The Ants Chapter 12

CHAPTER 12. SYMBIOSES AMONG ANT SPECIES The ant society is a decidedly more open system than the lower units of biological organization such as the organism and the cell. In the course of evolution the tenuous lines of communication among the members of the colonies have been repeatedly opened and extended to incorporate alien species. Many kinds of ants, for example, adopt aphids, mealybugs, and other homopterans as cattle to provide a steady source of honeydew; a few raid colonies of other species to acquire workers as domestic slaves, or utilize the odor trails of other species, or defend common nest sites. Just as frequently, the lines of communication have been tapped by other alien species that have insinuated themselves into the colony as inconspicuous social parasites. Taken together, the hundreds of cases of interspecific symbioses among ant species that have come to light encompass almost every conceivable mode of commensalism and parasitism. But true cooperation is rare or nonexistent. No verified examples of mutualism are yet known, in which two species cooperate to the benefit of both. All of the relationships carefully analyzed to date are unilateral, with one species profiting and the other species either remaining unaffected or, in the great majority of cases, suffering from the attentions of its partner.

The “ultimate” social parasite
There is no better way to begin a survey of the social symbioses than by considering the most extreme example known, that of the “ultimate” parasitic ant Teleutomyrmex schneideri. This remarkable species was discovered by Heinrich Kutter (1950a) at Saas-Fee, in an isolated valley of the Swiss Alps near Zermatt. Its behavior has been studied by Stumper (1950) and Kutter (1969), its neuroanatomy by Brun (1952), and its general anatomy and histology by Gösswald (1953). A second population has been reported from near Briançon in the French Alps by Collingwood (1956), a third in the French Pyrenees by Buschinger (1987c), and still others in the Spanish Sierra Nevada by Tinaut Ranera (1981). Appropriately, the name Teleutomyrmex means “final ant.”

The populations of Teleutomyrmex schneideri, like those of most workerless parasitic ant species (Wilson, 1963), are small and isolated. The Swiss population appears to be limited to the eastern slope of the Saas Valley, in juniper-Arctostaphylos woodland ranging from 1,800 to 2,300 m in elevation. The ground is covered by thick leaf litter and sprinkled with rocks of various sizes, providing, in short, an ideal environment for ants The ant fauna is of a typically boreal European complexion, comprising the following free-living species listed in the order of their abundance (Stumper, 1950):  Formica fusca, Formica lugubris, Tetramorium caespitum, Leptothorax acervorum, Leptothorax tuberum, Camponotus ligniperda, Myrmica lobicornis, Myrmixa sulcinodis, Camponotus herculeanus, Formica sanguinea, Formica rufibarbis, Formica pressilabris, and Manica rubida. For some unexplained reason this little assemblage is extremely prone to social parasitism. Formica sanguinea is a facultative slavemaking species, preying on the other species of Formica. Doronomyrmex pacis, a workerless parasite living with Leptothorax acervorum, was discovered by Kutter as a genus new to science in the Saas-Fee forest in 1945. In addition, Kutter and Stumper found Epimyrma stumperi in nests of Leptothorax tuberum, as well as two parasitic Leptothorax, goesswaldi  and kutteri, in nests of Leptothorax acervorum  (Kutter, 1969).

Teleutomyrmex schneideri is a parasite of Tetramorium caespitum  and Tetramorium impurum. Like so many other social parasites, it is phylogenetically closer to its host than to any of the other members of the ant fauna to which it belongs. In fact, it may have been derived directly from a temporarily free-living offshoot of this species, since Tetramorium caespitum and Tetramorium impurum (the host species at Briançon and in the Pyrenees) are the only nonparasitic tetramoriines known to exist at the present time through most of central Europe. It is difficult to conceive of a stage of social parasitism more advanced than that actually reached by Teleutomyrmex schneideri. The species occurs only in the nests of its hosts. It lacks a worker caste, and the queens contribute in no visibly productive way to the economy of the host colonies. The queens are tiny compared with most ants, especially other tetramoriines; they average only about 2.5 mm in total length. They are unique among all known social insects in being ectoparasitic. In other words, they spend much of their time riding on the backs of their hosts (Figure 12-1). The Teleutomyrmex queens display several striking morphological features that are correlated with this peculiar habit. The ventral surface of the gaster (the large terminal part of the body) is strongly concave, permitting the parasites to press their bodies close to those of their hosts. The tarsal claws and arolia are unusually large, permitting the parasites to secure a strong grip on the smooth chitinous body surface of the hosts. The queens have a marked tendency to grasp objects. Given a choice, they will position themselves on the top of the body of the host queen, either on the thorax or the abdomen. Deprived of the nest queen, they will then seize a virgin Tetramorium queen, or a  worker, or a pupa, or even a dead queen or worker. Stumper observed a case in which six to eight Teleutomyrmex queens simultaneously grasped one Tetramorium  queen, completely immobilizing her. The mode of feeding of the Teleutomyrmex is not known with certainty. The adults are evidently either fed by the host workers through direct regurgitation or else share in the liquid regurgitated to the host queen. In any case, they are almost completely inactive most of the time. The Teleutomyrmex adults, especially the older queens, are highly attractive to the host workers, who lick them frequently. According to Gösswald, large numbers of unicellular glands are located just under the cuticle of the thorax, pedicel, and abdomen of the queens; these are associated with glandular hairs and are believed to be the source of a special attractant for the host workers. The abdomens of older Teleutomyrmex queens become swollen with fat body and ovarioles, as is shown in Figure 12-1. This physogastry is made possible by the fact that the intersegmental membranes are thicker and more sclerotized than is usually the case in ant queens and can therefore be stretched more. Also, the abdominal sclerites themselves are widely overlapping in the virgin queen, so that the abdomen can be distended to an unusual degree before the sclerites are pulled apart. The ovarioles increase enormously in length, discard their initial orientation, and infiltrate the entire abdomen and even the postpetiolar cavity.

From one to several physogastric queens are found in each parasitized nest, usually riding on the back of the host queen. Each lays an average of one egg every thirty seconds. The infested Tetramorium colonies are typically smaller than uninfested ones, but they still contain up to several thousand workers. The Tetramorium queens also lay eggs, and these are capable of developing into either workers or sexual forms (Buschinger, personal communication). Consequently the brood of a parasitized colony consists typically of eggs, larvae, and pupae of Teleutomyrmex queens and males mixed with those of Tetramorium  workers.

The bodies of the Teleutomyrmex queens bear the mark of extensive morphological degeneration correlated with their loss of social functions. The labial and postpharyngeal glands are reduced, and the maxillary and metapleural glands are completely absent. The mandibular glands, on the other hand, are apparently normal. In addition, the queens possess a tibial gland, the function of which is unknown. The integument is thin and less pigmented and sculptured in comparison with that of Tetramorium; as a result of these reductions the queens are shining brown, an appearance that contrasts with the opaque blackish brown of their hosts. The sting and poison apparatus are reduced; the mandibles are so degenerate that the parasites are probably unable to secure food on their own; the tibial-tarsal cleaning apparatus is underdeveloped; and, of even greater interest, the brain is reduced in size with visible degeneration in the associative centers. In the central nerve cord, ganglia 9-13 are fused into a single piece. The males are also degenerate. Their bodies, like those of the males of a few other extreme social parasites, are “pupoid,” meaning that the cuticle is thin and depigmented, actually greyish in color; the petiole and postpetiole are thick and provided with broad articulating surfaces; and the abdomen is soft and deflected downward at the tip.

In its essentials the life cycle of Teleutomyrmex schneideri resembles that of other known extreme ant parasites. Mating takes place within the host nest. The fecundated queens then either shed their wings and join the small force of egg layers within the home nest or else fly out in search of new Tetramorium nests to infest. Stumper found that the queens could be transferred readily from one Tetramorium colony to another, provided the recipient colony originated from the Saas-Fee. However, Tetramorium colonies from Luxembourg were hostile to the little parasites. Less surprisingly, ant species from the Saas-Fee other than Tetramorium caespitum always rejected the Teleutomyrmex. However, Buschinger (personal communication) has pointed out that the Saas-Fee population could be caespitum or impurum, or a mixture of both. In other words, the transfer might have been attempted across species.

The kinds of social parasitism in ants
Social parasitism in ants is complicated, and its study has become virtually a little discipline of entomology in itself. The source of the complexity is first the large number of ant species that have entered into some form of parasitic relationship with each other. Second, at least two and possibly three major evolutionary routes lead to the ultimate stage of permanent, workerless parasitism. Finally, no two species are exactly alike in the details of their parasitic adaptation. Table 12-1 contains a list of the known parasitic ants, together with certain essential data concerning each of them. With this information readily at hand for constant reference, we will now present what is deliberately a rather didactic review of the entire subject, attempting to make it as orderly and clear as possible from the outset.

Wasmann (1891) distinguished two classes of consociations, or myrmecobioses as Stumper (1950) later dubbed them, that occur between different species of ants. These are the compound nests, in which two or more species live very close to each other, in some cases even running their nest galleries together, but keep their brood separated; and the mixed colonies, in which the brood are mingled and cared for communally. Compound nests are very common in nature. They reflect relationships that range, depending on the species involved, all the way from the accidental and trivial to total parasitism. Mixed colonies, on the other hand, almost always come about as a result of social parasitism. Forel (1898, 1901), who was the first to use the expression “social parasitism,” and Wheeler (1901a,b, 1910a) devoted a great deal of attention to compound and mixed nests and provided a useful classification of the underlying relationships, complete with a somewhat less useful set of Hellenistic terms to label the various categories. Let us examine this classification briefly. Then we will make more interesting use of it in tracing the evolution of parasitism and other forms of symbioses.

Compound nests
Plesiobiosis. In this most rudimentary association, different ant species nest very close to each other, but engage in little or no direct communication--unless their nest chambers are accidentally broken open, in which case fighting and brood theft may ensue. The less similar the species are to each other morphologically and behaviorally, the more likely they are to cluster together in an accidental, truly “plesiobiotic” relationship. Put the other way, closely related species of ants are the least likely to tolerate each other's presence.

Cleptobiosis. Some species of small ants build nests near those of larger species and either feed on refuse in the host kitchen middens or rob the host workers when they return home carrying food. R. C. Wroughton (quoted by Wheeler, 1910a) has described a species of Crematogaster in India whose workers “lie in wait for Holcomyrmex, returning home, laden with grain, and by threats, rob her of her load, on her own private road and this manoeuvre was executed, not by stray individuals, but by a considerable portion of the whole community.”  Workers of Conomyrma pyramica  in the southern United States collect dead insects discarded by colonies of Pogonomyrmex, including corpses of the Pogonomyrmex  themselves. Our impression in the field has been that some Conomyrma colonies obtain a large part of their food in this way, to the point of preventing kitchen middens from building up near the Pogonomyrmex  nests.

Lestobiosis. Certain small species, most belonging to Solenopsis and related genera, stay in the walls of large nests built by other ants or termites and enter the nest chambers of their hosts to steal food and prey on the inhabitants. For example, colonies of the “thief ants” of the subgenus Solenopsis (Diplorhoptrum), including especially Solenopsis fugax of Europe and Solenopsis molesta of the United States often nest next to larger ant species, stealthily enter their chambers, and prey on their brood. Species of Carebara in Africa and tropical Asia frequently construct their nests in the walls of termite mounds and are believed to prey on the inhabitants (see Chapter 15). The relationship is parasitic with respect to nest sharing and predatory with respect to brood theft.

Parabiosis. In this peculiar form of symbiosis, two or more species use the same nest and sometimes even the same odor trails, but they keep their brood separate. The situation is similar to the mixed foraging flocks of birds so prevalent in tropical forests, except that in some instances at least, one species dominates and exploits another.

Xenobiosis. This symbiotic state falls just short of a truly mixed colony. One species lives in the walls or chambers of the nests of the other and moves freely among its hosts, obtaining food from them by one means or another, usually by soliciting regurgitation. The brood is still kept separate. This relationship is truly parasitic.

Mixed colonies
The following phenomena are vital in the later stages of parasitic evolution. In a sense they form categories comparable to those just cited for compound nests, although they are less than ideal because they are not mutually exclusive. Nevertheless, we favor continuing to distinguish them on the grounds that the associated terminology is the familiar one in literature dating back over nearly a century and, more importantly, the classification can still be relied upon to serve as an adequate guide through the complex relationships as we understand them.

Temporary social parasitism. This symbiosis was first clearly recognized by Wheeler (1904b) as a result of his studies of the life cycle of members of the Formica microgyna group, especially Formica difficilis. It has since been discovered in a diversity of genera belonging to several subfamilies. The newly fecundated queen finds a host colony and secures adoption, either by forcibly subduing the workers or by conciliating them in some fashion. The original host queen is then assassinated by the intruder or by her own workers, who somehow come to favor the parasite. With the development of the first parasite brood, the worker force soon becomes a mixture of host and parasite species. Finally, since the host queen is no longer present to replenish them, the host workers die out, and the colony comes to consist entirely of the parasite queen and her offspring. Temporary social parasitism is generally considered to be preceded in evolution by the re-adoption of queens by colonies of their own species following the nuptial flights. Bolton (1986b) has referred to this condition as “autoparasitism.”

Dulosis (slavery). Certain ant species have become dependent on workers of other species which they keep as slaves. The slave raids of the evolutionarily advanced species are dramatic affairs in which the slavemaking workers go out in columns, penetrate the nests of colonies belonging to other, related species, and bring back pupae to their own nests. The pupae are allowed to eclose, and the workers become fully functional members of the colony. The workers of most slavemaking species seldom if ever join in the ordinary chores of foraging, nest building, and rearing of the brood, all of which are left to the slaves. Facultative inter- and intraspecific slavemakers also occur. These less specialized forms provide an illuminating glimpse into the likely early stages of dulosis.

Inquilinism (“permanent parasitism”). In this final, degenerate stage, the parasitic species spends its entire life cycle in the nests of the host species. Workers may be present, but they are usually scarce and display atrophied behaviors. In many of these species, as for example in Teleutomyrmex schneideri, the worker caste has been lost altogether. Wilson (1971) suggested the use of the term inquilinism in preference to the somewhat more familiar expression “permanent parasitism” since obligatorily dulotic species are also permanent parasites. Inquilinism and dulosis, on the other hand, form exclusive categories; they are meant to be the streamlined equivalents of Kutter's (1969) “permanent parasitism without dulosis” and “permanent parasitism with dulosis.” The queens of some inquiline species permit the host queen to live, while others either assassinate her or else somehow, and in a procedure yet to be firmly established by experiments, induce her own workers to accomplish the task.

The occurrence of social parasitism throughout the ants
A rich variety of new parasitic species, representing almost every conceivable evolutionary stage, has been added since the time of Wheeler's classic synthesis in 1910. They continue to be discovered at such a consistently high rate as to suggest that, at this moment, only a small fraction of the total world fauna of social parasites is known. The reason for the slow uncovering of the world fauna seems clear: parasitic species tend to be both rare and locally distributed. As a rule, moreover, the more advanced the stage of parasitism, the rarer the species. Thus, we find (Table 12-1) that temporary social parasites, such as members of the Formica exsecta and Lasius umbratus groups, are often nearly as widely distributed as their free-living congeners, and a few of the species are also very abundant. Species in which dulosis is weakly developed or even facultative, as, for example, the representatives of the Formica sanguinea group, are also relatively abundant and widespread. On the other hand, extreme dulotic species, such as the members of Strongylognathus, Polyergus, and Rossomyrmex, exist in more restricted, sparser populations. Finally, the extreme workerless parasites are, as a rule, both very rare and very locally distributed. Anergates atratulus comes closest to being an exception. It has been collected over a wide area from southern France to Germany, and it has even been accidentally introduced into the United States with its host Tetramorium caespitum. Yet everywhere within this range it is still a comparatively rare ant. The great majority of other workerless parasites have been found at only one or two localities and are extremely difficult to locate, even when a deliberate search is made for them in the exact spots where they were first discovered. Usually they give the impression, quite possibly false, of having no more than a toehold on their host populations and of existing close to the edge of extinction.

Most of the known parasitic species have been recorded exclusively from the temperate areas of North America, Europe, and South America. Almost certainly this reflects at least in part the strong bias of ant collectors, most of whom reside in these areas and devote a large part of their lives to a meticulous examination of local faunas. Switzerland, for example, is the present “capital” of parasitic ants for the simple reason that both Auguste Forel and Heinrich Kutter lived there. About one-third of the 110 Swiss species are parasitic (Kutter, 1969). Europe has received the attention of the expert collector, Alfred Buschinger, and his students for over twenty years. The United States has benefited similarly from the efforts of W. M. Wheeler, the Wesson brothers, and other more recent gifted collectors, while the rich trove of species uncovered in Argentina has been due to three men who spent a large part or all of their lives in the country--Carlos Bruch, Angel Gallardo, and Nicolás Kusnezov. We believe that as the huge and still little-known tropical ant faunas are more carefully worked (there are no resident myrmecologists on the Amazon!), many more parasitic species will come to light. Some evolutionarily advanced forms are already known from tropical regions. In a recent study, Wilson (1984c) recognized four tropical parasites in the genus Pheidole, including two new species and the Zairean Pheidole neokohli, which rivals Teleutomyrmex in the extremeness of its degeneration. Equally impressive are the strange postxenobiotic Kyidris parasites of New Guinea (Wilson and Brown, 1956). Wheeler (1925) pointed out that females of the numerous species of Crematogaster belonging to the “subgenera” Atopogyne and Oxygyne, groups widely distributed in Africa, Madagascar, and tropical Asia, have all of the morphological characteristics of northern ants known to be temporary parasites in that they tend to be small and shining and to possess falcate or very oblique mandibles and large postpetioles which are attached broadly to the gaster. The last of these characteristics is usually associated with physogastry, also a common but not diagnostic feature of social parasitism. Emery (1899) recorded a highly physogastric nest queen of Crematogaster ranavalonae from Madagascar. At least two species of the Neotropical dolichoderine genus Azteca (aurita and fiebrigi) possess some of these traits. The species of Rhoptromyrmex, found in South Africa, Asia, New Guinea, and Australia also possess them (Brown, 1964a; Bolton, 1986a). A special study of such species, and any others that can be found to possess various of the “temporary parasite syndrome” of characters, might prove very rewarding to future students of tropical myrmecology.

Even so, the vast differences in quality of sampling from the major parts of the world render the matter inconclusive, and there remains the possibility that life in certain climates and environments actually does predispose ant species toward parasitism. It is true, for example, that a disproportionate number of parasitic species, especially the complete inquilines, occur in mountainous and arid regions. We have already mentioned the extraordinary diversity of parasites found in the little forest of the Saas-Fee. Among numerous other examples that can be cited are the montane species Pheidole inquilina, Pheidole elecebra, Manica parasitica, Pogonomyrmex anergismus, Pogonomyrmex colei Doronomyrmex pocahontas, and Leptothorax faberi, which together make up about half of the known inquiline fauna of North America. Temporary social parasites, along with species that can be tentatively placed in this category by virtue of their morphology, are more abundant in the colder portions of Europe and North America than in the warm temperate and subtropical portions, even though the faunas of the two climatic zones are otherwise not radically different. Even more impressively, dulosis is a common phenomenon in the colder parts of Europe and Asia but rare in the warmer parts; and not a single example has ever been reported from the tropical or south temperate zones.

It is conceivable that cooler temperatures facilitate the introduction of parasitic queens in the early evolution of the phenomenon by dulling the responses of the host colonies. We have found, in general, that if ant colonies are first chilled in the laboratory they are more likely to adopt queens of their own species, which they would otherwise attack and destroy. In nature parasite queens need not wait for winter to utilize this effect. Some degree of chilling, say to 10° or 15°C, occurs commonly during the cool summer nights in mountainous regions, right in the middle of the season of nuptial flights. It should prove instructive to study the effects of various degrees of cooling of potential host colonies on the success of introduction of queens belonging to species at any early stage of inquilinism, such as Leptothorax faberi and Manica parasitica. Useful information might also be obtained from an analysis of the behavioral effects of cooling on ant groups that most commonly serve as hosts, such as the genera Leptothorax and Formica, as opposed to those that are relatively immune to social parasitism, such as the genus Camponotus.

Other clues to the origin of social parasitism can be found in the phylogenetic distribution of the phenomenon, which is remarkably patchy. The more advanced forms of parasitism, namely dulosis and inquilinism, are almost wholly limited to the subfamilies Myrmicinae and Formicinae and are furthermore heavily concentrated in certain genera, including Pheidole, Myrmica, Leptothorax, Tetramorium, Plagiolepis, Lasius, and Formica, and in the satellite parasitic genera derived from them. Two inquilines (Myrmecia hirsuta and [[Myrmecia inquilina) have been described from the primitive subfamily Myrmeciinae. In view of the relatively small number of species known in the Myrmeciinae (about 120) and the limited amount of field study devoted to it to date, parasitism in this group may eventually be found to occur at about the same level of frequency as in the Myrmicinae and Formicinae.  The only parasites known with certainty among the Dolichoderinae, on the other hand, are the temporarily parasitic species of Bothriomyrmex''.  This relative immunity is puzzling since the dolichoderines are a relatively large, numerically abundant group of advanced phylogenetic rank.  Perhaps the explanation lies in the fact that very few dolichoderine species range into the cooler portions of the North Temperate Zone where parasitic species are most likely to evolve.  Yet it is also true that a rich dolichoderine fauna exists in subtropical and temperate Argentina, where many myrmicine parasites have been discovered.  No parasitic species of any kind are yet known in the Ponerinae, Cerapachyinae, and Dorylinae.  One can speculate almost endlessly on why this is the case.  For example, the Ponerinae are primitive (but so are the Myrmeciinae, and in any case many ponerine species form large colonies with advanced social traits).  The Dorylinae engage in frequent nest changes (but many parasitic beetles, millipedes, wasps, and other arthropods emigrate with them along their odor trails).

An important lead from the phylogenetic distribution has emerged recently from studies by Buschinger, Alloway, and their co-workers on the myrmicine tribe Leptothoracini. Although leptothoracines represent fewer than 3 percent of the 8800 described ant species, they contain 30 (15 percent) of the 200 known parasitic species. From the research of Buschinger, Francoeur, and their co-workers, utilizing a combination of cytological, morphological, and behavioral traits, we can be reasonably sure that slavemaking alone has arisen a minimum of six times within the Leptothoracini: once each in the lines leading to Temnothorax duloticus, Harpagoxenus, Temnothorax americanus, Chalepoxenus, Epimyrma, and Myrmoxenus. As Buschinger has said, “The myrmicine tribe Leptothoracini comprises an astoundingly rich variety of socially parasitic genera and species. New species can be found nearly everywhere when populations of independent species are closely examined.”

What is the cause of the vulnerability of the leptothoracines to social parasites? On the other side, what inclines so many to turn into parasites? Buschinger (1970, 1986) and Alloway et al. (1982) believe that the key predisposing traits are polygyny, the regular occurrence of multiple laying queens in colonies, and polydomy, the spread of colonies to multiple nest sites. All these traits are developed in Leptothorax (= Myrafant) and Leptothorax (s.str.), the preeminent northern hemisphere representatives of the Leptothoracini and the stock group from which most of the parasitic genera and species have arisen. To be structured in this manner means that the colonies are relatively “open,” in other words they are more easily invaded by alien queens of the same or different species. Polygyny usually arises by the re-adoption of queens after they have mated outside the nest. This habit thus being fixed in the workers' repertory, colonies are more susceptible to invasion by “cuckoo” queens able to provide the right chemical cues. Polydomy often results in the creation of outlier nests containing only workers and immature forms. It is possible that these queenless fragments add to the general vulnerability of the colonies.

If the view be accepted that colony structure can predispose species for or against social parasitism, how might this explanation apply to the apparent scarcity of parasitism in the tropics? It is possible that for some reason ant species with leptothoracine-type biology are rare in warmer climates. However, our knowledge of the social organization and life cycle of the vast majority of tropical ant species is too meager to search for this correlation. Buschinger (personal communication) has offered one promising hypothesis: “In my opinion a very important factor might be that parasitism occurs most frequently when the host species form dense, large, and homogeneous populations. This is the case in many temperate-zone species, whereas in warmer areas a high species diversity is often combined with rarity and wide dispersal of nests of a given species. I had presumed this for a long time, and recently in Australia I found a perfect confirmation of this idea (I also did not find any parasites there, but did not check Myrmecia nests, which often form dense populations--they were too aggressive!).”

Another factor to examine is the means by which workers recognize colony odor at the species level. If they have a fully innate, “hard-wired” recognition separating individuals of the same species from those of different species, they will be very resistant in evolution to the intrusion of social parasites. If, on the other hand, they learn the species odor early in life, they can be more easily duped. A worker captured by a slavemaker while still in the pupal or callow (newly eclosed) adult stage can be imprinted on the odor of the captor. It will serve automatically as a slave thereafter.

The relative flexibility of early learning of conspecific brood labels might also play a role. Extensive research has been conducted on the recognition of brood (immature stages) by ants of the genus Formica (Jaisson, 1975, 1985; Le Moli and Passetti, 1977, 1978; Jaisson and Fresneau, 1978; Le Moli and Mori, 1982). The results demonstrate that young adult workers learn to recognize whatever species of brood they encounter within a period of approximately one week after their emergence from the cocoon. Unfamiliar brood pieces, whether of another species or their own, are rejected or destroyed. It has been suggested that an early learning of brood labels has favored the repeated evolution of slavemaking and social parasitism among ants. Le Moli (1980) and later Brian (1983) argued that the only species suitable as hosts or slaves are those in which brood recognition is based on learning without any bias favoring individuals of the same species. This preadaptive flexibility would ensure that immature slave or host ants, when eclosing, would learn the brood labels of their own species as well as those of their parasites. The Brian-Le Moli hypothesis is only partially supported by the evidence, however. Formica rufa and Formica lugubris, which do not exhibit a bias for learning conspecific brood labels, are not themselves victims of parasitism, although they are temporary social parasites of other Formica species (Gösswald, 1951a; Kutter, 1969). Formica polyctena, which is usually free-living, is occasionally parasitized by Formica truncorum (Kutter, 1969). Preference for early learning of conspecific brood labels has been found in Camponotus (Carlin et al., 1987a,b). It is tempting to attribute the “immunity” to social parasitism of the genus Camponotus to this bias of learning conspecific brood labels. These facts appear to support the Brian-Le Moli hypothesis. However, since the bias discovered in Camponotus is not exclusive, colonies should still be vulnerable to potential parasites. Le Moli cites as supporting evidence his discovery that Lasius niger does not learn brood labels at an early age, and he uses the trait to explain why Lasius niger is evidently immune against social parasitism. However, he overlooks the fact that Lasius niger serves as the host of the temporary social parasites Lasius umbratus (Crawley, 1909; Gösswald, 1938a; K. Hölldobler, 1953) and Lasius fuliginosus (Andrasfalvy, 1961).

To summarize, we can recognize several predisposing features toward social parasitism that might explain why it occurs in some ant genera and not others. Those most likely to take the step (1) live in cool or arid climates, (2) have multiple queens as a result of re-adoption of newly mated queens, (3) occupy multiple nests, some of which are at least temporarily without a resident queen, (4) live in dense populations, and (5) learn the species odor early in life.