Individual flowering phenology, plant size, and reproductive success in Linanthus androsaceus, a California annual.

For natural selection to shape population flowering phenologies, individual phenological variation must be correlated with variation in reproductive success. I therefore marked and followed individual plants of Linanthus androsaceus (a California grassland annual) throughout the flowering season, recording individual flowering phenology, flower number, mortality, and seed production. Although date of first flowering was unrelated to number of flowers, plants first flowering during an intermediate interval had a greater probability of setting seed, and a stronger relationship between seed number and flower number, than plants first flowering early or late in the season. The actual distribution of first flowering date in the population was clustered around this intermediate interval. In contrast with first flowering date, flowering duration was correlated with flower number, with a positively skewed distribution that reflected the skewed size structure of the population. The combined distributions of individual flowering dates and furations resulted in an overall population flowering curve that was skewed to the right. Within individual quadrats, Linanthus density was positively correlated with skewness if the quadrat flowering phenology curve, and negatively correlated with the percentage of plants in bloom at flowering peak. Thus, although individual variation in first flowering date was related to reproductive success in L. androsaceus, the size dependence of flowering duration provides a mechanism whereby ecological factors can shape population flowering phenologies without evolutionary change.