Amyrmex

Ward, P. S. & Brady, S. G., 2009, Rediscovery of the ant genus Amyrmex Kusnezov (Hymenoptera: Formicidae) and its transfer from Dolichoderinae to Leptanilloidinae., Zootaxa 2063, pp. 46-54 : 49-53

Redescription of male of Amyrmex Kusnezov

Head broader than long (CI 1.25-1.37), with large convex eyes that occupy the anterior two-fifths to one half of the sides of head (REL 0.43-0.57) (Figure 2); mandibles slender, elongate-triangular to sublinear, masticatory margin edentate and weakly differentiated from the unarmed basal margin; external margin of mandible curved basally, straight medially, and bent slightly mesad at apex (Figures 2, 4-5); mandible tips crossing at closure, mandible length subequal to eye length (ML/EL 0.88-1.10, ML/HW 0.35-0.41); genal teeth and hypostomal teeth lacking; clypeus short and transverse, bordered anteriorly by a thin clypeal lamella (apron); antennal sockets horizontal and exposed, and located close to the anterior clypeal margin; antenna 13-segmented, each segment longer than wide; scape of moderate length, SI 0.27-0.31; scape length subequal to or less than the length of ultimate antennal segment (SL/LA13 0.73-0.97), scape length 0.11-0.14* total length of antenna, less than twice the length of the second antennal segment (SL/LA2 1.50-1.78), and slightly more than half the combined length of the second, third and fourth antennal segments (SL/(LA2+LA3+LA4) 0.52-0.65); lateral ocelli separated from median ocellus by about their diameters.

Mesosoma with distinctive pronotum (Figure 3): U-shaped in dorsal view and reduced anteromedially to a thin horizontal strip, set well below the level of the dorsally protruding mesonotum; pronotum triangular in profile, with pointed posterior apex directed towards the wing base; mesonotum lacking notauli; parapsidal sutures very weakly impressed, barely discernable; axillae not meeting medially, connected by a narrow furrow; tegula very small and inconspicuous; mesopleuron lacking oblique transverse sulcus and hence not divided into anepisternum and katepisternum; mesoscutellum prominently bulging, as seen in lateral view; metapleural gland reduced and inconspicuous; propodeal spiracle small, circular, positioned at about midheight of propodeum and slightly posterior to the metanotum. Legs slender (LHT/HL ~1.2); mesotibia and metatibia each with a single short spur; tarsal claws lacking preapical tooth.

Wings with reduced venation (Figure 6); pterostigma present; forewing with elongate submarginal cell, four times longer than wide, and longer than basal cell; base of Rs weak, and absent in one species (Amyrmex BR02), resulting in confluence of the basal and submarginal cells; submarginal cell extending distad of stigma and terminating in an acute point; no free M vein after Rs+M; discal (medial) cell lacking, i.e., m-cu crossvein absent; A merging into cu-a, which curves anteriorly to join M+Cu at the point where veins M and Cu diverge, hence no free A vein distal to cu-a (Figure 6); hindwing lacking closed cells; anterior margin of hindwing with 1-4 hamuli; jugal lobe absent.

Metasoma slender in profile, obovate in dorsal view, widest at abdominal segment 5; abdominal segment 2 (petiole) subquadrate in profile (Figure 3), longer than high or wide, and only weakly constricted posteriorly, the helcium thus apparently quite broad; spiracle on abdominal segment 2 located on anterior third, near anterodorsal extremity; abdominal segment 3 larger than petiole, and not developed as postpetiole nor separated from abdominal segment 4 by a marked constriction; abdominal spiracle 3 located on anterior third of tergite; abdominal segments 2 and 3 with tergosternal fusion; abdominal segment 4 lacking tergosternal fusion; segment 4 with short but distinctly differentiated presclerites; spiracle present on anterior half of tergite 4; abdominal segments 5 and 6 lacking well differentiated presclerites, and not separated from succeeding segments by constrictions; abdominal spiracles 5 and 6 not discernable in specimens examined but possibly present at anterior margins of respective tergites; abdominal tergite 8 (pygidium) small and simple but visible dorsally, not wholly covered by abdominal tergite 7; cerci absent; subgenital plate (abdominal sternite 9) with posterior margin broadly concave but not bifurcate; basal ring not hypertrophied; paramere small and slender with rounded apex, paramere about 0.8* petiole length; volsella a simple, elongate-triangular lobe, lacking differentiated cuspis.

Body size very small; total length, excluding appendages, approximately 1.1-1.7 mm; HW 0.32-0.41, LHT 0.29-0.39; integument mostly smooth and shiny, with scattered piligerous punctures; pilosity common on most of body, suberect to decumbent. Color: body yellowish-brown to medium-brown, head and posterior margins of abdominal segments 4-7 darker, appendages (antennae, mandibles, legs) lighter.

Amyrmex golbachi , 1 male, ARGENTINA Formosa: Estancia Guaycolec, 25km N Formosa, 185m, 25°59'S 58° 12W, 17-20.xii.1998, Malaise trap, S. L. Heydon ( UCDC) (CASENT0106195).

Amyrmex BR01, 4 males, BRAZIL Rondonia: Fazenda Rancho Grande, 62km S Ariquemes, 165m, 12-22 November 1991, 10°18' S, 62°53'W, E. M. Fisher ( CASC, MZSP, UCDC) (CASENT0106161, CASENT0106183, CASENT0106184, CASENT0106185).

Amyrmex BR02, 1 male, BRAZIL Rondonia: Fazenda Rancho Grande, 62km S Ariquemes, 165m, 12-22 November 1991, 10°18' S, 62°53'W, E. M. Fisher ( UCDC) (CASENT0106186).

The male of Amyrmex golbachi from Argentina (Formosa) matches the original description (Kusnezov 1953) and the images of A. golbachi males from Tucuman on AntWeb (www.antweb.org). Differences between the three taxa are summarized in Table 1.

Relationship to other Leptanilloidinae

From the foregoing description the following differences emerge between the known males of Leptanilloides (Donoso et al. 2006; Ward 2007) and Amyrmex:

Amyrmex

Small body size, HW 0.32-0.41, LHT 0.29-0.39

Scape shorter: SI 0.27-0.31; scape less than twice the length of the second antennal segment (SL/LA2 1.5-1.8)

Legs shorter, LHT/HL ~1.2 Paramere small, ~0.8* petiole length Veins M and Cu diverging at crossvein cu-a

Submarginal cell elongate: about four times longer than wide, longer than the basal cell, extending distad of stigma and terminating in an acute point No free M vein after Rs+M

Leptanilloides

Variable body size, HW ~0.40-0.64, LHT 0.40-0.72

Scape disproportionately longer: SI 0.37-0.41; scape more than twice the length of second antennal segment (SL/LA2 ~2.2)

Legs disproportionately longer, LHT/HL ~1.5 Paramere large, ~1.5* petiole length

Veins M and Cu diverging distal to crossvein cu-a by a distance greater than the length of the crossvein Submarginal cell less elongate: no more than three times longer than wide, shorter than the basal cell, and terminating at level of stigma

Free M vein after Rs+M (may be weak)

However, in the UCDC collection there are several other male specimens that weaken these distinctions. First, there are seven additional leptanilloidine males -apparently representing two species -collected at Fazenda Rancho Grande, Rondonia, Brazil from the same series as Amyrmex BR01 and Amyrmex BR02 (12-22 November 1991, leg. E. M. Fisher) (CASENT0106187 to CASENT0106193). These are small (HW 0.22-0.38) and similar to Amyrmex except that (1) the forewing submarginal cell is less elongate, shorter in length than the basal cell, and does not exceed the stigma; (2) the mandibles are elongate-linear and bowed (i.e., falcate); and (3) the parameres are broad and paddle-shaped, subequal in length to the petiole. Thus, from a single collection from this one rainforest site in Rondonia there are males representing at least four species of Leptanilloidinae , two of them conforming to the strict Amyrmex diagnosis (above) and two other Amyrmex-like males. Second, a single male from Barro Colorado Island, Panama (12.viii.1978, leg. R. B. & L. S. Kimsey) (CASENT0106194) matches the Amyrmex description except that it is larger (HW 0.48, LHT 0.43) with disproportionately longer, falcate mandibles (ML/HW 0.52, ML/EL 1.33), and a short submarginal cell, not exceeding the stigma and approximately equal in length to the basal cell.

It is important to note that our concept of the male caste of Leptanilloides is based on only two species, L. mckennae Longino (Ward 2007) and L. nubecula Donoso , Vieira & Wild (Donoso et al. 2006), both with relatively large workers (HW 0.54-0.64). The L. mckennae males are also large (HW 0.59-0.64). The measurements given for the male of L. nubecula suggest a rather small male (HW 0.32, HL 0.32; see Donoso et al. 2006) but these may be in error. Based on the scale bar in the illustration of the male (Donoso et al. 2006, figure 25) HW should be about 0.37 and HL 0.29, and both of these values are unusually low in relation to the size of the workers (HW 0.54-0.56) and gyne (HW 0.74). In any event these two species do not adequately represent the spectrum of diversity within the genus. No males have been associated with workers of the smaller species of Leptanilloides such as L. biconstricta Mann and L. sculpturata Brandao , Diniz, Agosti & Delabie. The leptanilloidine males from Rondonia and Panama with short submarginal cells and falcate mandibles might belong here.

A further complication is that neither males nor DNA sequence data are available for the other workerbased leptanilloidine genus Asphinctanilloides Brandao, Diniz, Agosti & Delabie (1999), which was recovered as sister to Leptanilloides in morphological phylogenetic analyses (Brandao et al. 1999; Brady & Ward 2005). The three known species of Asphinctanilloides are associated with lowland Amazon and Atlantic coastal rainforest, which contrasts with the predominantly Andean and Central American distribution of Leptanilloides . Thus it seems quite possible that Amyrmex, known currently from the Amazon basin of Brazil and from northern Argentina, represents a senior synonym of Asphinctanilloides . For this reason we refrain from describing any of the Amyrmex-like males as new species since names may already be available for them (i.e., Asphinctanilloides anae Brandao , Diniz, Agosti & Delabie, A. amazona Brandao , Diniz, Agosti & Delabie and A. manauara Brandao , Diniz, Agosti & Delabie). Confirming this will require more extensive study. Since it may prove difficult to find worker-associated males of leptanilloidines in the field, DNA sequencing offers a reliable way of associating these disparate castes, if the material is sufficiently well preserved.

In the original description of Asphinctanilloides the workers were said to be distinguished from those of Leptanilloides by several features including (1) presence of a metanotal groove, (2) reduced postpetiole, smaller than the petiole as seen in profile, and (3) abdominal segments 5 and 6 lacking differentiated presclerites and hence without constrictions between abdominal segments 4 and 5, and 5 and 6 (Brandao et al. 1999). With the subsequent discovery of Leptanilloides species whose workers have a metanotal groove and a short postpetiole the first two characters have lost their diagnostic value (Longino 2003; Donoso et al. 2006). Brandao et al. (1999) also documented differences in the sting apparatus between Leptanilloides and Asphinctanilloides , however, and these have not been evaluated in the newly described species of Leptanilloides .

We examined two workers of Asphinctanilloides amazona (BRAZIL Amazonas: 28km N Manaus, 1.xii.1998, Berlese soil sample, leg. M. Verhaagh; CASENT0006016, CASENT0006815) ( CASC) and found that in both workers abdominal segment 5 has a differentiated presclerite, and is separated from abdominal segment 4 by a weak constriction. No such constriction occurs between abdominal segments 5 and 6. Further assessment of the morphological differences between the two genera is warranted. (Attempts to extract DNA from Asphinctanilloides workers belonging to this series were unsuccessful, apparently due to their initial collection into low concentration ethanol.)

Nevertheless, if a reduced worker postpetiole and an undifferentiated presclerite on abdominal segment 6 are derived features within Leptanilloidinae then it might still be possible to define Asphinctanilloides (or Amyrmex, if the two prove to be synonyms) in such a way that it is monophyletic, but this could also render Leptanilloides paraphyletic. These considerations, together with the discovery of a variable assortment of leptanilloidine males from scattered Neotropical localities, make the delimitation of genera in this group an ongoing challenge.

Conclusions

The ant genus Amyrmex Kusnezov was described over half a century ago from several males collected in Tucuman, Argentina, and it has been in a state of taxonomic limbo since then. Placed awkwardly in the subfamily Dolichoderinae -and even synonymized under the dolichoderine genus Forelius for a period of time -it is here shown to be a member of the subfamily Leptanilloidinae , within the dorylomorph clade. The association was established with DNA sequence data, but it is also supported by a reevaluation of morphological features. The relationship of Amyrmex, still known only from males, to the two worker-based leptanilloidine genera, Leptanilloides and Asphinctanilloides , remains uncertain, although the geographical distribution of Amyrmex suggests that it might be a senior synonym of Asphinctanilloides . Establishing generic limits within the Leptanilloidinae will require additional assessment of worker- and male-based material. It is becoming increasingly apparent that DNA sequences, in addition to resolving phylogenetic relationships among taxa, also offer a powerful source of evidence for correctly associating the male and female castes of ants such as leptanilloidines (Ward 2007) and dorylines (Schoning et al. 2008), where the workers tend to be subterranean and the males are collected separately at lights or in Malaise traps. This emphasizes the importance of collecting specimens into media, such as 95% ethanol, that provide adequate long-term preservation of DNA.