Microsatellite assessment of multiple paternity in natural populations of a live-bearing ®sh, Gambusia holbrooki

Females of many species mate with multiple males during the course of a single round of reproduction (Davies, 1991; Birkhead & Mùller, 1992). However, secure estimates of the frequency of concurrent multiple paternity in most cases awaited the application of molecular markers (Avise, 1994) that permit documentation of when copulations result in fertilizations such that individual broods are multiply sired. Genetic analyses have documented multiple paternity for a wide variety of animals and plants (e.g. Birdsall & Nash, 1973; Hanken & Sherman, 1981; Grif®ths et al., 1982; Ellstrand, 1984; Levin, 1988; Travis et al., 1990; Parker & Korn®eld, 1996). The analyses are facilitated when the mother is known with certainty, as for embryo broods carried internally by pregnant females in live-bearing animals. The frequency of multiple paternity is important for hypotheses concerning male and female ®tness advantages under alternative mating strategies. For example, mating with multiple males may enhance the probability that a female's genes are found in high-®tness genotypic combinations in her offspring, and/or it may provide a selective advantage in variable habitats by promoting genotypic variety among progeny within a brood (Lomann et al., 1988; Karron & Marshall, 1990). Access to multiple males could result in enhanced ®tness for any females who are sperm limited (Borowsky & Kallmann, 1976). For broods in which offspring interact extensively with one another, frequency-dependent selection might be envisaged (McCauley & O'Donnel, 1984) if multiply fathered broods of half-sibs have a selective advantage over broods of full-sibs (Antonovics & Ellstrand, 1984). Another possibility is that the presence of multiple insemination might not be adaptive either for females or their broods but instead might re ̄ect the outcome of competition among males over reproduction. The frequency of multiple paternity is potentially important also for its impact on population-level phenomena such as heterozygosity and colonization ability. The number of fathers for broods is expected to have a strong in ̄uence on how genetic variation is distributed within species (Chesser & Baker, 1996). Greater sire numbers are expected to increase effective population size and permit retention of higher genetic variation in colonizing species or populations exposed to strong ̄uctuations in abundance.