TAILSPOTS OF XIPHOPHORUS AND THE EVOLUTION OF CONSPICUOUS POLYMORPHISM.

Abstract

While attention has focused recently on the phenomenon of cryptic allozyme polymorphism, many of the problems of conspicuous morphological variation in animals remain unsolved. Numerous examples of species with genetic variation for color or morphology have been described (Mayr, 1963; Ford, 1975). Although there is general agreement among evolutionists that the morphs are subject to selection, clear demonstrations (Kettlewell, 196 1; Watt, 1968) of the adaptive significance of polymorphic variation are infrequent. Perhaps the most valid generalization about conspicuous polymorphisms is Ford's observation that the segregating alleles always seem to have correlated physiological effects (Ford, 1975, p. 1718). It will be argued in the discussion that such physiological effects are basic to conspicuous polymorphism, and that other correlated phenomena are usually secondary. Fishes of the genus Xiphophorus are well suited for the study of the significance of conspicuous variation because 1) they are highly polymorphic for yellow and red color and melanic patterns (tailspots and bodyspots), 2) they are easy to maintain and breed, and 3) the genetic bases of the polymorphism are known (Kallman, 1975). Of particular interest is the tailspot polymorphism which is controlled by a simple autosomal series of multiple alleles. Pattern alleles are codominant with respect to one another and dominant over the wild-type allele (+) which causes no pattern. In some populations of X. variatus, there are as few as two or three alleles, so the polymorphism is more amenable to analysis than other, more complex, systems. A potential factor selecting for tailspot variability in X. variatus is the variability and unpredictability of the environment (Borowsky, 1978a). During the rainy season, May through September, the rivers and upland arroyos flood. While they continue to flow freely during the first few months of the dry season, by January the flow has abated and the upland streams are reduced to a series of poorly connected pools. Superimposed upon this seasonal cycle of variation is a degree of unpredictability because the pools in which fish get trapped vary considerably in temperature, flow, size and other features. In addition to this spatial variation, there are differences among seasons because the amount of seepage during the dry season depends upon the amount of rainfall during the previous rainy season. The fate of any individual fish, therefore, is unpredictable. Previous work (Borowsky, 1978a) revealed a correlation between tailspot genetic variability and environmental variability. The populations studied contained three alleles +, C (=crescent), and Ct (=cut-crescent). Mean adult male size differed considerably among collections from the same populations and this phenotypic variation was interpreted as reflecting environmental variation. In collections where males were large, crescent males were above average in size while cut-crescent males were below average. In collections where males were small, cut-crescent males were above average in size while crescent males were below average. The mean size of wild-type males did not differ from the mean size of all males in any collection. Therefore, it was hypothesized that the crescent morph is favored under