3

MECHANISMS OF ANIMAL COMMUNICATION

II.

The recent resurgence of interest in proximate questions concerning animal communication has been in large part driven by the realization that perceptual processes have influenced the design of animal signals (reviews by Guilford and Dawkins, 1991, 1993; Dawkins, 1993;Pagel, 1993; Guilford, 1995; Dawkins and Guilford, 1996). Although we have known for some time that the structure of songs and calls can be accounted for, at least in part, by considering the physical characteristics of the signalling environment (e.g., Wiley and Richards, 1982; Forest, 1994) it is only quite recently that we have begun to understand the role played by the properties of signal receivers. Analyses of sexually-dimorphic 'advertisement' signals, such as the conspicuous visual ornaments of birds, have been particularly revealing (e.g., M. Andersson, 1982, 1994;Møller, 1988, 1990; Barnard, 1990; S. Andersson, 1992). Such signals are likely to have arisen as a consequence of sexual selection operating through female mate choice, and this is thought to have affected not only the size of the structures, but also their degree of bilateral symmetry (e.g., Møller, 1992; Møller and Pomiankowski, 1993). In some cases, male signals appear to have evolved in response to pre-existing biases in female receivers. Systematic comparative analyses provide compelling support for the hypothesis that the evolution of the receiver preferences pre-dates those of the male traits (Basolo, 1990, 1995a,b; Ryan et al., 1990; Ryan and Rand, 1993). These examples illustrate the considerable benefits that may accrue from a research strategy that integrates analyses of proximate factors, traditionally the province of experimental psychology and neuroscience, with an exploration of the functional problems that have always assumed central importance in behavioural ecology. Such an approach is more likely to provide a comprehensive account of behaviour than a narrow focus on the issue of adaptive significance (Dawkins, 1989; Stamps 1991).

Theoretical models of animal communication increasingly stress the concept of 'honest signalling' (Zahavi, 1975). This view suggests that many animal calls and displays are costly to produce, and that they consequently provide reliable information about attributes such as resource holding potential (e.g., Clutton-Brock and Albon, 1979) or resistance to parasites (Hamilton and Zuk, 1982; Zuk et al., 1990a,b). Formal mathematical and neural net models have been developed to describe the evolution of handicaps and ornaments (Arak and Enquist, 1993; Johnstone, 1994).

III.

REFERENTIAL SIGNALS

The theoretical and empirical advances described above have been almost exclusively concerned with displays that reflect the physical characteristics of the sender. This is not the only thing that animals signal about. Recent observational and experimental studies have demonstrated that some vocalizations encode not only individual attributes such as species, size, and motivational state, but also information about environmental events. Research on these systems, collectively termed 'referential signals', has been motivated by a very different set of theoretical concerns from those that provide the underpinning for studies of sexually-selected displays. The issues addressed have included the evolution of human language (e.g., Snowdon, 1993a; Evans and Marler, 1995) and the possibility that communication affords particular insights into the cognitive processes of non-human animals (e.g., Griffin, 1981, 1984, 1992; Cheney and Seyfarth, 1990, 1992). Analyses of referential signalling have proceeded along a parallel path, largely unaffected by recent developments in behavioural ecology. I suggest that this is unfortunate, both because studies of referential signals have not typically been concerned with issues of design, and because theoretical work on signal structure has largely ignored problems requiring an understanding of meaning.

I will not attempt a comprehensive summary of work on referential signalling, as this topic has been the subject of a number of recent reviews (Marler et al., 1992;Macedonia and Evans,