“Pheromone-induced accuracy of nestmate recognition in carpenter ants: Simultaneous decrease of Type I and Type II errors”
Natacha Rossi,* David Baracchi,* Martin Giurfa, and Patrizia d’Ettorre (Feb 2019)
Adding a new twist to the traditional view of alarm pheromones: formic acid improves nestmate discrimination in ants
*These authors contributed equally to this work
Being able to recognize friends and foes is essential for the evolution of sociality. Animals, including humans, cooperate with group-members but usually deter strangers, which represent a potential threat for the society. While humans and other animals often use visual recognition cues or sounds, insects communicate mostly via chemical messages. Ants are among the most ecologically successful social insects and represent the pinnacle of social life. Their colonies are called “eusocial”: the perfect society. Ant sociality is based on cooperation between group-members and on colony closure: the rejection of any individual that does not belong to the colony and that could steal the common colony’s resources.
In many ant species, when aliens are detected around the nest, colony guards release formic acid, a toxic chemical weapon that kills foes but also rings the alarm bell for group-members, which rush to attack the source of threat. Natacha Rossi and colleagues performed controlled experiments at the Laboratory of Experimental and Comparative Ethology (University of Paris 13), in collaboration with the CNRS Research Center on Animal Cognition (University Paul Sabatier, Toulouse), to test whether formic acid, besides its immediate alarm effect, modifies the recognition behavior of carpenter ants. They found that, after smelling formic acid, ants improve discrimination accuracy: they increase aggressive behaviors towards strangers, while at the same time decreasing aggressive acts erroneously directed towards group-members.
With this novel discovery, the function of alarm signals becomes more comprehensive than previously thought: not only do they provoke immediate alarm behavior and defensive responses, but they can reduce errors in recognition of identity by allowing enhanced discrimination between enemies, which should be deterred, and friends, which should be preserved during a contest. Taking into account the overall effect of alarm signals opens new perspectives for a deeper understanding of recognition systems and decision making in general.
The ecological and evolutionary success of social insects relies on their ability to efficiently discriminate between group members and aliens. Nestmate recognition occurs by phenotype matching, the comparison of the referent (colony) phenotype to the one of an encountered individual. Based on the level of dissimilarity between the two, the discriminator accepts or rejects the target. The tolerated degree of mismatch is predicted by the acceptance threshold model, which assumes adaptive threshold shifts depending on the costs of discrimination errors. Inherent in the model is that rejection (Type I) and acceptance (Type II) errors are reciprocally related: if one type decreases, the other increases. We studied whether alarm pheromones modulate the acceptance threshold. We exposed Camponotus aethiops ants to formic acid and subsequently measured aggression towards nestmates and non-nestmates. Formic acid induced both more non-nestmate rejection and more nestmate acceptance than a control treatment, thus uncovering an unexpected effect of an alarm pheromone on responses to nestmates. Nestmate discrimination accuracy was improved via a decrease of both types of errors, a result that cannot be explained by a shift in the acceptance threshold. We propose that formic acid increases the amount of information available to the ants, thus decreasing the perceived phenotypic overlap between nestmate and non-nestmate recognition cues. This mechanism for improved discrimination reveals a novel function of alarm pheromones in recognition processes and may have far-reaching implications in our understanding of the modus operandi of recognition systems in general.