The reproductive ecology of the pacific yew (Taxus brevifolia nutt.) under a range of overstory conditions in western Oregon

Abstract

The influence of overstory openness on the reproductive ecology of Pacific yew (Taxus brevifolia Nutt.) was investigated on 4 sites in western Oregon over 2 years. The breeding system of T. brevifolia was found to deviate from pure dioecy under a broad range of canopy and site conditions. Production of female strobili was observed on 17 of 58 predominantly male trees, while no male strobili were observed on 57 female trees. Genet sex ratios were significantly biased in only 1 population, where male genets outnumbered female genets by almost 2 to 1. Mean floral sex ratios were significantly male-biased in all populations and ranged from 5 to 12. Pollen-ovule ratios were in excess of 1,000,000 for all populations. In contrast, reproductive effort based on masses of mature strobili were female-biased by a factor of 1.1 to 5 for all sites. Seed masses also varied inversely with elevation. Pollination phenology varied with elevation and overstory openness. Pollen first began shedding at the lowest sites, and earlier in trees under open conditions than in trees with overstory canopy cover. The duration of pollen shedding varied from 3 to 20 days, and tended to be more protracted at lower sites and under open canopy conditions. Most of the variation in reproductive potential, as indexed by strobilus production, occurred within sites and within trees. Little variation between years was observed in male strobilus production during the three years of this study. Also, while female strobilus production was significantly greater in 1993 than in 1994, seed production did not differ between years. Overstory openness was positively associated with growth and reproductive potential of T. brevifolia. Specific leaf area was inversely correlated with overstory openness, and branching was positively correlated with overstory openness, suggesting that T. brevifolia adapts to overstory removal by producing denser foliar tissue and increased self-shading. In contrast to reproductive potential, seed production was not significantly associated with overstory openness during the two years of this study. Also, seed efficiency (the ratio of seed production to ovule production) was negatively associated with overstory openness. Seed efficiency ranged from 5 to 34%, and attrition occurred in two phases. The early phase occurred during the pollination period and was probably due in part to pollination failure. Supplemental hand-pollination resulted in a doubling of seed efficiency on two of the sites, but average seed efficiency was still less than 15% on branches receiving supplemental pollen. Other potential sources of early attrition included damage from phytophagous mites, pathogens, frost and genetic incompatibility. Attrition in the later stages of seed development was due in part to predation by vertebrate seed consumers. Predator exclusion significantly increased seed development efficiency (the ratio of seed production to developing ovules) on 3 of 4 sites over 2 years. Seed production was positively correlated with overstory openness on branches bagged to exclude vertebrates, suggesting that resource availability was important for seed production in the absence of predation. However, evidence for resource limitation of seed production was not consistent. Seed efficiency was not significantly associated with overstory openness in 1993, and no associations were detected between vegetative growth or previous reproduction and seed efficiency in 1994. Possible evolutionary explanations for low seed efficiency in T. brevifolia include the effects of sexual selection, stochasticity in pollination and predation, and the importance of excess ovaries as reserves that compensate for constant sources of mortality. Sources of seed attrition varied considerably among years and sites, emphasizing the importance of spatial and temporal variability in the reproductive ecology of T. brevifolia.