In this document one can find the data acquired from the experiences by visually counting the cockroaches at every time-step (30 min interval).


Time Time in minutes
Shelter : RG = Dark shelter that becomes Light shelter; GR = Light shelter that becomes Dark shelter; R = Dark shelter; G = Light shelter.
Experience : number id of the trial for a viben condition
Population size : number of indivduals
Condition : Inversion = Trials where the ligth colours are inverted between the shelter after 10 hours; Control = Trials where the light colours in the shelters remain unchanged.
Observed : Observed number of individuals at a given time-step
Color : Light colour of the shelter at a given time-step; Red (wavelngth pic at 630) ; Green Red (wavelngth pic at 560)

In this document one can find the data acquired from the experiences by visually counting the cockroaches at every time-step (30 min interval).


Time Time in minutes
Shelter : RG = Dark shelter that becomes Light shelter; GR = Light shelter that becomes Dark shelter; R = Dark shelter; G = Light shelter.
Experience : number id of the trial for a viben condition
Population size : number of indivduals
Condition : Inversion = Trials where the ligth colours are inverted between the shelter after 10 hours; Control = Trials where the light colours in the shelters remain unchanged.
Observed : Observed number of individuals at a given time-step
Color : Light colour of the shelter at a given time-step; Red (wavelngth pic at 630) ; Green Red (wavelngth pic at 560)

Data obtain from the experiments

Data obtain from the experiments

Data obtain from the experiments

Many social species are able to perform collective decisions and reach consensus. However, how the interplay between social interactions, the diversity of preferences among the group members and the group size affects these dynamics is usually overlooked. The collective choice between odourous and odourless shelters is tested for three groups of social cockroaches (Periplaneta americana) which are solitary foragers: naïve (individuals preferring the odorous shelter); conditioned (individuals without preference); and mixed (combining, unevenly, conditioned and naïve individuals). The robustness of the consensus is not affected by the naïve individuals’ proportion, but its rate and the selection frequency of the odorous shelter are correlated to this proportion. In mixed groups, the naïve individuals act as influencers. Simulations based on the mechanisms highlighted in our experiments predict that the consensus emerges only for intermediate group sizes. The universality of these mechanisms suggests that such phenomena are widely present in social systems.

Collective choices and consensus resulting from the competition of positive feedbacks, are the subject of numerous studies. However, how the interindividual diversity of preferences among the group members affects these dynamics is usually overlooked apart from vertebrates based studies. Gregarism is a useful model for studying how the interindividual diversity of preferences affects the collective choices and consensus. The decision-making of three types of groups of the cockroach Periplaneta americana is tested in a choice between an odorous shelter and an odourless one: naïves (individuals showing an inherent preference towards the odorous shelter); conditioned (individuals with the preference for the odorous shelter inhibited); and mixed (a combination of conditioned individuals and a minority of naïve ones). We show that while the consensus is robust for all groups (>90% of the total population is in the same shelter), the group composition determines the rate to achieve a consensus and the frequency of selection of the odorous shelter. Indeed, increasing the proportion of naïves in the group leads to an increasing number of consensuses towards the odorous shelter and to an increasing rate to reach them. We also show that the minority of naïve individuals in mixed groups act as leaders or influencers.