Although numerous studies have addressed plant compensatory responses to removal of vegetative tissue by herbivores, only a few have determined the effect of floral removal on subsequent seed production in natural populations. In Sanicula arctopoides, a perennial monocarp, removal of inflorescences (umbels), both naturally by deer (Odocoileus hemionus) and by artificial clipping early in the flowering season, led to no loss of maternal reproduction as measured by seed number and seed mass. Full compensation occurred when umbels bearing up to one—third of a plant's flowers were removed at the stage when plants are normally grazed. However, there are thresholds both in timing and severity of removal beyond which plants were unable to compensate fully. Compared to controls, artificial removal of developing fruits 20 d later than the normal stage resulted in a 42% decrease in seed production. Successive clippings that removed °55% of each plant's flowers resulted in a 52% decrease in seed production relative to controls; however, a single clipping of the same intensity resulted in full compensation. Differential reactions to clipping of different umbel ranks provided information about the mechanism of the response. Removal of secondary umbels resulted in decreased seed abortion rates in later developing umbels, while removal of tertiary umbels resulted in decreased abortion rates in both earlier and later formed umbels. These responses indicated that delayed abortion of developing seeds may be common in secondary umbels, that removal of secondary umbels by herbivores may relieve plants of some costs of abortion, and that S. arctopoides uses flexible allocation to achieve full compensation at typical levels of floral herbivory.

Seeds of Gossypium sturtianum and G. thurberi do not readily germinate under most conditions. Increased germination of G. sturtianum was associated in nature with the presence of a seed bug, Oxycarenus luctuosus. Experimental and descriptive studies were conducted in Australia and Arizona to test the hypothesis that more seeds of G. sturtianum and G. thurberi germinated during their first wet season from bolls that had been attacked by insects than from unattacked bools. In Australia, seeds from G. sturtianum bolls that had been caged with O. luctuosus were more likely to germinate than those caged without bugs. In Arizona USA, seeds of G. thurberi were attacked by boll weevils, Anthonomus grandis, and seed bugs, Sphyrocoris punctellus. Weevils destroyed 21%, scarified another 26%, and left undamaged the remaining 53% of the seeds in attacked bolls. Seeds of G. thurberi from bolls with weevils and seed bugs that were not destroyed by these insects were more likely to germinate during their first wet season than seeds from unattacked bolls. All three insects probably increase germination by breaking the impermeable seed coat. The ecological and evolutionary consequences of this interaction are not known.

Sexual expression in hermaphroditic plants is often a function of environmental factors affecting individuals before or during flowering. I tested for the effects of floral herbivory and lack of pollination in early umbels on the relative proportions of hermaphroditic and staminate (male) flowers produced on later umbels by Sanicula arctopoides, a monocarpic, andromonoecious perennial. Neither floral herbivory or lack of early pollination had a significant effect on the ratio of the two floral morphs, but the probability of producing staminate flowers on late umbels was strongly and positively related to plant size measured just prior to floral initiation and prior to herbivory. Plant size was also negatively correlated with flowering date. I suggest that producing staminate flowers on late umbels should benefit large early-blooming plants more than small late-blooming plants because more mating opportunities occur during the period when these flowers release pollen. Although herbivory did not cause labile changes of sex, whole plant phenotypic gender was still strongly affected by various forms of treatment. Sex-biased herbivory or lack of pollination rendered plants more or less phenotypically male, depending on which tissues were affected. Deer and pollen-feeding mites preferentially remove male tissues while hymenopteran seed predators preferentially remove female tissues. I conclude that combinations of herbivores could have counteracting or compounding effects on plant gender, and these effects may change the rankings of male and female reproductive success within populations.

No abstract

No abstract