Effectiveness of infrared sensor camera monitoring for detecting pollinator visits to <i>Cremastra appendiculata</i> var. <i>variabilis</i> in a temperate forest in Central Japan

Abstract Pollination is one of the most crucial plant–animal interactions. Although direct observation, continuous video recordings, scheduled cameras, and other strategies are often used to monitor flower visitors, long‐term observation remains challenging, especially at night. In this study, we evaluated the effectiveness of a commercially available infrared sensor camera (Ltl‐Acorn 6210) in detecting pollinator visits to Cremastra appendiculata var. variabilis , a forest orchid with a highly specialized pollination system. Over 4 years (2019–2022), 151 cameras were deployed across 151 inflorescences, yielding 2707 camera days and 885 videos of Bombus diversus queens, the only insect observed with attached pollinia. Fruit development occurred only at inflorescences where B. diversus queen visits were recorded, and pollinia removal was observed almost exclusively at such inflorescences, confirming that the sensor camera successfully detected visits contributing to pollination success. Other insects, including B. ardens , Syrphinae flies, and nocturnal moths contacted the flowers, but did not carry pollinia. Sensor cameras recorded visits successfully even during warm daytime hours, likely due to the high thoracic temperatures of bumblebee queens. Compared to scheduled cameras, sensor cameras produced fewer invalid recordings and more detections. However, a limitation was that small heterothermic insects were rarely detected. Our findings show that sensor cameras are a reliable, low‐maintenance tool for long‐term monitoring of effective pollinators in forest ecosystems. While not suited to comprehensive pollinator network studies, they are valuable for detecting large‐bodied insects, enabling efficient, long‐duration monitoring.