Scavenging by Pheidole pallidula: a key for understanding decision-making systems in ants

The usual evolutionary and ecological approaches to foraging in social insects often lack an investigation at the level of both individual behavioural complexity and social mechanisms ruling the emergence of adaptive collective strategies. The prey scavenging behaviour of the dimorphic ant Pheidole pallidula was used in this study to investigate (1) how individuals estimate the size of prey, (2) how they modulate their behaviour and communication and (3) how these modulations generate the diversity of collective foraging patterns. For a pile of small prey (fruit flies), the recruitment of foragers was slow because of the weak intensity of individual trail-laying behaviour and the long time spent by ants wandering around the food. In contrast, for a large prey item (a cockroach), strong recruitment was induced by ants that dashed back to the nest laying a more continuous chemical trail. Experiments with small immovable prey showed that the tractive resistance of prey was the key parameter the foragers used to estimate prey size and that it ruled their trail-laying intensity. These data allow us to generate a model about decision making in scavenging. The rules leading to collective choice in a foraging or an agonistic context are discussed. On the basis of these findings, some theoretical stances in sociobiology and some shortcomings in current approaches to cooperation in social insects are considered. ? 1997 The Association for the Study of Animal Behaviour An ant colony can demonstrate a wide range of strategies from solitary foraging to mass recruitment (see review by Passera 1984; Hölldobler & Wilson 1990). Foraging responses change depending on the type of food, for example insect versus seed or sugared water (Cammaerts 1980; Cammaerts & Cammaerts 1980; Breed et al. 1987), the food size (Wilson 1962; Traniello 1987), or the spatio-temporal distribution of available food (Hölldobler & Wilson 1970; Itzkowitz & Haley 1983; Hahn & Maschwitz 1985; Breed et al. 1987; Sundström 1993). Behavioural ecology provides a framework for describing and predicting these strategies in terms of Darwinian fitness. According to optimality theory, each selected foraging strategy represents the best achievable balance of costs and benefits that maximizes the net energetic yield to the colony. Many field and laboratory studies aim to Correspondence: C. Detrain, Laboratoire de Biologie Animale et Cellulaire, Université Libre de Bruxelles, Avenue F. D. Roosevelt 50, Bruxelles B-1050, Belgium (email: [email protected]). 0003–3472/97/030537+11 $25.00/0/ar960305 ? 1 537 assess these costs and benefits in terms of food intake, size of body reserves, energy expenditure of retrieving food, mortality risks from predators or competing species, and so on. However, this usual approach to foraging and other types of cooperation in social insects often remains at a rather theoretical level, describing the outcome of behaviour but not the mechanisms by which the outcome is achieved. Preoccupied with the adaptive value of the ‘chosen’ foraging strategy, sociobiology has more or less forgotten to investigate how an ant ‘measures’ complex parameters at the food source, ‘calculates’ costs and benefits of actions and ‘chooses’ one kind of recruitment system over another. Answering these questions needs a close interplay between optimality studies and analyses of behavioural mechanisms. Currently, we have little idea about the behavioural complexity needed at either an individual or a collective level, for an ant colony to achieve the best foraging pattern. In this respect, possibly the best known ants’ foraging behaviour is trail recruitment to a sucrose solution: faced with two 997 The Association for the Study of Animal Behaviour Animal Behaviour, 53, 3 538 food sources of different concentrations, a larger and more rapid recruitment is directed to the richer one (Pasteels et al. 1987; Beckers et al. 1990, 1993; for similar results with bees see Camazine & Sneyd 1991; Seeley et al. 1991). The resulting recruitment curves emerge from an interplay between positive feedbacks (trail laying) and from limiting factors such as the number of potential foragers and crowding around the food source (a non-feeding ant does not recruit, Wilson 1962). The intensity of trail-laying behaviour is modified according to the sucrose concentration perceived by the recruiting ant (Wilson 1962; Hantgartner 1969, 1970; Cammaerts 1977; Verhaeghe 1982; Pasteels et al. 1987; Beckers et al. 1990, 1993). The highest sucrose concentration results in the strongest trail reinforcement. These behavioural rules and social constraints can lead the colony to choose collectively to exploit the more energetically valuable food. In this case, we do know which information (here the sucrose concentration) causes the ants to alter their trail-laying behaviour and hence induces, at the colony level, the best response. This is one example in which an adaptive foraging pattern can be fully linked to behaviour and food assessment by the scouts. A wide array of foraging responses is also observed during prey scavenging but the mechanisms governing their emergence remain unclear. The scavenging patterns also range from individual exploitation of small items to strong trail recruitment for heavy prey. In Myrmica spp. (Cammaerts 1980; Cammaerts & Cammaerts 1980; de Biseau & Pasteels 1994) and Pheidole pallidula (Detrain & Pasteels 1991), the scavenging pattern is regulated through the use of chemical trails and recruitment displays in the nest, more intense for large cockroaches than for small fruit flies. These differences in recruiting behaviour result from a decision-making system that remains poorly understood. To fill this gap, we have studied the scavenging behaviour of the European ant P. pallidula. More precisely, we investigated how the P. pallidula colony ‘measures’ the size of food items and then ‘chooses’ a suitable scavenging pattern. The sterile caste of P. pallidula is dimorphic, majors accounting for 2–15% of the whole population. This species is a generalist and opportunistic ant, in terms of diet which varies in quality and size. Retrieved insects vary from 0.3 to 10 mg. One minor can carry small items while mediumsized prey or cumbersome body parts (such as legs) are retrieved by groups of cooperating minors. On average most scavenged insects individually retrieved weigh 0.86 mg. Larger prey (mean 3.5 mg) are carried back to the nest by the ants collectively (Detrain 1990). A very large food item induces a massive recruitment of both minors and majors which dissect the prey directly at the food site. Food discovery and trail recruitment is done exclusively by the minors, whose poison gland contains trail pheromone. Majors perceive and follow these trails but cannot produce the pheromone (Ali et al. 1988; Detrain & Pasteels 1991). The majors’ caste is only involved in recruitment for heavy prey which they cut into pieces. Scavenging in Pheidole species provides a good model for understanding how dead prey of different sizes induce different global foraging patterns and different levels of cooperation in the ant colony. It also raises several questions about the information content of the food itself, the relevant criteria and the measurement methods used by the forager at the food source. We also question how precisely the ant conveys information about the discovered food to the nest, and how the adaptive foraging response can emerge at the society level. The ultimate goal is to incorporate into the same framework this deeper understanding of decisionmaking systems in social insects and the more usual evolutionary and ecological approach to foraging.