A common idea in animal ecology is that adverse or stressful environmental conditions facilitate the evolution of social behaviour [1]. The formation of groups across a huge number of animal taxa is thus considered to have broad implications for the benefit of individuals, including mate finding, the efficient location and use of resources, thermoregulation, energetic benefits and defence against natural enemies or competitors [2, 3]. Basic proximate prerequisites for communal behaviour are cues indicating the location of conspecifics and the ability to receive and process information regarding these cues, which in turn induce inter-individual attraction [3]. Because the costs and benefits of communal behaviour typically vary with environmental conditions, the degree to which individuals are mutually attracted is regulated by signals indicating the presences of predators, food availability, etc. [4].

In the vast group of insects that depend on ephemeral resources, such as decaying plant tissues, dung and carrion, aggregation in the immature stages across resource patches is the result of the choice of a female to lay batches of eggs and/or to aggregate with conspecifics [5–8]. In studies of Drosophila as an ecological model system, one benefit that females flies seem to achieve by this spatial aggregation is that larval survival probability to the adult stage is highest at intermediate densities [9, 10], indicating the existence of so-called Allee effects [11]. Competing filamentous fungi co-occurring with Drosophila larvae on the same patches have been demonstrated to cause high rates of mortality when larvae feed solitarily or in small groups, whereas larger groups are able to hamper mould growth [12] (Fig. 1), which in turn increases larval survival [9, 13, 14]. Although the mechanisms leading to mould suppression are not fully understood, physical damage of the fungal tissue from the feeding (shovelling food with the mouth hooks) and locomotor (crawling and digging) behaviour of the fly larvae [15] seems to be the major cause of the repression of mould growth [12, 14].

Given the relationship between spatial oviposition patterns, Allee effects and the suppression of mould, spatial aggregation in Drosophila can be interpreted as an adaptive behaviour against competing fungi on larval feeding sites in order to enhance offspring survival. These ecological interrelationships might set conditions for facilitating social behaviour in the fly larvae because, at the level of larval behaviour, a more efficient strategy that might control the rapid establishment of noxious fungi would be to exert physical stress directly on fungal colonies. Thus, larvae should display an assortative behaviour on the site on which fungi are growing, rather than moving randomly and independently of each other across a resource patch, by which the fungal tissues might only incidentally be destroyed. In the present study, I have provided groups of Drosophila melanogaster Meigen (Diptera, Drosophilidae) larvae with fungal-infected (2-day-old colonies of Aspergillus niger van Tieghem) and uninfected (control patches) (F-C treatment) food patches and examined whether the distribution of larvae across the patches is driven by fungal infection. In comparison with this naturally occurring heterogeneity in patch quality, I have also studied the distribution of fly larvae when they were offered only infected (F-F treatment) or uninfected (C-C treatment) food patches in order to test for the existence of grouping behaviour in two types of homogenous larval environment. If grouping is irrelevant under the given experimental setting, no deviation from the regular larval distribution across the food patches would be expected, i.e larvae should distribute themselves across patches in order to minimise larval competition for food [16].

Although Drosophila is a thoroughly studied model organism in foraging biology [17, 18], knowledge about social interactions between the insect larvae is surprisingly limited. This is intriguing because drosophilids are also model organisms in spatial ecology in which Drosophila communities are characterised by strong intraspecific aggregation across patchily distributed substrates (e.g. decaying plant tissues) [19–21]. The lack of knowledge concerning social interactions among larvae and its possible role in competition with filamentous fungi have provided the specific impetus of the present study.

On the background of ecological interactions between insects and filamentous fungi on ephemeral resources, the experiment presented in this study was designed to test for social attraction in Drosophila larvae, an attraction that I hypothesised to be advantageous when larvae are confronted with noxious moulds. The results demonstrate that the fly larvae significantly aggregated on food patches on which young fungal colonies were growing (F-C treatment, Fig. 2a). Moreover, when provided with a homogeneous environment (F-F or C-C treatment), larvae displayed significant aggregation across the two food patches (Fig. 2b). In comparison, however, aggregation was significantly enhanced when larvae had the choice between a mould-free and a mould-infected site (Fig. 2b). Thus, the results suggest that grouping behaviour in Drosophila larvae involves both mutual attraction between group members [2] and the attraction of individuals to the same environmental stimulus [22], i.e. cues emitted by the fungi.

Group formation in eusocial insects and those living in groups for part or most of their lives is often mediated by pheromones, e.g. cuticular hydrocarbons that induce attraction between individuals [22, 23]. Chemical communication is also widespread in drosophilid behaviour, including those associated with spatial aggregation in adult flies [24–26]. Because several receptors on the cephalic lobe of Drosophila larvae have gustatory, mechanosensory and olfactory functions [27], both chemical and physical cues (e.g. substrate vibrations caused by larval movements) might be involved in mutual attraction. However, the mechanisms leading to grouping and group cohesion in Drosophila larvae are unknown. Interestingly, strong social attraction (communal digging) between Drosophila larvae is present in third-instar larvae, a behaviour regulated by a peptide neuromodulator (Drosophila neuropeptide F, dNPF) [28]; this shows striking similarities to the correlation of social feeding and the expression of a neuropeptide Y receptor homologue (NRP-1) in Caenorhabditis elegans [29]. Moreover, neurons that detect aversive environmental stimuli have been demonstrated to induce social feeding in C. elegans [30], thus providing support for the proposed relationship between environmental stress and social behaviour [31]. In contrast to C. elegans, the intimate communal digging behaviour in old third-instar Drosophila larvae does not occur in the context of food foraging behaviour but is part of the post-feeding phase prior to pupation [28]. Whereas downregulation of dNPF expression coincides with social behaviour in old and non-feeding Drosophila larvae, higher levels of dNPF expression in younger larvae seem to suppress strong larval aggregation and communal digging behaviour [28]. Therefore, it remains to be seen whether similar neural regulatory mechanisms are involved in earlier developmental stages of Drosophila larvae with respect to social affinity related to foraging for food and attraction to fungal competitors. With regard to the proximate causes of attraction to mould-infected sites, fungal-borne volatiles such as CO₂ or ethylene [32] might be perceived by fly larvae and might guide them to young mould colonies.

A general reason for the formation of social groups seems to be an adaptive response to stressful environmental conditions [1, 31]. As outlined in the introduction, filamentous fungi co-occurring with drosophilids on larval feeding sites can impede fly larval development; indeed, larval aggregations have been shown successfully to suppress mould growth [12] (Fig. 1). Consequently, the presence of fungi may indicate stressful ecological conditions that initiate attraction towards fungal patches and enhance the mutual attraction of Drosophila larvae (Fig. 2). In connection with the benefits that accrue from mould suppression, the present study demonstrates that the larval-driven inhibition of fungal development is not a mere by-product attributable to the maternal decision to aggregate eggs across patches but is the consequence of a positive response of individuals to conspecifics. Because adult density-dependent oviposition choices influence a larva's food quality and its susceptibility to natural enemies [33] or abiotic stress, as well as its probability of coming into contact with intra- and interspecific competitors, the present study demonstrates the possibility of adaptive behavioural relationships between the well-known adult gregariousness in Drosophila and communal behaviour in the immature stages. Further investigating this kind of behavioural adult-offspring correlations would strongly contribute to our understanding of the evolutionary costs and benefits of spatial aggregation in insect communities [34].

The study presented here demonstrates that fungal-infected food patches (1) attract first-instar Drosophila larvae and (2) enhance group foraging behaviour. Because the larval-driven reduction in mould growth [12] has been shown to improve the chance of larval survival to the adult stage [9, 13, 14], social attraction to fungal-infected sites may reflect an adaptive behavioural response. In connection with the maternal behaviour of aggregating eggs across substrate patches, the condition-dependent mutual attraction of larvae can be discussed as a communal defence behaviour against competing mould. Given that filamentous fungi seriously deteriorate the developmental conditions for insect larvae, mould may constitute one important ecological factor that has, at least in the larval stages, facilitated social attraction in Drosophila. Thus, this study highlights the largely unappreciated role of inter-kingdom competition as a potentially important driving force in the evolution of insect behavioural traits. In general, the group formation of Drosophila larvae in response to well-defined ecological conditions might be an interesting model system for the study of proximate and ultimate aspects of social biology.