NELSON R. CABEJ :: Epigenetic Principles of Evolution

Formation of species has been considered to be the crowning of a process of accumulation of genetic changes and inherited phenotypic changes in populations that, as a result of sudden geographic or geological transformations, are isolated from the rest of the parental population. According to this neoDarwinian view, the ultimate cause of species formation is external to the speciating populations. Evidently, such natural isolating events are not so frequent as to account for the evolution of many millions of extant animal species and many more extinct species.

In this chapter the speciation process will be considered as an intrinsically determined epigenetic process that can occur in sympatry. Sympatric speciation will be dealt with as a factor of evolution rather than a result of the evolutionary process. By creating in sympatry new mating systems within the parental panmictic mating system of the overall population, metazoans increase the probability of divergent evolution and may accelerate rates of evolutionary change.

Recognition of the Reality of Sympatric Speciation

The prevailing neoDarwinian concept is that speciation occurs under conditions of geographic isolation of populations, via gradual accumulation of changes in allele frequencies as a result of natural selection, gene mutations, genetic recombination, or gene drift. In the process, geographically separated populations become reproductively isolated, so that later, even on secondary contact populations evolve as separate species.

It is important to bear in mind, however, that this concept of the primacy of allopatric speciation was deduced from theoretical considerations (that speciation needs prevention of gene flow between populations and that under conditions of sympatry gene flow between populations will unavoidably occur) rather than from empirical evidence. Under this geographic orthodoxy, two basic conditions were sanctioned as necessary for speciation: physical separation of populations and the derived reinforcement upon secondary contact of incipient species.

For a long time since the 1940s, evolutionary studies have been focusing only on the geographical separation of populations, the “60-year-old blind alley” (Mallet, 2001), as almost the exclusive means of reproductive isolation that is necessary for speciation. Sympatric speciation has been rejected by evolutionists (Felsenstein, 1981) or has been considered only as a theoretical possibility of speciation with little, if any, role in the evolution of metazoans. While concluding that there is little evidence for sympatric speciation in island birds, Coyne and Price doubt whether it may represent an important mechanism of speciation (Coyne and Price, 2000).

In the last decades, however, despite the theoretical restrictions imposed by evolutionary genetics studies, investigators are increasingly accepting the possibility that sympatric speciation may have played a greater role in evolution and speciation than was generally assumed. Now it seems that the pendulum is swinging the other way and the gravity center of evolutionary studies is definitely shifting toward sympatric mechanisms of speciation (Via, 2001).

Experimental studies on the process of speciation under laboratory conditions have led to the conclusion that reproductive isolation, as a condition of speciation, may occur with or without allopatry (Rice and Hostert, 1993).

Most of theoretical studies still are guided by the assumption that disruptive selection within the population is necessary for the process of speciation via reproductive isolation to occur in sympatry. Seehausen and van Alphen (1999), e.g., think that disruptive selection acting on the existing color polymorphism might have been the cause of the rapid speciation in many cichlid fish species in East African lakes with relatively good visual conditions but they also believe that the cause of the disruptive selection is an intrinsic property, which ultimately is related with neurocognitive functions of neural circuits determining mate preferences and mate choices. However, all the experimental work for inducing reproductive isolation in laboratory populations via disruptive selection have failed, with probably a single exception (Thoday and Gibson, 1962; Fry, 2003). Under such circumstances, it is proposed that probably the only plausible model of sympatric speciation is the Bush model of populations using two habitats, of which one is used for mating. Examples of sympatric species that are reproductively isolated by host or habitat preferences may be considered proofs of sympatric speciation in nature (Fry, 2003).

Recently, models are presented of sympatric speciation without disruptive selection, in the absence of physical barriers or predators (Higashi et al., 1999). Most importantly empirical evidence is accumulating indicating that sympatric speciation occurred and is still occurring in nature. For, example, endemic cichlid fish species in some small crater lakes in Cameroon, colonized not long ago, are monophyletic and offer examples of such sympatric speciation (Schliewen et al., 1994). Studies in many insects provide strong evidence that a number of sensory (visual, olfactory, auditory) signal traits are used for creating separate mating systems within original populations leading to their reproductive isolation in the process of incipient sympatric speciation. Now, sympatric speciation is demonstrated in an adequate number of cases and is set on a firmer theoretical ground (Tregenza and Butlin, 1999).

Sympatric Speciation: Metazoans Erect Reproductive Barriers for Speciation without Changes in Genes

The bone of contention on sympatric speciation is the issue of reproductive isolation, which from the neoDarwinian view, cannot arise between two sympatric populations, i.e. in the absence of spatial separation and the resulting prevention of gene flow between them. From an orthodox neoDarwinian perspective, and from the view of the BSC (biological species concept), the sympatric speciation is impossible for the gene flow between populations in sympatry would prevent their divergent evolution (Felsenstein, 1981).

However, it may be argued, firstly, that to a certain extent, gene flow occurs (Bush and Smith, 1998 and references therein: Feder et al., 1995; Taylor et al. 1997) not only between populations of a species but even between species in nature, and there is no criterion on how much gene flow would be admissible for two populations to speciate or be considered true species. Secondly, that we know of numerous examples of changes in phenotypic characters that depend not on changes in genes but on changed patterns of gene expression alone.

Analysis of the mtDNA from six named subspecies of a wide-ranging species of cactus wren in southern California and the Baja California Peninsula revealed only two instead of six mtDNA groups. This phenomenon is common in birds in general. A survey of 41 named bird species revealed that only 3% of avian subspecies represent independent evolutionary units. It seems rational, after Crandall, to believe that subspecies represent inherited adaptive phenotypic variations of species that are not related with changes at genetic level. Indeed, if these bird subspecies would have evolved as independent evolutionary units it would be expected that they would have evolved in more than one character, which generally is not the case (Zink, 2004). Studies on continental European bird subspecies show that the independence of morphological evolution from changes in genes is by far a more widespread phenomenon than generally recognized:

97% of these species, distinct evolutionary units, failed the test of congruence, i.e. show no differences in their mitochondrial DNA. (Zink, 2004)

This observational evidence strongly suggests that morphological and genetic evolution are not necessarily related phenomena.

In general, there is no reliable evidence on existence of a relationship between the degree of morphological and physiological changes and speciation: sibling species both phenotypically and genetically are almost identical to each other and still represent separate species on their own, while different races of dog created by artificial selection are so distinctive from each other and still belong to a single species.

On theoretical grounds, it would be assumed that sympatric speciation requires a special mechanism of dividing a continuous panmictic population in two. The only mechanism that has been empirically substantiated, is that of sudden changes in the mating behavior of a group of individuals that leads to a separate mating system and reproductive isolation of the group from the rest of the sympatric population. That a behavioral mechanism takes the lead in the process of incipient speciation is not surprising if one would bear in mind that behavior is the most plastic component of the animal phenotype.

The change in animal mating behavior may arise in the form of a shift in mating preferences. Sudden shifts in mating preferences, affecting whole populations, are observed not only in the course of evolution but within the lifetime of an animal (see sections, Evolution of Receiver Biases and Evolution of Sender’s Signaling later in this chapter). Such a shift in mating preferences might automatically lead to formation of two separate mating systems and two reproductively isolated populations within the range of the original panmictic population.

Certainly, this transition is impossible to occur from the neoDarwinian view, according to which behaviors, as all other phenotypic traits, are determined by specific genes. Accordingly, under sympatric conditions, i.e., under conditions of panmixis, because of the gene flow between populations, changes in genes cannot lead to formation of two separate mating systems. Hence,

It is difficult to understand how genes for divergent ecology become correlated with genes for mate choice. (Emelianov et al., 2003)

But in view of the adequate observational and experimental evidence that sympatric speciation did occur and is still occurring, that belief is untenable. The belief stems from the faulty premise on the existence of the illusory “genes for mate choice”.

Sympatric speciation as a fact of nature invalidates all theoretical restrictions on its occurrence. Although unambiguous cases of sympatric speciation have been described about 40 years ago (Bush, 1969), because of the conceptual constraints imposed by the neoDarwinian paradigm, only recently the sympatric speciation has been recognized by most biologists as a real and widespread mechanism of speciation. This is what should be expected when theoretical concepts, built on insufficient, if any, empirical evidence, take priority over scientifically established facts.

Now, at the beginning of the 21st century, the situation has dramatically changed:

It has become clear that the traditional geographical classification of speciation modes is no longer appropriate to capture the essential complexity of many speciation processes (e.g., Doebeli and Dieckmann 2003; Mizera and Meszéna 2003). By emphasizing adaptive processes rather than restricting attention to biogeographical patterns of diversification, theoretical and experimental speciation research have taken off again to new shores. (Doebeli et al., 2005)

Speciation in sympatry is not a theoretical possibility but a demonstrated mechanism of evolution. Recent studies on 2 sibling species of cichlid fish of Lake Victoria, Africa, have shown that female mating preferences in these species are highly heritable, thus confirming the idea that female mate choice has been not only necessary but also sufficient for enabling reproductive isolation and the explosive speciation process that the cichlid fish have experienced very recently in East African lakes (Haesler and Seehausen, 2005).

Two incipient species of Drosophila melanogaster live in two slopes, the north-facing slope and the south-facing slope, of the “Evolution Canyon”, Israel. The canyon is only 100m wide at the bottom and 400m at the top, i.e. within the Drosophila flight range of several kilometers a day. Populations in two opposing slopes of the canyon, however have diverged in mate preference, body size, oviposition, and thermal preference (Michalak et al., 2001). There is no evidence on genetic differences between the two incipient species.

Sequence variation in the mtDNA of two morphologically similar, endemic cichlid (mbuna) species (Melanochromis auratus and M. heterochromis) from Lake Malawi, Africa, show that they have evolved in sympatry very recently, less than 10,000 years ago (Bowers et al., 1994).

Confronted with solid evidence on the occurrence of sympatric speciation, especially from studies on the explosive speciation of hundreds of cichlid species, with some of them having evolved several thousand years ago in East African lakes, many biologists are theoretically reconsidering the likelihood of sympatric speciation.

In their attempt, many were embarrassed by the still prevailing idea of the existence of particular genes for mate choice. The reasoning goes: genes for mate preference will be mixed up by genetic recombination during the earlier stages of the speciation process, when mating between individuals of different populations may still occur. To overcome this difficulty another assumption was made: genes for mate signaling and for mate preference are so close to each other in chromosomes that the probability of getting recombined is negligible (McCune and Lovejoy 1998).

No evidence has ever been presented to substantiate the idea of genes for mate preference or mate signaling.

However, granted that such genes really exist, it is highly improbable that nature would have arranged genes for both the mate choice and mate signaling in the same chromosome and even so close to each other as to prevent their recombination for the sake of the theory, to facilitate resolving our recombination problem. Certainly, even highly improbable events may happen once or two but rapid sympatric speciation has been far from a rare event: only in East African cichlid fish it has occurred repeatedly and independently in hundreds of cases.

It is the burden of the neoDarwinian school of thought to provide the proof that genes for mate choice do exist, that they are in the same chromosome and even so close to each other that the probability of their recombination is negligible. Indeed, even from a neoDarwinian point of view, it has been argued that

Stabilizing sexual selection generated by assortative mating can work against sympatric speciation by causing fixation at individual loci and by reducing the associations between alleles at different loci (linkage disequilibria) that are the genetic basis for sympatric speciation… Thus sympatric speciation by assortative mating is in a bind: strong assortment is needed to cause the population to fission, but it can also generate strong stabilizing selection. Stabilizing selection causes genetic variation to be lost, and it decreases the associations between alleles that are required for one population to split into two. (Kirkpatrick and Nuismer, 2004)

The cause for this insuperable difficulty for the neoDarwinian view of the speciation process seems to be the unproven, indeed false, premise on which that view is based: the assumed existence of closely linked genes for mate preference and mate signaling.

...............................................................................

Natural Selection and Sexual Selection in Sympatric Speciation

..............................................................................

Neurocognitive Populational Break-up

All the direct observational evidence on processes of sympatric speciation points to the fact that it starts with changes in sexual behavior, which make a group of individuals to exhibit mating biases toward members of the group and discriminate against the rest of individuals in the overall population. Such shifts in mating biases precede changes in genes, in morphology and physiology taking place in the process of sympatric speciation.

Adequate evidence on the occurrence of sympatric evolution during several last decades has shown that neither geographical isolation nor “genetic revolutions” are necessary for speciation (Bush, 1975) in metazoans.

Sympatric speciation, in distinction from allopatric speciation, takes place in spatially continuous populations. Separate mating systems in sympatric, spatially continuous, populations arise as a result specific changes in behavior of a group of individuals, whose preference for mating with individuals of the opposite sex displaying certain phenotypic characters, leads to their neurosensory isolation from the rest of the original population. Theoretically, formation of separate mating groups within the range of a species may result from:

- Sensory-determined preferences for particular individuals of the group, displaying specific distinctive phenotypic trait, or

- Sensory-determined preferences for specific niches of particular groups of individuals (ecological speciation?).

In all likelihood, the above neurally determined changes in mating preferences and/or neurally determined preferences for host plants, represent the basic mechanisms of sympatric speciation.

The neoDarwinian assumption that changes in mating preferences are related to changes in particular genes is rejected by the fact that in a number of well-known cases, mate preferences rapidly change and evolve without changes in genes (see Evolution of Receiver Biases later in this chapter).

Reproductive Isolation by Mate Choice

At the basis of sexual selection is the mate choice, i.e. the bias that the choosing sex (most of times females) exhibits for preferably mating with individuals of the opposing sex displaying particular phenotypic characters. A strictly preference-based, discriminatory mating occurring within a separate group of individuals of a population can, theoretically at least, lead to splitting of an original population in two reproductively isolated populations. Thus, it is possible that a speciation process may start in sympatry, with a change in mate preferences for certain sexual characters of a group of individuals within the original population. Indeed, empirical evidence suggests that sudden shifts in mating preferences of a group of individuals may be the most frequently used mechanism of reproductive isolation between populations in the initial stage of the process of speciation.

Mate choice is a neurobiologically determined decision made by an individual organism. That choice results from processing in mating behavior circuits of various sensory (visual, olfactory, acoustic, tactile, etc.) information received from prospective mates. We know that an intrinsic drive to influence these decisions stimulates the prospective mates of the opposite sex to display all their sexual characters and reproductive behaviors but we are still far from a real understanding of the nature of this decision-making information processing.

Changes in properties of neural circuits may lead to changes in mate preferences for certain phenotypic characters and to sexual isolation of populations of a single species. This suggests that mating behavior circuits in the CNS may be potential initiators of sexual isolation, reproductive isolation, and speciation. Mating decisions and their heritable changes may have important evolutionary effects (Phelps et al., 2006).

Mate choice implies mate recognition, which in turn is a function of the mate recognition system. There is no consensus among biologists on whether animals use the same traits for mate choice and for discriminating between conspecific and heterospecific individuals.

Some observations suggest that Drosophila heteroneura uses different traits for discriminating conspecifics from heterospecific individuals and for mate choice (Boake et al., 1997) but other evidence shows that mate recognition and reproductive isolation from heterospecific individuals may be based on the same traits (Ryan and Rand, 1993). Such, e.g., is the case with the wing melanin pattern in sympatric white butterflies, Pieris protodice and Pieris occidentalis, which do not hybridize under natural sympatric conditions although no postmating isolating mechanism exists for preventing their hybridization. P. protodice has a wing melanin pattern different from that of P. occidentalis but investigators have succeeded that by hybridizing the two naturally nonhybridizing species to increase wing melanization of P. protodice butterflies. They have shown that wing melanization pattern is used for both interspecific recognition and mate recognition by P. occidentalis females (Wiernasz and Kingsolver, 1992).

Experiments with two Drosophila species (D. montana and D. lummei) have shown that the male courtship song is used by females for both recognizing conspecifics and for mate choice (Saarikettu et al., 2005). Another example is that of male individuals of the repleta group of Drosophila flies, which likewise use their song both for recognizing conspecifics and for stimulating mating behavior in females (Ewing and Miyan, 1986).

Olfactory mating signals in Drosophila spp. are cuticular hydrocarbons. From studies on hybrids of two Drosophila species (D. serrata and D. birchii) it was concluded that the mechanism of mate recognition by cuticular hydrocarbons is responsible for both mate choice and evolution of reproductive isolation in Drosophila (Blows and Allan, 1998). Similarly, in frogs the advertisement call is used for both species recognition and mate choice and has functioned as a mechanism of the premating reproductive isolation of the group (Ryan and Rand, 2003)

From a strictly neoDarwinian perspective, it would be expected that traits (signals) used for mate choice have to be different from traits, whose evolution leads to reproductive isolation of diverging populations. This neoDarwinian prediction is not validated. Another neoDarwinian prediction is that the traits used for the reproductive isolation of the group would evolve gradually as a result of accumulation of genetic changes in populations. Again, contrary to this prediction, recent evidence shows that reproductive isolation between two populations may arise suddenly as a result of changes in reproductive behavior without changes in genes.

Evolution of Mating Preferences

Early evolutionary biologists doubted whether sexual selection through female choice occurred at all, because they did not think that female choice could evolve. In the past 20 years, however, evolution of female choice has been demonstrated to occur in a large number of species, including various insects, fish, birds, and reptiles.

Female (or male) choice is the behavioral expression of an intrinsic mating preference/bias for species-specific and individual (visual, auditory, olfactory) signals that the female receives via the sensory organs and perceives in its brain. Three main hypotheses for the evolution of female choice have been proposed: Fisher’s runaway hypothesis, the “good gene” hypothesis, and the sensory exploitation hypothesis. All of them focus on the preservation and spread of mating preferences by natural selection, paying very little attention, if not disregarding, the cause and mechanism of the change of mating preference, which is the key to the process of reproductive isolation and speciation.

Fisher’s Run-away Hypothesis

This is the first hypothesis for explaining the evolution of mate preferences and signal traits. The hypothesis predicts that females that display mating preference for, and mate with, males with particular trait(s) provide the male carrier of that trait with a selective advantage by producing more offspring with the same signal trait and similar mating preference. In turn, this leads to positive selection of the mating preference (hence the designation “runaway”) for exaggerated signaling traits (and exaggeration of the traits) up to a point where the cost of the development of the trait overweighs the advantage, i.e. “until checked by severe counterselection” or “disadvantageous consequences” (Lande, 1981; West-Eberhard, M.J., 1983; Kokko et al., 2003). Thus, the female preferences and male signal traits coevolve. Female preference may exert direct selection on male traits, especially by favoring female biases for male traits that may be related to fertility and parental care (Kokko et al., 2003). Evolution of female preferences is an indirect result of direct selection and evolution of male signaling traits.

It is not known whether runaway could occur in natural populations (Kirkpatrick and Ryan, 1991). Despite the large amount of empirical work on indirect benefits of mate choice, the fundamental prediction of the hypothesis that mating preferences increase the net offspring fitness has not been empirically tested. After two decades of experimental work in this field there is still no study showing that mean offspring fitness from mating with attractive males is elevated (Kokko et al., 2003).

Based on the relevant evidence, it is reasonable to consider that Fisher’s run-away hypothesis of evolution of mating preferences is still waiting to be validated.

The “Good Gene” Hypothesis

The “good gene” hypothesis holds that male attractiveness is an indicator of the presence of “good genes” for higher viability in males displaying attractive mating signals. Hence, the female mating bias is genetically correlated with male mating signals and is indirectly selected as a result of the direct selection of genes for increased fitness, which are correlated with male sexual traits. Evolution of “good genes”, thus, leads to selection and evolution of female preferences (indirect selection).

Both the runaway and good gene hypotheses hold that a genetic correlation exists between the male mating signals and female mating preferences. Kokko et al. have attempted to unify the run-away and good gene hypotheses in a hypothesis holding that the female preference and the offspring fitness may be related with each other.

There is no qualitative difference between these outcomes; rather, they are endpoints of a continuum. (Kokko et al., 2002)

The good genes and run-away hypotheses posit that female preferences may be influenced by natural selection if they lead to reproductive advantage, as it would occur when female preference may be correlated with signals of males of higher fitness qualities. Thus, selection for better fitness leads to indirect selection for male traits. This, however has not been possible to empirically substantiate.

Although it is possible that in some cases genes for hybrid inviability and mate recognition may be linked to the same chromosome, it is hard to believe that this genetic correlation may be a widespread phenomenon as to be speciationally relevant. Besides, evidence on the viability of hybrids and against reinforcement clearly overweighs the opposite evidence and the experimental support for the above hypothesis of indirect selection is inadequate (Kirkpatrick and Ryan, 1983).

Studies on the female preference for the swordtail in poeciliid fish, Xiphophorus helleri and Priapella olmecae, belonging to two sister genera, have shown that the female preference for swordtail is stronger in the species that has not evolved swordtail, what is the opposite of what would be expected from both Fisher’s runaway and good gene model (Basolo, 1998). Additional evidence on differential deposition of maternal hormones and other substances in eggs of birds mated with preferred males also contradicts the good gene model, and has led investigators to the conclusion that results of these experiments show that the father has no effect on the condition and fitness of the offspring (Balzer and Williams 1998; Cunningham and Russell, 2000).

Some cases of possible relevance in support of the good gene hypothesis come from comparative studies showing that the fluctuating asymmetry in elaborate feather ornaments in swallows (Hirundo rustica) is negatively related to the size of the ornament. In those studies it was observed that males with larger symmetric tails mated earlier and were reproductively more successful so that the female preference for males with larger and symmetric ornaments is believed to be related to higher quality of these males (Møller, 1992).

The hypothesis of good genes selection predicts that evolution of receiver biases is a by-product of natural selection because signals that are preferred are genetically correlated with other traits of higher evolutionary fitness. This prediction is refuted by experiments showing that

Trait evolution and preference evolution are often decoupled in sexual selection, that they need not evolve through genetic correlation, nor are the response properties of the receiver tightly matched to the properties of the signal, as a lock and key would be matched. Analogies between animal communication systems and human-engineered systems often stress the necessity of tightly matched signals and receivers. Studies of receiver biases suggest that such analogies might not be broadly applicable. The receiver’s past history might bias neural processing strategies toward those that are merely sufficient to enhance the receiver’s evolutionary fitness but are not optimal engineering solutions. Furthermore, tightly matched signal-receiver systems might have a selective disadvantage if they constrain the receiver’s ability to accommodate meaningful population variation. (Ryan, 1998)

The run-away and good gene hypotheses would predict that signals from the sender are necessary for maintaining female preferences. This prediction also is not substantiated, while several examples invalidating it have been presented. For example, the all-female species of the poeciliid fish, Poecilia formosa, has the same preference for body size that females of its ancestral dioecious species have and uses sperm from other species to fertilize its eggs but the heterospecific male genome is not incorporated in the genome of the offspring This indicates that P. formosa has been able to retain its ancestrally inherited mate preference, in the absence of male signaling, for evolutionarily long periods of time (Ryan, 1998).

Female guppies of the species Poecilia reticulata, prefer to mate with males with larger orange spots. In selection experiments for male attractiveness and female preferences with P. reticulata for three generations, Hall et al. also failed to obtain results that would be expected according to the above hypotheses (Hall et al., 2004). There is evidence that, contrary to the previous belief, the preference for males with large orange spots is not related to any higher fitness of these males, but it is related to a general sensory preference of females of this species for orange colored objects, including food (Rodd et al., 2002).

Physalaemus pustulosus is a monophyletic group of frogs in which females exhibit preferences for four call traits, suggesting that no correlation between the evolution of genes and the evolution of mating signals has taken place. On the contrary, the above fact lends support to the sensory exploitation hypothesis, which predicts that males evolve traits that match pre-existing female biases (Ryan and Rand, 2003).

In many animal species, during the breeding season, males form groups of mating displays (lek) where females have the opportunity to choose their mating partner. Since lekking males in general contribute only sperm, the neoDarwinian theory would predict that lekking males will not be “choosy”. Contrary to the prediction, males of the cichlid fish, Astatotilapia flaviijosephi show clear preference for larger females. Lek behavior

can make less attractive females more available to subordinate males, thereby increasing the contribution of the latter to the population gene pool and keeping genetic variability among males at a level that justifies female choice. (Werner and Lotem, 2006)

The same is true for a number of insects, fish, amphibians.

In bird species where males are the “choosy” sex, such as rock sparrows, males prefer to mate with females with large yellow breast patches, which are more likely to lay two broods a year (Griggio et al., 2005).

Generally, it may be said that there is no adequate evidence in support of the idea that preferred or “sexy” males are in possession of better genes and lead to an increase of fitness in the offspring.

Sensory Exploitation Hypothesis

According to the sensory exploitation (sensory bias) hypothesis of evolution of mate preferences, female preferences antecede male signaling traits and are males that evolve their signaling traits in order to match the female preferences. Hence, according to the hypothesis, female preferences evolve independently of male signaling and are not under direct or indirect action of natural selection. The sensory exploitation hypothesis seems to have found considerable empirical support.

This is a noncorrelational hypothesis. It holds that males evolve signaling traits for exploiting preexisting mating biases of the females and no genetic correlation is necessary for evolution of female mating preferences and male mating signals. The preexisting female bias toward certain characters determines evolution of male sexual traits so that they match female mating biases. The prediction of the hypothesis that receiver’s (mostly females) bias precede evolution of respective signaling traits of the mate is substantiated in a number of cases and is in line with the present knowledge on the evolution of the mate recognition system.

It is generally believed that no genetic correlation is necessary for male signaling traits to evolve in response to specific female preferences.

The sensory exploitation hypothesis suggests that, contrary to coevolution through genetic correlation, a trait and a preference in sexual selection - or, more generally, a signal and a receiver in animal communication - can evolve out of concert, with the evolution of one component lagging behind that of the other. If a receiver has a bias toward responding to certain signal parameters, such as louder sounds or brighter colors because they are easier to detect, we would expect the evolution of louder or brighter signals without assuming the need for genetic correlations between trait and preference, as required by indirect selection. (Ryan, 1998)

The receiver bias is a product of the neurosensory system (the system of perception of the external world by senses including neural pathways of transmission of information and perception produced in respective regions of the CNS), which evolves independently of evolution of genes. The hypothesis would predict that the sender of the signal evolves its signals for exploiting preexisting mate preferences of the receiver.

Female mate preferences evolve for female’s own benefit, generally for increasing female’s fitness. Empirical evidence in support of this hypothesis for species where males form displaying groups (leks) in front of females still lacks, but there is some evidence that this occurs sometimes in nonlekking species. So, e.g., females of several species prefer mating males that provide nest sites or care for the young (Kirkpatrick and Ryan, 1991). While this hypothesis is more attractive to many biologists, an explanation has not been provided on how the male can evolve mating signals to satisfy female mating biases.

The visual, auditive and other signals, emanated from the sender (most of times a male), are received by sensory organs of the receiver, which converts them into patterns of electrical signals, transmit them for processing in neural circuits, where the input is compared with the neurocognitive mating preference standards before making any mate choice decision. This suggests that properties of the neural circuits are essential for generating mating preferences. This gives us some hints on the surprising and still unexplained phenomenon of the sudden appearance and changes in mating preferences, in particular on the divergence of mate preferences without changes in genes. Such sudden changes in mating preferences could be predicted in view of the neurally determined high plasticity of the function of the “mate recognition system”.

Sensory drive may do more than offer a quirky exaptive alternative for how mate choice and mating biases evolve. It may provide the initial ‘nudge’ often required to initiate choice-display. If males are not advertising, females are unable to choose; and it does not pay males to advertise unless females are choosing. Something needs to happen to make females choose and thereby make it worthwhile for males to display. (Kokko et al. 2003)

While we are not aware of what exactly happens, certainly we know where it happens: in the the neural circuits for mate choice and there is where the “nudge” comes from. Both male display and female choice are mating behaviors that, like any other behavior, have a nongenetic, neural substrate and neurocognitive basis.

All three models (Fisher’s run-away, the “good gene” and the sensory exploitation) would predict that over time a correlation between the female preferences and male signals will evolve.

Models have shown that all three hypotheses for the evolution of preference are internally valid; i.e. they could work. Testing the external validity, of these hypotheses, that is the probability for generalization, however, has proven troublesome. (Ryan and Rand, 1993)

All three hypotheses are still waiting to be validated.

Neural-cognitive Mechanisms of Sympatric Reproductive Isolation

The reproductive isolation, as the first stage in the process of speciation, offers no selective advantage and natural selection can explain neither why speciation happens nor why and how mating biases change and evolve. An understanding of sympatric reproductive isolation and speciation requires a comprehensive view on the sensory bias as inducers of reproductive isolation and sympatric speciation rather than their influence on evolution of secondary sexual traits that is still dominating studies on “sexual selection”.

A satisfactory knowledge of the mechanisms of reproductive isolation is a crucial element in understanding the process of evolution in metazoans. From the fact that many species, under appropriate conditions, interbreed and produce fertile hybrids, may be inferred that premating isolation, based on sensory-cognitive mechanisms enabling them to avoid mating heterospecific individuals of the opposite sex, is an important mechanism of reproductive isolation and of the initiation of the speciation process in sympatry. This underlines the need for a better understanding of the animal recognition system as a function of the central nervous system.

Evolution of Receiver Biases

Mating choices are expression of sensory biases. As behavioral characters, sensory biases represent behavioral output of the processing of mate sensory input in neural circuits. While we know that these biases are inherited and evolve as species-specific characters, we do not know how they are fashioned in the nervous system. Mating biases or preferences, as neurobiological products of the activity of the nervous system, underlie the mating behavior and behavioral characters (see also Neural Basis of Animal Behavior and Animal Behavior Is not Determined by Genes in chapter 9).

Based on general considerations on receiver’s biases as manifestation of neurobiological processes, one would agree with Enquist and Arak that these biases are a necessary outcome of sensory processing in neural circuits (Enquist and Arak, 1993; Enquist and Arak, 1994). The intrinsic properties of these circuits might determine the curious phenomena of preferences for exaggerated sender’s signaling traits and the origin attractivity, including preference for symmetry.

While the role of selection in evolution of female preferences and corresponding male mating signals has been investigated intensively one has to admit that the role of selection is not the crux of the problem, which lies deeper than selection, in the underlying epigenetic neurobiological mechanisms that generate new or modified mate preferences.

Female preferences are neurally determined mating biases resulting from integration and processing of sensory (visual, auditory, olfactory, tactile, etc.) stimuli communicated by the sender (usually the male) to the receiver (the female). Male stimuli may be perceived by the female as attractive, unattractive or even repelling.. We have no real idea on the brain mechanisms that assign certain traits the attributes of being attractive or, “sexy”. We do not know how these sensory stimuli are manipulated in the brain of these species to produce the attractiveness, mating biases, and preferences on which mating decisions are made.

In general terms, it may be predicted that males with “attractive” characters, characters that induce female’s pleasure, will have greater chance of mating and leaving offspring, consequently will increase their representation over generations and lead to maintenance or evolution of the male “attractive” characters, such as singing, color and color patterning of the body, etc.

Focusing almost entirely on male signals, sexual selection theory, in general, did not deal with the origin of female biases and preferences. Understanding signal evolution under sexual selection is not the whole issue. There is no doubt that female choice can drive in males evolution of traits that are more attractive to females. The difficult and quite controversial issue is why females have evolved receiver properties that make one trait more attractive than another (Ryan and Rand, 1993).

What essentially is the attractiveness of individuals of one sex, which stimulates sexual preference of the other?

From our human experience we know that anything that is attractive to our sight or hearing is beautiful but we have only vague ideas on those subjective, nongenetic rules or criteria on forms, colors, and patterns that produce the sense or feeling of beauty and attractiveness.

While no attempts, to my knowledge at least, have been made to study these rules and criteria, it is not difficult to observe that sexually selected traits are among the most beautiful forms and patterns we see in the world of animals. This is a general indication that some attraction biases in sensory perceptions of the receiver are shared by humans and animals.

A human endeavour for perfection has found its expression in the development of fine arts where exaggeration of real proportions and patterns has always been, and still is, an essential part of artistic trends, which are reminiscent of the exaggeration of the signaling traits in the animal world.

Symmetry in signaling and other traits, is an essential component of the sender’s attractiveness and mate choice. It has been described for females but evidence that males prefer symmetry has also been presented (Hansen et al., 1999). In regard to the origin of the preference for symmetry, Enquist and Arak reason:

One problem faced by animals is the need to recognize objects in different positions and orientations in the visual field. An object viewed from a particular location is focused on the retina as an ‘image’, which is a geometrical transformation of the object itself. An intriguing idea is that the need to generalize many such transformations of the same object may lead to preferences for symmetry and the evolution of symmetrical signals. (Enquist and Arak, 1994)

They experimented with artificial neural networks and found that images of patterns or “signals” consist of colored squares in a grid. Then they pasted images onto an artificial retina and allowed them to coevolve for successive generations producing mutations, which were artificially selected to allow to evolve only signals that elicited the highest output. They observed that the coevolved signals consisted of purer, brighter and less variable colors than random patterns. These signals also showed marked symmetry. This result may be alternatively explained

1. by a higher ability of symmetrical signals to stimulate the network, or

2. the network has evolved toward sensitivity for symmetry.

In experiments for determining which of the above was the real cause of the evolved symmetry they observed that networks trained to recognize signals projected randomly on retina developed strong preferences for symmetries. In contrast, only weak preference for symmetry developed in networks trained to recognize signals projected in different locations of the retina. This suggests that preference for symmetry evolved from an evolutionary pressure for recognizing signals, regardless of their position in the visual field (Enquist and Arak, 1994). Commenting on the results of their experiments on female biases in artificial neural networks, and their relation to the female preference in nature, they pointed out:

It is an interesting thought that all nervous systems built for recognition may share certain general biases which result from hidden properties of the recognition system. Indeed, many elaborate signals that occur in nature are often as impressive to human observers as they appear to be to the intended recipient. Darwin’s idea that a ‘sense of the beautiful’ is an inherent, aesthetic property of animal nervous systems may be not far from the truth. In Darwin’s own words “When we behold a male bird elaborately displaying his graceful plumes or splendid colours… it is impossible to doubt that [the female] admires the beauty of her male partner”. (Enquist and Arak, 1994)

According to an hypothesis, preferences for symmetry evolve as a by-product of cognitive processes where the sum of fluctuating asymmetries is averaged to zero asymmetry, which is symmetry (Enquist and Arak, 1994). The hypothesis seems to have found empirical support in a few experiments. Untrained starlings exhibit no preference for symmetry (Swaddle, 1999) but when these birds were trained to detect a set of asymmetric images they develop preference for asymmetry. Such examples suggest that preference for symmetry may develop on the basis of learning mechanisms, without any correlation between this trait and the fitness (Swaddle et al., 2004). So, e.g., no relation with male fitness or quality has been found in experiments on symmetry and attractiveness of human face (Rhodes et al., 1999).

A generalization probably could be made for birds, where many studies have shown that they exhibit a clear preference for symmetry (Swaddle and Cuthill, 1994; Enquist and Arak, 1994) and for what humanly are considered to be ornaments in their morphological characters. It has been hypothesized, and limited evidence is presented, that both these preferences are used as indicators of male quality. Since developmental asymmetry is considered to be a negative indicator of the fitness and since in experiments an inverse correlation has been found to exist between the size of the ornament and the asymmetry, it has been concluded that female preference for exaggerated mating signals may be an indicator of male good quality (Møller, 1992). This sensory bias for symmetry “may account for the observed convergence on symmetrical forms in nature and decorative art” (Enquist and Arak, 2002).

Idea has been expressed that mate preference for symmetry is evolutionarily selected and the prevailing explanation of the evolution of preferences for symmetry in bilateral traits is that it results from selection against fluctuating asymmetry, which in turn is a consequence of developmental disturbances and developmental stress. However, experiments with artificial neural networks suggest that these preferences may evolve in the absence of any correlation between the symmetry of male signaling traits and the male fitness, and that the female preference for symmetry evolved as a result of selection for mate recognition (Johnstone, 1994).

Attempts have been made to relate general preference for symmetry with the concept of Gestalt (German for shape) by relating the preference for symmetry to the Gestalt law of simplicity. According to the Gestalt concept, percepts precede sensory experience, with the latter acting as a stimulus for activating preexisting percepts. These percepts may be related with the specificity of organization of neural structures in the brain. Some evidence in support of this explanation comes from studies on infants, which seem to validate Helmholtz’s prediction that it is the organization of the inner ear that makes humans find pleasure in harmony and dislike discordance in music (Ryan, 1998).

Let’s return, now, to the general problem of the attractiveness. In a model by Phelps et al. (2006) animals convert the input of visual, auditory, olfactory signals (morphological and courtship signals) from possible mates into neural equivalents suitable for comparison before deciding to choose the most attractive mate that exceeds some minimal criterion. In their model, the mate choice and species recognition are fundamentally similar, in the meaning that both rely on the same mechanisms (Phelps et al., 2006). The authors wonder

whether shifts in preference acuity or threshold reflect changes in the number or nature of neurons that process sensory information or assign it affective value. (Phelps et al., 2006)

The answer to the question is “Yes”, based not only on theoretical consideration on the behavior as a product of the activity of neural circuits, but also on empirical evidence. So, for instance, the marine plainfin midshipman fish, Porichthys notatus, has two male morphs displaying differences in body size, gonad/body weight index, reproductive tactics, and vocal motor traits. The type I males reach reproductive maturity at smaller body size than the type II males and females. Immunocytochemical examination revealed that the size and number of GnRH neurons in the hypothalamic POA (preoptic area) was 50-100% greater in adults than in juveniles and changes in the POA phenotype are correlated with the development of type I and type II males in P. notatus. The temporal pattern of such changes in POA may represent the proximate mechanism for the development of alternative male reproductive morphs (Grober et al., 1994).

Phelps and Ryan trained neural networks to accept spectrograms, visual representations of ‘conspecific’ whining túngara frog calls and reject visual representations of a noise (the higher the intensity of the sound the darker the image) (figure 20.1). The ‘conspecific’ signal was changed over time approximating the reconstructed evolution of túngara frog call from its beginning to the present form. First neural networks were trained to recognize the primitive ancestral call then successively to learn two intermediate calls and finally the present-day túngara call. In the control group, the selected neural networks were also trained to recognize calls but not in the reconstructed evolutionary sequence.

Investigators observed differences in the patterns of biases and in preferences for novel stimuli between the two groups. The group of the neural networks trained to recognize the first call, which evolved similarly to the reconstructed evolution of túngara frogs, were best at evolving to recognize subsequent calls and reproducing the biases of real túngara frogs.

.......................................................................

These results suggest that historical processes “shape the design of communication systems” and determine receiver biases. Based on analogies between the behavior of neural networks and female responses, investigators believe that circuits are represented by “reciprocal connections of the torus semicircularis and auditory thalamus”. (Phelps and Ryan, 1998; Phelps and Ryan, 2000).

In most of cases the “choosy” sex is the female, and male signaling traits are selected under the influence of female preferences for certain male traits. However, “exceptions to the rule” are known. There are cases in insect, fish, amphibian and bird species, when males are also attracted to, or show preference for, females displaying specific traits. For example, both males and females of the rock sparrow, Petronia petronia have a yellow breast patch. In females, the size of the patch is correlated with the fecundity (the annual number of broods), and males prefer mating females with larger patches, implying that the female trait is sexually selected. Experimental reduction of breast patches in female rock sparrows leads to declining sexual interest of males for these females (Griggio et al., 2005).

Mating preferences are expressed not only in choice of the mating partner but, in birds, they also appear in the form of a clear tendency to invest more in eggs when mated with males displaying more attractive visual (Gil et al., 1999) or acoustic (Kolm, 2001; Gil et al., 2004) signals.

Controversy exists over the possibility of coevolution of male signaling traits and female preferences. In a comparison between two sister species of the swordtail fish, Xiphophorus multilineatus and X. nigrensis, Morris and Ryan observed that females of X. nigrensis responded similarly to males with vertical bars of X. multilineatus and to conspecific males without bars. In contrast, females of X. nigrensis found more attractive males with bars of the sister species than those without bars of their own species (Morris and Ryan, 1996), what speaks against coevolution of male signaling and female preferences, that male traits and female preferences “do not coevolve via genetic correlations” (Ryan, 1998).

As for males, evolution of their responsiveness to bars seems to have been congruent: males with bars of the X. multilineatus are responsive to bars whereas bar-less nigrensis males do not respond to bars. These observations suggest that the attraction of the signal receiver to vertical bars may have been a mate bias in females of their common ancestor, that has been conserved even after the loss of vertical dark bars in X. nigrensis males. The presence of this bias in X. nigrensis females represents an opportunity for “sensory exploitation”. If at some point in the species’ phylogeny the males would evolve vertical bars, females would be in possession of the corresponding sensory bias. The reverse phenomenon of incongruence of female preferences for male signals is also observed in other fish species (Morris and Ryan, 1996).

However, examples of correlation and coevolution of male traits and female preferences have also been presented. For example, the stalk-eyed fly, Cyrtodiopsis dalmanni, exhibits strong sexual eye dimorphism. Artificial selection for long and short eyespan in males for 13 generations, led to a shift of the female preference for long eye-span males in females of the long eye-span line and unselected line, while females of the short eye-span line continued to prefer short eye-span males (Wilkinson and Reillo, 1994). Experiments with artificial neural networks also suggest that female biases and male signals may coevolve.

Enquist and Arak (1993) have proposed a hypothesis on the evolution of the preference for exaggerated male secondary sexual traits, based on experiments with artificial neural networks. They found out that artificial neural networks may be trained to recognize certain images. Although no recognition system can be trained to identify all the possible variations (they are too many to specifically be identified) of those images they acquire the ability to “generalize”, that is to classify new variants of images into groups of images they have recognized before and to respond to them. When the network trained to recognize a bird image was presented with images of exaggerated long-tailed (and long-winged) conspecific males, gave a weaker response to shorttailed, heterospecific males and did not respond at all to tailless bird images. This implies that biases in neural networks arise unavoidably. Investigators also attempt to explain the conclusion of a number of studies that female biases precede the evolution of specific male signals:

Our models suggest that biases in response to signals inevitably exist as a fundamental consequence of the context in which recognition occurs. Because the number of forms that a signal can take is almost infinite, the recognition mechanism is always likely to show a greater response to some variants of the male signal not yet in existence. Such unexpressed, or ‘hidden’, female preferences will change continuously as a side effect of selection for improved recognition, by genetic drift (many solutions exist for a given recognition problem) and because of correlated effects of selection acting on male signals. (Enquist and Arak, 1993)

It is assumed that sexual preferences of receivers for specific sexual traits of the sender arise and evolve via selection but the substantiating evidence is scarce and equivocal. Moreover, evidence that evolution of receiver sexual preferences is not related to selection has also been accumulated, and studies with neural networks have shown that these biases in the animal’s brain might arise as an emergent property of neural processing of external signals in the absence of any training selective context. In many respects, artificial neural networks of neuronoid units have often been found to behave like real nervous systems. For example, in experiments with túngara frog calls these neural networks have been able to respond to the call and, based on such experiments, it is suggested that the call may have evolved as a by-product of a sensory system for species recognition. In the context of the role of mating calls in the reproductive isolation, this implies that higher speciation rates may be related to the rapid evolution of mating signals (Phelps and Ryan, 1998).

While many authors believe that female biases may result from natural selection, others like West Eberhard (1984), Basolo (1990) and Ryan (Ryan, 1998) believe that they evolve not for their own sake, but as a byproduct of selection for other characters correlated with sensory biases. However, numerous examples of female biases for heterospecific mate signals and also cases of the loss of existing preferences suggest that receiver biases may evolve independently of the natural or sexual selection, for reasons that may be related with the properties of neural networks determining the biases.

Most investigators believe that mate preference precedes the evolution of mating signals and it is the latter that evolves to match the existing mate preference. Besides the example of evolution of preference for swordtail in Priapella olmecae, mentioned earlier, in favor of this hypothesis testify a number of other examples on the occurrence of female preferences in species with males lacking the preferred trait. So, e.g., females of common grackles (Quiscalus quiscula) have preference for repertoires of four song types although the conspecific males sing only one song type. It is believed that female preference antedates the evolution of male song repertoires (Searcy, 1992). No role for natural selection or sexual selection can explain the evolution of such examples of female biases.

Female biases for mating signals may change over evolutionary time, and they may be related to the diversity of male mating signaling. An example of evolution of female biases comes from a study on females of two sister genera, Xiphophorus and Priapella. Xiphophorus males with swordtail attract females of both Xiphophorus helleri and Priapella olmecae, even though swords in Xiphophorus evolved after divergence of the two genera. This suggests that Priapella females may have evolved the preference for the swordtail even though its conspecific males have not evolved swordtail.

Receiver biases can even change during the lifetime of a single individual as it is the case with females of the satin bowerbirds (Ptilonorhynchus violaceus), which display different mating preferences during different stages of life (Coleman et al., 2004).

Although female preferences are innate traits, they can be modified by experience. Exposure of female fish of the green swordtail to predation, for instance, makes them to switch mate preference to swordless fish from the original state of long sword preference (Johnson and Basolo, 2003). Early life experiences in guppies modify female mate preference to orange male coloration. It is noteworthy that even cases of reversal of the lost ancestral mate preferences are described. Visual exposure of the female guppy, Poecilia reticulata, to its cichlid predator Cichlasoma biocellatum induces her to revert to the initial preference for brighter males or become unreceptive (Gong and Gibson, 1996).

Mating preferences can be experimentally changed by administration of hormones. For example, administration of HCG (human chorion gonadotropin), a ligand of the pituitary LH (luteinizing hormone) in neotropical túngara frogs increases the female receptivity (measured by the number of responses and the time lapse between call and response) and permissiveness (the likelihood of responding to the less attractive calls) (Lynch et al., 2006).

From a neoDarwinian standpoint, i.e. from the view that mate preferences are determined by genes, such sudden changes are inexplicable. But such changes become quite understandable from the epigenetic view that mate preferences are function of neural circuits, which display a relative plasticity at both the developmental and evolutionary levels.

Evolution of Sender’s Signaling

In most cases females are receivers of male signals, i.e. they are the choosy sex, whereas males are senders of signals and objects of female mating preferences. As signals serve sexual phenotypic traits, which may be acoustic, olfactory, behavioral and, above all, visual signals related to the shape, size, colors, and patterning of these traits.

Evolution of male signaling may be a direct (sensory exploitation hypothesis) or indirect (“good gene” or runaway hypotheses) result of selection on male sexual phenotypic traits, which show a high level of plasticity. In some described cases, males change their mating signals in response to environmental stress. For example, male individuals of the firefly Photinus greeni, on detecting the presence of the predator Photuris, decrease their flash pattern rate and only males that maintain a relatively high flash pattern rate succeed in eliciting female attraction and mating. With no differences in male morphological traits, Photuris females prefer higher flash pattern rates (Demary et al., 2006), whereas the male lantern size plays no role in mate choice (Cratsley and Lewis, 2003).

Mating signals can also evolve independently of the female biases. An example of evolution of mating signals decoupled from female biases is observed in experiments with two sibling species of the northern swordtail fish of the genus Xiphophorus, Xiphophorus multilineatus and X. nigrensis (figure 20.2). These species have a patern of dark vertical bars that functions as signal for attracting females and for deterring rival males Vertical bars are present in the species X. multilineatus but absent in X. nigrensis. Nevertheless, females of the latter respond to this signal and show a preference for it, although it is absent in males of their own species. Given that both species come from a common ancestor, the presence/absence of bars in them may have resulted from the loss of bars in X. nigrensis or gain of bars in X. multilineatus.

Females of the satin bowerbird, Ptilonorhynchus violaceus, at different ages show different preferences for male displays.......

......................................................................

Mate Recognition System

.........................................................................

Evolution of the Mate Recognition System

Neurocognitive Mechanisms of Reproductive Isolation

Visual-cognitive Mechanism of Reproductive Isolation

Neural Reception of Visual Signaling

Visual perception of signaling traits is one of the most important means for discriminating between conspecific and heterospecific individuals as well as for mate choice. Visual stimuli are converted into electrical signals in the retina and from there transmitted for processing to lateral geniculate nucleus and still higher to the primary and secondary visual cortex. In the process, the light stimuli are filtered and further transformed in little known ways, in order to produce the visual perception, which is a neurobiological interpretation of the light information. Different animal species are receptive to different wavelengths of light reflected by the Umwelt. Hence, the same object may be perceived in different ways by different animals.

Fish use sensory recognition and sensory preferences for identifying and choosing conspecific individuals of the opposite sex as a mechanism of reproductive isolation under sympatric conditions. The visual system of fish is adapted to the environment. In cichlid fish the system has evolved to adapt to the spectral transmission in the aquatic environment in general but is also adapted to the degree of the transparency of the water, in order to better identify mating color signaling and body patterning.

There is evidence that visual cues, mainly male coloration, are used for mate recognition in cichlid fish (Kornfield and Smith, 2003). Four closely related species of the cichlid fish Pseudotropheus zebra group, are morphologically indistinguishable and show no differences in their courtship behavior, suggesting that the color pattern may be the only distinctive component of their mate recognition system (Couldridge and Alexander, 2002).

Body color and patterning are key visual cues used by female cichlid fish of the great East African lakes for discriminating between conspecific males and males of closely related species that morphologically are very similar. Females also use those cues for mate choice (Carleton et al., 2005).

Under natural conditions, adult color pattern is an important visual cue for mate recognition in mbuna fish of Lake Malawi as well as in rock-dwelling cichlids in Lake Victoria, East Africa. In experiments that eliminated the differences in male body color by using monochromatic light, females preferred the larger and more active males. These experiments suggest that, despite the role of body size, the male body color is the most important of the visual cues these fish species use for mate recognition and is believed to have played a leading role in the extraordinary rapid evolution of these East African lake species (Danley and Kocher, 2001).

Males of the cichlid fish Pseudotropheus callainos court only conspecific females. It was assumed that males of this species during the courtship emit sounds, which are detectable by females but the use of auditory cues in mate recognition in this species seems to be unlikely. The fact that males of this species cannot discriminate between conspecific and heterospecific females that are similar in body color indicates that no acoustic or chemical cues are used for mate recognition by East African cichlid fish (Knight and Turner, 1999).

Males of two closely related fish species of the Haplochromis nyererei complex have distinct body colors. Males of one species are blue and the other’s are red. Each of them mates with conspecific females only. When color distinction between species was experimentally masked by monochromatic light they mated non-assortatively indicating that body coloration is the only cue used for mate recognition. However, females of both species mated more frequently with blue males, which are larger and have higher display rates, suggesting that in the absence of the color cue, females use body size and display rates as mating criteria (Seehausen and Alphen, 1998). Two morphospecies of the Carribean fish of the genus Hypoplectrus (H. unicolor and H. gemma) show only minimal genetic difference but display distinct color patterns. Living in sympatry, they mate exclusively like with like (Barreto, F. S. and McCartney, 2008).

The dragon lizard, Ctenophorus decresii, consists of a monophyletic group of sibling species (C. decresii, C. fionni, C. rufescens, C. tjantjalka, and C. vadnappa). These species use complex displays for social and sexual communication behavior. While these displays involve ventral side of the body, throat, chest, etc. which are conspicuously colored, the colors and patterns of the dorsal part of their body, head included, have evolved to match the habitat background for reducing the avian predation risk (Stuart-Fox et al., 2004). This is considered to be an example of combined effects of the natural and sexual selection acting respectively on the dorsal and ventral parts of the body for body color and patterning.

In birds especially, visual cues are the most important element of mate choice decisions and it is believed to have played an important role in evolution of the class of Aves.

………………………………………….

Figure 20.7. Illustration of the scheme for NO interaction with mitochondria in the mechanism for oxygen gating for on/off switching of firefly flashing (From Aprille et al., 2004).

Behavioral studies on Photinus ignites have shown that females of this species display preference for longer and brighter male flash signals, and these traits are positively correlated with the size of spermatophore at the beginning of the mating season. However, the preference declines sharply when the duration of the flash exceeds the range of species-specific norm (Cratsley and Lewis, 2003), thus preventing the possibility of mating with heterospecific males.

It is interesting to also point out that not only do bioluminescent insects of the Lampyridae family have species-specific models of encoded flash patterning in the central nervous system but they can imitate heterospecific flashing patterns for luring individuals of prey species of the Photinus group. This is the case with the predator female fireflies, Photuris versicolor, which are able to modulate their flash signal to mimic female flash responses of their prey, the firefly Photinus tanytoxus (Trimmer et al. 2001).

From an evolutionary point of view, the central control of the flashing patterns in these insects represents a potential mechanism for reproductive isolation between sympatric populations found under different evolutionary pressures and for ensuing evolutionary and speciation processes. This is what seems to have occurred in the case of the neurally-induced reproductive isolation of three morphs of fish of Brienomyrus species flock of the Gabon river, west-central Africa, which are believed to be at the stage of incipient speciation and a similar neural mechanism of reproductive isolation has been operational in the cases of reproductive isolation of the sibling species of the Rhagoletis complex (to briefly be discussed later in this chapter).

Given that in these bioluminescent insects

1. Changes in the signals that lead to reproductive isolation are products of the brain activities, and

2. The changes in the bioluminescent flashing patterns involve no changes in genes, it is logical to assume that evolutionary changes in bioluminescent flashing patterns leading to reproductive isolation and incipient speciation result from epigenetic changes in the function of neural circuits determining generation of these signals.

In view of the above-mentioned fact that insects are capable of mimicking flash patterns of their preys, it is not difficult to imagine how the CNS and the central generator of the flash pattern, in a whole population might shift to a new pattern under specific changes and new evolutionary pressures in the environment. In response to such changes, in principle at least, bioluminescent insects can adaptively change the structure and synaptic morphology of the relevant neural circuit, and consequently the properties of the circuit, which determine the flashing patterns. Needless to say, all these changes are epigenetic by the origin and nature and require no changes in genes.

Olfactory-cognitive Mechanism of Reproductive Isolation

Neural Reception and Processing of Olfactory Signals

Olfactory system provides animals with a key sense not only for detecting predators, preys and food but also for discriminating between conspecific and heterospecific individuals in the environment. In many aquatic and terrestrial species, mate choice is determined, partially or totally, by olfactory cues.

Olfactory signals may also have been important factors in the reproductive isolation, speciation, and evolution of many species such as Drosophila spp. and probably in the rapid formation of hundreds of species of cichlid fish in East African lakes. A good idea about the importance of the system in the life and evolution of animals may be obtained from the fact that as much as 4% of genes in higher eukaryotes code for olfactory proteins (Firestein, 2001).

Olfactory system emerged early in the evolution of metazoans as a part of the nervous system and in its main features are conserved across metazoans, from insects to mammals. Axons of the ORN (olfactory receptor neurons) that express the same receptor molecules, in Drosophila converge to the same glomeruli in the antennal lobe, thus forming a spatial odour map in the fly brain. Odorants bind multiple olfactory receptors; hence it is possible that the representation of olfactory stimuli is combinatorial (Jefferis et al., 2004).

In Drosophila, ORN axons connect with dendrites of the second order projection neurons (PNs) in the antennal lobe in a strictly determined pattern (figure 20.8). Projection neurons, in turn, send axons to higher brain centers. Contrary to what was generally believed, and to what is described for other insects and vertebrates, it is not the axons of the ORNs that select the dendrites or determine the synaptic morphology there; PN (projection neurons) specify their dendrites and create a prototype of the adult glomerular map before their presynaptic partners, axons of the ORNs, reach these neurons. It may be possible, however, that axons of the ORNs help to refine the specificity of the preformed connections (Jefferis et al., 2004).

Figure 20.8. Organization of the mature Drosophila antennal lobe (AL) with vertebrate counterparts in parentheses. Each differently shaded circle represents olfactory receptor neurons (ORNs) expressing a particular seven-transmembrane span receptor or their post-synaptic projection neuron (PN) partners (From Jefferis et al., 2004).

Prespecification of dendrites in PNs determines the hardwiring of the olfactory system and the resulting behavioral responses of flies to odorants (Jefferis et al., 2001). How the ORN axons find their corresponding dendrites among the numerous PN dendrites is unknown. According to one hypothesis,

ORN axons and PN dendrites have substantial autonomous patterning ability…the two proto-maps interact during development to generate the final mature glomerular organization. (Jefferis et al., 2004)

Hypotheses for explaining this experience-independent prepatterning of synaptic morphology are not lacking but one might still wonder where the information for patterning synaptic morphology in the antennal lobe comes from.

In a study of the anatomy of the antennal lobe of more than 30 endemic Hawaiian species of the family Drosophilidae it was found that 2 (out of 51 identifiable) glomeruli were enlarged in the antennal lobe (AL) of males only. This trait evolved independently in 37 Drosophilidae species of two genera (Drosophila and Scaptomyza, derived from the first Drosophila species that migrated to Hawaii between 1 and 2 million years ago). This sexual dimorphism of Drosophila brains is estimated to have arisen sometime between 0.4 and 1.9 million years ago (Kondoh et al., 2003). The macroglomeruli DA1 and VA1, differ from all other glomeruli that are innervated from projection neurons of a single group: DA1 is innervated by the lateral and ventral group of neurons, while VA1 is innervated from the dorsal and ventral groups. It is hypothesized that projection neurons of the ventral group innervating both DA1 and VA1, which have similar axon patterns in both glomeruli, are responsible for the sexual dimorphism of antennal lobe in species of the Drosophilidae family (Kondoh et al., 2003).

The idea that the brain dimorphism is related to male specific genes (Kondoh et al., 2003) is rejected by earlier experiments of Schneiderman et al. (1986) and Rossler et al., (1999), which have shown that, in Manduca sexta, females lacking such genes also develop macroglomeruli in the antennal lobe when they receive grafts of male antennae (Schneiderman, et al., 1986; Rossler et al., 1999) and, to the contrary, males that normally develop the macroglomerular complex do not develop it when are innervated by olfactory receptor neurons from a grafted female antenna. The pheromone-related sexual behavior of such gynandromorphic males and females depends on graft origin rather than on the host itself (Schneiderman et al., 1986; Rossler et al., 1999). Based on the results of such experiments it is concluded that not differences in any sex-specific genes but male and female ORC axons are involved in determining the position, anatomical features, and innervation of sexually dimorphic glomeruli. (Schneiderman et al., 1986)

The olfactory system is an important component of the mate recognition system with crucial functions for the mate choice. It detects and recognizes species-specific olfactory signals, above all pheromones that are used for coordinating the reproductive behavior and activity of mating partners in invertebrates and vertebrates.

Secretion of pheromones in insects is neurally regulated by cerebral neuropeptides, pheromonotropic hormones, of which the most extensively studied is PBAN (pheromone biosynthesis activating neuropeptide) family of neuropeptides secreted by neurons in the insect brain (Altstein et al., 1993). The neurohormone is also involved in determining the body color in the moth Spodoptera littoralis: the insect larvae double the biosynthesis and secretion of PBAN in hemolymph when reared in light and/or, under stress conditions of crowding, induce dark coloration in larvae. By contrast, the larvae decrease secretion of PBAN and have lighter color when reared singly or in darker environment (Altstein et al., 2002).

Another pheromonotropic hormone, isolated from the moth Helicoverpa zea, genus Heliothis, is Hez-PMP (pheromonotropic melanizing peptide), which induces pheromone release and melanization of the insect in a dose-dependent mode. Besides ganglia, the neuropeptide is also released by the esophageal nerve (Raina et al., 2003).

The biosynthesis and secretion of the sex pheromone in the hemolymph of the noctuid insects, Spodoptera littoralis and Mamestra brassicae, also may involve PBAN or PBAN-like neuropeptides. The control and regulation of the synthesis and secretion of these neuropeptides, is determined by a “neural input from the ventral nerve cord” on the pheromone gland (Iglesias et al., 1998).

In Drosophila melanogaster, olfactory signals (pheromones) released by males are received by sensory nerves and transmitted to the brain (antennal lobe) for processing in the olfactory neural circuit. The olfactory circuit in Drosophila not only is responsible for identification of odors but it is also necessary for the courtship behavior of the fly that is determined by another circuit, which, in male flies, comprises a number of neurons of a larger group of neurons that have been identified to express two specific genes (Stockinger et al., 2005). This is one of the best known neural circuits in Drosophila (figure 20.9).

Figure 20.9. Wiring Diagram: Adult versus Larval Olfactory System of D. melanogaster. Adult and larval olfactory pathways share the same general design. However, there are twice as many primary “olfactory identities” (ORN types or AL glomeruli are differently shaded) in the adult. Moreover, in the adult AL, the different types of ORNs (open circles) and PNs (filled circles) that innervate a particular AL glomerulus occur in multiple copies, whereas larval ORN and PN types are unique, resulting in an almost complete lack of cellular redundancy. Thus, the adult olfactory pathway is characterized by converging and diverging connectivity in the AL, whereas the larval pathway is organized as straightforward channels in which ORNs, LAL glomeruli, PNs, and calycal glomeruli are related essentially in a 1:1:1:1 fashion (ratios indicated in black refer to the features shown in the preceding line). The larval MB calyx retains a strong spatial organization that is not obvious in the adult (note: adult MBs include, apart from MB γ neurons, additional classes of intrinsic neurons).

Abbreviations: AL, antennal lobe; LAL, larval antennal lobe; LNs, local modulatory neurons; MB, mushroom body; OR, odor receptor; PNs, projection neurons (From Ramaekers et al., 2005).

The fly has ~1300 ORNs (olfactory receptor neurons) in the antennae and maxillary palps. Each of these neurons expresses only one of the 43 types of OR (olfactory receptor molecules). Each ORN projects an axon to one of the specific glomeruli in AL (antennal lobe), the equivalent of the vertebrate olfactory bulb. From the antennal lobe, 150-200 PN (projection neurons) transmit olfactory signals to the MB (mushroom body) calyx and to the lateral horn. The circuit is cytologically completed and operational only after metamorphosis (Marin et al., 2005). When a group of antennal ORNs bind a specific molecule, it activates a corresponding group of projection neurons in the AL glomeruli, which is then inhibited by inhibitor circuits that form combinatorially by local interneurons (Ng et al., 2002).

Surprisingly, the basic organization of the olfactory system in metazoans, from insects to humans, is well conserved. Olfactory receptor neurons (~1300 in Drosophila but millions in mammals) express a number of olfactory receptors but each of them expresses only one receptor molecule type. Axons of ORNs that express a particular receptor converge to only one or two of the glomeruli in the AL (antennal lobe). The number of glomeruli in Drosophila is 43, but it numbers into the thousands in mammals. Via projection neurons, glomeruli send olfactory information to the mushroom body calyx and the lateral horn (Ramaekers et al. 2005). The structure of the olfactory system in insects is qualitatively determined in the larval stages of development and its further development into the adult system basically consists almost entirely in quantitative changes.

Key elements in the olfactory pathway are projection neurons (PNs) which connect antennal lobes (AL) with the lateral protocerebrum (LPR) not only directly but indirectly as well. In order for a male insect to initiate and perform male courtship behavior, it is necessary to receive the specific olfactory and chemosensory cues or visual cues of a conspecific virgin female. The initiation of the male courtship behavior is not related to the presence or activity of the mushroom body (that behavior is also initiated in the mushroom body-ablated males) but to a cluster of some excitatory and some inhibitory neurons in the lateral protocerebrum. These neurons seem to integrate the sensory information that triggers the male courtship behavior (Broughton, 2004). Thus, the chemosensory and visual information for activating the male courtship behavior in Drosophila reaches the lateral protocerebrum in two ways: the indirect pathway via the mushroom body calyx (responsible for the experience-dependent olfactory processing) and a direct pathway, which is responsible for experience-independent olfactory processing. Ablation of the mushroom body calyx, i.e. the block of the indirect pathway does not impair the experience-independent functions of odor detection and male courtship behavior but affects the experience-dependent functions of odor- and courtship-conditioning. Unlike the ablation of mushroom body calyx, the block of synaptic transmission by TeTxLC (tetanus toxin light chain) impairs both odor detection and courtship behavior (figure 20.10).

Olfactorily Determined Reproductive Isolation in Sympatry

Chemosensory signals and the receptive olfactory system in insects play an extensively demonstrated role not only for discriminating between the conspecific and heterospecific individuals and for assessing mate quality but also for the initiating the process of reproductive isolation between sympatric populations.

Experiments on two recently diverged sympatric cichlid fish species (Pseudotropheus emmiltos and P. fainzilberi) from Lake Malawi have shown that P. emmiltos females use olfactory signals and preferences, to a much larger extent than the male body color, for discriminating their conspecific mating partners from P. fainzilberi males.

Figure 20.10. Schematic representation of different PN target regions and their role in olfactory behavior. In MB/AL ablated flies, a cluster of PNs lateral to the AL and collaterals of the persisting PNs in the MB calyx are missing, whereas the direct PN pathway to the LPR remains intact. Such males show normal experience-independent olfactory responses (“odor detection”) and normal experience-independent courtship behavior (“courtship”). However, experience-dependent modulations of olfactory responses (“odor conditioning”) and of courtship behavior (“courtship conditioning”) are impaired. When blocking synaptic transmission in GH146/TNT, information processing via the direct and indirect pathway is largely impaired. As a consequence, these flies are impaired in odor detection and courtship. These results suggest that the direct pathway from the AL to the LPR is sufficient and necessary for experience-independent odor detection and courtship, whereas the indirect pathway is necessary for the experience-dependent modulation of these behaviors.

Abbreviations: AL, antennal lobe; act, antennocerebral tract; cx, mushroom body calyx; GH146/TNT, tetanus toxin light chain; gl , antennal lobe glomeruli; MB, mushroom body; LPR, lateral protocerebrum (From Heimbeck et al. 2001).

Based on the results of these experiments and other relevant evidence on the role of olfactory and auditory cues in the Mexican pupfish (Strecker and Kodric-Brown, 1999) and Drosophila (Ortiz-Barrientos et al. 2004), it has been concluded that changes in olfactory signals and olfactory preferences have played a greater role that is generally believed in the rapid speciation of these fish (Plenderleith et al., 2005) and Drosophila.

It is curious to know that females of the poeciliid swordtail species Xiphophorus cortezi not only discriminate between the conspecific and heterospecific male olfactory cues but they show a stronger response to male olfactory cues of the closely related Xiphophorus nigrensis than to the more distant species Xiphophorus montezuma (McLennan et al., 1997), suggesting that olfactory cues have diverged in the course of the evolution of these species.

Male plethodontid salamanders during the courtship, before releasing sperm, deliver to the females pheromones secreted by the mental gland. These pheromones (sodefrin and sodefrin-like peptides) increase female sexual receptivity. Most of salamander species use a 50 to 100 million year old ancestral “scratching” behavior for delivering pheromones to females. They scratch female’s back with the protruding premaxillary teeth (PPT) and then rub the mental gland on the scratched region. Despite the long evolutionary stasis of the courtship behavior and the teeth used for transferring the pheromone to females, in most of Plethodon spp., about 19 million years ago, species of the eastern Plethodon clade entered a period of divergent evolution (figure 20.11). While Plethodon cinereus generally retained the ancestral mode of pheromone delivery, males of P. welleri and P. wehrlei lost the protruding premaxillary teeth and show a posterior dislocation of the mental gland (MG), which is characteristic of the olfactory delivery; these males deliver pheromones by application of the mental gland to the nares of the female or by head-rubbing behavior. Males of another plethodontic salamander species, P. glutinosus, also have lost the premaxillary teeth but they deliver pheromones by slapping their posteriorly displaced mental glands on the nares of the female salamanders (Palmer et al., 2007).

Male salamanders of the Plethodontidae family secrete a protein pheromone that contains the Plethodontid receptivity factor(PRF) (Rollmann et al., 1999), which increases female receptivity during courtship interactions. It is a signaling cytokine protein expressed only in the mental gland, which has evolved by gene duplication in almost all Plethodontides. The structure of this protein has diverged during the phylogeny of the group (Palmer et al., 2005).

Chemically, PRF shows sequence homology with neurotropins (growth factors involved in survival and differentiation of neurons). The pheromone also contains another protein, known as PMF (plethodon modulatory factor). These two proteins represent 85% of the total protein content of the pheromone.

Figure 20.11. Cladogram showing the relationships of various clades of plethodontid salamanders and the evolution of characters involved in courtship pheromone delivery. Approximate divergence times are shown at bottom. Small rectangular boxes show the point of origin (solid) or loss (open) of various characters.

Abbreviations: SPF, sodefrin-like precursor factor; MG, mental gland; PPT, protruding premaxillary teeth; SD, scratching delivery of courtship pheromones; PRF, plethodontid receptivity factor; OD, olfactory delivery of courtship pheromones (From Palmer et al., 2007).

Studies on Plethodon shermani have shown that male salamanders release the pheromone in response to the sensory (tactile) stimulation during contact of the male mental gland with female nares, i.e. via a neural sensory pathway. While PRF has a stimulating effect on female receptivity (expressed in the form of shorter courtship and mating time), PMF (plethodon modulating factor) has the opposite effect of prolonging the courtship and mating time. Each of the two main components of the pheromone, PRF and PMF, bind to receptors in separate vomeronasal neurons, which transmit their separate information for processing in the brain, thus regulating mating behavior (Wirsig-Wiechmann et al., 2006).

Another essential pheromone component in plethodontides is the protein SPF (sodefrin precursor-like factor), which has been recruited for a pheromonal function in this group about 50-100 Mya, that is much earlier than PRF, which is believed to have been recruited ~27 Mya (Palmer et al., 2007). It is not understood why SPF had to be replaced by PRF in plethodontids, but the fast evolution of the molecular pheromonal factors (and the replacement of SPF by PRF) in this group is in contrast with the evolutionary stasis of the morphological (loss of premaxillary teeth) and behavioral elements (delivery behavior) of the pheromonal system in plethodontides. This supports the notion that evolution of genes in metazoans is decoupled from the evolution of morphology and behavior (Palmer et al., 2007).

That the evolution of the pheromone system is not related to changes in pheromonal genes is also corroborated by the fact that all plethodontid salamenders have the genes for SPF and PRF, regardless of whether they express them or not. The difference in this respect consists in the fact that some 27 Mya, salamanders switched to expression in the mental gland of the PRF instead of the SPF that was expressed by their ancestors. This evolutionary change came as a result of an epigenetic switch. The evolutionary transition from the use of SPF to the use of PRF in salamanders is correlated with a change in the properties of the olfactory circuits, which evolved for receiving and perceiving the PRF instead of SPF as a pheromone.

In principle, we know that such changes in perception of olfactory signals in the CNS are not related to changes in genes, but evolve intrinsically as a result of changes in neural circuits and in their synaptic morphology. Furthermore, the effect of both these pheromonal substances for stimulating the sexual activity in females is not a direct one. Each of them does not act on the reproductive system directly. They bind to specific receptor molecules in the olfactory system and are processed in respective neural circuits for inducing the reproductive behavior and reproductive physiology in female salamanders.

Mammals also use olfactory cues for mate choice. For example, both humans and mice prefer odor of individuals that possess dissimilar antigens [cell surface glycoproteins, which differ in the peptide-binding region (PBR) of their molecules] of the MHC (major histocompatibility complex), a group of extremely polymorphic genes involved in the immunological mechanism of recognition of “self from non-self”. Mice can recognize individuals that differ only in their MHC system, but otherwise are genetically identical to them. Recognition of their own individual MHC phenotype in other individuals allows them to avoid inbreeding (Penn and Potts, 1998; Penn and Potts, 1999). The olfactory sense of discrimination between similar and dissimilar MHC peptides (antigens) is so strong that mice can discriminate between individuals that differ in only one antigenic determinant.

Conflicting ideas and experimental results are obtained in attempts to experimentally determine whether the ability to discriminate MHC-similar from MHC-dissimilar individuals (“functional genome analysis by the nose”) is inherited or is learned (imprinted) after birth. It is possible that olfactory circuits for identifying odors are established before birth but may be fine tuned postnatally. In some cases, it has been possible to experimentally prove that olfactory discrimination is imprinted early during postnatal life (Penn and Potts, 1998).

Female three-spined sticklebacks (Gasterosteus acculeatus) are able to achieve greater MHC diversity in the offspring by receiving MHC-related olfactory signals for mate choice. Identification of these signals in the olfactory neural circuit allows them to choose mates that differ from them in MHC antigens (Boehm and Zufall, 2006). The MHC system, however, may not be a universal system of mate recognition in vertebrates (Penn and Potts, 1999).

As pointed out earlier, the olfactory system is remarkably conserved not only across vertebrates. A study on 12 Drosophila species, which evolved and diverged for ~63 million years, which is a little less than, but comparable to, the time of divergence of eutherian mammals (~100 million years), has shown that the repertoire of OR genes in these species remained almost unchanged (Nozawa and Nei, 2007).

Despite the similarities in the basic olfactory system in insects and mammals [like insects, in vertebrates olfactory receptor neurons (ORNs) send information for processing to the olfactory bulb, the vertebrate equivalent of the insect antennal lobe], in many mammals, along the main olfactory epithelium, evolved the vomeronasal organ (VNO) as a second olfactory organ. It resides within a thin bone capsule and its nerves (vomero-nasal nerves) follow a pathway to the brain that is different from that of the olfactory nerves. It is believed that this specialized olfactory organ in mammals serves for intraspecific communication, including mating.

Auditory-cognitive Reproductive Isolation

Neural Reception and Processing of Acoustic Signals

Male singing in crickets is regulated by a brain center located in the command neurons of the anterior protocerebrum. This song center is experimentally activated by administration of cholinergic neurotransmitters and inhibited by GABA (gamma-butyric acid). During the processing of song patterns, changes in the cytosolic Ca²⁺ occur in parallel with the chirp rhythm in the auditory interneurons. Acoustic signals of the male song are used as mating signals by female crickets (Hedwig, 2006). There is empirical evidence on the role of courtship auditory cues in mate choice and discrimination of conspecific from heterospecific individuals in insects.

In amphibians, the receiver processes auditory signals (mating calls) in a part of the midbrain, torus semicircularis, which is considered to be homologous to the mammalian inferior colliculus. Coding of acoustic communication signals is also believed to take place in this part of the brain.

The response of the auditory midbrain of female tungara frogs, Physalaemus pustulosus, to mating calls of conspecific, heterospecific, and irrelevant calls has been evaluated by the expression of the gene egr-1 in various regions of the torus semicircularis. The pattern of expression in response to those stimuli is different and specific for the laminar, midline and principal nuclei. Within these nuclei a difference is observed in the patterns of egr-1 expression in the midline and principal nuclei, on the one hand, and the laminar nucleus on the other (Hoke et al. 2004). Another part of the amphibian brain that is crucially involved in the signal receiver response to auditory mating calls is the hypothalamus, which is anatomically and functionally connected to torus semicircularis. Several hypothalamic nuclei are involved in neurohormonal regulation of reproductive behavior (figure 20.12).

A correlation exists between the levels of egr-1 expression in the auditory midbrain and in the hypothalamus. Differences in egr-1 expression, in response to auditory signals are also observed between the different regions of the hypothalamus that are involved in the behavioral responses to auditory inputs (Hoke et al., 2005).

Figure 20.12. Functional connectivity of the hypothalamus. Gray arrows show significant relationships in frogs that heard irrelevant acoustic stimuli (P. enesefae whine and chuck-only), and black arrows indicate relationships in frogs exposed to behaviorally relevant stimuli (conspecific whine and whine–chuck). (A) egr-1 levels in midbrain and thalamic nuclei implicated in auditory processing are significant predictors of hypothalamic expression patterns. Relationships between auditory and hypothalamic regions do not vary with relevance of stimulus. (B) egr-1 correlations between hypothalamic regions differ based on behavioral relevance of acoustic stimulus (From Hoke et al., 2005).

Specific responses of individual hypothalamic regions are derivative of the responses of the auditory nuclei to which they are connected, whereas the functional connectedness within the hypothalamus is an emergent property modulated by the relevance of the social context. The principles of parallel processing and distributed functional networks in the frog hypothalamus are remarkably similar to those processes posited in cognitive neuroscience including perception, memory, and decision-making. (Hoke et al., 2005; figure 20.12).

Acoustic signals in frogs are received by two peripheral auditory organs, the amphibian papilla for low frequency sounds and basilar papilla for higher frequencies. The auditory input is processed in the hypothalamic auditory nuclei, the auditory nuclei of the midbrain and the thalamus.

In the majority of birds it is the male that performs singing and it is observed that males have two main song nuclei in the forebrain, the high vocal center (HVC) and the nucleus robustus archistriatalis. In the course of evolution, with the increase of the song repertoire, a corresponding increase in the volume of HVC occurred. For example, in the sedge warbler only male birds sing, and their HVC volume is 7 times greater than in females. The sexual dimorphism of the brain arises experience-independently, and brain differences are observed between birds reared in isolation and those that are exposed to songs (Leitner et al., 2002). In all likelihood these remarkable changes in the brain structure between males and females are epigenetically determined and inherited for no genes in sex chromosomes have been identified to be involved in the process of development of brain sexual dimorphism. Not only the brain centers but the muscles regulating singing in males as well are larger than in females.

In singing birds in general brain nuclei responsible for singing increase their volume during the breeding season.

The Song Circuit in the Brain of Birds

The song production is function of the “song system”, a group of functionally related brain nuclei and respective connecting pathways, which are similar in all groups of birds that have evolved learned song (songbirds, parrots, and hummingbirds). Two distinct song circuits or impulse pathways that start from the HVC (high vocal center), and have their homologues in the mammalian brain, have evolved within the song system in birds: the PDP (posterior descending pathway) or song production circuitry (HVC à RA à, nXIIts àmuscles of syrinx, the song organ) which is responsible for both the acquisition and production of learned song, and the AFP (anterior forebrain pathway) or song learning circuitry (HVC à Area X à DLM àLMAN à RA), which is necessary for song acquisition only (Nottebohm, 2005; figure 20.13).

The neural circuitry in the brain song control system is determined by brain intrinsic mechanisms and recent evidence shows that the song repertoire in birds is an automatic result of the spontaneous activity of neural circuits in the brain song nuclei. This conclusion has been drawn from the fact that even when kept in complete acoustic isolation, or when birds are experimentally deafed, they are still able to develop the species-specific song repertoire in an experience-independent way (Leitner et al., 2002). However, this may not be universally true; other experiments with fledgling zebra finches have shown that formation of the motor phase of the song neural circuitry is determined within the 10 days after birth,when it learns tutor’s song, and becomes fully functional by 35 days of age (Roper and Zann, 2006).

Figure 20.13. The Song System of Songbirds. Nucleus HVC feeds information into two pathways that ultimately lead to the neurons in the tracheosyringeal half of the hypoglossal nucleus (nXIIts) that project to vocal muscles. HVC projects to nucleus RA directly (PDP), and indirectly via Area X, the dorsolateral anterior thalamic nucleus (DLM), and LMAN (AFP) in a manner that shares similarities with the mammalian pathway cortex→basal ganglia→ thalamus→ cortex.

Abbreviations: AFP, anterior forebrain pathway; DLM, dorsolateral anterior thalamic nucleus; LMAN, lateral magnocellular nucleus of the nidopallium; RA, robust nucleus of the arcopallium; nXIIts, the tracheosyringeal motor nucleus in the brain stem; PDP, posterior descending pathway (From Nottebohm, 2005).

Evidence on factors involved in regulation of the song control system in birds is also contradictory. While some reports suggest that animals respond to the gonadal testosterone by modifying the structure and function of the song control system (Brenowitz, E.A. 2004), other studies show that the development of the neural song system in zebra finch males is regulated by brain factors rather than influenced by circulating hormones (Wade and Arnold, 2004; Gahr, 2004). Besides the neurosteroid pathway in the brain, which regulates the production of sex hormones, it is important to remember that even changes in gonadal testosterone levels, ultimately, represent downstream results of neuroendocrine cascades starting in the brain, along the hypothalamic-pituitary-gonadal axis. Song-related daily release of steroid hormones is regulated by the circadian system under control of specific hypothalamic centers in the avian brain, with the release of melatonin as a central regulatory element. All the brain centers of the song control in male (not female) house sparrows, the HVC (high vocal center), magnocellularis anterior (MAN), and Area X, contain elevated levels of melatonin, which ultimately is a mediator of the central neurally determined circadian mechanism (Whitfield-Rucker and Cassone, 1996).

Neural song circuits are determined before hatching and their establishment does not depend on experience. How are these neural song circuits in male birds determined and established? There is no reason to believe that they are established differently from how other neural circuits in the brain form: they are designed during the individual development and fine tuned during the later life by learning/experience. Their establishment during the individual development seems to take place experience-independently, “according to the brain’s best guess”.

In some female birds, a special relationship has been observed to exist between the reception of songs and the size of eggs they produce. So, e.g., the playback of male “sexy” syllables (songs) in canaries induces production of eggs that are larger than when they don’t hear male syllables or hear only no “sexy” syllables (Leitner et al., 2006).

Neural song circuits in the brain of birds show remarkable plasticity both in the size of song control nuclei (which become larger during the breeding season) and the synaptic connections, plasticity that manifests itself in changed patterns and duration of songs during the breeding and nonbreeding seasons. Changes in neural song circuits are cause of the observed seasonal variations in songs (Brenowitz, 2004).

The evolution of songs and the increase of the complexity of birds’ songs has been related to the female preference for complex songs. For explaining the preference, an “anti-monotony” hypothesis has been put forward, according to which complex vocalizations reduce or prevent habituation of females. The innate preference of birds for complex vocalizations is illustrated by the example of grackles: although males of this bird lack complex songs in their repertoire, females are more attracted to heterospecific complex songs and even to the artificially constructed ones (Ryan, 1998).

The fact that birds can so easily, without changes in genes, modify their neural song circuits and vocalizations in response to seasonal changes suggest that birds potentially can change mate preferences and mate choice, inducing thus reproductive isolation within evolutionarily short periods of time.

Acoustically-determined Reproductive Isolation in Sympatry

One of the first forms of interindividual communication in metazoans has probably been based on the production and perception of vibrations and sounds. A colerrhynchan insect, Hackeriella veitchi, in Northeastern Australia, believed to be a Gondwanan relict insect lineage, emits vibrational signals for interindividual communication. This suggests that vibrational signaling by a simple tymbal (sound-producing organ) observed in this insect, and probably other mechanisms of signal production, (percussion and stridulation) evolved in four groups of Hemiptera as early as 230 Mya (Hoch et al., 2006).

Some insects have evolved special organs for producing ultrasounds of above 20,000 vibrations per minute, and use them as cues in mate choice. For example, males of the lesser wax moth emit such ultrasonic signals in 100-120 pairs of pulses/sec, which have a phonotactic effect on females. The female brain has evolved such a high acoustic resolution power that it can distinguish even acoustic signals that last for as little as 150 microseconds. Female preferences for pulse amplitude, rate, length, and length intervals are greater than the average values of male populations. However, when the pulse rate increased to 42 pairs/sec, the female preference levelled off and later decreased because such sounds last less than the length of an action potential and auditory neurons of A. grisella females cannot generate 142 action potentials for second (Jang and Greenfield, 1996). Male crickets produce their song by rhythmic wing movements under control and regulaton of the command neurons descending from a particular center in the anterior protocerebrum (Hedwig, 2005). This song is received by interneurons in the acoustic sensory organs of female crickets.

R.M. Hennig has observed that females of two sibling cricket species, Teleogryllus oceanicus and T. commodus, have evolved two different temporal filters, the first species - a period filter and the second - a pulse duration filter (Hennig, 2003). The rapid evolution of homologous circuits in two sibling species of such different properties of pattern analysis and temporal filters, after diverging from the common ancestor, is believed to have resulted from changes in the properties of neural circuits responsible for conspecific call recognition. There is no evidence on any changes in genes being involved in this rapid evolutionary divergence.

Gray and Cade have studied the acoustically-determined speciation process in the case of the North American field cricket species, Gryllus texensis (formerly Gryllus integer) and Gryllus rubens, two cryptic sister species, that are morphologically indistinguishable, living both in sympatry and allopatry. They are prezygotically completely isolated from each other while showing no postzygotic isolation. They do not hybridize under natural conditions although they can produce viable hybrids with equal fertility to that of parental species. The prezygotic reproductive isolation and the evolution of these sister species seems to have been based exclusively on the evolution of divergent male songs and respective female preferences in two sympatric populations of the common ancestral species. These cryptic species differ between them in both male songs (their trills are distinct, with 80 pulses/sec and 56 pulses/sec respectively) and female preferences. Based on the presence of prezygotic and absence of post-zygotic reproductive isolation, in the absence of character displacement in male signals and female responses as well as in positive correlation between male signals and female responses, investigators have concluded that evolution of these two species from their common ancestor is the result of sexual selection (Gray and Cade, 2000).

Males of insect sibling species Neoconocephalus robustus and N. bivocatus generate exceptionally fast calls with pulse rates of ~200 s^-1 and ~175^-1respectively. Females of both species are highly selective about conspecific male calls but recognition mechanisms are “strikingly different” despite their close relationship as sibling species. Females of N. robustus respond to a continuous, very fast conspecific call of 200 pulses/s with very short intervals, which is recognized as such but may be impossible to be faithfully encoded by their sensory system. N. bivocatus females, on the other hand, are attracted to their own conspecific calls with a pulse rate of 175 pulses/s, but the sensory system of the insect merges every two consecutive pulses into one pulse pair, by ignoring the interval between each pair, thus, creating 87 pairs of pulses. By halving that fast and difficult to recognize pulse rate, the insect is able to recognize its conspecific call (Deily and Schul, 2004). Obviously, no changes in genes can be related to the neurobiologically determined the ability bility of insects to “halve” the pulse rate.

Two subspecies of the grasshopper, Chorthippus parallelus, live on both sides of Pyrenees (South France and Iberian Peninsula). Their mate recognition system uses acoustic signals, and possibly olfactory signals. A study on the female preference of the grasshoppers on both sides of the hybrid zone has shown that the female homogamic preference differs abruptly over a distance of 1 km, suggesting that selection for female preference operates in the hybrid zone (Butlin and Ritchie, 1991).

Visual signals, color and patterning of the body, are certainly the main cues used by female fish in mate recognition and choice but recent studies have shown that at least three sympatric cichlid species (Pseudotropheus zebra, P. callainos and P. “zebra gold”) in Lake Malawi (formed ~ 1 million years ago), generate different species-specific acoustic signals that are distinct with regard to the pulse rate and frequency and are used as cues for mate choice. Although there is additional evidence that these acoustic signals are used in mate choice and species recognition, there is no evidence that these courtship acoustic signals have led to sympatric speciation in cichlid fish of Lake Malawi (Amorim et al., 2004).

Males of the túngara frog, Physalaemus pustulosus, add a chuck to the basic whine component of their mating call. Adding of this suffix is characteristic of bigger males, so that by preferring calls with chucks, indirectly, females choose larger male mates. The whine component is processed in the amphibian papilla, auditory structure of the inner ear, and chuck component of higher frequency is processed in the basilar papilla of the female frogs. The amphibian papilla may function for species recognition and basilar papilla may be specifically involved in determining mate preferences. The tuning of the basilar papilla for sensing chucks represents an ancestral state, so the preference for chuck evolved before the evolution of chucks in male calls and, most importantly, its evolution involved no changes in genes.

Females gain a reproductive advantage by preferring whines with lower frequency chucks, but selection for this female preference is not the evolutionary cause. Instead, the reproductive advantage gained by preference for whines with lower frequency chucks is an evolutionary effect due to the way frog brains work and how they have evolved. (Autumn et al., 2002)

In a series of studies it is proven that evolution of courtship song and reproductive isolation in the group Drosophila willistoni (consisting of six sibling species) (Gleason and Ritchie, 1998) and evolution of mating calls in túngara frogs are not related to genetic evolution (Pröhl et al., 2006). In this context, it is important to bear in mind that songs themselves (except probably for some stereotypic sounds which may have a genetic basis) are behavioural phenomena, which cannot be traced back directly to genetics.

Birds may be hearing-only birds, or hearing-and-vocalizing birds (hummingbirds, parrots and songbirds). The latter are capable of vocal learning by imitating other birds or even non-avian species in the case of parrots. This capability is related to the evolution in these birds of specialized forebrain structures (figure 20.14).

Seven brain centers are related to vocalization of which five centers in telencephalon (caudomedial neostriatum, or NCM; caudomedial hyperstriatum or CMHV; dorsocaudal neostriatum or Ndc; intermediate archistriatum or Ai; and caudal paleostriatum or PC) one in thalamus (dorsointermediate nucleus of the posterior thalamus or DIP) and one in mesencephalon (dorsal part of the lateral mesencephalic nucleus or MLd) (Jarvis et al., 2000).

In vocal learning birds, such as hummingbirds, songbirds, and parrots these structures have evolved independently and, curiously enough, they are well conserved among species and in all of them comprise seven vocal control nuclei in seven different regions of the forebrain (in hummingbirds there is an additional, eighth, nucleus in the mesencephalon). Not surprisingly, these structures are absent in hearing-only, vocal non-learners, avian species. Given the incredible fact that a group of seven structures has independently and similarly evolved in 3 of 23 avian orders, it is suggested that

The evolution of these structures is under strong epigenetic constraints; in which case, similar structures may have also evolved in vocal learning mammals (humans, cetaeans and bats). (Jarvis et al., 2000)

Figure 20.14. The brains of songbirds and non-songbirds differ. These schematic diagrams of parasagittal views of the brains of a songbird (a) and a non-songbird (b) illustrate the dramatic differences between them. Songbirds have an elaborate network of interconnected forebrain nuclei that form an interface between auditory input (which converges on field L, the primary auditory projection region in the avian forebrain) and vocal output, which is produced in the syrinx, the avian vocal organ. Non-songbirds also have field L, and they can produce vocalizations in the syrinx, but they do not have the network of forebrain nuclei that songbirds have.

Abbreviations: DLM, nucleus dorsolateralis anterior, pars medialis; DM, dorsomedial nucleus of the midbrain nucleus intercollicularis; HVC, high vocal center; lMAN, lateral magnocellular nucleus of the anterior nidopallium; mMAN, medial magnocellular nucleus of the anterior nidopallium; NIF, nucleus interface of the nidopallium; nXIIts, tracheosyringeal portion of the nucleus hypoglossus; RA, robust nucleus of the arcopallium; RAm, nucleus retroambigualis; rVRG, rostro–ventral respiratory group; X, Area X. Adapted. (From Bolhuis and Gahr, 2006).

Evolution of complex structures of the song circuit in birds has been epigenetically determined and involved no changes in genes.

Evidence presented in this subsection shows that neural mechanisms of releasing, receiving and processing of auditory signals determine reproductive isolation between a number of closely related species of insects. That evidence also suggests that these neural mechanisms change and determine reproductive isolation of natural populations without changes in genes.

Electrocognitive Mechanism of Reproductive Isolation

Electrogenesis in Fish

In 1951, H.W. Lissmann suggested that the ability of a fish occurring in the Nile River, Gymnarchus niloticus, to avoid obstacles when

swimming backward, i.e. outside its visual field, may be related to the presence of an electric organ, which “may enable the animal to detect objects in the vicinity of its body” (Lissmann, 1951). Ever since the study of electric signals in fish has been object of a special interest.

Besides visual, acoustic, and olfactory signals, elasmobranch fish (rays, skates, and sharks), a group of ~800 species, use favorable aquatic medium for electrical prey detection and intraspecific communication. Electrogenesis, production and emission of electrical signals, is a defining feature of tropical mormyriform fish (the genus Mormyridae with about 200 electric species and the genus Gymnarchidae, with only one species of electric fish) in Africa and gymnotiformes with 115 nominal species in seven fish families of the South and Central America, as well as a number of catfish (Hopkins, 1999).

These teleost fish have specialized electrical organs consisting of muscle-derived cells or electrocytes, which in mormyrids are represented by multinucleated structures of several centimeters in diameter. The electric organs produce weak electric pulses and form an electromotor system for transmitting electrical signals for electrical guidance, reproductive behavior, and intraspecific communication.

Electrical signals used for interindividual communication in mormyrid fish consist of electric organ discharge (EOD) pulses, with a constant waveform that is believed to serve species recognition, while the sequence of pulse intervals (SPI) is variable (100-300 ms) and expresses sender’s identity and motivation (Carlson and Hopkins, 2004).

The electrical organ is innervated by electromotor neurons of the spinal cord, which receive signals from medullary neurons that, in turn, are induced by signals from the fish brain. Ultimately, the electric organ discharges (EODs) are determined by specific central nuclei in the CNS. A medullary command nucleus (CN) receives inputs from the precommand nucleus (PCN) and the dorsal posterior nucleus (DP) (Carlson, 2002).

A simple neural circuit that starts with the CN, via the medullary relay nucleus (MRN) activates spinal electromotor neurons (EMNs), triggering production of EODs (figure 20.15).

Figure 20.15 Sagittal schematic showing the functional neuroanatomy of the mormyrid electromotor system. Excitatory terminals are identified by flat lines, inhibitory terminals by solid circles. Black denotes medullary electromotor nuclei, dark gray denotes mesencephalic and diencephalic electromotor nuclei, and light gray denotes corollary discharge nuclei.

Abbreviations: BCA, bulbar command-associated nucleus; C3, third cerebellar lobule; CN, command nucleus; DP, dorsal posterior nucleus of the thalamus; EGp, eminentia granularis pars posterior; EL, exterolateral nucleus of the torus semicircularis; ELL, electrosensory lateral line lobe; EMN, electromotor neurons; IL, inferior lobe of the hypothalamus; MCA, mesencephalic command-associated nucleus; MRN, medullary relay nucleus; OB, olfactory bulb; PCN, precommand nucleus; Tel, telencephalon; TM, tectum mesencephali; Val, valvula of the cerebellum; VPd, dorsal subdivision of the ventroposterior nucleus of the torus semicircularis (From Carlson and Hopkins, 2004).

Regulation of the normal baseline rhythm of 100-300 ms EOD intervals is function of PCN (precommand nucleus) and DP (dorsal posterior nucleus of the thalamus), which, in turn, receive input from the VPd (dorsal subdivision of the ventroposterior nucleus of the torus semicircularis). (Carlson and Hopkins, 2004).

Electroreception is widespread among metazoans. Besides fish, electroreception has independently evolved in a number of amphibian taxa and it is known to have also evolved in at least one mammalian species, the platypus (Ornithorhynchus anatinus) in Australia and Tasmania (figure 20.16).

…………………………………………..

The Structure of the System of Electroreception in Fish

Mormyriform fish have cutaneous electroreceptors of three types: ampullary electroreceptors, used for electrolocation of preys and predators, mormyromast electroreceptors (electrolocation within close range) and Knollenorgans (Germ. for tuberous organ), electroreceptors for reception of EODs from other fish, which are used for interindividual communication (Hopkins, 1999).

The first anatomic descriptions of ampullary organs in elasmobranch fish (sharks, skates, and rays) appeared by the middle of the 17 century, but it took three centuries until, by early 60es of the 20^th century, biologists discovered that the ampullary organ has electrosensory functions (Tricas, 2001).

Knollenorgan electroreceptors are broadly tuned to the species-specific EOD spectrum. Damages of the Knollenorgan pathway to the brain prevent fish from performing various communication behaviors. When electric fish emit EODs, the Knollenorgan pathway in the CNS is blocked so that its own EODs are not transmitted for processing in the brain (Arnegard et al., 2005).

The electromotor system has coevolved with the sensory electroreceptor system in these species. Besides, the mormyrid Gnathonemus petersii, as well as some other South American and African species, are capable of emiting electrosignals and perceiving their reflection in the skin electroreceptors. These fish have two small electroceptive pits or foveae (Lat. pl. of fovea, small pit) resembling the visual fovea in the retina) in the Schnauzenorgan (German Schnauze - muzzle), which help them to form an “electrical image” of objects with strong contextual effects, by determining the distance and the three-dimensional shape of the object (von der Emde, 2006; Caputi and Budelli, 2006). Electrosensory units of these fish are ampullae of Lorenzini (figure 20.17). Each of these units consists of a group of subdermal alveoli that are connected to the environment via a 1mm long canal. The structure of the ampullary canals enables elasmobranchs to distinguish between small electric fields of living organisms and large uniform electric fields of abiotic environment (Tricas, 2001). The lining of the ampulla consists of a layer of sensory cells, with their kinocilia projecting to the lumen of the ampulla, and the supporting cells. The receptor cells are innervated by primary afferent neurons, which encode the amplitude and frequency of electrical signals and transmit them to the brain (Tricas, 2001) where a perception of the source of electrical stimuli is generated.

It is believed that electric signaling, based on the ampullary electrosensory systems such as receptors of ampullae of Lorenzini, appeared very early in vertebrate evolution (according to Bullock et al., 1982, in Bodznick et al., 2003). In fish it has evolved at least 6 times. In marine species these ampullae are grouped in clusters distributed mainly on the head region.

Figure20.17. The ampulla of Lorenzini. a – the ampulla of the barndoor skate, Raja laevis, is formed by several alveoli that share a continuous lumen (L) and a subdermal canal that has a single pore on the skin. The sensory epithelium (SE) forms the highly resistive ampulla wall that connects with the canal epithelium (CE) at the marginal zone (MZ). The ampulla and canal are filled with a highly conductive gel. This arrangement forms an electrical ‘core conductor’ in which the potential within the ampulla lumen is isopotential with that at the surface pore. The sensory epithelium is innervated by primary afferent neurons that conduct electrosensory information to the brain. b – in most elasmobranch species, the sensory epithelium is a layer of receptor cells (RC) and support cells (SC). Tight junctions between these cells form a high electrical resistance barrier between the lumen of the ampulla and basal portion of the receptor cells. The difference between lumen voltage and reference voltage (VREF) stimulates the small apical surface of the receptor cells and controls release of neurotransmitter onto primary afferent neurons (N) (From Tricas, 2001).

The electrosensory system allows fish and amphibians to recognize prey, predators, conspecifics, and mates as well as regulate social behavior. The primary afferent neurons encode the data on electrical amplitude and frequency and transmit these data to the brain (Tricas, 2001), in the dorsal octavolateral nucleus (DON) on the dorsolateral wall of the hindbrain (Bodznick et al., 2003).

………………………………………….

Electrosensory Communication in Social and Reproductive Behavior

EOD diversity in electric fish stems from differences in the shape and duration of the wave but also from the pulse polarity. The form of the wave depends on the type of cells in the electric organ. The duration of the wave varies between 100 microseconds and 10 milliseconds (Hopkins, 1999).

In the electric fish that show sexual dimorphism for EOD, differences are also observed in the morphology of electrocytes (Hopkins and Bass, 1981). So, e.g., the South American gymnotiform fish, Hypopomus occidentalis is sexually dimorphic both in the morphology of electrocytes and in the frequencies of EODs, with females having higher peak power frequencies (Hagedorn and Carr, 1985). However, there are other species of electric fish in which the morphology of electrocytes in the electric organs seems to be similar in males and females but sexual dimorphism, i.e. differences in the EODs of adult males and females also occur (Mills et al., 1992).

Mormyrid fish have an electric organ in their caudal peduncle, which generates pulses of EODs (electric organ discharges) (Arnegard et al., 2006). Species of this group are nocturnal fish that use EODs for locating objects in darkness.

…………………………………………

Duration of EODs reflects the social status (high-ranking males have longer EODs and lower-ranking ones have shorter EODs) of male fish, as assessed by males themselves, and is used in the communication between them. Changes in the social status as well as changes in hormonal levels are correlated with respective changes in the EOD duration, making this trait very plastic (Carlson et al., 2000).

Studies on the form and duration of EOD (electric organ discharge) waveforms in fish led to the identification of new species within what previously, based on morphological criteria, have been considered to be single species (figure 20.21).

Figure 20.21. Systematists disagree on the number of species in the genus Campylomormyrus, but recent recordings of EOD waveforms suggest there may be even more species than the latest revision, which recognized 18 species. Only four of the nine species represented here can be unambiguously identified by reference to the type specimens, most of which are housed in the Musee Royal de l’Afrique Central in Tervuren, Belgium. The remainder represent forms whose identity is uncertain (From Hopkins, 1999).

Changes in the electrosensory system of fish also occur during their individual development. The sensitivity of the electrosensory system in the Atlantic stingray, Dasyatis sabina, changes with age so that enables the developing fish to avoid predators when still young, and detect prey and choice mates in adulthood (Sisneros and Tricas, 2002). Obviously, these changes in electrosensitivity are epigenetically determined, i.e. involve no changes in genetic information. It has been shown that such changes in the properties of the electrosensory system are correlated with the sequential ontogenetic changes in the electrosensory primary afferent neurons in embryos, neonates, juveniles, and adult individuals of the clearnose skates, Raja eglanteria (Sisneros et al., 1998).

It is also noteworthy that neonates of the bonnethead shark, Sphyrna tiburo, have well-developed electrosensory system and show innate electrosensory feeding behavior by biting electrodes without previous association of electric stimuli with food (Kajiura, 2003).

Stingrays produce a standing dc (direct current) bioelectric field, which they use for detecting and locating conspecifics or mates. The highest frequency sensitivity of the primary afferents in the clearnose skate (Raja eglanteria) and the little skate (Raja erinacea) is similar to the pulse rate of their respective electric organ discharges (EOD), indicating the importance of electro-communication in their social and reproductive behaviors (Tricas and Sisneros, 2004). The round stingray, Urolophus halleri, has not a special electric organ but its bioelectric field arises from standing ionic potentials at various sites on the skin and buccal epithelium and this bioelectric field during mating season is used by males as the primary cue for localizing their female mates buried under the sandy bottom of shallow shoreline (Tricas et al., 1995).

In males of the Atlantic stingrays, Dasyatis sabina, a correlation exists between the seasonal changes in the androgen levels in body fluids and changes in the dental sexual dimorphism, sexual behavior, aggressive behavior, as well as the neurophysiological changes in the electrosensitivity. These electrosensory changes enhance the ability of juvenile fish to avoid predators, improve the ability of adults to locate prey, coordinate social behavior in both sexes, and help adult males to detect and identify mates (Sisneros and Tricas, 2000).

All but species of genus Stomatorhinus in the family of mormyrid fish, have three separate zones in the rhombencephalic ELL (electrosensory lateral line lobe) (figure 20.22)

where all electroreceptor primary afferents project: the VLZ (ventrolateral zone), which receives ampullary organ afferents, the MZ (medial zone) where mormyromast A-cell afferents terminate, and DLZ (dorsolateral zone), which receives B-cell afferents.

In electric fish of the Ivindo River in Gabon, West Central Africa, only two ELL zones (VLZ and MZ) exist. The loss of DLZ is an anatomic change clearly related to the termination of the electrosensory input from the B-cell in species of the genus Stomatorhinus (McNamara et al., 2005). This fact illustrates the close relationship between the sensory input and organization of the central nervous system.

Arnegard and coll. (2005) examined pairs of three sympatric morphs (type I, II and III) of the Brienomyrus flock of species of mormyrid fish that have radiated recently in East Africa. These morphs are genetically and morphologically indistinguishable, but are clearly distinct in their EOD patterns (rate and waveform). Moreover, they use the same microhabitat what obviously excludes the possibility of existence of different gene pools, and action of different evolutionary pressures:

Type I and type II/III are conspecific signal types comprising undifferentiated gene pools in each regionally defined and phenotypically polymorphic population. (Arnegard et al., 2005)

Figure 20.22. Schematic comparison between P. marchei and Stomatorhinus ivindoensis. A summary of the differences found between the mormyromasts of a species with a three-zone ELL (P. marchei) and Stomatorhinus, missing the B-afferent and, hence, the DLZ in the ELL.

Abbreviations: A, A-cell; ac, accessory cell; B, B-cell; bm, basement membrane; e, epidermis; i.c., inner chamber; i.e.c., intraepidermal cavity; m, myelin sheath; n, afferent nerve (From McNamara et al., 2005).

C. numenius displays a common EOD waveform type in juveniles but adult individuals generate three different EOD types. It is observed that these distinct waveform EOD types are each correlated with a different fish morphotype (morphotypes A, B, and C). These sympatric morphotypes are reproductively isolated. Based on the study of microsatellite loci investigators have concluded that these “morphotypes” of C. numenius in fact represent three cryptic species (Feulner et al., 2006).

Reproductive isolation in sympatry of fish of the Campylomormyrus genus of Central African river basin is correlated with differences that these species show in waveform types of their EODs. Campylomormyrus tamandua is characterized by common EOD in young and adult, in male and female individuals (figure 20.23)

Figure 20.23. Overlays of amplitude-normalized EODs (n D number of individuals per overlay). (A) Common EOD type of C. tamandua. (B) Common juvenile and different adult EOD types of three different morphs (A, B, and C) of C. numenius. Note the different time scales (From Feulner et al., 2006).

Evolution of Electrosignals, Electrocognitive Isolation of Populations, and Sympatric Speciation in Fish

Behavioral studies have shown that the ability of electric fish to distinguish between species-specific and nonspecific EODs as well as between female and male EOD waveforms (Hopkins and Bass, 1981; Arnegard et al., 2005) may have played a role in evolution of electric fish; it may have been used not only to recognize conspecifics and for social interactions between them but also as a mechanism of reproductive isolation between groups of individuals or populations in the process of speciation. This seems to be supported by observations on more than 20 sympatric species of the Brienomyrus fish stock in the Gabon River, with each of them producing a species-specific EOD waveform.

Nested within the Gabon Brienomyrus species

flock of the Gabon River is the magnostipes complex consisting of sympatric morphologically indistinguishable morphs (type I and type II/III) of the same size. Reproductively, morphotypes of this complex are not isolated but express distinct EODs although they are genetically indistinguishable at several nuclear loci. Each morph of the magnostipes complex generates its specific EOD (figure 20.24).

Figure 20.24. Adult sizes of sympatric magnostipes-complex morphs from the Makokou region of the Ivindo River. Photographs of a type II male (above: specimen 5945; SL=105·mm) and a type I male (below: specimen 5944; SL=147·mm) collected from Loa-Loa Rapids, showing their elongated EODs (Arnegard et al., 2006).

However, males of the type II, when exposed to playback EODs of both types I and II respond preferably to type II, whereas type I males show no preference (Arnegard et al, 2005; Arnegard et al., 2006). In view of the striking morphological and genetic similarities between sympatric morphs, while they are still interbreeding, the dramatic differences in the EOD waveform and EOD-mediated species recognition during courtship within the Brienomyrus stock suggest that they are in the process of sympatric speciation (Arnegard et al., 2005; figure 20.25).

Two lines of evidence and a well-studied central nervous system mechanism suggest a general role for EODs in species isolation in the Gabon Brienomyrus flock. Firstly, each of the two dozen or more morphologically distinct species that have already been discovered produces a different, species-typical EOD waveform. Secondly, in situ electrical playback experiments have demonstrated EOD-mediated species recognition in the context of courtship in Brienomyrus sp. vad. Lastly, the EODs of type I and type II/III appear to differ sufficiently for waveform discrimination via a neural pathway described by Xu-Friedman and Hopkins (1999; and references therein). Therefore, instances of dramatic signal difference between sympatric morphs appear somewhat paradoxical in this group of fish when genetic evidence for reproductive isolation between them is lacking. (Arnegard et al., 2005)

Figure 20.25. Examples of sympatric assemblage of morphologically similar mormyrids from the Brienomyrus species flock of Gabon. Photographs are shown next to voltage traces of electric organ discharges (EODs) recorded from the same individuals: (A) adult male; (B) adult female or nonbreeding male; (C) adult female or nonbreeding male; and (D) adult female or nonbreeding male. Scale bars of 20 mm and 1msec are indicated (From Arnegard et al. 2005).

Evolution of electrogenic and electrosensory organs in gymnotiform fish of South America and in freshwater mormyrid fish of Africa represents a striking example of evolutionary convergence. That evolution has occurred independently in both groups and yet they exhibit striking similarities. In both groups the production of EODs is controlled by a ventral midline nucleus, which is known as PN (pacemaker nucleus) in gymnotiforms and CN (medullary command nucleus) in mormyrids. In both groups neurons of this nucleus project to adjacent relay neurons whose axons innervate electromotor neurons in the spinal cord (Carlson and Hopkins, 2004).

Studies conducted on these fish seem to exclude genetic factors from a possible involvement in the process of the evolutionary divergence of these morphs in sympatry:

Instances of dramatic signal difference between sympatric morphs appear somewhat paradoxical in this group of fish when genetic evidence for reproductive isolation between them is lacking…Striking genetic similarity between coexisting morphs provides the signature of a fully sympatric process, whether it involves incipient speciation or the maintenance of phenotypic dimorphism by some other mechanism. An additive, polygenic signal architecture, if found, would cast doubt on the viability of a scenario of postdivergence hybridization (upon secondary contact) as an alternate explanation for the genetic patterns we describe. Under such an architecture, introgression sufficient to homogenize allele frequencies across neutral loci would have almost certainly led to a breakdown of signal differences, rather than the discrete signal classes we observe. (Arnegard et al., 2005)

Neodarwinian Explanation of Sympatric Reproductive Isolation and Speciation of Brienomyrus spp.

From the paradigmatic neoDarwinian view, it would be predicted that evolution of Brienomyrus spp. in sympatry, is animprobable, if not impossible, event because:

1. Inherited phenotypic changes, including behavioral changes in electrical communication, require changes in genes or allele frequencies and such changes are not there.

2. Reproductive isolation between electrogenic Brienomyrus fish populations in sympatry is impossible because of the unavoidable gene flow between populations that use the same microhabitat.

Both the above neoDarwinian predictions are invalidated by the lack of genetic changes and by the demonstrated occurrence of the evolutionary divergence of these fish in sympatry.

Epigenetic Explanation of Sympatric Reproductive Isolation and Speciation of Brienomyrus spp.

Repeated evolution of electrolocation in fish is related to the suitability of aquatic environment for electric communication. Although obvious, this statement does not tell anything why the electrolocation evolved first in fish, and in other aquatic vertebrates (aquatic amphibians and even in a semiaquatic mammal) but no aquatic invertebrate is known to have evolved it. Probably, a plausible explanation would be that electrolocation requires a complex algorithm that probably is beyond the computational capabilities of the invertebrate CNS/neural net. Additionally, recall that to the analysis and interpretation of electric signals fish devote a considerable part of the brain that invertebrates probably cannot afford.

Independent evolution of electric communication in several fish groups and its evolution exhibits a high degree of convergence. Clear parallels are observed in the evolution of the mechanisms of rhythm control and the computational algorithms used for processing temporal cues. Remarkably, solutions to the problems related to evolution of electric discharges have been in certain respects similar but in others different for different groups (Hopkins, 1999), with no evidence on involvement of gene mutations or any changes in the genetic informationor allele frequencies in populations.

How could such electrically distinct morphs evolve within panmictic populations of a common genotype that use the same microhabitat? How is the reproductive isolation initiated and maintained in populations of the same genotype in sympatry? We have already an empirically corroborated answer: the differentiation of populations is based on the fact that each population emits strictly determined EOD patterns, can discriminate between conspecific and heterospecific EODs and has evolved the preference for species-specific EODs. As we have already argued and substantiated, different neuro-cognitive properties of incipient species result from modifications in the structure of respective neural circuits, which often exhibit a high evolutionary plasticity. There is no reason to believe that the different electrogenic and electrosensory properties of fish populations of Brienomyrus species in sympatry and the respective neural circuits that are developmentally very plastic could not also be evolutionarily equally plastic.

The repeated and independent evolution of electrogenic and electrosensory apparatus in fish, the convergence during the evolution of the algorithms that made the electrocognitive behavior possible, evolution in all of them, often independently, of the adaptive cancellation (dissociation of the sensory information coming from the environment from the one that results from their own activity), are based on high computational capabilities of special neural circuits in the central nervous systems, with no proven or hypothetical involvement of genetic mechanisms.

Evolution of neural circuits that may lead to evolution of animal behavior, including electrogenic and electrosensory behavior, may be related to changes in the number of neurons, which the respective circuits consist of, or to patterns of synaptic connections (establishment of new connections and loss of existing connections), for the properties of neural circuits and their output depend on the structure of circuits.

Neural circuits may not only evolve suddenly but often they prove to be evolutionarily extremely conservative. So, e.g., circuitries responsible for feeding behavior in gastropods might have been conserved for hundred millions of years (Gillette, 1991).

The behavioral output of neural circuits in response to environmental stimuli is usually adaptive. It is product of computational activity of the circuits, which, in response to the sensory input and the sensory feedback, can change without changes in genes:

This flexibility in the output of neuronal networks has two evolutionary consequences. First, there is no need to evolve a completely new circuit to produce a new behavior. Second, the fact that a network must play roles in many different behaviors or at different developmental stages may constrain it from being altered because changes in the network that would be advantageous for one behavior might be disastrous for another. (Katz and Harris-Warrick, 1999)

Weak synaptic connections observed in circuitries of jamming avoidance response (JAR) may be remnants of lost ancestral functions of these circuitries (Striedter, 1992; Heiligenberg et al., 1996). In studies on jamming avoidance response in Apteronotus and Eigenmannia it was observed that these species show differences in location and expression of receptor types by neurons of the pacemaker nucleus (Pn).

By varying the relative expression or location of receptor types over the course of evolution the temporal dynamics of specific pacemaker modulations could be easily altered…our findings indicate that relatively small changes in the strength of connections between nuclei as well as in the anatomical, physiological and pharmacological properties of individual neurons can lead to extensive diversification of behaviors in closely related genera. (Heiligenberg et al. 1996)

According to Arbas et al. (1991),

Central pattern generators that retain the flexibility to subserve multiple behavioral functions provide ample raw material for evolutionary change. (Heiligenberg et al., 1996).

Epigenetic changes in the structure and properties of neural circuits and central pattern generators are the only changes necessary for inducing differentiation of electrogenic and electrosensory properties of sympatric populations, which enable their reproductive isolation and the beginning of the speciation process in sympatry.

......................................................................

Neuro-hormonal Correlates of Mate Choice and Mate Recognition System

.....................................................................

Brain, Behavior and Evolution

Animal behavior is product of the computational and motor activity of specific circuits and is closely related with cognitive functions (learning, memory, decision-making, etc.) of the brain. The idea that a change in behavior is the first stage in the process of speciation is neither new nor consistently rejected (West-Eberhard, 1989; Price et al., 2003). As early as 1963, Ernst Mayr, would admit the role of behavior in initiating the process of speciation:

A shift into a new niche or adaptive zone is, almost without exception, initiated by a change in behavior. The other adaptations to the new niche, particularly the structural ones, are acquired secondarily. With habitat and food selection - behavioral phenomena - playing a major role in the shift into new adaptive zones, the importance of behavior in initiating new evolutionary events is self-evident. (Mayr, 1963a)

By moving to another area or by expanding their range animals may be subject to the speciational effects of geographic isolation and gene drift.

Being the most plastic of phenotypic characters in metazoans, as a rule, animal behavior precedes the morphological and physiological adaptation:

The impression that behavior takes the lead in evolution is commonplace…Adaptive behavioral plasticity is expected to evolve more readily than does adaptive morphological plasticity, because of the greater abundance of potential cues for regulating the expression of an immediate (behavioral) adaptive response… The evolution of an adaptive plastic morphological response in animals, by contrast, requires a cue operating early enough in ontogeny to trigger the development of the appropriate morphology….Behavior during development can extensively influence morphology. (West-Eberhard, 1989)

In line with this idea, the evolutionary change starts with neural processing of the external/internal stimuli from which the new adaptive behavior arises. However, from a conventional view, it is difficult to see how this new behavior and the neural mechanism that produces it may be related to the future change in morphology.

While changes in behavior allow animals to rapidly adapt to the environment, such changes may also enable them to enter new niches or adaptive zones and expand their geographical range. In doing so, animals are subject to new selection pressures that facilitate divergent evolution and speciation processes (West-Eberhard, 1989; Wcislo, 1989). Indeed, experiments on Drosophila kept under total darkness for 800 generations led to a number of inherited changes in behavioral (phototaxis, olfaction, daily rhythms) and morphological traits (Wcislo, 1989). However, idea has been expressed that behavioral plasticity and innovativeness, by adapting metazoans to the changed conditions in the environment, can both enhance or inhibit the rate of evolutionary change (Price et al., 2003; Paenke et al., 2007).

If animal behavior has such an important role in metazoan evolution, then the brain as the determinant of animal behavior has to be crucially involved in their evolution. Hence, logically, it may be predicted that a correlation should exist between the evolution of the brain and metazoan evolution in general. This prediction is at the core of the hypothesis of the behavioral drive.

Behavioral Drive or Brain Size-Environmental Change (BS-EC) Hypothesis

C. Darwin may have been the first to observe a positive correlation between the position of the species in the tree of life and the rate of evolutionary change:

The productions of the land seem to change at a quicker rate than those of the sea, of which a striking instance has lately been observed in Switzerland. There is some reason to believe that organisms, considered high in the scale of nature, change more quickly than those that are low: though there are exceptions to this rule. (Darwin, 1859h)

Looking upon the anatomical evolution of birds in the context of the low accumulation of point mutations causing amino acid substitutions in the class of Aves and the time elapsing since their first appearance, Wyles et al. (1983) came to the conclusion that morphological evolution of birds has been much faster than anatomical evolution of other vertebrate classes except for mammals. To explain their observation on the unusually rapid evolution of bird morphology they proposed the hypothesis of “behavioral drive”, according to which the relatively higher proportions of brains in relation to body weight in birds and mammals, in comparison with fish, reptiles and amphibians, makes these classes behaviorally more innovative. Behavioral drive hypothesis predicts that large-brained animals, being behaviorally more flexible and innovative, have comparatively higher evolutionary rates and will have more clades. Accordingly, acquisition of new behaviors allows them to extend the range of their habitat and adopt new habitats, finding thus themselves under new evolutionary pressures.

The brain is still bigger in songbirds and primates, whose rates of anatomical evolution are especially high. The individuals in these two groups are notable also for their mobility and ability to communicate over long distances, both visually and vocally. The genus Homo is at the top of the scale in regard to rate of anatomical evolution, relative brain size, and the capacity for rapid behavioral shifts throughout large populations. From the strength of the correlation (r>0.97) between the two sets of values in the table 20.3, we conclude that most of the variation in rate of anatomical evolution among vertebrates is associated with, and thus may be due to, variation in relative brain size. (Wyles et al., 1983)

Neurocognitive sympatric speciation in insects may be determined by changes in sexual behavior (sexual neuro-cognitive sympatric speciation) or by changes in nonsexual behaviors (non-sexual neuro-cognitive sympatric speciation) of groups of individuals within a population.

Among insects, the genus Drosophila offers some of the most spectacular examples of sympatric speciation. Out of the total of 1,500 Drosophila species described throughout the world, almost 500 have sympatrically evolved in Hawaii islands within about one to several million years since these islands emerged. This explosive speciation took place in sympatry as a result of changes in song patterns and courtship patterns in Drosophila populations.

1. Zimbabwe (Z) lines of Drosophila melanogaster in East Africa represent one of the most authentic cases of incipient speciation in nature. Females of these lines, under natural conditions, do not mate with males from cosmopolitan (from other continents) M (from melanogaster) lines of D. melanogaster. In contrast, females of the cosmopolitan races mate with males of their own and males of Zimbabwe populations (Takahashi and Ting, 2004). Their reciprocal crosses display almost no reproductive isolation and produce viable fertile hybrids (Hollocher et al., 1997). The viability and fertility of their hybrids in F₁ and F₂and the polymorphism of sexual behavior in males and females of Zimbabwe populations suggest that they may be in an initial stage of speciation determined by changes in the mating preferences of female African flies (Wu, C. et al., 1995).

The material basis of the changed mate preference of the Z line is still unknown. A group of investigators reported that in experiments of substitutions of ds2 (desaturase-2) locus with ds2^Z allele of Z flies a decline in cold resistance and an increase of starvation resistance occurs in M (cosmopolitan line) flies and the substitution may be responsible for the sexual isolation between them (Greenberg et al., 2003). Investigators believe that the reproductive isolation is a pleiotropic effect of the ds2 on the cuticular hydrocarbon profile of the flies. Repeating these experiments, another group of researchers failed to confirm any effect of ds2 on cold resistance and starvation resistance (Coyne and Elwyn, 2006a). As for the influence of cuticular carbons on the sexual behavior of M and Z populations repeated experiments also show no indication of such an influence (Takahashi and Ting, 2004; Coyne and Elwyn, 2006a). Most recently it is suggested that the reason for the failure to reproduce results in repeated experiments may have been the fact that investigators had not properly controlled the genetic background according to a protocol (Greenberg et al., 2006b), but that protocol was not used in the original experiments as well (Coyne and Elwyn, 2006b).

Despite the controversial evidence on the cause of reproductive isolation between Drosophila M and Z lines, by consensus it is admitted that these lines represent two reproductively isolated populations in the stage of incipient species occurring in sympatry.

In most cases of sympatric speciation as a result of neurocognitive reproductive isolation in Drosophila species, genetic differences are of the order of the genetic differences observed within populations of Drosophila species. This, as well as the fact that no genes for reproductive isolation have been possible to be identified so far in the Animal kingdom in general, casts doubt on the neoDarwinian prediction on a possible involvement of genes in reproductive isolation in the process of sympatric speciation.

2. North American field crickets Gryllus texensis and Gryllus rubens are two cryptic sister species living in sympatry and allopatry in an area extending from south central to the south eastern part of the North American continent, the first being dominant species in the western part of the area and the second on the eastern part. Males of both species are morphologically indistinguishable, with the pulse rate of calling song being the only known distinctive character on which their reproductive isolation is based. Although viable and fertile hybrids between two species may be produced in laboratory, hybridization in nature does not take place or occurs very rarely and cannot succeed because the intermediate variant of the hybrid song does not attract males of either species. Even if the reproductive isolation is not absolute and a low degree of gene flow between populations of the two species does occur, prezygotic mechanisms are much more important than postzyotic mechanisms. The neoDarwinian prediction of hybrid infertility or inviability as well as character displacement (predicted by the reinforcement model) has not been shown to exist in nature (Gray and Cade, 2000).

Based on experimental results, investigators conclude against the allopatric origin of these species. Quoting them in extenso:

Thus, of the two major predictions of reinforcement models, (i) postzygotic isolation equal to or exceeding prezygotic isolationin allopatry and (ii) prezygotic isolation mechanisms enhancedin sympatry, neither is supported by the currently available data.Of the predictions of speciation by sexual selection models (i) preeminence of prezygotic mechanisms in both allopatry and sympatry,(ii) no character displacement in male signals or female responses,and (iii) a positive genetic correlation between male signalsand female responses, all are supported by present knowledge.We note that our finding of a positive genetic correlation meansthat runaway sexual selection could have occurred but does not demonstrate that it did occur; moreover, the reinforcement modelin no way precludes a genetic correlation, and not all modelsof speciation by sexual selection require runaway. Nonetheless,our study is among the most complete in providing empirical evidencefavoring the sexual selection model. The evidence presented here should be taken as encouragementof research examining speciation by sexual selection. Our resultsand discussion are not intended to be in any way critical of thewidely accepted allopatric model with no involvement of sexualselection (i.e., vicariance) nor of the allopatric model invoking reinforcement. Instead, we are willing to suppose that becausethere are many millions of animal species on earth, there mayhave been more than one mechanism ofspeciation. (Gray and Cade, 2000)

Changes in non-sexual behaviors in insects may determine a tendency for shifting to new hosts for mating and living. Such shifts isolate reproductively particular groups from the rest of the population and lead to rapid formation of new races and species. Due to the fact that speciation in such cases takes place under conditions of spatial isolation (different tree species) this often is considered to be an ecological mechanism of reproductive isolation.

The ecological nature of the reproductive isolation is hard to argue for the mechanism of the reproductive isolation in these cases is ultimately determined not externally by the host but by an intrinsically neurocognitively determined preference and a behavioral shift to the new host.

Others consider such cases of host shifting to be a form of reproductive isolation in allopatry. Again, the deepest causal basis for the switch of insect populations from the ancestral host plant to a new host plant is not a spatial separation or geographic isolation but a new behavior that is neurobiologically determined by processing of sensory (visual, olfactory, etc.) cues in neural circuits.

1. Rhagoletis pomonella is an apple fruit fly. It has evolved from Rhagoletis hawthorn feeding flies, very recently, in less than 2 centuries, after the introduction of apple trees in America. The shift of R. pomonella from its original hawthorn host to apples was first described, more than one century ago, by Walsh in 1867, as a probable case of sympatric speciation but it was Bush who argued that apple and hawthorn flies and four or more other groups represent a species complex that has sympatrically radiated via host-plant shifts (Linn et al., 2003).

The reproductive isolation of R. pomonella, from the original hawthorn feeding race is based not on genetic changes but on a neurally determined increased preference for apple fruit volatile compounds and decreased preference for hawthorn volatiles. This epigenetic switch of preferences from hawthorn to apple volatile compounds made it possible the evolution of a new sibling species from the original stock of hawthorn flies within 150 years (Linn, Jr. et al., 2003).

Ripening apple fruits emanate volatile compounds (with butyl hexanoate being the key olfactory signal), which serve Rhagoletis apple flies as a long-range cue for detecting apple trees, whereas visual cues become dominant at shorter distances for localizing apple fruits. In a process of specialization for using different host plants, by evolving new host preferences, one (or more) ancestral population of R. pomonella entered a stage of reproductive isolation from the rest of populations. Over time the isolated population succeeded in adapting its life history to the phenology of the host plants, leading thus to present time status of sibling species.

It is very intriguing the fact that within the extremely short time of existence as a separate species, R. pomonella has evolved two distinct morphs: one with longer ovipositors when reared on hawthorn and one with shorter ovipositors when reared on apples (Bush, 1969). The apple race of R. pomonella exhibits a high host fidelity that strongly restricts interbreeding with the hawthorn race to as little as 4-6% per generation (Feder et al. 1994, 1998).

Experimental evidence shows that at least some phytophagous insects (R. pomonella and Cornus florida) display aversive behavior and avoid the odors of nonnatal plants (Forbes et al., 2005). So, e.g., populations of apple, hawthorn and dogwood flies show aversion to volatiles of nonnatal hosts, what suggests that their ancestral state has been avoidance, and host shifting during evolution is associated not only with preference for the new host plant but also with aversion to other plants.

Now Rhagoletis pomonella is a group of six sibling species believed to have evolved in sympatry, very recently in the evolutionary time scale, from a common ancestor, via reproductive isolation determined by adapting to living on different host plants. Each of these species feeds on a particular nonoverlapping species of plants, using volatile compounds as olfactory cues for recognizing its respective host plant.

2. In contrast, Rhagoletis suavis comprises a group of an identical number (six) of species, each of them specialized on a distinct host walnut species of the genus Juglans only, without host shifts beyond Juglans, and has evolved allopatrically over time only in a number of walnut species in North America.

Attempts have been made to explain the different rates of speciation in these two groups of fruit fly species by assuming that different mechanisms of speciation have been in action in each of these species complexes. The rapid sympatric evolution of R.. pomonella is related to the ability of the ancestral species for host shifting (Dres and Mallet, 2002). For Rhagoletis suavis it is hypothesized that adaptation to walnut husks, which contain juglone, a very potent phenolic toxin, may have required specializations that have prevented any host shift beyond the Juglans genus, a phenomenon that also occurred in Drosophila pacea, which, for normal larval development, requires a sterol found in its senita cactus host plant (Bush and Smith, 1998 and references therein). Coevolution of suavis species and their respective walnut hosts is excluded as a possibility because all the modern species of Juglans genus arose ~40 million years ago, much earlier than an estimated 2-5 million years since the appearance of suavis species (Bush and Smith, 1998).

In contrast to the evolution of suavis species, evolution of 6 sibling species of Rhagoletis pomonella took place within 2-3 last centuries.

3. The larch budmoth, Zeiraphera diniana (Lepidoptera: Tortricidae), has two host races feeding respectively on the European larch (Larix decidua) and Cembran pine (Pinus cembra) in mountains of Europe. During outbreaks, the larch race undertakes long-range migrations. Both races interbreed in captivity and may also hybridize in nature. These races differ in a number of traits, but the most important for mate choice and reproductive isolation between them are differences they show in mating signals, pheromones released by females and in the male response to them. The host plant of the larch budmoth, Zeiraphera diniana, serves as an indirect cue for mate finding and assortative mating. Males respond to the female pheromone but the response is stronger when the call comes from a tree of their own or from a neighbourhoood consisting of such trees than when coming from nonspecific host plants (Emelianov et al., 2003).

4. A stem-galling tephritid fly, Eurosta solidaginis, is a species with two host races living on two goldenrod hosts of the genus Solidago (S. altissima and S. gigantea) growing in the same habitat. Both races are electrophoretically distinct only at the level of host races. Their hybrid offspring are viable and fertile. Females inject an egg into an unexpanded leaf of the host plant.

It is noteworthy that both sympatric goldenrod species show greater genetic distance than those living in the same hosts in different geographic areas. Both species differ in the time of emergence but difference in phenology of emergence cannot sustain the observed reproductive isolation. They show strong mating and oviposition preferences and these seem to be the main factors for their reproductive isolation. Populations associated with one host emerge 10-14 days earlier than the populations of the other host, thus increasing chances of maintaining the reproductive isolation of two host races. Investigators assessed that the two races are in an intermediate stage of the sympatric speciation (Craig et al., 1993).

5. It may be argued that because of the incomplete prevention of interbreeding between races in host plants, populations may not attain complete reproductive isolation. At least in a number of well studied cases, the nature has overcome this difficulty by complementing host plant specialization with an additional assortation mechanism, for facilitating sympatric speciation. Such is the case, e.g., with the European corn borer, Ostrinia nubilalis (Lepidoptera: Crambidae) introduced from the American continent almost 5 centuries ago. Now this species consists of two sympatric species/host races, one feeding on maize (Zea mays) and the other on mugwort (Artemisia vulgaris). No genetic differences exist between their mitochondrial DNAs and their pheromone binding proteins are identical. Both host races show diverging male mating preferences: the eastern population with males attracted to females with trans-11-tetradecenyl acetate as prevailing pheromone and western population males attracted to females with the isomere cis-11-tetradecenyl acetate as main component of the pheromone mix. The two sibling species do not interbreed under natural conditions (Bush, 1975). Very low level of gene flow and hybridization between the two races seems to take place in nature. Both the difference in the time of emergence (the maize race emerges 10 days later than the mugwort race), and different pheromones released by females of both races allow males to recognize females of their own race, thus promoting their reproductive isolation (Thomas et al., 2003).

6. Two closely related Australian fruit fly species, members of the family Drosophilidae, Scaptodrosophila hibisci and S. aclinata, courtship and mate on flowers of different Hibiscus species (rosemallow, flowering plants of the family Malvaceae) hosts.

Two pairs of these sympatric S. hibisci populations breeding on different Hibiscus species show genetic differences of the same magnitude with allopatric populations of S. hibisci, suggesting that these sympatric populations are in the process of incipient speciation (Barker, 2005).

7. A monophyletic group of forest-dwelling Laupala crickets in Hawaii Island experienced a speciation rate of 4.17 species per million year, which is 26 times faster than the average speciation rate of speciation of arthropods and is second faster rate described after the one observed in evolution of cichlid fish of the East African lakes. The explosive speciation of these crickets is still going on. It is based on evolution of female preference to mate conspecific male crickets whose courtship song’s pulse rate is characteristic for each species. Curtship songs, thus, enable crickets to avoid interbreeding with heterospecific crickets of the Laupala group. It is believed that this divergence in mate preferences and in the sexual behavior is the cause of the rapid evolution of crickets in this group (Mendelson and Shaw, 2005).

8. Aphid species Dysaphis anthrisci majkopica, cannot survive and reproduce on the plant, Chaerophylum maculatum. However, after experimentally rearing it for 4 to 8 generations on that plant, a proportion of aphids displayed adaptive changes in morphology (rostrum and body size) and adapted to live on the hostile plant (Shaposhnikov, 1965).

The neoDarwinian theory for a long time has considered sympatric speciation as a highly unlikely mode of speciation. According to the paradigm, the first step in the process of speciation is accumulation of differences in the gene pool of two populations, as a result of the prevention of the gene flow between two populations that are geographically isolated from each other. Over time, it is argued, the accumulation of genetic changes eventually leads to genetic incompatibility and postzygotic reproductive isolation of populations. The majority of biologists dealing with the problem of speciation still regard the gradual accumulation of favorable genetic variation as a necessary condition of speciation:

Mating discrimination, sterility and inviability of hybrids, all increase gradually with time as estimated by genetic distance. This confirms the impression gained from the study of geographical variation that speciation (at least in Drosophila) is not a sudden event. (Maynard Smith, 1989)

2. long periods of time for accumulation of the extremely rare “useful mutations” in relevant genes.

Evolution of Rhagoletis pomona complex from the original hawthorn Rhagoletis spp. within an evolutionary instant of one to two centuries and the fact that this evolution took place in sympatry refutes both neoDarwinian predictions. Attempts have been made, and continue to be made, for explaining the contradiction between the neoDarwinian paradigm and the facts on sympatric speciation. Feder et al. (2005) have developed the “reticulate” model of speciation of Rhagoletis pomonella sibling species complex. Contrary to the previous evidence, they speculate in length that the case of Rhagoletis speciation may involve allopatry, which may have produced the necessary genetic variation for later host shifting (Feder et al. 2005):

In the case of R. pomonella, the relationship involves a likely sequence of geographic isolation, life history adaptation, secondary contact, differential introgression, inversion clines, and sympatric host shifts. The evolution of reinforcement can be viewed in analogous manner, involving non-host-related traits affecting prezygotic isolation rather than ecological adaptation per se. Also, there is no reason to presume that host-related differences that originated in sympatry cannot be solidified by periods of geographic isolation between host-associated populations. (Feder et al., 2005)

Lack of substantiating evidence that this long sequence of hypothetic events might have happened devoids the hypothesis of any explanatory power. Besides, the time required for the presumed sequence of events required by the reticulate model is incompatible with the extremely short period of time within which the Rhagoletis pomonella species complex (only one century from the introduction of apples from Europe to the first record of the new host species) evolved. If none of the events in the assumed sequence of events in this reticulate scenario has been substantiated in the case of speciation of the Rhagoletis group, it is hard to understand why should one resort to such a highly speculative model when experimental evidence clearly suggests that the insect has shifted the host under conditions of sympatry.

While the theoretical possibility of the reticulate scenario cannot be rejected, it is the burden of the authors to argue that this scenario is not a pure conjecture. Until this is done Occam’s razor suggests that speciation occurred in sympatry by host shifting.

Bush and Smith see the evolution of the Rhagoletis species complex as a case of ecological speciation in sympatry (Bush and Smith, 1998). They believe that the ecological speciation resulted from a sudden shift of the insect’s host preference. But they do not deal with the origin of the new behavior for host shifting, which implies two neurocognitive phenomena, the preference for the new host and recognition of the host. Essentially, that shift, as already pointed out, is a behavioral shift that results from changes in properties of the neural circuits determining that change in insect’s behavior.

Genes at relatively few key loci affecting mate recognition, habitat choice and fitness in alternative habitats can be sufficient to initiate the process of ecological race formation and speciation. (Bush and Smith, 1998)

This assertion raises serious objections. Firstly, neither population genetics knowledge nor empirical evidence would support the idea that several sibling species of Rhagoletis pomonella and necessary changes in genes could evolve from a common ancestral hawthorn-infesting species within an extraordinarily short period of time of ~150 years/generations.

Secondly, no changes have been identified in any of the “relatively few key loci” and electrophoretic analysis has only identified several differences in not so key allozyme loci (Berlocher, 1999).

Thirdly, but not less importantly, host preferences and especially mate preferences are determined not by genes or their products, but are epigenetically determined by neural-cognitive processes taking place in the CNS.

However, Bush and Smith come very close to accepting the role of the neural circuits in the insect brain in determining the host shift, reproductive isolation and speciation of the Rhagoletis species complex:

Host and mate choice are therefore tightly linked, a common feature in the mating behavior of many parasites and other habitat specific organisms. Host selection is mediated by both visual and olfactory cues. Ultimately selection of a host for mating and oviposition is determined by specific chemical cues emanating from the host plant and their fruits. Changes in host perception can have a direct effect on mate choice. (Bush and Smith, 1998)

Needless to say that “changes in host perception” are not related to any changes in genes but are neurally, epigenetically determined in neural circuits.

We know, in principle at least, how these host cues are perceived: first, the olfactory signals emanated from the prospect host are received by the insect olfactory receptor neurons (ORNs), transformed into neural (electrical) signals that are transmitted for processing in the CNS, first to the antennal lobe then to the mushroom body. The output of the neural processing of the olfactory stimuli emanating from the apple trees, represents a neurally determined attraction that manifests itself in the form of the observed tendency of these insects to migrate to apple trees and use these trees as their hosts for mating and oviposition. (see for details Neural Reception and Processing of Olfactory Signals in the Mate Recognition System earlier in this chapter). Thus, the change in perception and the shift in the host choice behavior has been result of change in the processing of olfactory signals emanating from the apple trees, a shift that was based on a change in computational properties of neural circuits of the original stock of the hawthorn Rhagoletis species.

It is a general neurobiological knowledge that computational properties of circuits, where olfactory and visual cues are perceived, are not functions of genes. We do not know of any gene that could computate external information for perceiving sensory cues and triggering behavioral response at the organismic level (in fact we cannot even imagine how this could ever happen). With no genes involved in the behavioral shift to new hosts, it is not surprising that no reasonable neoDarwinian explanation of sympatric speciation by host shifting has ever been presented.

The key element for initiating host shift-related sympatric speciation is the emergence of the neurally-determined preference for a new host, which leads to adoption of the new plant as a natural host.

The shift to a new host plant of a group of individuals or a population may lead to its reproductive isolation from the rest of the original population. Once this occurs, then phenotypic changes of different kinds will unavoidably take place in the process of the evolutionary adaptation to the new host. That this adaptation may occur rapidly is suggested by the fact that the larvae of parasitic insects from several genera can develop in laboratory on unnatural plant hosts (Dres et al., 2002).

The apple race R. pomonella evolved its preference for a blend of apple fruit volatiles within one and half a century (Linn et al., 2003). It is demonstrated that hybrid flies between various Rhagoletis sibling species, lose the ability to respond to the volatile substances emanated by the host tree in doses that elicit that response in parents and, consequently cannot recognize or distinguish them from volatiles of other plants. Investigators have argued that this is not a result of any developmental imbalance related to hybridization because:

If the altered behavior of F₁is a byproduct of a general developmental imbalance, then hybrids would be expected to exhibit various other phenotypic abnormalities with detrimental fitness consequences, in addition to their reduced olfactory response. (Linn, C.E. et al. 2004)

They believe that this may result from a rise in the threshold of the behavioral response by olfactory circuits in hybrid flies (Linn et al., 2004). In view

of the fact that flies recognize their host trees based on processing of olfactory and visual cues as well as on learning, it is logical to conclude that in F₁ something has changed in the properties of the neural circuits where the processing of these signals takes place. Indeed, there is experimental evidence substantiating this prediction. Olfactory receptor neuron (ORN) responses of F₁ hybrids of Rhagoletis races show ORN response profiles that are different from those of each of the parents, a fact that may contribute to the reproductive isolation by decreasing mating chances of hybrids (Olsson et al., 2006a).

Given the absence of evidence on changes in relevant genes in these cases of incipient speciation in populations of R. pomonella, it has been hypothesized that changes in host preferences may be related to changes in the number or types of neurons that receive olfactory signals. Empirical evidence clearly shows that, contrary to the hypothesis, differences in host preference among Rhagoletis populations are not related to any alterations in the number or class of receptor neurons responding to host volatiles (Olsson et al., 2006b). There is evidence that generally these receptor neurons are resistant to evolutionary change. Studies in nine sibling species of the Drosophila melanogaster subgroup have shown that ligand affinity, and action potential amplitude has been conserved to a remarkable degree over millions of years across changing environments (Stensmyr et al., 2003).

Empirical evidence also shows that the only observable differences in the case of apple, hawthorn, and flowering dogwood-origin populations of R. pomonella is only an epigenetic change in the sensitivity of ORNs (olfactory receptor neurons) and in the firing patterns of these neurons, suggesting that these changes may be the cause of the divergent host preferences and host plant shifts of these populations (Olsson et al., 2006c). Hence, there is reason to believe that host shifts in Rhagoletis spp. are consequences of nongenetic changes in the structure (and properties related to that structure) of the neural circuits that process the host plant volatile substances. We have already seen that neurocognitive processes on which insect preferences are based are plastic and that plasticity may be a source of the evolutionary rapid host plant shifts:

Each fly taxon has a similar capacity for detecting all volatiles, regardless of host species. Thus, prior to host shifts, fly populations appear to have already possessed the ability to detect novel host volatiles and did not evolve new odor receptors for these volatiles. Instead, switched preferences for novel fruit volatiles and avoidance of ancestral hosts may have been established through alterations in the central processing centers of the brain (mainly the antennal lobe and protocerebrum) or modifications in glomerular innervation and connectivity to higher processing centers. Thus, many insects may have an innate chemosensory potential for rapidly changing their host affiliation that, when coupled with appropriate variation in host-related performance (survivorship), could help trigger rapid race formation and speciation. (Dambroski et al., 2005)

It may be safely said that host plant shifts are not related to any changes in genes but only to a changed behavior and preference for the new host. This changed preference is an epigenetic affect of “alterations in the central processing centers of the brain”.

Two populations of the Vogelkop bowerbirds (Amblyornis inornatus) build bowers and decorate them in strikingly different ways. Females from each of the two populations prefer bowers and decorations of males from their own population, establishing thus a behaviorally conditioned reproductive isolation in this initial stage of the speciation process that is actually taking place in two populations of the same species (Uy and Borgia, 2000). Individuals of both populations are morphologically identical, show only little genetic differences of recent origin, but there is no evidence on any differences in genes related to the bower-building behavior.

Rare cases of the “reverse” process of two metazoan species merging into one are observed in nature. This seems to occur exclusively between sympatric species. Kraak and coll. (2001) reported an increase in the proportion of hybrids between a pair of “benthic” (bottom-dwelling) and “limnetic” (top-dwelling) species of the three-spined sticklebsacks Gasterosteus aculeatus, in the Lake Enos, southeastern Vancouver Island, Canada, from 1% to 12-17% within 7-15 years and concluded that the benthic and limnetic forms of the fish were collapsing, and a “new” species was evolving from their hybridization (Kraak et al., 2001). Their conclusion on the collapse of the “benthic” and “limnetic” species and evolution of a “new” hybrid species in the lake were corroborated later by another group (Taylor et al., 2006). The species’ breakdown seems to be asymmetrical in the meaning that the limnetic form is introgressed into the benthic form (Gow et al., 2006).

Investigators have identified the introduction to the lake of the American signal crayfish (Pascifasticus lenisculus) in late 80s as the agent of the detrimental environmental effects on aquatic plants and faunas that caused the collapse of the pair of species. Movements of the crayfish increased turbidity of the water, making thus difficult for three-spined stickleback pair species identification of visual cues necessary for performing their elaborate courtship ritual and mating discrimination (Taylor et al 2006).

Fom a theoretical point of view, it would be predicted that if the singing patterns of two species would converge, these species would tend to hybridize because they would find it impossible to distinguish between the conspecific and heterospecific individuals of the opposite sex. This prediction has been proven to be true in a number of cases, including the case of two species of passerine birds, the collared flycatcher, Ficedula albicollis, and the pied flycatcher, Ficedula hypoleuca. Under sympatric conditions, the pied flycatchers learn, and include in their song repertoire, parts of the song of the collared flycatcher. Males of the collared flycatcher respond to the mixed song similarly to their species specific song, while their females pair with male pied flycatchers only if they sing the mixed song. The inclusion of the heterospecific song in the song repertoire of the male pied flycatchers is the reason why in sympatry, as it was predicted, ~ 20% of male flycatchers pair (hybridize) with female collared flycatchers (Qvarnström et al., 2006).

A neoDarwinian explanation of the phenomenon of species diffusion would hardly be applicable in the cases of species diffusion for the sterility and low viability of hybrids predicted by the paradigm would lead not to fusion but to gradual extinction of both species.