Delayed germination and dispersal in desert annuals: Escape in space and time.

A model is developed to consider the interplay between dispersibility and delayed germination in desert annuals. The model explores the effect of low levels of dispersal, considered realistic for annual plants, on optimal germination fraction. The model also demonstrates the effect of the amount and accuracy of "predictive" (responsive to the environment) dormancy on the optimal innate germination fraction (not responsive to environmental conditions).Optimal germination fraction is found to be very sensitive to changes in despersibility especially at the limited dispersibilities that are realistic for annual plants. As dispersibility increases, optimal germination fraction increases. If plants make two kinds of seeds with differing despersibility, reproduction is maximized if the low dispersal seeds have delayed germination and the high dispersal seeds have quick germination. If dormancy mechanisms permit seeds to germinate when environmental conditions allow successful maturation, and remain dormant when environmental conditions do not permit successful maturation, what fraction of seeds should remain dormant under predicted good conditions as a hedge against inaccurate prediction of the environment? If environmental cues that break dormancy are uncorrelated with environmental conditions that permit successful maturation, predictive dormancy has little or no effect on the optimal innate germination fraction. When predictive dormancy lowers the probability of germinating when environmental conditions preclude successful maturation, the optimal innate germination fraction increases with increasing germination control by predictive dormancy. With a moderate degree of germination control by predictive dormancy, the optimal innate dormancy is still sensitive to changes in dispersal in the low dispersal ranges characteristic of annual plants.Evidence is presented from plant species that have both dispersal and germination dimorphisms to support the predicted correlation of high germination fractions with high dispersal.