Native pollinators play a crucial role in flowering plant reproduction.  Several native North American pollinator species are already federally listed as endangered, and in Europe, there is evidence that many pollinator species are undergoing population declines.  Restoring lost pollinators to ecosystems may positively impact the plant communities in question.  However, we currently lack evidence that manipulating pollinator abundance will alter pollination services to plants, especially when most such interactions are situated in complex webs containing alternative pollinating species.  Moreover, the pollinator niche breadth of a plant may limit the benefits it receives from restoration of any one pollinator taxon. 
I tested these ideas using a small network of two alpine plants (Trifolium dasyphyllum and parryi) and their two bumblebee pollinators (Bombus kirbyellus and sylvicola).  I conducted experiments to determine the pollinator niche breadth of the plants, and to examine how pollination benefits varied with abundance of alternative pollinator species.  Preference and pollination efficiency trials, as well as field observations, showed that both bees are equally efficient pollinators of the two Trifolium, but T. dasyphyllum has a broader pollination niche than T. parry; Queens of both bumblebee species visit T. dasyphyllum equally while visits to flowers of T. parryi are almost completely provided by B. kirbyellus.  Additions of Bombus foragers to naturally occurring  mixed species arrays showed that increasing the abundance of B. kirbyellus enhances pollination services to T. dasyphyllum only slightly, but enhances services to T. parryi substantially.  Results are a step toward quantifying the impact of pollinator restoration on plant population growth, within a realistic network of interacting partners that vary along the facultative-obligate continuum.