R. Haven Wiley

Physical constraints on acoustic communication in the atmosphere: Implications for the evolution of animal vocalizations

Summary1.Acoustic communication requires not only detection of the signal but also discrimination of differences among signals by the receiver. Attenuation and degradation of acoustic signals during transmission through the atmosphere will impose limits on acoustic communication. Attenuation of sound during atmospheric transmission results primarily from atmospheric absorption, ground attenuation, scattering of a sound beam, and deflection of sound by stratified media. For maximum range of detection, therefore, animals should favor optimal positions in their habitat and optimal weather conditions. Frequency-dependent attenuation seems not to differ consistently among major classes of terrestrial habitats, such as forests and fields. Increased scattering of higher frequencies from vegetation in forests is in part matched by scattering from micrometerological heterogeneities in the open.2.In addition to frequency-dependent attenuation, two kinds of degradation during atmospheric transmission will limit a receivers ability to resolve differences among acoustic signals: the accumulation of irregular amplitude fluctuations from nonstationary heterogeneities, often atmospheric turbulence, and reverberation. Both types of degradation affect temporal patterns of amplitude or intensity modulation more than patterns of frequency modulation. Both effects should increase with carrier frequency, as they depend on the relationship between wavelength and the dimensions of scattering heterogeneities. Irregular amplitude fluctuations are more severe in open habitats and primarily mask low frequencies of amplitude modulation; reverberations are more severe in forested habitats and primarily mask high frequencies of amplitude modulation and rapid, repetitive frequency modulation. This difference between forested and open habitats could explain previous reports that birds in the undergrowth of tropical forests avoid rapid frequency modulation in their long-range vocalizations.3.Maximum range of detection is probably not the primary selection pressure on many animal vocalizations, even for territorial advertisement, except perhaps in tropical forests. Instead, acoustic signals might incorporate features that degrade predictably with range to permit a receiver to estimate the signalers distance. Future investigations might explore the propagation of animal vocalizations in relation to the usual spacing of animals in their habitat. Features that encode different kinds of information, such as individual and species identity, might propagate to different distances.4.Measurements of the transmission of sound in natural environments have often not controlled several important parameters. First, the effects of gound attenuation and scattering are not linear with range; consequently measurements of excess attenuation over different ranges in the same environment might differ. Second, the directionality of speakers and microphones will affect measurements of attenuation and reverberations in scattering environments. Third, as stationary waves shift with frequency, any single microphone placement will lie in a null for some frequencies and in a maximum for others.

Reverberations and Amplitude Fluctuations in the Propagation of Sound in a Forest: Implications for Animal Communication

Effective communication requires that the receiver not only detect the presence of a signal but also discriminate significant variations in signals. Consequently, both attenuation and degradation of the structure of acoustic signals during transmission will limit the range of communication. In this study we document two primary sources of degradation of acoustic signals during propagation through natural environments, irregular amplitude fluctuations and reverberations. Amplitude fluctuations arise especially from atmospheric turbulence, while reverberations also result from scattering surfaces, such as vegetation. Both primarily mask information coded in amplitude modulation of the signal and repetitive frequency modulation, like the trills in the songs of many passerine birds. Irregular amplitude fluctuations primarily mask low frequencies of amplitude modulation in signals. Atmospheric turbulence from wind is the primary determinant of the intensity of irregular amplitude fluctuations, although amplitude fluctuations also increase with carrier frequency and range. In contrast, reverberations depend primarily on carrier frequency and range. Reverberations are least at intermediate frequencies (2-8 kHz). At lower frequencies reverberations in sound transmission near the ground often take the form of discrete echoes, probably from canopy foliage or from the change in acoustic impedance between air in the canopy and overlying air masses. At higher frequencies reverberations usually consist of a steady decay in acoustic energy. Consequently, in contrast to irregular amplitude fluctuations, reverberations primarily mask high rates of amplitude modulation and repetitive frequency modulation in acoustic signals. Intermediate frequencies (2-8 kHz) are most suitable for long-range acoustic communication, because irregular amplitude fluctuations, reverberations, and attenuation increase with carrier frequency, while reverberations and attenuation from ground interference increase at low frequencies. The great majority of animals that engage in long-range acoustic communication use this middle range of frequencies. Perhaps because of the increase in reverberations at low carrier frequencies, the songs of rufous-sided towhees show a correlation between trill rate and the minimum frequency in trills. To minimize the effects of amplitude fluctuations and reverberations on long-range acoustic communication, signals should encode information either in frequency modulation or in repetitive amplitude modulation that allow enough redundancy or signal averaging to permit recognition of signals by receivers. Because reverberations are more severe in environments with many scattering surfaces, long-range acoustic communication in forests, as opposed to open environments, should avoid rapid amplitude modulation or repetitive frequency modulation. Among North Carolina passerine birds, species that breed in forests tend to avoid rapid repetition rates at any given frequency in their long-distance songs. The directionality of both the broadcast and reception of acoustic signals will influence the effects of scattering on reverberations and attenuation of acoustic signals. In scattering environments, the optimal directionality of sound production and reception will require compromises.

ASSOCIATIONS OF SONG PROPERTIES WITH HABITATS FOR TERRITORIAL OSCINE BIRDS OF EASTERN NORTH AMERICA

To investigate adaptations for long-range acoustic communication in birds, I analyzed associations between broad categories of habitats and properties of territorial songs for eastern North American oscines. From published recordings. I obtained three frequency properties (maximal, minimal, and dominant) and three temporal properties of songs (presence of sidebands, presence of buzzes, minimal period of repeated elements). Sidebands and buzzes indicated rapid amplitude modulation of a carrier frequency. Habitats occupied by territorial males were classified into six categories (broad-leaved or mixed forest, coniferous forest, parkland or forest edge, shrubland, grassland, and marshes). Frequencies in songs correlated strongly with body size, which varied among habitats. Analysis of covariance and phylogenetic regression, after controlling for body size, revealed an association of maximal but not dominant or minimal frequencies with habitat. In contrast, the temporal properties of song were all strongly associated with habitat, even within phylogenetic groupings. These results suggest that the temporal properties of songs of many oscines have evolved to reduce the effects of reverberation in forested habitats. Exceptional species might have retained features of song subject to degradation to permit listeners to judge distances to singers. In addition, adaptations for acoustic communication in different habitats might include differences in the perception of songs.

PERSPECTIVE: INDIRECT MATE CHOICE, COMPETITION FOR MATES, AND COEVOLUTION OF THE SEXES

When Darwin first proposed the possibility of sexual selection, he identified two mechanisms, male competition for mates and female choice of mates. Extending this classification, we distinguish two forms of mate choice, direct and indirect. This distinction clarifies the relationship between Darwins two mechanisms and, furthermore, indicates that the potential scope for sexual selection is much wider than thus far realized. Direct mate choice, the focus of most research on sexual selection in recent decades, requires discrimination between attributes of individuals of the opposite sex. Indirect mate choice includes all other behavior or morphology that restricts an individuals set of potential mates. Possibilities for indirect mate choice include advertisement of fertility or copulation, evasive behavior, aggregation or synchronization with other individuals of the same sex, and preferences for mating in particular locations. In each of these cases, indirect mate choice sets the conditions for competition among individuals of the opposite sex and increases the chances of mating with a successful competitor. Like direct mate choice, indirect mate choice produces assortative mating. As a consequence, the genetic correlation between alleles affecting indirect choice and those affecting success in competition for mates can produce self‐accelerating evolution of these complementary features of the sexes. The broad possibilities for indirect mate choice indicate that sexual selection has more pervasive influences on the coevolution of male and female characteristics than previously realized.

Signal Detection and Animal Communication

Publisher Summary This chapter provides an introduction to signal detection theory and its applications in psychophysics, and the objective is to identify general principles for the study of adaptations in animal communication. These principles can clarify the properties of signals that affect a receivers performance. It also suggests ways to extend the theory to the classification, as well as detection of signals. To apprehend the essential features of signal detection theory, it helps to consider a simple situation. Suppose an individual listens for a conspecific vocalization characterized by some feature, such as a particular frequency. In this case, the signal has a single feature, a particular frequency, which varies along a single dimension, its intensity. The earliest application of signal detection theory to a behavioral problem was the determination of human sensory thresholds. Detectability is a measure of a receivers ability to separate a signal from background stimulation; the analogous measure of ability to separate two signals is discriminability. Signal detection theory makes it clear that any receivers performance in detecting or discriminating signals has limits. Although in many situations it is reasonable to assume that an animals task involves no more than detection of an appropriate signal, in others some classification of a stimulus is essential. Signal detection theory describes decisions based on the outputs of perceptual channels. Detection and discrimination, the focus of discussion so far, suggest that the perceptual channels under consideration are sensory receptors and their immediate neural connections. Signal detection theory suggests ways that receivers and signalers could coevolve. Signal detection theory involves a level of abstraction unfamiliar in field studies of animal communication. In all of these ways, signal detection theory can advance our understanding of both the physiology and the evolution of communication.

Mechanisms and Evolution of Spacing in Animals

Animals of the same species are rarely distributed randomly. Each individual’s movements are influenced by those of its neighbors, with the result that any population exhibits a characteristic pattern of individuals’ locations and activities in space.1 In this chapter, we discuss in turn three approaches to understanding individuals’ spatial relationships: quantitative specification of patterns of spacing; analysis of the behavioral mechanisms that control spacing; and identification of the effects of natural selection on the evolution of spacing. This division separates discussion of the proximate controls of spacing, in our initial sections, from consideration of the ultimate controls, with which we conclude.

Correlates of dominance in wintering white-throated sparrows: age, sex and location

Abstract Correlates of dominance were investigated in a wintering population of white-throated sparrows, Zonotrichia albicollis , in 3 successive years at four locations along a 400-m hedgerow. Multiple regression analyses were used to detect characteristics correlated with a sparrows ability to dominate conspecifics. Age, but not length of time resident in an area, and sex, but not size, both affected dominance. Individuals had higher dominance near the centres of their ranges, and dominance in a birds first winter was related to dominance in later winters. The presence of site-dependent dominance in this species shows that dominance relations observed at any one location resulted from overlapping dominance fields of the individuals occurring there.

The use of genetic markers to estimate the frequency of successful alternative reproductive tactics

SummaryThis paper outlines a method for estimating rates of successful alternative reproductive tactics from parental exclusions known through the use of genetic markers. We review a method for calculating the probability of excluding a putative father when he is not the actual father. We adapt this method to model two mating tactics of concern to sociobiologists: extrapair copulations (EPCs) and intra-specific egg parasitism (egg-dumping). Four different types of parental exclusions are possible (both male and female, male only, female only, and ambiguous). The two models predict different proportions of each type of exclusion. Models are also generated for the case when the putative mothers or fathers genotypes are not available.We used parental exclusions from an electrophoretic study of indigo buntings (Westneat 1987b) to demonstrate these methods. The distribution of parental exclusions in the buntings departed significantly from the predictions of the egg-dumping model, but agreed well with those of the EPC model. The probability of detection for the EPC model (0.401) was then used to estimate the actual rate of extra-pair fertilizations (0.421 of all the young sampled). We present a method for calculating a confidence interval on this estimate, which ranged from 0.247 to 0.659. We concluded that these methods will allow the quantitative study of the success of alternative reproductive tactics in a wide variety of species.

Is there an ideal behavioural experiment

M uch of the field of animal behaviour rests on experimental studies of the responses of animals to different classes of stimuli. Playback experiments, which compare responses of animals to tape recordings of different sounds, are a prime example. Consequently, the proper design of these experiments is central to the scientific study of animal behaviour. Over a decade ago, a discussion of the design of behavioural experiments focused on the problems of pseudoreplication (Kroodsma 1989a, b, 1990; Searcy 1989; McGregor et al. 1992; Weary & Mountjoy 1992), and a recent paper has reviewed subsequent progress in avoiding pseudoreplication in experimental studies of bird song (Kroodsma et al. 2001). In this context, pseudoreplication consists of repeatedly presenting the same stimulus, repeatedly using the same subject, or pooling the results from presentations of similar stimuli, all problems identified by Hurlbert (1984) in some ecological experiments. Kroodsma et al. (2001) advocate a nested analysis of variance (ANOVA) to avoid these problems of pseudoreplication. In the proposed design, each subject receives a single presentation and each exemplar of a stimulus is used only once. This proposed design implies that there is only one experimental design ideally suited for comparisons of responses to different stimuli. My objective here is not to challenge nested ANOVA but to expand the discussion of behavioural experiments. To this end, I identify some compromises any experimenter must make in justifying the biological independence of subjects, the external validity of conclusions, the multiple use of exemplars and subjects, and the effects of sample size on unsuspected bias. These compromises make it less clear that any one design is universally optimal. In some circumstances, it is appro-

The constraints of digestive rate: An alternative model of diet selection

SummaryA general model is developed to predict diet selection when digestive capacity is limited and when food items differ in digestibility and digestive turnover time. Under these conditions, in order to maximize the rate of energy acquisition, animals should maximize digestive rate, by selecting food with high digestibility and rapid passage through the digestive tract. They should spend as much time as they have available to search for this food. These and other predictions differ from those of the widely used Contingency Model, which maximizes the rate of ingestion of energy. Many experiments in the literature have not discriminated between predictions of these two models. Moreover, clarification of the conditions under which these two general models apply leads to a new perspective on the diversity of foraging behavior in animals.