Search

Summary Dispersal and dormancy are important risk-reducing strategies in unpredictable environments. Negative covariation between these strategies is theoretically expected, but empirical evidence is limited and inconsistent. Moreover, covariation may be affected by other life-history traits and may vary across levels of biological organization. We assessed dispersal (vertical fall time of fruits, a proxy for wind dispersal ability) and dormancy (germination fractions measured during germination trials) in populations of 15 annual and 12 perennial wind-dispersed species in six Asteraceae genera from South Africa. Dormancy was higher in annuals than in perennials, whereas fall time was largely determined by evolutionary history. Controlling for phylogeny, dispersal and dormancy was negatively associated across species and life-history categories. Negative covariation between dispersal and dormancy was not evident at either the individual level (except for seed heteromorphic species) or the population level. Our study provides rare empirical support for the theoretical expectation of tradeoffs between dormancy and the alternative risk-reducing strategies, perenniality and dispersal, but refutes the expectation of increased dispersability in perennials. Although negative covariation between dispersal and dormancy at the species level appears not to be a simple consequence of upscaling individual-level mechanistic tradeoffs, our findings suggest that selection for one strategy may constrain evolution of the other.

Summary Since co-flowering plants often share pollinators, their fecundity is likely affected by pollen transfer within and among plant species. Changes in pollinator visitation rates (e.g. through facilitation or competition for pollinators) are usually thought to exert the predominant influence on plant fecundity. However, the transfer of conspecific relative to heterospecific pollen between individuals may also play an important role. Indeed, the relative importance of these determinants of fecundity is expected to depend on the density, relative abundance and spatial arrangement of con- and heterospecifics. We investigated the effects of con- and heterospecific density and spatial distribution pattern on pollinator visitation and plant fecundity within and between two annual, self-incompatible co-flowering species [Dimorphotheca pinnata and Ursinia cakilefolia (Asteraceae)] by manipulating their relative abundance, overall patch density and dispersion patterns in experimental arrays in Namaqualand, South Africa. We quantified pollinator visitation rates and fruit set in arrays of varying density and aggregation. This enabled us to determine which mechanism(s) were driving variation of fecundity, particularly through their influence on visitation rates, mate availability and heterospecific pollen interference. To test whether autogamy offers reproductive assurance when individuals are scattered among a dense population of heterospecifics, we included an autogamous species (U. anthemoides) in a separate experiment. We found that increased fecundity with increasing conspecific density was not the result of higher visitation rates, but rather increased mate availability. Furthermore, increased spatial aggregation of conspecifics at low density significantly increased fecundity through reduced heterospecific interference. In contrast to results for self-incompatible species, fruit set in U. anthemoides was consistently high and unaffected by scattered distribution patterns. This suggests that autogamy may offer reproductive assurance when mates are limited and the potential for interspecific pollen transfer (IPT) is high. Synthesis. In this study of annual daisies, variation in fruit set is primarily driven by factors that affect the transfer of conspecific relative to heterospecific pollen, independent of pollinator visitation rate. Our findings demonstrate that mate limitation and IPT negatively affect fruit set and that these effects can be mitigated by intraspecific aggregation and the ability to autonomously self-pollinate.

 Premise of the study: The showiness of fl displays is usually explained as an adaptation to attract pollinators. However, selection for less attractive displays imposed by non-pollinating agents, particularly herbivores, may balance pollinator-driven selection for highly visible infl orescences. We investigated whether infl orescence architecture, particularly the unusual groundlevel fl owering associated with a specialized bird perch in Babiana ringens may have originated, in part, as an adaptive response to mammalian herbivory.  Methods: We measured levels of herbivory by antelope in populations of B. hirsuta , the putative sister species of B. ringens , which possesses the likely ancestral form of infl orescence architecture. To test for position-dependent effects of herbivory on fl owers, we compared the herbivory rates and seed production of manipulated infl orescences in a fi eld experiment. We predicted that fl owers at the base of infl orescences would suffer less herbivory than those in apical positions.  Results: We found herbivore damage to flin 50% of naturally occurring B. hirsuta plants. Manipulated infl orescences with only basal fl owers, and consequently similar infl orescence architecture to B. ringens , experienced signifi cantly lower herbivory and higher seed set than infl orescences manipulated to have only apical fl owers.  Conclusions: Our results are consistent with the hypothesis that position-dependent herbivory on infl orescences could have played a role in the evolution of infl orescence design. More specifi cally, position-dependent herbivory may have selected for the loss of apical fl owers. Position-dependent herbivory may have contributed toward the evolution of a naked infl orescence axis, a structure that characterizes B. ringens and functions as a bird perch facilitating cross-pollination by sunbirds.