Clearly "Should I use male or female flies for behavior?" is a trick question because there really is nothing limiting you from testing both male and female flies.
Sexually dimorphic behaviors like courtship, have been studied since the 1960’s (Erhman, 1964), and the neurogenetics of courtship is probably one of the best understood behavioral mechanisms in Drosophila (Villella & Hall, 2008; Rings & Goodwin, 2019). Most of what we know about courtship behaviors focuses on the behavior of male flies, but more recently people have started looking at the role of females (for example, Shao et al, 2019; Kim et al, 2024). Aggression is also a sexually dimorphic behavior that was largely focussed on male behavior, but is beginning to be studied in females (Palavicino-Maggio & Sengupta, 2022; Chiu et al 2023).
Like other subfields of behavioral neuroscience, I honestly think that focussing on just one sex or the other is largely historical. Certain behavior assays were run with a certain sex and because all of the initial characterization was in that sex, people didn’t bother running the other sex. And, because for a long people people viewed Drosophila as little in vivo test tubes, they didn’t thing sex of the fly was necessarily relevant. Until recently, details about the sex of the flies used in a study were often not even included in the methods (Note - always write the sex of flies used into your study!).
I think this is also because people tested male and female flies without necessarily caring about sex (depending of course on which chromosome their gene of interest was). In the associative memory field (ie Pavlovian memory for an odor associated with shock or sugar), scientists typically sorted out a mixed bunch of ~100 flies and rarely looked at sex differences or recorded what proportion of males and females are used. In fact, this is still standard practice for these types of studies that use large numbers of flies.
Courtship suppression memory is different (Raun et al, 2019) because the focus is looking at how males learn to suppress their courtship for a previously mated female. Alcohol behaviors have always looked at one sex or the other and not necessarily mixed males and females. This was originally because male and female flies metabolize alcohol differently and later because there was a neurogenetic basis for sex differences in alcohol responses (Devineni & Heberlein, 2012). Interestingly, Drosophila courtship suppression and alcohol behaviors were largely studied by women in the field, but I have no idea whether that played a role in the decision to look at both male and female flies.
Behaviors where you need to tether a fly, like flight, visual learning or instrumental conditioning, have historically used female flies (Wolf & Heisenberg, 1990; Dill et al, 1997). My understanding is females were chosen because they are a bit bigger than male flies and their flight can be sustained longer. That is, they last longer in these assays so you can get a bit more data out of them. I’m not sure if it’s still practice to use females or to look at both male and female flies in these paradigms.
So, why isn’t it a standard practice in Drosophila behavior to test both male and female flies and look at sex differences in behavior? Since flies are literally less than a dime a dozen, why wouldn’t you test for sex differences? In some cases, it is true that it is extra work, and doubles your sample size, but in my opinion it’s worth it. And, because we can typically test pretty high sample sizes, your studies should be well-powered enough to look for sex differences. A short aside: I have probably lost some fly people here and they are asking ‘what does well-powered mean’? In (non-fly) behavior studies what is often done is a pilot test with a smaller sample size, then using this preliminary data to run a power analysis, which will tell you what sample size you need in order to be able to reliably pull out statistically significant differences between groups.
So this leads to another question: should we separate our males and females before performing the experiments? Typically we do this for logistical reasons - we separate out males and females when we collect our flies so that we can test the effects of both males and females in our behavior assay later. However, we have no idea how being isolated in same-sex groups affects the behavior of the flies. Clearly, group dynamics certainly affects behavior so we can only presume that this will impact behavior in one way or another (Jezovitz et al, 2021; Yadav et al, 2024; for more on this check out ‘Does isolating flies affect behavior?’). With the advent of more sophisticated tracking systems, we can also look at the dynamics of male and female behaviors in same-sex or mixed-sex groups, so I suspect we will start to see information on this emerging (for more on this check out ‘Which tracking system should I use?).
For those people currently funded by the NIH, you are probably aware of the Sex as a Biological Variable (SABV) mandate which requires everyone who is performing vertebrate animal research to run male and female animals and do the appropriate tests to see if there are sex differences in whatever they are studying. Since flies are not vertebrate animals, this mandate does not apply to Drosophila research. That being said, a lot of grant reviewers do not necessarily know that SABV applies to vertebrate animals only, and flag lack of SABV compliance in their reviews if it isn't included in your grant. I saw an example in a grant I reviewed that directly addressed this by saying something like ‘Since Drosophila are not vertebrate animals, SABV does not apply, but we will be testing both male and female flies with sufficient sample size to detect sex differences’. Ever since then I include a statement like this in my grants - and I think it helps.
So my overall recommendation: Test both male and female flies. Maybe test them together if you are using a tracking system that can pick up on sex differences, and maybe test them apart if you want to look at how they act in same-sex groups or if mixing them complicates the behavior you want to measure because they keep courting, etc. Make sure to detail out everything in your methods:
Which sex was used?
Were they isolated in a same-sex group in advance of the study?
Were they tested in a same-sex group or a mixed-sex group?
How many were tested at one time?
Did you do a power analysis to see if you could pull out sex differences from your study? If not, what was your rationale for your sample size?
Do you have more tips for looking at sex differences? I’d love to hear them!