How to efficiently obtain accurate estimates of flower visitation rates by pollinators

Regional declines in insect pollinators have raised concerns about crop pollination. Many pollinator studies use visitation rate (pollinators/time) as a proxy for the quality of crop pollination. Visitation rate estimates are based on observation durations that vary significantly between studies. How observation duration relates to the accuracy of the visitation rate estimate is, however, unknown. We studied this relationship using six day-long observations (06:00. h-19:00. h) in leek-seed production fields (totalling 78. h). We analysed beyond which point in time observing longer did not significantly improve the accuracy of the visitation rate estimate (minimum observation duration). We furthermore explored the relationship between the minimum observation duration and visitation rate, time of day and temperature. We found that the minimum observation duration (mean. ±. SD: 24. ±. 11.9. min) was significantly related to visitation rate, where the observation time required to obtain accurate estimates decreased with increasing visitation rate. Minimum observation duration varied greatly between days and between fields but not within days. Within days, the visitation rates differed significantly only between the hour-intervals 06:00. h-07:00. h (lowest visitation rate) and 09:00. h-11:00. h (highest rate). Minimum observation duration decreased up to around 22. °C beyond which it remained fairly stable. Surprisingly, even after three day-long observations on the same plant we found new pollinator species visiting the flowers, suggesting that species-richness estimates based on plant observations alone probably underestimate true species richness. Because especially between-day variation in visitation rate on single plants can be large, reliable estimates of the pollinator visitation rate during the plant's flowering time require observations on multiple days. Standardising the number of pollinators rather than the time to observe (standardised pollinator timing approach: time to n-pollinator visits) may provide more consistent accurate assessments of visitation rate, especially for studies that use gradients in visitation rates to examine the contribution of pollinators to crop pollination