To control flight, insects rely on the pattern of visual motion generated on the retina as they move through the environment. When light levels fall, vision becomes less reliable and flight control thus becomes more challenging. Here, we investigated the effect of light intensity on flight control by filming the trajectories of free-flying bumblebees (Bombus terrestris, Linnaeus 1758) in an experimental tunnel at different light levels. As light levels fell, flight speed decreased and the flight trajectories became more tortuous but the bees were still remarkably good at centring their flight about the tunnel's midline. To investigate whether this robust flight performance can be explained by visual adaptations in the bumblebee retina, we also examined the response speed of the green-sensitive photoreceptors at the same light intensities. We found that the response speed of the photoreceptors significantly decreased as light levels fell. This indicates that bumblebees have both behavioural (reduction in flight speed) and retinal (reduction in response speed of the photoreceptors) adaptations to allow them to fly in dim light. However, the more tortuous flight paths recorded in dim light suggest that these adaptations do not support flight with the same precision during the twilight hours of the day.

Night sky orientation with diurnal and nocturnal eyes: dim-light adaptations are critical when the moon is out of sight

In comparison to other insects, like honeybees, bumblebees are very effective pollinators. Even though landing is a crucial part of pollination, little is known about how bumblebees orchestrate the final, critical moments of landing. Here, we use high-speed recordings to capture the fine details of the landing behaviour of free-flying bumblebees (Bombus terrestris), while landing on a flat platform with different orientations. We find that the bees have a fairly constant body and head orientation at the moment of leg extension, irrespective of platform tilt. At the same moment in time, the distance to the platform is held constant at around 8 mm (with the exception of low platform tilts). The orientation of the antennae and the first appendage that touches the platform vary between platform orientations, while the duration of the hover phase does not. Overall, the final moments of landing in bumblebees and their close relatives, the honeybees, are similar. However, the distance to the platform at the moment of leg extension and the duration of the hover phase are different in bumblebees and honeybees, suggesting that they are primarily adapted to land on surfaces with different orientations.

To make a smooth touchdown when landing, an insect must be able to reliably control its approach speed as well as its body and leg position—behaviors that are thought to be regulated primarily by visual information. Bumblebees forage and land under a broad range of light intensities and while their behavior during the final moments of landing has been described in detail in bright light, little is known about how this is affected by decreasing light intensity. Here, we investigate this by characterizing the performance of bumblebees, B. terrestris, landing on a flat platform at two different orientations (horizontal and vertical) and at four different light intensities (ranging from 600 lx down to 19 lx). As light intensity decreased, the bees modified their body position and the distance at which they extended their legs, suggesting that the control of landing in these insects is visually mediated. Nevertheless, the effect of light intensity was small and the landings were still well controlled, even in the dimmest light. We suggest that the changes in landing behavior that occurred in dim light might represent adaptations that allow the bees to perform smooth landings across the broad range of light intensities at which they are active.

1. The ability to measure flying insect activity and abundance is important for ecologists, conservationists and agronomists alike. However, existing methods are laborious and produce data with low temporal resolution (e.g. trapping and direct observation), or are expensive, technically complex, and require vehicle access to field sites (e.g. radar and lidar entomology). 2. We propose a method called "camfi" for long-term non-invasive monitoring of the activity and abundance of low-flying insects using images obtained from inexpensive wildlife cameras, which retail for under USD$100 and are simple to operate. We show that in certain circumstances, this method facilitates measurement of wingbeat frequency, a diagnostic parameter for species identification. To increase usefulness of our method for very large monitoring programs, we have developed and implemented a tool for automatic detection and annotation of flying insect targets based on the popular Mask R-CNN framework. This tool can be trained to detect and annotate insects in a few hours, taking advantage of transfer learning. 3. We demonstrate the utility of the method by measuring activity levels and wingbeat frequencies in Australian Bogong moths Agrotis infusa in the Snowy Mountains of New South Wales, and find that these moths have a mean wingbeat frequency of 48.6 Hz (SE = 1.4), undertake dusk flights in large numbers, and that the intensity of their dusk flights is modulated by daily weather factors. Validation of our tool for automatic image annotation gives baseline performance metrics for comparisons with future annotation models. The tool performs well on our test set, and produces annotations which can be easily modified by hand if required. Training completed in less than 2 h on a single machine, and inference took on average 1.15 s per image on a laptop. 4. Our method will prove invaluable for ongoing efforts to understand the behaviour and ecology of the iconic Bogong moth, and can easily be adapted to other flying insects. The method is particularly suited to studies on low-flying insects in remote areas, and is suitable for very large-scale monitoring programs, or programs with relatively low budgets.