Much work on sex ratios in dioecious plants has involved the genus Silene. The classic literature indicates that female-biased sex ratios result from excess pollen because pollen tubes from female-determining pollen reach the ovary more quickly, on average, than pollen tubes from male-determining pollen. I proposed the "weak male hypothesis" to account for female-biased ratios from limited amounts of pollen. This hypothesis assumes that male-determining pollen either germinates at a lower rate or experiences greater pollen tube attrition than female-determining pollen. To test this hypothesis, limited numbers of pollen grains were placed at one of three positions along the style. According to the hypothesis, if male-determining pollen is handicapped in germination, seed set and progeny sex ratios, which should all be female biased, should not be affected by where on the style the pollen is placed. If handicapping occurs during pollen tube growth, seed set should decrease and sex ratios should become increasingly female biased as the distance traveled by the pollen tubes increases. Seed set exhibited a nonsignificant decrease as style distance increased; none of the sex ratios was female biased. Therefore, the weak male hypothesis is rejected. In another experiment, pollen from individual anthers was separated into nonoverlapping small and large size classes. Sex ratios resulting from the two size classes did not differ, indicating that there is no sex-related bimodality for pollen size. In two other experiments, excess pollen was placed at one of two positions on the style. Of 48 resulting sex ratios, only six deviated significantly from 1:1; of these six, three were female biased and three were male biased. Style position had a significant effect in one experiment (progeny from distant pollinations contained excess males), but not the other. Sex ratios were not affected by either parent in these experiments. The data indicate that female-biased sex ratios do not necessarily result from excess pollen. Therefore, gametophytic competition cannot be assumed to explain the preponderance of females in natural populations. Reasons for the differences between my results and results of early experiments are discussed.

This study uses habitats that have been naturally fragmented throughout the Holocene to examine the effects of habitat fragmentation on a much longer time scale. To investigate the long-term effects of habitat fragmentation on population genetic structure of Carex bigelowii, Diapensia lapponica , and Minuartia groenlandica, isozyme variation in fragmented patches of alpine tundra was compared to variation in sites from continuous alpine tundra. The questions posed by this study include: Is genetic variation reduced in fragmented habitat? Does gene flow occur between habitat fragments? What is the effect of fragment size on genetic variation? Have populations from fragments diverged more than populations separated by a similar distance in continuous habitat? Some of the results agree with the predictions of population genetic theory, and some do not. Dipensia lapponica and C. bigelowii have significantly less genetic variation in fragmented habitat. Contrary to theoretical predictions, M. groenlandica has significantly more genetic variation in fragmented habitat. Gene flow between populations of D. lapponica and C. bigelowii in fragmented habitat appears to be absent or reduced compared to unfragmented habitat. Gene flow between populations of M. groenlandica does not seem to have been affected by the scale of habitat fragmentation in this study. Correlation of genetic variation with fragment size is positive for C. bigelowii, not significant for D. lapponica and negative for M. groenlandica. Contrary to theoretical predictions, populations of M. groenlandica have significantly more genetic variation in small habitat patches than in large ones. Populations of D. lapponica and C. bigelowii have diverged more in fragmented habitat than in unfragmented habitat, possibly as a result of genetic drift. Minuartia groenlandica populations have not diverged more in fragmented than in unfragmented habitat. Diapensia lapponica, C. bigelowii, and M. groenlandica have virtually identical distributions in the northeast, and must have migrated into the region contemporaneously (on a geologic time scale) from southern refugia, yet have strikingly different patterns of genetic variation. This observation should serve as a cautionary note in attempts to infer past migration patterns from current patterns of genetic variation. (Abstract shortened by UMI.)