Partitioning of pollinators during flowering in an African Acacia community

Ecology, Dec, 1998 by Graham N. Stone, Pat Willmer, J. Alexandra Rowe

RESULTS

The distribution of Acacia flowering patterns in seasonal time

Months during which flowering by component Acacia species was observed are summarized in Table 3. The highest diversity of flowering Acacia species in this habitat was present in December and January, following the autumn rains, when up to eight species flowered concurrently. In contrast to the majority of Acacia species, A. thomasii and A. etbaica flowered only during the dry season, at a time when very few individuals of other species were in flower. These species occupy seasonal flowering niches in which competition for pollinators with other Acacia species is either much reduced or avoided altogether.

Intraspecific variation in diurnal timing of pollen release

Intraspecific variation at a single site on a single day. - Intraspecific patterns of pollen release for Acacia species flowering together after the autumn rains are shown in Fig. 2. For all except A. drepanolobium [ILLUSTRATION FOR FIGURE 2D OMITTED], peaks for individuals of each species illustrated in Fig. 2 are significantly aggregated within the time period between dawn and dusk (A. nilotica V = 0.177, P [less than] 0.01; A. tortilis V = 0.09, P [less than] 0.05; A. zanzibarica V = 0.035, P [less than] 0.025; A, senegal V = 0.028, P [less than] 0.025). Intraspecific variation was lowest in the early morning dehiscence of A. nilotica and A. tortilis [ILLUSTRATION FOR FIGURE 2A, B OMITTED], and more marked in the later dehiscing A. drepanolobium and A. senegal [ILLUSTRATION FOR FIGURE 2D, E OMITTED]. A different pattern was seen in A. thomasii, flowering alone in the dry season. The four A. thomasii trees sampled showed far more variation in patterns of pollen availability than any of the co-flowering species (compare [ILLUSTRATION FOR FIGURE 2 OMITTED] with [ILLUSTRATION FOR FIGURE 3A OMITTED]), with no statistical support for aggregated timing of pollen release. Means across trees showed no significant peak at any time period [ILLUSTRATION FOR FIGURE 3B OMITTED].

Intraspecific variation among sites within a single season. - Peaks of pollen availability in A. zanzibarica occurred at very similar times in all the sites sampled [ILLUSTRATION FOR FIGURE 4A OMITTED] and peaks of pollen availability across sites were significantly aggregated in time (V = 0.065, P [less than] 0.025). Variation across sites was no greater than was found among individuals at a single site (Compare [ILLUSTRATION FOR FIGURE 2 OMITTED] with [ILLUSTRATION FOR FIGURE 4A OMITTED]). What differences there were in the timing of peak pollen availability between days at the same site, and among sites, were explained well by variation in relative humidity; peak pollen availability occurred later on days or at sites with a higher mean relative humidity over the period within which dehiscence was initiated [ILLUSTRATION FOR FIGURE 4B OMITTED].

Unlike A. zanzibarica, Acacia senegal showed no aggregation in timing of peak pollen availability among sites (V = 0.028, P [much less than] 0.05; [ILLUSTRATION FOR FIGURE 5A OMITTED]). Peak pollen availability occurred 2-4 h earlier at Zange and Vitiwini than at Kisima and Ubani, and the onset of dehiscence at Kisima was delayed relative to the other sites. Analysis of variance of the pollen-to-anther ratios indicates a significant difference between two pairs of sites (Zange and Vitiwini in one group, Kisima and Ubani in the other) over time (time x group interaction term [F.sub.8, 125] = 2.6, P = 0.012). A possible explanation could be differences in daily patterns of microclimate between these pairs of sites: were a critically low value of relative humidity to be the cue for anther dehiscence, earlier peaks of pollen availability would be expected at sites whose relative humidity falls to this level earlier in the day. Zange and Vitiwini showed earlier anther dehiscence despite higher relative humidity during the period preceding dehiscence [ILLUSTRATION FOR FIGURE 5B OMITTED], suggesting that some other difference between these populations is responsible for differences in the timing of pollen release. A consistent difference between the two pairs of sites is the presence/absence of Acacia zanzibarica in the co-flowering assemblage: early release in A. senegal coincides with absence of Acacia zanzibarica from Zange and Vitiwini, and later release coincides with presence of A. zanzibarica at Kisima and Ubani (Table 1).