Summary1. Pollinator-mediated selection on traits associated with mechanical fit and attraction of pollinators were investigated through both sexual functions in Cyclopogon elatus (Sw.) Schlechter (Orchidaceae). 2. Only halictid bees, principally Augochlora nausicaa , were observed as pollinators. The pollinarium becomes attached to the ventral surface of the bee's mouthparts (labrum) when the proboscis, which closely matches the length of floral tube, is projected into the flower to reach nectar. 3. We detected directional selection on nectary depth, with deeper nectaries favoured only through male fitness, because this trait affects pollinaria removal but not deposition. Correlational selection was detected through male function between nectary depth and the number of flowers in an individual's floral display. These traits affect pollination in a multiplicative way: flower number is positively related to the number of bee visits, and nectary depth positively affects the effectiveness of pollinaria removal at each visit. 4. We also detected stabilizing selection on display size. For smaller displays there was a strongly positive association between number of flowers and overall reproductive success, which can be attributed to a simple numerical effect. However, the expected performance of individual flowers is impaired in large displays by pollinator limitation, because bees visiting the display pollinate few flowers per visit and each bee carries, at most, one pollinarium.

BackgroundSome species of long-spurred orchids achieve pollination by a close association with long-tongued hawkmoths. Among them, several Habenaria species present specialized mechanisms, where pollination success depends on the attachment of pollinaria onto the heads of hawkmoths with very long proboscises. However, in the Neotropical region such moths are less abundant than their shorter-tongued relatives and are also prone to population fluctuations. Both factors may give rise to differences in pollinator-mediated selection on floral traits through time and space.Methodology/Principal FindingsWe characterized hawkmoth assemblages and estimated phenotypic selection gradients on orchid spur lengths in populations of three South American Habenaria species. We examined the match between hawkmoth proboscis and flower spur lengths to determine whether pollinators may act as selective agents on flower morphology. We found significant directional selection on spur length only in Habenaria gourlieana, where most pollinators had proboscises longer than the mean of orchid spur length.Conclusions/SignificancePhenotypic selection is dependent on the mutual match between pollinator and flower morphologies. However, our findings indicate that pollinator-mediated selection may vary through time and space according to local variations in pollinator assemblages.

By mimicking shape and female mating pheromones, flowers of sexually deceptive orchids attract sexually excited males which pollinate them while trying to copulate. Although many studies have demonstrated the crucial importance of odour signals in these systems, most flowers pollinated by pseudocopulation resemble, at least superficially, an insect body and these visual cues may be important to cheat pollinators. In this 2‐year study, we show that the shape of the labellum of Geoblasta pennicillata is a target of pollinator‐mediated natural selection. Contrary to our expectations, plants with a labellum shape more similar to female wasps were not favoured. The strength and pattern of phenotypic selection varied between study years and sexual functions. Although selection through female success was probably associated to the fine‐tuning of the mechanical fit between flower form and male wasp, shape was the target of natural selection through male success in both study years indicating that male wasps use this trait when choosing flowers. The imperfect mimicry and patterns of selection observed indicated that an exact imitation is not needed to attract and deceive the pollinators and they suggested a receiver perceptual bias towards uncommon phenotypes.

No abstract

Cape Verde is a tropical oceanic ecosystem, highly fragmented and dispersed, with islands physically isolated by distance and depth. To understand how isolation affects the ecological variability in this archipelago, we conducted a research project on the community structure of the 18 commercially most important demersal fishes. An index of ecological distance based on species relative dominance (Di) is developed from Catch Per Unit Effort, derived from an extensive database of artisanal fisheries. Two ecological measures of distance between islands are calculated: at the species level, ΔDi, and at the community level, ΔD (sum of ΔDi). A physical isolation factor (Idb) combining distance (d) and bathymetry (b) is proposed. Covariance analysis shows that isolation factor is positively correlated with both ΔDi and ΔD, suggesting that Idb can be considered as an ecological isolation factor. The effect of Idb varies with season and species. This effect is stronger in summer (May to November), than in winter (December to April), which appears to be more unstable. Species react differently to Idb, independently of season. A principal component analysis on the monthly (ΔDi) for the 12 islands and the 18 species, complemented by an agglomerative hierarchical clustering, shows a geographic pattern of island organization, according to Idb. Results indicate that the ecological structure of demersal fish communities of Cape Verde archipelago, both in time and space, can be explained by a geographic isolation factor. The analytical approach used here is promising and could be tested in other archipelago systems.

Morphological variations of the serranid fish Cephalopholis taeniops were studied in relation to habitat fragmentation in the Cape Verde Archipelago. While a significant allometric effect existed (11·5% of total body‐shape variation), differences in morphology associated with sex and ontogeny were not significant. MANOVA followed by CVA showed that each island presented a particular allometric pattern. Average body shape for all islands was well discriminated with CVA models. Pair‐wise comparisons of the ontogeny of morphological change between islands revealed that northern islands (Santo Antão, São Vicente and Santa Luzia) along with Boavista Island showed a similar direction in shape ontogeny, while all other paired comparisons indicated different ontogenetic patterns. When comparing directions of inter‐population shape changes, individuals from Fogo Island, the southernmost locality, departed far from the orthogonal relation, suggesting that they were undergoing independent body‐shape trajectories. Physical isolation by geographic distance and depth was positively correlated with morphological divergence among populations from different islands. This finding supports the hypothesis that habitat fragmentation in the Cape Verde Archipelago can be interpreted in terms of marine population structure.