Type material.

Amblyopone australis Erichson, 1842. Lectotype [here designated]. Worker. AUSTRALIA: Tasmania. Scheyer [sic] (misspelling of Schayer) [Museum für Naturkunde der Humboldt-Universität Berlin (ZMHB): 7228, CASENT0104575. Examined based on images. Two specimens were found with same locality and collector information; one was labeled “Type.” The original description of this species does not include detailed type information; however, Erichson mentions in the same paper that Schayer provided much material from Tasmania [19]. Original labels on this specimen seem to have been replaced by new labels, and the collector on the new label is listed as “Scheyer,” a misspelling of Schayer, the collector's name.]

Non-Malagasy material.

Am. australis Erichson, 1842 [w]; Am. australis [m: The Australian National Insect Collection (ANIC) 32016150, CASENT0172246. Examined based on images].

Type material.

Stigmatomma denticulatum Roger, 1859. Holotype. Worker. Griechenland (Greece): Monte Scapo (Mt. Scapo), Insel Zante (Zante Is.). Von Kiesenwelter. [Museum für Naturkunde der Humboldt-Universität Berlin (ZMHB): 18992, CASENT0104566. Examined based on images. Only one specimen from Greece is in ZMHB and it is labeled “typus.” Original labels on this specimen seem to have been replaced by new labels. The new label mistakenly lists the collector as “Roger, S.” instead of Von Kiesenwelter.]

Malagasy material.

S. mg01 [g: CASENT0083104]; S. mgm01 [m: CASENT0111612]; S. mgm02 [m: CASENT0109221]; S. mgm03 [m: CASENT0007087]; S. mgm04 [m: CASENT0059491].

Non Malagasy material.

S. pallipes (Haldeman, 1844) [w]; S. denticulatum (Roger, 1859) [w].

Type material.

Stigmatomma (Xymmer) muticum Santchi 1914. Lectotype [here designated]. Worker. NIGERIA: Nigerie S. lbadan (Silvestri). Det. Santchi 1913 [The Naturhistorisches Museum Basel (NHMB): CASENT0217322 not dissected)].

Non type Malagasy material.

X. mgm01 [m: CASENT0083010]; X. mgm02 [m: CASENT0052348]; X. mgm03 [m: CASENT0218103]; X. mgm04 [m: CASENT0113147]; X. mgm05 [m: CASENT0080666]; X. mgm06 [m: CASENT0083494]; X. mgm07 [m]; X. mgm08 [m]; X. mgm09 [m].

Non type material from other than the Malagasy region (not dissected).

In addition to above, several undetermined male specimens were available from Central African Republic (X. cf01: CASENT0087255), Gabon (X. gam01: CASENT0247582), and Thailand (X. th01: CASENT0119328); X. mutica worker specimens were available from Ivory Coast (CASENT0006952), Cameroon (CASENT0009095), Central African Republic (CASENT0417121), and Gabon (CASENT0004308).

Tergosternal fusion.

In this study, tergosternal fusion is considered to occur when tergite and sternite are fused to each other in a manner different from the fusion seen in the proceeding posterior segment. The degree of fusion in males seems to be highly variable, with many intermediates. According to Bolton [29], the tergosternal fusion seen in abdominal segments III and IV was pointed out in Gotwald [30] for the first time. Bolton classified segments as “fused” when “the tergite and sternite of any given segment either meet edge to edge, or narrowly overlap and are immovably welded together.” In the Amblyoponinae the level of fusion is difficult to ascertain, especially when the segments overlap. Therefore, we use a relative comparison with abdominal segment V which is never fused in ants.

Abdominal sternum IX (Figure 7).

We prefer the term abdominal sternum IX rather than the subgenital plate and the hypopygium, the terms used for this sclerite in previous studies, because the latter two terms are ambiguous in homology; the subgenital plate and the hypopygium are not homologous in males and females of the same taxon. Even within the same sex, the subgenital plate is not consistent between taxa. (More detailed information is provided in the terminology section and table 1 of Yoshimura & Fisher [5]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Character matrix for males of Malagasy Amblyoponinae.

https://doi.org/10.1371/journal.pone.0033325.t001

Pygostyles (Figure 4).

Pygostyle is used to refer to the pair of appendages on tergum X of the abdomen of male Hymenoptera. Cerci, on the other hand, should refer to appendages on abdominal tergum XI, not X (see also table 1 in Yoshimura & Fisher [5] and Yoshimura & Fisher [3]).

Venation and cells on the wings (Figures 10, 11).

We use veins, not cells, to describe wing characters and to discuss the homology of wing characters between amblyoponine genera and genera in other subfamilies. In previous studies of Amblyoponinae [6], [8], [12] other than Brown's works [9], [17], the number of closed cells on the fore- and hindwings was used to highlight differences between taxa. The focus on cells, however, diverts attention away from careful comparative studies of venation patterns as mentioned in Brown & Nutting [31] (see also table 1 in Yoshimura & Fisher [4] and the terminology section in Yoshimura & Fisher [5]).

In the present study, the homology of each vein across taxa was determined based on comparative studies of taxa with well-developed veins, such as species in the genus Mystrium (Figures 10C, 11C). Terms used to define veins follow the recommendation of Wootton [26], Gauld and Bolton [21], and Serna et al. [27]. However, we use 1A on the forewing instead of A; 1rs-m on the hind wing instead rs-m. Our naming system for wing veins is summarized in Figures 10C (forewing) and 11C (hindwing). Cell names as indicated in Gauld and Bolton [21] are also given in Figure 10C. We provide cell names only in reference to this earlier study, and not as an endorsement for their use.

The names of wing cells in ants have been inaccurately interpreted due to homonymy. Submarginal cells 1 and 2 in Figure 10C were referred to as cubital cells in previous works for Amblyoponinae [6], [8], [12], [17]; however, cells formed between cubitus and anal, i.e. subbasal and subdiscal in Figure 10C, were also referred to as 1st and 2nd cubital cells by Huber and Sharkey [23]. Nichols defines the cubital cells as wing cells bounded anteriorly by the cubitus or one of its branches [32], and agrees with the naming system by Huber and Sharkey [23]. On the other hand, “cubital cells” in Amblyoponinae [6], [8], [12], [17] are bounded anteriorly by subcosta and pterostigma, or Rs+M and radial sector, demonstrating the inconsistent application of the term “cubital cell(s)” in ant taxonomy.

Serna et al. [27] revised wing cell names proposed in their previous paper [33]. Their new system, however, still disagrees with Gauld and Bolton's system in some areas. For example, Serna et al. [33] used the terms submarginal cell 3 and discal cell 2 to identify forewing cells instead of marginal 1+2 and subdiscal 1 respectively, as in Gauld and Bolton [21]. According to Nichols's definition [32], cells immediately distal to pterostigma and bordering the costal margin are marginal cells. The marginal cells in Gauld and Bolton [21] are identical with the Nichols's definition, but the one in Serna et al. [27] is not. Creating new homonymies in terminology is not preferred because it could lead to more confusion in future comparative studies. Therefore, marginal 1+2 and subdiscal 1 should be used instead of submarginal cell 3 and discal cell 2 in the sense of Serna et al. [33].

Remarks.

We propose that the process on the basal portion of the cuspis on the volsella (Figure 9) is a character unique to Amblyoponinae. It cannot be used as a complete diagnostic character due to the reduction of this process in some males of Prionopelta. The presence of this process has been illustrated and/or mentioned in several previous studies for Amblyopone sensu Brown (1960) (e.g. figure 24 A in Brown [9]; figure 21 in Gotwald [35]; figure 9 in Ogata [11]), and for Myopopone (diagnosis and figure 13 in Brown [9]). We confirmed the absence of this process in Proceratiinae and Ponerinae; these data and a polymorphism in Prionopelta suggest that reductions of this process in some, but not all, Prionopelta constitute a case of secondary loss.

The above character, which is difficult to confirm without dissection, is the only one unique to the subfamily Amblyoponinae. However, combinations of characters under diagnosis clearly separate the Amblyoponinae from the six other subfamilies in the Malagasy region. In the Malagasy region, Amblyoponinae differ from Cerapachyinae in lacking a bispinose abdominal sternum IX (Figure 7) and in having a moderate volsella; from Dolichoderinae in having a broader and higher attachment of the petiole to abdominal segment III and poorly differentiated basimere compared with the harpago; from Formicinae in having the petiole as above and the shorter scape of the antenna never reaching the posterior margin of the head (Figure 1); from Myrmicinae and Pseudomyrmecinae in having a slightly reduced or same-sized abdominal segment III compared with IV (as in Figure 2); from Ponerinae in having a broader and higher attachment of the petiole to abdominal segment III; and from Proceratiinae in having the dorsal margin of abdominal segment IV nearly the same length as its ventral margin in lateral view (in Proceratiinae, the dorsal margin of abdominal segment IV in lateral view is at least twice as long as its ventral margin) (Figure 2).

Amblyoponinae males have a broad, dorsal attachment of the petiole to abdominal segment III. However, this character is sometimes difficult to differentiate from the attachment in Proceratiinae, which may also appear broadly attached. The differences in the shape of abdominal segment IV mentioned above, however, will easily separate the two subfamilies.

Additional genus characters other than generic diagnoses.

In full-face view, head wider than long when eyes included, ocelli well-developed, eyes well-developed and protruding laterally, situated on middle to anterior portion of lateral margin of head (as in Figure 1A). Occipital carina absent. Anterior clypeal margin distinctly or weakly convex anteriorly (Figure 1A). Antenna stout and not extremely long, scape short, never reaching posterior margin of head in full-face view. Mandibles cross when fully closed.

Mesopleural oblique furrow very weak or absent, reaching anterior margin of mesopleuron far ventrally from posteroventral corner of pronotum when the furrow is visible (Figure 2A). With mesonotum in dorsal view, parapsidal furrows deeply impressed (Figure 3A); the axilla clearly divided. Propodeum without teeth or spines on its posterodorsal portion.

Subpetiolar process developed to various degrees, absent in several species. The distal margin of abdominal tergum VIII relatively flat, not strongly protruding on middle portion.

Body sculpture weak. Body color yellow to blackish brown.

Remarks.

A male synopsis of the genus Adetomyrma is provided for the first time based on nine morphospecies. Genus Adetomyrma is endemic to Madagascar. Males of Adetomyrma are distinguished easily from the other four Malagasy amblyoponine genera by lack of a constriction between petiole and abdominal segment III in dorsal view (Figure 3A), lack of a transverse furrow dividing pre- and posttergite of abdominal segment IV (Figure 6A), radial sector on forewing failing to reach the costal margin, and absence of 2rs-m on forewing (Figure 10A).

Other than characters uniquely observed in Adetomyrma, absence of 1rs-m on the hindwing is shared with Xymmer.

In addition to the separable characters above, two additional characters can be useful to distinguish Adetomyrma from the other Malagasy amblyoponine males: the posterior margin of abdominal sternum IX shallowly to deeply concave (Figure 7A), and a triangular expansion on the aedeagus (Figure 8A). However, several exceptions to these features were observed. A convex posterior margin of abdominal sternum IX is common among amblyoponine males in this region, while a concave posterior margin is observed in only a few males of Stigmatomma other than Adetomyrma. A basicoventral, triangular-like expansion on the aedeagus is unique to Adetomyrma, and this character easily separates Adetomyrma from the other genera. However, only one large morphospecies, Adetomyrma mg03, has extraordinarily specialized genitalia, and the expansion in this species only is vestigial. The bizarre genital features in Ad. mg03 are not similar to any other known males in Amblyoponinae. A basicoventral expansion on the aedeagus is confirmed also in Mystrium (Figure 8C); however, the expansion in the latter has two differences from the former: 1) serrate dents are present on its ventral margin; and 2) distal margin of the expansion is rounded, never triangular.

New palpal formulae in the genus Adetomyrma were found in this study. We observed three palpal formulae in Adetomyrma males, 3,3 (Figure 13A)/2,3/2,2, although only one formula 3,3 has been reported in Adetomyrma based on workers (Ward [28], original description). According to Brown's observation, the palpal formula observed in males of Amblyopone sensu Brown [9] is frequently the same as that in conspecific workers [9]. We confirmed this tendency in Mystrium and Prionopelta as well through observation. Two other types of palpal formulae may be observed in the workers of Adetomyrma with further observation, although 3,3 is the most common.

A lack of tergosternal fusion in abdominal segment IV was observed only in Adetomyrma among amblyoponine males in the Malagasy region. Interestingly, we found polymorphisms in this character within Adetomyrma. Tergosternal fusions in abdominal segments II (petiole) to IV were focused on in Gotwald [30], and redefined by Bolton [29]. Since its discussion in Bolton [29], this character has been regarded as an important diagnostic character for poneromorph subfamilies (sensu Bolton [10]). Ward proposed the lack of tergosternal fusions in abdominal segments III and IV to be a probable case of generic autapomorphy in workers of Adetomyrma [28]. In this work, Ward raised an interesting question for ant systematics in the form of a pair of hypotheses: either a tergosternal fusion in ants occurred independently in several lineages, or represents a reversal. Saux et al. inferred that Adetomyrma had undergone secondary reversal of tergosternal fusion, based on the distribution of character states on their phylogenic tree [36]. In our observations, interestingly: 1) tergosternal fusion in abdominal segment III was confirmed in all males of morphospecies examined, although larger portions of tergite and sternite overlap more than in other amblyoponine males; 2) in one of these morphospecies (Adetomyrma mg02), the associated workers lacked tergosternal fusion in abdominal segment III; and 3) tergosternal fusion in abdominal segment IV is distinct in males of five of the nine morphospecies, while the other four are difficult to judge but also seem to lack tergosternal fusion in abdominal segment IV. The fact that the fusion of abdominal segment III is inconsistent between males and workers in at least one species suggests that the lack of tergosternal fusion in Adetomyrma workers is secondary. This result is consistent with the explanation proposed by Saux et al. [36].

Mystrium Roger, 1862. (Figures 1C, 2C, 3C, 5B, 7C, 8C, 9C, 10C, 11C, 12C, 13C)

With characters of Amblyoponinae. Frontal carinae present. Anterior margin of clypeus with dent-like projections. Antenna consisting of 13 segments. Mandible with single, blunt, apical tooth (Figure 12C). Palpal formula 4,3 (one specimen each of eight morphospecies dissected: Figure 13C). Notaulus absent or indistinct in many cases, but distinct in several species. Mesepimeron with distinct posterodorsal lobe (epimeral lobe). Mesotibia with one or two spurs in most cases, rarely indistinct. Metatibia with two spurs (Figure 5B). In dorsal view, distinct constriction present between petiole and abdominal segment III (Figure 3C). Abdominal segment IV with tergosternal fusion. Pretergite of abdominal segment IV distinctly differentiated from posttergite; a deep transverse furrow divides them. Pygostyles absent.

Distal margin of abdominal sternum IX convex (Figure 7C). Separation between basimere and harpago distinct. Basal projection on cuspis well-developed (Figure 9C). The basicoventral portion of aedeagus extended basally, distal margin of extension rounded (Figure 8C). Aedeagus in lateral view, serrate denticles present on basal portion of its ventral margin.

On forewing (Figure 10C), pterostigma well-developed, radial sector fully present and reaches to costal margin, 2r-rs connected with radial sector posterior to pterostigma, 2rs-m present, cu-a position variable, located close to or far from junction between media and cubitus. On hindwing (Figure 11C), radius present, 1rs-m present, media present apical to 1rs-m.

Remarks.

The only unique character of males of Mystrium found in this study was the basicoventral portion of the aedeagus roundly and strongly extending basally (Figure 8C) (see also the genus Adetomyrma). It is the first time a unique character for male Mystrium has been provided that distinguishes it from the other amblyoponine genera, although this character is difficult to observe without dissection.

In practice, males of Mystrium are distinguished easily from Adetomyrma by having a distinct constriction between the petiole and abdominal segment III (Figure 3C), from Stigmatomma by the lack of the pygostyle, from Prionopelta by having two metatibial spurs (Figure 5B), and from Xymmer by a complete radial sector on the forewing (Figure 10C).

In addition to a unique character in the aedeagus, the presence of clypeal dent-like projections, a palpal formula of 4,3, and a distinct epimeral lobe are provided as new diagnostic characters.

Several diagnostic characters provided in previous studies require updating. First, Emery described the notaulus in Mystrium as deeply impressed [6]; however, eight species/morphospecies out of the eleven examined lack the furrow. Second, Emery described the number of mesotibial spurs in Mystrium as two [6], and Wheeler used this number in his key to separate Stigmatomma, which had a single mesotibial spur [7]. However, our observations revealed the number of mesotibial spurs is not useful as a separable character at the genus rank in Amblyoponinae, because intra-generic variation was found in Mystrium, Stigmatomma, and Xymmer (see Character 10 in Table 1). In fact, Bolton recorded the number of mesotibial spurs in Mystrium as one [10]. Therefore, this character is not useful for distinguishing genera, although it may be useful at the species level.

Prionopelta Mayr, 1866. (Figures 1D, 2D, 3D, 5A, 7D, 8D, 9D, 10D, 11D, 12D, 13D)

With characters of Amblyoponinae. Frontal carinae present. Anterior margin of clypeus with dent-like projections. Antenna consisting of 13 segments. Mandible with two sharp teeth (Figure 12D). Palpal formula 2,2, one specimen each of one species and five morphospecies dissected (Figure 13D). Notaulus distinct (Figure 3D). Mesepimeron without distinct posterodorsal lobe (epimeral lobe). Mesotibia with single spur. Metatibia with single spur (Figure 5A). In dorsal view, distinct constriction present between petiole and abdominal segment III. Abdominal segment IV with tergosternal fusion. Pretergite of abdominal segment IV distinctly differentiated from posttergite; a deep transverse furrow divides them. Pygostyles present.

Distal margin of abdominal sternum IX convex (Figure 7D). Separation between basimere and harpago distinct. Basal projection on cuspis usually well-developed, but reduced in size in several species (as in Figure 9D). Basicoventral portion of aedeagus not extraordinarily expanded (Figure 8D). Aedeagus in lateral view, serrate denticles present on basal portion of its ventral margin.

On forewing (Figure 10D), pterostigma reduced in size, radial sector absent between M+Rs and 2r-rs, radial sector reaches to costal margin, 2r-rs connected with radial sector distal to pterostigma, 2rs-m present, cu-a located far from junction between media and cubitus. On hindwing (Figure 11D), radius present in most cases, but absent in one species, 1rs-m present, media present apical to 1rs-m.

Remarks.

Males of Prionopelta are distinguished from the other Malagasy amblyoponine genera by having two teeth on the mandible (Figure 12D), a single metatibial spur (Figure 5A), the pterostigma reduced in size, and the 2r-rs connecting with the radial sector distal to the pterostigma (Figure 10D). The pterostigmal character is proposed as a new diagnostic character.

We found useful male-based generic characters of Prionopelta have been provided in three previous studies [8]–[10]. Brown included the male characters in a diagnosis of Prionopelta [9], Bolton listed several male characters in his Appendix 2 [10], and Kusnezov proposed several characters in his male key to separate Prionopelta from Stigmatomma (as Ericapelta Kusnezov) and Paraprionopelta Kusnezov [8]. From these studies, our observation confirmed 2,2 in palpal formula (Character 4 and 5 in Table 1) [8]–[10], two-toothed mandible (Character 3 in Table 1) [8], [9], presence of notaulus (Character 7 in Table 1) [9], single tibial spur (Character 11 in Table 1) [10], and wing venations on fore- and hindwings (Characters 22–30 in Table 1) [9] as useful generic characters.

Of these characters, the two-toothed mandible (Character 3) may not separate Prionopelta from several groups in another region (see discussion below).

Reduction in size of the basal projection on the cuspis is observed only in Prionopelta as an intra-generic variation, and seems to be the result of a reversal. Brown's diagnosis of Prionopelta included this reduction as a lack of a “second cusp.” We observed reductions in this projection in three males of Malagasy Prionopelta. However, we also observed the following: 1) all amblyoponine genera examined, other than Prionopelta, have a well-developed projection on the basal portion of the cuspis; 2) the distinct projection was observed even in males of Malagasy Prionopelta except in the three males mentioned above. These data suggest the occurrence of this reduction is limited to a portion of the genus Prionopelta.

Stigmatomma Roger, 1859 stat. rev. (Figures 1B, 2B, 3B, 4, 6B, 7B, 8B, 9B, 10B, 11B, 12B, 13B, 14B, 16)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 16. Mandibular transformation in Stigmatomma and Amblyopone.

16A, 16-a1-a5, Malagasy Stigmatomma (Stigmatomma mg01: CASENT0227519); 16B, 16-b1-b5, Amblyopone australis (CASENT0172268). 16A, 16-a1-a4, 16B, 16-b1-b4, worker mandible; 16-a5, 16-b5, male mandible. Series represents the transformations in workers' mandibles to SS form (16-a1-a5), with males having a single-toothed mandible, or AD form (16-b1-b5), with males having double-toothed mandible; mandibular teeth are blue and basal denticles are red.

https://doi.org/10.1371/journal.pone.0033325.g016

Stigmatomma Roger, 1859: 250 [37], original description, type-species: Stigmatomma denticulatum Roger, 1859, by subsequent designation of Bingham, 1903: 36 [38] [Lectotype examined. See also Materials and Methods.]

Arotropus Provancher, 1881: 205 [39]. syn. nov.

Fulakora Mann, 1919: 279 [40].

Lithomyrmex Clark, 1928: 30 [41]. syn. nov.

Ericapelta Kusnezov, 1955: 273 [8]. syn. nov.

Amblyopone (in part): Brown, 1960 [9]; Bolton, 1995 [42]; 2003 [10].

With characters of Amblyoponinae. Frontal carinae absent. Anterior margin of clypeus with dent-like projections. Antenna consisting of 13 segments. Mandible with single, sharp, apical tooth (Figure 12B). Palpal formula 4,3/4,2/3,2 (one specimen each of five morphospecies dissected: Figure 13B). Notaulus distinct (Figure 3B). Mesepimeron with distinct posterodorsal lobe (epimeral lobe) in most cases, but rarely indistinct. Mesotibia with one or two spurs. Metatibia with two spurs. In dorsal view, distinct constriction present between petiole and abdominal segment III (Figure 3B). Abdominal segment IV with tergosternal fusion. Pretergite of abdominal segment IV distinctly differentiated from posttergite; a deep transverse furrow divides them (Figure 6B). Pygostyles present (Figure 4).

Distal margin of abdominal sternum IX concave or convex (Figure 7B). Separation between basimere and harpago usually distinct but indistinct in one species. Basal projection on cuspis well-developed (Figure 9B). Basicoventral portion of aedeagus not extraordinarily expanded. Aedeagus in lateral view, serrate denticles present on basal portion of its ventral margin (Figure 8B).

On forewing (Figure 10B), pterostigma well-developed, radial sector fully present and reaching to costal margin, 2r-rs connected with radial sector posterior to pterostigma, 2rs-m present, cu-a located close to or far from junction between media and cubitus. On hindwing (Figure 11B), radius present or absent, 1rs-m present, media present apical to 1rs-m.

Remarks.

We resurrect Stigmatomma here as an independent genus from a synonymy with Amblyopone based on their worker mandible characters (see Discussion of morphological evolution below for detail). No unique male character has been found in Stigmatomma. However, males of Stigmatomma are distinguished easily from Adetomyrma by having a sharper mandible (Figure 12B), constrictions between the petiole and abdominal segment III (Figure 3B) and between AIII and AIV (Figure 6B), serrate denticles on basal margin of the aedeagus (Figure 8B), a complete radial sector and 2rs-m on forewing (Figure 10B), and the 1rs-m on hindwing (Figure 11B); from Mystrium by absence of the frontal carina, sharper apex of the mandible, presence of the pygostyle (Figure 4), and less expanded posteroventral portion of the aedeagus; from Prionopelta by absence of the frontal carina, more maxillary palpal segments (two in Prionopelta, three or four in Stigmatomma), two metatibial spurs (one in Prionopelta), developed pterostigma, and complete radial sector on forewing; and from Xymmer by presence of anterior clypeal conical setae, presence of the pygostyle, having complete radial sector and 2rs-m on forewing, and having 1rs-m on hindwing.

Previous studies most often include diagnostic characters for Stigmatomma and/or Amblyopone [6], [9]–[12], [17]. Many of these characters are still useful for distinguishing Stigmatomma from the other amblyoponine genera, although some taxonomic changes have been made since the information was given. Among the previous diagnostic characters, one should be updated. The number of mesotibial spurs was described as single in Emery [6], but should be updated to one or two, as Bolton mentioned [10]. Brown proposed a new subgeneric classification for Amblyopone and discussed the usefulness of wing venation as a diagnostic character of the genus [17]. Brown rejected this classification in 1960 [9] because of an exception in the wing characters observed in Amblyopone australis, and proposed a new concept for the genus Amblyopone [17]. Yet our results suggest the wing character proposed by Brown [17] is useful to separate among amblyoponine genera (see also the discussion of morphological evolution below).

Combination in Stigmatomma.

The following names are transferred from Amblyopone to Stigmatomma as comb. rev.: amblyops, armigerum, bellii, bierigi, bruni, celata, chilense, denticulatum, elongatum, emeryi, feae, impressifrons, luzonicum, minuta, normandi, oregonense, pallipes, quadratum, reclinatum, rothneyi, santschii, saundersi, silvestrii, zwaluwenburgi.

The following names are transferred from Amblyopone to Stigmatomma as comb. nov.: agostii, annae, besucheti, boltoni, caliginosum, cleae, crenatum, degeneratum, egregium, electrinum, eminia, exiguum, falcatum, ferrugineum, fulvidum, gaetulicum, gingivalis, glauerti, gnoma, gracile, groehni, heraldoi, lucidum, lurilabes, monrosi, mystriops, noonadan, octodentatum, ophthalmicum, orizabanum, papuanum, pertinax, pluto, punctulatum, rubiginoum, sakaii, smithi, trigonignathum, trilobum, wilsoni, zaojun, and testaceum [testaceum is a nomen dubium; however, we preliminarily transfer it to Stigmatomma based on its locality (Sri Lanka)].

The following species, which are all restricted to Australia, New Caledonia, New Guinea, and New Zealand, remain in Amblyopone: aberrans, australis, clarki, hackeri, leae, longidens, mercovichi, michaelseni.

Xymmer Santschi, 1914 stat. rev. (Figures 1E, 2E, 3E, 7E, 8E, 9E, 10E, 11E, 12E, 13E, 14A–C)

Stigmatomma (Xymmer) Santschi, 1914: 311 [43], original description, type species: Stigmatomma (Xymmer) muticum Santschi, 1914 [43] [lectotype examined (Figure 14A–C). See also Materials and Methods.]

Stigmatomma (Ximmer [sic]): Emery, 1919 [44].

Xymmer: Wheeler, 1922 [7], in key.

Amblyopone (Xymmer): Clark, 1934 [45].

Amblyopone (Stigmatomma) (in part): Brown, 1949 [17].

Amblyopone (in part): Brown, 1960 [9]; Bolton, 1995 [42]; 2003 [10].

With characters of Amblyoponinae. Frontal carinae absent. Anterior margin of clypeus flat, without dent-like projections. Antenna consisting of 13 segments. Mandible with single tooth, apex sharp or blunt (Figure 12E). Palpal formula 3,3 (Figure 13E)/3,2/4,3, one specimen each of six morphospecies dissected. Notaulus distinct (Figure 3E). Mesepimeron with distinct posterodorsal lobe (epimeral lobe: Figure 2E). Mesotibia without spur in most cases, with single spur in one species. Metatibia with two spurs. In dorsal view, distinct constriction present between petiole and abdominal segment III. Abdominal segment IV with tergosternal fusion. Pretergite of abdominal segment IV distinctly differentiated from posttergite; a deep transverse furrow divides them. Pygostyles absent.

Distal margin of abdominal sternum IX convex (Figure 7E). Separation between basimere and harpago distinct. Basal projection on cuspis well-developed (Figure 9E). Basicoventral portion of aedeagus not extraordinarily expanded (Figure 8E). Aedeagus in lateral view, serrate denticles present on basal portion of its ventral margin.

On forewing (Figure 10E), pterostigma well-developed, radial sector absent between M+Rs and 2r-rs, radial sector reaches to costal margin, 2r-rs connected with radial sector posterior to pterostigma, 2rs-m present, cu-a located far from junction between media and cubitus in most cases, located close to junction in one species. On hindwing (Figure 11E), Sc+R1 and radius (R1) absent, 1rs-m absent, media absent apical to 1rs-m.

Remarks.

We resurrect Xymmer here as an independent genus from a synonymy with Amblyopone based on a morphological examination of males of nine morphospecies. Males of Xymmer can be distinguished from other amblyoponine males in the Malagasy region by a lack of dent-like projections on the anterior margin of the clypeus. In addition to this unique character, Xymmer can be separated from Adetomyrma by fewer mesotibial spurs (two in Adetomyrma, 0–1 in Xymmer), absence of the pygostyles, and abdominal sternum IX convex distally; from Stigmatomma by lack of the pygostyles, the radial sector on the forewing between Rs+M and 2r-rs, the media on the hindwing apical to the cubitus, and 1rs-m on the hindwing; from Mystrium by a lack of the frontal carinae, radial sector on forewing partially absent, an absence of 1rs-m, and media apical to cubitus on hindwing; from Prionopelta by a lack of the frontal carinae, having a distinct epimeral lobe, lack of pygostyles, and lack of a vein on hindwing between radial sector and cubitus. The mesotibial spur may be useful as a separable character: the spur is absent in most males of Xymmer (only X. mgm01 has the spur on the mesotibia out of the nine morphospecies examined), while at least a single spur is present in all males of Stigmatomma.

Few morphological discussions exist regarding the taxonomic status of the name Xymmer. Xymmer was described by Santschi as a monotypic subgeneric taxon under Stigmatomma [43]. Since Santschi's original description, Xymmer was raised to genus by Wheeler in his identification key for African Amblyoponini [7]. Clark regarded Xymmer as a subgenus in Amblyopone [45] following Wheeler's suggestion [46]; however, distinguishing characters were not discussed in their treatments. Brown [9], [17] discussed separable characters for Xymmer (as a junior synonym under the subgenus Stigmatomma) for the first time since Santschi's original description. Brown regarded Stigmatomma and its related names as junior synonyms of Amblyopone at that time [9], and this treatment constituted the recent concept of the genus Amblyopone.

We decided that the male characters proposed above and associations between the worker lectotype and our materials are sufficient to warrant designating Xymmer as an independent genus from Amblyopone. Santschi's description [43] misidentifies the criterion for separating Xymmer from Amblyopone as having conical setae just present or absent, as discussed in Brown ([9]: p.165). Although Santschi proposed a lack of conical setae on the anterior clypeal margin in workers as a diagnostic character [43], this description did not provide enough information on the uniqueness of the character observed in Xymmer. Our examination of the lectotype of Stigmatomma (Xymmer) muticum found a distinct plate-like projection just dorsal to the junction between the clypeus and labrum, and the projection widens distally with a flat distal margin (Figure 14C). This projection seems to be located along a line slightly ventral to where the conical setae are arranged (Figure 14D), and could replace the function of the conical setae; therefore, it is different from a typical mid-clypeal projection in Amblyopone, which bears the conical setae on its anterior margin. This specialized projection in Xymmer should be emphasized as a unique character to differentiate Xymmer from Amblyopone and Stigmatomma. The same clypeal character observed in the lectotype of S. (X.) muticum was found in some worker specimens in the Malagasy region, and we confirmed that male specimens identified as Xymmer (e.g. CASENT0007085) were congeneric with worker specimens identified as Xymmer (e.g. CASENT0007090) (unpublished COI sequence data). The Malagasy workers have a “mutica-like” clypeal character.

Combination in Xymmer.

A single species name, Xymmer muticus (Santschi, 1914) comb. rev., is here transferred from the genus Amblyopone to Xymmer. The genus name Xymmer should be treated as masculine in accordance with Article 30.2.4 of the ICZN Code [34], because this name is an anagram for “Myrmex.”