This is a very common species throughout the lowlands of Costa Rica, and in mid-elevation areas where it is open. The workers are large, conspicuous ants that are often seen foraging on low vegetation. They are generalized arboreal predators, preying on diverse arthropods. They are opportunistic cavity nesters, nesting in dead wood, knots, dead cavities in live branches, bromeliad bases, and abandoned or peripheral cavities in myrmecophytic Cecropia. Nests may be in the canopy or near ground level. Workers will aggressively defend their nest site when disturbed.
Colony founding may be haplometrotic (by a single queen) or pleometrotic (by multiple queens). In a study on villosa from the state of Bahia, Brazil, Trunzer et al. (1998) reported
"40 percent of all collected founding colonies were pleometrotic associations with two (24 percent) or three (16 percent) queens. The egg-laying rate of solitary founding queens did not differ from that of queens in the two-queen groups, whereas individual queens in the three-queen groups laid significantly fewer eggs. Twenty-one weeks after collection pleometrotic associations contained significantly more workers than nests founded by haplometrosis. During the founding phase only one queen per founding association left the nest to forage. Within pleometrotic founding groups the number of eggs laid by individual queens did not differ, and no aggressive behavior occurred, even in the presence of workers. Pachycondyla villosa seems to be one of the rare cases, where pleometrosis can lead to primary polygyny at least in the laboratory, with the queens coexisting in mature colonies without antagonistic behavior."
Once colonies are established, workers exhibit polyethism, falling into three general behavioral classes: nurses, inactives, and sentinal-foragers (Perez et al. 1985). Queens participate in brood care (Perez et al. 1985). Heinze et al. (1996) studied dominance interactions and reproduction among workers, and reported
"Workers in queenless groups of the ant Pachycondyla villosa engage in antennal boxing and biting and by these interactions establish social dominance hierarchies, in which several high-ranking individuals may lay eggs. We observed egg cannibalism by dominant workers. The presence of larvae negatively effects the number of worker-produced eggs in queenless colony fragments in two ways. First, larvae feed on the eggs, and second, workers lay fewer eggs when larvae are present. In queenright colonies, workers lay eggs at a low rate, which are eaten by the queen or fed to the brood, but apparently they do not compete aggressively for egg-laying."
Camargo and Caetano (1995) report that reproductive eggs and trophic eggs of workers are histologically disinct.
Dejean (1990) has studied nest site selection, and demonstrated a form of imprinting, in which workers exhibit a preference for their natal nest substrate. He carried out the study in the Sian Ka'an Biosphere Reserve, in the state of Quintana Roo, Mexico. He found that in the field villosa nested almost exclusively in Aechmea bracteata (Bromeliaceae). When offered choices, colony fragments from the field preferred nest sites with pieces of Aechmea. Lab colonies reared with fragments of Schomburgkia tibicinis (Orchidaceae) preferred nest sites with Schomburgkia. Lab colonies reared without plant contact showed no preference among nest sites with and without plant fragments.
Valenzuela et al. (1994) examined activity patterns and foraging habits in cacao plantations in the Soconusco region, Chiapas, Mexico. They reported
"Most foraging was in cacao trees, although nests were always located in other trees. Foraging occurred exclusively between 03:00-04:00 and 19:00-20:00 h. Out of 467 samples of material taken to the nest by foraging ants, 43 percent was made up of solids and 57 percent of liquids, the latter obtained mainly from the extra-floral nectaries of the cacao. Solid foods were mostly insects belonging to eight orders, the most frequent being Homoptera and Lepidoptera. Homopteran prey were mainly nymphs and some adults of Clastoptera globosa (Cercopidae), which can be an important pest of cacao trees. Most of the insects prey were phytophagous and only a small fraction were entomophagous. These results show that P. villosa can be an important natural control of the populations of some phytophagous insects that are potential pests for the cacao in this region."
Workers typically antennate potential prey, but when attacking dangerous prey (such as termite soldiers), workers exhibit "prudent" behavior, in which they hold the antennae back out of the way. The prudent behavior is learned, and develops after first contact with dangerous prey (Dejean et al. 1990). Workers harvest extrafloral nectar and carry it as a droplet between the mandibles. These droplets are shared among all colony members, including larvae (Dejean and Corbara 1990).
4-methyl-3-heptanone is an alarm pheromone produced in the mandibular glands (Duffield and Blum 1973).
Mathias et al. have done histological studies on the mandibular glands (Mathias et al. 1991) and female reproductive system (Mathias and Caetano 1992a, b).
Note added 23 March 2007: Bill Mackay alerted me to the paper by Lucas et al. (2002). They found that "villosa" was composed of three sympatric cryptic species at a site in Brazil. The anterior face of the petiole is flat in P. villosa s.s., and it is concave in the other two. The two with concave face were separated by relative height and length of the petiole. One of the concave forms they referred to P. inversa (Smith 1858), the other they thought was possibly new. Mackay (pers. comm.) pointed out that P. villosa curvinodis Forel 1899 is another candidate. Among the small amount of Costa Rican material here in my collection, I was able to separate collections cleanly into two forms, with flat versus concave anterior petiolar face. I was not able to separate the latter group into two forms. For now I am using the names villosa and curvinodis for these two forms. They are broadly sympatric in Costa Rica and do not seem to separate by habitat.
After comparative observations of multiple specimens of both Pachycondyla villosa and P. curvinodis, I found two other characters confirming the villosa/curvinodis division. The first is the postero-lateral ridges of the propodeum: on curvinodis, the ridges were more pronounced and were crenulated where the setae were inserted. On villosa, these ridges are rounder and if crenulated, were subtly so. The second, and in my opinion, far better character is the antero-ventral process on the first gastrite, near the insertion of the petiole. This process was acute and tall on all of the workers of villosa, while broad and shorter on the curvinodis specimens. I had no queens of villosa, while I had four of curvinodis. Two of the queens could be determined at curvinodis using this character, whereas the other two could not.