Key points are not available for this paper at this time.
We used a maximum likelihood (ML) method to examine hybridization and introgression in mixed populations of black oaks ( Quercus kelloggii and Q. wislizenii var. frutescens ), manzanitas ( Arctostaphylos patula and A. viscida ), and irises ( Iris fulva and I . hexagona ). This method uses codominant molecular markers, such as allozymes and restriction fragment length polymorphisms, to estimate the frequencies of six genealogical classes of individuals (parental species, F 1 and F 2 hybrids, and first generation backcrosses) present in a mixed population. We contrasted these estimates with those from the traditional “direct method” which estimates hybrid frequencies directly from observed frequencies of molecular phenotypes. Both methods indicated that the putative hybrid oaks were all F 1 hybrids. For the manzanitas, only ML estimates could reject the hypothesis that introgression was limited to first generation backcrossing. At least 27% of the individuals were estimated to represent second or later generation introgressants. Analysis of the irises indicated bidirectional localized introgression. These methods, in conjunction with clinal variation in the frequency of marker alleles, indicated that this pattern of introgression probably resulted from asymmetric density‐dependent gene flow by pollen. The ML method provides a more reliable description of the pattern of hybridization and introgression because it accounts for overlap among parental species and hybrid phenotype distributions.
Nason et al. (Wed,) studied this question.