Pachycondyla apicalis

Wild, A. L., 2005, Taxonomic revision of the Pachycondyla apicalis species complex (Hymenoptera: Formicidae)., Zootaxa 834, pp. 1-25 : 13-15

[[ Pachycondyla apicalis species complex ]]

Species boundaries

The morphological evidence considered in conjunction with geography supports the separation of the P. apicalis complex into three species. All three are broadly sympatric from northern South America to southeastern Brazil, and the two more common species, P. apicalis and P. verenae , co-occur in Central America north to southern Mexico (Fig. 11). It is notable that every record of the rarer species P. obscuricornis involves sympatry with P. apicalis (2 sites), P. verenae (2 sites), or both (3 sites). Furthermore, 29 of 81 records of P. apicalis (35.8%) show sympatry with at least one of the other two species, and 29 of 60 records of P. verenae (48.3%) indicate sympatry as well. This pattern is likely to be an underestimate of local sympatry given that many records result from haphazard hand collecting and not from thorough myrmecological surveys. Sites that have been extensively surveyed (e.g., La Selva Biological Station in Costa Rica and the Mbaracayú Reserve in Paraguay) usually uncover at least 2 of the species.

Figures 12-17 illustrate some of the morphometric differences among these species. These differences are consistent across the entire range of the apicalis complex, including a number of sites where two or more species locally coexist. In particular, P. obscuricornis consistently has shorter antennal scapes than P. apicalis and P. verenae (Figs. 12, 13), a relatively broader head (Fig. 14), and a more pubescent hypopygium (Fig. 8). Pachycondyla apicalis can be separated from P. verenae by a taller petiolar node (Fig. 17), by a broader head (Fig. 16), by the lack of strong margination of the petiolar node, by the coloration of the antennal apices, and in South America (but not Central America) by the more extensive pilosity. Pachycondyla apicalis is usually larger in most measurements than P. verenae , but there is enough overlap that size alone is not always diagnostic. Given the strength and consistency of the morphological differences across multiple instances of sympatry, significant gene flow between these species is unlikely.

The common species P. verenae and P. apicalis both show considerable variation across their ranges in a number of characters, including petiole shape, head shape, scape length, eye size, and pilosity. This variation is either localized or broadly allopatric over a north-south cline, and thus in spite of the overall amount of variation there appears to be no justification for further division of the complex. In contrast, P. obscuricornis shows almost no variation across its range, suggesting a recent origin or a population bottleneck.

Phylogenetic relationships among the three species are unknown. Some characters, such as antennal scape length and hypopygial pubescence, suggest a closer relationship between P. apicalis and P. verenae , while others, such as mesosomal and petiolar configuration, indicate a relationship between P. obscuricornis and P. apicalis . Molecular genetic data will likely provide the most satisfactory resolution to the problem. Given the variation in the life-histories of these ants (e.g., monogyny in P. apicalis and polygyny in P. verenae ) and the amount of published biological information, a phylogeny could be of great utility.

Nomenclature

The amount of nomenclatural confusion in this complex is surprising considering the small number of species involved. Latreille (1802) described two similar species from South America, Formica flavicornis and F. apicalis . I have not seen type specimens of either, but the description of the shape of the node and the coloration of the antennal apices leaves little doubt about the identity of apicalis , a conclusion also reached by Brown (1957). The identity of flavicornis is not as clear. The name flavicornis is preoccupied by an older Fabricius (1798) species, and Forel (1905) provided Neoponera latreillei as a replacement name. Brown (1957) placed flavicornis and latreillei as a synonyms of apicalis on the basis of antennal coloration, a decision that I see no reason to challenge.

Emery was alone among previous workers in correctly recognizing three distinct entities, although he misapplied the name apicalis . In his 1890 paper “Voyage de M. E. Simon au Venezuela”, he discussed a form with a strongly marginate petiolar node (“ apicalis ”), a form with a more rounded node and yellow antennal tips (“ flavicornis ”), and a new variety of flavicornis with dark antennal tips (“ obscuricornis ”). Material determined by Emery in MHNG confirms that these names correspond to P. verenae , P. apicalis , and P. obscuricornis , respectively. Emery was content to keep the two forms with the more rounded node as varieties of a single species rather than as separate species, so he placed obscuricornis as a variety of flavicornis , and later latreillei as a variety of obscuricornis in “Genera Insectorum ” (Emery 1911) once Forel (1905) proposed latreillei as a replacement name for the preoccupied flavicornis .

Forel’s handling of the group parallels Emery’s but is less consistent. Examining the Forel collection at MHNG, I found that Forel largely followed Emery’s treatment of the species with the marginate node as apicalis and that with the yellow antennal apices as flavicornis / latreillei . Forel also provided new names for relatively slight variants, verenae from Panama as a variety of apicalis , and latocciput from Ecuador as a race of obscuricornis . Why he assigned latocciput to obscuricornis instead of latreillei is not clear, all the more so since he acknowledges both taxa in the brief description. It is apparent, however, that he put more consideration into form than coloration.

Brown’s (1957) primary contribution was the realization that Latreille’s original description of Formica apicalis matched the species that Forel and Emery had been calling first flavicornis and then latreillei . He put these into synonymy under apicalis . Brown evidently did not examine relevant specimens in either Forel or Emery’s collections, however, so he did not know that Emery and Forel used the name apicalis in a sense opposite to his own. This led him to infer erroneously that Forel’s Neoponera apicalis var. verenae must be a variety of Latreille’s apicalis instead of the oldest available name for the species with the marginate petiole.

Brown apparently did not see sufficient South American material prior to his 1957 publication to uncover the rarer species P. obscuricornis , as he lays out a case for the existence of “two and only two species” (Brown 1957, pg 231). Brown’s two species, “ apicalis ” and “ obscuricornis ”, are unambiguously apicalis Latreille and verenae Forel . Under his two-species dichotomy, Brown argued in the absence of type material that Emery’s obscuricornis could not be apicalis because of the antennal coloration. By default he treated it as the valid name for the other widespread species ( P. verenae [Forel] in the present study). Apparently Brown did arrive much later at a three-species conclusion that was never published, although it is unlikely he recognized the third species as being Emery’s obscuricornis (C. R. F. Brandão, pers. comm).

Brown’s reorganization was perpetuated in an unpublished but widely circulated manuscript key to Neotropical Pachycondyla . Consequently, the common species P. verenae has been misdiagnosed consistently as P. obscuricornis in the literature and in museum collections. In those cases where voucher specimens of “ obscuricornis ” were available in this study (Kempf 1972, Duelli & Duelli-Klein 1976, Wild 2003, Longino 2004), as well as in Traniello and Hölldobler (1984) which contained sufficiently detailed photographs, all records actually pertain to P. verenae . Since there are no collection records of true P. obscuricornis from Central America, it is almost certain that other MesoAmerican studies of “ obscuricornis ” (e.g., Fresneau 1984, Oliveira & Hölldobler 1991, Lommelen et al 2002) also pertain to P. verenae . More ambiguous are South American references (e.g., Wheeler et al 1999, Düssman et al 1996). Given the relative rarity of true P. obscuricornis , most South American records may also refer to P. verenae , but the identity of these cannot be verified without the examination of specimens.

Unfortunately, taxonomic instability may continue to persist for some time in the Pachycondyla apicalis complex. While the species-level nomenclature may stabilize with the recognition of three species, it is almost certain that the heterogeneous genus Pachycondyla is paraphyletic with respect to much of the rest of the tribe Ponerini (C. Schmidt, unpublished molecular data). As generic taxonomy falls in line with new phylogenetic hypotheses, it is entirely possible that various generic names, including Neoponera for the species discussed here, will be resurrected to retain monophyly of the ponerine genera.