THE EVOLUTION OF APOSEMATISM IN MARINE GASTROPODS.

Abstract

In a recent review of the literature on marine gastropods, Rosenberg (1989) concludes that in this group,.. . as in the lepidopterans, . . . individual is the predominant mode of evolution of warning coloration. Rosenberg's analysis raises interesting issues concerning the evolution of aposematism generally, and in marine gastropods in particular. We tackle two issues here. (1) What kind of evidence actually demonstrates aposematism? We are not convinced that sufficient evidence has been reported for marine gastropods. (2) What are the selective conditions that can favor aposematism? We argue that the reliance on the distinction between kin and individual characteristic of this (Rosenberg, 1989), and many other studies (Engen et al., 1986; Leimar et al., 1986; Roper and Redston, 1987; Grober, 1988), is misleading. We argue that individual is never likely to be a sufficient explanation for the evolution of warning coloration in any species, because selection (here, the favoring of like-phenotypes) is always likely to play some role, at least when warning colors reach high frequency. Neither Rosenberg (1989) nor his sources provide direct evidence that coloration functions aposematically in the seven gastropod species considered. The adaptive theory of warning coloration (Wallace, 1867) posits that conspicuousness evolves in unpalatable forms because it is a more effective warning signal than the alternative, crypsis. Except where aversion is innate, the critical prediction is that colors must be especially potent in promoting the acquisition and maintenance of learned aversions (Gittleman and Harvey, 1980; Harvey and Paxton, 1981; Mallet and Singer, 1987; Roper and Redston, 1987; C-uilford, 1988; Guilford and Cuthill, 1989), or their effective use by predators (Guilford, 1990). The experimental studies cited by Rosenberg (1989; Table 1) demonstrate that fish find these conspicuous gastropods unpalatable, but not that fish are more likely to reject (unpalatable) conspicuous prey than (unpalatable) cryptic prey. Though unfortunate, it is not unusual to assume function without these critical data (Guilford, 1990). Circumstantial evidence for aposematism comes from the apparent correlation between unpalatability and conspicuousness, at least in insects (Edmunds, 1974), birds (Cott, 1947; Baker and Parker, 1979), amphibia (Cott, 1940), and reptiles (Greene, 1988). On its own such evidence is weak because it does not eliminate cases in which unpalatable prey are conspicuous merely because they can afford not to be cryptic, or cases in which conspicuous prey are secondarily unpalatable because they are otherwise vulnerable. In neither case need conspicuous coloration have been selectively favored for an function (Guilford, 1988). However, for marine gastropods it is not yet clear that even this weak comparative evidence exists. Many ovulids have brightly colored mantles, but most are apparently cryptic against their coral hosts (Rosenberg, 1989). And one study cited by Rosenberg (Thompson, 1960) concludes that conspicuous gastropods are no more or less acceptable to fish than the cryptically colored species. All those tested were distasteful, all were rejected. Thompson (1960) cautions against the terms aposematic and coloration, and rejects this explanation (Cott, 1940) for the bright colors of opisthobranch mantles. We can deduce that fish reject noxious gastropods, some of which are conspicuous, but not that the conspicuousness of some species evolved to advertise unpalatability. Even though the controlled experiments still need to be done, aposematism seems a reasonable hypothesis for the very striking pattems of at least some unpalatable gastropods, such as Cyphoma gibbosum and Crenavolva tigris (Rosenberg, pers. comm.). How, then, might aposematism evolve? It is usually assumed that in marine gastropods (Rosenberg, 1989), as in other taxa (Engen et al., 1986; Leimar et al., 1986; Roper and Redston, 1987; Grober, 1988), evolution must proceed either by individual or by kin selection. We argue that neither provides a complete explanation. Consider first the evolution of waming coloration in a species with family grouped population structure. It has long been recognized that clustering of similar phenotypes could favor the spread of warning signals that advertise unpalatability, through generalization by a locally resident predator (Fisher, 1930; Harvey et al., 1982; Mallet and Singer, 1987). However, while similar phenotypes are more likely to cooccur in family groups, the benefits of predator avoidance go predominantly to individuals that share the same phenotype. Dissimilar kin do not receive equivalent protection from predators. Thus, even with family grouping the evolution of waming colors does not proceed by classical kin (Hamilton, 1964; Maynard Smith, 1964). Where the benefits from the death or injury of one wamingly colored individual go preferentially towards other wamingly colored individuals, standard relatedness calculations will not predict inclusive fitness consequences correctly because is synergistic (Queller, 1984; Guilford, 1988; though