Foraging behaviors in Poneroids and Ectatomminae

Diversity and evolution of the foraging behaviors in Poneroids and Ectatomminae ants (Hymenoptera: Formicidae)

(This article has been prepared by Benoit Guenard.)

Introduction
To fill the need of reproduction, individuals have to ensure their energy requirements during their development and adult-life by collecting food resources (Schatz et al. 1997). From this perspective, insect societies can be perceived according to the superorganism theory as a sum of individuals forming a colony which are governed by similar rules as solitary individuals (Wheeler 1960). However, the fact that numerous individuals cooperate together in a nest allows the colony to share different tasks and maximize the use of space and time. While a solitary individual has to perform the seeking of a mate, the reproduction, the collection of food, and the rearing of its offspring at different time periods and at a singular location for each of them, the insect societies can perform all or most of those tasks at the same time and at different locations (Hölldobler and Wilson 1990, Passera and Aron 2006). This task partitioning allows the development and the evolution of new strategies especially to collect food resources.

Contrary to other social Hymenopterans or termites, ants represent a perfect system to study the foraging strategies in social insects due to their non-flying moves and non-cryptic foraging behavior (for most of the species). The Formicidae family appeared approximately 150 million years ago (Moreau et al. 2006) and colonized most terrestrial ecosystems where they play important ecological roles. They are considered as dominant among arthropods in many ecosystems (Hölldobler and Wilson 1990, Passera and Aron 2006). The recent phylogenetic studies have divided the Formicidae in three distinct clades that differentiate from each other between 120 and 150 million years ago (Moreau et al. 2006, Brady et al. 2006). The first separation occured when the Leptanilloide clade differentiated from the other Formicidae, and then Formicoid clade differentiated from the Poneroid clade.

Because of their ecological and taxonomical dominance (representing respectively 74% and 79% of the known living genera and species), many experimental studies were performed on the subfamilies that constitute the Formicoid clade, and many books and reviews give a specific focus on the foraging strategies in this group (Carroll and Janzen 1973, Traniello 1989, Hölldobler and Wilson 1990, Ruano et al. 2000 Passera and Aron 2006). The three subfamilies that formed the Formicoid clade, the Myrmicinae, the Formicinae and Dolichoderinae are especially interesting because of their morphological and sociobiological traits. Those clades are generally referred as the “advanced” Formicidae. Poneroid clade is considered as the “primitive” or “archaic” clade of Formicidae (Traniello 1989, Baroni-Urbani 1993). However, the Poneroids still represent close to a quarter of the known living genera and a fifth of the species.

For this review, emphasis will be given on the Poneroid clade which includes the Ponerinae, the Amblyoponinae, the Proceratiinae, the Paraponerinae and the Agroecomyrmecinae (see table 1 for their relative importance in the Formicidae). Because of its recent creation and phylogenetical change from the Poneroid to the Formicoid clade, specific attention will be given to the Ectatomminae subfamily (Bolton 2003). Finally, references to other Formicoid species will be used to bring theoretical reflection.

The foraging behavior of ants can be described as a sum of complex behaviors that can be broken in several chronological sequences (Dejean 1991): territory exploration to find food resource, resource localization, identification and the attack or collection of resource. Then according to the foraging behavior, which is function of several factors, the prey can be carried to the nest, or the scout can go back to the nest to recruit other workers. With the exception of nomad ants (Brady 2003), the three first sequences, territory exploration, food localization and its collection can be considered as relatively similar for most of the species of ants, although some interesting nuances exist. For this reason, this review will focus on the last sequence which is the recruitment or non recruitment of nestmates and on the different recruitment strategies found. However, as it is impossible to consider the evolution of a behavior without integrating the other sequences, I will refer in some parts to previous sequences. Finally, the different factors that influence the emergence of specific strategies and its evolution through the phylogeny of ants will be discussed.

Generally, four main strategies are recognized in ants (Carroll and Janzen 1973, Beckers et al. 1989, Passera and Aron 2006). The first one which does not imply any recruitment is called solitary foraging. Then, three main modes of recruitment are usually distinguished based on the solicitation of the food discovered (also called scout) and the importance of the chemical communication (Passera and Aron 2006). Those three modes are tandem-running, group-foraging and mass-foraging. However, inside each type of strategy, different patterns can emerge. In this paper, I will briefly refer to a fourth mode of recruitment recently discovered: tandem-carrying, also called carrying behavior (Guenard and Silverman, in prep.).

Table 1: Number of genera and species in each subfamily and their respective proportion in the Formicidae family. (F) Formicoid, (P) Poneroid, (L) Leptanilloid. In bold are the subfamilies which are the focus of this review.

Solitary foraging
Also called individual foraging, solitary foraging is a strategy in which a worker will discover, capture and transport the food without any systematic cooperation or communication with other nestmate workers (Beckers et al. 1989). This foraging strategy is found in species of all the different subfamilies of ants (Jaffe 1984, Beckers et al. 1989, Peeters and Crewe 1987, Hölldobler and Wilson 1990, Baroni Urbani 1993), with the exception of the army ants (Aenictinae, Dorylinae and Ecitoninae) (Brady 2003) and is considered as the ancestral state of the different foraging behaviors and the simplest strategy possible (Baroni Urbani 1993).

If solitary foraging is ubiquitous in ants, it is probably because this strategy is more efficient in collection of low densities food resources that are easy to capture and collect by a single worker (Caroll and Janzen 1973, Bernstein 1975). In such situations, the recruitment of other workers will only increase the cost of foraging without increasing the benefit of the resource collected. For this reason, many species have developed other strategies to recruit workers to collect larger prey or more difficult resources (Table 2). However, some species use only solitary foraging, even in presence of an important resource. In general, strict solitary foraging is associated with species collecting scattered and unstable food sources (Caroll and Janzen 1973, Bernstein 1975). In the Amblyoponinae subfamily, this is the case in the Neartic centipede specialist Stigmatomma pallipes (as Amblyopone pallipes) (Traniello 1978), the Ethiopian Stigmatomma pluto (as Amblyopone pluto) (Beckers et al. 1989), or the East Paleartic Stigmatomma silvestrii (as Amblyopone silvestrii) (Ito 1993). Many species of the Stigmatomma genus, are specialist predator, some like Stigmatomma pallipes (as Amblyopone pallipes) and Stigmatomma silvestrii (as Amblyopone silvestrii) are specialist predator of centipedes (Ito 1993). In the Ponerinae subfamily, exclusive solitary foraging is found in the Oriental Diacamma rugosum (Fukumoto and Abe 1983), in three species of the Neotropical Dinoponera, Dinoponera australis (Fowler 1985), Dinoponera gigantea (Fourcassie and Oliveira 2002) and Dinoponera quadriceps (Araujo and Rodrigues 2006), the oriental species Harpegnathos saltator (Maschwitz and Schonegge 1983), the neotropical Odontomachus bauri (Oliveira and Hölldobler 1989, Ehmer and Hölldobler 1995) and Odontomachus chelifer (Raimundo et al. 2009) and the Ethiopian Odontomachus troglodytes (Dejean and Bashingwa 1985), in the Neotropical Neoponera apicalis (Fresneau 1985), Neoponera obscuricornis (Traniello and Hölldobler 1984) and Pachycondyla villosa (Dejean et al. 1990), the Ethiopian Pachycondyla berthoudi (Peeters and Crewe 1987), Pachycondyla hottentota (Dean 1989) Pachycondyla sennaarensis (Lachaud and Dejean 1994) and Pachycondyla soror (Dejean 1991), the Ethiopian Platythyrea conradti (Peeters and Crewe 1987), the Neartic Ponera pennsylvanica (Pratt et al. 1994) and the Neotropical polyxenid specialist Thaumatomyrmex contumax (Brandao et al. 1991). For the Ectatomminae, only the Neotropical Ectatomma opaciventre was described as strictly solitary forager (Pie 2004).

It is interesting to note that if in some species a strict solitary behavior is expressed during foraging activities, other recruitment behaviors are expressed during colony migration, especially tandem-running or carrying behavior. This was found in Diacamma rugosum (Fukumoto and Abe 1983), Neoponera apicalis (Fresneau 1985), Neoponera obscuricornis (Hölldobler and Traniello 1980), Pachycondyla sennaarensis (Lachaud and Dejean 1994) and Ponera pennsylvanica (Pratt et al. 1994). Recruitment strategies are possible in those species but are never expressed for foraging which demonstrates some flexibility abilities for recruitment, but some ecological constraints in their expression to forage (Lachaud and Dejean 1994).

Several traits associated with solitary foraging, but not only, were noticed by many studies. One of them is the spatial fidelity (also called “Ostrue” or path fidelity) observed in the Ectatomminae Ectatomma opaciventre (Pie 2004), the Paraponerinae Paraponera clavata (Harrison et al. 1989), the Ponerinae Diacamma rugosum (Fukumoto and Abe 1983) Dinoponera gigantea (Fourcassie and Oliveira 2002), Odontomachus bauri (Ehmer and Hölldobler 1995), Neoponera apicalis (Fresneau 1985) and Pachycondyla tarsata (Dejean et al. 1993). In those cases, foragers specialize in particular foraging zones on which they orientate using visual cues (Dejean et al. 1993a, b). Foragers explore recurrently the same “sub-territory” which enables them to orient quickly to food resources and to return to the nest quickly. Also, this enables them to reduce the time effort in food research and also to reduce the external constraints such as disorientation, competition and predation (Harrison et al. 1989, Dejean et al. 1993b, Pie 2004).

Short range recruitment
Although the recruitment of workers from the nest to the food resource by a successful scout is absent in solitary forging, two types of indirect recruitment can be distinguished. The first one is the short distance recruitment also called “kinopsis” by Stager (1931) (in Dejean 1991). The successful scout who finds food recruits workers that can be at the vicinity of the discovered food resource. This behavior was observed in Stigmatomma pallipes (as Amblyopone pallipes) (Traniello 1982) and Pachycondyla soror (Dejean 1991). In the last example, the scout turns around the prey (termite piles or grasshopper) touching the ground with the apex of the gaster. This behavior attracts the surrounding workers that were hunting and then attack the prey. Finally, the workers can cooperate to transport the prey until the nest (Dejean 1991).

Social facilitation
The second type of indirect recruitment is the social facilitation. In that case, the successful scout goes back to the nest and few after, the number of foragers observed leaving the nest increase. In this case, the foragers don’t show a specific direction toward the food resource, and some of them won’t find it (Lachaud and Dejean 1994). This behavior was observed in Ectatomma permagnum (Cogni and Oliveira 2004b), Ectatomma ruidum (Lachaud 1985), Gnamptogenys horni (Pratt 1994), Gnamptogenys moelleri (Cogni and Oliveira 2004 a,b), Paraponera clavata (Breed and Bennet 1985), Odontomachus bauri (Cogni and Oliveira 2004b), Odontomachus chelifer (Lachaud and Dejean 1994), Odontomachus troglodytes (Peeters and Crewe 1987), Pachycondyla berthoudi (Peeters and Crewe 1987), Pachycondyla sennaarensis (Lachaud and Dejean 1994), and Pachycondyla soror (Dejean 1991). This behavior was qualified by several authors as an archaic/primitive form of mass recruitment (Lachaud 1985, Lachaud and Dejean 1994). This strategy was described as efficient even if all the workers do not discover the initial food resource because it allows the discovery of other alternative food resources when those ones are distributed at random or in the form of small groups arranged in a heterogeneous way in the foraging area (Lachaud and Dejean 1994).

Prey chain transfer
If the two behaviors above implied primitive form of recruitment, the following behavior shows a primitive form of cooperation. In the prey chain transfer, the scout finds and collects the food. After the transport of this one on a certain distance, the food is transmitted to another worker which will bring until the nest (Levieux 1966, Lopez et al. 2000). This behavior is different from the group retrieval in which two or more workers cooperate to bring home a food object too large to be managed by a single individual (Hölldobler and Wilson 1990). In the prey chain transfer, although different nestmates collaborate in the retrieval of the food object, it is neither simultaneous nor a consequence of the prey weight (Lopez et al. 2000). This behavior was observed in several species, as Ectatomma ruidum (Schatz et al. 1996), Paraponera clavata (Fewell et al. 1992), Odontomachus troglodytes (Dejean and Bashingwa 1985), Pachycondyla analis (Levieux 1966) Pachycondyla sennaarensis (Levieux and Diomande 1978) and Pachycondyla tarsata (Lopez et al. 2000). In the case of the granivorous Pachycondyla sennaarensis, the seed is transmitted between the workers after palpation, which suggest a real communication between the workers (Levieux and Diomande 1978). In Pachycondyla analis, a similar observation was made, but in its polymorphic species, the termite was always transmitted from a small worker to a bigger one, but, group-raiding strategy is used in this species (Levieux 1966).

Carrying behavior
It is usually recognize that tandem-running is the most primitive foraging strategies expressed in ants (Hölldobler and Wilson 1990). However, another behavior can be described as intermediate between solitary foraging and tandem-running. In carrying behavior, a worker will carry another one between its mandibles to bring this one to a specific location. This behavior was described many times during colony movements in different subfamilies of ants (Moglich and Hölldobler 1974, Fukumoto and Abe 1983, Peeters and Crewe 1987, Hölldobler and Wilson 1990, Pie 2002, Aron and Passera 2006), or during the early phase of attack in the slaving-ant Rossomyrmex (Moglich and Hölldobler 1974, Aron and Passera 2006). However, this behavior was never described in foraging activities. Carrying behavior was recently discovered as foraging strategy of the Ponerine, Brachyponera chinensis (Emery) which use it to recruit workers on food resource (Guenard and Silverman in prep.). Preliminary results indicate that this strategy is used for food resource which are too important to be carried by a single worker or that can not be moved.

Tandem running
Tandem running is considered by many authors as the simplest recruitment behavior (Lenoir and Jaisson 1982, Agbogba 1984, Passera and Aron 2006). Jaffe (1984) distinguishes two main types of tandem running. The first one occurs when the scout attracts a nesmate using antennal contact and leads it in tandem to the food resource. The second type is the tandem running with odor signals, also called tandem calling, in which a scout uses chemical cues for either orienting itself back to the food source or to help attract nesmates. In both case, the two ants keep a physical contact with each other (Beckers et al. 1989). Tandem-running is found in the African species Hypoponera eduardi (Beckers et al. 1989) and Hypoponera sp. (Agbogba 1984), the Neotropical Odontomachus haematodus (Hölldobler and Engel 1978), the African Pachycondyla caffraria (Agbogba 1992), and Pachycondyla crassa (Hölldobler and Engel 1978), the Neotropical Pachycondyla harpax (Hölldobler and Engel 1978) and the oriental Pachycondyla tesseronoda (Jessen and Maschwitz 1985, 1986).

The emergence of tandem running can be found in Pachycondyla havilandi in which a primitive form of tandem-running can be found through a social facilitation behavior (Duncan and Crewe 1994a). Duncan and Crewe (1994a) have noticed at many occasions a duo of ants leaving the nest by tandem-running. The duo will stay formed for approximately 5 minutes and then separate. The same duo of ants could be observed at different moments of the day performing the same behavior.

Group foraging
With group foraging, also called group recruitment, the scout guides a group of nestmates (two or more individuals) at a time, laying a pheromone trail to the nest in some cases (Beckers et al. 1989). Chemicals signals are used but other signals, such antennation contacts, etc, are also important (Jaffe 1984). The size of the groups is variable between species but also between the different group formed for a given species. In Pachycondyla analis, the size of the raiding group can reach 650 individuals (Lepage 1981).

Group foraging is used in the Amblyoponinae subfamily by Amblyopone australis (Hölldobler and Palmer 1989), one species of the A. reclinata group (Ito 1993), Mystrium rogeri (Hölldobler et al. 1998); in Ectatomminae by Ectatomma brunneum (Pie 2004), Ectatomma ruidum (Schatz et al. 1997), Gnamptogenys horni (Pratt 1994), Gnamptogenys menadensis (Johnson et al. 2003), Gnamptogenys sulcata (Daly-Schveitzer et al. 2007), Rhytidoponera chalybaea (Ward 1981), Rhytidoponera purpurea (Beckers et al. 1989); in Ponerinae by Centromyrmex bequaerti (Dejean and Feneron 1999), Leptogenys attenuata (Maschwitz and Schonegge 1983), Leptogenys chinensis, Leptogenys intermedia (Maschwitz and Schonegge 1983), Leptogenys intermedia (as Leptogenys nitida) (Duncan and Crewe 1994b), Leptogenys sp. 13 (Steghaus-Kovac and Maschwitz 1993), Neoponera commutata (Mill 1984), Pachycondyla ilgii (Peeters and Crewe 1987), Neoponera laevigata (Hölldobler and Engel 1978) Neoponera marginata (Leal and Oliveira 1995), Pachycondyla tarsata (Dejean et al. 1993a), Platythyrea modesta (Djeto et al. 2001), Plectroctena mandibularis (Peeters and Crewe 1987), Plectroctena minor (Dejean et al. 2001) and Simopelta oculata (Peeters and Crewe 1987). However, some of these examples are doubtful. The distinction between group foraging and mass foraging strategies are not always clear, and several authors can make confusion between the two behaviors (Beckers et al. 1989, Passera and Aron 2006).

A primitive form of group foraging can be found in Pachycondyla caffraria which use usually the tandem-running to recruit (Agbogba 1984). However on some circumstance, groups of three individuals could be observed. In those cases, the formation of the “threesome” appears by the addition of a third individual to the existing duo or by direct formation of the “threesome” after antennal contacts (Agbogba 1984).

Mass recruitment
The chemical mass recruitment is considered to be the most complex recruitment strategy (Passera and Aron 2006). Scout ant guides hundreds or thousands of nesmates to the food source by chemical means only (Oster and Wilson 1978, Jaffe 1984).

A special case of mass recruitment can be found in the arboreal Ponerinae Platythyrea modesta. In this species, when a large prey item is captured by a worker, this one will recruit the whole colony to exploit it. In this case, even the larvae are transported outside of the nest to feed on the prey item (Djieto-Lordon et al. 2001). Similar examples were found in different species of the Amblyoponinae subfamily (Djieto-Lordon et al. 2001).

Mass recruitment was observed in the Amblyoponinae of the Australian genus Onychomyrmex, Onychomyrmex doddi, Onychomyrmex hedleyi and Onychomyrmex mjobergi (Wilson 1958, Brady 2003); in the Paraponerinae Paraponera clavata (Breed and Bennett 1985), in the Ponerinae Leptogenys diminuta (Wilson 1958), Leptogenys kitteli (Duncan and Crewe 1994b), Leptogenys processionalis (Beckers et al. 1989), Leptogenys purpurea (Wilson 1958), and Leptogenys sp. (Maschwitz et al. 1989).

Obligate collective foraging is a strategy described in army ants, in which the food resource are never discover by a solitary scout, but by a group which will discover, capture and retrieve the prey. The group is totally leaderless and only pheromonal trail are used for orientation between the hunting area and the nest (Brady 2003). This behavior was described in some Ponerinae species, Leptogenys processionalis distinguenda (as Leptogenys distinguenda) (Witte and Maschwitz 2000), Leptogenys intermedia (as Leptogenys nitida) (Duncan and Crewe 1994) and several other species of Leptogenys and probably Simopelta (Brady 2003). This behavior is also suspected in some Amblyoponinae species of the Onychomyrmex genus (Brady 2003). However, although those species express an obligate collective foraging, they don’t fill at least one of the two other characteristics of army ants (nomadism and dichthadiigyny) (Brady 2003).

Ecological factors affecting foraging strategies
Strict solitary foraging seems to be very common in species which have a specialist diet on solitary non-social (Peeters 1997). At the opposite, Wilson (1958) hypothesizes that group and mass foraging in Poneroid is an advantageeous strategy to exploit colonies of social insects, especially termites or brood in ant colonies, and large prey items. However, several strict termitophagous species use solitary foraging or tandem running (Peeters and Crewe 1987). According to Carroll and Janzen (1973), mass recruitment can appear by selection in five situations: 1) if the size of the ant colony is large enough to avoid a depletion of the population during recruitment phases, 2) if the pheromone contribution by worker is weak, 3) if in non territorial habitat, the mass recruitment ensure the protection of the food resource, 4) if the dispersion of many small worker can be the best method to locate and capture moving prey, 5) if large items of food appear erratically within a few meters of the nest. The point 1 is coherent with the finding of Beckers and collaborators (1989) in which the complexity of foraging was positively correlated with the mean colony size of the different species. For instance, in Stigmatomma, species with a colony size of one or two dozens of workers express solitary foraging, while a species of reclinatum group, which has 100 workers per colony is able to retrieve centipeds cooperatively by group foraging (Ito 1993, Peeters 1997).

Poneroid and Ectatomminae clades express the complete diversity of foraging strategies that can be found in the Formicoid clade, and even more (carrying behavior). If many species express primitive behaviors, such solitary foraging, carrying behavior or tandem running, those also exist in the Formicoid clade (Carroll and Janzen 1973, Traniello 1989, Hölldobler and Wilson 1990). Ecological constraints seem to be the main factors that regulate the foraging strategies expressed in ants (Jaffe 1984, Peeters and Crewe 1987, Baroni Urbani 1993). For instance, in a study on the effect of high temperatures in ant foraging in Formicinae and Myrmicinae, Ruano et al. (2000) found that solitary foraging was very common is such habitat. This can be explained by the fact that collective foraging strategies use pheromonal trails for recruitment and orientation. However, the molecules used are highly volatile and their efficiency duration decreases with temperature. In such habitat, foraging strategies seem then more adapted. Biotic interactions could also regulate the foraging strategies. Hunt (1983) proposes that the foraging strategies evolution could be mediated by predation.

High flexibility of foraging strategies is also found for a given species. Factors such geographic location (Lepage 1981), the season (Lepage 1981, Fewell et al. 1996, Cogni and Oliveira 2004a), the brood presence (Lachaud and Dejean 1994), the distribution of food (Longhurst and Howe 1979), the food quality (Cogni and Oliveira 2004), quantity/weight of food (Dejean 1991, Schatz et al. 1997, Dejean et al. 2001, Johnson et al. 2003, Dali-Schveitzer et al. 2007), the distance to the food (Dejean 1991, Fewell et al. 1992) modify the foraging strategies in Poneroid and Ectatomminae clades. Such flexibility has sometimes disorientated some authors on previous observations, like in Ectatomma ruidum (Schatz et al. 1997) or Paraponera clavata (Breed and Bennett 1985). For food retrieval, some species express a graduated recruitment, e.g. Paraponera clavata, (Breed et al. 1987), Ectatomma ruidum (Schatz and al. 1997). Graduated recruitment is the individual decision of the scout to recruit and to adjust the number of responding foragers in a predictable manner depending upon resource characteristics (Breed and al. 1987). This strategy maximizes the food collected with a minimum of cost.

Conclusion
This review presents various studies demonstrating that the poneroid and the Ectatomminae clades show a large range of foraging recruiting behavior. This idea is supported by the fact that those subfamilies are considered in ants as the ones having the most diverse recruitment and trail pheromone glands (Janssen et al. 1999). The behaviors expressed range from the simplest solitary foraging all along a gradient of complexity until the mass foraging behavior. The full gradient of complexity can be found inside a same genus (e.g: Ectatomma, Leptogenys, Pachycondyla, or even in the primitive Amblyopone + Stigmatomma, see table 2). Previous studies seeking for a correlation between phylogeny and a gradient of complexity within foraging behaviors concluded that the ecological conditions shaped the response to the ants and no clear direct evolutive pathways could be developed (Jaffe 1984, Peeters and Crewe 1987, Baroni Urbani 1993, Passera and Aron 2006). In 1987, Peeters and Crewe conclude: “the occurrence of simple or complex hunting strategies does not reflect phylogenic relationship”. The recent revision of the Poneroid clade (Bolton 2003) and the new phylogeny (Brady et al. 2006, Moreau et al. 2006) could change this conclusion. However, although more data analysis would be necessary, it does not seem obvious in the light of this review that a correlation between phylogeny and foraging complexity could be established.