HETEROZYGOTE ADVANTAGE AT THE FRUIT WING LOCUS IN PLECTRITIS CONGESTA (VALERIANACEAE).

Abstract

The genus Plectritis consists of four species of winter annual plants occurring along the Pacific coast of North America and intermittently inland as far as the Rocky Mountains, with an additional disjunct species occurring in Chile. The North American species exhibit a characteristic dimorphism in the fruits, with individual plants producing either winged or wingless fruits (Neilsen, 1949; Dempster, 1958; Morey, 1959, 1962). Populations may be either monomorphic or polymorphic for this fruit wing character. In the two closely related species P. congesta (Lindl.) DC. and P. brachystemon F. & M., the fruit wing phenotype is controlled by two alleles at a single locus, with the allele for winged fruits dominant to the allele for wingless fruits (Ganders et al., 1977a, 1977b). Presumably this is the case in the other species as well. A very similar fruit wing polymorphism also occurs in some species of the related genus Valerianella. In V. ozarkana the polymorphism is also controlled by a single locus with the allele for winged fruits dominant (Ware, 1969). Progeny tests using the fruit polymorphism as a genetic marker have provided estimates of outcrossing rates and genotype frequencies in several natural populations of P. congesta. Comparisons of observed with expected genotype frequencies showed that heterozygote frequencies were higher than expected at the estimated outcrossing rates, and in some cases higher than expected even if the outcrossing rate had been 100% (Ganders et al., 1977b). Ganders et al. (1977b) concluded that heterozygote advantage at the winged fruit locus was responsible for maintaining the polymorphism in populations. Subsequent progeny tests of other populations have produced estimates of outcrossing rate in good agreement with previous studies, and have also indicated excess heterozygotes in populations (Table 1). The method used to estimate expected and observed genotype frequencies, and hence relative fitness of genotypes, was that of Harding (1970). Such fitness estimates obtained by comparing observed and expected genotype frequencies are subject to a number of assumptions and problems. One is that the joint estimate of outcrossing rate and genotype frequencies is for a single year. Since the observed genotype frequencies in a population of annuals are a function of the outcrossing rate the previous year, while the expected genotype frequencies are based on the current year outcrossing rate, the method therefore assumes that outcrossing rate, are essentially constant from year to year. If the outcrossing rate had been higher the previous year, there would be more heterozygotes than expected at the current year outcrossing rate. The observed excess of heterozygotes in P. congesta is unlikely to be the result of fluctuating outcrossing rates for two reasons. First, it seems highly unlikely that either a random or systematic reduction in outcrossing rate would have occurred in all seven populations sampled in three different years (Table 1). Second, in some populations heterozygotes were present at frequencies higher than expected even if outcrossing had been 100% (Ganders et al., 1977b). Demonstrating excess heterozygotes by comparing observed and expected genotype frequencies also assumes that the polymorphism is stable and not transient. Continuing selection against one homo-