The Ants Chapter 17

CHAPTER 17. THE FUNGUS GROWERS

Members of the myrmicine tribe Attini share with macrotermitine termites and certain wood-boring beetles the sophisticated habit of culturing and eating fungi. The Attini are a morphologically distinctive group limited to the New World, and most of the 12 genera and 190 species occur in the tropical portions of Mexico and Central and South America. Besides their unique behavior and the many peculiar behavioral and physiological changes associated with it, the Attini are distinguished from other ants by a distinctive combination of anatomical traits, including the shape of the antennal segments; a less-than-absolute tendency toward hard, spinose, or tuberculate bodies; and a proportionately large, casement-like first gastral segment.

It is conceivable that fungus growing originated only once in a single ancestral attine living in South America during that continent's long period of geological isolation from late Mesozoic times to approximately four million years ago. Exactly when the event occurred is open to conjecture, but it was almost certainly prior to the Miocene Epoch. Extinct but modern-looking species of Trachymyrmex (Baroni Urbani, 1980) and Cyphomyrmex (Wilson, 1985h) have been found in the Dominican amber, which is believed to be either late Oligocene or early Miocene in age.

In Africa, southern Asia, and other parts of the Old World tropics, the Attini are replaced by fungus-growing termites (Macrotermitinae), which in their turn do not occur in the New World. No one can be sure whether this complementary global pattern is due to a mutual preemption involving competitive exclusion of one group by another or whether it is simply one more accidental outcome reflecting the extreme rarity of the evolutionary origin of fungus gardening. The latter is more likely to be the case, meaning that if attines were to be introduced today into the range of the macrotermitines, or vice versa, the two kinds of insects could coexist with little interference. This is possible because attines utilize insect excrement and fresh plant material for the most part, while the macrotermitines use dead plant material. Also, fungus-growing ants forage above ground, often even in trees, while fungus-growing termites are primarily subterranean.

The Attini are an enormously successful group where they exist. One species, Trachymyrmex septentrionalis, ranges north to the Pine Barrens of New Jersey, while in the opposite direction several species of Acromyrmex penetrate to the cold temperate deserts of central Argentina. In the vast subtropical and tropical zones in between, attines are among the dominant ants. Many of the species gather pieces of fresh leaves and flowers to nourish the fungus gardens, and Atta and Acromyrmex rely on this source exclusively. Since they attack most kinds of vegetation, including crop plants, they are serious economic pests. The species of Atta in particular are among the scourges of tropical agriculture. They are familiar to local inhabitants as the wiwi in Nicaragua and British Honduras, the bibijagua in Cuba, the hormiga arriera in Mexico, the bachac in Trinidad, the bachaco in Venezuela, the saúva in Brazil, the cushi in Guyana, the coqui in Peru, and the leaf-cutting or parasol ant in most English-speaking countries, the last name alluding to the fact that an Atta worker holding a leaf fragment over its head gives the impression of carrying a parasol. The problems of agriculture in Atta country have been humorously epitomized in the following anecdote by V. Wolfgang von Hagen, in connection with his attempt to grow a vegetable garden in Belize:

My Indian servants, dusky, kinky-haired Miskito men, lamented all this work. It was useless, quoth a toothless elder, to plant anything but bananas or manioc, as the Wiwis were sure to cut off all the leaves. Without the slightest encouragement the Miskito Indians would launch forth on the tales of the ravages of the Wiwi Laca, but unswayed by the illustrations, like Pangloss I could only remark that all this was very well but let us cultivate our garden. In two weeks the carrots, the cabbages, the turnips were doing well. The carrots had unfurled their fernlike tops, the cabbage grew as if by magic. From our small palm-thatched house my wife and I cast admiring eyes over our jungle garden. Our minds called forth dishes of steaming vegetables to replace dehydrated greens and the inevitable beans and yucca. Even the toothless Miskito elder came by and admitted that white man's energy had overcome the lethargy of the Indian. Then the catastrophe fell upon us. We arose one morning and found our garden defoliated: every cabbage leaf was stripped, the naked stem was the only thing above the ground. Of the carrots nothing was seen. In the center of the garden, rising a foot in height, was a conical peak of earth, and about it were dry bits of earth, freshly excavated. Into a hole in the mound, ants, moving in quickened step, were carrying bits of our cabbage, tops of the carrots, the beans--in fact our entire garden was going down that hole. I could see the grinning face of the toothless Miskito Indian. The Wiwis had come.

In fact, the leaf-cutting ants of the genera Atta and Acromyrmex are the leading agricultural pests of Central and South America. They were preadapted for this role by their ability to use many plant species with the aid of their symbiotic fungi, which serve as a sort of ancillary digestive system. They also build up high population densities, such as 5 colonies per hectare in Atta vollenweideri and 28 per hectare in Atta capiguara, with each colony containing a million or more workers (Fowler et al., 1986a,b).

Leafcutters are the dominant herbivores of the Neotropics, consuming far more vegetation than any other group of animals of comparable taxonomic diversity, including mammals, homopterans, and lepidopterans. The amount of vegetation cut from tropical forests by Atta alone has been calculated on the basis of twelve studies to lie between 12 and 17 percent of leaf production (Cherrett, 1986). Grass-cutting species of Atta, which are distinguished from other members of the genus by their short, massive mandibles, are equally voracious. Each colony of Atta capiguara uses about 30 to 150 kilograms of dry matter each year; the figure of Atta vollenweideri is 90 to 250 kilograms per year. Atta capiguara reduces the commercial carrying capacity of pastureland, measured by the number of head of sustainable cattle, by as much as 10 percent (Fowler et al., 1986a).

Because of the catholicity of their diets, or rather the diets of their fungus, leafcutters have an extraordinarily diverse impact on agriculture. It includes the direct distribution of most kinds of crops, loss of land surface to the large nests (from 30 to 600 square meters per nest, when soil erosion is included), accidents caused to animals and agricultural machinery, and highway and other right-of-way damage from the excavation of the huge nests. Because of the variation in damage from one country to the next, the total loss due to the ants is impossible to calculate, but it is probably in the billions of dollars. Yet research on leafcutters remains relatively neglected. According to Cherrett (1986), by the early 1980s only 1250 articles had been written on Atta and Acromyrmex, as opposed to 10,000 on locusts.

Leafcutting ants have been important to the economy of Latin America throughout historical times. The early Portuguese colonists, who dubbed Brazil the kingdom of the ants, left behind such testaments to the saúva as the following: “If there is not much wine in this land it is because of ants which strip the leaves and fruit” (1587), “In a word, it is the worst scourge that farmers have” (1788), and “Either Brazil kills the saúva or the saúva will kill Brazil” (1822) (cited by Mariconi, 1970). Deep within their huge nests, able to multiply themselves many times each year, the leafcutters are nearly invulnerable to anything but massive poisoning.

Because so many species thrive in cleared land and secondary forests, leafcutters as a whole have benefitted by the advent of European civilization. The ubiquitous Atta cephalotes, for example, is specialized to live in forest gaps, and as a consequence it is able to invade subsistence farms and plantations from Mexico to Brazil (Cherrett and Peregrine, 1976). Prior to 1954 Atta capiguara around São Paulo was limited to a small savanna south of the city and had little or no economic impact in the area. When nearby forests were cleared for conversion to coffee plantations and then pastureland, the species spread rapidly and reached pest proportions. After Acromyrmex octospinosus was accidentally introduced into the West Indian island of Guadaloupe, shortly before 1954, it spread rapidly to become an important agricultural pest (Therrien et al., 1986). If any leafcutter ants, and especially Atta, were to be established in sub-Saharan Africa or some other part of the Old World tropics, the result might be an ecological catastrophe. The terrestrial ecosystems of these continents are unprepared for a herbivore with the resiliency and proficiency of these highly organized insects.

Yet in spite of the problems leafcutters cause, it would be a mistake to think of them as the uncompromising enemy of humankind. During millions of years of coevolution with their natural environment, they have become an integral part of the ecosystems of the New World tropics and warm temperate zones. They supplant to a large extent the populations of herbivorous mammals, which are relatively sparse through most of the New World tropics. They prune the vegetation, stimulate new plant growth, break down vegetable material rapidly, and turn and enrich the soil. In the tropical moist forests, Atta are major deep excavators of soil and stimulators of root growth (Haines, 1978). If leafcutters were to be extirpated, a profound readjustment of the structure of forests and grasslands would result, including the extinction of at least a few species of plants and animals. Such considerations have led Harold G. Fowler and his co-workers in Brazil (personal communication) to call for the protection of Atta robusta, a local forest-dwelling species in São Paulo State now endangered by rapid deforestation within its range.

Leafcutting ants are among the most advanced of all the social insects. They arose in the New World, most likely in South America when that continent was isolated from the remainder of the Western Hemisphere. During the past ten thousand years, a mere eyeblink in geological time, these insects have encountered the most advanced product of mammalian evolution from the Old World, Homo sapiens. Certain difficulties have arisen from this contact, with the great bulk of the losses occurring on the human side. In order to redress the balance, we need to learn a great deal more about the biology of our adversaries, paying particular attention to the weak points that undoubtedly occur in their complicated social systems. But the goal should be intelligent management of their populations and never their complete eradication. Our advantage--and responsibility--lies in the fact that we can think about these matters and they cannot.

Fungus culturing
What happens to the vegetation after the Atta workers have carried it down their holes is a fascinating story that has been worked out through many decades of research. Bates, in his book The Naturalist on the River Amazon (1863), suggested that the ants use the leaves “to thatch the domes which cover the entrances to their subterranean dwellings, thereby protecting from the deluging rains the young broods in the nests beneath.” Other early observers believed that the leaves are eaten or used to maintain a constant nest temperature by heat of fermentation. Thomas Belt was the first to surmise the far stranger truth. In The Naturalist in Nicaragua (1874) he described the garden chambers deep within the Atta nests as being “always about three parts filled with a speckled brown, flocculent, spongy-looking mass of a light and loosely connected substance. Throughout these masses were numerous ants belonging to the smallest division of the workers, and which do not engage in leaf-carrying. Along with them were pupae and larvae, not gathered together, but dispersed, apparently irregularly, throughout the flocculent mass. This mass, which I have called the ant-food, proved, on examination, to be composed of minutely subdivided pieces of leaves, withered to a brown colour, and overgrown and lightly connected together by a minute white fungus that ramified in every direction throughout it. . . That they do not eat the leaves themselves I convinced myself; for I found near the tenanted chambers deserted ones filled with the refuse particles of leaves that had been exhausted as manure for the fungus, and were now left, and served as food for larvae of Staphylinidae and other beetles.”

It was left to Alfred Möller (1893) to observe for the first time the actual eating of the fungi. He found that the tips of the hyphae produce peculiar spherical or ellipsoidal swellings (Figure 17-1) which are plucked and eaten. Möller called these objects “heads of Kohlrabi” because of their fancied resemblance to the vegetable. Later Wheeler relabeled them gongylidia (singular: gongylidium), and this name has stuck. A group of gongylidia, to complete the terminology, is sometimes referred to as a staphyla (plural: staphylae), while a piece of the peculiar morel-like fungus of Cyphomyrmex rimosus is called a bromatium. The gongylidial clusters of Atta and Acromyrmex, averaging about half a millimeter in diameter, were later observed to be eaten both by adult workers and larvae. The structures are rich in glycogen, in a form readily assimilated by the ants (Quinlan and Cherrett, 1979; Febvay and Kermarrec, 1983). Kermarrec et al. (1986) have described the gongylidium of the Acromyrmex octospinosus fungus as “a goat-skin bottle which has a thick wall covered with mucilage and is filled with a finely granulated mictoplasm that maintains its turgidity.” It is a tank “filled with glygogen, hydrolases, and viral particles.”  About 56 percent of the dry weight of the mycelium as a whole, or interconnected mass of hyphae, of the Atta colombica fungus is available in the form of soluble nutrients, which include 27% carbohydrates, 4.7% free amino acids, 13% protein-bound amino acids, and 0.2% ergosterol and other lipids. The carbohydrates include trehalose, mannitol, arabinitol, and glucose but no detectable polysaccharides (Martin et al., 1969a). Why the Acromyrmex fungus has abundant glycogen while the Atta fungus lacks it, if this reported difference actually exists, is not known.

Yet in spite of the problems leafcutters cause, it would be a mistake to think of them as the uncompromising enemy of humankind. During millions of years of coevolution with their natural environment, they have become an integral part of the ecosystems of the New World tropics and warm temperate zones. They supplant to a large extent the populations of herbivorous mammals, which are relatively sparse through most of the New World tropics. They prune the vegetation, stimulate new plant growth, break down vegetable material rapidly, and turn and enrich the soil. In the tropical moist forests, Atta are major deep excavators of soil and stimulators of root growth (Haines, 1978). If leafcutters were to be extirpated, a profound readjustment of the structure of forests and grasslands would result, including the extinction of at least a few species of plants and animals. Such considerations have led Harold G. Fowler and his co-workers in Brazil (personal communication) to call for the protection of Atta robusta, a local forest-dwelling species in São Paulo State now endangered by rapid deforestation within its range.

Leafcutting ants are among the most advanced of all the social insects. They arose in the New World, most likely in South America when that continent was isolated from the remainder of the Western Hemisphere. During the past ten thousand years, a mere eyeblink in geological time, these insects have encountered the most advanced product of mammalian evolution from the Old World, Homo sapiens. Certain difficulties have arisen from this contact, with the great bulk of the losses occurring on the human side. In order to redress the balance, we need to learn a great deal more about the biology of our adversaries, paying particular attention to the weak points that undoubtedly occur in their complicated social systems. But the goal should be intelligent management of their populations and never their complete eradication. Our advantage--and responsibility--lies in the fact that we can think about these matters and they cannot.

As fresh leaves and other plant cuttings are brought into the nest, they are subjected to a process of degradation before being inserted into the garden substratum. First the ants lick and cut them into pieces 1-2 mm in diameter. Then they chew the fragments along the edges until the pieces become wet and pulpy, sometimes adding a droplet of clear anal liquid to the surface. Then, using side to side movements of the fore tarsi, they carefully insert the fragments into the substratum. Finally, the ants pluck tufts of mycelia from other parts of the garden and plant them on newly formed portions of the substratum. A newly inserted single leaf section 1 mm in diameter receives up to ten such tufts in five minutes. The transplanted mycelia grow rapidly, as much as 13 µ in length per hour. Within 24 hours they cover most of the substratal surface.

Michael Martin and his co-workers discovered that Atta workers contribute digestive enzymes in the fecal droplets they deposit on the fungus, including a chitinase, an a-amylase, and three proteinases (Martin, 1970; Martin et al., 1973). Subsequently, Boyd and Martin (1975) showed that the proteinases originate in the fungus and pass unaltered through the digestive tract of the ants back to the fungus. The ants avoid digesting fungal enzymes by the simple expediency of not secreting any digestive enzymes on their own. Acromyrmex octospinosus also lacks proteinases, but these smaller leafcutters produce their own chitinases in the labial glands (Febvay and Kermarrec, 1986). The metabolic capabilities of the attine ants and their symbiotic fungi have yet to be worked out in detail, but it is at least clear that the ants have lost some key enzymes. They depend heavily on their symbionts for many of their nutrients, while the fungi in turn depend on the ants for care and the recycling of some of the enzymes.

Still another chapter in leafcutter biology began with the revelation, by Barrer and Cherrett, that Atta and Acromyrmex workers feed directly on plant sap. As much as a third of radioactivity in experimentally labeled leaves is absorbed directly by the ants as a result (Barrer and Cherrett, 1972; Littledyke and Cherrett, 1976). It might seem possible at first that the ants merely contribute the liquid to the fungus as added nutrient rather than assimilate it themselves. However, the sap must be crucial to the workers because, as Quinlan and Cherrett (1979) found, only 5 percent of their energy requirements are met by ingestion of juice of the fungal staphylae. In contrast, the larvae are able to subsist and grow entirely on the staphylae. These findings suggest that adult workers use only the juice of the staphylae, while the larvae use every part of the staphylae. The queen, to complete the story, is known to obtain at least a substantial part of her food from trophic eggs laid by workers and fed to her at frequent intervals.

In summary, the main properties of the leafcutter-fungus symbiosis can be stated as follows. Adult ants are fundamentally nectar feeders, predators, and scavengers. Their entire digestive system, from their peculiar infrabuccal and proventricular filters to the delicate midgut and limited spectrum of digestive enzymes, is geared to this dietary commitment. They are ill-suited to be herbivores. The fungus, in exchange for protection and cultivation, digests the cellulose and other plant products normally inaccessible to leafcutters and shares part of the assimilable metabolic products with them. In the case of the “lower” attines, which do not cut leaves but use insect remains and excrement, the fungus converts the chitin and other products otherwise less available to ants.

Curiously, the single outstanding problem of attine biology all along, the identity and biological qualities of the symbiotic fungus, remains wrapped in mystery. The principal difficulty has been the reluctance--indeed, the near inability--of the fungus to form sporophores, the elaborate fruiting structures required for taxonomic diagnosis. Evidently the ants do not permit the fungi to form the mushrooms or other spore-bearing bodies under natural conditions. Instead the ants feed exclusively on the special gongylidial tips of the elementary mycelial clusters, a preference that appears to have resulted in the loss of the ability on the part of the fungus to produce sporophores. Reciprocally, the fungi utilize the ants for transport and do not have to depend on windborne spores to transfer themselves from nest to nest. Although Moeller did not clarify this problem in Atta, he was lucky enough to discover sporophores growing from abandoned Acromyrmex nests on four separate occasions. These proved to be agaracine mushrooms, wine-red in color, which Moeller formally named Rozites gongylophora. Mycologists have since confirmed their placement in the basidiomycete family Agaricaceae, but transferred the species gongylophora to the genus Leucocoprinus (Heim, 1957; Kermarrec et al., 1986). Subsequent attempts by entomologists to locate sporophores in abandoned attine nests and to culture them in the laboratory from the gardens of various attine genera have rarely succeeded. The most notable single advance was Weber's (1957b) use of a medium of sterile oats to rear sporophores of an apparent Leucocoprinus (= Lepiota) from mycelia originating from a Cyphomyrmex costatus garden. If future mycologists ever succeed in isolating a plant hormone or nutrient combination that enhances sporophore formation in fungi, dramatic further progress can be expected in this field.

The current evidence overall seems to support Roger Heim's opinion that the symbiotic fungus cultivated by all the attine ants is Leucocoprinus gongylophorus. The identification of this species or at least a set of closely similar forms placed variously in the basidiomycete genera Leucocoprinus, Lepiota, and Rozites has been confirmed by the rearing of sporophores from the garden mycelia in the attine genera Atta, Cyphomyrmex, and Myrmicocrypta. These ants represent almost the entire phylogenetic spread of the tribe Attini. According to Heim, it is unlikely that different attines picked up various leucocoprines here and there in the course of their evolution. In the absence of opposing strong evidence, this parsimonious hypothesis seems preferable to that of Weber (1979), who preferred to place the fungi of the lower attines (such as Cyphomyrmex and Myrmicocrypta) in a separate genus, Lepiota. Lehmann (1975, 1976) offered a third, truly radical opinion, that the attine fungus is not a basidiomycete at all but an ascomycete in the genus Aspergillus close to the symbiont of the fungus-growing beetles and Old World fungus-growing termites. This conclusion is probably too parsimonious. It is based on several tenuous morphological comparisons, including the supposedly primitive ascomycete appearance of the gongylidial swellings. It also flies in the face of contrary evidence based on sporophores cultured from attine mycelia. Weber (1979) has in fact identified an unusual Aspergillus in abnormal gardens of Atta and Acromyrmex, but it was strongly avoided by the ants and appeared to be a contaminant of unusually wet gardens. The only clear exception to the strict conformity of attines to Leucocoprinus or a closely related group of leucocoprine genera is the cultivation of a yeast by Cyphomyrmex rimosus (Wheeler, 1907b; Weber, 1979).

This last example may prove to be the tip of an iceberg in the study of the secondary microflora of the attine gardens. While it is true that the ant cultures are dominated by a single fungus species, microorganisms also exist and might even participate in the symbiosis. Recent research to this end, reviewed by Kermarrec et al. (1986), has revealed that both yeasts and bacteria are in fact present in substantial numbers. The metabolic activity of these microorganisms remains uncertain, however. It is possible that bacteria assist in the lysis of cellulose into products that are more readily utilized by the fungi. At least six species of Bacillus have been identified inside nests of Atta laevigata, and the ants appear to inoculate fresh vegetation with these microorganisms during preparation of the substrate. On the other hand, it is equally possible that the bacteria parasitize the ant-fungus symbiosis, draining away some of the energy that would otherwise flow directly through the fungus to the ants. The parasitism hypothesis gains credence from a finding by Kermarrec and his co-workers that the symbiotic fungus of Atta and Acromyrmex secrete substances antagonistic to bacteria and other kinds of fungi.

Upon reflection it is impressive how nearly pure the ants keep the fungal growth in their nest chambers. They build this monoculture by a variety of techniques: the plucking out of alien fungi, the frequent inoculation of the Leucocoprinus mycelia onto fresh substrate, the manuring of the substrate with enzymes and nutrients to which the Leucocoprinus are especially adapted, the production of antibiotics to depress competing fungi and microorganisms, and the production of growth hormones. These last two methods entail an instinctive form of chemical engineering on the part of the ants. Maschwitz et al. (1970) and Schildknecht and Koob (1970) identified phenylacetic acid, D-3-hydroxydecanoic acid (“myrmicacin”), and indoleacetic acid in the secretions of the metapleural glands of Atta sexdens workers. They suggested that these compounds play different roles in the purification of the symbiotic fungus culture: phenylacetic acid suppresses bacterial growth, D-3-hydroxydecanoic acid inhibits the germination of spores of alien fungi, and indoleacetic acid, a plant hormone, stimulates mycelial growth. As pointed out by Weber (1982), this interpretation can be confirmed only by a demonstration that the components of the metapleural gland are actually present in the fungus gardens at bacteriostatic and fungistatic levels (Weber, 1982).

The life cycle of leafcutter ants
Leafcutting ants comprise 24 known species of Acromyrmex (Table 17-1) and 15 of Atta (Table 17-2). Because the Atta workers are so large and spectacular in their behavior, many entomologists have set out to study their life cycle and biology. They include Moeller, the pioneer in the subject, Forel, Goeldi, Huber, von Ihering, and Wheeler, all of whose publications are exhaustively reviewed in the classic 1907 study of the North American Attini by Wheeler. More recent researchers have included Autuori, Bitancourt, Bonetto, Borgmeier, Eidmann, Fowler, Geijskes, Gonçalves, Jacoby, Kerr, Moser, Stahel, Weber, and others; their work is carefully reviewed in Weber (1972, 1982) and the symposium volume Fire ants and leaf-cutting ants edited by Lofgren and Vander Meer (1986).

All of the Atta species appear to have basically the same colony life cycle. The nuptial flights of some species, such as the infamous sexdens of South America, take place in the afternoon, while texana of the southern United States and a few others hold their flights at night (Autuori, 1956; Moser, 1967a). Because the ponderous females work their way high into the air before the males approach them, actual matings have not been observed. Nevertheless, Kerr (1962), by counting sperm from the spermathecae of four newly mated sexdens queens with the aid of a hemocytometer, was able to show that each individual is inseminated by at least three to eight males. The actual estimated numbers of sperm varied among the queens he examined from 206 to 320 million, seemingly more than enough to last an individual the ten or more years speculated to be the normal life span of an Atta queen.

During the nuptial flight and immediately afterward, as the queens attempt to start new colonies, mortality is extremely high. Out of 13,300 Atta capiguara founding colonies in Brazil, only 12 were alive three months later (Fowler et al., 1986b). From a start of 3,558 incipient Atta sexdens rubropilosa colonies, only 90 or 2.5 percent were alive after three months (Autuori, 1950). The survivorship of rubropilosa during the same time interval was 6.6 percent (Jacoby, 1944), while figures of 10 percent were obtained for Atta cephalotes and zero percent for Atta capiguara in Central America and Brazil respectively (Fowler et al., 1986b).

In 1898 von Ihering made the important discovery of how the fungus is transferred from nest to nest. Prior to departing on the nuptial flight the Atta sexdens queen packs a small wad of mycelia into her infrabuccal chamber,a cavity located (in all ants including Atta) beneath the opening of the esophagus just to the rear of the base of the labium. Following the nuptial flight, which in Brazil may occur anytime from the end of October to the middle of December, the queen casts off her wings and quickly excavates a little nest in the soil. When finished, the nest consists of a narrow entrance gallery, 12-15 mm in diameter, which descends 20-30 cm to a single room 6 cm long and somewhat less in height. Onto the floor of this room, according to Jakob Huber (1905) and Autuori (1956), the queen now spits out the mycelial wad. By the third day fresh mycelia have begun to grow rapidly in all directions, and the queen has laid the first 3 to 6 eggs. In the beginning the eggs and little fungus garden are kept apart, but by the end of the second week, when more than 20 eggs are present and the fungal mass is ten times its original size, the two are brought together. At the end of the first month the brood, now consisting of eggs, larvae, and possibly pupae as well, is embedded in the center of a mat of proliferating fungi. The first adult workers emerge sometimes between 40 and 60 days. During all this time the queen cultivates the little fungus garden herself. At intervals of an hour or so she tears out a small fragment of the garden, bends her abdomen forward between her legs, touches the fragment to the tip of the abdomen, and deposits a clear yellowish or brownish droplet of fecal liquid onto it (Figure 17-2). Then she carefully places the mycelial fragment back into the garden. Although the Atta sexdens queen does not sacrifice her own eggs as a culture medium, she does consume 90 percent of the eggs herself, and, when the larvae first hatch, they are fed with eggs thrust directly into their mouths. The queen apparently never consumes any of the growing fungus during the rearing of the first brood. Instead, she subsists entirely on her own catabolizing fat body and wing muscles. Soon after the first workers appear, they begin to feed themselves on the gongylidia. They also manure the fungal garden with their fecal emissions and feed their sister larvae with eggs laid by the mother queen. The eggs given to the larvae are larger than those permitted to hatch; a histological study by Bazire-Benazet (1957) has shown that they are in fact “omelets” formed in the oviducts by the fusion of two or more distinct but ill-formed eggs. After about a week the new workers dig their way up through the clogged entrance canal and start foraging on the ground in the immediate vicinity of the nest. Bits of leaves are brought in, chewed into pulp, and kneaded into the fungus garden. About this time the queen ceases attending both brood and garden. She turns into a virtual egg-laying machine, in which state she remains for the rest of her life. Now for the first time the workers begin to collect gongylidia from the fungal mass and to feed them directly to the larvae.