K E Y W O R D Sflower production, fruit set, phylogenetically independent contrast, pollen limitation, pollen removal, pollination by deceit, reproductive ecology

A combination of mechanical isolation and incomplete ethological isolation eliminates the possibility of pollen transfer between the species. These results do not support either the facilitation or competition hypothesis regarding the effect of nearby rewarding flowers on non-rewarding plants. The absence of a significant effect of non-rewarding P. chusua on rewarding G. diantha can be ascribed to low levels of overlap between the pollinator pools of two species.

No abstract

The presence of such a large number of terrestrial orchid species in a small area (ca. 1 km 2 ) of the Huanglong valley in southwestern China is uncommon for this country. Studying the relationship between the distribution patterns of these orchid species and their microenvironments may help us understand this uncommon phenomenon. We established 662 1 m £ 1 m plots, measured the cover of each species and found that there were 33 orchid species distributed mainly in two diVerent habitats, i.e. travertine areas and forest. In the travertine areas, 30 orchid species were found; the six most common ones being Cypripedium bardolphianum, Cypripedium Xavum, Cypripedium tibeticum, Galearis diantha, Ponerorchis chusua and Phaius delavayi. However in the forested habitat, we found 21 orchid species; the most common ones being Tipularia szechuanica and Goodyera repens. Travertine areas had a higher number of orchid species as well as higher numbers of orchid species per plot as compared to forest. Light availability seems critical to the performance and distribution of orchid species. Stream Xow through the travertine area during the