The genus Arisaema (Araceae) has rapidly diversified in Japan, and multiple species often coexist in the field. Although Japanese Arisaema species hybridize from artificial crossing, hybrid individuals are rare in mixed populations; suggesting the presence of effective pre-pollination barriers. We examined the following reproductive barriers between A. sikokianum and A. tosaense: habitat, phenology, and pollinator isolations. METHODS:Habitat isolation was examined by interspecific comparisons of microhabitat conditions at a mixed site and of altitude at the sampling site of herbarium specimens. Phenological isolation was evaluated by comparing seasonal transition in apparent spathe condition and frequency of insect visitation. Pollinator isolation was examined by comparing floral visitor assemblages between the two Arisaema species. To avoid overestimation of pollinator isolation due to seasonal changes in insect assemblages, we also compared visitor assemblages between natural and late-flowering A. sikokianum, where the latter was experimentally introduced and blooming with a natural A. tosaense population. RESULTS:Microhabitat conditions and sampling elevations of herbarium specimens overlapped between the two Arisaema species. At the population level, A. sikokianum and A. tosaense flowered for 39 and 52 days, respectively, with 13 days overlap. Insect visitation in A. sikokianum decreased before the seasonal overlap. Floral visitor assemblages differed between the two Arisaema species, while the difference between natural and lateflowering A. sikokianum was less distinct. CONCLUSIONS:Phenological and pollinator isolation contribute to reproductive isolation between the two Arisaema species and should enable the two species to coexist in this area.

Background and Aims Interspecific difference in pollinators (pollinator isolation) is important for reproductive isolation in flowering plants. Species-specific pollination by fungus gnats has been discovered in several plant taxa, suggesting that they can contribute to reproductive isolation. Nevertheless, their contribution has not been studied in detail, partly because they are too small for field observations during flower visitation. To quantify their flower visitation, we used the genus Arisaema (Araceae) because the pitcher-like spathe of Arisaema can trap all floral visitors. Methods We evaluated floral visitor assemblage in an altitudinal gradient including five Arisaema species. We also examined interspecific differences in altitudinal distribution (geographic isolation) and flowering phenology (phenological isolation). To exclude the effect of interspecific differences in altitudinal distribution on floral visitor assemblage, we established 10 experimental plots including the five Arisaema species on high- and low-altitude areas and collected floral visitors. We also collected floral visitors in three additional sites. Finally, we estimated the strength and contribution of these three reproductive barriers using the unified formula for reproductive isolation. Key Results Each Arisaema species selectively attracted different fungus gnats in the altitudinal gradient, experimental plots, and additional sites. Altitudinal distribution and flowering phenology differed among the five Arisaema species, whereas the strength of geographic and phenological isolations were distinctly weaker than those in pollinator isolation. Nevertheless, the absolute contribution of pollinator isolation to total reproductive isolation was weaker than geographic and phenological isolations, because pollinator isolation functions after the two early-acting barriers in plant life history. Conclusions Our results suggest that selective pollination by fungus gnats potentially contributes to reproductive isolation. Since geographic and phenological isolations can be disrupted by habitat disturbance and interannual climate change, the strong and stable pollinator isolation might compensate for the weakened early-acting barriers as an alternative reproductive isolation among the five Arisaema species.

No abstract

No abstract

No abstract