Size-dependent sex allocation in a gynodioecious wild strawberry: The effects of sex morph and inflorescence architecture

Some models of sex allocation predict that if male fitness gains decelerate faster than female fitness gains with increasing size, then plants should increase allocation to female function with increasing plant size. We teste d this prediction in a gynodioecious (females and hermaphrodites) wild stra wberry by manipulating plant size in clonally replicated genotypes, by hand -pollinating with pollen from unrelated donors, and by measuring sex alloca tion response at two levels: among flowers and within flowers. These method s avoided potential confounding factors that can occur in studies of size-b ased allocation under natural conditions, such as genotype-based variation in size, and size-based variation in pollen limitation or self-pollination. They also allow us to determine if inflorescence architecture represented a constraint on the response to plant size. We found that a 75% increase in the vegetative size of hermaphrodites resulted in 78% more flowers, 91% mo re fruits, 10% more pollen per flower, and 16% more ovules per flower. Thes e responses to plant size represent a significant modification of phenotypi c gender at the within-flower level (pollen : ovule ratio) but not at the a mong-flower level (fruit : flower ratio). We also found that while female p lants enhanced flower and fruit allocation with plant size to the same degr ee as hermaphrodites, they increased ovule number per flower to a greater d egree. Sex differential effects of plant size such as those just described could influence breeding system evolution if they alter the relative seed f ertility of the sex morphs. We also found that inflorescence architecture m odified the response to plant size. Specifically, the effect of plant size on ovule number was strongest in basal positions and was absent at the most distal position. This, combined with a trend in the opposite direction for pollen production, resulted in position-dependent pollen : ovule ratios. S mall plants had higher pollen : ovule ratios than large plants at the prima ry (basal) position, equivalent ratios at intermediate positions, and lower ratios at the quaternary (distal) position. We suggest that this architect ural variation in response to plant size may result from structural limitat ions or physiological trade-offs within the inflorescence.