Colony Foundation

AntWiki: The Ants --- Online

Colony Founding and Growth

The Founding Stage

As soon as the queens are inseminated, they shed their membranous wings by raking the middle and hind legs forward and snapping the wings free at the basal dehiscent sutures. Over the coming weeks the alary muscles and fat bodies are metabolized and converted into eggs, as well as food to rear the first batch of larvae. The latter nutrient materials are packaged as either trophic eggs (eggs that cannot develop but are used exclusively as food), specialized salivary secretions, or both. The basic process was first described in the classic study of the formicine ant Lasius niger by Charles Janet (1907). More recently, it has been found that the esophagus of the queen expands into a "thoracic crop" in which the converted tissues are temporarily held in liquid form. In Pharaoh's ant (Monomorium pharaonis), the esophagus diameter widens from 7-10 micrometers to 265 micrometers. The thoracic crop has been demonstrated in five genera of Myrmicinae and Formicinae so far (Petersen-Braun and Buschinger, 1975).

In the case of Solenopsis invicta at least, the conversion process is mediated by the corpora allata. Allatectomized queens fail to cast their wings or undergo wing muscle histolysis, while treatment of these operated individuals with juvenile hormone causes both processes to proceed (Barker, 1979).

The external trigger for wing shedding and histolysis is not exclusively insemination, as might be guessed. When virgin queens of some species are taken from the presence of other queens, they drop their wings after about a day and begin to behave more in the manner of inseminated nest queens. In the case of Solenopsis invicta in particular, the crucial signal is a relatively nonvolatile pheromone produced by queens and conveyed to the virgin alates (Fletcher and Blum, 1981; Fletcher et al., 1983).

The conversion of body tissues into food for the larvae was a vital evolutionary advance in the Formicidae. Wheeler (1933b) suggested that partially claustral colony founding (where the queen still leaves the nest to obtain some of the food) is the primitive state and fully claustral colony founding was derived from it. This inference is based on sound logic: if the ants did originate from predatory vespoid wasps, as the anatomical evidence suggests, they or their immediate ancestors were likely to have passed through a stage in which foundresses still captured insect prey and transported them to preexisting nests. In other words, the earliest ants are likely to have been partially claustral.

It is further true that species of the relatively primitive subfamily Ponerinae display finely graded steps leading from the partial to the fully claustral mode, from dependence on outside foraging to more or less complete freedom from it. The queens of Pachycondyla (= Bothroponera) soror, for example, are in an exactly intermediate stage. They forage outside the nest, but their wing muscles are still reduced and metabolized in the manner of the higher ants (Haskins, 1941). The queens of Odontomachus haematodus are capable of rearing their first brood at least partially with their own oral secretions, but, in one experiment performed by Haskins and Haskins (1950b), the larvae still failed to reach maturity. Finally, these authors found that the unusually bulky queens of Pachycondyla (= Brachyponera) lutea are able to rear the first brood all the way through solely with their own secretions, even though they continue to forage outside the nest when given the opportunity. From Pachycondyla lutea it is but a short step to the condition typifying most myrmicines, formicines, and other "higher" ants, in which complete claustral colony founding is the mode.

At least one substantial advantage of claustral colony founding is obvious and may in fact have played a role in its general adoption among the phylogenetically more advanced ants. Social insect workers suffer their highest mortality during foraging trips (Porter and Jorgensen, 1981; Schmid-Hempel, 1984), and it is probable that the same is true of founding queens forced to leave their nests to search for food.

The Ergonomic (exponential) Stage

The first workers produced by the queen are typically "nanitics" or "minims," that is, miniature forms somewhat smaller than the smallest workers encountered in older colonies of the same species (Figure 3-16). They are characteristically timid in behavior but otherwise perform the same repertory of tasks as do workers in older colonies. In the fire ant Solenopsis invicta at least, they differ from other worker castes in venom composition, specifically in the relative proportions of piperidine alkaloids (Vander Meer, 1986b). In the case of dimorphic species, the first-brood nanitics possess the basic anatomical structure of the minor caste. Major workers usually do not appear until later, and even then are initially smaller in average size. Minims are a general, perhaps universal phenomenon in ants, occurring not only in the "higher" subfamilies and genera, but also in the primitive Australian genus Myrmecia (see review in Wood and Tschinkel, 1981).

Ergonomic models designed to calculate the net energetic yield and hence growth rates of colonies support the intuition that the small size and timidity of the first workers represent prudent features built into the investment strategy of colonies as a whole (Oster and Wilson, 1978). A newly founded colony should strive to maximize the number of workers and their initial survival rate at the expense of everything else. The intuited reasons are as follows. With the queen's internal resources exhausted, there is a minimum number of workers needed to accomplish an adequate performance in each of the vital tasks--a certain number to enlarge the nest, a certain number to nurse the second brood, a certain number to forage, and so forth. There should also exist an optimum number, above this minimum, since adult mortality is probable before the second brood reaches adulthood. The optimum number of nanitics can be defined as that above which the survival probability of the queen can no longer be significantly increased and in fact is likely to be decreased. Because the biomass of adult workers that can be produced by the founding queen is very limited, it is efficient for the queen to divide it into many small workers, as documented by Porter and Tschinkel (1986) in the fire ant Solenopsis invicta. But this advantage is easily reversed, because to raise a great many such individuals would necessitate the production of excessively small nanitics unable to exploit the food items and nest sites for which the species is anatomically and behaviorally adapted. As a result, there should be an optimum number of nanitic workers, determined by the balance between the advantages of a larger initial worker force and the disadvantages of a smaller body size (see Figure 3-17). This result has been confirmed experimentally in Solenopsis invicta by Porter and Tschinkel (1986). They found that nanitics are less efficient on an individual basis than ordinary minor workers in rearing brood, but more efficient as a group than a group of minor workers of equal combined weight--because they are more numerous. On the other hand, they are energetically more expensive to maintain and thus are superior only for the brief period of colony founding.

Moreover, the small size of the incipient colony seems to dictate that its members be relatively timid in behavior. Suppose that an encounter with a single enemy such as a group of foragers from an alien colony results in the loss of five workers. For a mature colony containing thousands of members, this sacrifice is not only tolerable but desirable, if it clears enemy scouts from the territory on which the population depends for food. But for an incipient colony of only ten workers, the loss could be fatal. Furthermore, the potential gain from expelling territorial intruders is expected to be less, because the incipient colony is still living on a fraction of the available food supply yielded by the surrounding terrain.

The nanitics appear to owe their miniature size at least partially to the meager nutrients supplied them by the founding queen in their larval stage. Pheromonal or other programmed stimuli from the queen may also be important. When Wood and Tschinkel (1981) introduced newly inseminated Solenopsis invicta queens into groups of workers with differing numbers, the workers in the first brood increased in average size according to the number of attending workers in the adoptive group. Yet none of these sets of offspring were as small in average size as the nanitics of normal incipient colonies.

When the first brood of workers reaches the adult stage, the new colony undergoes a radical transformation. If the queen has been conducting the ordinary chores of the colony, she now stops in order to devote herself exclusively to egg-laying. The workers take over all of the remaining tasks, including the feeding of the queen herself. For a few worker generations, the average number of which varies among species, no new reproductive forms are reared. Also, with the exception of raiding species such as Myrmecocystus mimicus and Solenopsis invicta, few if any reproductive individuals or alien workers are adopted from the outside. Thus the colony is a semi-closed system devoted to its own exponential growth (Brian, 1957b, 1965b, 1983; Wilson, 1971; Oster and Wilson, 1978). The colony can in fact be viewed in this middle, ergonomic stage as a growth machine: its hypothesized "purpose" is to proliferate workers as quickly and safely as possible. The growth function is implemented chiefly by division of labor--the right number of foragers to harvest the surrounding terrain, the right number of nurses to stoke larval growth, and a sufficient but not excessive number of defenders and auxiliaries standing by for emergencies. The focus of the colony is not yet reproduction or dispersal. New nest sites are sought only when the old ones become environmentally untenable or too small to hold the expanding colony.

During the ergonomic stage competition within the colony is at a minimum (see our analysis of competition in Chapter 6). However, the beginning of the stage, or more precisely the transition to this stage from the preceding, founding stage, is sometimes accompanied by hostile interactions. In Lasius flavus, Messor pergandei, Myrmecocystus mimicus, and Solenopsis invicta, pleometrotic laboratory groups revert to monogyny when the first brood or mature workers appear. The Lasius queens fight with one another and then break apart into single-queen units (Waloff, 1957). Those of Myrmecocystus form dominance hierarchies, with the supernumerary individuals eventually being driven out by the workers (Bartz and Hölldobler, 1982). When multiple Solenopsis invicta queens are introduced to queenless workers, the latter usually execute all but one, both in the laboratory and under natural conditions (Wilson, 1966; Fletcher and Blum, 1983; Tschinkel and Howard, 1983). In the carpenter ants Camponotus herculeanus and Camponotus ligniperda, large colonies often contain several queens, but these individuals are intolerant of one another and maintain territories within the diffuse nests, a condition referred to by Hölldobler (1962) as oligogyny. The same phenomenon occurs in the Australian meat ant Iridomyrmex purpureus: queens that cooperated amicably during nest founding become antagonistic after the first workers appear, and in the end permanently separate within the nest (Hölldobler and Carlin, 1985).

If the colony survives the precarious period during which the first and second worker broods are being reared, it is likely to enjoy an interval of sustained exponential growth. But this growth, like that in all populations, can be expected to slow with time and eventually to come to a halt. Most data on the course of colony growth in social insects generally suggest curves that are roughly sigmoidal (hence "logistic") in form (Brian, 1965b, 1983). In a recent, thorough study of the fire ant Solenopsis invicta, Tschinkel (1988a) was able to show that colonies under natural conditions grow logistically, attaining the maximum worker population of about 220,000 in 4 to 6 years. This is the expected result, but the underlying density-dependent controls are more complex than those determining typical logistic growth in nonsocial insects. The theory of colony growth, based on the concepts of economies of scale and evolutionary optimization, has been developed in some detail by Oster and Wilson (1978). We will return to the subject repeatedly in later chapters as part of our analysis of caste, division of labor, foraging strategies, and defense.

The Reproductive Stage

If a monogynous colony were to maintain its worker population at zero population growth, it would be unable to reproduce, since total investment means by definition that no production of virgin queens and males is possible. Consequently, at some point short of its maximum possible size, the colony should devote part of its production to the creation of virgin queens and males. The timing of the conversion varies among species according to the special adaptations the species have otherwise made to their environments (Wilson, 1971; Oster and Wilson, 1978). Within species, the production of reproductives increases as a function of colony size. Among species of the genus Myrmica at least, the average worker stature (related to colony size) also has a particularly strong influence on queen production but not male production (Elmes and Wardlaw, 1982). It is a common event, as documented in Myrmica rubra by Brian (1957a,b), for males to appear in the nest prior to the females--and sometimes at erratic intervals, possibly as a consequence of workers laying eggs in competition with the queen. Brian has termed the interval of early male production the "adolescent" period of colony growth, coming between the "juvenile" (= ergonomic) period during which the worker population expands and the "mature" period during which new, virgin queens are produced. In truth, very little information has been published on the timing of these key events in the colony life cycle, so it is impossible to generalize about the sequence in which males and queens appear. In many, perhaps even most ant species, some colonies rear only males in a given season, others rear only queens, while still others rear a mixture (Nonacs, 1986a,b). Whether individual colonies change their strategy from one year to the next is not known. Table 3-2 gives most of the available data on colony size in various species of ants. A few additional data have been compiled by Baroni Urbani. These raw numbers tell us nothing directly about the growth rates or factors limiting mature colony size. Also, the numbers often represent underestimates of the typical mature colony size, because colonies of all ages were censused and colonies in many wild populations tend to be young. Nevertheless, the data do permit some inferences when comparisons are made between major groups:

1. No apparent correlation is yet evident between the size of the mature colony and the longevity of the colony, at least as measured in monogynous species by the life span of the queen (see also Table 3-3). However, the data are so few, especially of colony life spans, that it would be premature to draw any firm conclusion. There is a clear need for more longevity studies of ants of all castes because of the relevance of such data to population dynamics.

2. There is no clear relation between climate and colony size. If anything, temperate species tend to have somewhat larger colonies on the average. This is due to a special ecological effect connected with constraints in the nest site of many tropical species, as follows.

3. There is a strong relation between preferred nest site and mature colony size. Among the ants of New Guinea rain forests, for example, species that nest in rotting logs and other pieces of decaying wood on the ground (almost all Ponerinae and the majority of Myrmicinae) form smaller colonies than those living in less restricted nest sites, such as the open soil of the forest floor (Acidomyrmex, Pheidologeton, Leptomyrmex, Pseudolasius, Acropyga, most Paratrechina), open air at the ground surface (Aenictus), and various parts of the tree canopy (most Crematogaster, Iridomyrmex, Camponotus, Polyrhachis, and Oecophylla).

4. The most elaborate caste and communication systems occur in species with large, perennial colonies, for example the legionary ants (Dorylus, Eciton), leafcutter ants (Acromyrmex, Atta), and the marauding ants of the genus Pheidologeton.

5. The great variation in colony size among species belonging to the same taxonomic group (for example, the Myrmicinae) attests to the capacity of this population trait to evolve with relative speed. Small alterations in the physiological parameters of individual ants such as mean worker life span and thresholds in queen determination of larvae can bring about major differences in mature colony size. Given this potency to adapt colony size to local environmental conditions at the species level, we should feel encouraged about the possibility of inferring which conditions have been critical in evolution.

Two Contrasting Strategies of Colony Foundation

The ability to disperse and establish new colonies is crucial to the ecological success of ants. However, founding queens are often alone, which means that they face the same dangers as solitary insects. Colony-founding behaviours have diversified tremendously across the ants, and underlie their capacity to colonize almost all terrestrial habitats. Unlike social wasps and bees, many ant species evolved (1) queens that can store large amounts of metabolic reserves, (2) workers that are much smaller than their queens. Both these adaptations make it possible to raise the first workers without outside food (claustral foundation), and this is arguably central to the tremendous success of ants belonging in subfamilies Dolichoderinae, Formicinae and Myrmicinae.

During independent colony founding (ICF), lone queens need to raise the first generation of offspring without the help of nestmates. Following aerial dispersal and mating with a foreign male, they locate a suitable nesting site and break off their wings. Foundresses excavate or take over an existing shelter, and lay a first batch of eggs. During the next few weeks (sometimes months), they must guard and feed their offspring. ICF is a risky strategy, especially when queens need to forage outside the nest (non-claustral foundation). ICF necessitates the annual production of a large number of sexuals, and this ability is affected by colony size. In many phylogenetically independent lineages, ICF has been replaced by a strikingly different founding mode, whereby queens are not alone. In such dependent colony founding (DCF) species, existing colonies divide into two or more daughter groups which soon become autonomous. Thus the queens disperse on foot to a new nesting site together with nestmate workers, and the latter feed and protect the brood.

A unique feature of ants is the evolution of permanently wingless queens in many species exhibiting DCF. Many social wasps and bees start new colonies by DCF, but there is universal retention of flying workers and queens. This flight constraint is absent in ants: wingless workers are the major players during the division of existing colonies, and since dispersal occurs exclusively on foot, the production of winged queens is selected against. Accordingly, ergatoid (no wings) queens (see list of species) and brachypterous (short-winged) queens (list of species) evolved convergently in species belonging to over 50 genera (Peeters 2012). Nonetheless, non-flying queens are independent founders (ICF) in a few species, e.g. Plectroctena mandibularis, Myrmecia regularis and Pogonomyrmex laticeps.

Mutualism with Fungi or Sap-Feeding Insects

In a restricted number of genera, claustral ICF is possible due to trophic interactions with a mutualist, hence foundresses lack large metabolic reserves. During the mating flight of Acropyga and Tetraponera binghami, foundresses carry a gravid pseudococcid (scale insect) to the stem cavity where they settle - being clonal, the sap-suckers multiply and supply sufficient honeydew to feed the first worker brood. Similarly in Azteca, dispersing queens carry an ascomycete fungus in their infrabuccal pocket; after settling in a domatium, this fungus is used to process parenchyma tissue scraped away from the plant. This is then used to feed the first larvae, meaning that the foundress does not need to forage outside. Acquisition of appropriate scale insects is an important step in the successful establishment of a new colony on a host plant. In contrast to the above examples of vertical transmission of the mutualistic partner, foundresses in other species need to find scale insects or fungi near the newly established nest (horizontal transmission). Cladomyrma lives together with mealybugs (also a scale insect), and first instars ('crawlers') disperse to settle into incipient nests of the ants. Melissotarsus and Rhopalomastix live in a mutualism with diaspidid scale insects inside living trees, and it is likely that crawlers locate tunnels newly chewed by founding queens.

Parasitism is also a Strategy of Independent Founding

In several lineages, newly mated queens attempt to enter existing colonies of their own or different species, following which they exploit local resources to raise their offspring. As discussed by Peeters & Molet (2010), social parasitism is a form of ICF because queens are not helped by nestmate workers. They disperse alone and take considerable risks when trying to enter the host colony. Host colonies are nothing more than a resource of the environment to be exploited, similar to insect prey. Many authors have considered parasitism to be DCF because queens ‘depend’ on their host colonies, however claustral queens also 'depend' on their metabolic reserves, and non-claustral queens 'depend' on the food they gather outside. What is crucial in DCF species is that queens depend on closely related nestmate workers, i.e. there is convergence of genetic interests.

The founding strategies of a majority of ant species remain unstudied. Indeed, there are no published data about colony foundation for many large genera. This scarcity of information stems from the necessity of studying colony foundation in the field. DCF is especially difficult because it can occur unpredictably over the year, and it needs to be distinguished from simple nest emigrations (Cronin et al. 2013). In species where the colony founding strategy is unknown, data on queen morphology (alive or dead) can generate testable hypotheses. The thorax architecture of queens and workers reflects the size and geometry of head muscles (Keller et al. 2014). Among species with flying queens, some have strong neck muscles (similar to those of workers) because they need to forage outside the incipient nests (non-claustral ICF), and their prothorax is larger than in claustral species. In non-flying queens, the absence of wing muscles is associated with a simplified thorax (i.e. fusion of sclerites), and such queens generally perform DCF.

Survey of the various founding strategies documented across 49 ant genera. Compiled by Christian Peeters

Mating Strategies are distinct from Founding Strategies

Mating usually precedes colony founding, but these are two distinct events affected by very different selective pressures (Peeters & Molet 2010). ICF queens in some species mate close to their natal colony (“female-calling”) before dispersing, while in other species they mate far away (“male-aggregation”). Female-calling is associated with lower mortality since flight means increased risks of predation or getting lost. DCF queens mate near their natal colony, or even inside it Ophthalmopone berthoudi. Female-calling is typical of poneroid ants but it is also found in various formicoids, e.g. desert-dwelling Cataglyphis. The evolutionary loss of flying queens is characteristic of female-calling species. In some DCF species where queens retain flying ability, these also show female-calling since they will re-enter the natal nest once mated (this is incompatible with mating in far away male-aggregations).