Nocturnal bees and wasps. (A) The Central American sweat bee Megalopta genalis (Halictidae), whose sensitive apposition eyes allow them to forage at night by visually learning landmarks along the foraging route and around the nest entrance. Reproduced with the kind permission of the photographer, Dr Michael Pfaff. (B,C) The pale-yellow coloured Central American paper wasp Apoica pallens (Vespidae), which congregates on the outside of the nest (B) to create a distinctive pale object that may be visible to returning foragers at night. Photographs in B and C were reproduced with the kind permission of the photographer, Gillian Little, and Daniel Marlos from `What's That Bug?', a website devoted to popular entomology (www.whatsthatbug.com).

The daily timing of foraging flights in the nocturnal bee Megalopta genalis. Megalopta is active during two short time intervals each day, once in the morning (A) and once in the evening (B), when light levels at the nest (blue symbols) can be lower than 10^–4 cd m^–2. The number of bees leaving the nest (green bars) and returning to the nest (red bars) in 5 min intervals is shown relative to sunrise or sunset (0:00 h). Data were collected over several nights from several nests during two successive years on Barro Colorado Island (Panama). (A) The timing of foraging trips in the morning. The grey area indicates the time before astronomical twilight. (B) The timing of foraging trips in the evening. The grey area indicates the time after astronomical twilight. During the time interval between the onset of astronomical twilight in the evening and its offset in the morning, no light from the sun is present in the night sky. Modified with kind permission from Kelber et al. (Kelber et al., 2006).

Nocturnal landmark orientation in the nocturnal halictid bee Megalopta genalis. (A) A typical nocturnal orientation flight, as seen from below. The bee leaves her nest, and quickly returns to face the nest entrance. Flying in short arcs, she investigates the nest entrance and a neighbouring landmark to learn their spatial arrangement before departing on her foraging trip. Each `ball-and-stick' represents the position of the head (ball) and body (stick) at 40 ms intervals. (B,C) Landmark learning. Bees leaving for a foraging trip learn the position of their nest relative to others (B), or learn the presence of a white square card attached to their nest (C). Upon return, bees enter the nest marked by the landmarks they have previously learned, not their actual nests (which are marked by stars). The rear side of the square card was attached to a Perspex cylinder that slipped neatly over the end of the nest stick to hold the card in place over the nest entrance. Times and light intensities at departure and return are also shown. Adapted from Warrant et al. (Warrant et al., 2004).

The flight paths of four individual nocturnal bees (Megalopta genalis) returning to the nest at dimmer (A,D) and brighter (B,C) light levels during dawn (A,B) and dusk (C,D). In each plot, the three-dimensional flight path (black) is shown together with two-dimensional projections (grey shadows on the right, left and bottom walls) of the flight path onto the imaginary walls of a cubical space centred on the nest entrance (which is shown as a cylinder on the right wall). Luminance in the early dawn (A) was 1.1×10^–4 cd m^–2 and the landing lasted 11.4 s; late dawn (B), 1.9×10^–3 cd m^–2 and the landing lasted 4.7 s; early dusk (C), 3.9×10^–3 cd m^–2 and the landing lasted 1.8 s; and late dusk (D), 3.9×10^–4 cd m^–2 and the landing lasted 16.2 s. Note that during both the dawn and the dusk, landing flights were more circuitous and took longer in dimmer light. Each grid square is 10 cm×10 cm. Adapted with kind permission from Theobald et al. (Theobald et al., 2007).

Adaptations for nocturnal vision in the photoreceptors of the nocturnal sweat bee Megalopta genalis, as compared to photoreceptors in the closely related diurnal sweat bee Lasioglossum leucozonium. (A,B) Responses to single photons (or `photon bumps': arrowheads) recorded from photoreceptors in Megalopta (A) and Lasioglossum (B). Note that the bump amplitude is larger, and the bump time course much slower, in Megalopta than in Lasioglossum. (C–F) Average contrast gain as a function of temporal frequency in Megalopta (blue curves, N=8 cells) and Lasioglossum (red curves, N=8 cells) at different adapting intensities, indicated as `effective photons' per second in each panel for each species [for each species, each stimulus intensity was calibrated in terms of `effective photons'; that is, the number of photon bumps per second the light source elicited, thereby eliminating the effects of differences in the light-gathering capacity of the optics between the two species, which is about 27 times (Lillywhite and Laughlin, 1979)]. In light-adapted conditions (C,D), the two species reach the same maximum contrast gain per unit bandwidth although Lasioglossum has a broader bandwidth and a higher corner frequency (the frequency at which the gain has fallen off to 50% of its maximum). In dark-adapted conditions (E,F), Megalopta has a much higher contrast gain per unit bandwidth. All panels adapted with kind permission from Frederiksen et al. (Frederiksen et al., 2008).

The average rates of information transmission (in bits s^–1) in the photoreceptors of the nocturnal and diurnal sweat bees Megalopta genalis (blue curves, N=8 cells) and Lasioglossum leucozonium (red curves, N=8 cells). (A) When the photoreceptors alone are considered (via a light source calibration in `effective photons' absorbed by the photoreceptor per second), it is evident that at all intensities Lasioglossum has a higher information rate than Megalopta. (B) When light sources are instead calibrated to external ambient intensities (a normalised intensity of 100 corresponds to the light intensity on an overcast day, or around 180 cd m^–2), Megalopta has a higher information rate in dim light. This, however, is due to its 27 times more sensitive optics and is not due to an intrinsic adaptation present within the photoreceptors. Error bars show ±s.d. Both panels adapted with kind permission from Frederiksen et al. (Frederiksen et al., 2008).