Social learning is learning from the observation of how others interact with the environment. However, in nature, individuals often need to process serial social information and may either favour the most recent information (recency bias), constantly updating knowledge to match the environment, or the information that appeared first in the series (primacy bias), which may slow down adjustment to environmental change. Mate-copying is a widespread form of social learning in a mate choice context related to conformity in mate choice, and where a naïve individual develops a preference for a given mate (or mate phenotype) seen being chosen by conspecifics. Mate-copying is documented in most vertebrate taxa and in the fruit fly Drosophila melanogaster. Here, we tested experimentally whether female fruit flies show a primacy or a recency bias by presenting pictures of a female copulating with one of two contrastingly coloured male phenotypes. We found that after two sequential contradictory demonstrations, females show a tendency to prefer males of the phenotype preferred in the first demonstration, suggesting that mate-copying in D. melanogaster is not based on the most recently observed mating and may be influenced by a form of primacy bias.

A mate-copying experiment in D. melanogaster consists of two distinct phases. First, a demonstration phase where the observer female can acquire social information from demonstrators, followed by a mate choice test of the observer. We used the same set-up as previous studies. It was made of two adjacent plastic tubes (0.8 cm x× 3 cm each) that wasere plugged on one side with cotton, and on the other by a microscopy cover slide (1.6 cm x× 1.6 cm) that served as a window from which the observer female could watch the demonstration. Another similar glass partition was placed between the two tubes, separating them. The observer female was placed in the tube adjacent to the side of the demonstration. Demonstrations consisted inof one life-sized picture of a female mating with one of the coloured males with a male of the other colour standing by, as previously done by Nöbel et al.. Pictures were approximately 0.5 cm from the focal female tube. We used 26 different pictures to reduce pseudo-replication, and we found that the picture ID did not explain the variation in our data. Each trial consisted of two 30 minute demonstrations in two successive slots followed by a mate choice test. Demonstrations were followed by a 15 minute resting time during which the observer female remained in the experimental device. Treatments consisted of a main treatment and two controls and ended with a 30 minute mate choice test, where the observer female had the choice between a pink and a green living male. These males were initially added to the tube separated by the glass partition after the demonstrations were finished, which was removed at the start of the mate choice test.

The main treatment consisted of two sequential demonstrations providing conflicting social information (i.e., the demonstrator female copulated with the male of one colour in the first demonstration and with the male of the other colour in the second. To control for the time gap between each demonstration and the mate choice test, we replicated that protocol in the controls but provided no visual stimulus in the second (“‘Control 1”’) or first demonstration (“‘Control 2”’). These controls also tested whether flies were learning socially with the current protocol.