Simon Goss

Collective patterns and decision-making

Autocatalytic interactions between the members of an animal group or society, and particularly chemically or visually mediated allelomimesis, can be an important factor in the organisation of their collective activity. Furthermore, the interactions between the individuals and the environment allow different collective patterns and decisions to appear under different conditions, with the same individual behaviour. While most clearly demonstrable in social insects, these principles are fundamental to schools of fishes, flocks of birds, groups of mammals, and many other social aggregates. The analysis of collective behaviour in these terms implies detailed observation of both individual and collective behaviour, combined with mathematical modelling to link the two.

Trails and U-turns in the selection of a path by the ant Lasius niger

A series of laboratory experiments show how colonies of the ant Lasius niger systematically select the shorter of two paths of varying length and form between nest and foraging area, and do so with a large majority of foragers. Three different mechanisms were considered to contribute to this selection, and were evaluated by comparing mathematical models with experimental data. Bi-directional trail laying was shown to contribute much less than U-turns, where the proportion of ants that turned back depended on the paths geometry and length, and also on the trail strength. A strong modulation of the amount of pheromone deposited per ant as a function of the paths geometry was found. This modulation can also contribute to the selection of the shorter branch in the experiments where the rate of U-turns is initially the same. The selection of the path is shown to be a collective process whereby trail laying and following amplifies small initial differences in the traffic on each path caused by these three mechanisms. The foragers show no significant tendency to follow the path they used previously.

Modulation of trail laying in the ant Lasius niger (Hymenoptera: Formicidae) and its role in the collective selection of a food source

Foragers of the ant Lasius nigerexploiting a 1 Msugar source were found to lay 43 %more trail marks than those exploiting a 0.05 or a 0.1 Msource. The trail laying per forager decreased during the course of individual recruitment episodes, and the mean lifetime of the trail pheromone was estimated to be 47 min. A mathematical function describing the probability that a forager chooses one of two paths in relation to the amount of trail pheromone on them closely fitted experimental data. These results were incorporated into a model describing the recruitment dynamics of L. niger.Simulations of this model showed that the observed modulation of trail laying with respect to food source quality is sufficient in itself to account for the systematic selection of the richer source seen in the experiments.

The blind leading the blind in army ant raid patterns: Testing a model of self-organization (Hymenoptera: Formicidae)

We present field experiments and analyses that test both the assumptions and the predictions of a model that showed how the swarm raids of the army ant Eciton burchellimight be self-organizing, i.e., based on hundreds of thousands of interactions among the foraging workers rather than a central administration or hierarchical control. We use circular mill experiments to show that the running velocity of the ants is a sigmoidal function of the strength of their trail pheromones and provide evidence that the swarm raid is structured by the interaction between outbound and inbound forager traffic mediated by the pheromones produced by both of these sets of ants. Inbound traffic is also affected by the distribution of prey, and hence, sites of prey capture alter the geometry of the raid. By manipulating the prey distributions for E. burchelliswarms, we have made them raid in a form more typical of other army ant species. Such self-organization of raids based on an interaction between the ants and their environment has profound consequences for interpretations of the evolution of army ant species.

How Trail Laying and Trail Following Can Solve Foraging Problems for Ant Colonies

One of the most striking features of an ant colony’s behaviour is its capacity for the spatial organisation of foraging activity. The use of trail pheromone to guide fellow workers in the nest to a large food source or rich foraging zone has been extensively studied (e. g. Wilson 1971) and obviously contributes to foraging efficiency. We have recently, however, been able to show that trail laying and trail following behaviour are more than just a means of communicating a food source’s location. When more than one trail is present at a time, the interactions between foragers and the trails can lead to the collective selection of the shortest path or the best food source, despite the fact that individual foragers have no means of making such choices.

Random behaviour, amplification processes and number of participants: how they contribute to the foraging properties of ants

Abstract Two major types of foraging organisation in ants are described and compared, being illustrated with experimental data and mathematical models. The first concerns large colonies of identical, unspecialised foragers. The communication and interaction between foragers and their randomness generates collective and efficient structures. The second concerns small societies of deterministic and specialised foragers, rarely communicating together. The first organisation is discussed in relation to the different recruitment mechanisms, trail-following error, quality and degree of aggregation of food-sources, and territorial marking, and is the key to many types of collective behaviour in social insects. The second is discussed in relation to spatial specialisation, foraging density, individual learning and genetic programming. The two organisations may be associated in the same colony. The choice of organisation is discussed in relation to colony size and size and predictability of food sources.

Dynamics of Collective Exploration in the Ant Pheidole Pallidula

Foragers of large ant societies tend to be integrated into a web of chemical communication, and, we conjecture, to show extensive trail laying behaviour outside food recruitment. This factor, combined with their large number, leads to the development of a network of marks and trails in the foraging area which serve to organise their foraging activity. The foragers orient themselves principally with reference to these chemical signals, exploration being a collective rather than an individual process. This is well documented for Iridomyrmex humilis (Deneubourg et al 1990; Aron et al, in press) and the army ants (Rettenmeyer 1963; Schneirla 1971; Topoff et al 1980; Franks and Fletcher 1983; Deneubourg et al 1989). Other species may be reasonably thought to follow the same logic (Pheidologeton diversus: Moffett 1984, 1988; Solenopsis sp.: Tschinkel 1987, Wilson and H611dobler pers. comm.; Iridomyrmex pruinosus: H611dobler pers. comm.). This type of organisation is often correlated with polydomy, which requires permanent links between the different sub-nests. The polymorphic ant Pheidole pallidula (Nyl.) is one of the most abundant Mediterranean species, whose opportunism is characterised by its night and day activity, the diversity of its diet and the flexibility of its foraging techniques which can include participation by the majors (Detrain, 1990). Because of its polydomous nature, its large colony size (several hundred individuals per subnest), and its strong trail recruitment to food sources, we felt that Ph. pallidula fitted into this pattern, contributing to its ecological success, and would therefore form a collective exploratory trail even though such behaviour has not previously been reported. This

Functional Self-Organisation Illustrated by Inter-Nest Traffic in Ants: The Case of the Argentine Ant

In many ant species the colony is not a single structure but rather a number of decentralised nests linked together by a network of trails. This network can be extended to include trails that form between the nests and long-lasting food sources or rich foraging areas. The formation of inter-nest networks was studied with laboratory colonies of the Argentine ant Iridomyrmex humilis. Bridges were placed to link isolated nests, with branches of equal length arranged in a triangle and a square linking three and four nests respectively, and two branches of different length linking two nests. The resulting ant traffic connected all the nests, neglecting redundant bridges, forming a minimum spanning tree. Cutting a frequented bridge caused the traffic to divert to a neglected bridge. Visual cues appear not to be essential as similar networks were generated both in light and darkness. Rotating the bridges caused the traffic to change correspondingly indicating a primary role of chemical cues with respect to memory or visual cues. Where the bridges between two nests were of unequal length, the ants neglected the longer one.

Error, communication and learning in ant societies

Abstract When large food sources are discovered, many species recruit using different mechanisms. Numerous recruits do not reach the food source and explore the foraging area. These ants make new discoveries. The interplay between communication and exploration generates collective and efficient structures. This organisation is discussed in relation to the mechanisms of recruitment, quality and degree of aggregation of food-sources and territorial marking. This first organization, characteristic of large colonies, is compared to another foraging organisation: small societies of deterministic and specialised foragers rarely communicating together.

Collectively Self-Solving Problems

Observing insect societies, we are deeply impressed by their capacity to build structures and solve problems, and it is not really surprising that popular litterature is full of anthropomorphic explanations stressing the capacities of individual ants. In such blueprints, however, a central unit manages the whole system, and to achieve this it must collect all the data needed. The algorithms to treat these data are necessarily complex and therefore highly specific. They can tolerate neither internal errors, inexact or incomplete information, nor changes in the problem which is assumed to be stable. The consequences of this type of organisation are such that each solution must be constantly monitored and overhauled to cope with unforseen events, leading to a spiral of mutually increasing complexity and fragility.