I originally published this post on May 19, 2007.

As we mentioned just the other day (May 2007), studying animal behavior is tough as "animals do whatever they darned please". Thus, making sure that everything is controlled for in an experimental setup is of paramount importance. Furthermore, for the studies to be replicable in other labs, it is always a good idea for experimental setups to be standardized. Even that is often not enough. I do not have access to Science but you may all recall a paper from several years ago in which two labs tried to simultaneously perform exactly the same experiment in mice, using all the standard equipment, exactly the same protocols, the same strain bought from the same supplier on the same date, the same mouse-feed, perhaps even the same colors of technicians' uniforms and yet, they got some very different data!

The circadian behavior is, fortunately, not chaotic, but quite predictable, robust and easily replicable between labs in a number of standard model organisms. Part of the success of the Drosophila research program in chronobiology comes from the fact that for decades all the labs used exactly the same experimental apparatus, this one, produced by Trikinetics (Waltham, Massachusetts) and Carolina Biologicals (Burlington, North Carolina):

This is a series of glass tubes, each containing a single insect. An infrared beam crosses the middle of each tube and each time the fly breaks the beam, by walking or flying up and down the tube, the computer registers one "pen deflection". All of those are subsequently put together into a form of an actograph, which is the standard format for the visual presentation of chronobiological data, which can be further statistically analyzed.

The early fruitfly work was done mainly in Drosophila pseudoobscura. Most of the subsequent work on fruitfly genetics used D.melanogaster instead. Recently, some researchers started using the same setup to do comparative studies of other Drosophila species. Many fruitfly clock labs have hundreds, even thousands, of such setups, each contained inside a "black box" which is essentially an environmental chamber in which the temperature and pressure are kept constant, noise is kept low and constant ("white noise"), and the lights are carefully controlled - exact timing of lights-on and lights-off as well as the light intensity and spectrum.

In such a setup, with a square-wave profile of light (abrupt on and off switches), every decent D.melanogaster in the world shows this kind of activity profile:

The activity is bimodal: there is a morning peak (thought to be associated with foraging in the wild) and an evening peak (thought to be associated with courtship and mating in the wild).

The importance of standardization is difficult to overemphasize - without it we would not be able to detect many of the subtler mutants, and all the data would be considered less trustworthy. Yet, there is something about standardization that is a negative - it is highly artificial. By controlling absolutely everything and making the setup as simple as possible, it becomes very un-representative of the natural environment of the animal. Thus, the measured behavior is also likely to be quite un-natural.

Unlike in the lab, the fruitflies out in nature do not live alone - they congregate with other members of the species. Unlike in a 'black box', the temperature fluctuates during the day and night in the real world. Also unlike the lab, the intensity and spectrum of light change gradually during the duration of the day while the nights are not pitch-black: there are stars and the Moon providing some low-level illumination as well. Thus, after decades of standardized work, it is ripe time to start investigating how the recorded behaviors match up with the reality of natural behavior in fruitflies.

Three recent papers address these questions by modifying the experimental conditions in one way or another, introducing additional environmental cues that are usually missing in the standard apparatus.


In the first paper, Nocturnal Male Sex Drive in Drosophila (Current Biology, Volume 17, Issue 3 , 6 February 2007, Pages 244-251), by Fujii, Krishnan, Hardin and Amrein, the problem of isolation was dealt with. In order to do this, a different apparatus had to be used, in this case a bunch of petri-dishes placed under the watchful eye of a video camera:

Two flies at a time were placed in each petri-dish: either two males, two females or one male and one female. Their general locomotor activity was compared to that of isolated insects of both sexes. In addition, some more concrete behaviors - "close-proximity" (i.e., two individuals approaching each other), courtship and mating were monitored as well.

So, what did they find? Putting two males or two females together did not change the activity patterns much. But putting one male and one female together provoked a large change of behavior - most of the approaches, courting and mating occured during the night!

Here are individual males:

And here are male-female pairs:

Additionally, they found that the flies revert back to their standard patterns if placed back into isolation (so being virgin vs. experienced does not matter and the shift in patterns of behavior is not permanent).

Also, by entraining males and females 11 hours out of phase with each other and then placing them together, they discovered that the males drove the couples' behavioral patterns - the pair always assumed the phase of the male.

Finally, experiments with various genetic knock-outs and mutants (I'll spare you the tedious details) revealed that a) the males drive the pattern due to their perception of the females' smell and b) intact circadian pacemakers in both the brain and the antennae (the fly equivalent of the olfactory bulb in some sense) are necessary for the shift of behavior to a nocturnal pattern.


So, if the first paper suggests that the smell of virgin females can lure males to get active in pitch darkness, would a low-level light at night also encourage flies to stay up all night and party? The role of dim light during the nights was studied in the second paper, Moonlight shifts the endogenous clock of Drosophila melanogaster (PNAS, February 27, 2007, vol. 104, no. 9, pp. 3538-3543) by Wolfgang Bachleitner, Lena Kempinger, Corinna Wülbeck, Dirk Rieger, and Charlotte Helfrich-Förster.

In an earlier paper from the same lab, shutting down the activity in one of the clock genes eliminated the morning peak of activity, but the evening peak remained for quite a few days (even weeks) afterwards (does that mean that eating is less important than mating?), suggesting that the two peaks are driven by different clocks.

In this paper, the authors used an artificial light equivalent in intensity to a quarter-moon light. They compared activity patterns as well as patterns of clock-gene expression in standard light-dark cycles, in constant dark, in constant moonlight, and in a light-moonlight cycle.

Both the activity and the gene expression changed dramatically when moonlight was present. The morning peak started earlier, during the latter portion of the moonlit night, while the evening peak extended into the early portion of the next moonlit night:

The pattern of cycling of clock-proteins followed the same pattern as activity - advanced (and broadened) in the part of the brain thought to house the morning oscillator and delayed (and broadened) in parts of the brain thought to house the evening oscillator (though the literature is still not clear on their exact location).

Furthermore, deletion of an important photopigment (cryptochrome, which is not a clock gene in flies) only slightly raised the treshold of sensitivity to light - the activity patterns changed in the same way as in wildtype flies once the intensity of moonlight was raised to 0.5 lux. Being a blue-light pigment, cryptochrome may be imporant in detection of changing spectra during dawn and dusk and thus involved in the measurement of photoperiod (daylength).

But, deletion of a gene that results in lack of compound eyes (but not ocelli) made the flies blind to moonlight (not daylight, though). So, light detection by compound eyes is necessary for the changes in activity patterns in the presence of moonlight. Also, the peripheral clock in the compound eyes did not switch patterns of gene expression under moonlight in wildtype flies.

Hot Summer Nights

I initially intended to include another paper in this review - Integration of Light and Temperature in the Regulation of Circadian Gene Expression in Drosophila (PLoS Genet 3(4): e54 doi:10.1371/journal.pgen.0030054) by Boothroyd CE, Wijnen H, Naef F, Saez L, Young MW - but now that I have read it, I realize that it may be better to write about it on its own (and not just because it is huge, but also because it is conceptually complex) or together with another paper that recently saw some press and deserves some coverage by me as well. So, for now, yes, one can use temperature cycles to alter the patterns of fruitfly activities (and move it somewhat into the night) but that is not the main finding of the paper.

In summary, what a simple light-dark cycle in the laboratory does to isolated fruitflies is "box" their activity entirely within the light phase of the cycle. In other words, it exerts a strong masking effect of light on activity. Out in nature, presence of dim light, temperature cycles and conspecifics allows these insects to spread their activity into the night. While we still agree that the pattern, being bimodal, is that of a crepuscular animal, these new findings suggest that the fruitflies are not predominantly diurnal as thought to date, but flexible in their behavior and under some conditions even strongly nocturnal animals.