*2.2. Behavior*

The behavior of single tester males towards two Di2 headless target flies (a male and a female) was measured over a 5 min period. All flies were isolated 0–4 h after eclosion under CO2 anaesthesia. A few minutes prior to the test, anaesthetized target flies were decapitated with a razor blade (cleaned with ethanol between each sex). In all courtship tests, we used headless flies which do not copulate: therefore, decapitation allows courtship duration to be standardized. Moreover, some tests were performed under red light (25W with a Kodak safe-light filter no. 1) under which flies are virtually blind. In these conditions, most behavioral, visual and acoustic variables associated with the object fly are removed, this enhancing the behavioral e ffect of pheromones [55,56]. Only upright headless flies were used for the test. Tester males were individually aspirated (without anaesthesia) under a watch glass used as an observation chamber (1.6 cm3). After 5 min, the two headless target flies were simultaneously introduced and the courtship index (CI) toward each target was measured (CIf = courtship towards female flies; CIm = courtship towards male flies). CI corresponds to the fraction of the time spent courting relative to the total amount of time multiplied by 100. We only took in account male active courtship (wing vibration, licking and attempting copulation; the very brief episodes of tapping behavior were rarely seen under red light and thus was not included in the CI calculation). No qualitative di fference was noted between the di fferent courtship sequences shown by the di fferent tester male genotypes. The discrimination measure was based on the comparison between CIf and CIm for each genotype and condition used. *n* > 30. Locomotor activity, resulting of the cumulation of the activity of single flies, was measured during five periods of 10 s (total of 50 s), over five minutes (*n* > 24). All tests took place 1–5 h after lights on, in a room at 24 ± 0.5 ◦C with 65 ± 5% humidity. Tests were performed over several days to randomize the experimental variation of uncontrolled environmental parameters.
