Insect. Soc. 55 (2008) 296-300 0020-1812/08/030296-5 DOI 10.1007/s00040-008-1005-6 ? Birkhauser Verlag, Basel, 2008 | Insectes Sociaux Research article Parabiosis between basal fungus-growing ants (Formicidae, Attini) C.E.D. Sanhudo, T.J. Izzo and C.R.F. Brandao* Museu de Zoologia da Universidade de Sao Paulo, Av. Nazare 481, Sao Paulo, SP, Brasil, 04263-000, e-mail: crfbrand@usp.br, Fax: 55-11-61658115 Received 30 July 2007; revised 1 February and 7 April 2008; accepted 18 April 2008. Published Online First 14 May 2008 Abstract. We describe the first observation of parabiosis between two Attini ants (Apterostigma urichii Forel and Cyphomyrmex faunulus Wheeler) found in northern Manaus, AM, Brazil. Complete, mature colonies of both species were found in a single cavity inside a rotten log, sharing and tending a single combined fungus garden, made up of two distinct halves, each cultivated by one species. Workers of one species often antennated workers of the other species and showed no aggression toward each other or toward each other's workers, queens, or immatures. Laboratory observations suggest that imma- tures of both species feed on hyphae from either half of the fungus garden. We were not able to find other parabiotic pairs involving the same species in the same locality, although we found colonies of both species sharing trails and foraging territories. Keywords. Parabiosis, basal Attini, ants, fungus-growing. Introduction Parabiosis is defined as a particular form of facultative or obligatory symbiosis in which two or more species utilize the same nest and sometimes even the same odour trails, but nevertheless keep their broods separate (Holldobler and Wilson, 1990). It was first described by Forel (1898), based on his observations of the association between Crematogaster limata parabiotica and Dolichoderus de- bilis in Colombia. Forel {op. cit.) emphasized the facul- tative nature of this particular association, pointing out that, in general, neither of the participating species is dependent on the association; although parabiotic spe- cies often live together, they may also be found nesting ' Author for correspondence. and living independently. However, in some cases, the relationships between parabiotic species are not neces- sarily amicable and mutualistic, but instead may be parasitic, with one of the species driving away the other at the food source (Swain, 1980). Parabiosis has been described for ant species from all over the globe (e.g. Santschi 1910; Greaves and Hughes, 1974; Greenslade and Haliday, 1983; Kaufmann and Maschwitz, 2006), but most records of this association report facultative relations and are concentrated in the Neotropics, where several parabiotic ant associations, generally involving members of different subfamilies, have been recorded (Mann, 1912; Wheeler, 1921; Weber, 1943; Orivel etal., 1996; Davidson, 1988; Ipinza-Regla etal., 2005). Until now, the only case of parabiosis between ants belonging to the same subfamily, but to different tribes, involves the Formicinae Brachymyrmex and Camponotus (Errard et al., 2003). We describe here the first observation of parabiosis between two closely related ant species, of Apterostigma and Cyphomyrmex, both fungus-growing ant genera belonging to the tribe Attini (Myrmicinae). Most fungus-growing ants are obligatory cultivators of basidiomycete fungi, on which they depend for nourish- ment (Holldobler and Wilson, 1990). Recently, Mikheyev et al. (2006) challenged the paradigm that fungus-growers ecological success depends largely on ancient fungus vertical transmission allied to suppression of fungal sexuality. Green et al. (2002) and Mikheyev et al. (2006) were able to show evidence of frequent cultivar recombi- nation in sympatric attines and long-distance horizontal transmission of symbionts independent of their ant hosts. As in most well characterized mutualisms, the general principle shifted from one-to-one specificity to diffuse coevolution between participants. Mikheyev et al. (2007) show that garden exchange likely happens continuously at the population level. However, no behavioural mech- anism enabling the ants to either switch to a different Insect. Soc. Vol. 55, 2008 Research article 297 cultivar or regain their cultivar after loss has been clearly identified thus far. In nature, fungus garden loss is due to pathogens or other causes. Experimentally, loss of the fungus garden induces fungus-growing ants to obtain a replacement cultivar from a neighbouring colony, either by joining the other colony and cooperating in the cultivation of a common garden, or by stealing or aggressively usurping a neighbour's garden (Adams etal.,2000). Additionally, fungus-growing ants may share nests with non-attine ants, at least temporarily. Adams et al. (1998-1999; 2000) discovered colonies of Megalomyr- mex wettereri Brandao in Panama containing healthy fungus gardens of Cyphomyrmex longiscapus; laboratory observations revealed that the Megalomyrmex "consume the fungus by cropping mycelium from the garden substrate. However, they do not forage for and add nutrient substrates, or otherwise tend the fungus garden; thus, when the garden becomes depleted, M. wettereri must locate and usurp new gardens in other attine colonies". Other Megalomyrmex species of the Silvestrii species group have been found within the fungus gardens of various basal Attini (Brandao, 2003). Attini is a diverse array of predominantly Neotropical genera (13 accepted as valid presently), which have been traditionally (and artificially) divided into the "Lower" attines, sometimes called basal, which includes the so- called Paleoattini (Apterostigma, Mycocepurus and Myr- micocrypta) and six genera of unknown affinities, and the Neoattini (also known as the "Higher" attines), which includes the leaf cutting Atta and Acromyrmex, and, for some authors, Trachymyrmex and Sericomyrmex as well. The Lower Attini live in colonies with at most a few hundreds of workers, in simple nests, whereas the Higher attines live in much larger colonies inhabiting complex nests. In general, basal Attini display little or no inter- or intraspecific aggression towards other ant species sharing their foraging territories, as already pointed out by Weber (1941), who noted how closely together different basal fungus-growing species may nest. Because basal Attini use very abundant organic detritus as substrates for fungus gardening, including pieces of arthropod carcasses and faeces (Leal and Oliveira, 2000), and the species are generalist foragers, it is believed that competition for resources is very low among them, and thus that aggression is also low (Tallamy and Wood, 1986). The results of a study on the foraging ecology of attine ants in a Brazilian savannah, in special as to the seasonal use of fungal substrate (Passos and Oliveira, 2003), revealed that the basal attines are very opportunistic and utilize plant parts according to the phenology of the vegetation. These attines preferred to collect available items from nearby plants and always harvested the fungal substrate on the ground. Recently-fallen flowers and fruits com- prised the major part of the fungus-substrate used by all genera studied there. Tougher plant parts, such as leaves and seeds, insect remains, and other material were also used by the ants, particularly in the dry season, probably to compensate for the lower availability of flowers and fruits during this period. However, in the more derived leaf cutters Acromyrmex and Atta, which cultivate fungi on macerated fresh vegetation, competition for resources may be high, with consequent territoriality and aggres- sion among neighboring colonies, which may escalate into deadly wars among colonies, as seen in several ant species (Holldobler and Wilson, 1990). Obligatory parabiosis has been described for several pairs of ant species, usually involving species from distant phylogenetic lineages. For instance, Crematogaster limata parabiotica (Myrmicinae) and Camponotus femoratus (Formicinae) share nests and the same trail systems, although they utilize different food resources (Holldobler and Wilson, 1990). Species with similar resource require- ments are potential competitors and so are not expected to enter into parabiotic associations. Methods One parabiotic colony of Apterostigma urichii Forel and Cyphomyrmex faunulus Wheeler was found by the first two authors in a rotten log cavity in the Reserva Dimona, Projeto Dinamica Biologica de Fragmentos Florestais (PDBFF), northern Manaus, AM (02?20'S; 65?05'W), Brazil, between November 27 and 29, 2004. The collectors actively searched for more colonies of both species in Dimona Farm, Colosso Farm (02?24'S; 59?52'W) and Km 41 Reserve (02?26'S; 59?46'W), all administrated by PDBFF (Bierregaard et al., 2002). In four weeks of intense field work they found several colonies of C. faunulus and A. urichii nesting in isolation in the three localities, but no other parabiotic pair. They found, in the Colosso Farm, two more heterospecific pairs of the cited species nesting below the bark of two different tree logs, but in both cases each species occupying its own chamber, less than 30 cm apart of the other species chamber. The parabiotic colonies and the unique fungus mass were trans- ferred to an artificial nest constructed from two 10 cm2 acrylic boxes linked by a plastic tunnel; one box served as the foraging arena and the other as the colony chamber. We offered as garden substrate ground dried orange rind, dried oats, and the frass of an Urbanus sp. butterfly (Hesperidae), all of which were accepted by workers of both species. Unfortunately, the workers were unable to fully reorganize the fungus garden (see below), and its health rapidly declined after arriving in the laboratory; workers of both species began to die upon arrival in the lab. The colonies survived for one week in artificial conditions. Observation The Apterostigma uruchii and Cyphomyrmex faunulus colonies living in parabiosis had one dealated queen each and contained, respectively, 23 and 16 workers, two and one males, and multiple larvae and pupae. The single fungus garden filled up almost half of the cavity, and was clearly divided into two distinct halves, a reddish one inhabited by C faunulus and a greenish one inhabited by A. urichii (Fig. 1). Approx- imately one-third of the combined fungus garden showed signs of deterioration and contained no immatures. The garden substrate in both halves consisted of unidentified anthers, small seeds, and frag- ments of insect exoskeletons (elytra, heads, antennae, and wings). The total volume of the fungus garden was 350 ml and the total weight, including the ants from both colonies, was 120 g. In the half of the garden occupied by A. urichii, we found the pupa of a Muscidae fly (Helina sp.), from which an adult emerged immediately after the arrival of the colony in the lab. However, the adult fly died soon after emergence and, because its features had not completely sclerotized still, species-level identification was impossible. 298 C.E.D. Sanhudo et al. Parabiosis in basal Attini Figure 1. Double-origin combined fungus garden shared by Cypho- myrmex faunulus (left) and Apterostigma urichii (right) found at Reserva Dimona Projeto Dinamica Biologica de Fragmentos Florestais - PDBFF northern of Manaus, AM, Brazil. Note at the left half, the C faunulus dealate gyne (middle) and a worker (above) and, at the right half, a worker and larvae (behind and at the down corner) of A. urichii. Picture taken in the lab by Domingos J. Rodrigues. As is typical for live-collected attine colonies, the fungus garden almost completely broke apart while being transported from the field to the lab. Ants of both species worked together to rebuild it, and even transported each other's immatures to the newly partially rebuilt fungus garden. We cannot assure that each species tended the immatures of the partner species, but our impression is that immatures of both species were fed and transported by workers of both species, based on the significantly greater size of the Apterostigma larvae, in relation to those of Cyphomyrmex. Workers of both species frequently antennated each other and walked freely over each other's halves of the fungus garden. On several occasions, we observed a worker of A. urichii antennating the C. faunulus dealate queen. We never observed even a single instance of aggression between the species during the week in which they were maintained in the lab. The heterospecific pairs found nesting very close in Colosso Farm tree logs, although occupied different chambers, shared natural trails formed by crevices below the bark of the rotting trunks. Examination of the logs suggests that workers of both species even visited each other nests, and also do not show any aggressive behaviour towards each other. Discussion Colonies occupying different nest chambers inside rotten logs can be easily mixed together when the log is opened in the field. We rule out this possibility in the present case and are sure that the co-occurrence of two basal attines in a single nest chamber does not represent a collecting artefact. The fact that complete colonies were sharing a unique fungus mass inside a single chamber, with a single entrance, and the lack of aggressiveness among workers of the two species, indicates that the parabiotic colony and the combined fungus garden were truly living together. Also, we had the opportunity to observe other hetero- specific pairs of these species sharing trails, but living in distinct nests. We reject also the option that we had found, by chance, a momentary raid of a parasite or lestobiotic (when colonies of a small species nest in the walls of a larger species and enter the chamber of the larger to prey on brood or to rob food, Holldobler and Wilson, 1990) species attacking its host, because we found dealated reproductive females and mature males, along with immatures, in both colonies of the parabiotic pair; raids are normally conducted only by workers. Furthermore, we never observed these ants attacking each other, as would be the case if we had by chance found one of the species attacking the nest of the other. It is important to distinguish the case described here from the sharing of a fungus garden induced by the loss of a garden, e.g., due to pathogens (Adams et al., 2000), in which the colony of one attine species invades another colony (conspecific or of another species or even genus) in order to acquire a replacement cultivar. In the case reported here, both cultivated fungi are assumed to have been present and to have together composed a combined fungus garden (see below). As expected from such distantly related fungi, each was distinct and the two were not fully integrated, either in natural or laboratory conditions, even if the recomposition of the fungus in the laboratory was only partial. According to Adams et al. (2000), usurpation involves the stinging and ultimately killing of the host species, contrariwise to the complete lack of aggresiveness between members of the different colonies we observed in the field and laboratory. Apterostigma urichii belongs to the so-called Paleo- attini and Cyphomyrmex faunulus is a basal member of the Neoattini. We would not expect leaf-cutters (derived members of Neoattini, Atta and Acromyrmex) to engage in parabiosis because leaf cutters compete for and defend their resources, resulting on territorialism, aggression, and even wars among neighbouring colonies. Even in permanent parabiotic relationships, there is a conflict of interest between partners (Vantaux et al., 2007) and one species can eject the other from the common nest (Swain, 1980). Because both Apterostigma urichii and Cypho- myrmex faunulus are commonly found nesting alone, the parabiotic association between them is obviously facul- tative. Similar facultative associations have been record- ed between Crematogaster limata parabiotica and either Pachycondyla goeldii or Camponotus femoratus (David- son, 1988; Orivel et al., 1996). If we accept the putative absence of competition among basal Attini ants, this absence per se, cannot fully explain the mutualistic relationship observed in the case described here. In addition to the requirement that both species do not compete for resources, the mutualism also requires that the association is selectively advantageous to both partners, which however, it is not always possible to ascertain. In the present case, favourable nesting sites could be limited (see Kaspari, 1996), leading these ants, that do not compete for foraging resources, to share nests, as they live in relatively small colonies and can not efficiently guard nesting sites. When nesting sites are the Insect. Soc. Vol. 55, 2008 Research article 299 limiting resource, Solenopsis invicta queens cooperate to found colonies pleometrotically (Tschinkel, 1998). Sim- ilar benefits derived from primary pleometrosis could help to explain the parabiotic association studied here, as, for instance, faster worker production in recently founded colonies and improved group defence against alien workers from mature colonies (reviewed in Holldobler and Wilson, 1990). Besides the facultative nature of the observed parabiosis, the relation can also be only temporary. The case described here is unusual because it involves a parabiosis between two ant species from two different genera that reportedly cultivate two very distinct fungal groups. Apterostigma species of the Pilosum group, which includes A. urichii, are exceptional among Attini in cultivating wood-decomposing (coral-mushrooms) Pter- ulaceae fungi (called G2 and G4), which are only distantly related to the litter-decomposing Lepiotaceae (Basidio- mycota: Agaricales) that are universally cultivated by all other more than 200 species of fungus-growing ants, including the most basally diverging Apterostigma species in the Auriculatum group (Villesen etal., 2004). Culti- vation of Pterulaceae represents a unique symbiont switch, with basal Apterostigma species retaining ances- tral lepiotaceous fungiculture and with none of the monophyletic, pterulaceous-cultivating Apterostigma species known to have reverted back to lepiotaceous fungiculture. Munkacsi et al. (2004) showed that the Pterulaceae are morphologically and phylogenetic differ- ent from the Lepiotaceae, and discussed how the two fungi converged to the same evolutionary interactions with ants, despite their distinct origins. The Apterostigma urichii gardens analyzed so far came from Kurupukari, Guyana, and are clearly veiled G2 fungus gardens hanging on tree trunks (Villesen et al., 2004). In general, G2 pterulaceous fungus gardens are protected by a mycelial veil, but when G2-cultivators nest in enclosed spaces (such as a chamber in a log or in the ground), the veil may be inconspicuous or easily damaged by the collector. Pterulaceous fungus gardens in the G4 group, in contrast, do not have veils. That is, the Pilosum group Apterostigma appears to grow the veil only in response to desiccation, or to partition off the garden in a larger space. In short, presence of a veil is a good indicator that the Apterostigma species is in the Pilosum group, but the absence of a veil does not necessarily mean that the ant does not belong to the Pilosum group nor that it is not cultivating a pterulaceous garden (Mueller, pers. comm.; Villesen et al., 2004). Attine ants are able to differentiate between closely related cultivar strains. Switches to distantly related cultivars are behaviourally unlikely, but, hypothetically, may occur under constrained con- ditions in which colonies are forced to import novel cultivars (Mueller et al., 2004). The most surprising observation, however, is the combined nature of the fungus garden cultivated by the parabiotic colony. Unfortunately we were not able to culture it and we do not know whether or not the two garden halves consisted of the same or of closely related fungi in the Leucocoprineae or whether they consisted of two very different fungi, one in the Pterulaceae and one in the Leucocoprineae. It is reasonable to expect that such distantly related fungi, when brought into contact, would have antagonistic reactions to each other, "battling" with antibiotic chemicals produced by different bacterial strains, which would result in reduced garden productivity (Lenoir et al., 2001). In nature, the unique fungal mass reared by the parabiotic colony presented clearly two halves distinct by different colours, even when observed in different lightning and shades. Although the ants were not able to fully recompose the fungal mass after trans- port to the laboratory, they tried to do so, and the partially recomposed fungus also presented the two very distin- guished halves in laboratory conditions (Fig. 1). If we accept the hypothesis that the fungal mass actually belongs to a single fungus species, the two different colours could be the result from the ants adding different substrates for the same fungus in each one's half, which would imply in fungus malleability regarding substrate. If the fungus mass is the result of two combined fungus species, each one brought and reared by one of the ant species, this would imply in the compatibility of fungi from very different origins, and may thus represent an important observation in the context of the evolutionary history of the fungus-ant associations understanding. Even if parabiosis between other pairs of basal attines has not been registered until now, it may be common. For example, Dr. Ulrich Mueller (pers. obs.) recorded two attine species, both in the genus Mycocepurus, which also live together. MacKay et al. (2004) briefly mentions that Mycocepurus curvispinosus apparently lives within nests of M. smithii, in Veracruz, Mexico, with whom they share trails with. The purpose of the present report is to encourage colleagues to search for additional evidence of parabiosis among basal attines, in order to further corroborate the observation reported and to test the hypotheses proposed here, providing highly relevant contributions for the knowledge on the biology of the Attini, and on the evolution of interspecific relationships. Acknowledgments Antonio Mayhe-Nunes (UFRRJ) identified the Attini species. Dr Marcia S. Cury identified the Helina fly. Ulrich Mueller and Ted Schultz made important suggestions regarding this finding and to the manu- script. Mario de Pinna, Eliana Cancello, Rogerio Rosa da Silva and Rodrigo M. Feitosa critically read the manuscript. The authors are grateful to Nicole Gerardo, whose critical comments significantly improved the manuscript. Thanks are also due to an anonymous referee for his (her) critical review and suggestions. We thank the "Projeto Dinamica Biologica de Fragmentos Florestais - PDBFF" and CNPq (Conselho Nacional de Desenvolvimento Cienti'fico e Tecnologico, Brazil) for partial financing. This is manuscript number 511 of the PDBFF technical series. 300 C.E.D. Sanhudo et al. Parabiosis in basal Attini References Adams R.M.M., Mueller U.G. and Schultz T.R. 1998-1999. Nest usurpation and fungivory in Megalomyrmex sp. nov. (Silvestrii species group): lestobiotic or predatory relations with fungus- growing ants ? In: Insects at the Turn of the Millennium (Schwarz M.P. and Hogendoom K.E., Eds). Proc. Int. Congr. IUSSI, Adelaide, Australia, p 27 Adams R.M.M., Mueller U.G, Green A.M. and Narozniak J.M. 2000. Garden sharing and garden stealing in fungus-growing ants. Naturwissenschaften 87: 491-493 Bierregaard R.O., Gascon C, Lovejoy T.E. and Mesquita R. (Eds.) 2002. Lesson from Amazonia: the Ecology and Conservation of a Fragmented Forest. Yale University Press. New Haven. 496 pp Brandao C.R.F. 2003. Further revisionary studies on the ant genus Megalomyrmex Forel (Hymenoptera : Formicidae : Myrmicinae : Solenopsidini). Pap. Avuls. Zool. 43:145-159 Davidson D.W. 1988. Ecological studies of Neotropical ant gardens. Ecology 69:1138-1152 Errard C., Ipinza Regla J. and Hefetz A. 2003. Interspecific recognition in Chilean parabiotic ant species. Insect. Soc. 50: 68-73 Forel A. 1898. La parabiose chez les fourmis. Bull. Soc. Vaud. Set. Nat. 34: 380-384 Greaves T. and Hughes R.D. 1974. The population biology of the meat ant. /. Aust. ent. Soc. 13: 329-351 Green A.M., Mueller U.G. and Adams R.M.M. 2002. Extensive exchange of fungal cultivars between sympatric species of fungus- growing ants. Mol. Ecol. 11:191 -195 Greenslade P.J.M. and Halliday R.B. 1983. Colony dispersion and relationships of meat ants Iridomyrmex purpureus and allies in an arid locality in South Australia. Insect. Soc. 30: 82-99 Holldobler B. and Wilson E.G. 1990. The Ants. The Belknap Press of Harvard University Press. Cambridge, Mass. 732 pp Ipinza-Regla J., Fernadez A. and Morales A.M. 2005. Hermetismo entre Solenopsis gayi Spinola, 1851 y Brachymyrmex giardii Emery, 1894 (Hymenoptera, Formicidae). Gayana 69: 27-35 Kaufmann E. and Maschwitz U. 2006. Ant-gardens of tropical Asian rainforests. Naturwissenschaften 93: 216-227 Kaspari M. 1996. Testing resource-based models of patchiness in four Neotropical litter ant assemblages. Oikos 76: 443-454 Leal I. and Oliveira PS. 2000. Foraging ecology of attine ants in a Neotropical savanna: seasonal use of fungal substrate in the cerrado vegetation of Brazil. Insect. Soc. 47: 376-382 Lenoir A., Ettorre P., Errard C and Hefetz A. 2001. Chemical ecology and social parasitism in ants. Annu. Rev. Entomol. 46: 576-599 Mackay W.P., Maes J.-M., Fernandez PR. and Luna G. 2004. The ants of North and Central America: the genus Mycocepurus (Hymenop- tera: Formicidae). /. Insect Science, 4:1-7 Mann W.M. 1912. Parabiosis in Brazilian ants. Psyche 19: 36-41 Mikheyev A.S., Mueller U.G. and Abbot P. 2006. Cryptic sex and many- to-one coevolution in the fungus-growing ant symbiosis. Proc. Natl Acad. Sci. 103: 10702-10706 Mikheyev A.S., Mueller U.G. and Boomsma J.J. 2007. Population genetic signatures of diffuse co-evolution between leaf-cutting ants and their cultivar fungi. Mol. Ecol. 16: 209-216 Munkacsi A.B., Pan J.J., Villesen P., Mueller U.G, Blackwell M. and McLaughlin! D.J. 2004. Convergent coevolution in the domestica- tion of coral mushrooms by fungus-growig ants. Proc. R. Soc. Lond. 271:1777-1782 Mueller U.G., Poulin J. and Adams R.M.M. 2004. Symbiont choice in a fungus-growing ant (Attini, Formicidae). Behav. Ecol. 15: 357- 364. DOI: 10.1093/beheco/arh020 Orivel J., Errard C and Dejean A. 1996. Ant-gardens: interspecific recognition in parabiotic ant species. Behav. Ecol. Sociobiol. 40: 87-93 Passos L. and Oliveira PS. 2002. Ants affect the distribution and performance of seedlings of Clusia criuva, a primarily bird- dispersed rain forest tree. /. Ecol. 90: 517-528 Santschi F. 1910 ("1909"). Formicides nouveaux ou peu connus du Congo francais. Annls Soc. ent. Fr. 78: 349-400 Swain R.B. 1980. Trophic competition among parabiotic ants. Insect. Soc. 27: 377-390 Tallamy D.W. and Wood T.K. 1986. Convergence patterns in subsocial insects. Annu. Rev. Entomol. 31: 369-390 Tschinkel W.R. 1998. An experimental study of pleometrotic colony founding in the fire ant, Solenopsis invicta: what is the basis for association? Behav. Ecol. Sociobiol. 43: 247-257 Vantaux A., Dejean A., Dor A. and Orivel J. 2007. Parasitism versus mutualism in the ant-garden parabiosis between Camponotus femoratus and Crematogaster levior. Insect. Soc. 54: 95-99 Villesen P., Mueller U.G, Schultz T.R., Adams R.M.M. and Bouck A.C. 2004. Evolution of ant-cultivar specialization and cultivar switching in Apterostigma fungus-growing ants. Evolution 58: 2252-2265 Weber NA. 1941. The biology of the fungus-growing ants, Part VII: The Barro Colorado Island, Canal Zone, species. Revta Ent. 12:93-130 Weber NA. 1943. Parabiosis in neotropical "ant gardens". Ecology 24: 400-404 Wheeler W.M. 1921. A new case of parabiosis and the "ant gardens" of British Guiana. Ecology 2: 89-103 To access this journal online: http ://www.birkhauser.ch/IS