Plant-pollinator interactions database for construction of potential networks

Biological recording data were obtained from the Bees, Wasps and Ants Recording Society (BWARS), Butterflies for the New Millennium (BNM, Asher 1997) and Hoverfly Recording Scheme (HRS), with interactions recorded as incidental metadata submitted alongside occurrence records. Pollinator biological records with associated metadata were algorithmically screened for text matching a valid scientific or vernacular plant name (or a widely used synonym or abbreviation of either). In most cases, such matches record the visitation of a flower by a pollinating insect. However, there are other reasons for a valid match in the comments field, including descriptions of local habitat (e.g. 'English oak woodland'), pollinator behaviour (e.g. 'flying near blackthorn hedge') or sampling methods (e.g. 'swept from Arrhenathrum eliatus sward'). In order to exclude these, all plant species producing a match were screened according whether they produce flowers which are known to be visited by insects and whether such visits are likely to result in beneficial pollen transfer (i.e. the plant is primarily entomophilic). These criteria excluded the majority of trees and grasses, which are likely to derive from habitat descriptions or interactions other than flower visits (e.g. honeydew feeding). Even though some primarily wind-pollinated grass flowers are known to be visited by certain pollinators, their impact on pollen transfer is likely to be very slight in temperate climates (Clifford 1964; Adams et al. 1981). The data contains only plant species not defined as 'casuals' in PLANTATT ('casual is a plant that is present only as populations which fail to persist in the wild for periods of more than approximately five years' Hill et al. 2004).

For a full description of methods, please download the supporting documentation (plant_pollinator_interactions_supporting_documentation.rtf)