Resource discovery versus resource domination in ants: a functional mechanism for breaking the trade‐off

New Perspectivesis intended to allow the communication of comments, viewpoints and speculative interpretation of issues in ecology pertinent to entomology, and to fuel discussion and debate.

Abstract

The importance of competition in structuring natural communities is often a contentious issue, but not for myrmecologists (Hölldobler & Wilson 1990). Perhaps because eusociality buffers colonies against predation and the physical environment, and because so many ants act as generalized scavengers and exudate-feeders, competition appears to be ubiquitous in ant communities. In Hölldobler and Wilson's review of this subject, for example, factors ranging from overt aggression to over-dispersion of colonies, over-dispersion and complementarity of traits, exclusion experiments, and natural and experimental introductions, are cited as evidence for competition. With the importance of competition largely resolved, a question arising naturally is whether there exist regularities in the assembly of communities of ants contesting many of the same resources. Wilson (1971) anticipated this question in his early recognition of three categories of ants: opportunists, insinuators, and extirpators. Opportunists specialize in exploitative competition, i.e. in discovering food quickly and exploiting it before other ants arrive. In contrast, extirpators dominate food finds aggressively, and insinuators insert themselves inconspicuously to thieve food from the behavioural dominants. As Wilson was quick to point out, the strategies can vary intraspecifically and be context dependent (see also, e.g. Morrison 1996). A decade ago, working with a community of woodland ants, Fellers (1987) took these ideas further by postulating an evolutionary trade-off between resource discovery (finding and using resources in advance of competitors) and behavioural dominance of resources, once located (attacks and/or avoidance) (Fig. 1). Fellers equated discovery ability with the more familiar exploitative competition, and recognized behavioural dominance at baits as encounter competition by Schoener's (1983) classification (Box 1). Encounter competition was considered to be a component of interference competition, which also includes territoriality, a trait not exhibited by any of Fellers’ ants. In her study of temperate woodland ants, Fellers found that, except at distances very near the colony, rates of discovery of food baits were independent of distance to the colony but characteristic of particular species. A suite of three exploitative competitors found baits rapidly, but nearly always relinquished them to more dominant species. A pair of encounter species was usually the last to arrive at baits, but almost invariably controlled them once found. The three remaining species exhibited intermediate characteristics in both resource discovery and behavioural dominance at baits (Fig. 1). Species interacted frequently over food baits; more than half of the interspecific encounters in Fellers’ study resulted in some form of encounter competition, evaluated as aggression (fights or chases) or avoidance. In data from Fellers (1987; Table 17), dominance rankings appear to be inversely correlated with discovery rankings. The pattern suggests that an evolutionary trade-off between interference and exploitative competitive ability (interference including only behavioural dominance in encounter interactions) may facilitate coexistence of ants in a temperate woodland community. Evolutionary trade-offs between exploitative competitive ability and behavioural dominance may be common in ant communities, as well as occurring in social bees (Nagamitsu & Inoue 1997) and even solitary vertebrates (Brown et al. 1994). Such trade-offs may permit the coexistence of species that span the range of expression of these negatively correlated traits (Fellers 1987; Savolainen & Vepsäläinen 1989; Morrison 1996; Holway in pressa). For example, a pattern similar to that in Fig. 1 has recently been documented by Holway (in pressa) for native ants in riparian woodlands within California's Central Valley. Interestingly, however, a non-native invader of these communities (Linepithema humile, the introduced Argentine ant) was not bound by the trade-off between exploitative and encounter competition, and was gradually replacing the native ant fauna. Not only did L. humile find artificial baits more quickly, but it dominated them more consistently than did the natives. Moreover, behavioural dominance by L. humile was correlated with numerical dominance to consolidate the ecological dominance of this species (Box 1). Indeed, wherever L. humile had replaced the native ants, total worker numbers exceeded those summed over all species in adjacent areas lacking L. humile (Holway in pressb). The capacity of Argentine ants to break the trade-off between exploitative competitive ability and behavioural dominance can be viewed from the broader perspective of other exotic or native ant species that may have done the same. Thus, to the extent that coexistence in native ant communities depends on the trade-off, the replacement of native ant communities by exotic species may imply that, like L. humile, other exotics may have broken the trade-off. For example, where introduced to the southern United States, fire ants (Solenopsis invicta) are replacing many native ants in the introduced range of this species (Porter & Saviganano 1990). A possible factor in the success of this pest is that, whereas native ants are beset by various natural enemies, like species-specific parasitoids (Diptera, Phoridae), which reduce their capacity to dominate baits, the range expansion by S. invicta has occurred in the absence of natural enemies that restrain its activities in its native range (Orr et al. 1995). An absence of natural enemies in the introduced range might also explain the success of Argentine ants in North America (Orr & Seike 1997). If behavioural dominance (or encounter ability) increases conspicuousness to natural enemies, e.g. if recruitment and/or alarm pheromones attract such enemies, then the trade-off between exploitative and interference competition might be accounted for wholly or in part by a trade-off between behavioural dominance and susceptibility to enemies (see also F. R. Adler, unpublished data). In the absence of natural enemies, ants are then free to evolve behavioural dominance as well as exploitative competitive ability. This explanation, however appealing for invasive ant species, is less likely to account for cases where certain ant species have become ecologically dominant in their own native habitats, i.e. in the presence of natural enemies with which they have evolved. For example, at boreal latitudes, communities of taiga ants typically include not only aggressive, non-territorial species (Fellers 1987, encounter species and Wilson 1971, extirpators) and relatively submissive species (including Fellers’ exploitative competitors and Wilson's opportunists and insinuators), but also territorial dominants in the Formica rufa-group (Vepsäläinen 1982; Box 1). Whereas submissives defend just their nest sites (type I territoriality), and extirpators defend both nests and food finds (type II territoriality), territorial dominants defend real spatial territories against most extirpators and some submissives (type III territoriality, Vepsäläinen 1982). As in L. humile, behavioural dominance is correlated with numerical dominance (Vepsäläinen 1982; Savolainen & Vepsäläinen 1989). Similarly, in lowland tropical rain forests, a suite of ecologically dominant ants defends interspecific territories that contribute to a mosaic of different ant communities in the arboreal zone (reviewed in Wilson 1971; Hölldobler & Wilson 1990). Each of the various dominants is associated with a particular set of subordinate ant species, as well as with other arboreal arthropods, which have been ecologically sorted or have evolved for their abilities to coexist with the ecological dominants (see also Morrison 1996). Thus, some subset of Wilson's (1971) extirpators and Fellers's (1987) encounter species has evolved the capacity to defend space as well as food resources. In these species, territoriality augments encounter interactions as another form of interference competition (Morrison 1996; Davidson 1997; Box 1). Behavioural dominance is again positively correlated with numerical dominance (Vepsäläinen 1982; see also Morrison 1996). Ecological dominance of ant communities by spatially territorial species, to the exclusion of many or even all other ants, is also characteristic of other natural ant faunas. It occurs, for example, in mangroves (Cole 1983) and in arid Australia (Greenslade 1979), as well as on islands invaded by, or comprising entirely, introduced species (Fluker & Beardsley 1970; Clark et al. 1982; Morrison 1996). Do these ecological dominants use territoriality to break the trade-off between exploitative competitive ability and behavioural dominance? This appears likely for dominant ant-garden ants of western Amazonia. Together, Camponotus femoratus and Crematogaster limata var. parabiotica co-inhabit arboreal carton nests and forage together in a relationship called parabiosis (Davidson 1988), a term that is rightly still non-committal as to the fitness consequences for the two species. In 1984, D. W. Davidson (unpublished data) placed a total of forty cheese baits and forty honey baits out in the lower arboreal zones (< 3 m high) of five different ant-garden areas (eight baits per area), and found that one or the other or both of these species were the first ants to discover thirty-seven of forty cheese baits and thirty-four of thirty-six honey baits (four honey baits were not discovered during the census). Despite their proficiency in locating baits (and despite attacks by phorid flies), by the end of 60-min sampling periods these same ants eventually dominated all thirty-six of the honey baits and thirty-nine of forty cheese baits (personal observation). Together with exploitative and encounter competition, interspecific territoriality by one or both of the ant-garden ants (Davidson 1988, 1997) almost certainly contributed to the abilities of these species to both discover and control baits. Across ants as a whole, defence of actual spatial territories is highly correlated with ecological dominance. It is typical of many exotic pest species (articles in Williams 1994), including L. humile (Holway in pressa), as well as of dominants in rain forest canopy and taiga communities. If territoriality enables ants to break the trade-off between resource discovery and resource domination, one might ask why all ants are not territorial. In the following section, I explore functional constraints other than those involving increased vulnerability to natural enemies (see, e.g. F. R. Adler, unpublished data). Are there constraints on the evolution of territoriality and ecological dominance, and, if so, what mechanistic features of ant biology might be involved? Neither large worker body size, nor potent chemical weaponry alone is sufficient to account for behavioural dominance. Thus, although behavioural dominance was somewhat correlated with body size in Fellers's (1987) study, rapid recruitment by ants often compensates for small size, allowing behavioural dominance to parallel colony size closely (e.g. Holway in pressa). Fellers also noted that two of the dominant formicines employed offensive chemical weaponry (formic acid) in contests with other ants. However, a dolichoderine (Tapinoma sessile), also chemically defended, remained the second most subordinate species. Futhermore, in Holway's (in pressa) study, the offensive and defensive exocrine product of L. humile was no more powerful or effective in repelling native ants than was the similarly concentrated exocrine product of another chemically-defended species. Chemical weaponry is widespread among ant taxa, and the possession of such weaponry cannot in itself ensure behavioural dominance. Older and recent studies are converging on the theory that phylogeny influences the capacity for breaking the trade-off between exploitative and encounter competition, and for gaining ecological dominance. Thus, in various geographical realms, dominants include representatives of a restricted number of ant subfamilies and genera (Vepsäläinen 1982; Savolainen & Vepsäläinen 1989; Anderson 1992; Davidson & Patrell-Kim 1996; Davidson 1997). At relatively high latitudes, most are members of the Formicinae [wood ants in the Formica rufa group (Vepsäläinen 1982; Savolainen & Vepsäläinen 1989)], and formicines are also prominent among encounter species in temperate woodlands [species of Camponotus, Formica, Lasius and Prenolepis (Fellers 1987)]. In both taiga and lower latitude temperate woodlands, many of the encounter species are also formicines, whereas myrmicines play subordinate or submissive roles (Fellers 1987; Savolainen & Vepsäläinen 1989). Ecological dominants of arid Australia come mainly from the Dolichoderinae (mostly Iridomyrmex), and again myrmicines tend to be subordinate (Greenslade 1979; Anderson 1992). Finally, in the tropical arboreal zone, ecological dominants include both formicines and dolichoderines, but also two myrmicine genera (Crematogaster and Myrmicaria, the latter in localized areas only). At least one ponerine (Paraponera clavata) may be evolving in that direction (Davidson 1997). What characteristics particular to these taxa might have facilitated the evolution of behavioural and numerical dominance? Eisner (1957) showed many years ago that formicines and many dolichoderines are particularly well adapted to benefit from liquid foods. Together with a distensible crop, modifications to the proventriculus (= gizzard) from the plesiomorphic state in more predatory ants, allow energetically efficient storage of large volumes of such foods. These key innovations to the structure connecting the crop (or social stomach) to the worker's individual stomach, or midgut, almost certainly evolved independently in the two subfamilies. They differentiate these taxa from most other ants and aculeate Hymenoptera, which require relatively expensive muscular contractions to hold liquid food. More recently, DeMoss (1973) observed functionally similar, but morphologically distinctive, adaptations in the proventriculus of Crematogaster, a myrmicine relying extensively on liquid foods (plant and homopteran exudates). The proventriculus has not been studied in Myrmicaria or Paraponera clavata, but because P. clavata transports exudates in the mandibles, modifications to the proventriculus are not expected (Davidson 1997). The evolution of more efficient means of storing and processing liquid food may have primed these ant taxa for increasing specialization on liquid diets, which are high in carbohydrates but very low in protein and amino acids (nitrogen) (Davidson 1997; see also White 1993). All of the territorial dominants whose diets are well studied are known to feed extensively on exudates (Davidson & Patrell-Kim 1996; Davidson 1997). This is true even for myrmicine genera (e.g. Pheidole and Monomorium), in which just a subset of species are exudate-feeders (Jahn & Beardsley 1994; Vail & Williams 1994). Despite very low concentrations of nitrogenous compounds in plant tissues (e.g. White 1993), comparisons of δ15N values provide evidence that a number of such specialized exudate-feeders obtain their nitrogen lower in the trophic chain than do more predatory species in the same habitats (Davidson & Patrell-Kim 1996). This result is consistent with Tobin's (1994) hypothesis that the extraordinary abundance of ants in tropical canopy samples occurs because the ants feed mainly as herbivores, rather than as predators. Moreover, a few ant species contribute disproportionately to the abundance of ants in canopy samples, and numerical dominance occurs mainly among heavy exudate users (Davidson & Patrell-Kim 1996; Davidson 1997). Across these same taxa, reductions in protein-rich exoskeleton may also facilitate rapid colony growth, despite nitrogen-poor diets (Davidson & Patrell-Kim 1996). How might all of this relate to trade-offs between rates of discovery and dominance of food resources? In processing the exudate volumes needed to concentrate nitrogen from the low amounts in exudates, ants obtain large quantities of carbohydrates, in excess of those that can be paired with protein for colony growth. Faced with this resource imbalance, ants might use excess carbohydrates at little cost for activities that enhance protein gain. A resource-balance model (Davidson 1997), similar to those used by plant biologists (Bryant et al. 1985), suggests that, among ecologically dominant ants of the tropical arboreal zone, excess carbohydrate appears to be directed toward: (1) high tempo activity (rapid motion; Oster & Wilson 1978), associated with high dynamic densities (Hölldobler & Wilson 1990) of workers and, presumably, rapid rates of resource discovery, (2) defence of absolute spatial territories, and (3) investment in nitrogen-free (or mainly carbon-based) offensive and defensive chemical weaponry (Fig. 2). (Contributions to carbon-based recruitment chemistry are also a possibility; Davidson 1997.) Whereas the latter two functions enhance interference ability, the first should be correlated directly with rates of resource discovery, i.e. with exploitative competitive ability. Thus, evolutionary innovations providing carbohydrates in excess may enable these ants to be masters of all trades in competition for limiting nitrogen. Although these arguments were made for ecological dominants of rain forest canopies, they might also apply to temperate-zone dominants and many exotic pest species that rely heavily on homopteran exudates (Savolainen & Vepsäläinen 1989; Porter & Saviganano 1990; Williams 1994; Holway in pressa). Two contemporary phylogenetic hypotheses (after Baroni-Urbani et al. 1992; Shattuck 1992) for relationships among the subfamilies of ants (after Davidson 1997). Extinct ant subfamilies were omitted from the tree on the left. Notations to the right of ant taxa: E = exudate-feeders; P = proventriculus modified (or possibly modified in Crematogaster) for handling large liquid or = structure = exocrine are nitrogen-free (or mainly carbon-based in and may or may not be taxa, some of the a common However, Eisner (1957) that modifications of the proventriculus evolved independently in the two because in the two subfamilies (formic in formicines, and and in and are of different in formicines and in nitrogen-free evolved independently in the two is not to ants. White for example, has that it is characteristic of and White nitrogen as a mainly to which have adapted in and independent However, many might that in of nitrogen the evolutionary of competition for limiting nitrogen. The from ants appears to that of carbohydrates might have been employed in evolutionary over limiting allowing particular taxa to the trade-off (Fig. and to the This might have in the evolutionary of on Thus, in the from to and then rates have increased in a from to more recent This of an per of body may a evolutionary for to that have evolved to and/or use to more for limiting protein (see also and to break the trade-off. within ant taxa by a proventriculus modified for in the of behavioural and numerical dominance (Fellers 1987; Savolainen & Vepsäläinen 1989; Davidson & Patrell-Kim 1996; Davidson 1997). For example, just a small subset of species in such taxa has the of ecological dominance typical of territorial species in the tropical dominance excess carbohydrates appear to be a but not sufficient for behavioural and numerical dominance (Davidson 1997). Although are as factors likely to be are nests per colony, ubiquitous in ecological dominants of both and arboreal (Savolainen & Vepsäläinen 1989; Williams 1994; Davidson 1997; Holway in pressa) and an ability to nest (Hölldobler & Wilson 1990). the of the colony to the resource (Hölldobler & including relatively and of worker numbers to resource abundance Moreover, it enables high nest densities and of space that might be to (Hölldobler & Wilson 1990). In the rain forest nest of carton or or of bound by carton or allow colonies of arboreal species to food resources rather than the of nests (Davidson 1997), e.g. in which may be most at low and relatively of the species, the expression of dominance may be on the of nest sites (Fluker & Beardsley Finally, in to many of the ant species, and most of the invasive species, are by per mainly or by or with lower than by solitary and lacking among Williams 1994; Davidson 1997; Holway in pressa). tropical ecological and some appear to be with to all three characteristics (Davidson it may be for the evolution of worker that may be in activities to resource and relatively rapid and of space in with by (Hölldobler & Wilson 1990). and high colony and worker densities in species the lower investment in both and competitive more study is needed to the extent to which carbohydrate nitrogen and or in ecological dominance and allow particular species to evolutionary trade-offs between exploitative and interference competition. data are both introduced Argentine ants and native ant-garden ants appear to have broken the trade-off, and these as well as many other spatially ecological dominants rely extensively on exudate resources to Williams 1994; Davidson 1997). The of large homopteran may be to the of a large crop of whose activities are mainly by Moreover, homopteran and territoriality may even be to and because the of territories of competitors and in for members of the to with their colonies, rather than independent colony et al. 1993). In at least one species, food have also been to break the of colonies 1993), and one might that a similar might with the of homopteran Finally, of the trade-off on competitive for breaking the trade-off. from the relationship in Fig. 1 as a trade-off between ability and vulnerability to natural enemies (Orr et al. & Seike 1997; F. R. Adler, unpublished data). (unpublished data) has also evolutionary of the trade-off, including the very different consequences of competition involving interference exploitative I to Holway for of the ideas for key and for his review of an of the Patrell-Kim Fig. and to the ideas