The ants comprise a single family, the Formicidae, within the superfamily Vespoidea and the order Hymenoptera. There are 20 extant subfamilies of ants with a total of 288 extant genera. Some 9000-10,000 species of ants have been described, and it is estimated that there may be 15,000 species of ants alive in the world today. The earliest known fossil ants are from the Cretaceous (ca 120 mya), but ants probably did not become common until the Eocene (ca 45 mya).
EVOLUTION AND ECOLOGICAL SUCCESS
Ants are now extremely successful ecologically. Currently, they may even equal the biomass of humanity. They dominate, at their size scale, many terrestrial ecosystems from latitudes north of the boreal tree line to such southern climes as Tierra del Fuego, Chile. In certain tropical forests the contribution of ants to the biomass is spectacular. In Brazilian rain forests, for example, the biomass of ants has been estimated as approximately four times greater than the biomass of all of the vertebrates combined.
One of the reasons ants are so successful is that their colonies have extremely efficient divisions of labor: they evolved factories millions of years before we reinvented them. Another reason is that they can modify their immediate environment to suit themselves, much as we do. Leafcutter ants (Atta) , for example, evolved agriculture tens of millions of years before humanity developed agronomy. Furthermore, leafcutter ants also use antibiotics and symbiotic bacteria to protect the crop of fungi they grow on the leaves they collect. By contrast, weaver ants (Oecophylla) fashion homes from living leaves by sowing them into envelopes, using their larvae as living shuttles and the silken thread they produce as glue. Ants can also dominate areas by mobilizing large numbers of well-coordinated foragers; indeed, an ant colony’s foragers can be so numerous and well organized that they give the impression of being everywhere at once.
Ants can also be important as seed distributors and as seed harvesters, in the turnover of soils, and in the regulation of aphid numbers and the minimization of outbreaks of defoliating insects. Economically important pest species include the imported fire ant (Solenopsis invicta) in North America and leafcutter ants (such as Atta) in the neotropics. There are also many ecologically destructive “tramp” ants or invasive species that have been distributed to alien habitats by human commerce.
Ants and plants often have closely coupled ecological relationships. Certain plants even encourage ants by producing rewards such as energy-rich elaiosomes on their seeds to encourage seed dispersal, nutritious Beltian bodies and extrafloral nectaries to entice ants to visit their leaves and shoots (hence to remove the plant’s natural enemies while there), or even by supplying preformed homes (domatia) to invite ants directly to inhabit and thus better protect them. Although many ants are hunter-gatherers, very many species tend aphids for the excess honeydew they excrete. By “milking” aphids in this way, ants can in effect become primary consumers of plant products and by thus operating at a lower trophic level they can build up a larger biomass than obligate carnivores would be able to do. Yet most ants mix their diet by also consuming animal protein; for example, they will devour their own aphid milk cows if the latter become sufficiently abundant.
Arguably, the best evidence of the ecological success of ants is that their worst enemies are other ants.
EUSOCIALITY, SOCIAL ORGANIZATION, AND SOCIAL DIVERSITY
Except for a few species that have secondarily lost the worker caste, all ants are eusocial: they have an overlap of adult generations, cooperative brood care, and reproduction dominated by a minority of the colony’s members. Typically, an established ant colony consists of one or more queens (each of which may have mated with one or more winged males on a nuptial flight), an all-female set of wingless workers, and the colony’s brood of eggs, larvae, and pupae. The majority of queens mate only before they establish a colony. Thereafter, they store the sperm they have received.
All ants have haplodiploid sex determination. This property probably had a major role in the evolution of their eusociality through kin selection. Males are haploid, having only a single set of chromosomes, and thus the sperm that individual males produce is genetically homogeneous. Hence, the (diploid) daughters of the same mother and father are unusually closely related to one another. This is likely to have favored the evolution of female workers. Nevertheless, there can be continuing conflicts within colonies between the workers and the queen (or queens) over the sex ratios they produce and which colony members produce the males. Queens can choose to produce either unfertilized (haploid) eggs destined to become males or fertilized (diploid) eggs. The latter may develop into workers or potential new queens (gynes) generally depending on how much food they receive as larvae. The workers may or may not be sterile. Fertile workers produce viable (unfertilized) haploid eggs that can develop into males. Hence, there can be conflict both among the workers and between the workers and the queen over whose sons the colony produces. Indeed, in many species of ants with only small numbers of workers in their mature colonies, there are dominance hierarchies among the workers, who fight one another over egg production. Sometimes the queen moves with active aggression against the most dominant worker to curtail its production of sons in favor of her own. In addition, even when workers are sterile and serve one, singly mated queen, they may prefer to raise more of the queen’s daughters, to whom they are more closely related, than the queen’s sons. It is clear, though, that the apparent social cohesion of ant colonies is often partly an illusion. For all these reasons, the study of ants has had a major impact on recent pioneering evolutionary biology because these insects provide test cases by which the evolutionary resolution of the tension between cooperation and conflict can be explored.
Among ants, there is a diversity of mating systems and social organizations. So even though it is tempting to think of the typical ant colony as having a single, singly mated queen and occupying a single nest site, the diversity of social systems among the ants is in fact huge. For example, many ant species consist of facultatively multiqueened (polygynous) colonies. Indeed, roughly half of European ant species exhibit polygyny, and there seems to be no reason to regard this as an unusual proportion. Some ant colonies are founded by solitary queens; some by groups of unrelated queens that may later fight over who will be the one to succeed. Other colonies simultaneously occupy multiple nests (polydomy), a habit often associated with polygyny, while others exhibit colony fission, with both daughter colonies usually being monogynous. Most persistent polygyny is associated with the secondary adoption of queens. Unusual social systems include queenless ants, workerless ants (inquilines),and slave-making ants. In certain queenless species, the workerlike females produce other diploid females through a parthenoge-netic process called thelytoky. By contrast, certain inquilines have dispensed with the worker caste, and queens infiltrate and exploit established colonies of other species. Slave making may occur both intraspecifically and interspecifically. Interspecific slave making is also associated with nonindependent colony foundation in which slave-maker queens infiltrate established colonies of their host species, kill the host queen or queens, and produce workers that are reared by currently available host workers. The slave-maker workers raid other neighboring host colonies to capture large larvae and pupae. Such raids thus replenish the stocks of slave workers, which do all the foraging and brood rearing for the slave makers. There are also ant species in which there are polymorphic queens, others in which there are polymorphic males, and many in which there are polymorphic workers.
One of the outcomes of eusociality is that established colonies can be well defended by the workers against enemies. Thus, ant colonies are relatively K-selected; that is, they are selected to hold onto resources and to persist for long periods rather than being ephemeral, here-today-gone-tomorrow, r-strategists. Associated with this trait is the extreme longevity of ant queens. It is estimated that they can live 100 times longer than other solitary insects of a similar size. Worker populations in mature, well-established monogynous colonies range from a few tens to 20 million, and certain so-called super-colonies consist of a huge network of linked nests each with many queens. One supercolony of Formica yessensis in Japan may have as many as 300 million workers. Given such longevities and densities, it is clear that ants may also prove to be important model systems for understanding the spread of disease or the evolution of mechanisms to minimize the spread of disease among viscous populations of close kin. It is even possible that polygyny and multiple mating (polyandry) have evolved, at least in part, to promote genetic heterogeneity within colonies and thus help to minimize disease risks.
DIVISION OF LABOR
The relatively large biomass of ants in many ecosystems can be attributed not just to the way in which the ants interact with other organisms but to the way in which they interact with their nestmates in general and, in particular, to efficiencies that accrue from divisions of labor. One of the most dramatic traits associated with the division of labor among the workers is physical polymorphism, which is the presence of different physical worker forms within the same colony. In the African army ant, Dorylus wilverthi, for example, the smallest workers at 0.12 mg dry weight are only 1% of the dry weight of the largest workers (soldiers), and this relatively great size range is exceeded in certain other species (e.g., in Pheidologeton diversus, the smallest workers have a dry weight that is about 0.2% that of the largest majors). It is not just the size range that is impressive in such species but also the degree of polymorphism among the workers. Darwin, writing in On the Origin of Species, seemed well aware not only of the phenomenon but also of its implications. Indeed, one of Darwin’s most penetrating insights in his 1859 masterpiece was his suggestion that sterile forms evolved in social insects because they are “profitable to the community” and that “selection may be applied to the family, as well as to the individual.” He further suggested that once such colony-level selection had begun, the sterile forms could be molded into distinct castes “Thus in [the army ant] Eciton, there are working and soldier neuters, with jaws and instincts extraordinarily different” (Fig. 1A and 1B).
FIGURE 1 The army ant, Eciton burchellii. (A) Head of major worker. (B) Head of minor worker. (C) Head width vs. pronotum width allometry for workers. (D) Frequency-dry weight histogram for a large sample of workers. The allometrical relationship has a slope greater than 1, so larger workers (such as majors) have disproportionately large heads. The size frequency distribution is skewed to the right so relatively few of these very large majors are produced. (Drawings © Nigel R. Franks.)
Such worker polymorphism is now known to be associated with the differential growth rates of different putative tissues and body parts during the preadult stages. Indeed, the study of ants made a major contribution to the development of the concept of allomet-ric growth (Fig. 1C and 1D). Notably polymorphic genera include the army ants Eciton and Dorylus, leafcutter ants (Atta), carpenter ants (Camponotus), and members of the genera Pheidole and Pheidologeton. Indeed, Camponotus and Pheidole are the two most species-rich ant genera.
However, genera with polymorphic workers are in the minority. Approximately 80% of ant genera consist entirely of species with mon-omorphic workers, most of the remaining genera consist of species in which there are at most only two easily recognizable worker morphs, and only about 1% of genera have species in which three or more worker morphs can be relatively easily recognized within colonies.
Polymorphism among the workers is mostly associated with extreme physical specialization. Thus, Eciton majors have ice tong-like mandibles and are specialist defenders of the colony against would-be vertebrate predators or thieves (Fig. 1A). It has been shown that colonies of Pheidole pallidula can produce more defensive majors in response to stresses induced by conspecific competitors. Majors are not always for defense: large-headed majors in Pheidole and Messor serve as specialist grinders of harvested seeds. Even among such polymorphic species, however, the majority of workers belong to castes of generalists, which give their colonies an ability to respond rapidly to changes in the environment. Such generalists show behavioral flexibility not possible with the extreme morphological specialization of certain physical castes. Nevertheless, divisions of labor also occur within the majority generalist caste. Such workers typically specialize in different tasks at different times during their lives. This is known as temporal polyethism, in contrast to physical polyethism.
The sophisticated divisions of labor in monomorphic ants are being investigated. In Temnothorax albipennis, the workers show very little size variation, and colonies consist of, at most, a few hundred such workers living in flat crevices between rocks. Individual workers could easily roam all around such nests within a minute, but instead they have spatial fidelity zones; that is, they remain faithful to certain parts of the nest and the segregated tasks within such areas for months on end. The workers can even reconstruct their own spatial fidelity zones relative to one another if, and when, their colony is forced to emigrate to a new nest site because of the destruction of the old site. In this (and many if not all) ant species, younger workers tend to work deep within the nest at its safe center, tending the queen, the eggs, and the larvae. As they get older, workers tend to move progressively out from the center of the nest, and toward the end of their lives, they eventually engage in the most dangerous task of foraging in the outside world, where they are likely to meet predators and other hazards. However, the correlation between age and task is often very weak, and in an increasing number of species, it has been shown that the division of labor among monomorphic workers is extremely flexible. Workers can respond to the removal of other workers by reverting to tasks that they did earlier in their lives or, if need be, they may begin foraging even when they are very young. Thus, though age may influence what workers do, it is unlikely to be the organizing principle of the division of labor in many species. Rather, it seems that workers are continuously monitoring their workloads and the delays they experience while waiting to interact with their nestmates and will flexibly change their tasks accordingly to maximize their productivity.
COMMUNICATION AND PHEROMONES
Ants have diverse systems of communication, but by far the most important medium for signaling involves the chemicals known as pheromones. Ants can deposit chemical trails to recruit nestmates to discoveries of food, but they also use pheromones during exploration of potential foraging areas and nest sites. Many ants can also produce highly volatile chemicals to signal alarm when they encounter dangerous predators or other hazards. Different ants in different subfamilies use a remarkable diversity of glandular structures even just to produce recruitment pheromones. These may be produced from cloacal glands, Dufour’s glands, the hindgut, poison glands, pygidial glands, rectal glands, sternal glands, or even tibial glands on the back legs. Furthermore, many pheromones appear to be complex mixtures of many chemical compounds.
Pheromones can be effective in minute quantities; it has been estimated that 1 mg of the trail substance of the leafcutting ant, Atta texana . if laid out with maximum efficiency, would be sufficient to lead a colony three times around the world.
Nestmate recognition is another important aspect of communication in ants. A pleasing metaphor for the ant colony is a factory inside a fortress. Ant colonies are dedicated to the production of more ants; but workers need to “know” that they are working for their natal colony, and colonies also need to be well defended against other ants and against infiltration by other arthropods, which might tap into their resources. Ant colonies employ colony-specific recognition cues as one of their defense systems. These are often in the form of cuticular hydrocarbons that can be spread throughout the colony both by grooming and trophallaxis (the latter is usually associated with liquid food exchange). Slave-making ants circumvent the recognition cues of their slaves by capturing them as larvae and pupae—these captives are not yet imprinted on their natal colony odor but later become imprinted on the odor of the colony that kidnapped them after they have metamorphosed into adult workers. Sometimes colony-specific odors also can be influenced by chemicals picked up from the colony’s environment. Nevertheless, countless species of arthropods from mites to beetles have infiltrated ant colonies. For example, more than 200 species of rove beetle (Staphylinidae) are associated with New World army ants alone, and other groups such as mites are probably even more species rich. Often these infiltrators are called guests simply because their relationships with their host ant colony and to its resources are unknown( Fig. 2 ).
FIGURE 2 Scanning electron micrograph of a worker of Lasius flavus with a kleptoparasitic mite, Antennophorus grandis, gripping its head. The mite steals food when two workers exchange nutritious liquids during trophallaxis.
SELF-ORGANIZATION, COLLECTIVE INTELLIGENCE, AND DECISION MAKING
A rapidly developing approach to the study of ants and other social insects is the application of self-organization theories. Here self-organization can be defined as a mechanism for building spatial structures and temporal patterns of activity at a global (collective or colony) level by means of multiple interactions among components at the individual (e.g., worker) level. The components interact through local, often simple, rules that do not directly or explicitly code for the global structures. The importance of studies of such self-organization is that they can show how very sophisticated structures can be produced at the colony level with a fully decentralized system of control in which the workers have no overview of the problems they are working to solve.
A simple and very intuitive example of how ants use self-organization is found in their ability to select shortcuts. Certain ants can select the shortest paths to food sources. Indeed, where there is a short and a long path to the same food source, the decision-making mechanism can be surprisingly simple. The ants that happen to take the shorter path get there and back more quickly than the ants that happen to take the longer path. All the ants lay attractive trail phe-romones, and such pheromones are reinforced more rapidly on the shorter path simply because that path is shorter and quicker. In such cases, individual ants do not directly compare the lengths of the two paths, but the colony is able to choose the shorter one. Sometimes the shorter path is used exclusively, while at other times a small amount of traffic may continue to use the longer path. Having some traffic that continues to use the longer path is likely to be costly in the short term, but it may represent a beneficial insurance policy if the shorter path becomes blocked or dangerous. Self-organization also has a major role in such phenomena as brood sorting, rhythms of activity within nests, and building behaviors. This new approach may help to answer, at least in part, the age-old challenge of how ant colonies are organized.
