Learning and evolution

Summary of the PhD-thesis by Joost Beltman
Universiteit Leiden
Thursday 21 october 2004

Evolution by Learning

Learning is widespread in the animal kingdom. It is difficult to give a rigorous and all-encompassing definition of learning (Papaj & Prokopy 1989), but one that suffices for the purpose of this thesis is "a more or less permanent change in behavior that occurs as a result of experience". Still, there exists a large variety in the details of different types of learning processes such as habituation, imprinting, and conditioning. For example, differences occur in what is learned, from whom it is learned, and when it is learned. In this thesis I focus on a process similar to sexual imprinting, which is the acquisition of mate preferences through learning the characteristics of parents or siblings early in life (Lorenz 1937).

It has been known for a long time that learning can influence the speed and the outcome of evolution. One of the first researchers who recognized the potential role of learning in evolution was Baldwin (1896). He suggested that learning might result in fast adaptation (now often called the Baldwin effect). In that case learning is said to "guide evolution" (Hinton & Nowlan 1987; Maynard Smith 1987). On the other hand, learning could potentially constrain evolutionary change, by masking fitness differences between individuals (Papaj 1993). Egas et al. (2004) argue that learning might be most likely to guide evolution under regimes of disruptive selection, while the constraining effect of learning would be more probable under regimes of stabilizing selection.

Apart from possibly affecting the speed of evolution, learning has also been proposed to inuence the outcome of evolution. It has been hypothesized to inuence the evolution of interspecific avian brood parasitism (ten Cate & Vos 1999), the evolution of conspicuous characteristics through sexual selection (ten Cate & Bateson 1988), and speciation (Grant & Grant 1997; Irwin & Price 1999; Owens et al. 1999; Slabbekoorn & Smith 2002; Egas et al. 2004). Given the abundance of ideas on the impact of learning on evolution, it is perhaps surprising that these ideas have often not been tested with the aid of theoretical models (but see for example Laland 1994a,b; Aoki et al. 2001; Ellers & Slabbekoorn 2003). In this thesis such formal models are used to study the role of learning in two of the above mentioned processes, namely the evolution of interspecific avian brood parasitism, and the evolution of new species.

The brood-parasitic viduid finches.
Although the ideas that underly the hypotheses tested in this thesis are relevant for a broad range of animal species, the initial source of inspiration was one particular group of bird species. They are finches of the genus Vidua (whydahs and indigobirds, family Viduidae) that live in Africa. The majority of species in this genus of brood parasites is associated with a single estrildid finch host species. The young of the viduid finches show a striking resemblance to the young of their hosts: both the mouth patterns and the plumage characteristics of the foster species are mimicked (Nicolai 1964; Payne 1973b). This species-specific mimicry gives the nestlings an increased survival probability in comparison with nestlings that do not mimic the host young (Payne et al. 2001), possibly because the mimicry influences the amount of food received from the foster parents. Among the five avian clades in which obligate interspecific brood parasitism occurs, the Viduidae are special because they are the only examples in which the cross-fostered young learn the heterospecific songs of their foster species (Davies 2000). This occurs in both sexes: males learn to produce the heterospecific songs of their foster father (Payne 1973b; Payne et al. 1998), and females develop a mating preference for males that sing a song resembling that of their foster father (Payne 1973a; Payne et al. 2000). Furthermore, mature females use the learned song to locate host nests for parasitization (Payne et al. 2000).

Song learning and the evolution of brood parasitism.
The phenomenon of sexual imprinting, which is widespread among birds (ten Cate & Vos 1999), renders a potential problem for brood-parasitic species. Because obligate interspecific brood parasites are raised by heterospecifics, imprinting on characteristics of the foster species could lead to difficulties with finding a conspecific partner at a sexually mature age. It can be expected that brood parasites have evolved mechanisms to circumvent this potential problem. Finding out what these mechanisms might consist of is an interesting research question. For instance, Hauber et al. (2001) gathered evidence that broodparasitic cowbirds (Molothrus ater) use a particular nonlearned vocalization as a kind of password to find conspecifics.

Even though present-day brood-parasitic species are expected to have solved the problem of "misimprinting" on the foster species, it is conceivable that such misimprinting may have been an obstacle at the time of the evolution of their brood-parasitic behavior (Sorenson 1994; Slagsvold & Hansen 2001). In the Viduidae, this evolutionary obstacle may have been diminished because sexual imprinting occurs on song characteristics rather than on visual characteristics. This is different because males are not able to adopt the visual characteristics of other species, but they could learn to sing the heterospecific songs of the foster species and thus attract conspecific females who are imprinted on these heterospecific songs. Hence, males and females that have learned the heterospecific songs may mate with each other, maybe after having been rejected by the foster species. For this reason, some authors suggested that a flexible song learning mechanism as it occurs in the Viduidae promotes the evolution of broodparasitic behavior (Nicolai 1964; ten Cate & Vos 1999). This hypothesis raises two interesting questions: First, has this mechanism played a role in the origin of the brood-parasitic behavior of the Viduidae? A second question is whether flexible song learning indeed promotes the evolution of brood-parasitic behavior. One of the subjects of this thesis concerns the latter, more general, of the two questions which is analyzed by means of population dynamical models.

Learning of habitat features and speciation
In theoretical studies that deal with speciation, researchers have often restricted themselves to the investigation of the plausibility of sympatric speciation (as opposed to allopatric speciation), usually ignoring learning processes. A problem with sympatric speciation is that random mating will normally prevent the splitting of a single gene pool in two separate pools, even if disruptive selection on an ecological trait would favor the evolution of extreme phenotypes. Learning could possibly assist speciation through the generation of assortative mating, as can be explained with the example of the brood-parasitic finches.
Speciation in the Viduidae might be promoted according to the following hypothetical scenario (Payne 1973b; Klein & Payne 1998; ten Cate 2000; Sorenson & Payne 2001): Suppose that some females accidentally lay their eggs in the nests of a new foster species. When these eggs are accepted and the young are raised successfully, the cross-fostered young will learn the new songs. That is, the males will produce the new songs, and the females will develop a preference for the new songs. Therefore, the young will most likely mate assortatively with each other, and the females will prefer to lay their eggs in the nests of the new foster species. Hence, the result of these events could be a new specialist broodparasitic species. The finding that egg-laying mistakes occur occasionally in the wild (Payne 1973b), and that these sometimes result in hybrids (between different parasitic species) with a history of several generations (Payne & Sorenson 2004) is consistent with the above scenario.
Speciation following the learning of foster species' song that is thought to occur in the Viduidae is probably a more general phenomenon. The essential ingredients of the process are: First, a previously unused host or habitat is being colonized (referred to as habitat in the rest of this thesis). Second, features of the habitat are learned, and this causes habitat-specific assortative mating and a tendency to produce offspring in that habitat. In the example of the brood-parasitic finches the heterospecific songs comprise the learned host features.
Besides the viduid finches there exist many species in which the learning of features of the habitat influences mating and the location where young are produced. Usually this occurs through a preference for the habitat which animals experience at a young age: the impact on the mating behavior and on the location where young are produced is then a side-effect of the habitat preference. Many insect species are prime candidates for this, because mating and oviposition often take place on the exploited host. Besides insects, some fish species also imprint on their local habitat (e.g., salmon) to return to it for reproduction. Migration in birds provides another example: migration routes are often learned at a young age, and if individuals end up by accident in a new geographical area, this may lead to reproductive isolation from the ancestral group. References to specific examples where the learning of habitat features is known to influence mating behavior and the location where young are produced are given in chapter 3. Notice that I do not aim to provide an exhaustive review; rather the purpose is to show that the learning of habitat features is a widespread phenomenon.
I will refer to the above discussed scenario as "speciation through a learned habitat preference". Note that this phrase is not meant to imply that learning is a sufficient cause of speciation; rather the name is introduced to make a distinction between a learned habitat preference and other factors that potentially play a role in speciation. This speciation mechanism has been suggested by several researchers, in general as well as in specific contexts (e.g., Thorpe 1945; Maynard Smith 1966; Payne 1973b; Rice 1984; Kondrashov & Mina 1986; West-Eberhard 2003). Despite this general acceptance of its importance, the speciation mechanism has not received much theoretical attention. For example, Diehl & Bush (1989) study the effect of sympatric and parapatric conditions on specation, and they note the similarity between geographical separation and a learned habitat preference (see also Kondrashov & Mina 1986, and chapter 3 of this thesis). Furthermore, a learned habitat preference is included in the sympatric speciation models of Kawecki (1996, 1997): One of the habitat choice mechanisms that he examines concerns the invasion of an allele that causes a preference for the natal habitat (note that he does not mention the role of learning in the development of this preference). However, rather than to focus on the role of learning, his aim is to demonstrate that a transient instead of a balanced polymorphism at loci affecting fitness in different habitats can be sufficient for the evolution of a habitat preference (and thus for the initial stages of sympatric speciation). The purpose of this thesis is to concentrate on the role of a learned habitat preference. In particular, it is examined under what circumstances speciation through a learned habitat preference could take place, with the aid of formal models. Notice that the models are not tailored to one particular species, because the purpose is not to find out whether speciation through a learned habitat preference occurs in a specific example.

Although the presence of a learned habitat preference may assist speciation, this is certainly not the only possibility to attain assortative mating. For instance, the same could be achieved by means of a genetically determined habitat preference. Speciation through the evolution of such a genetic habitat preference has been studied using formal models (e.g., Rice 1984; Diehl & Bush 1989; Johnson et al. 1996; Kawecki 1996, 1997; Fry 2003). These approaches come close to that of this thesis, differing mainly in whether the habitat preference is determined genetically or by learning. A logical follow-up question is thus what will happen when both mechanisms are considered at the same time. This is another goal of this thesis: to make a comparison between speciation through a genetic and through a learned habitat preference.

Outline of the thesis.
The thesis can be divided in two parts. In the first part, the impact of the learning of foster species' song on the evolution of specialist avian brood parasitism is considered. In the second part, the effect of a learned habitat preference on speciation is examined, as well as the evolution of such a learned habitat preference in comparison with the evolution of a genetically determined habitat preference.

Part I: the evolution of brood parasitism.

In chapter 2 I examine the hypothesis that a flexible song learning mechanism (copying the heterospecific songs of the foster species) facilitates the evolution of brood-parasitic behavior. This is done by means of two versions of a population dynamical model, a recurrence equation model and an individual-based model. The recurrence equation model is used to study the dynamics of a population of brood parasites that compete with their nestbuilding ancestors. The influence of the exibility of song learning on the permanent establishment of the brood parasites is examined. The results show that, contrary to what other researchers hypothesized, flexible song learning is an obstacle to the evolution of brood parasitism. This is because a high flexibility of song learning makes mate acquisition difficult: males that are preferred by brood-parasitic females are initially rare, which results in a low fitness of brood parasites due to costly mate choice. This problem is not present when individuals learn only songs that they are genetically predisposed to learn, because in that case all males are equally attractive to brood-parasitic females. Results from a, more realistic, individualbased model, where the brood-parasitic trait can evolve more gradually, confirm these findings. However, it is also shown that the obstacle of flexible song learning can be overcome quite easily when males also are carriers of the brood-parasitic trait. This is probably a consequence of brood parasitism being a neutral trait in males: this increases the number of mutants carrying genes for brood parasitism, and thus makes the female task of finding suitable partners easier.

Part II: speciation.

In chapters 3 and 4, it is examined under what circumstances speciation through a learned habitat preference can take place. This can occur in three steps: (i) colonization of a previously unused habitat, (ii) genetic divergence of the groups by adaptation to the habitats, and (iii) a decrease of genetic mixing between the lineages. These steps are studied in detail with a gene-culture coevolutionary model. It is assumed that density is regulated separately in each of two habitats, and that the viability of an individual depends on its genotype as well as on which habitat it exploits.

In chapter 3, I focus on the first steps, colonization of a new habitat, and ecological adaptation to the new habitat. The effect that the learning of habitat features has on mate choice and on the location where young are produced is modelled as two separate parameters. When these effects are strong, genetic divergence can be either easier or more difficult than when the effects are weak, depending on female fertility and on the viability of heterozygotes in the ecological trait.

In chapter 4 I examine the completion of the speciation process by the decrease of genetic mixing between the two adapted lineages. Hence, in this chapter the selection pressures on the strengths with which learning influences mate choice and the location where young are produced are studied, both before and after genetic divergence. It is shown that, as soon as two genetically different lineages have evolved, the decrease of genetic mixing is straightforward: There is then selection toward producing young more frequently in the habitat that individuals exploit themselves, and toward stronger assortative mating between individuals that use the same habitat. When genetic divergence between the groups fails, speciation still often occurs because selection favors producing young in the own habitat. As a result, conditions may be created that allow for genetic divergence at a later evolutionary stage, after which a decrease of genetic mixing completes speciation as before.

In conclusion, chapters 3 and 4 show that speciation through a learned habitat preference is an extremely effective mechanism. In chapters 3 and 4 speciation is investigated under the assumption that a learned habitat preference is already present at the onset of the speciation process. This leaves several open questions: First, what is the effect of different learning abilities on speciation? Second, can the learning of habitat features evolve due to disruptive selection on an ecological trait that favors specialization on two habitats? Third, is speciation in the latter case more likely to occur through the evolution of a genetic or learned habitat preference? These topics are covered in chapters 5 and 6, in which I consider the evolution of both an ecological trait, and two traits that determine respectively the genetic and learned component of the habitat preference.

In chapter 5, an adaptive dynamics approach is used incorporating the effect of habitat choice on the location where young are produced, but not its impact on mating. With this model the course of evolution until the presence of disruptive selection can be examined (from that moment on mating can no longer be neglected). It is shown that when the learning ability is high, disruptive selection on the ecological trait occurs in a larger area of the parameter space than when the learning ability is low. Because disruptive selection facilitates speciation, it can be concluded from this analysis that learning promotes speciation. However, a learned habitat preference is unlikely to evolve before disruptive selection is present, because selection on the genetic component of the habitat preference is stronger.

In chapter 6, this analysis is extended to investigate the complete speciation process by constructing an individual-based version of this model. The results show that, when loci are completely unlinked and learning confers only little cost, the presence of disruptive selection on the ecological trait is most likely to lead to speciation via the simultaneous evolution of a learned habitat preference. For high costs of learning, speciation is more likely to occur through the simultaneous evolution of a genetic habitat preference, but only when the number of loci coding for the traits is small. For large numbers of loci, recombination prevents completion of speciation: when the (physical) linkage between loci is strong, speciation through a genetically determined habitat preference occurs more readily than when the linkage between loci is weak.

The Viduidae revisited.
Before discussing some of the more general implications of the studies presented in this thesis, let us consider how the results can help to improve our understanding of the evolutionary events in the Viduidae and their ancestors. Using recent molecular phylogenies (Sorenson & Payne 2001, 2002) one can construct the most probable order of evolutionary events in this group (Payne, R. B., personal communication): The closest relative of the Vidua finches is the cuckoo finch (genus Anomalospiza). This species is a generalist brood parasite, and it does not mimic the songs of its foster species (Davies 2000). Furthermore, the species Vidua macroura and Vidua hypocherina, which are the basal clades in the phylogeny of the Vidua finches, are brood-parasitic but do not mimic the songs of their foster species (Klein & Payne 1998; Nicolai 1964). This suggests that brood parasitism evolved only once in this group (Sorenson & Payne 2002), and that the first brood parasites were generalists that did not learn the songs of their foster species (Payne, R. B., personal communication). Necessarily, in this scenario the mimicry of the heterospecific songs would have evolved at a later evolutionary stage than the brood-parasitic behavior. The colonization of new foster species would subsequently have resulted in several speciation events.

The results from this thesis are in accordance with the scenario suggested by the molecular phylogenies. In chapter 2, it is shown that if flexible song learning had been present from the onset of the brood-parasitic behavior, it would have been an obstacle for brood parasitism evolving.

In chapter 6, it is shown that the learning of foster species' song (a specific form of the learning of habitat features) can evolve in response to disruptive selection on ecological specialization on two foster species. In the case of the Viduidae the ecological specialization could for example result from selection on matching the mouth patterns of the parasitic young with those of the foster species' young. It seems likely that the learning of foster species' song could have evolved due to disruptive selection, because the Viduidae are songbirds that have to learn their song. The machinery in the brain needed for song learning was thus already present, meaning that the cost of learning will have been low or absent. Under these conditions, speciation is more likely to occur through a learned host preference than through a genetic preference (chapter 6). Finally, the results from chapter 5 show that once the learning of foster species' song is in place, it promotes the occurrence of new speciation events. Even though molecular phylogenies as well as results from this thesis suggest the same scenario of evolutionary events in the ancestors of the Viduidae, this cannot be conclusive on the basis of these studies alone. The suggestion that the learning of foster species' song has promoted speciation in the Vidua finches seems well-founded by experiments, observations, and genetic analyses (Payne 1973b; Payne & Payne 1995; Payne et al. 1998, 2000, 2002; Sorenson et al. 2003; Payne & Sorenson 2004). However, the question whether the copying of foster species' song or brood-parasitic behavior was the first to evolve in the ancestors of the Viduidae has not been adressed using experiments. One way to improve our insight in the behavior of the ancestors of the Viduidae would be to examine in more detail the song learning process in the closest relatives of the viduid finches. These include the generalist brood parasites that do not copy heterospecific song as well as species of Estrildidae, that are closely related to the viduids but are not brood-parasitic. Cross-fostering experiments are needed to determine what changes in the learning process may trigger these birds to learn the songs of other species. Also, an interesting option is to simulate parasitism in this family using similar experiments as used by Slagsvold & Hansen (2001).

Concluding remarks.
This thesis shows that learning can influence the outcome of evolution. Although this is not a new insight, most researchers, maybe especially theoreticians, have a tendency to disregard the role of learning. This seems unjust given that mating preferences, which obviously are crucial in determining the course of evolution, are affected by experience probably more often than not. It is important to incorporate such assumptions into formal models to check the validity of hypotheses on the impact of learning. When only verbal models are used, the conclusions drawn may be wrong or incomplete (see for example chapter 2 and Laland 1994a). One of the subjects where learning may be more important than normally realized concerns the origin of new species. In this thesis this idea is confirmed for one particular way in which learning may affect speciation, namely by means of a learned habitat preference (assuming that mating and production of young take place in the preferred habitat). Naturally, there are many other mechanisms through which learning could influence speciation, such as the development of mate preferences through sexual imprinting or mate choice copying. Although for some cases a start has been made (e.g. Laland 1994a), most of the possible scenarios still await a theoretical analysis. It seems especially rewarding to incorporate several mechanisms, including genetic counterparts of learned mate preferences, within the same modelling framework (as in chapter 6 of this thesis and in Johnson et al. 1996). Only then can a fair comparison between different mechanisms potentially causing speciation be made. Using such methods, this thesis shows the importance of a learned habitat preference in the formation of new species. The results suggest that learning in general may turn out to crucially determine the outcome of evolution.

Literature cited.
Aoki, K., Feldman, M. W. & Kerr, B. 2001. Models of sexual selection on a quantitative genetic trait when preference is acquired by sexual imprinting. Evolution 55:25-32.

Baldwin, J. M. 1896. A new factor in evolution. Am. Nat. 30:441-451. Davies, N. B. 2000. Cuckoos, cowbirds and other cheats. T. & A. D. Poyser, London.

Diehl, S. R. & Bush, G. L. 1989. The role of habitat preference in adaptation and speciation. In: Speciation and its consequences (D. Otte & J. Endler, eds), pp. 345-365. Sinauer Associates, Sunderland, Massachusetts.

Egas, M., Sabelis, M. W., Vala, F. & Lesna, I. 2004. Adaptive speciation in agricultural pests. In: Adaptive Speciation (J. A. J. Metz & U. Dieckmann, eds). Cambridge Univ. Press, in press, Cambridge.

Ellers, J. & Slabbekoorn, H. 2003. Song divergence and male dispersal among bird populations: a spatially explicit model testing the role of vocal learning. Anim. Behav. 65:671-681. Fry, J. D. 2003. Multilocus models of sympatric speciation: Bush versus Rice versus Felsenstein. Evolution 57:1735-1746.

Grant, P. G. & Grant, R. B. 1997. Hybridization, sexual imprinting, and mate choice. Am. Nat. 149:1-28. Hauber, M. E., Russo, S. A. & Sherman, P. W. 2001. A password for species recognition in a brood-parasitic bird. Proc. R. Soc. Lond. B 268:1041-1048.

Hinton, G. E. & Nowlan, S. J. 1987. How learning can guide evolution. Complex Systems 1:495-502.

Irwin, D. E. & Price, T. 1999. Sexual imprinting, learning and speciation. Heredity 82:347-354.

Johnson, P. A., Hoppensteadt, F. C., Smith, J. J. & Bush, G. 1996. Conditions for sympatric speciation: a diploid model incorporating habitat fidelity and non-habitat assortative mating. Evol. Ecol. 10:187-205.

Kawecki, T. J. 1996. Sympatric speciation driven by beneficial mutations. Proc. R. Soc. Lond. B 263:1515-1520. Kawecki, T. J. 1997.

Sympatric speciation via habitat specialization driven by deleterious mutations. Evolution 51:1751-1763. Klein, N. K. & Payne, R. B. 1998.

Evolutionary associations of brood parasitic finches (Vidua) and their host species: analyses of mitochondrial DNA restriction sites. Evolution 52:566-582. Kondrashov, A. S. & Mina, M. V. 1986.

Sympatric speciation: when is it possible? Biol. J. Linn. Soc. 27:201-223.

Laland, K. N. 1994a. On the evolutionary consequences of sexual imprinting. Evolution 48:477-489.

Laland, K. N. 1994b. Sexual selection with a culturally transmitted mating preference. Theor. Popul. Biol. 45:1-15.

Lorenz, K. 1937. The companion in the bird's world. Auk 54:245-273.

Maynard Smith, J. 1966. Sympatric speciation. Am. Nat. 100:637-650.

Maynard Smith, J. 1987. When learning guides evolution. Nature 329:761-762.

Nicolai, J. 1964. Der Brutparasitismus der Viduinae als ethologisches Problem. Z. Tierpsychol. 21:129-204.

Owens, I. P., Rowe, C. & Thomas, A. L. 1999. Sexual selection, speciation and imprinting: separating the sheep from the goats. Trends Ecol. Evol. 14:131-132.

Papaj, D. R. 1993. Automatic behavior and the evolution of instinct: lessons from learning in parasitoids. In: Insect learning: ecology and evolutionary perspectives (D. R. Papaj & A. C. Lewis, eds), pp. 243-272. Chapman and Hall, London.

Papaj, D. R. & Prokopy, R. J. 1989. Ecological and evolutionary aspects of learning in phytophagous insects. Ann. Rev. Entomol. 34:315-350.

Payne, R. B. 1973a. Vocal mimicry of the paradise whydahs (Vidua) and response of female whydahs to the songs of their hosts (Pytilia) and their mimics. Anim. Behav. 21:762-771.

Payne, R. B. 1973b. Behaviour, mimetic songs and song dialects, and relationships of the parasitic indigobirds (Vidua) of Africa. Ornithol. Monog. 11:1-333.

Payne, R. B. & Payne, L. L. 1995. Song mimicry and association of brood-parasitic indigobirds (Vidua) with dybowski's twinspot (Eustichospiza dybowskii). Auk 112:649-658.

Payne, R. B. & Sorenson, M. D. 2004. Behavioral and genetic identification of a hybrid Vidua: maternal origin and mate choice in a brood-parasitic finch. Auk 121:156-161.

Payne, R. B., Payne, L. L. & Woods, J. L. 1998. Song learning in brood-parasitic indigobirds Vidua chalybeata: song mimicry of the host species. Anim. Behav. 55:1537-1553.

Payne, R. B., Payne, L. L., Woods, J. L. & Sorenson, M. D. 2000. Imprinting and the origin of parasite-host species associations in brood-parasitic indigobirds, Vidua chalybeata. Anim. Behav. 59:69-81.

Payne, R. B., Woods, J. L. & Payne, L. L. 2001. Parental care in estrildid finches: experimental tests of a model of Vidua brood parasitism. Anim. Behav. 62:473- 483.

Payne, R. B., Hustler, K., Stjernstedt, R., Sefc, K. M. & Sorenson, M. D. 2002. Behavioural and genetic evidence of a recent population switch to a novel host species in brood-parasitic indigobirds Vidua chalybeata. Ibis 144:373-383.

Rice, W. R. 1984. Disruptive selection on habitat preference and the evolution of reproductive isolation: a simulation study. Evolution 38:1251-1260.

Slabbekoorn, H. & Smith, T. B. 2002. Bird song, ecology and speciation. Philos. Trans. R. Soc. Lond. B 357:493-503.

Slagsvold, T. & Hansen, B. T. 2001. Sexual imprinting and the origin of obligate brood parasitism in birds. Am. Nat. 158:354-367.

Sorenson, M. D. 1994. Is sexual imprinting an obstacle to the evolution of interspecific brood parasitism? In: Abstracts of the North American Ornithological Conference, Abstract 132, University of Montana, Missoula, Montana.

Sorenson, M. D. & Payne, R. B. 2001. A single ancient origin of brood parasitism in African finches: Implications for host-parasite coevolution. Evolution 55:2550- 2567.

Sorenson, M. D. & Payne, R. B. 2002. Molecular genetic perspectives on avian brood parasitism. Int. Comp. Biol. 42:388-400.

Sorenson, M. D., Sefc, K. M. & Payne, R. B. 2003. Speciation by host shift in brood parasitic indigobirds. Nature 424:928-931.

Ten Cate, C. 2000. How learning mechanisms might affect evolutionary processes. Trends Ecol. Evol. 15:179-181.

Ten Cate, C. & Bateson, P. 1988. Sexual selection: The evolution of conspicuous characteristics in birds by means of imprinting. Evolution 42:1355-1358.

Ten Cate, C. & Vos, D. R. 1999. Sexual imprinting and evolutionary processes in birds: a reassessment. Adv. Study Behav. 28:1-31.

Thorpe, W. H. 1945. The evolutionary significance of habitat selection. Anim. Ecol. 14:67-70.

West-Eberhard, M. J. 2003. Developmental plasticity and evolution. Oxford Univ. Press, New York.