Evolutionary aspects of acoustic communication in Ribautodelphax planthoppers (Homoptera, Delphacidae)

Delphacidae (Homoptera), commonly referred to as planthoppers, are herbivores, which usually feed on grasses and sedges. During sexual behaviour males and females communicate by exchanging low-frequency vibrational signals, which are transmitted through the substrate, normally the host plant. This thesis deals with the acoustic behaviour of one planthopper genus, Ribautodelphax, where both male and females have been found to produce species-specific calls, which differ between species in temporal parameters. As in other planthoppers, the acoustic signals of males and females are rather different. Male calls are more complicated, and consist of at least two structurally different elements, a variable number of 'chirps', followed by a 'buzz' of variable length, hence termed the 'chirp-section' and the 'buzz-section', respectively. The female call consists of a series of simple pulses, which differs between species in interpulse interval length ( IPI ), signal duration, and modulation of pulse repetition rate within the signal.
This study was aimed at answering the following central, interrelated questions:

1. Do planthopper calls have a function in species recognition (sexual isolation) and mate preference?
2. What forces have caused differentiation of acoustic signals?
3. What is the relationship between divergence in acoustic signals and speciation?.

It was confirmed that, during the first phase of the sexual behaviour (the socalled distant calling phase), these calls are especially important in bringing potential mating partners together. Males called first, and mating-receptive females responded acoustically. The male then started searching actively for the female and continued to exchange calls with her, during which the female remained sedentary until the male came in close range. Females only responded to conspecific male calls when they are virgin and old enough. In populations of two closely related species, the development of female responsiveness with age corresponded fairly well with that of insemination levels, which shows acoustic response levels to be good indicators of mating receptiveness.

At close range (courtship in the strict sense), males remained acoustically active, but the female signal length and calling frequency tended to decrease, and some females ceased responding altogether. This suggests that during courtship the male call serves in maintaining and enhancing the female's receptiveness, and that the female call is less important. At this stage of sexual behaviour females appeared to be rather cautious, and usually only allowed copulation after many refusals. Courtships were clearly shorter when a female was confined with two males instead of one. Females seemed not to mate randomly with the available males, leaving open the possibility of some form of sexual selection. In the absence of other obvious cues, it seems possible that females might prefer males on the basis of their acoustic signals, but possibly due to the limited number of observations no trend in preference was found. However, females did not actively choose between males, nor did males behave in any way aggressively towards each other. After both males called initially, usually only one male continued courtship, leaving the possibility that males first assess their relative attractiveness or social status m their calls.

Different closely related Ribautodelphax species performed the same behaviours during courtship, but differed more or less in the frequencies of transitions between behavioural events. A more distantly related species deviated more strongly in transition frequencies, as well as by exhibiting a behavioural element not shared by the other species.

Many combinations of Ribautodelphax species are known to be able to produce viable and fertile interspecific hybrids under no choice conditions. However, when both conspecific and heterospecific partners are available, interspecific matings rarely take place, if at all. After rearing two species together for 10 generations no indication was found of introgression having occurred. Thus, recognition of conspecifics takes place before mating. Most females exposed to playbacks of heterospecific male calls responded about as well as to conspecific calls. In contrast, males were found to approach only playbacks of conspecific female calls, or, in a two- way choice experiment, chose significantly more often for the conspecific call. In an additional experiment males were continously exposed to either a conspecific or a heterospecific female playback call during their development from egg to adult. After this treatment both types of males preferred the conspecific female call over the heterospecific one, but males with experience of the conspecific call did this significantly more often. Males primed with the heterospecific female call performed similarly to acoustically naive males. This shows that recognition of conspecific female signals by males is largely genetic, but can be improved to some extent by previous experience of the conspecific signal, whereas the recognition mechanism is not affected by heterospecific signals. Thus, the acoustic communication between the sexes forms at least part of the specific mate recognition system of these species. Apparently, species recognition in Ribautodelphax results primarily from the male preference for conspecific female calls. This is a surprising result, because females only mate once during their lives, in contrast to males. In the absence of any obvious male parental investment, apart from costs involved in searching, the females would be expected to be more selective.

At close range, most interspecific courtships observed did not result in copulation, which is likely to be a by-product of within-species choosyness by females towards acoustic or non- acoustic performances of males. Such interspecific encounters are unlikely to occur under natural conditions, because at that stage recognition has already taken place.

Artificial bi-directional selection for large and small IPI s in the female call of R. imitans was very successful, resulting in non-overlapping distributions of the character after only five generations. The mean of realized heritability estimates of all selection lines over this period was above 80%, and still above 50% over 10 generations. IPI proved to be a polygenic character, controlled by at least six independently segregating genetic factors. Other female call characters, like signal duration, and modulation of pulse repetition rate within the signal exhibited correlated responses. After the experiment a significant degree of symmetrical assortative mating was found in mate preference tests between males and females from oppositely selected lines, but co-selected males did not show a significant preference for female playback calls with IPI s close to those occurring in their selection lines. Some characters of the chirp-section of the male call also appeared to exhibit a correlated change, suggesting that male and female call characters do not evolve independently. It seems possible that the assortative mating among individuals of the selection lines is due to female preference for the changed male call parameters, rather than to the preference of males for changed IPI s in the female calls.

The genetic control of male call characters in R. imitans was studied by father-sons regression. Heritability estimates of the chirp-section characters were statistically significant (0.44-0.54), in contrast to those of the buzz-section (0.09-0.28). Phenotypic, genetic, and environmental correlations calculated among male call characters suggest that chirp-section and buzz-section characters vary independently. One chirp-section character, number of chirps, appeared to be influenced by sex-linked loci. This means that the heritability estimate obtained by father-sons regression probably underestimates the true heritability of this character.

The possibility that the calls have evolved as adaptations to prevent hybridization (reinforcement) appears to be unlikely, for reasons like the apparent genetic plasticity of call characters, the observation that females inseminated by heterospecific males produce both viable and fertile offspring, and the fact that these species live ecologically isolated. It seems more probable that the calls obtained their species-specificity as the result of selection and chance, e.g. after founder events. Potentially, sexual selection might also have contributed to the differentiation of at least the male call. In view of the genetic correlation between some male and female call characters found in R. imitans, the possibility that change in the call of one sex might affect that of the other cannot be excluded.

The observation that, during distant calling, males are much more selective than females with respect to calls of other species appears to be best explained from the need to to be as efficient as possible in finding a proper conspecific mate. Because Ribautodelphax species are confined to different host plants, the chances of meeting other related species are slim, and selection for precise species recognition is likely to be weak. In order to attract as many as possible males to chose among, it might be sufficient for females to recognize the calls of congenerics, which have basically the same structure. For males precise recognition is likely to be more important. There are usually less mating-receptive females than males present in a population, because females mate only once during their lives, in contrast to males. Tuning to a specific call type would increase the chances of finding as many as possible attractive females, as the fraction of such females (i.e. presumably those with call characteristics close to the population mean), is likely to decrease as the season progresses. In species living syntopically on the same host plant, females might be expected to be at least as selective as males towards heterospecific calls of the other sex, because that would seem to be the most economic and safe way to avoid heterospecific encounters. In two species of the planthopper genus Prokelisia, which share the same hostplant, this indeed seems to be the case.

In populations of two Ribautodelphax species, where diploid males and females live associated with triploid gynogenetic females, a peculiar use of the acoustic communication system in a within species context appears to occur. The triploids occasionally arise spontaneously in diploid populations, and need to mate with diploid males, but produce only identical triploid females. Because of their two-fold reproductive advantage, triploids potentially can outcompete the diploid females, which would also lead to their own demise. However, the ratio diploid: triploid females in the field was reported to be stable over time. A model is suggested explaining this stable coexistence by a dynamic 'armsrace' involving the female calls, enabled by the genetic plasticity of the female call, and driven by the selection pressure on males to prefer female calls deviating from the population mean, thereby avoiding mating with the otherwise indistinguishable triploids. This would also explain the peculiar occurrence of several different female call types within and between populations of the species in which such triploids occur.

In view of the potential effect of acoustic signals in species recognition, the evolution of the acoustic communication system might be the primary force behind speciation in planthoppers. However, a confounding factor in Ribautodelphax is that, although the species studied live potentially sympatrically, each is confined to one particular host plant species, on which they feed and oviposit and therefore are unlikely to meet related species in the field. Hence, call differentiation could have taken place as the result of isolation after a change to a new host. However, comparative evidence from related genera shows that acoustic differentiation can also occur without a host plant shift. It therefore seems inevitable to conclude that the change of the acoustic communication system in allopatry is indeed the main factor in planthopper speciation. Although speciation will be facilitated by a host plant shift, because it obstructs secondary contact, it appears to be no prerequisite. Species of genera living syntopically on the same host appear to have developed more rigid recognition systems than members of genera which are ecologically or geographically isolated, because interspecific inseminations were reported to be extremely rare or non-existent, even under no-choice conditions. In these species the specific mate recognition system has apparently changed sufficiently in allopatry to enable coexistence with congenerics after secondary contact. The process of speciation is viewed as undirected change of the specific mate recognition system in small isolated populations up to the point where other populations are no longer recognized, which is in accordance with Paterson's recognition species concept, except that the mate recognition system, at least some of its components, appears to be less evolutionary stable than envisaged by that author.