Andrew M. Stoehr

Immune defense and host life history.

Recent interest has focused on immune response in an evolutionary context, with particular attention to disease resistance as a life‐history trait, subject to trade‐offs against other traits such as reproductive effort. Immune defense has several characteristics that complicate this approach, however; for example, because of the risk of autoimmunity, optimal immune defense is not necessarily maximum immune defense. Two important types of cost associated with immunity in the context of life history are resource costs, those related to the allocation of essential but limited resources, such as energy or nutrients, and option costs, those paid not in the currency of resources but in functional or structural components of the organism. Resource and option costs are likely to apply to different aspects of resistance. Recent investigations into possible trade‐offs between reproductive effort, particularly sexual displays, and immunity have suggested interesting functional links between the two. Although all organisms balance the costs of immune defense against the requirements of reproduction, this balance works out differently for males than it does for females, creating sex differences in immune response that in turn are related to ecological factors such as the mating system. We conclude that immune response is indeed costly and that future work would do well to include invertebrates, which have sometimes been neglected in studies of the ecology of immune defense.

Sex, size, and plumage redness predict house finch survival in an epidemic

Mycoplasma gallisepticum is a well–known disease of poultry but until 1994 had not been observed in passerine birds. From 1994 to 1996, tens of millions of house finches (Carpodacus mexicanus) are believed to have died in an epidemic of mycoplasmal conjunctivitis, similar to ‘pinkeye’ in humans. The outbreak of Mycoplasma gallisepticum affected finches of both sexes but disproportionately killed males, shifting the sex ratio from male–biased to female–biased. This differential male mortality is consistent with a cost of testosterone, which is a key prediction of the immunocompetence handicap hypothesis. Males and females that survived the epidemic weighed significantly less and had significantly shorter wing chords, tarsi, and bills than did individuals before the epidemic. Male survivors also had significantly redder plumage than males that did not survive, supporting the idea that plumage brightness serves as an indicator of condition, as proposed by the honest advertisement model of sexual selection.

Testosterone and the allocation of reproductive effort in male house finches (Carpodacus mexicanus)

Abstract Testosterone has been proposed to serve as the mediator that controls the relative effort that an individual male bird will devote to mating effort versus parental effort. Here, we demonstrate a testosterone-influenced trade-off between parental and mating efforts in male house finches. Male house finches with experimentally elevated testosterone fed nestlings at a significantly lower rate, but sang at a higher rate than males without manipulated testosterone levels. Females mated to testosterone-implanted males fed nestlings at a significantly higher rate than females mated to males without testosterone implants, resulting in similar feeding rates for both treated and untreated pairs. The effects of testosterone on male house finches, however, were not as dramatic as the effects of testosterone observed in some other socially monogamous species of birds. Because extra-pair copulations are uncommon in house finches and males provide substantial amounts of parental care, these more modest effects may be due to differences in how the allocation of reproductive effort affects the costs and benefits of different reproductive behaviors.

THE EVOLUTION OF SEXUAL SIZE DIMORPHISM IN THE HOUSE FINCH. II. POPULATION DIVERGENCE IN RELATION TO LOCAL SELECTION

Recent colonization of ecologically distinct areas in North America by the house finch (Carpodacus mexicanus) was accompanied by strong population divergence in sexual size dimorphism. Here we examined whether this divergence was produced by population differences in local selection pressures acting on each sex. In a long‐term study of recently established populations in Alabama, Michigan, and Montana, we examined three selection episodes for each sex: selection for pairing success, overwinter survival, and within‐season fecundity. Populations varied in intensity of these selection episodes, the contribution of each episode to the net selection, and in the targets of selection. Direction and intensity of selection strongly differed between sexes, and different selection episodes often favored opposite changes in morphological traits. In each population, current net selection for sexual dimorphism was highly concordant with observed sexual dimorphism—in each population, selection for dimorphism was the strongest on the most dimorphic traits. Strong directional selection on sexually dimorphic traits, and similar intensities of selection in both sexes, suggest that in each of the recently established populations, both males and females are far from their local fitness optimum, and that sexual dimorphism has arisen from adaptive responses in both sexes. Population differences in patterns of selection on dimorphism, combined with both low levels of ontogenetic integration in heritable sexually dimorphic traits and sexual dimorphism in growth patterns, may account for the close correspondence between dimorphism in selection and observed dimorphism in morphology across house finch populations.

Are significance thresholds appropriate for the study of animal behaviour

Recent papers in Animal Behaviour concerning statistical power analysis have generated an interesting discussion concerning how the results of a significance test should be interpreted (Greenwood 1993; Thompson & Neill 1993; Johnsson 1996; Thomas & Juanes 1996). These discussions are welcome and overdue, and have surely improved the quality of manuscripts submitted to Animal Behaviour. I believe, however, that there is still room for improvement, and that this improvement should be relatively painless to authors. This improvement depends upon a basic understanding of the relationship between statistical power, the strength of the biological association or effect present, and P values, and what this means for the way in which the results of a significance test are interpreted. Power is defined as the probability of rejecting a false null hypothesis and depends upon three related variables: the sample size (N), the magnitude of the effect or association in the population (i.e. the size of the differences between group means, and the variance associated with these differences, or the strength of the correlation), often referred to as the effect size, and the ‘significance criterion’ (Cohen 1992). The significance criterion here refers to the preset type I error rate (alpha), or what Animal Behaviour’s ‘Instructions to Authors’ refer to as a significance threshold. The relationship between power, sample size, effect size and significance threshold implies several things. Because power increases with sample size, for a given effect, larger sample sizes will generally result in lower P values. Thus, if sample sizes for similar experiments are different, P values will differ even if both experiments have detected the same effect. Similarly, identical P values will rarely mean identical effect sizes. Additionally, power increases as measurement error decreases (something often not mentioned in discussions of power) so that two experiments that both result in a mean difference between experimental and control groups of (say) 10 g may result in very different P values if one study was conducted in the field on wild animals and another conducted under semicontrolled conditions with captive animals (Hurlbert 1994). What this means is that stating that the results of one study are statistically significant while another’s are nonsignificant, ‘moderately significant’, or ‘highly significant’ says very little about the biological conclusions one should draw from these experiments. Understanding the relationship between power and sample size, and thus P values, is important. However, power is always defined as the probability of getting a significant result at a prespecified significance threshold, traditionally set at 0.05. Animal Behaviour’s ‘Instructions to Authors’ state that ‘departure from a significance threshold of 0.05 should be stated and justified . . .’. I would like to argue that in the field of animal behaviour, preset significance thresholds and the accept/reject decisions that accompany them are not logical, and should perhaps be abandoned. This does not mean that I advocate eliminating significance tests, as I will explain, but rather that perhaps there are more reasonable ways to use them in the study of animal behaviour. My primary purpose in writing this commentary is not to suggest that what we are currently doing in the field is bad science, but simply that there is room for improvement.

Inter‐ and intra‐sexual variation in immune defence in the cabbage white butterfly, Pieris rapae L. (Lepidoptera: Pieridae)

Abstract 1. Immune defence imposes fitness costs as well as benefits, so organisms are expected to optimise, not maximise, their immune responses. This should result in variation in immune responses under varying environmental conditions.

The Number of Provisioning Visits by House Finches Predicts the Mass of Food Delivered

Abstract One classic means of assessing variation in avian foraging success and provisioning effort, counting the number of trips to the nest, assumes that parents bring equal amounts of food during each trip. We tested this assumption in male House Finches (Carpodacus mexicanus) by using video cameras to record both the number of nest visits and the mass of food delivered as measured by an electronic balance. We compared the number of feeding visits and mass of food delivered at each of three stages in the nest cycle: incubation, young nestlings, and older nestlings. The number of provisioning trips was significantly correlated with the mass of food provided by a male to his mate or to their offspring during each stage of the nesting cycle. Furthermore, this correlation became stronger as the breeding cycle progressed. These observations support the assumption that, for this species and perhaps others that carry food in their crop, the number of provisioning visits to the nest is a reasonable predictor of the mass of food provided. El Número de las Visitas del Aprovisionamiento de Carpodacus mexicanus Predice la Masa del Alimento Entregada Resumen. Un método clásico para establecer la variación en el éxito de forrajeo y esfuerzo de aprovisionamiento en aves, que consiste en contar el número de viajes hacia el nido, asume que los padres llevan igual cantidad de alimento durante cada viaje. Probamos este supuesto en machos de Carpodacus mexicanus mediante cámaras de video que registraron tanto el número de visitas al nido como el peso del alimento entregado, este último medido mediante una balanza electrónica. Comparamos el número de visitas al nido y el peso del alimento entregado en cada uno de los tres estados del ciclo de nidificación: incubación, pichones jóvenes y pichones tardíos. El número de viajes de aprovisionamiento se correlacionó positivamente con el peso del alimento provisto por un macho a su pareja o a su progenie en cada estado del ciclo de nidificación. Además, esta correlación se hizo más fuerte a medida que el ciclo de nidificación progresó. Estas observaciones apoyan el supuesto que, al menos para esta especie y otras que transportan alimento en el buche, el número de visitas de aprovisionamiento al nido predice razonablemente bien la masa de alimento provista.

Seasonal phenotypic plasticity of wing melanisation in the cabbage white butterfly, Pieris rapae L. (Lepidoptera: Pieridae)

Abstract 1. Effective thermoregulation is crucial for the fitness of small flying insects. Phenotypic plasticity of the ventral hindwing of pierid butterflies is widely recognised as adaptive for effective thermoregulation. Butterflies eclosing in cooler environments have more heavily melanised wings that absorb solar radiation, thus allowing flight under these cool conditions.

Ultraviolet Reflectance of Color Patches in Male Sceloporus undulatus and Anolis carolinensis

toads of the Western Ghats, India. Part I. Cobra 27: 1-25. . 1997b. A field guide to the frogs and toads of the Western Ghats, India. Part II. Cobra 28:1-24. . 1997c. A field guide to the frogs and toads of the Western Ghats, India. Part III. Cobra 29:113. DE SA, R. 0., R. WASSERSUG, AND A. I. KEHR. 1997. Description of tadpoles of three species of Scinax (Anura: Hylidae). Herpetol. J. 7:13-17. DUELLMAN, W. E., AND L. TRUEB. 1986. Biology of Amphibians. McGraw-Hill, New York. GOSNER, K. L. 1960. A simplified table for staging anuran embryos and larvae, with notes on identification. Herpetologica 16:183-190. HIRAGOND, N. C., AND S. K. SAIDAPUR. 1999. Description of tadpole of Rana temporalis from South India. Curr. Sci. 76:442-444. JENNINGS, R. D., AND N. J. SCOTr JR. 1993. Ecologically correlated morphological variation in tadpoles of the leopard frog, Rana chiricahuensis. J. Herpetol. 27:285-293. KAISER, H., AND R. ALTIG. 1994. The atypical tadpole of the dendrobatid frog, Colostethus chalcopis, from Martinique, French Antilles. J. Herpetol. 28:374378. LAVILLA, E. O., R. 0. DE SA, AND I. DE LA RIVA. 1997. Description of the tadpole of Atelopus tricolor. J. Herpetol. 31:121-124. ORTON, G. L. 1953. The systematics of vertebrate larvae. Syst. Zool. 2:63-75. RAO, C. R. N. 1914. Larva of Rana curtipes. Boul. Rec. Ind. Mus. 10:256. SANTOS, C. S., A. C. R. ALVES, AND S. P. DE CARVALHO E SILVA. 1998. Description of the tadpoles of Hyla giesleri and Hyla microps from southeastern Brazil. J. Herpetol. 32:61-66. SEKAR, A. G. 1990. Notes on morphometry, ecology, behaviour and food of tadpoles of Rana curtipes Jerdon, 1853. J. Bom. Nat. Hist. Soc. 87:312-313.

PARENTAL CARE IN RELATION TO BROOD SIZE IN THE HOUSE FINCH

Abstract Studies that quantify parental care in birds are often faced with the confounding effects of variation in brood size. That is, nestlings from broods of varying sizes may receive different quantities of food for reasons not entirely related to the parental quality of the adults. To control for variation in brood size, researchers often divide feeding visitation rates by brood size to yield a per-nestling feeding rate. This presents problems, however, if adults adjust food load size in response to variation in brood size. We examined the relationship between brood size and parental care in the herbivorous House Finch (Carpodacus mexicanus) by considering not only visitation rates but also food load sizes. As brood size increased, the overall visitation rate increased but both per-nestling visitation rate and per-nestling food load rate decreased. The relationship between brood size and per-nestling care was similar regardless of whether we considered visitation rate or actual food loads, suggesting that in the House Finch per-nestling visitation rate serves as a reasonable index of total mass of food received by the nestlings. However, we urge caution in assuming that per-nestling visitation rate is an adequate measure of parental care in other species.