IntroductionSampling for describing plant–pollinator interaction networks has been performed using techniques that either focus on the plants (with flower-visit data) or the animals (with analyzing pollen on the body surface of flower visitors). The differences in the structure of the networks obtained using these methods likely influences our understanding of the contribution of nocturnal pollinators, yet this key finding has yet to be the focus of study.MethodsIn this study, we conducted an intensive diurnal field survey in the subalpine meadows of the Dajiuhu Wetland and supplemented the data with an analysis of diurnal and nocturnal pollen data to examine the changes in pollination networks.ResultsWe observed 41 plant and 154 pollinator species, corresponding to 665 specific interactions. Visitation and pollen analyses showed significant differences in the composition and interaction between network plants and pollinators, resulting in important structural changes in the network. Given that the diurnal pollen data showed new links that were preferentially attached to highly connected nodes, the level of asymmetric specialization did not decrease; however, nestedness increased 1.3-fold, and mean pollinator connectivity from 3.1 to 5.1. As the behaviors of nocturnal pollinators tended to be more specialized, the inclusion of nocturnal pollen data led to an increase in the number of extreme-specialist pollinator species. Consequently, nestedness decreased 0.8-fold, but mean plant connectivity went from 14.2 to 16.2.DiscussionThese findings suggest that the structure of pollination networks is influenced by the sampling methods and the level of detail of the investigation. Our study has strong implications for the development of monitoring schemes for plant–pollinator interactions. Due to the practical difficulties of nocturnal field visitation, when conducting research, combining diurnal field visitation with both diurnal and nocturnal pollen analyses is the most convenient and realistic method to capture the full complexity of these networks.
