggs in this group. Here we test the incubation-limitation hypothesis in a field experiment, in redshank Tringa totanus. We created five-egg clutches by adding one egg from another nest to a just completed four-egg clutch. Four-egg control clutches were created by replacing one of the eggs by an egg from another nest. All egg removals, additions and replacements were done before incubation started. Incubation time in five-egg clutches increased by 1 day to 24.3ǂ.23 days, compared to 23.3ǂ.32 days in four-egg clutches. Egg hatchability and nest predation rates did not differ significantly between treatments. On average five-egg clutches produced one extra chick at hatching (4.5ǂ.26 chicks) compared to four-egg clutches (3.5ǂ.27 chicks). Also when several additional costs from incubating enlarged clutches are added, redshanks by laying a fifth egg would on average increase their reproductive success at hatching by an estimated 22%. The incubation-limitation hypothesis therefore is clearly rejected in this species. Possible mechanisms behind the four-egg clutch limit in waders and ways of testing the alternatives are discussed.     

The cost of incubation in relation to clutch-size in the Collared Flycatcher Ficedula albicollis

This study examines the importance of avian incubation costs as determinants of clutch-size variation by performing clutch-size and brood-size manipulations in the same population of Collared Flycatchers Ficedula albicollis during the same breeding season. In 2 5 cases when three or more clutches of the same size were completed on the same day, we moved two eggs on the day after the last egg had been laid from one randomly selected clutch (C) to another (C) and moved two other eggs from this to a third clutch (C⁺). In 20 other cases of simultaneously completed clutches of the same size, we moved two randomly selected young from one brood to a second and from that moved two other young to a third (B, B and B⁺groups). Most females were weighed the day after completion of the clutch and 1–4 days before hatching of the young, and some of them also 10–14 days after hatching of the young. We measured the daily energy expenditure of females incubating manipulated clutches of 4, 6 and 8 eggs by means of the doubly-labelled water (D₂¹⁸O) technique and also recorded their nest attendance. Hatching success of fertilized eggs was reduced in the enlarged clutches compared with control and reduced clutches. Females expired on average 3142.6 ml CO₂ and expended 78.6 kJ per day while incubating, which corresponds to a metabolic intensity of 3.3 times BMR. Daily energy expenditure increased with clutch-size due to higher costs while incubating, and not because of changed activity patterns. There were no significant differences in length of incubation, female mass or mass changes between phases for the C, C and C⁺groups. In both the C and B groups, enlarged broods produced significantly more fledged young than control broods, and those significantly more than reduced broods. Fledgling tarsus-length and mass did not differ significantly between treatments in either the C or B groups. There was no significant difference in breeding success between clutch and brood manipulations. In this season, incubation costs did not entail significant fitness losses, expressed either as fledgling production or female condition. Also, control females could have raised more young to fledging age than they did with no apparent costs.     

Differential investment and costs during avian incubation determined by individual quality: an experimental study of the common eider (Somateria mollissima)

Individuals of different quality may have different investment strategies, shaping responses to experimental manipulations, thereby rendering the detection of such patterns difficult. However, previous clutch-size manipulation studies have infrequently incorporated individual differences in quality. To examine costs of incubation and reproductive investment in relation to changes in clutch size, we enlarged and reduced natural clutch sizes of four and five eggs by one egg early in the incubation period in female common eiders (Somateria mollissima), a sea duck with an anorectic incubation period. Females that had produced four eggs (lower quality) responded to clutch reductions by deserting the nest more frequently but did not increase incubation effort in response to clutch enlargement, at the cost of reduced hatch success of eggs. Among birds with an original clutch size of five (higher quality), reducing and enlarging clutch size reduced and increased relative body mass loss respectively without affecting hatch success. In common eiders many females abandon their own ducklings to the care of other females. Enlarging five-egg clutches led to increased brood care rate despite the higher effort spent incubating these clutches, indicating that the higher fitness value of a large brood is increasing adult brood investment. This study shows that the ability to respond to clutch-size manipulations depends on original clutch size, reflecting differences in female quality. Females of low quality were reluctant to increase investment at the cost of lower hatch success, whereas females of higher quality apparently have a larger capacity both to increase incubation effort and brood care investment.     

CLUTCH-SIZE, INCUBATION AND HATCHING SUCCESS IN THE HOUSE SPARROW AND TREE SPARROW PASSER SPP. AT OXFORD

Clutch-size, incubation and hatching success were studied in P. domesticus and P. montanus in 1961 and 1963–64 at Oxford. The most frequent clutch-size was four eggs in P. domesticus and five eggs in P. montanus . With one exception, colonies of P. domesticus showed no significant annual or local variations in its mean clutch-size; in P. montanus , however, there were significant annual variations in the mean clutch-size. Both species showed a seasonal increase followed by a decrease in their mean clutch-sizes.
Partial incubation occurred during the laying period of the clutch; sufficient incubation for continuous development of the embryo was apparently achieved when the last egg had been laid in clutches of two and three eggs in P. domesticus and in clutches of four eggs in P. montanus , but when the penultimate egg had been laid in larger clutches of both species. On average, hatching in P. domesticus occurred more or less synchronously in all eggs in clutches of two and three eggs, and in all eggs except the last one laid in larger clutches; the last egg in the larger clutches hatched up to a day after the others. It is suggested that this pattern of hatching was brought about by the pattern of incubation during the laying period.
P. domesticus had a lower hatching success than P. montanus , probably because fewer of its eggs were fertile.     

Costs and response to conspecific brood parasitism by colonial red-breasted mergansers

Costs of conspecific brood parasitism (CBP) are expected to be influenced by a species’ life history traits. Precocial birds lay large clutches, and clutches that have been enlarged by CBP can affect host fitness through a longer incubation period, displaced eggs, and lower hatching success. We examined costs and response to CBP by hosts in a population of colonial red-breasted mergansers (Mergus serrator; n?=?400 nests over 8 years) within which 29% of parasitized clutches were enlarged considerably (≥?15 eggs). Length of the incubation period did not increase with clutch size. The mean number of eggs displaced from a parasitized nest during incubation (2.8) was 2×?greater than at an unparasitized nest (1.4). Hatching success declined by 2% for each additional egg in the nest. Thus, for a nest with?≥?15 eggs, one or more fewer host eggs hatch relative to an unparasitized nest with the same number of host eggs, assuming equal probability of success for all eggs. Hosts were 40% more likely to desert nests receiving 2 or 6 experimental eggs relative to unparasitized control nests, although it is unknown whether hens deserting a nest renested elsewhere. Our study indicates that costs of CBP to hosts during nesting may be limited to those red-breasted mergansers incubating the largest clutches (≥?15 eggs), and it raises questions about the adaptive significance of deserting a parasitized clutch.     

Clutch predation in great tinamous Tinamus major and implications for the evolution of egg color

Egg camouflage has been found to reduce predation in several ground‐nesting species. Therefore, the evolution of eggs that lack camouflage in ground nesting birds is puzzling. Even though clutch predation in the tropics is high, tinamous are the only tropical ground‐nesting birds that do not build a nest and do not lay cryptic eggs. I studied predation of great tinamou clutches in a lowland tropical forest and found that risk of predation was higher during incubation when the eggs are covered by the parent, than during laying when they are exposed, suggesting that predators primarily use cues from the incubating males to locate the clutch and not cues from the eggs. Clutch size had no effect on predation rate, even though larger clutches are more conspicuous to a human observer. Predation by visual cues is likely reduced during incubation by the camouflaged plumage and high nest attendance of males. If most predators use cues from the incubating male and not the eggs to locate clutches, then conspicuous egg color may have evolved in great tinamous as an intra‐specific signal. I evaluate hypotheses that may explain the maintenance of conspicuous egg color in tinamous.     

Does the cost of incubation set limits to clutch size in common eiders Somateria mollissima?

We examined the effect of natural clutch size on the cost of incubation in a population of common eiders Somateria mollissima nesting in Tromsø, northern Norway. The body condition of females at day 5 in the incubation period was not related to clutch size (3–6 eggs), but females incubating large clutches lost more mass and had a lower body condition at day 20 in the incubation period than females incubating small clutches. Females incubating large clutches had a slightly shorter incubation period and a lower egg predation rate. The results do not support the hypothesis that the female's ability to produce eggs is the only ultimate control of clutch size in eider. Instead the results suggest that there may be an interaction between the allocation of body reserves to eggs and incubation, and that females producing large clutches allocate more of their body reserves to incubation than females producing small clutches, in order to shorten the incubation period and to minimise the risk of predation on eggs.     

Intraclutch egg-size variation in the Barnacle Goose Branta leucopsis: an egg-removal experiment

We carried out an egg-removal experiment on Barnacle Geese Branta leucopsis to test the hypothesis that tactile contact between the brood patch and eggs accumulating in the nest during laying was causally related to intraclutch egg-size variation, acting through the development of incubation behaviour. Egg removal had no effect on mean egg mass, relaying interval or clutch size: mean (±s.e.) clutch sizes for first clutches were 4.9 ± 0.9 and 4.7 ± 0.9 eggs for control and experimental females, respectively. We therefore conclude that Barnacle Geese have a determinate pattern of egg laying. Daily removal of all eggs from the first-laid egg did not affect the within-clutch pattern of egg-size variation: in both control and experimental clutches the first egg was relatively small, the second or third egg was largest and there was a linear decline in egg size to the smallest, last-laid egg. There was no significant difference in either the absolute or relative size of the last-laid eggs in control and experimental clutches. Tactile contact with eggs in the nest, therefore, is not required for the expression of intraclutch egg-size variation in Barnacle Geese. In this respect, determinate-laying Barnacle Geese differ from species with indeterminate laying patterns (e.g. gulls) where egg removal does affect the normal pattern of within-clutch egg-size variation. If hormonal changes (e.g. prolactin) associated with onset of incubation are causally related to the intraclutch decrease in egg size, then the stimuli involved appear to differ between determinate and indeterminate layers.     

BREEDING OF THE ADÉLIE PENGUIN PYGOSCELIS ADELIAE AT CAPE BIRD

Observations were made during four seasons (1967–68, 1968–69, 1969–70, 1970–71) on the breeding of Adelie Penguins at Cape Bird, Ross Island, Antarctica. Breeding data from individuals were related to date of return, laying date, clutch-size, nest location, and change of mate. Some females consistently laid near the mean date of laying, while others were consistently early or late layers. Laying date may be under direct genetic control, or may reflect feeding ability. The mean clutch-size was smaller in peripheral compared to central nests, and smallest of all in isolated nests. Clutches laid late in the season were smaller than those laid near the peak date, and small, late clutches were laid in peripheral rather than central nests. These differences may reflect age and/or feeding ability. Penguins that are better able to find food will return earlier, obtain central sites, and have larger clutches than those with a lesser ability. The two main causes of egg and chick losses were predation and parental failure. Losses were highest in single-egg clutches, at isolated and peripheral nests, and among eggs laid late in the season. These results may be partly related to the age and experience of the penguins. However, regardless of age, peripheral nesting and late laying were always disadvantageous. The sex ratio of adults in the colonies was 117♂:100♀. This may be explained by the higher mortality of females. Some males could not find partners, but females that did not breed had probably been unable to obtain sufficient food for gonad development. The return of penguins, incidence of non-breeding, adult mortality, clutch-size, and breeding success at Cape Bird were all markedly different in 1968–69 compared to the other three seasons studied. This season was marked by the persistence of sea-ice along the northwestern shores of Ross Island. The low reproductive output in 1968–69 was thought to result from a shortage of food for egg-laying and incubation.     

Predator‐mediated interactions between preferred,alternative and incidental prey in the arctic tundra

Apparent competition between prey is hypothesized to occur more frequently in environments with low densities of preferred prey, where predators are forced to forage for multiple prey items. In the arctic tundra, numerical and functional responses of predators to preferred prey (lemmings) affect the predation pressure on alternative prey (goose eggs) and predators aggregate in areas of high alternative prey density. Therefore, we hypothesized that predation risk on incidental prey (shorebird eggs) would increase in patches of high goose nest density when lemmings were scarce. To test this hypothesis, we measured predation risk on artificial shorebird nests in quadrats varying in goose nest density on Bylot Island (Nunavut, Canada) across three summers with variable lemming abundance. Predation risk on artificial shorebird nests was positively related to goose nest density, and this relationship was strongest at low lemming abundance when predation risk increased by 600% as goose nest density increased from 0 to 12 nests ha^?1. Camera monitoring showed that activity of arctic foxes, the most important predator, increased with goose nest density. Our data support our incidental prey hypothesis; when preferred prey decrease in abundance, predator mediated apparent competition via aggregative response occurs between the alternative and incidental prey items.     

Why do grebes cover their nests? Laboratory and field tests of two alternative hypotheses

Egg predation is a common feature influencing the reproductive success of open nesting birds. Evolutionary pressure therefore favours building cryptic, inconspicuous nests. However, these antipredatory pressures may be in conflict with thermoregulatory constraints, which select for dry nest material maintaining optimum temperature inside a nest cup during the absence of incubating parents. Here we examined possible trade-offs between nest crypsis and thermoregulation in Little Grebes (Tachybaptus ruficollis), which lay their eggs in floating nests built from wet plant material. As this species regularly covers its eggs with nest material, we experimentally examined (1) the rates of egg predation on covered and uncovered artificial nests and (2) possible thermoregulatory costs from nest covering by comparing temperature and relative humidity changes inside the nest cup. Results revealed that covering clutches is beneficial in terms of deterring predators, because uncovered eggs were more vulnerable to predation. Moreover, covering clutches also had thermoregulatory benefits because the mean temperature and relative humidity inside nest cups covered by dry or wet materials were significantly higher for covered compared to uncovered treatments. Covering clutches in Little Grebes therefore does not pose thermoregulatory costs.     

Incubation capacity limits maximum clutch size in black-tailed gulls Larus crassirostris

Factors determining clutch size of birds have long been the central issue in studies in life histories. It is assumed that the configuration of brood patches could limit the maximum clutch size. To test this hypothesis we manipulated clutch sizes and measured egg temperature as well as reproductive consequences in black-tailed gulls Larus crassirostris , which usually lay two egg clutches and have three brood patches. Mean egg temperature in 4-egg clutches (32.6±1.0°C) was significantly lower than in 2-egg (34.6±0.4°C) and 3-egg clutches (34.1±0.4°C), because egg temperature of the coolest egg within a 4-egg clutch was often substantially lower than the other three eggs. The proportion of eggs hatching from 4-egg clutches (11.6%) was lower than those of 2-egg (49.1%) and 3-egg clutches (52.0%). Four-egg clutches had longer incubation periods (29.6±1.3 day) than 2-egg (28.1±1.7 day) and 3-egg clutches (28.0 ±1.3 day). The results indicate that incubation capacity, which may be determined by the configuration of brood patches, limits the maximum clutch size in black tailed gulls, but not the actual clutch size typically laid.     

Multiple maternity in black-headed gull Larus ridibundus clutches as revealed by protein fingerprinting

Social monogamy with biparental care is the norm in gulls Laridae , but egg colour variation suggests that some nests may contain mixed clutches laid by more than one female. Here we use protein fingerprinting of egg albumen to assess the occurrence of mixed maternity clutches in three colonies of black-headed gulls. Among 160 analysed clutches with >1 egg, 34% contained eggs from more than one female, and 15% of the eggs in clutches >1 came from other females than the major female (laying most eggs in nest). Among clutches with 2–3 eggs 28% were mixed, and among clutches with 4 or more eggs 89% contained eggs from two or more females. There were significantly fewer eggs from the major female in mixed nests (mean=2.06±0.63 SD) than in non-mixed nests (mean=2.82±0.43 SD). In nests without evidence of female conflict, hatching success of minority eggs was similar to that of eggs from the major female (12.5 and 8.4%, respectively). In 21% of mixed maternity nests, one or more minority eggs was buried or punctured, and 25% of eggs from major females were also found evicted, suggesting conflict between females and rejection of eggs. Intra-specific nest parasitism seems the most likely cause of mixed clutches, but there are also other possible causes.     

An analysis of clutch-size in New World passerine birds

Variation in clutch-size among New World passerine birds was analysed with respect to four variables: body-mass, geographic latitude, the frequency of nest predation, and the structure of the nest. Data were analysed separately by averaging traits at three taxonomic levels: species, genus, and subfamily. Allometric scaling with body-mass did not account for significant variation in clutch-size regardless of the taxonomic level of analysis. Latitudinal effects on clutch-size were highly significant at all taxonomic levels. For species building small-pensile nests and open-cup nests, nest predation had a significant partial effect on clutch-size, with latitude held constant. When nest predation and latitude were held constant, clutch-size was significantly different among species building small-pensile nests, open-cup nests, and domed nests. These results suggest that New World passerine clutch-size is related to at least three variables: latitudinal effects, nest predation, and nest structure.     

Costs of large communal clutches for male and female Greater Rheas Rhea americana

The breeding system of the Greater Rhea Rhea americana is almost unique among birds as it combines harem polygyny and sequential polyandry, with communal egg-laying and uniparental male care. In this species, large communal clutches (more than 30 eggs) are rare and have a lower hatching success than smaller clutches. Here we analyse the proximate causes of hatching failures and the costs of large communal clutches (and therefore the costs of extensive polygyny) for males and females. We evaluated if length of the nesting period, egg viability, egg losses during incubation and male parental activity at the nest were affected by clutch size. We also evaluated if chicks hatched from large clutches have a lower survival during the first 2 months after hatching. Large clutches had longer nesting period and lower hatching success, mainly as a result of bacterial contamination of the eggs and increased hatching asynchrony. In addition, large clutches tended to lose more eggs as a result of accidental breakage or predation. Male activity at the nest and chick survival were not related to clutch size. Low hatching success, nest predation risk and energetic costs associated with large clutches penalize females that join large harems and males that accept additional eggs into the nest.     

Egg coloration and selection for crypsis in open-nesting blackbirds

High variation in egg coloration among birds has traditionally been explained as adaptation for camouflage. We tested this hypothesis by conducting reciprocal clutch exchanges (n=301) among Brewer's blackbirds Euphagous cyanocephalus , red-winged blackbirds Agelaius phoeniceus , and yellow-headed blackbirds Xanthocephalus xanthocephalus . We predicted that clutches placed against their natural nest backgrounds would have higher survival rates than heterospecific clutches. Intraspecific clutch exchanges were used as a control. Clutch survival was monitored for a 9-d period at all nests, during which time incubation rhythm and nest defense were quantified. Intraspecific clutch exchanges did not influence incubation or nest defense. For two of the species, intraspecific clutch exchanges did not influence clutch survival; in red-winged blackbirds, however, intraspecifically exchanged clutches had somewhat depressed survival curves relative to control clutches (P=0.08). The effect of interspecific clutch exchanges differed by host species. In Brewer's nests, eggs of the yellow-headed blackbird had lower survival than Brewer's eggs (P=0.02), but survival of red-winged blackbird eggs did not differ from Brewer's eggs (P=0.50). In nests of red-winged blackbirds, all three clutch types had approximately equal survival. In yellow-headed blackbird nests, eggs of the red-winged blackbird had lower survival than yellow-headed blackbird eggs (P=0.06), and survival of Brewer's eggs did not differ from yellow-headed blackbird eggs (P=0.31). These findings support a role for egg coloration as camouflage in two of the three species studied.     

Plasticity in incubation behavior and shading by king rails Rallus elegans in response to temperature

King rails experience a wide range of temperatures during the course of the breeding season throughout their rapidly contracting geographic range. Incubating parent birds are adapted to keep their eggs within a temperature range appropriate for embryo development, but king rail clutches are at risk of exceeding lethal temperatures in the latter half of the nesting season. We investigated whether behavioral plasticity during incubation enables parents to maintain clutch temperature within tolerable limits for embryo development. Video revealed that king rail parents interrupted incubation to stand above and shade their eggs. We tested the hypothesis that the onset of shading was a direct response to ambient temperature (adaptive plasticity). We monitored clutch temperature directly by experimentally adding into clutches a model egg embedded with a programmable iButton. We measured ambient temperature at the nest site simultaneously. Parents spent proportionately more time shading and less time incubating their eggs at higher ambient temperatures. Shading may primarily function in cooling the parent. The frequency and duration of shading bouts were significantly greater at higher ambient temperatures. Parents also took more frequent but shorter recesses in hotter conditions. Diurnal recesses exposed eggs to direct sunlight, and the highest clutch temperatures were recorded under these conditions. Complete hatching failure in at least one nest was attributable to high clutch temperature for an extended period. Because mean ambient temperature increases throughout the breeding season, we investigated seasonal patterns in onset of incubation and its effect on hatching rate. Later in the season, parents tended to initiate incubation earlier, and hatching asynchrony increased significantly. Together these results suggest that breeding king rails may be constrained in their ability to cope with sustained high temperatures should seasonal averages continue to rise as predicted.     

Clutch size manipulation, hatching success and offspring phenotype in the ball python (Python regius)

In a diverse array of avian and mammalian species, experimental manipulations of clutch size have tested the hypothesis that natural selection should adjust numbers of neonates produced so as to maximize the number of viable offspring at the end of the period of parental care. Reptiles have not been studied in this respect, probably because they rarely display parental care. However, females of all python species brood their eggs until hatching, but they do not care for their neonates. This feature provides a straightforward way to experimentally increase or reduce clutch size to see whether the mean clutch size observed in nature does indeed maximize hatching success and/or optimize offspring phenotypes. Eggs were removed or added to newly laid clutches of Ball Pythons ( Python regius ) in tropical Africa (nine control clutches, eight with 50% more eggs added, six with 42% of eggs removed). All clutches were brooded by females throughout the 2-month incubation period. Experimental manipulation of clutch-size did not significantly affect the phenotypes (morphology, locomotor ability) of hatchlings, but eggs in 'enlarged' clutches hatched later, and embryos were more likely to die before hatching. This mortality was due to desiccation of the eggs, with females being unable to cover 'enlarged' clutches sufficiently to retard water loss. Our results support the notion of an optimal clutch size, driven by limitations on parental ability to care for the offspring. However, the proximate mechanisms that generate this optimum value differ from those previously described in other kinds of animals. © 2003 The Linnean Society of London . Biological Journal of the Linnean Society 2003, 78 , 263–272.