PREDICTING EVOLUTIONARY CONSEQUENCES OF GREATER REPRODUCTIVE EFFORT IN TRIPSACUM-DACTYLOIDES, A PERENNIAL GRASS

To test the prediction that a mutation causing greater reproductive effort will result in reduced vegetative vigor, we compared the seed production, growth, and carbohydrate status of normal and pistillate genotypes of the grass Tripsacum dactyloides differing in seed production by up to fourfold. We evaluated the costs of reproduction by two methods: experimental manipulation of reproductive effort and comparison of high- and low-yielding genotypes. Despite the large difference in seed production, normal and pistillate (high-yielding) genotypes did not differ in growth rate over a 1-yr period. Contrary to predictions, carbohydrate reserves in the high-yielding genotype were significantly higher at the peak of the seed ripening period. Stalk defoliation and complete removal of seed stalks reduced plant growth rate, total aboveground biomass, and biomass of vegetative storage organs, especially when plants were also under stress from biweekly defoliation. However, stalk removal caused increased rates of growth in the year after experimental manipulations. Reproductive tillers were costly to the plant in terms of lost meristems and therefore future plant size, but seed costs were contained within reproductive tillers of both genotypes, explaining the lack of a trade-off in the high-yielding, pistillate genotype. Although experimental reduction of reproductive effort revealed a cost of reproduction, this result could not be used to correctly predict the consequences of a gene for greater seed production.