Department of Zoology and Entomology, Pretoria University, Pretoria, South Africa, fax 27 12 3625242, e-mail JWHFerguson@zoology.up.ac.za

The concept of species has been a long-standing problem in evolutionary biology. As early as 1974, Sokal published a paper entitled ‘The species problem reconsidered’ and, more recently, Zink (1996) considered it as one of the most enduring unresolved problems in biology. The reasons for the persistence of this problem are multifarious, but two main causes are responsible.

Firstly, there is an apparent conflict between biologists over a fundamental approach to species. Some, for instance Mayr (1963) and Paterson (1981), emphasised the evolutionary processes giving rise to new species that culminated in what I term Speciation Theories. In this approach one of the fundamental questions asked is: how does a new species arise? On the other hand other workers, e.g. Simpson (1953) and Cracraft (1983), emphasised the interpretation of the pattern of biodiversity observed in nature. Here the fundamental question is: how do we recognise a species? I term this latter approach as deriving Species Definitions. All the speciation theories include a species definition, but the general shortcoming has been that these definitions are not operational (Sokal & Crovello 1970). This is exemplified in cases of what Templeton (1989) described as ‘too little sex’ (application of speciation theories to asexual and parthenogenetic taxa) and ‘too much sex’ (application of speciation theories to instances of hybridisation). Indeed, Mayr’s (1963) Biological Species Concept and Paterson’s (1981) Recognition Species Concept only apply to sexual taxa. In contrast, Species Definitions are operational and useful for deciding on the specific status of a wide variety of taxa in many situations. However, species definitions could be highly artificial and there is no reason to assume that a particular species definition has any correspondence with the underlying evolutionary processes which gave rise to the observed pattern. There is a critical need to arrive at a species definition which is widely applicable but which is connected tightly to a testable theory of speciation. (This, off course, assumes the existence of universal evolutionary processes which could apply to all situations, including asexual taxa).

A second reason for the persistence of the species problem becomes evident when considering the different speciation theories. Paterson (1981) termed this a lack of distinction between, on the one hand, adaptations shaped for particular evolutionary functions and, on the other hand, animal characteristics which are incidental by-products of other evolutionary processes. For instance, genetic incompatibility between species has been reported frequently. Viewing incompatibility as a major evolutionary force which facilitates divergence between incipient species results in a totally different perspective on the speciation process compared to viewing it as the fortuitous consequence of genetic differentiation observable after incipient species have become separate taxa.

Since 1980, a great deal of attention has fallen on the importance of animal communication systems in giving rise to the biodiversity we observe. Both Paterson’s (1981, 1985) Recognition concept and Templeton’s (1989) Cohesion concept emphasised the role which animal communication systems play in the genetic cohesion of a species. More recently there has been a flood of research on the mechanisms of sexual selection, most of which comprises aspects of signalling between the sexes or within a sex. Moreover, several authors recently have supported Fisher’s (1930) idea that a genetic correlation between female preference for male traits and the corresponding male traits could result in the origin of new species. My contribution falls in the category of speciation theory. I survey the evidence in favour of sexual selection as an evolutionary mechanism of speciation. I also survey the evidence that sexual selection is the major evolutionary force shaping pre-mating behaviour. I conclude that other mechanisms appear to play important roles in shaping pre-mating behaviour. A strict distinction between evolutionary mechanisms and incidental consequences in interpreting the evolutionary significance of animal communication systems is vital.

Nearly all of the biologists who investigated the importance of sexual selection in speciation, worked within the framework of the Biological Species Concept (BSC; Mayr 1963). The BSC distinguishes several kinds of isolating mechanisms which separate different species. These include premating and postmating isolating mechanisms (Mayr 1963). One of the important premating isolating mechanisms is ethological isolation, the failing of heterospecific pairs to recognise one another. Lande (1981, 1982) and West-Eberhard (1983) suggested that sexual selection often brings about ethological isolation between closely-related taxa. However, sexual selection need not be associated automatically with the BSC. In fact, arguments about the importance of sexual selection in speciation could be advanced under the umbrella of most of the current speciation theories. For instance, interspecific differences in courtship behaviour could be interpreted as premating isolating mechanisms under the BSC, but could also be termed cohesion mechanisms under the Cohesion Concept (Templeton 1989).

Several models have been advanced to explain speciation on the grounds of sexual selection. Fisher (1930) presented a brief description through which rapid phenotypic change in a signalling system (a runaway process) resulted from the genetic correlation between female preferences and the corresponding preferred male traits. This correlation does not imply genetic linkage of the loci coding for these traits, but a linkage disequilibrium brought about by the fact that females with a particular preference mate with males having a particular expression of a secondary sexual trait. Given that the traits are autosomal, this means that their offspring are likely to have a combination of the particular male trait and female preference. This correlation increases with each generation and results in more extreme expression of the female preference. If females in different geographical areas have slightly different preferences, the Fisherian runaway process can give rise to large differences in the pre-mating behaviour in different populations. West Eberhard (1983) offered a verbal argument which supported Fisher’s view. The Fisherian approach depends only on the genetic correlation between male and female traits: natural selection on male or female traits are not involved. Only two parameters, the female preference (p) and male trait (t) are involved, the correlation being represented by r _pt.

In contrast, the ‘good genes’ approach proposes that females prefer particular male traits which indicate the fitness of individual males as exemplified by Zahavi’s (1975) handicap principle. In this case, secondary sexual traits of males are badges which represent a fitness increase obtained by a female who prefers a particular badge. Three parameters are involved: female preference p, male trait t and male fitness v, with some correlation (r _tv) between t and v. Accordingly the evolution of male-female interaction is shaped by the correlation (r _vp) between female preference and male fitness. However, this correlation is not direct, but results from two separate correlations r _pt and r _tv (Iwasa et al. 1991).

Kirkpatrick (1982) modelled the Fisherian argument based on a two-locus model and suggested that several neutrally stable equilibria exist for female/male communication systems in which there is not an automatic tendency for particular male traits to become fixed in the population. However, these equilibria are disturbed by pleiotropy, natural selection, genetic drift or mutations which enter the population and which can cause the male/female interaction to evolve rapidly in an unpredictable way, giving rise to new systems of male-female interaction which could represent the origin of new species. Significantly, highly deleterious male traits could become fixed in a population in this way. The model was altered to represent a good genes situation by incorporating direct fitness effects such as male care associated with particular secondary sexual traits. In this case the outcome was deterministic and advantageous male traits became fixed in the model populations. By removing the unpredictability of the model in this way, geographic variation in female preferences would largely be reduced, thus limiting the probability of the origin of new male-female interactions which would represent new species.

Lande (1981, 1982) presented analytical quantitative genetic models which formalised Fisher’s arguments. These models suggest that several metastable equilibria exist for male-female premating communication which were disturbed by genetic linkages, genetic drift or mutation and resulted in the rapid changes in male-female communication. The system became stable and predictable when sexual selection was balanced by the male trait being subject to natural selection with a steeper fitness-phenotype relationship than is the case for the female preference. A major factor determining the stability of equilibria is the ratio of the genetic correlation (B) to the genetic variance (G) in the male trait with large values of B enhancing stability. Lande proposed that speciation can take place by sexual selection on male traits subject to weak natural selection and large variance in female preference. Lande (1982) modelled the above situation incorporating a cline in natural selection on a male trait. He again found metastable equilibria at which slight environmentally-induced geographic differences in male traits are amplified by sexual selection, causing rapid divergence of these traits along the cline. Tomlinson (1989) created a two-locus model of a cline in female preference for a male trait. He found that the cline in male-female interactions is unstable, eventually reduced to that observed in the absence of preference.

A common prediction of the above models is that the Fisherian process can drive male characteristics to such extreme forms that the fitness of these males is strongly negatively affected. This reduced-fitness effect could be observed if, for instance, the Fisherian process gives rise to long-tailed male birds in which the long tails cause the birds to be subject to intense predation and adverse energetic consequences. In fact, Kirkpatrick’s (1982) model predicted that near-lethal alleles could become fixed in the population.

Three models investigated sympatric speciation as a result of sexual selection. All of these assumed a single-locus female preference trait. Wu (1985) created a two-locus model (single-locus female preference, single-locus male signal) with populations subject to genetic drift. He found that genetic drift occasionally depleted all females with intermediate mate choice, which enabled a Fisher runaway process to operate on each extreme male preference characteristic, resulting in two separate species. Payne & Krakauer (1997) investigated a similar two-locus model, but in this model non-preferred males migrate away from the mating areas where they were rejected by females. Within the marginal areas these males encounter females that do not have a preference for male traits encountered in the core areas. This gives rise to genetic correlation between male traits and female preference in the marginal areas, causing the Fisherian process to bring about separate species in the core and marginal areas. Turner & Burrows (1995) created a model with a single locus female preference, but in which there is an interaction between the Fisherian process and a ‘good genes’ process via female choice for more fit males (because of the reduces-fitness effect, referred to above). Their conclusion was that sexual selection for more fit males can result in a new mate preference trait, separate from the original male-female interaction, thus giving rise to a new species. In a way their results run counter those of Fisher, Kirkpatrick and Lande who predicted that sexual selection can maintain disadvantageous traits in a population. Also, the proposed divergence in mate interaction could not be generated unless a novel allele (i.e. selecting for fitter males) was introduced at allele frequencies of 2% or higher.

A general weakness of the models proposed above is their evolutionary instability of male-female interactions which are strongly affected by initial conditions and by extraneous genetic and environmental factors. Many of the authors assumed that this instability can explain the origin of new signallng systems and species. However, if environmental conditions change continuously in the real world and if species were defined in terms of sexually selected signalling systems, we would expect larger numbers of continuous speciation events. In fact, Raikow’s (1986) question could be inverted to: why are there not many more kinds of passerines? This instability of male-female interactions has troubled biologists. Pomiankowski et al. (1991) and Iwasa et al. (1991) argued that the cost of female preference was not taken into account in previous models. They suggested that if the cost of increased predation, disease and energy expenditure associated with mate preference were taken into account, male-female interactions have very stable equilibrium points. However, if this were the case, the very stability of such systems would minimise the probability of divergence in signalling systems and speciation.

The critical parameters of most of these models have not been measured in the wild because these are difficult to obtain. For instance, the genetic correlation between male signal and female preference has not been measured quantitatively. The main reason for this is because the female response characteristic is extremely difficult to quantify. Another weakness of the existing models is that many are based on extremely simplified systems such as single-locus or haploid systems. The shortcomings of this approach are obvious in sexual taxa with complex and diverse communicatory characteristics. In addition, only a single communicatory aspect has been modelled, whereas most animal communicatory acts comprise several steps, often involving more than one sensory system. There is a fundamental lack of field data supporting the models of speciation by sexual selection. For instance, speciose taxa such as the passerine birds (Mitra et al. 1996) and the African cichlid fishes (Payne & Krakauer 1997) have been discussed within the context of speciation by sexual selection. However, field or laboratory measurements of the genetic characteristics of signalling systems in these taxa have not been performed to corroborate the predictions and applicability of the above models.

It is clear that the models of speciation by sexual selection that have been put forward are tentative and need to be tested through observations and experiments. In particular, measurements of genetic correlation and linkage disequilibrium between female preference and male traits need to be performed and the cost of female choice needs to be quantified.

If sexual selection is important in speciation, taxa subject to intense sexual selection should be more speciose than those in which sexual selection is not rife. We expect a greater intensity of sexual selection in polygynous and promiscuous taxa than in monogamous taxa. A more stringent test of the theory is possible if one could compare promiscuous taxa with nonpromiscuous equivalents. Mitra et al. (1996) found that promiscuous bird taxa are more speciose than non-promiscuous groups. However, their result is also complicated by the accumulating evidence of extensive promiscuity among monogamous birds (Birkhead et al. 1988) and that cryptic sexual selection is rife among all forms of mating systems (Eberhard 1997). Mitra et al. (1996) cautioned that their result may be the consequence of the covariance between avian mating system and an unknown factor which affects speciation rate. Sexually dichromatic coloration is one such factor. Barraclough et al. (1995) tested the hypothesis that taxa which are sexually dichromatic are more speciose than those with monomorphic plumage and found such a correlation. However, they cautioned that, even though their result is consistent with the predictions from theory, other explanations for this phenomenon also need to be investigated. The testing of predictions of models of speciation by sexual selection performed over broad ranges of taxa have therefore yielded results consistent with those expected, but which are also tentative because of the alternative explanations which could explain the observations.

A considerable body of literature exists, dealing with the recognition between conspecific mates (e.g. Templeton 1981, 1989; Paterson 1981, Gerhardt 1982). Here the emphasis is on signalling which allows conspecifics to recognise one another as appropriate partners of the same species. Paterson (1985) termed this specific-mate recognition, distinct from other forms of mate recognition, e.g. between members of an established pair of monogamous birds or mammals. I use the term ‘mate recognition’ here strictly in Paterson’s sense of specific-mate recognition. During this process an animal does not evaluate the quality or fitness of a potential partner but merely detects the partner as an individual with whom mating could take place (i.e. as an appropriate partner). Mate selection, on the other hand, is a behavioural response indicating that an individual tends to mate with one individual rather than another (Ryan & Rand 1993). This implies mate choice. Within the context of mate selection the behaviour of the signaller is used by the receiver to judge the fitness or quality of the signaller. Within the scenario of sexual selection, mate recognition is not the issue; it is implied to have occurred before sexually selected signalling takes place.

Several authors have argued that mate recognition and signalling under sexual selection should not be seen as two separate phenomena. Ryan & Rand (1993) took a moderate stance which emphasised the interaction between properties which are sexually selected and those which bring about specific-mate recognition: ‘There is considerable debate as to what evolutionary forces bring about species recognition and the preferences involved in sexual selection, but it seems clear that selection in one circumstance can have unintended consequences in another circumstance. For example, there might be selection to mate with conspecifics to avoid hybridisation with heterospecifics, but if there is variance in the species-specific signals of males or the species-specific preference function of females this should generate sexual selection’. An example of the more extreme approach is that of Polakov et al. (1995) who forcefully argued: ‘Mate recognition will almost always appear, superficially, to be an all-or-none quantum process.... However, there appears no ‘appropriate’ signal that stimulates a female to respond, but rather a gradient of signals that provide varying levels of stimulation to the female’s auditory systems. Therefore, what appears superficially as a quantum process need not be so........ At the proximate, mechanistic level, the processes of mate selection and mate recognition are essentially identical.’

Firstly, I wish to argue that mate recognition does not automatically imply mate preference. Species with strong male-male competition often have social systems where females mate with the most dominant male(s). The lek system is one in which strong male-male competition results in the most dominant males occupying positions on the lek. Bradbury & Davies (1987) surveyed male-female interactions at leks and found that, in many cases, mating at the lek is controlled by the males and that no clear female preference is measurable. In addition, monogamous social systems, observed in many bird groups, do not offer the possibility of frequent female evaluation of male quality. In all these cases mate recognition systems exist, but the prevalence or importance of mate preference systems is unclear.

Secondly, I argue that not all courtship signals function solely for mate selection. Several authors showed that mate recognition systems in vertebrates comprise a fairly complex set of characteristics. Aubin & Bremond (1983) showed that a number of characteristics of the song of the Skylark Alauda arvensis including temporal, frequency and modulation characteristics are required for species-specific recognition. Baker (1991) showed that a combination of auditory and visual signals are important for species-specific recognition in Indigo and Lazuli Buntings Passerina cyanea and P. amoena. Similarly, Ratcliffe & Grant (1983, 1985) showed that species-specific recognition in the Darwin’s Finches complex Geospiza spp. depends on a complex of vocal, visual and morphological characteristics. This suggests that mate recognition in most vertebrates is a fairly complex process, often dependent on several sensory modes. On the other hand, most of the studies quantifying female mate selection for male traits suggested that females select relatively simple traits e.g. tail length (Andersson 1982, Moller 1988) or call repertoire (Searcy & Andersson 1986). This indicates that mate selection by females involves a subset of the display characteristics of males. The argument that courtship signals do not solely function for mate selection has another side. Marler (1957) was one of the first biologists who suggested that many, of not most, bird song contains species-specific information. Since then Emlen (1972) and Shiovitz (1975) have experimentally shown this to be true for several American bird species. For instance the American buntings Passerina spp. belong to a complex of sibling species. Shiovitz found that the indigo bunting has a very complex song which has a characteristic preamble which is species-specific and which differs in each of the closely-related taxa. Both the above studies showed that the remainder of the indigo bunting song is highly variable. Emlen argues that this variable part of the song functions to indicate the motivational status of the signaller. However, more recent studies indicated that hypervariability in bird song functions to indicate male quality to females (Searcy 1984). If this argument held, the preamble of the indigo bunting song contains information useful in mate recognition while the remainder of the song contains information useful for mate selection. The above experiments were performed using male-male interactions, which does not automatically imply that females would perceive the calls in the same way. Although definitive experiments, contrasting mate recognition with mate selection in birds, have not been conducted, these have been performed on anurans and insects (Ryan & Wilczynsky 1991, Rand et al. 1992, Boake et al. 1997). This suggests that mate recognition and mate selection constitute separate processes in many taxa.

Thirdly, I argue that recognition of exaggerated male signalling traits is not a universal phenomenon and that recognition of modal male signalling characteristics is also common. Many authors, e.g. Andersson (1982; widowbirds), Moller (1988; swallows), Zuk et al. (1990; jungle fowl) Vonschantz et al. (1994; grouse), indicated that females prefer males with extreme characteristics, e.g. tail length, spur size or colouration. These conclusions are mostly based on the correlation between male characteristics and male reproductive success. However, there is another level of comparison which is not easily measured: comparing male signalling traits with female sensory characteristics. With birds, an approach to this is measuring the behavioural response of hormonally-treated females towards male signals (Searcy & Marler 1981). In invertebrates for which female response can be measured more easily and accurately, several studies (e.g. Schildberger et al. 1989, Ritchie 1996). have shown that the shape of the female sensory characteristic matches the male signal and that there is not a directional female sensitivity for exaggerated male signals. Butlin et al. (1985) showed that grasshopper females Chorthippus brunneus prefer males with modal call characteristics and that males with deviant call characteristics are avoided. This would suggest that ,in these cases, directional selection, as observed when correlating mating success with male signal characteristics, might operate through a different set of sensory characteristics. It also suggests that stabilising selection, and not directional selection as expected in most sexually selected traits, might be applicable to the sensory components of many communication systems.

The extremist argument that mate selection and mate recognition result from the same process (Polakov et al. 1995) is therefore simplistic and probably erroneous. The moderate stance of Ryan & Rand (1993) needs to be seriously considered here. Although the latter authors advocated mate recognition and mate selection as a unitary problem, they did not imply that the two processes cannot be separated. Their statement is firstly, that events in the mate selection domain can have all sorts of fortuitous consequences in the mate recognition domain, and vice versa, secondly that the mate recognition and the mate selection approaches are not mutually exclusive and, thirdly, that these two facets of communication should not be treated separately. Enough evidence exists indicating that mate recognition is an aspect of biology that is worth studying in its own right. It is easily appreciated that the above two aspects of animal communication do interact. However, relatively little attention has been paid to mate recognition. Very little theory has been developed over the last two decades, suggesting what the evolutionary dynamics of mate recognition systems might be. Since mate recognition is one of the crucial factors in defining the boundaries of a species (Paterson 1985, Templeton 1989), it is likely that a good understanding of the evolutionary processes affecting mate recognition systems will enable a better understanding of the process of speciation. Models of speciation through sexual selection and mate selection systems (Lande 1981, Kirkpatrick 1982 and others cited above) may be realised in nature. However, as indicated, these models are very tentative and, even though they were created more than a decade ago, no experimental measurement of the critical parameters in any of these models have been attempted. This is for good reason since the critical parameters of these models are almost invariably very difficult to measure in nature. However, the existence of these models of speciation should not impede investigations about the evolutionary importance of mate recognition systems as a problem worth studying in its own right.

Mate recognition systems exist as separate, identifiable objects of study in animal communication. As pointed out by Ryan & Rand (1993), mate recognition systems are probably affected by mate selection systems. For instance, such interaction could take place through sensory biases (Ryan & Keddy-Hector 1992). This interrelationship should be seen in the same light as the interrelationship between the physiological characteristics of animals and their ecological characteristics: a change in thermal tolerance affects the available habitats for a particular animal. This does not negate the independent characteristics of each of these fields. The study of the evolutionary dynamics of mate recognition systems is likely to yield valuable insights into the process of speciation. A process-oriented approach towards the understanding of speciation is fundamental for interpreting phylogenetic patterns and the evolutionary interactions which generated the biodiversity we observe.

Andersson, M. 1982. female choice selects for extreme tail length in a widowbird. Nature 299: 818-820

Aubin, T. & Bremond, J.C. 1983. The process of species specific song recognition in the Skylark Alauda arvensis. An experimental study by means of synthesis. Z. Tierpsychol. 61: 141-152.

Baker, M. 1991. Response of male Indigo and Lazuli Buntings and their hybrids to song playback in allopatric and sympatric populations. Behaviour 119: 225-242.

Barraclough, T.G., Harvey, P.H. & Nee, S. 1995. Sexual selection and taxonomic diversity in passerine birds. Proc. R. Soc. Lond. B. 259: 211-215.

Birkhead, T.R., Pellatt, J.E. & Hunter, F.M. 1988. Extra-pair copulations and sperm competition in zebra finch. Nature 334: 60-62.

Boake, C.R., Deangelis, M.P. & Andreadis, D.K. 1997. Is sexual selection and species recognition a continuum? Mating behaviour of the stalk-eyed fly Drosophila heteroneura. Proc. Natl. Acad. Sci. USA. 94: 12442-12445.

Bradbury, J.W. & Davies, N.B. 1987. Relative roles of intra- and intersexual selection. In: Bradbury, J.W. & Andersson, M.B. (eds) Sexual selection: testing the alternatives. Chichester. John Wiley & Sons.

Butlin, R.K., Hewitt, G.M. &.Webb, S.F. 1985. Sexual selection for intermediate optimum in Chorthippus brunnues (Orthoptera: Acrididae). Anim. Behav. 33: 1281-1292.

Cracraft, J. 1983. Species concepts and speciation analysis. Curr. Ornithol. 1:159-187.

Eberhard, W.G. 1997. Female control. Princeton, NJ. Princeton University Press.

Emlen, S.T. 1972. An experimental analysis of parameters of bird song eliciting species recognition. Behaviour 41: 130-171.

Fisher, R.A. 1930. The genetical theory of natural selection. Oxford. Clarendon.

Gerhardt, H.C. 1982. Sound pattern recognition in some North American treefrogs (Anura: Hylidae): Implications for mate choice. Am. Zool. 22: 581-595.

Iwasa, Y., Pomianskowski, A. & Nee, S. 1991. The evolution of costly mate preferences II. The ‘handicap’ principle. Evolution 45: 1431-1442.

Kirkpatrick, M. 1982. Sexual selection and the evolution of female choice. Evolution 36: 1-12.

Lande, R. 1981. Models of speciation by sexual selection on polygenic traits. Proc. Natl. Acad. Sci. USA. 78: 3721-3725.

Lande, R.. 1982. Rapid origin of sexual isolation and character divergence in a cline. Evolution 36: 213-223.

Marler, P. 1957. Specific distinctiveness in the communication signals of birds. Behaviour 11: 13-39.

Mayr. E. 1963. Animal species and evolution. Cambridge, Mass. Belknap Press.

Mitra, S., Landel, H. & Pruett-Jones, S. 1996. Species richness covaries with mating systems in birds. Auk 113: 544-551.

Moller, A.P. 1988. Female choice selects for male sexual tail ornaments in the monogamous swallow. Nature 322: 640-642.

Paterson, H.E.H. 1981. The continuing search for the unknown and the unknowable: a critique on contemporary ideas on speciation. S. Afr. J, Sci. 77:113-119.

Paterson, H.E.H. 1985. The recognition concept of species. In: Vrba, E.S. (ed.) Species and speciation. Transvaal Museum, Pretoria.

Payne, R.J.H. & Krakauer, D.C. 1997. Sexual selection, space and speciation. Evolution 51: 1-9.

Polakov, D.A., Backwell, P.A., Caithness, N. &. Jennions, M.D. 1995. Stabilising or directional selection in signalling systems: Investigations in a population of painted reed frogs, Hyperolius marmoratus. Sth. Afr. J. Sci. 91: 270-273

Pomiankowski, A., Iwasa, Y. & Nee, S. 1991. The evolution of costly mate preferences. I. Fisher and biased mutation. Evolution 34: 1422-1430.

Raikow, R.J. 1986. Why are there so many kinds of passerine birds? Syst. Zool. 35: 255-259.

Rand, A.S., Ryan, M.J. & Wilczinski, W. 1992. Siganl redundancy and receiver permissiveness in acoustic mate recognition by the Tungara frog, Phaesalaemus pustulosus. Am Zool. 32: 81-90.

Ratcliffe, L.M. & Grant, P.R. 1983. Species recognition in Darwin’s Finches (Geospiza, Gould) I. Discrimination by morphological cues. Anim. Behav. 31: 1139-1153.

Ratcliffe, L.M. & Grant, P.R. 1985. Species recognition in Darwin’s finches (Geospiza, Gould) III. Male responses to playback of different song types, dialects and heterospecific songs. Anim. Behav. 33: 290-307.

Ritchie, M.G. 1996. The shape of female mating preference. Proc. Natl. Acad. Sci. USA. 93: 14628-14631.

Ryan, M.J. & Rand, A.S. 1993. Species recognition and sexual selection as a unitary problem in animal communication. Evolution 47: 647-657.

Ryan, M.J. & Wilczynsky, W. 1991. Evolution of intraspecific variation in the advertisement call of a cricket frog (Acris crepitans, Hylidae). Biol. J. Linnean Soc. 44: 249-271.

Ryan, M.J. & Keddy-Hector, A. 1992. Directional patterns of female mate choice and the role of sensory biases. Am. Nat. 139: S4-S35.

Schildberger, K., Huber, F. & Wohlers, D.W. 1989. Central auditory pathway: Neuronal correlates of phonotactic behaviour. In: Huber, F., Moore, T.E. & Loher, W. (eds) Cricket behaviour and neurobiology. Ithaca, Comstock.

Searcy, W.A. 1984. Song repertoir size and female preferences in Song Sparrows. Behav. Ecol. Sociobiol. 14: 281-286/

Searcy, W.A. & Andersson, M. 1986. Sexual selection and the evolution of song. Ann. Rev. Ecol. Syst. 17: 507-533.

Searcy, W.A. & Marler, P. 1981. A test for responsiveness to song structure and programming in female sparrows. Science 213: 926-928.

Shiovitz, K.A. 1975. The process of species-specific song recognition by the Indigo Bunting, Passerina cyanea, and its relationship to the organisation of avian acoustical behaviour. Behaviour 55: 128-179.

Simpson, G.G. 1953. The major features of evolution. New York. Columbia University Press.

Sokal, R.R. 1974. The species problem reconsidered. Syst. Zool. 22: 360-374.

Sokal, R.R. & Crovello, T.J. 1970. The biological species concept: A critical evaluation. Am. Nat. 104, 127-153.

Templeton, A.R. 1981. Mechanisms of speciation - a population genetic approach. Annu. Rev. Ecol. Syst. 12, 23-48.

Templeton, A. R. 1989. The meaning of species and speciation: a genetic perspective. In: Otte, D. & Endler, J.A. (eds) Speciation and its consequences. Sunderland, Mass. Sinauer.

Tomlinson, I.P.M. 1989. Models of clines in sexual selection by female choice. Biol. J. Linnean Soc. 36: 331-348.

Turner, G.F. & Burrows, M.T. 1995. A model of sympatric speciation by sexual selection. Proc. R. Soc. Lond. B. 260: 287-292.

Vonschantz, T., Grahn, M. & Goransson, G. 1994. Intersexual selection and reproductive success in the pheasant Phasianus colchicus. Am. Nat. 144: 510-527.

West Eberhard, M.J. 1983. Sexual selection, social competition, and speciation. Q. Rev. Biol. 58: 155-181

Wu, C-I. 1985. A stochasic simulation study on speciation by sexual selection. Evolution 39: 66-82.

Zahavi, A. 1975. mate selection-a selection for a handicap. J. Theor. Biol. 53, 205-214.

Zink, R.M. 1996. Species concepts, speciation, and sexual selection. J. Avian Biol. 27,1-6.

Zuk, M., Thornhill, R., Ligon, J.D., Johnson K., Austad S., Ligon, S.H., Thornhill N.W. & Costin, C. 1990. The role of male ornaments and courtship behaviour in female choice of Red Jungle Fowl. Am. Nat. 136: 459-473.