Two papers published in Nature use different techniques to show that separate groups of circadian neurons control the characteristic morning and evening bouts of activity in the fruitfly Drosophila melanogaster.

When flies are kept in a cycle consisting of 12 h of light and 12 h of darkness, their activity peaks twice a day, around the times that the lights are switched on (morning) or off (evening). They become more active before the change in lighting, showing that their internal circadian clock allows them to anticipate the change in illumination.

Circadian rhythmicity in D. melanogaster depends on about 100 'clock' neurons, which are found in six clusters in the brain. Stoleru et al. and Grima et al. used targeted gene expression techniques to investigate the specific contributions of different groups of cells to the morning and evening peaks in activity. Stoleru and colleagues used the cryptochrome (cry) gene driver to drive expression of the proapoptotic gene hid, to kill circadian neurons that expressed the circadian photoreceptor cry, which produced a severe lack of circadian rhythmicity. The flies did not show anticipatory activity before lights-on or lights-off, and in constant darkness they were arrhythmic.

Ablation of one subset of neurons — the ventral–lateral LNv neurons, which express the neuropeptide PDF (pigment-dispersing factor) — suppresses the morning anticipatory peak of activity. When the authors generated flies in which only the rest of the cry-expressing neurons — those that did not also express PDF — were ablated, these flies had a normal activity peak in the morning, but did not show activity in anticipation of lights-off in the evening.

These results indicate that the LNv neurons are responsible for the morning activity peak, and the other cry-expressing neurons, including the dorsal–lateral LNd neurons, are responsible for the evening activity peak. Grima and colleagues came to the same conclusion, by using cell-specific drivers to rescue expression of the circadian protein Per in subsets of neurons in per-null flies. When Per expression was rescued in the LNv neurons, the morning activity peak was restored, but when it was rescued in both the LNv and LNd neurons, both the morning and evening activity peaks were normal.

Both groups also found that the 'morning' oscillator, maintained by the LNv neurons, can be maintained in sustained darkness, whereas the 'evening' oscillator cannot, although this is difficult to reconcile with findings that the morning peak, but not the evening peak, is progressively lost when wild-type flies are placed in darkness. It is likely that these two oscillators, although they can act independently, are functionally coupled under normal circumstances to allow them to be coordinated and to respond flexibly to challenges such as seasonal changes.