Fecundity characteristics, phenology, and behavior of insect flower—visitors were studied for 7 early flowering woodland herbs: Claytonia virginica, Dentaria laciniata, Dicentra canadensis, Dicentra cucullaria, Erythronium albidum, Isopyrum biternatum, and Sanguinaria canadensis. Sanguinaria canadensis is facultatively autogamous, the Dicentras are obligate outcrossers, and the remainder are self—compatible, at least within a stem. All are insect pollinated except sometimes S. canadensis. The numbers of ovules per flower and flowers per stem tended to be inversely correlated, and large—seeded species (S. canadensis, E. albidum, I. biternatum) had lower numbers of potential seeds per stem than did small—seeded species. Flowering of all species typically occurred during the first prolonged period of weather suitable for pollinator activity and ceased by the time the canopy closed. Annual differences in flowering times were associated with differences in average temperatures (i.e., early blooming in a warm, early spring), but cumulative degree—hours or degree—days of air or soil temperatures were not well correlated with flowering times. Other constraints on flowering phenology are discussed, including the predictability of suitable conditions, a proposed "fail—safe" mechanism that may assure flowering before canopy closure even if temperatures are abnomally low, and the importance of nontemperature factors in defining suitable conditions. Flowering time was not very finely tuned to the temperature regime and pollinator activity; flowers blooming during the flowering peak often had low seed production and the fertilization rate of most species was low. Evidence that seed production may have been pollinator limited for several species was obtained by comparing the success of hand pollination and of natural pollination, rarity of certain specialized pollinators, and estimates of the abortion rates of fertilized ovules. We suggest that flowering in early spring is a high—risk option in terms of insect—mediated sexual reproduction. Certain flower—visiting insects favored D. laciniata out of proportion to its abundance, but no effect on seed set of other species was detectable. Honeybees were abundant and active flower visitors with the potential for disrupting ecological/evolutionary relationships between native insects and flowers.

No abstract

No abstract

No abstract

In the eastern Andes of Colombia, the shrub Befaria resinosa (Ericaceae) has peaks of flowering that are separated by extended periods of low flower production. The effect that these fluctuations in flower production have on pollen flow was investigated by using fluorescent dye as a pollen analog. Dye applied to open flowers was dispersed over long distances more often during low flower production than during high flower production. Whether enhanced pollen dispersal during flowering lows is of benefit to individual plants is not clear. The proportion of flowers that set fruit is positively correlated with flower abundance, negating the possibility that increased pollen dispersal results in a higher rate of fruit production due to outbreeding effects. It is also difficult to attribute the pattern of fruit production to changes in pollinator visitation rates, which are negatively correlated with flower abundance in the case of hummingbirds and not correlated at all with flower abundance in the case of insects. An opportunistic, large-bodied hummingbird (Colibri coruscans) visits B. resinosa during high flowering and may be a particularly effective pollinator, accounting for some of the increase in the proportion of flowers setting fruit. Rainfall is positively correlated with flower production and may be an important factor in shaping flowering phenology, but it is not significantly correlated with the proportion of flowers setting fruits. The possibility that low-level flowering may counteract inbreeding that results from peak flowering is discussed.