COMPONENTS OF FLOWERING TIME VARIATION IN A DESERT ANNUAL.

How much of the variation seen in life histories is consistent with adaptive hypotheses, and how much requires other kinds of explanation? Differences in flowering time between Sonoran (earlier flowering) and Chihuahuan Desert (later flowering) populations of the desert annual Eriogonum abertianum Torr. (Polygonaceae) are significant, repeatable between greenhouse experiments, and persist into a second greenhouse generation. These apparent genetic differences are consistent with a hypothesis of local adaptation: field demographic studies (Fox, 1989b) show that many fewer Sonoran than Chihuahuan Desert plants survive to the summer rainy season, suggesting selection for earlier flowering in the Sonoran Desert. Within natural populations there is considerable phenological complexity: time of first flowering varies by up to six months, and individuals may have zero, one, or several reproductive episodes. Greenhouse sib analyses revealed only marginal among-family genetic variation for flowering size. The resemblance between parents and offspring for size and time of flowering varied with growth conditions, suggesting that this marginal variation among families may be at least partly due to factors other than additive genetic variance. On the other hand, moisture limitation significantly delayed the onset of flowering in two independent experiments. Variation in moisture availability in both time and space is characteristic of desert environments. The phenological complexity in natural populations may thus be generated by random variation in moisture availability, possibly in conjunction with variation in germination date and plant size. The results call into question the claim that drought generally induces flowering in desert annuals.