Definitions pertain to all adult forms unless otherwise noted.

HW Head Width, maximum width of cranium in full-face view, including compound eyes if present;

HL Head Length, maximum length of head in full-face view from anterior margin of head capsule to cranial vertex;

EW Eye Width, maximum breadth of compound eye measured perpendicular to anteroposterior axis of head (male);

EL Eye Length, maximum length of compound eye measured parallel to anteroposterior axis of head (male);

SL Scape Length, maximum length of scape in medial view, excluding bulbus;

LF2 Third Antennomere Length, length of the basal flagellomere;

ML Mandible Length, maximum length of mandible from view orthogonal to lateral mandibular margin, measured from ventral mandibular articulation to mandibular apex;

MaL Mandalar Length, maximum length of mandalus, measured along proximodistal axis of mandible;

WL Weber’s Length, maximum diagonal distance measured from most anterior extent of pronotum excluding (female) or including (male) cervical shield to most posteroventral extremity of the mesosoma, including propodeal lobes if present;

PrW Pronotal width, maximum width of pronotum, measured in dorsal view;

MW Mesonotal width, maximum width of mesonotum in dorsal view, measured immediately anterior to mesocoxal foramina;

MSW Mesoscutal width, maximum width of mesoscutum in dorsal view (male);

MSL Mesoscutal length, maximum length of mesoscutum in dorsal view (male);

PTL Petiolar length, maximum length of petiole in dorsal view, not including presclerites;

PTH Petiolar height, maximum height of petiole in profile view, including sternal process and dorsal node, if distinct;

PTW Petiolar width, maximum width of petiole in dorsal view orthogonal anteroposterior axis;

PPL Postpetiolar length, maximum length of postpetiole in dorsal view, not including presclerites;

PPW Postpetiolar width, maximum width of postpetiole in dorsal view;

PPH Postpetiolar height, maximum height of postpetiole in profile view, including sternal process and dorsal node, if distinct;

TW4 Width of abdominal tergite IV, maximum width of abdominal tergite IV measured in dorsal view.

CI (HW / HL) × 100;

SI (SL / HW) × 100;

MI (ML / HW) × 100;

OI (EW / EL) × 100;

MSI (MSW / MSL) × 100;

PI (PTW / PTL) × 100;

PPI (PPW / PPL) × 100;

TI1 (PPW / TW4) × 100.

Nomenclature for sculpture and setation combines Harris (1979), Wilson (1955), and Boudinot et al. (2020). Notational conventions for palp and tibial spur formulae follow Bolton (2003). Cephalic nomenclature follows Richter et al. (2021) and Boudinot et al. (2021). Mesosomal nomenclature follows Liu et al. (2019); metasomal, Lieberman et al. (2022). Male genital nomenclature follows Boudinot (2018). Descriptive terms for larval morphology follow Wheeler and Wheeler (1986, 1976). Wing venation is described using Brown and Nutting (1949) and Ogata (1991), with interpretation of homologies in male wing venation following Boudinot (2015) in some ambiguous cases observed in Leptanilla. Any morphological terms unaddressed in these publications follow the Hymenoptera Anatomy Ontology (Yoder et al. 2010). Glossaries of external morphological terms for worker and male Leptanillinae are summarized in Figs 1–3. In instances where the homology of the terminal abdominal sternite is ambiguous, this sternite is termed a hypopygium.

I here follow Barraclough (2019) in treating a species as an evolutionarily independent population of organisms that is genetically and phenotypically distinct from other such populations (Simpson 1961). In sexually reproducing organisms, such as the Leptanillinae (so far as is known), reproductive isolation sufficient to maintain interspecific distinctiveness—in other words, the absence of genotypic and phenotypic intermediates—is an expected property of species. Mechanically incompatible genitalia are an expected corollary of reproductive isolation, and thus would indicate interspecific differentiation, but may only be asserted to be so for sibling populations that occur in sympatry and exhibit consistent phenotypic differentiation. The degree of differentiation between such species serves as a “yardstick” by which to assess whether allopatric populations diverge sufficiently in phenotype to be considered heterospecific (Tobias et al. 2010; Ward and Branstetter 2022). Scenarios that allow this calibration of phenotypic difference are fulfilled thrice among the leptanilline morphospecies for which UCEs have been successfully enriched: one instance being Leptanilla najaphalla sp. nov. and Leptanilla zhg-my05 (Sabah, Malaysia); another, Leptanilla charonea and Leptanilla cf. zaballosi (Madrid, Spain); and the last, Leptanilla zhg-bt01 and -02 (Bhutan). In all cases the two putative sympatric species are recovered as closely related terminals by phylogenomic inference (Griebenow 2020, 2021; Griebenow et al. 2022), and males of each species pair exhibit a phenotype uniformly distinguishable across all available specimens by the proportions of the genitalia. Variation among the syntopic specimen series assigned to these morphotypes is bimodal, with the exceptions to this bimodality not constituting intermediates. Thus, there is no indication that any differentiation in genital shape among these sympatric species can be considered intraspecific.

Protanilla beijingensis Man, Ran, Chen & Xu, 2017.

Protanilla concolor Xu, 2002.

Protanilla eguchii Satria, Putri & Ahda, 2023.

Protanilla flamma Baidya & Bagchi, 2020.

Protanilla furcomandibula Xu & Zhang, 2002.

Protanilla jongi Hsu et al., 2017.

Protanilla lini Terayama, 2009.

Protanilla rafflesi Taylor in Bolton, 1990b.

Protanilla schoedli Baroni Urbani & de Andrade, 2006.

Protanilla tibeta Xu, 2012.

Protanilla wardi Bharti & Akbar, 2015.

Protanilla bicolor Xu, 2002.

Protanilla gengma Xu, 2012.

Protanilla boltoni (Borowiec, Schultz, Alpert & Baňař, 2011), comb. nov.

Protanilla helenae (Borowiec, Schultz, Alpert & Baňař, 2011), comb. nov.

Protanilla taylori (Taylor in Bolton, 1990b), comb. nov.

Protanilla izanagi Terayama, 2013.

Worker diagnosis.

1. Medial mandibular surface with or without (Protanilla taylori species group) multiple rows of peg-like chaetae.
2. Medial mandibular margin with regularly spaced denticles.
3. Medial mandibular margin without teeth.
4. Ventromedial mandibular margin with or without subapical teeth.
5. Labrum with peg- or pencil-like chaetae (Fig. 21A, B).
6. Maxillary palp 4-merous.
7. Labial palp 2- or 1-merous.
8. Clypeus distinct, with epistomal sulcus present (Fig. 22A).
9. Dorsal mandibular articulation apparent in full-face view (Fig. 24B) or rarely not so (Fig. 24A) (Protanilla concolor).
10. Medial chaetae on second protarsomere (Fig. 25A).
11. Meso-metapleural suture present, strongly impressed, scrobiculate.
12. Subpetiolar process present.
13. Abdominal segment III narrowly or broadly conjoined to abdominal segment IV.
14. Length of abdominal postsclerites IV greater than that of abdominal postsclerites V–VI.
15. Somal sculpture largely absent, if present then irregularly reticulate to rugose (Protanilla boltoni (Borowiec et al., 2011), comb. nov.).

Gyne diagnosis.

As in worker, but alate or rarely ergatoid; with compound eyes and 3 ocelli. If alate then venation Ogata Type IVb. M + Cu and Rsf3 absent; Rs + M, Cuf2-3, and 1m-cu spectral or absent.

Male diagnosis.

1. Maxillary palp 4-merous.
2. Labial palp 2- to 1-merous.
3. Clypeus distinct.
4. Ocelli present, not set on tubercle.
5. Pronotum not anteroposteriorly prolonged.
6. Mesoscutum not anteroposteriorly prolonged.
7. Notauli present or absent.
8. Pterostigma present.
9. 1A in hindwing present or absent.
10. Upper metapleuron distinct from metapectal-propodeal complex.
11. Lower metapleuron indistinct from metapectal-propodeal complex.
12. Abdominal segment II petiolate.
13. Abdominal segment III petiolate or not.
14. Cupula present.
15. Volsellae present, parossiculus and lateropenite distinct.
16. Penial sclerites medially articulated.

Larval diagnosis.

Habitus pogonomyrmecoid. Cranium subelliptical in full-face view. Mandibles typhlomyrmecoid, without teeth, lateral surfaces smooth. Setae short, suberect. Ventral prothoracic process absent; larval hemolymph tap apparently absent.

Remarks.

The tribe Anomalomyrmini was erected by Taylor in Bolton (1990b) to include Anomalomyrma and Protanilla, which were both monotypic when established. Boudinot et al. (2022) merged the tribe into Leptanillini, although the Anomalomyrmini and Leptanillini sensu Bolton (1990b) are indubitably reciprocally monophyletic. All molecular phylogenetic inference (e.g., Borowiec et al. 2019; Griebenow 2020, pers. obs.) indicates the paraphyly of Protanilla relative to Anomalomyrma, with statistical support of varying strength. Anomalomyrma is therefore here synonymized with Protanilla (see “Protanilla taylori species group” for explanation of nomenclatural priority). The phylogeny of Protanilla remains debatable (pers. obs.), with morphological diagnoses formulated below for the major lineages revealed by these analyses, here treated as informal monophyletic species groups. These lineages are recovered on deeply separated internal nodes (pers. obs.). Protanilla izanagi Terayama is left unplaced to species group due to an absence of molecular data for this species and bizarrely modified mandibles which exclude it from the species groups as diagnosed here. The position of Protanilla zhg-th02, known only from a single male specimen, is unstable across different phylogenomic analyses (pers. obs.), but is always situated on a long branch. This morphospecies does not conform to the male-based diagnoses of any of the species groups here delimited for which male morphology is known and does not represent the as-yet unknown male of the Protanilla taylori species group. Based on this evidence, Protanilla zhg-th02 represents a major subclade of Protanilla for which workers remain to be discovered.

The Protanilla rafflesi species group is further divided into three species complexes, with two distinctive species left unplaced to species complex. Species boundaries in Protanilla require further inquiry, with it being possible that the clade is over-split; each species complex may respectively represent a widespread, geographically variable species. Both sexes are notably conservative in terms of morphology. Robust species delimitation, reciprocally illuminated by morphometric and molecular data, is impossible with material as scanty as is available for Protanilla, so no revisions to species-level taxonomy within this clade are made here.

Worker diagnosis.

1. Medial mandibular surface armed with peg-like chaetae.
2. Mandible straight, not bowed along anteroposterior axis of cranium.
3. Vertical dorsal lamella absent from mandible (Fig. 26A).
4. Laterodorsal longitudinal groove present.
5. Clypeal surface flattened.
6. Median clypeal ridge externally visible.
7. Outline of clypeus in full-face view campaniform to oblate-trapezoidal.
8. Pronotal breadth subequal to propodeal breadth in dorsal view.
9. Mesotibia without spurs.
10. Petiole sessile.
11. Subpetiolar process with fenestra.
12. Abdominal sternite III convex, linear, or concave in profile view.
13. Abdominal segments II–III without tergotergal or sternosternal fusion.
14. Abdominal segments III–IV narrowly or broadly conjoined.
15. Anterior margin of abdominal post-tergite IV linear to strongly emarginate in dorsal view.
16. Soma concolorous.

Gyne diagnosis. As for genus, alate or ergatoid; if ergatoid than alar sclerites present.

Male diagnosis.

1. Distal 3 maxillary palpomeres of unequal lengths (Griebenow 2020: fig. 10A).
2. Labial palp 2- or 1-merous.
3. Antennomere 3 shorter than scape.
4. Antero-admedian signum present or absent; if present, then unsculptured.
5. Notauli present or absent; if present, then unsculptured.
6. Parapsidal lines present or absent.
7. 1A present in hindwing.
8. Abdominal segment III not petiolate.
9. Length of abdominal segment IV subequal to, or less than, respective lengths of abdominal segments V–VII.
10. Cupula non-annular.

Larval diagnosis. As for genus.

Remarks. This clade shows striking morphological conservatism in the worker caste and males, with their possibly being many cryptic species. Protanilla jongi deviates from most of the clade in having broadly conjoined abdominal segments III–IV, and a ventral subapical mandibular tooth but is robustly confirmed to be nested well within the P. rafflesi species group by phylogenomic inference (pers. obs.). I therefore also place P. furcomandibula Xu & Zhang, 2002 in the P. rafflesi species group, as this species appears to be a close relative of P. jongi (Hsu et al. 2017), with the ventral subapical mandibular tooth being hypertrophied, and abdominal sternite II concave in profile view rather than linear to convex. The concavity of abdominal sternite II in profile view is homoplasious with the Protanilla taylori species group, as is the broad connection of abdominal segments III–IV.

A 4,2 palpal formula was confirmed for the worker of Protanilla lini by examination with micro-CT (Richter et al. 2021), while the palpal formula of the conspecific male was tentatively interpreted as 4,1 by Griebenow (2020). The palpal formula of the worker in the Protanilla rafflesi species group, and indeed Protanilla as a whole, has largely gone unreported, with this study being the first to confirm the palpal formula of any representative of the Protanilla taylori species group. Palpal formula across the Formicidae shows sexual monomorphism, with few exceptions (Bolton 2003; see sections on the Protanilla bicolor species group and Leptanilla thai species group below), meaning that the interpretation by Griebenow (2020) of the male labial palp in P. lini as 1-merous was in error.

Three species complexes are hereby recognized in the Protanilla rafflesi species group: the rafflesi complex (Protanilla rafflesi Taylor in Bolton, 1990b, P. schoedli, and Protanilla wardi Bharti & Akbar, 2015); the concolor (Protanilla concolor Xu, 2002; Protanilla tibeta Xu, 2012; and Protanilla eguchii Satria, Putri & Ahda, 2023); and the lini complex (P. lini, P. beijingensis, P. flamma, and P. wallacei). Each of these complexes consist of species that are extremely similar, but for which material is too scarce to query interspecific boundaries. Protanilla furcomandibula and P. jongi are presumably close relatives, but are readily distinguishable based on known specimens, and so are not consigned to a species complex. Without phylogenomic inference, it is unclear if these species complexes are reciprocally monophyletic. Protanilla wallacei sp. nov. based upon worker specimens is recovered as sister to P. lini (pers. obs.), as would be predicted based on observed worker phenotype.

A single specimen (CASENT0842639) of Protanilla beijingensis is herein reported from Khyber Pakhtunkhwa, Pakistan, in a remarkable range extension for a species heretofore known only from Beijing, China (Man et al. 2017). CASENT0842639 qualitatively differs from the type series in possessing a pair of peg-like chaetae on the labrum rather than a single median chaeta, but it is unknown whether this constitutes intra- or interspecific variation in Protanilla. This specimen is part of a series figured by Bolton (1990b: figs 1–6), for which coordinates are unavailable. Despite this, it appears that the collection was made at an elevation of 2400–2700 meters, in a cold temperate climate resembling that of the type locality.

Dias et al. (2019: 164) described the worker of Protanilla schoedli from ten specimens collected across Sri Lanka, based on “overall similarity in … general appearance” to the holotype gyne (CASENT0911228) and the implicit assumption that multiple Protanilla spp. cannot occur in sympatry. However, the putative worker P. schoedli display no more affinity to CASENT0911228 than to other members of the Protanilla rafflesi species group, with the anterior margin of the petiolar node being straight (Dias et al. 2019: 164) rather than concave in profile view, as in CASENT0911228 (Baroni Urbani and de Andrade 2006: 46). The morphology of the petiolar node is not dimorphic between worker and gyne in Protanilla. This lack of concavity excludes these worker specimens from the Protanilla rafflesi species complex to which P. schoedli belongs. The putative workers of P. schoedli (Dias et al. 2019) more closely resemble Protanilla flamma Baidya & Bagchi, 2020, but the difference in reported ranges of CI, SI, and PI between these two series supports their heterospecificity, if these morphometric differences reflect species boundaries. In this study, the putative P. schoedli (Dias et al., 2019) are regarded as an undescribed species belonging to the Protanilla lini species complex. While neither P. schoedli nor P. flamma have been sequenced, other members of their respective species complexes have (P. wardi vs. P. lini and P. wallacei), with phylogenomic inference therefrom supporting their heterospecificity (pers. obs.).

The Protanilla rafflesi species group contains some of the only Protanilla spp. for which bionomic data are available, with micro-computed tomographic studies of cephalic skeletomusculature in P. lini demonstrating the existence of “trap-jaw” capabilities in that species (Richter et al. 2021). The existence of putative trigger hairs across Protanilla (Griebenow et al. 2022: table 4) suggests that trap-jaw biology is a synapomorphy of the genus and paralleled in the Leptanillinae only by Leptanilla laventa (Griebenow, Moradmand, & Isaia in Griebenow, Isaia, & Moradmand, 2022), comb. nov.

Worker diagnosis.

1. Medial mandibular margin armed with peg-like chaetae.
2. Mandible straight, not bowed along anteroposterior axis of cranium.
3. Vertical dorsal lamella absent from mandible.
4. Laterodorsal longitudinal groove absent.
5. Clypeal surface concave.
6. Median clypeal ridge not externally visible.
7. Outline of clypeus in full-face view campaniform.
8. Breadth of pronotum subequal to propodeum in dorsal view.
9. Mesotibia with 1 spur.
10. Petiole sessile.
11. Subpetiolar process with fenestra.
12. Abdominal sternite III convex in profile view.
13. Abdominal segments II–III without tergotergal and sternosternal fusion.
14. Abdominal segments III and IV narrowly joined.
15. Anterior margin of abdominal post-tergite IV linear to slightly emarginate in dorsal view.
16. Soma bicolored, rarely concolorous.

Gyne diagnosis. As for genus, ergatoid, without alar sclerites (pers. obs.).

Male diagnosis.

1. Distal 3 maxillary palpomeres subequal in length (Griebenow 2020: fig. 10B).
2. Labial palp 2-merous.
3. Antennomere 3 longer than scape.
4. Antero-admedian signum absent.
5. Notauli present, scrobiculate.
6. Parapsidal lines absent.
7. 1A absent from hindwing.
8. Abdominal segment III petiolate.
9. Length of abdominal segment IV subequal to, or exceeding, combined length of abdominal segments V–VIII.
10. Cupula annular (Griebenow et al. in press).

Larval diagnosis. Larva unknown.

Remarks. Phenotypic differentiation between the Protanilla bicolor and Protanilla rafflesi species groups in the worker caste is comparatively slight, but the two clades are discretely distinguishable by tibial spur formula. The strong concavity of the anterior clypeal margin referred to in previous descriptive literature more correctly refers to the face of the clypeus: the anterior margin itself is in fact no more emarginate in this clade than in the Protanilla rafflesi species group. The morphology of Protanilla TH03, a male singleton attributable to this clade by molecular data (e.g., Borowiec et al. 2019), and that of male P. gengma (Griebenow et al. in press) differs from all other known males of Protanilla in multiple respects, most conspicuously in petiolation of abdominal segment III: this condition is unique among male Leptanillinae.

Workers of the Protanilla bicolor species group are unique among examined Protanilla workers in exhibiting a mesotibial spur, an apparent symplesiomorphy of this clade. Palpal formula could not be assessed in the worker caste due to a lack of fresh specimens, but given sexual monomorphism of palpal formula across the Formicidae save for the Ponerini, Typhlomyrmex (Bolton 2003), and probably the Leptanilla thai species group as well (this study), it is sound to predict a 4,2 formula.

Species boundaries in the Protanilla bicolor species group remain unclear. Specimens identified as P. gengma are known to vary in labral chaeta count according to geographical origin (Aswaj et al. 2020), but the relevance of this trait to species delimitation is unknown. Protanilla VN03 appears transitional in morphometric terms between Protanilla bicolor Xu, 2002 and P. gengma, but PTL in Protanilla VN03 falls outside the range observed in either of these species.

Worker diagnosis.

1. Medial mandibular surface without peg-like chaetae.
2. Mandible straight, not bowed along anteroposterior axis of cranium.
3. Vertical dorsal lamella absent or present (Protanilla taylori (Taylor in Bolton, 1990b)) on mandible.
4. Laterodorsal longitudinal groove present.
5. Clypeal surface concave.
6. Median clypeal ridge not externally visible.
7. Outline of clypeus in full-face view an oblate trapezoid.
8. Pronotal breadth greater than propodeal breadth in dorsal view.
9. Mesotibia without spurs.
10. Petiole subsessile.
11. Subpetiolar process with fenestra present or absent.
12. Abdominal sternite II convex in profile view.
13. Abdominal segments II-III without tergotergal and sternosternal fusion.
14. Abdominal segment III broadly joined to abdominal segment IV.
15. Anterior margin of abdominal tergite IV entire in dorsal view.
16. Soma concolorous.

Gyne diagnosis. As for worker, but, alate. Pencil-like chaetae present on mandible; two or three rows of cuticular denticles along masticatory margin.

Male diagnosis. Male unknown.

Larval diagnosis. Larva unknown.

Remarks. Anomalomyrma was established for Protanilla taylori comb. nov. by Taylor in Bolton (1990b) on account of derived mandibular morphology and the tergotergal and sternosternal fusion of abdominal segments II–III, a character state unique among the Formicidae (Bolton 1990b, Borowiec et al. 2011). While P. taylori is known only from the gyne, Borowiec et al. (2011) described Protanilla boltoni (Borowiec, Schultz, Alpert & Baňař, 2011), comb. nov. and Protanilla helenae (Borowiec, Schultz, Alpert & Baňař, 2011), comb. nov. based on worker material, and refined the diagnosis of Anomalomyrma, demonstrating that the presence of a vertical mandibular lamella was of no diagnostic utility in the Anomalomyrmini at the genus level, and predicting that the resemblance between the mandibles of Anomalomyrma and the then-undescribed Protanilla izanagi (see below) was homoplasious. This hypothesis has not yet been tested with phylogenomic inference.

Given the paraphyly of Protanilla relative to Anomalomyrma under phylogenomic inference from several differently curated datasets (pers. obs.), the latter genus is synonymized under Protanilla. These names were established in the same publication (Bolton 1990b), and the latter is here given precedence as permitted in Article 24.2 of the International Code of Zoological Nomenclature. The Protanilla taylori species group is equivalent to the former genus Anomalomyrma.

The vertical dorsal lamella in Protanilla taylori and P. izanagi has few parallels within the Formicoidea, being comparable to the morphology observed in both female and male beast ants (Camelomeciidae: Camelosphecia), which are known only from Cretaceous burmite (Boudinot et al. 2020). Among extant formicoids, the mandible of these two Protanilla spp. is most reminiscent of that observed in armadillo ants (Agroecomyrmecinae: Agroecomyrmecini: Tatuidris tatusia Brown & Kempf, 1968), which is likewise bowed, but with the masticatory margin armed with a brush of robust feathery setae (Brown and Kempf 1968: fig. 3) rather than peg-like chaetae, cuticular denticles, or both.

The feeding ecology of P. taylori and P. izanagi may therefore resemble that of the armadillo ants. Brown and Kempf (1968: 189) hypothesized that armadillo ants feed on “slippery or active arthropod prey”, with William Brown speculating that these ants were specialist predators of oligochaetes (P. S. Ward, pers. comm. 2021). Given that known ant specialists on oligochaete prey, such as Psalidomyrmex procerus Emery (Formicidae: Ponerinae: Ponerini) (Lévieux 1983; Déjean et al. 1999), have mandibles quite unlike those of armadillo ants, this seems improbable. Food court experiments to determine the diet of these ants were unsuccessful, but isotopic analysis of armadillo ant tissue suggests that the unknown prey is itself predatory (Jacquemin et al. 2014: 5).

Protanilla taylori and Protanilla id01 differ notably from the species known only from workers in the presence of two and three ranks, respectively, of produced denticles on the mandible (Bolton 1990b; this study), as opposed to the condition observed in most Protanilla; with the presence of pencil-like chaetae on the mandible, which are absent in the worker-based species. The worker and gyne caste remain unassociated in all three described species of the Protanilla taylori species group, plus Protanilla id01. Until the female castes respectively unknown from these species are discovered, we cannot determine whether observed mandibular differences are to be credited to allospecificity, or to caste dimorphism.

Protanilla izanagi Terayama.

Worker diagnosis.

1. Medial mandibular surface with peg-like chaetae.
2. Mandible bowed along anteroposterior axis of cranium (Fig. 17B).
3. Vertical dorsal lamella present on mandible (Fig. 26B).
4. Laterodorsal longitudinal groove present on mandible.
5. Clypeal surface flattened.
6. Median clypeal ridge not externally visible.
7. Outline of clypeus in full-face view an oblate trapezoid.
8. Pronotal breadth greater than propodeal breadth in dorsal view.
9. Mesotibia without spurs.
10. Petiole sessile.
11. Subpetiolar process with fenestra present.
12. Abdominal sternite II convex in profile view.
13. Abdominal segments II–III without tergotergal and sternosternal fusion.
14. Abdominal segment III narrowly joined to abdominal segment IV.
15. Anterior margin of abdominal tergite IV entire in dorsal view.
16. Soma concolorous.

Gyne diagnosis.

As for genus, alate.

Male diagnosis.

Male unknown.

Larval diagnosis.

Larva unknown.

Remarks.

Prior to formal description, this peculiar species from southern Honshu was cited by Hölldobler and Wilson (1990) and Imai et al. (2003) as Anomalomyrma (the former authors referring to it under the nomen nudum Anomalomyrma kubotai), due to the presence of an erect mandibular lamella. Borowiec et al. (2011) concluded that this character state alone was insufficient to place the morphospecies in Anomalomyrma, with its habitus being otherwise consistent with that of Protanilla. Terayama (2013) accordingly described Protanilla izanagi in that genus. The presence of distinct posterior faces on the dorsal petiolar and post-petiolar nodes, with abdominal segments III and IV not being broadly conjoined, shows an affinity to the Protanilla rafflesi and Protanilla bicolor species groups, but these character states are plesiomorphic for Protanilla (pers. obs.). It is likely that the similar mandibular morphology of P. izanagi and the Protanilla taylori species group reflects similar diet (see “Remarks” for the Protanilla taylori species group above) and is therefore homoplasious (Borowiec et al. 2011). Terayama (2013) describes the compound eye as being absent in the worker, but the specimens that I examined are remarkable in the retention of two ommatidia (Fig. 17A). The presence of any trace of the compound eye in the worker is unique among the Leptanillinae. No molecular data are available for P. izanagi, and so in the absence of compelling morphological evidence, this species must be left unplaced to species group within Protanilla. I predict, however, that molecular data will demonstrate that Protanilla izanagi belongs within the Protanilla rafflesi species group.

Leptanilla argamani (Kugler, 1987 (“1986”)), comb. nov.

Leptanilla belantan sp. nov.

Leptanilla escheri (Kutter, 1948).

Leptanilla indica (Kugler, 1987 (“1986”)), comb. nov.

Leptanilla judaica Kugler, 1987 (“1986”).

Leptanilla kunmingensis Xu & Zhang, 2002.

Leptanilla lamellata Bharti & Kumar, 2012.

Leptanilla laventa (Griebenow, Moradmand, & Isaia in Griebenow, Isaia, & Moradmand, 2022), comb. nov.

Leptanilla thai Baroni Urbani, 1977.

Leptanilla ujjalai Saroj, Mandi & Dubey, 2022.

Leptanilla anomala (Brues, 1925), comb. nov.

Leptanilla copiosa (Petersen, 1968), comb. nov.

Leptanilla havilandi Forel, 1901.

Leptanilla bethyloides sp. nov.

Leptanilla najaphalla sp. nov.

Leptanilla acherontia sp. nov.

Leptanilla africana Baroni Urbani, 1977.

Leptanilla alexandri Dlussky, 1969.

Leptanilla astylina Petersen, 1968.

Leptanilla australis Baroni Urbani, 1977.

Leptanilla besucheti Baroni Urbani, 1977.

Leptanilla bifurcata Kugler, 1987 (“1986”).

Leptanilla boltoni Baroni Urbani, 1977.

Leptanilla buddhista Baroni Urbani, 1977.

Leptanilla charonea Barandica, López, Martínez & Ortuño, 1994.

Leptanilla doderoi Emery, 1915.

Leptanilla exigua Santschi, 1908.

Leptanilla hunanensis Tang, Li & Chen, 1992.

Leptanilla islamica Baroni Urbani, 1977.

Leptanilla israelis Kugler, 1987 (“1986”).

Leptanilla japonica Baroni Urbani, 1977.

Leptanilla javana (Wheeler & Wheeler, 1930).

Leptanilla kubotai Baroni Urbani, 1977.

Leptanilla macauensis Leong, Yamane, & Guénard, 2018.

Leptanilla minuscula Santschi, 1907.

Leptanilla morimotoi Yasumatsu, 1960.

Leptanilla nana Santschi, 1915.

Leptanilla oceanica Baroni Urbani, 1977.

Leptanilla okinawensis Terayama, 2013.

Leptanilla ortunoi López, Martínez, & Barandica, 1994.

Leptanilla plutonia López, Martínez, & Barandica, 1994.

Leptanilla poggii Mei, 1995.

Leptanilla revelierii Emery, 1870.

Leptanilla swani Wheeler, 1932.

Leptanilla taiwanensis Ogata, Terayama & Masuko, 1995.

Leptanilla tanakai Baroni Urbani, 1977.

Leptanilla tanit Santschi, 1907.

Leptanilla tenuis Santschi, 1907.

Leptanilla theryi Forel, 1903.

Leptanilla vaucheri Emery, 1899.

Leptanilla yunnanensis Xu, 2002.

Leptanilla zaballosi Barandica, López, Martínez & Ortuño, 1994.

Leptanilla butteli Forel, 1913.

Leptanilla clypeata Yamane & Ito, 2001.

Leptanilla hypodracos Wong & Guénard, 2016.

Leptanilla kebunraya Yamane & Ito, 2001.

Leptanilla palauensis (Smith, 1953).

Leptanilla santschii Wheeler & Wheeler, 1930.

Worker diagnosis.

1. Medial mandibular margin without peg-like chaetae.
2. Medial mandibular margin with or without denticles, if present then irregularly spaced.
3. Medial mandibular margin with at least one subapical tooth.
4. Ventromedial mandibular margin without subapical teeth.
5. Labrum without peg-like chaetae.
6. Maxillary palp 1- to 2-merous.
7. Labial palp 1-merous.
8. Clypeus indistinct.
9. Dorsal mandibular articulation not visible in full-face view.
10. Medial chaetae absent from second protarsomere (Fig. 25B).
11. Meso-metapleural suture usually vestigial to absent, rarely present; if present then unsculptured.
12. Subpetiolar process present or absent.
13. Abdominal segment III narrowly joined to abdominal segment IV.
14. Length of abdominal postsclerites IV longer than or subequal to that of abdominal postsclerites V–VI.
15. Somal sculpture present and widespread, never punctate.

Gyne diagnosis.

Dichthadiiform, and therefore lacking wings and axillary sclerites. Mandibles edentate or with three teeth (Leptanilla kubotai) (Terayama and Kinomura 2015). Compound eyes repressed or present; if present then consisting of one or two ommatidia. Abdominal segment III never petiolate.

Male diagnosis.

1. Maxillary palp 1- to 2-merous.
2. Labial palp 1-merous.
3. Clypeus distinct or indistinct.
4. Ocelli present or absent (Leptanilla TH03, Leptanilla zhg-bt03); if present then set on tubercle or rarely not (e.g., Leptanilla najaphalla sp. nov.).
5. Pronotum anteroposteriorly prolonged.
6. Mesoscutum anteroposteriorly prolonged.
7. Notauli absent.
8. Pterostigma absent.
9. 1A absent from hindwing.
10. Upper metapleuron distinct from metapectal-propodeal complex (Leptanilla thai species group, Leptanilla bethyloides sp. nov., Leptanilla zhg-th01) or indistinct.
11. Lower metapleuron indistinct or distinct from metapectal-propodeal complex (Leptanilla havilandi species group, Leptanilla bethyloides sp. nov., Leptanilla zhg-th01).
12. Abdominal segment II petiolate or not (e.g., Leptanilla TH02).
13. Abdominal segment III not petiolate.
14. Cupula present or absent; if present, then annular.
15. Volsellae present or absent (Leptanilla havilandi species group, Leptanilla bethyloides species group), if present then parossiculus and lateropenite indistinct (Griebenow et al. in press).
16. Penial sclerites medially fused or articulated (Leptanilla astylina Petersen, 1968), rarely partly articulated (Leptanilla TH03).

Larval diagnosis.

Habitus leptanilloid. Cranium subpyriform in full-face view. Mandibles leptanilloid, with teeth, lateral surface shagreened with spinules. Setae short and suberect or flexuous, elongated, and subdecumbent to erect. Ventral prothoracic process and hemolymph taps present.

Remarks.

The four genera known solely from males at the time of Bolton (1990b) were provisionally retained in the Leptanillini by that author, with the knowledge that at least some would prove to be satellite genera of Leptanilla. The phylogeny of the Leptanillini is now robustly resolved with phylogenomic and total-evidence approaches: Leptanilla s. l. (Griebenow 2020, 2021) includes Scyphodon and Noonilla (= Scyphodon s. l.; Griebenow et al. in press), with Leptanilla s. str., with which Phaulomyrma was synonymized (Griebenow 2021); and is sister to a well-supported clade first recovered by Borowiec et al. (2019) and identified as Yavnella by Griebenow (2020, 2021).

The question of the formal rank of major subclades in the Leptanillini depends upon practical utility. For generic ranking of subclades to be useful, these clades must be distinguishable based upon the morphology of both the male sex and available female castes. Yavnella and Leptanilla s. l. are readily diagnosed based upon males, as are the subclades of Leptanilla s. l. (pers. obs.). The taxonomic problem then lies in whether these groups can be distinguished based upon worker morphology.

Using phylogenomic inference, Griebenow et al. (2022) identified the worker of Yavnella, while Leptanilla havilandi Forel is sister to Scyphodon s. l. (in those analyses represented only by Noonilla spp.) and Leptanilla thai is robustly recovered within Yavnella as well (pers. obs.). The morphological similarities between Leptanilla laventa (Griebenow et al. 2022), comb. nov. and L. thai to the exclusion of Leptanilla s. str., such as the emarginate frontoclypeal process, cannot be interpreted as synapomorphic. L. havilandi and thai are extremely close morphologically, as noted by Baroni Urbani (1977). In this study, I find that these two species are discriminated by areolate sculpturation of the torulus in L. thai (no such sculpture is observed in L. havilandi; Fig. 27), different mandibular dentition, and a more elevated frontoclypeal process in L. havilandi. Sculpturation requires scanning electron microscopy to be assessed, while elevation of the frontoclypeal process and mandibular dentition are difficult to accurately assess with light microscopy (as evidenced by the incorrect accounting of mandibular teeth in the description of L. thai (Baroni Urbani, 1977)), making these characters impractical for identification of leptanilline workers to genus. This impracticality, and lack of consistent morphological distinction between the worker castes across all Yavnella and Leptanilla, argues against maintaining the two as separate genera.

Therefore, the most conservative course of nomenclatural action is to synonymize Scyphodon, Noonilla, and Yavnella under Leptanilla. The diversity of Leptanilla is here organized in informal species groups, for which diagnoses based upon all known castes are provided below. Wherever sampling of molecular data across Leptanilla is sufficient for phylogeny of these species groups to be known, these are delimited to be monophyletic. Several aberrant species for which molecular data are unavailable are left unplaced to species group.

Worker diagnosis.

1. Mandible with 3–4 teeth.
2. Maxillary palp 1- to 2-merous.
3. Frontoclypeal process present, apex emarginate.
4. Lateral clypeal teeth absent.
5. Meso-metapleural groove absent or present (Leptanilla kunmingensis Xu & Zhang, 2002).
6. Mesotibia with two spurs.
7. Metatibia with 1–2 spurs.
8. Length of abdominal segment II subequal to width in dorsal view, or length much greater than width (Leptanilla laventa).
9. Anterior of abdominal tergite IV lateromedially constricted in dorsal view (Leptanilla laventa) or not lateromedially constricted.
10. Length of abdominal tergite IV greater than combined length of posterior abdominal tergites in dorsal view.

Gyne diagnosis. As for genus, but petiole longer than broad in dorsal view, outline rectangular (Leptanilla escheri) to subpyriform (Leptanilla belantan). Placement of these two species in the Leptanilla thai species group is provisional (see Remarks).

Male diagnosis.

1. Mandalus ≥ 0.5× length of that of the mandible.
2. Mandible fused to cranium, rarely articulated.
3. Anteromedian ocellus orthogonally dorsal to compound eye in profile view.
4. LF2 > SL, rarely LF2 ≈ SL.
5. Distal transverse carina absent from procoxa.
6. Protrochanter not elongated.
7. Profemur not enlarged, sometimes proximally kurtotic.
8. Arcuate medial carina absent from profemur.
9. Apicoventral hook absent from profemur.
10. Ventromedian carina absent from protibia.
11. Protibial comb absent.
12. Antero-admedian signum present or absent.
13. Pronotum and mesoscutum not anteroposteriorly prolonged.
14. Mesoscutellum without recurved posteroventral process.
15. Adventitious spectral M+Cu absent from forewing.
16. Upper metapleuron distinct from metapectal-propodeal complex or indistinct.
17. Lower metapleuron indistinct from metapectal-propodeal complex.
18. Propodeal declivity concave in profile view.
19. Petiole without distinct dorsal node.
20. Abdominal sternite II without ventral process.
21. Abdominal tergite VIII broader than long in posterodorsal view.
22. Abdominal sternite IX posteriorly separate from gonocoxites.
23. Mulceators absent.
24. Cupula present.
25. Gonopodites inarticulate.
26. Gonocoxites with partial ventromedian fusion.
27. Gonocoxites without or rarely with dorsomedian fusion (Leptanilla TH03).
28. Gonocoxites partly fused to penial sclerites or unfused.
29. Gonostyli present or rarely absent (Leptanilla TH03).
30. Volsellae present.
31. Volsellae medially separate.
32. Volsella furcated, sometimes entire (Leptanilla TH03, Leptanilla zhg-bt03).
33. Penial sclerites usually with complete median fusion, rarely with partial median fusion.
34. Penial sclerites dorsoventrally compressed or not (Leptanilla TH03).
35. Phallotreme apical.
36. Phallotreme dorsal.
37. Dense phallotremal vestiture of setae absent.

Larval diagnosis. As for genus. Larva is known only in Leptanilla escheri and Leptanilla judaica, the placement of which in this species group has not been confirmed by molecular phylogenetic inference.

Remarks. Leptanilla escheri, L. judaica, Leptanilla kunmingensis Xu & Zhang, 2002, Leptanilla lamellata Bharti & Kumar, 2015, L. ujjalai, and L. belantan sp. nov. are placed in this species group with some caution, given a lack of molecular data for these species. These four species bear some resemblance to Leptanilla laventa comb. nov. (e.g., in the palpal formula being 2,1), which differs from them only in the elongation of the appendicular sclerites. Since worker morphology in Leptanilla is often indecisive when inferring phylogeny, or downright misleading (pers. obs.), these species may belong elsewhere within Leptanilla. With only species included in phylogenomic analysis under consideration, the Leptanilla thai and Leptanilla havilandi species groups are mutually indistinguishable based upon worker morphology without examination of cranial microsculpture. However, male specimens of the Leptanilla havilandi species group are known only from the Sundan region, and so extralimital worker specimens that conform to the worker-based morphological diagnosis of that species group presented here are instead referred to the Leptanilla thai species group. These two clades are only definitively known in sympatry from peninsular Malaysia (Fig. 28). Since phylogenomic inference confirms the position of L. thai within the former genus Yavnella, and this is the oldest species name assigned to that clade for which that hypothesized placement can be confirmed with molecular data, this clade is informally exemplified by that species.

As noted in Griebenow et al. (2022), the anatomical identity of the frontoclypeal process observed in the Leptanilla thai species group, the Leptanilla havilandi species group, Leptanilla clypeata and Leptanilla hypodracos Wong & Guénard, 2016 is unclear. Prior authors assumed a clypeal origin, which may be in part correct, but this hypothesis cannot be tested with external examination due to the absence in worker Leptanilla of apparent anterior tentorial pits or an unequivocal epistomal sulcus. Elision of the boundaries between the frons and clypeus also occurs in Discothyrea (Proceratiinae) and Aulacopone relicta Arnol’di, 1930 (Ectatomminae: Heteroponerini), likewise involved in an anteromedian projection from the cranium in full-face view (Taylor 1979). Detailed micro-CT study of the shelf-like frontoclypeal process in the Discothyrea oculata and Discothyrea traegordhi species complexes was able to confirm the identity of this process as a mosaic of the frons and clypeus (Hita-Garcia et al. 2019), and only similar data can possibly be used to clarify the anatomy of the frontoclypeal process in Leptanilla.

The palpal formula in the worker caste of L. thai and L. laventa is 2,1 (Griebenow et al. 2022), which, among those species that have been confirmed to belong to the Leptanilla thai species group by phylogenomic inference, are the only ones for which the worker caste is known. All known males of the Leptanilla thai species group examined in this study possess a 1-merous palp (cf. Kugler 1987), meaning that it is probable that the Leptanilla thai species group shows sexual dimorphism in palpal formula. This would be only confirmed by definitive association of conspecific worker and male specimens belonging to this clade. If confirmed, the Leptanilla thai species group would constitute only the third independent origin in the Formicidae of decoupled palpal formula between the sexes (Bolton 2003). Curiously, this would run opposite to the tendency in other cases of decoupling within the Formicidae, in which the palpomere counts of the worker are reduced relative to those in the male.

The Leptanilla thai species group is broadly distributed across southern Asia (Griebenow et al. 2022: fig. 20), with males being more diverse and abundant than any other leptanilline clade in Malaise trap residues from mainland Southeast Asia. An undescribed male morphospecies is recorded from Sana’a, Yemen (Collingwood and Agosti 1996), meaning that the Leptanilla thai species group extends at least to the extreme northeastern corner of the Afrotropics, but within that ecozone is perhaps restricted to the southern Arabian Peninsula. No specimens are yet known from the Eastern Palaearctic, with the nearest examples being L. kunmingensis and an undescribed worker specimen (CASENT0064302), both from Yunnan Province, China. This absence from the Eastern Palaearctic is notable given the thorough myrmecological sampling of Japan and to a lesser extent Taiwan. Better sampling of the Sundan region is needed, but members of the Leptanilla thai species group are conspicuously rare in collections from this area compared to mainland Southeast Asia, with only two male morphospecies being known from a single locality south of the Pattani-Kangar Line (Whitmore 1988), with Leptanilla belantan, which may represent the worker of either of these. It may be surmised from the distribution of the Leptanilla thai species group that this clade originated in subtropical seasonal forests of mainland Southeast Asia or the Indian subcontinent, explosively radiating in the former region and arid habitats of the Western Palaearctic and (marginally) the Afrotropics. The Leptanilla thai species group appears to have been mostly unsuccessful in penetrating perhumid equatorial rainforests. I propose that preoccupation of ecological niche space in the Sundan region by the Leptanilla havilandi species group is perhaps responsible, given the close functional similarities between the worker phenotypes in these two clades to the exclusion of confirmed worker morphology in the Leptanilla revelierii species group.

Worker diagnosis.

1. Mandible with three teeth.
2. Maxillary palpomere 2-merous.
3. Frontoclypeal process present, apex emarginate.
4. Lateral clypeal teeth absent.
5. Meso-metapleural suture absent.
6. Mesotibia with two spurs.
7. Metatibia with two spurs.
8. Length of abdominal segment II subequal to width in dorsal view.
9. Anterior of abdominal tergite IV not lateromedially constricted in dorsal view.
10. Length of abdominal tergite IV greater than combined length of posterior abdominal tergites in dorsal view.

Gyne diagnosis. Gyne unknown.

Male diagnosis.

1. Mandalus ≥ 0.5× length of the mandible or < 0.5× length of mandible.
2. Mandible never fused to cranium, fully articulated.
3. Anteromedian ocellus orthogonally dorsal to compound eye in profile view or posterior to compound eye.
4. LF2 < SL, rarely LF2 ≈ SL (Leptanilla copiosa (Petersen, 1968)).
5. Distal transverse carina present on procoxa (Fig. 29A).
6. Protrochanter not elongated.
7. Profemur not enlarged, or moderately enlarged, sometimes proximally kurtotic.
8. Arcuate medial carina absent from profemur.
9. Apicoventral hook absent from profemur.
10. Ventromedian carina present on protibia.
11. Protibial comb absent.
12. Antero-admedian signum present or absent.
13. Pronotum and mesoscutum anteroposteriorly prolonged.
14. Mesoscutellum without recurved posteroventral process.
15. Adventitious spectral M+Cu absent from forewing.
16. Upper metapleuron indistinct from metapectal-propodeal complex.
17. Lower metapleuron usually distinct from metapectal-propodeal complex, rarely (L. anomala (Brues, 1925)) indistinct.
18. Propodeal declivity convex in profile view.
19. Petiole reduced, without distinct dorsal node.
20. Abdominal sternite II without ventral process.
21. Abdominal tergite VIII distinctly longer than broad in posterodorsal view.
22. Abdominal sternite IX completely fused to gonocoxites.
23. Mulceators absent.
24. Cupula absent.
25. Gonopodites articulate.
26. Gonocoxites with complete ventromedian fusion.
27. Gonocoxites with complete dorsomedian fusion.
28. Gonocoxites completely fused to penial sclerites.
29. Gonostyli present.
30. Volsellae absent.
31. Inapplicable .
32. Inapplicable .
33. Penial sclerites with complete median fusion.
34. Penial sclerites not dorsoventrally compressed.
35. Phallotreme preapical.
36. Phallotreme dorsal.
37. Dense phallotremal vestiture of setae present or absent.

Larval diagnosis. Larva unknown.

Remarks. This clade is restricted to the Sundan region and the Philippines (Fig. 28). Most known specimens are Bornean in origin. The bizarre males of the Leptanilla havilandi species group were first described as the genera Scyphodon and Noonilla, with Leptanilla anomala (Brues, 1925) being regarded as Hymenoptera incertae sedis (Brues 1925). Male morphospecies attributable to Noonilla in addition to the type species (L. copiosa) were identified and sequenced by Griebenow (2020, 2021). Griebenow et al. (in press) treats this clade as Scyphodon s. l., despite not yet having subjected the position of Scyphodon relative to Noonilla to phylogenetic analysis. Nonetheless, Bayesian total-evidence inference confirms the monophyly of Scyphodon s. l. inclusive of L. havilandi (pers. obs.), here formally synonymized with Leptanilla.

The worker of L. havilandi bears a striking resemblance to L. thai, including in the presence of an emarginate frontoclypeal process, but is distantly related, demonstrating the morphological conservatism of the worker caste in Leptanilla. Leptanilla clypeata and L. hypodracos are sympatric with the Leptanilla havilandi species group, and morphologically like L. havilandi, introducing the possibility that these are members of this clade. Given the lack of phylogenetic signal in leptanilline worker morphology, however, this hypothesis must be tested with molecular data.

The close affinity of L. anomala and L. copiosa, to the exclusion of other described Leptanillinae, was not suggested by previous authors who argued for the placement of L. anomala within the Leptanillinae (Petersen 1968; Boudinot 2015). This is in part due to the preservation in balsam of the type series of L. anomala, a status that conceals autapomorphies of the Leptanilla havilandi species group, namely phallotremal setae and the distal transverse carina on the procoxa: examination of CASENT0106168 revealed these character states. In addition, the discovery of additional undescribed male morphospecies within the Leptanilla havilandi species group (Griebenow 2020, 2021; Griebenow et al. 2022) revealed intermediates in morphospace, juxtaposing the dorsoventrally compressed head and mesosoma of L. anomala with the nub-like, non-spatulate mandibles of L. copiosa.

Worker diagnosis. Worker unknown.

Gyne diagnosis. Gyne unknown.

Male diagnosis.

1. Mandalus ≥ 0.5× length of the mandible.
2. Mandible never fused to cranium, fully articulated.
3. Anteromedian ocellus posterior to compound eye.
4. LF2 < SL.
5. Distal transverse carina absent from procoxa.
6. Protrochanter not elongated.
7. Profemur not enlarged.
8. Arcuate medial carina absent from profemur.
9. Apicoventral hook absent from profemur.
10. Ventromedian carina absent from protibia.
11. Protibial comb absent.
12. Antero-admedian signum absent.
13. Pronotum and mesoscutum anteroposteriorly prolonged.
14. Mesoscutellum with or without recurved process.
15. Adventitious spectral M+Cu absent from forewing, or present (Leptanilla TH01).
16. Upper metapleuron distinct from metapectal-propodeal complex or indistinct.
17. Lower metapleuron distinct from metapectal-propodeal complex or indistinct.
18. Propodeal declivity convex in profile view.
19. Petiole well-developed, with or rarely without distinct dorsal node (Leptanilla TH07).
20. Abdominal sternite II with or without ventral process.
21. Abdominal tergite VIII broader than long in posterodorsal view.
22. Abdominal sternite IX posteriorly separate from gonocoxites.
23. Mulceators absent.
24. Cupula present (Griebenow et al. in press).
25. Gonopodites articulate.
26. Gonocoxites without ventromedian fusion.
27. Gonocoxites without complete dorsomedian fusion.
28. Gonocoxites unfused to penial sclerites.
29. Gonostyli present.
30. Volsellae absent.
31. Inapplicable .
32. Inapplicable .
33. Penial sclerites with complete median fusion.
34. Penial sclerites dorsoventrally compressed.
35. Phallotreme apical.
36. Dense phallotremal vestiture of setae absent.

Larval diagnosis. Larva unknown.

Remarks. This species group is restricted to mainland Southeast Asia north of the Pattani-Kangar Line (Fig. 30), with the type locality of L. bethyloides being their northernmost known extent. Like the Leptanilla najaphalla species group, the Leptanilla bethyloides species group is known only from male specimens. These are never abundant in known collections, with it therefore appearing that this species group exhibits genuine rather than artifactual rarity; no exemplars of this clade were described in detail by Griebenow et al. (in press), meaning that the male genital skeletomusculature of the Leptanilla bethyloides species group is more poorly understood than that of any other major leptanilline clade.

Volsellae are completely absent in Leptanilla zhg-mm03 (CASENT0842829), in a homoplasy with the Leptanilla havilandi species group (Griebenow et al. in press). The total absence, as opposed to extreme reduction, of the volsellae cannot yet be definitively confirmed for any other representatives of the Leptanilla bethyloides species group due to a lack of specimens for study.

The Leptanilla bethyloides species group qualitatively possesses male morphological diversity disproportionate to the depauperation of known lineages: the condition of the metapleuron varies from completely indiscernible (Leptanilla TH07) to both the upper and lower metapleuron being completely visible (e.g., L. bethyloides). However, the lower metapleuron is never distinct from the metapectal-propodeal complex in the absence of the same distinction for the upper metapleuron, as in most of the Leptanilla havilandi species group. Other conditions unusual among Leptanilla that are sporadically observed in the Leptanilla bethyloides species group include elongated antennomeres, a posteriorly recurved mesoscutellum (both only observed in Leptanilla zhg-th01), and a dorsomedian penial carina (Leptanilla TH01).

Worker diagnosis. Worker unknown.

Gyne diagnosis. Gyne unknown.

Male diagnosis.

1. Mandalus ≥ 0.5× length of the mandible.
2. Mandible never fused to cranium, fully articulated.
3. Anteromedian ocellus posterior to compound eye.
4. LF2 < SL.
5. Distal transverse carina absent from procoxa.
6. Protrochanter not elongated.
7. Profemur enlarged, sometimes markedly constricted proximally.
8. Arcuate medial carina absent from profemur.
9. Apicoventral hook present or absent from profemur.
10. Ventromedian carina absent from protibia.
11. Protibial comb present.
12. Antero-admedian signum absent.
13. Pronotum and mesoscutum anteroposteriorly prolonged.
14. Mesoscutellum without recurved posterodorsal process.
15. Adventitious spectral M+Cu present in forewing.
16. Upper metapleuron indistinct from metapectal-propodeal complex.
17. Lower metapleuron indistinct from metapectal-propodeal complex.
18. Propodeal declivity convex in profile view, with distinct dorsal and posterior faces, dorsal face parallel to craniocaudal axis.
19. Petiole well-developed, with distinct dorsal node.
20. Abdominal sternite II with or without ventral process.
21. Abdominal tergite VIII broader than long in posterodorsal view.
22. Abdominal sternite IX with narrow posteromedian fusion to gonocoxites.
23. Mulceators present.
24. Cupula absent or present (Leptanilla zhg-id01), if present then fused anteriorly to abdominal sternite IX and posteriorly to gonocoxites (Griebenow et al. in press).
25. Gonopodites inarticulate.
26. Gonocoxites with complete dorsomedian fusion.
27. Gonocoxites with complete ventromedian fusion.
28. Gonocoxites fused to penial sclerites or unfused.
29. Gonostyli present or absent.
30. Volsellae present.
31. Volsellae medially fused.
32. Volsella never furcated, although paired, recurved cuticular processes may be present at apex.
33. Penial sclerites with complete median fusion.
34. Penial sclerites lateromedially compressed or subcircular in cross-section.
35. Phallotreme apical or subapical.
36. Phallotreme dorsal or ventral.
37. Dense phallotremal vestiture of setae absent.

Larval diagnosis. Larva unknown.

Remarks. This clade remains known only from males, necessitating the regrettable description of a species based solely upon male material (L. najaphalla) to provide the “Bornean morphospecies group” (Griebenow 2020, 2021) with an informal species group name. The males of the Leptanilla najaphalla species group are flagrantly bizarre, defined by such autapomorphies as a protibial comb composed of parallel-sided cuticular processes (previously misidentified as setae; Griebenow 2020, 2021), the complete median fusion of the volsellae at the base, and the presence of mulceators. It appears that the protibial comb is serially homologous with the probasitarsal comb, a structure synapomorphic for the Hymenoptera (Basibuyuk and Quicke 1995). While the protibial comb and mulceators are unparalleled in the Hymenoptera, the medial fusion of the volsellae is also observed in Sceliphron caementarium (Drury, 1773) (Sphecidae: Sceliphrini) (Schulmeister 2003: fig. 11C).

Micro-CT scans reveal that all 7 morphospecies sampled in Griebenow et al. (in press) (including L. najaphalla, as Leptanilla zhg-my02) show posteromedian fusion of abdominal sternite IX to the gonocoxites, an apomorphy apparently derived independently from the anatomical condition observed in the Leptanilla havilandi species group (Griebenow et al. in press). This species group is robustly supported as sister to the Leptanilla havilandi species group (Griebenow 2020, 2021; Griebenow et al. 2022), which likewise is restricted to the Sundan region. Despite this phylogenetic position, no unequivocal male morphological synapomorphies are known for the two clades, with the fusion of S9 to the gonocoxites, and medial fusion of the gonocoxites, being perhaps homoplasious between the two according, given a lack of the Remanean homology criterion of “special quality” (Griebenow et al. in press). Further Winkler and pitfall sampling in the Sundan region, particularly Borneo, will be required to collect the unknown female castes of the Leptanilla najaphalla species group. It is also possible that Leptanilla butteli Forel, 1913 and Leptanilla kebunraya Yamane & Ito, 2001, the worker morphology of which is aberrant among Leptanilla, are representatives of this clade.

Worker diagnosis.

1. Mandible with 3–4 teeth.
2. Maxillary palpomere 1-merous.
3. Frontoclypeal process absent or present, never emarginate.
4. Lateral clypeal teeth absent.
5. Meso-metapleural suture absent or present (Leptanilla hunanensis).
6. Mesotibia with 0–1 spur.
7. Metatibia with two spurs.
8. Length of abdominal segment II subequal to width in dorsal view.
9. Anterior of abdominal tergite IV not lateromedially constricted in dorsal view.
10. Length of abdominal tergite IV equal or less than combined length of posterior abdominal tergites in dorsal view.

Gyne diagnosis. As for the genus, but petiole quadrate to distinctly broader than long in dorsal view.

Male diagnosis.

1. Mandalus ≥ 0.5× length of the mandible.
2. Mandible never fused to cranium, fully articulated.
3. Anteromedian ocellus posterior to compound eye.
4. LF2 < SL.
5. Distal transverse carina absent from procoxa.
6. Protrochanter rarely elongated (Leptanilla ci01) (Fig. 31) or not elongated.
7. Profemur enlarged or not enlarged.
8. Arcuate medial carina present on profemur (Leptanilla ci01) (Fig. 31) or absent from profemur.
9. Apicoventral hook absent from profemur.
10. Ventromedian carina absent from protibia.
11. Protibial comb absent.
12. Antero-admedian signum absent.
13. Pronotum and mesoscutum anteroposteriorly prolonged.
14. Mesoscutellum without recurved posterodorsal process.
15. Adventitious spectral M+Cu absent from forewing.
16. Upper metapleuron indistinct from metapectal-propodeal complex.
17. Lower metapleuron indistinct from metapectal-propodeal complex.
18. Propodeal declivity convex in profile view.
19. Petiole well-developed, with or without distinct dorsal node.
20. Abdominal sternite II with or without ventral process.
21. Abdominal tergite VIII broader than long in posterodorsal view or rarely longer than broad in posterodorsal view (Leptanilla ci01).
22. Abdominal sternite IX posteriorly separate from gonocoxites.
23. Mulceators absent.
24. Cupula absent or present (L. astylina).
25. Gonopodites articulate, rarely inarticulate (Leptanilla exigua Santschi, 1908).
26. Gonocoxites with ventromedian fusion partial to complete (L. astylina).
27. Gonocoxites without complete dorsomedian fusion.
28. Gonocoxites unfused to penial sclerites.
29. Gonostyli present.
30. Volsellae present.
31. Volsellae medially separate.
32. Volsella entire.
33. Penial sclerites with complete median fusion.
34. Penial sclerites dorsoventrally compressed, rarely lateromedially compressed (L. astylina, Leptanilla zhg-na01).
35. Phallotreme apical or subapical.
36. Phallotreme dorsal.
37. Dense phallotremal vestiture of setae absent.

Larval diagnosis. As for genus.

Remarks. The Leptanilla revelierii species group is by far the most geographically widespread clade within the Leptanillinae and correspondingly is the most speciose. Leptanilla revelierii Emery was the first species within the Leptanillinae to be scientifically described, while Leptanilla japonica Baroni Urbani is the leptanilline species that has been subjected to the most bionomic study. This is the only leptanilline clade to have expanded its range west of the Arabian subcontinent, radiating extensively throughout the Afrotropics and the Mediterranean Basin (Fig. 32). It does not appear that this species group extends into temperate latitudes of the Western Palaearctic, but Leptanilla alexandri Dlussky, 1969 is reported from Uzbekistan (Dlussky 1969). The Leptanilla revelierii species group, with the Protanilla rafflesi species group, are the sole leptanilline clades confirmed to range into the Eastern Palaearctic and occupy fully temperate climates (Fig. 33). In addition, the Leptanilla revelierii species group is so far the only clade within the Leptanillinae known to have traversed Wallace’s Line. The apparent ease with which this clade has radiated across the Old World is striking when compared to its sister, which remains restricted to only a portion of the Indo-Malayan ecoregion.

Leptanilla swani Wheeler is the sole species of Leptanilla to be described from Australia, although the undescribed species-level diversity of Leptanilla from that continent is conspicuous, with richness highest in Queensland. Male specimens are known from as far south as the Australian Capital Territory. Leptanilla zhg-au06 is known from a single male specimen collected on Christmas Island, in what may be a human-mediated introduction. Contrary to the suggestion of Wheeler (1932) that Leptanilla are relict elements of the Australian ant fauna, the Leptanilla revelierii species group can be assumed to be recent arrivals to Australasia from the Indo-Malayan ecoregion. There is also a great undescribed diversity of the Leptanilla revelierii species group in the Afrotropics, with no fewer than nine male morphospecies purportedly being collected at the Brandberg Massif in Namibia (Robertson 2000). Malaise trapping in conjunction with syntopic soil sampling in the Afrotropics and Australasia will surely yield a large trove of new species belonging to the Leptanilla revelierii species group. Collections of the Leptanilla revelierii species group in the Indo-Malayan ecoregion remain scanty compared to sympatric members of other species groups of Leptanilla.

Leptanilla ci01 is here provisionally considered to belong to the Leptanilla revelierii species group, despite its extreme deviation from the male morphology observed in the rest of that clade, since (1) Bayesian total-evidence inference excludes this aberrant morphospecies from all other major Leptanilla clades with posterior probability greater than 0.95 (pers. obs.) and (2) no other clade of Leptanilla is known to exist in sub-Saharan Africa. Bayesian total-evidence inference likewise excludes L. astylina from all clades within the Leptanillinae besides the Leptanilla revelierii species group, with high posterior probability (pers. obs.). What were interpreted as “medially fused volsellar plates” by Petersen (1968: 581) appear in fact to be the gonocoxites, with the “large, valve-like” sclerites interpreted as the gonocoxites (Petersen 1968: 581) therefore corresponding to the gonostyli—the putative absence of gonostyli referred to by the specific epithet of L. astylina is therefore false. Even with this reinterpretation, the male genitalia in L. astylina deviate from what is observed in the rest of the Leptanilla revelierii species group, conspicuously in the presence of a cupula (Ogata et al. 1995), complete ventromedian fusion of the gonocoxites and the medial separation of the penial sclerites shown in Petersen (1968: figs 3, 4), which could not be confirmed by examination of the holotype. The medial concavity and ellipsoid outline of the gonostylus (Petersen 1968: fig. 3) is also aberrant among the Leptanilla revelierii species group, as is the lateral concealment of the gonocoxite by the gonostylus (Petersen 1968: fig. 5) and the exposure of the volsellae. Leptanilla astylina may be sister to the remainder of the Leptanilla revelierii species group.

Despite the variety and vast geographical range of the Leptanilla revelierii species group, male morphology within the clade is quite homogeneous relative to the other major subclades of Leptanilla for which males are known, particularly when compared to the species-poor Leptanilla havilandi and Leptanilla najaphalla species groups. The dramatic innovation observed across the male phenotype of Leptanilla ci01 is striking when considered in this context.

Molecular data are unavailable for these species of Leptanilla; even with the contextualization of leptanilline morphology onto a well-resolved phylogeny inferred from molecular data or jointly from those data and discretized male morphology (Griebenow 2021), these species cannot be confidently placed to the species groups delimited here, due to morphological evidence that is equivocal in phylogenetic signal or too aberrant to be of comparative use. Leptanilla clypeata Yamane & Ito, 2001 is known from both the worker and gyne; Leptanilla palauensis (M.R. Smith, 1953) from the male alone; and the remaining species only from the worker caste. Most of these morphospecies are known only from the Indo-Malayan ecoregion.

Leptanilla clypeata and L. hypodracos are very similar to one another, and closely conform to the worker-based diagnosis of the sympatric Leptanilla havilandi species group and the parapatric Leptanilla thai species group. The palpal formulae of these species would provide further evidence as to their phylogenetic position, but have not been described, and I was not able to obtain specimens for study. These species differ from the Leptanilla havilandi and thai species groups only in the emargination of the anterior petiolar margin in dorsal view. Worker morphology is quite invariable across Leptanilla, and so the phylogenetic significance of this character state cannot be extrapolated; given the relative morphological conformity of the worker caste between the phylogenetically distant L. havilandi and L. thai, even the phylogenetic affinity of L. clypeata and L. hypodracos with one another cannot be assumed without corroboration.

Leptanilla butteli resembles the Leptanilla revelierii species group overall but differs from the members of that clade in having two mandibular teeth rather than three or four, and abdominal sternite II projecting distinctly below the level of abdominal sternite III along the dorsoventral axis (Baroni Urbani 1977: fig. 25). Leptanilla kebunraya joins L. butteli in being one of the only two Leptanilla species in which the worker mandible has two teeth, but otherwise bears little apparent resemblance to L. butteli to the exclusion of other Leptanilla. L. kebunraya is unique among known Leptanilla in having anterolateral frontoclypeal projections, which invite comparison with the lateral clypeal teeth of Feroponera ferox Bolton & Fisher, 2008 (Ponerinae: Ponerini). This is of no help in inferring the function of these structures in L. kebunraya since the biology of F. ferox is largely unknown (Bolton and Fisher 2008).

Leptanilla palauensis was described as the first known male of Probolomyrmex Mayr (Proceratiinae: Probolomyrmecini), without associated workers or gynes (Smith 1953), and is still known only from the holotype. Taylor (1965) tentatively transferred the species to Leptanilla, with Petersen (1968) following this classification with some reservation, noting that William Brown and Edward O. Wilson doubted it was even an ant. Griebenow (2021) briefly mentioned L. palauensis, noting that examination of the holotype confirmed its placement within Leptanilla s. l. (Griebenow 2021: 628). This phylogenetic position is confirmed by Bayesian total-evidence inference (pers. obs.); however, the exact phylogenetic position of this morphospecies within Leptanilla remains poorly resolved, and the combination of character states observed in in L. palauensis excludes the species from all species groups of Leptanilla here delimited. The lateromedial compression of the penial sclerites, in conjunction with well-developed volsellae, perhaps implies a phylogenetic relation with the Leptanilla najaphalla species group, or with Leptanilla zhg-my08 (for which molecular data are unavailable), also incertae sedis; both these lineages are known only from Borneo. L. palauensis is a striking biogeographical outlier among the Leptanillinae, being known only from the volcanic island of Babeldaob in Palau, and therefore the only known leptanilline from Oceania (Fig. 33). All known Leptanilla gynes are flightless, limiting their dispersal capabilities, but the remote location of L. palauensis is paralleled by the presence of Leptanilla oceanica Baroni Urbani in the Ogasawara Islands (Baroni Urbani 1977).

Almost nothing is known of the biology of Leptanilla butteli, L. kebunraya, and L. hypodracos. Among Leptanilla, our biological knowledge of L. clypeata is second in comprehensiveness only to that available for L. japonica, with Ito and Yamane (2020) providing observations of live colonies, including feeding and egg-laying behavior. Billen et al. (2022) and Billen and Ito (2022) thoroughly described the exocrine glands of worker L. clypeata, with the dorsoproximal intramandibular gland discovered in this species being novel for the Formicidae.

Molecular data are unavailable for Leptanilla santschii Wheeler & Wheeler, 1930, which is known only from the male holotype. The club-like volsellae and absent gonostyli of Leptanilla santschii (Wheeler and Wheeler 1930: fig. 2D; Petersen 1968) would exclude this species from the Leptanilla revelierii species group, if the description of Wheeler and Wheeler (1930) is accurate, but with the holotype missing (Stefan Cover, pers. comm. 2020), morphological data are too limited to permit Bayesian total-evidence inference to test this hypothesis.

Most subclades of the Leptanillinae show strong morphological conservatism in the worker caste. It is consequently difficult to assess the scope of intraspecific phenotypic variation in workers, and the sparseness of collected specimens prevents algorithmic species delimitation using molecular data. Therefore, morphospecies known only from a single specimen are excluded from the following keys, even if phylogenomic data are available therefrom and no new species are described in this study based upon worker singletons. Any such species hypothesis would be weak due to lack of comparative context, and be falsifiable simply by the discovery of additional specimens (Bond et al. 2022).

Protanilla taylori comb. nov. and the undescribed Protanilla id01 are known only from the gyne, and thus excluded from this key. It does not appear that either P. taylori or Protanilla id01, which are known only from Borneo, represent the gyne of P. boltoni or P. helenae (Borowiec et al. 2011).

Protanilla schoedli Baroni Urbani & de Andrade, 2006 is known only from the gyne (Baroni Urbani and de Andrade 2006) and is excluded from the key. Dias et al. (2019) described the putative worker; however, given known morphological variation in the worker caste among described species of Protanilla, I here consider this as representing an undescribed species, related to Protanilla flamma Baidya & Bagchi, 2020.

The following keys are corrected and extended from Griebenow (2020), with updated generic assignments for undescribed morphospecies; concordances of these morphospecies identifiers with previous publications are provided in Table 1. Respective male-based keys to each of the major subclades are subsequently provided.

These include all described species for which males are known, and all undescribed male morphospecies for which molecular data are or soon will be available, except for Leptanilla ZA01 (for which only genital morphology is known), Leptanilla TH07 and Leptanilla zhg-mm14 (for which genital morphology is unknown). Based on phylogenetic inference from both molecular and morphological data (Griebenow 2021; pers. obs.), these three morphospecies belong to the Leptanilla revelierii species group, the Leptanilla bethyloides species group, and the Leptanilla thai species group, respectively. Leptanilla zhg-au04 and zhg-au06, of the Leptanilla revelierii species group, are also excluded due to lacking observations of the gonopodital apex, making it impracticable to include these morphospecies in the male-based key to that clade.