A revision of the giant Amazonian ants of the genus Dinoponera ( Hymenoptera , Formicidae )

Dinoponera Roger 1861 has been revised several times. However, species limits remain questionable due to limited collection and undescribed males. We re-evaluate the species boundaries based on workers and known males. We describe the new species Dinoponera hispida from Tucuruí, Pará, Brazil and Dinoponera snellingi from Campo Grande, Mato Grosso do Sul, Brazil and describe the male of Dinoponera longipes Emery 1901. Additionally, we report numerous range extensions with updated distribution maps and provide keys in English, Spanish and Portuguese for workers and known males of Dinoponera.

introduction Dinoponera Roger 1861 is a strictly South American genus in the subfamily Ponerinae, tribe Ponerini, commonly called tocandiras or giant Amazonian ants (Zahl 1959, Fourcassié and Oliviera 2002, Haddad et al. 2005).These ants are generally less well known than Paraponera clavata (Fabricius 1775), the bullet ant or hormiga bala (Haddad et al. 2005), yet Dinoponera workers may surpass 3 cm in total body length, making them the largest in the world.The genus has been found from montane rainforest on the eastern slope of the Andes in Perú, Ecuador and Colombia to savannah and lowland rainforest in Brazil, Guyana, south through Bolivia, Paraguay and Argentina.Roger (1861) described Dinoponera based on Ponera gigantea (Perty 1833) and the synonym Ponera grandis (Guérin-Méneville 1838).Several authors described additional species (Santschi 1921, Borgmeier 1937, Kempf 1971), subspecies (Emery 1901) and varieties (Santschi 1921, Santschi 1928), resulting in several tetranomials and taxonomic confusion.Kempf (1971) provided a revision of the genus and with additional corrections (Kempf 1975), that formed the basis for the current taxonomy of the genus (Table 1).
Species limits among Dinoponera are difficult to define (Kempf 1971).Furthermore, revisionary studies have been hampered by availability of specimens and collection bias.Previous works (Perty 1833, Guérin 1838, Roger 1861, Emery 1901, Santschi 1921, Santschi 1928, Borgmeier 1937) were based on limited numbers of specimens.Most specimens have been collected only at the fringes of the Amazon basin, or along major rivers (Fig. 13).Currently, six species are recognized in the genus Dinoponera (Table 1), as well as two questionable subspecies of Dinoponera australis based on male coloration (Borgmeier 1937).Kempf (1971) had doubts as to the status of several species largely because uncollected areas suggested that geographic variation was poorly understood.Character integration between what appeared to be distinct species was a possibility, based on the lack of collections in key areas.In particular, Kempf (1971) recognized that because of gaps in distribu-table 1.Current taxonomy of Dinoponera as suggested by Kempf (1971Kempf ( , 1975) ) and modified by this publication, including known castes of each taxa. 1 This taxa or caste was described by this publication.Female is used instead of the caste designation of gyne and worker because they are morphologically indistinguishable. 2 Validity of these taxa is questionable and could not be addressed in this study as types were unavailable.

Taxa
Described castes Dinoponera australis Emery, 1901 Male/Female D. a. bucki Borgmeier, 1937 2  Male/Female D. a. nigricolor Borgmeier, 1937 2  Male/Female Dinoponera gigantea (Perty, 1833) Male/Female Dinoponera hispida sp.n. 1  Female 1 Dinoponera longipes Emery, 1901 Male 1 /Female Dinoponera lucida Emery, 1901 Female Dinoponera mutica Emery, 1901 Female Dinoponera quadriceps Santschi, 1921 Male/Female Dinoponera snellingi sp.n. 1  Male 1 tional data, character integration was a possibility in the uncollected areas between D. longipes, D. mutica, D. quadriceps and D. gigantea where sympatry could exist.If sympatry is demonstrated between different Dinoponera species and character integration is not found, this lends support to the recognition of taxa as distinct species.Kempf (1971) suggested that a study incorporating more specimens and quantitative characters could be more effective at defining species limits.Current analyses such as DNA bar-coding or other molecular investigation could solve this dilemma.This study seeks to determine whether the current taxonomy of Dinoponera reflects the actual species richness.We re-evaluate the alpha taxonomy of Dinoponera using morphological characters (including male genitalia), and interpretation of distribution and sympatry for workers and males.

specimens and methods
We examined 345 workers and 11 males of Dinoponera, including type specimens for D. australis, D. longipes, D. lucida, D. mutica and D. quadriceps.We were unable to acquire the type specimens of Dinoponera gigantea, D. australis nigricolor and D. australis bucki.In the material examined section for each species, female workers including possible gamergates are designated with a "w", and male specimens with a "m" following the number of specimens for each.Collections cited in this study are abbreviated as follows: AMNH American Museum of Natural History CASC California Academy of Sciences Entomological Collection CUIC Cornell University Insect Collection FMNH Field Museum of Natural History FL1 First funicular segment length.Maximum length of second antennal segment, for males only.FL2 Second funicular segment length.Maximum length of third antennal segment, for males only.

HL
Head length.Midline measured from the distal edge of the clypeus at the median (not including clypeal teeth) to the occipital margin of the head at its median (not including ocelli in males).HW Head width.Maximum width of the head in full-face view including eyes.

EL
Eye length.Maximum length of the eye measured along its longer vertical axis.

EW
Eye width.Maximum width of the eye measured along its shorter horizontal axis.

OD
Ocelli diameter.Maximum diameter of the medial ocellus, for males only.

WL
Weber's length.Distance measured between the anterior margin of the pronotum to the posterior margin of the metapleural bulla in lateral view.FWL Forewing length.Maximum length of the forewing measured from the base of the costal vein to the wing apex, for males only.HWL Hindwing length.Maximum length of the hindwing measured from the base of the costal vein to the wing apex, for males only.

PL
Petiole length.The maximum measurable longitudinal distance between the anterior and posterior extensions of the petiolar node in lateral view.

PH
Petiole height.Height of the petiole measured laterally from the median of the subpetiolar process viewed laterally to the median of the dorsum of the petiole.

PW
Petiole width.Maximum width of the petiole measured in dorsal view.

GL
Gaster length.Maximum longitudinal distance from articulation with petiolar helcium to distal edge of hypopygidium (subgenital plate in males) measured in lateral view.HFL Hind femur length.Maximum length of the posterior femur measured from its basal articulation with the trochanter to its apex at the articulation with the tibia.TBL Total body length (sum of MDL, HL, WL, PL, GL).
Label data were used from all specimens to plot distributions.These data were combined with localities derived from literature (Kempf 1971, Kempf 1975, Araujo et al. 1990, Peeters et al. 1999, Monnin and Peeters 1999, Fourcassié and Oliviera 2002, Monnin et al. 2003, Mariano et al. 2004, Araújo and Rodriques 2006, Marques-Silva et al. 2006) to supplement specimen distribution data.GPS coordinates were recorded from labels, or estimated from Google Earth (http://earth.google.com/)or the Global Gazetteer Version 2.1 Directory of cities and towns in the world (http://www.fallingrain.com/world/index.html).Maps were constructed using the software package ArcGIS version 9.2 (ESRI 2007).A complete table of localities including estimated geographic coordinates is provided in supplementary file 1.
Description of the worker.Abundant setae; black integument, ranges from smooth and shiny with no microsculpturing, to finely micropunctate or scaled depending on species (Fig. 12).Head: Mandibles long and curved posteriorly in side view; seven large teeth; erect setae on dorsum.Ventral surface of head with sparse decumbent and subdecumbent setae; may have fine striations depending on species; Papal formula 4, 4; large bilobed labrum.Clypeus with two laterally projecting teeth on anterior edge, clypeus bulging medially, extending posteriorly between frontal lobes, anterior edge with row of long setae; sparse appressed setae from distal edges to medial area of clypeus.Area posterior to clypeus with varying amounts of striation.Tentorial pits apparent.Frontal lobes raised and conspicuous, with striations at posterior constriction.Antennae: geniculate, 12 segments, all with flagellate setae; scape long, extending past posterior border of head; funiculus covered in minute appressed pubescence.Gena depressed medially of eye; dense appressed setae on the antero-lateral sides of the head; covered in conflected punctulate sculpturing.Eyes large, elliptical with slight depression (ocular ring) around circumference.Frons with large pads of long flagellate pubescence (lost in older or poorly curated specimens).Median furrow running from posterior termination of clypeus, between frontal lobes to center of frons, terminates in shallow pit in most specimens.Entire head covered in long flagellate subdecumbent setae (Fig. 1A).Mesosoma: in lateral view weakly convex; covered in long subdecumbent to erect flagellate pilosity and dense pubescence; pronotal disc with slight bulges; promesonotal suture distinct, suture between mesopleuron and propodeum distinct; mesonotum fused with propodeum and episternum, separated by slight furrows; basilar sclerite large, ovaloid; propodeum with broadly rounded dorsal outline, dorsal surface gradually curves into posterior face (Fig. 2); propodeal spiracle forms nearly vertical slit; sulcus running from center of propodeum along lower edge of propodeal spiracle to posterior edge of propodeum at dorsal edge of bulla, patches of short white pubescence at curved posterior border of pronotum and basilar sclerite.Legs long, covered in long setae with short, stiff pubescence.One well-developed, antennae cleaning, comb-like spur on foreleg; one spine-like appendage and one less developed denticular comb on mesothoracic tibia; one spine and one comb-like spur on hind tibia.Posterior side of fore leg basitarsus with dense pads of golden setae; tarsal claws bidentate.Petiole: node large and tabular in lateral view, narrow attachments at base to propodeum and gaster; in dorsal view largest width less than propodeum and gaster, varies from ovate rectangular to ovate triangular in outline; covered in long subdecumbent to erect flagellate pilosity; pubescence on anterior face and ridges of subpetiolar process; subpetiolar process reduced, slightly variable between species.Gaster: typical of ponerines; covered with flagellate setae with short pubescence; small protuberance at articulation of gastric sternite III and the petiole; stridulatory file of varying size on acrotergite of gastral tergum II; posterior edges of the pygidium and hypopygidium with characteristic rows of minute spines.
Description of the male.Integument: smooth and nitid; reddish to dark brown/ black.Head: Mandibles greatly reduced, rounded, spoon shaped, lacking teeth; palps elongated, maxillary palps 4 segmented, labial palps 3 segmented; labrum reduced, rounded to truncate, emarginated distal margin in D. snellingi and D. longipes covered with setae.Clypeus large, triangular, bulging medially; anterior tentorial pits large; frontal lobes absent; antennal sockets almost touching, located at posterior apex of clypeus.Antennae: geniculate, 13-segmented, pilosity varies from fine pubescence to long setae in different species; scape shorter than second funicular segment, but shorter than 1 st , 1 st funicular segment reduced.Compound eyes large, along lateral side of head, deeply emarginated medially.Three ocelli at posterior margin of head, bulging beyond margin of head in all species except D. australis.Entire head immaculate, covered in fine pubescence and long erect setae (Fig. 3).Mesosoma: pronotum triangular, exposed narrowly dorsally anterior to scutum; scutum large, bulging antero-dorsally,  with 3 longitudinal carina; small tegula over insertion of forewing; scutellum domed, side with vertical carina, dorsal surface smooth; basilar sclerite under hind wing reduced; fused mesopleuron, separated by furrow into anepisternum and katepisternum; metanotum exposed between scutellum and propodeum, reduced; dorsal face of propodeum shorter than posterior face, rounded into posterior face; coxa large, conical (Fig. 3).Wings: covered in minute pubescence, venation as shown in Figure 5. Legs: one well-developed, antennae cleaning, pectinate spur on foreleg; one spine-like and one less developed denticular comb on mesothoracic tibia; one spine and one comblike spur on hind tibia.Posterior side of fore basitarsus with dense pads of golden setae; tarsal claws bidentate.Petiole: narrow attachments at base to propodeum and gaster; petiolar node humped dorsally, subpetiolar process anteriorly triangular.Gaster: large, cylindrical, covered in fine silvery pubescence; pygidium terminating in spine posteriorly, with short cerci; hypopygidium with long fine erect setae, tabular subgenital plate with posterior end truncated, often emarginated.Genitalia : basal ring with dorso-anterior loop structures; parameres long, rounded, with emarginated ventro-basal edge (Fig. 9); volsella articulated with basiparamere along ventral edge, lateral finger-like cuspis volsellaris, medial digitus volsellaris with distal wide toothed cusp, basal medial lobe with tooth-like structures varying with species (Fig. 10); penis valve of aedeagus roughly triangular and rounded, aedeagal apodeme curved horn-like antero-lateral arm structure arising from mid-valve ridge, terminating at interior surface of basiparamere (Fig. 11).
Description of the larvae.A basic description of the larva of D. quadriceps (cited as D. grandis mutica) is present in Mann (1916).A detailed description of the egg and all larval stages of Dinoponera gigantea are present in Wheeler and Wheeler (1985).The following generic description of Dinoponera larvae is from their work: "Profile pogonomyrmecoid (i.e., diameter greatest near the middle of abdomen, decreasing gradually toward anterior end and more rapidly toward posterior end, which is rounded; thorax more slender than abdomen and forming a neck, which is curved ventrally).Body with numerous (114-160) mammiform tubercles, each with 2-25 short simple hairs; body hairs lacking elsewhere.Cranial hairs lacking.Mandible dinoponeroid (i.e.narrowly subtriangular in anterior view; anterior portion curved posteriorly; with or without medial teeth.)" Discussion.Dinoponera's status as a genus is validated as several characters differentiate it from other genera.Size (TBL>2.3cm) is the most obvious character distinguishing Dinoponera.The only other ants with a worker caste approaching this size are Paraponera clavata (Fabricius) and the larger Pachycondyla such as P. crassinoda (Latreille 1802), P. impressa Roger 1861 and P. villosa (Fabricius 1804).Paraponera clavata is easily separated by its anvil shaped petiole with a spine on the ventral surface, highly sculptured body and deep antennal scrobes.Pachycondyla is regarded as the sister taxa to Dinoponera (Kempf 1971, Schmidt 2010).Dinoponera, in addition to their size, are distinguishable from Pachycondyla by the presence of two laterally projecting clypeal teeth (Fig. 1A) and rows of spines on the pygidium and hypopygidium.Several (n=6) specimens have been observed to have a single ocelli in the pit at the termination of the median furrow.These anomalous specimens were previously thought to be queens (Borgmeier 1937) but as it has been shown that Dinoponera lacks queens, the presence of the ocelli is hypothesized to be the result of a Mermis Dujardin 1842 nematode parasite (Kempf 1971).
Dinoponera biology.Dinoponera is one of the roughly 10 ponerine genera in which some species have secondarily lost the typical morphologically specialized queen caste for a reproductive worker known as a gamergate (Haskins and Zahl 1971, Araujo et al. 1990, Paiva and Brandão 1995, Monnin and Peeters 1998, Peixoto et al. 2008).Conflict over dominance is intense in colonies with younger workers usually joining a linear hierarchy of one to five workers depending on colony size.The gamergate, or alpha female has the highest ranking (Monnin andRatnieks 1999, Monnin et al. 2003).The alpha female mates with non-nestmate males at night at the entrance of the nest (Monnin andPeeters 1998, Monnin andPeeters 1999).After copulation the female bites through the male's gaster to release herself and pulls out the genital capsule which acts as a temporary sperm plug (Monnin and Peeters 1998).After mating the female is unreceptive to other males and remains monandrous (Monnin and Peeters 1998).The gamergate maintains dominance with ritualized behaviors such as antennal boxing and biting, 'blocking', as well as gaster rubbing and curling (Monnin and Peeters 1999).Lipid stores within Dinoponera australis females were found to be strongly related to foraging activity and reproductive status within the colony, ranging from 1-39% of an individual's dry mass (Smith et al. 2011).It is uncertain, however, whether nutritional differences between females is a cause or consequence of rank.Gamergate females possess a higher concentration of a cuticular hydrocarbon (9-hentriacontene, 9-C 31 :1) that indicates rank and is passed onto gamergate-laid egg cuticles (Monnin and Peeters 1997, Monnin et al. 1998, Peeters et al. 1999).Additionally, alpha females may 'sting smear' a competing female with secretions from the Dufour's gland, triggering the lower ranking workers to immobilize the marked female (Monnin and Ratnieks 2001).Subordinate females (beta, gamma, or delta) may produce unfertilized eggs but these are usually consumed by the alpha female in a form of "queen policing" (Monnin and Peeters 1997).Egg recognition in D. quadriceps was found to be due to differences in cuticular hydrocarbons, and only workers engaged in brood care could distinguish non-nestmate eggs (Tannure- Nascimento et al. 2009).Cuticular hydrocarbons are also used to distinguish adult nestmates from non-nestmates, however, this is only effective with non-nestmate foragers (Nascimento et al. 2012).Nascimento et al. (2012) found that brood-caring workers from different colonies had very similar hydrocarbon profiles and were more often accepted into alien colonies.
Males are born throughout most of the year in tropical species (Araujo andJaisson 1994, Monnin andPeeters 1998), however D. australis which lives in the more temperate south was found to only produce males in May-July (Paiva and Brandão 1995).When the alpha declines reproductively or dies, she is replaced by a high-ranking worker (Monnin and Peeters 1999).
New colonies are founded by fission, a process in which a beta female is fertilized in the natal nest (Monnin and Peeters 1998).This new alpha female then leaves the nest with a cohort of workers to found an incipient colony, sometimes employing tandem running (Overal 1980).
The nest consists of large chambers and tunnels in the soil possibly with an earthen mound and can be 0.10-1.2mdeep (Araujo et al. 1990, Morgan 1993, Fourcassié and Oliviera 2002, Vasconcellos et al. 2004).Nests are deeper in D. australis and D. quadriceps than in D. gigantea, Monnin et al. (2003) suggests that deeper nests are a possible adaptation to seasons and aridity.Dinoponera gigantea nests may have up to eight entrances and can be weakly polydomous (Fourcassié and Oliviera 2002), whereas 1-30 openings with an average of 11 were recorded for D. longipes (Morgan 1993).Nesting density and spatial distribution varies depending on habitat (Fowler 1985, Vasconcellos et al. 2004).Density ranges from 15-40 nests per ha -1 (Vasconcellos et al. 2004) to 80 nests per ha -1 (Paiva and Brandão 1995).Morgan (1993) measured a spacing between nests for D. longipes with a median of 35m (n=22, range 14-69.5m).Dinoponera australis and D. gigantea usually nest at the base of trees (Paiva andBrandão 1995, Fourcassié andOliviera 2002).Observations of D. quadriceps nests show that in more arid Caatinga and Cerrado habitats, nests are predominantly constructed under trees, whereas in Atlantic forest 60% of nests were 3m away from any tree (Vasconcellos et al. 2004).
Workers lower in the hierarchy forage individually for food items on the substrate and do not recruit other nestmates to assist with food transport (Fowler 1985, Fourcassié et al. 1999, Fourcassié and Oliviera 2002, Araújo and Rodrigues 2006).Although foraging workers do not recruit nestmates, Nascimento et al. (2012) found a positive feedback between incoming food and stimulation of new foragers as well as task partitioning once food was brought into the nest.Lower ranking females processed protein resources while higher ranking females handled small food pieces and distributed them to the larvae.Fourcassié and Oliviera (2002) found D. gigantea foraging to be concentrated in the early morning and afternoon but did not sample at night.Morgan (1993) observed the highest activity at night in D. longipes.Dinoponera quadriceps has a marked seasonal pattern in activity.It is most active in May-August , the late rainy season to early dry season in the semiarid Caatinga (Medeiros et al. 2012).Activity was strongly negatively correlated to temperature and positively correlated to prey abundance (Medeiros et al. 2012).The diets of both D. gigantea and D. quadriceps have been shown to be predominantly scavenged invertebrates, but include live prey, seeds and fruits (Zahl 1959, Fourcassié and Oliviera 2002, Araújo and Rodrigues 2006).Araújo and Rodrigues (2006) state that the taxonomic diversity of prey is comparable to other tropical ponerines, but has an optimal prey size of 2-3 cm in Dinoponera.Diet seems to be very similar across the genus, regardless of habitat (Araújo and Rodrigues 2006).
Despite their large size and strong venom, Dinoponera are likely preyed on by many vertebrate and invertebrate species across South America.Like many other ant species, Dinoponera can be infected by the entomopathogenic fungi Codyceps sp.(Evans 1982).Buys et al. (2010) discovered a Kapala sp.eucharitid wasp emerging from the puparia of Dinoponera lucida.
Anatomy has been described several times.Marques-Silva et al. (2006) studied of the sensilla and glands of the antennae.Anatomy of the venom apparatus and mandibular glands of Dinoponera gigantea is presented in Hermann et al. (1984).Further studies of the mandibular glands and its contents were presented by Oldham and Morgan (1993) and Oldham et al. (1994).Oldham et al. (1994) found that the mandibular gland secretions of workers differed greatly from those of gamergates, which were 98% dimethylalkylpyrazine and lacked the four other pyrazines and 50 times more volatiles than those found in worker secretions.The post-pharyngeal gland morphology was examined by Schoeters and Billen (1997).The cuticular hydrocarbons used in nestmate recognition may be produced by epidermal glands which Serrão et al. (2009) found in the epidermis of abdominal sternites in D. lucida.
For subduing large live prey and defense (Morgan, 1993), workers possess a sting that has been known to cause severe pain lasting up to 48 hours, lymphaedenopathy, edema, tachycardia and fresh blood to appear in human victim feces are common symptoms (Haddad et al. 2005).In gamergates the venom sac is empty (Monnin et al. 2002).Workers may have 60-75 unique proteinaceous components in the venom (Morgan et al. 2003, Johnson et al. 2010).The convoluted gland within the venom system of Dinoponera australis has been found to possess close similarities to those of vespine wasps (Schoeters and Billen 1995).The contents of D. australis venom have been found to be similar to those of Pachycondyla spp.(Cruz López 1994, Johnson et al. 2010).Billen et al. (1995) studied the morphology and ultrastructure of the pygidial gland of D. australis.Due to the high diversity of compounds and systemic effects found by Haddad et al. (2005), venom of Dinoponera could be of use to the pharmaceutical industry.For instance, Sousa et al. (2012) demonstrated in mice that venom from D. quadriceps had antinociceptive properties.The authors note that the local population of northeast Brazil uses dry crushed D. quadriceps ants to treat earaches, and the stings of live ants are administered for back pain and rheumatism.
Several studies of the cytogenetics of Dinoponera species have been conducted.Dinoponera lucida may have the highest number of chromosomes within the Hymenoptera however the karyotype is variable between populations (2n=106-120) (Mariano et al. 2004, Mariano et al. 2008, Barros et al. 2009).Mariano et al. (2008) interpreted the karyotype differences between populations as being due to a division of the species into allopatric populations during the Quaternary.Variability in the karyotype within a described species has been found in the Pachycondyla as well, and may represent cryp-tic species (Mariano et al. 2012).Descriptions of the banding patterns on Dinoponera chromosomes are provided by Barros et al. (2009) andde Aguiar et al. (2011).
Dinoponera belongs to the tribe Ponerini in the subfamily Ponerinae.The evolutionary position of the genus within Ponerinae was resolved by Schmidt (2010).Based on the phylogenetic analysis of Schmidt (2010) and karyotype analysis by Mariano et al. (2012), Dinoponera's closest living relatives are in the Pachycondyla species group consisting of P. crassinoda, P. harpax (Fabricius 1804), P. impressa, P. metanotalis Luederwaldt 1918, andP. striata Smith 1858.Prior to the generation of well supported phylogenies other associations had been proposed.Carpenter (1930) suggested that the fossil Archiponera wheeleri Carpenter from the Miocene Florissant shale of Colorado may be an ancestor of Dinoponera and Streblognathus aethiopicus Smith 1858.Molecular data has shown that Carpenter's (1930) hypothesis is false (Schmidt 2010).Streblognathus is not closely related to Dinoponera, and its morphological similarity is purely convergence.The placement of Archiponera wheeleri is still questionable.Clave para la identificación de las obreras de Dinoponera 1 Esquina antero-inferior del pronoto con proceso en forma de diente distinto (Fig. 1D Pigidio terminando en un espina, alargada y delgada (Fig. 4M); válvula peneal del aedeago en forma de cuña (Visto en perfil) (Fig. 11E) .....longipes Emery -Pigidio terminando en una espina corta y ancha, en forma de triángulo (Fig. 4N); válvula peneal del aedeago con una reborde distal y un lóbulo triangular ventral (Fig. 11B Pigidio acabado numa espinha, alongada e delgada (Fig. 4M); válvula penal do aedeago em forma de cunha (visto em perfil) (Fig. 11 E) ........longipes Emery -Pigidio acabado numa espinha apara e alarga, em forma de triângulo (Fig. 4N); válvula penal do aedeago com um reborde distal e um lóbulo triangular ventral (Fig. 11B Worker diagnosis.This species is most easily recognized by the antero-inferior corner of pronotum having a distinct tooth-like process (Fig. 1D), the pilosity being short and relatively sparse and the integument being finely micro-sculptured and dull (Fig 12B).In addition the scape length is shorter than the head width and the total body length is under 30 mm.Dinoponera lucida could be confused with D. australis in that it also has a tooth-like process on the pronotum and can have a TBL under 30 mm, but differs in having the smooth and shiny integument, long flagellate hairs on lobe and forward slanting dorsal edge of petiole.

Key to the workers of
Description of the worker.Measurements (mm) (n=21) TBL: 23.42-29.31(26.21);MDL: 3.59 -4.31 (3.88); HL: 4.51-5.64(4.99); HW: 4.31-5.74(4.89); SL: 4.31-5.02(4.73); WL: 6.25-7.69(7.12); PL: 1.79-2.26(2.03); PH: 2.56-3.28(2.90); PW: 1.59-1.95(1.75); GL: 7.28-9.64(8.20); HFL: 5.54-6.66(6.16).A description of the external morphology of the worker is given by Kempf (1971): "Antennal scape length equal to, or shorter than head width.Pubescence on front of head short and inconspicuous.Gular face of head subopaque, finely reticulate-punctate throughout; the fine, arcuate striae variably developed from completely covering the undersurface of head to only vestigially shown antero-laterally or nearly absent.Sides of head reticulate-punctate, subopaque.Antero-inferior corner of pronotum dentate.Pronotal disc superficially reticulate and quite shining; paired swellings either feeble or distinct.Length of hind tibia equal to or less than head length.Petiole, in dorsal view, subquadrate, width over length proportion always more than 0.80, notably shorter and broader than in the other species; its shape…resembling that of mutica, with the upper anterior and posterior corners equally rounded; finely reticulate, somewhat shining; vertical sulcus on posterior surface either absent of more rarely vestigial to feebly developed.Terga I and II of gaster either reticulate-punctate or more superficially reticulate (in the southern range of the territory) and accordingly either subopaque or somewhat shining: fine appressed pubescence lacking completely on disc of the terga, present on the sides.Stridulatory file on acrotergite of tergum II of gaster well developed, broad and triangular, extending back to the acrotergite for about one half to two thirds of its length." Male diagnosis.Dinoponera australis males are recognized by their rounded head, with compound eyes, reduced ocelli and the posterior margin around the ocelli not protruding as in other species (Fig. 4E).This species is also characterized by the short, broad pygidial spine (Fig. 4O), volsella with tear-drop shaped basal lobe covered in minute teeth (Fig. 10D) and aedeagus with a latero-apical fold, notches and teeth along ventral edge as shown in Fig. 11D.
Description of the male.A description of the external morphology of the male is given in Kempf (1971): "Head…with smaller eyes, the maximum interocular width being greater than their diameter; with smaller ocelli not protruding above the posterior border of head when seen in full-face view; antennal scape very short, less than twice as long as broad; funiculi without standing hairs; petiole distinctly shorter although variable in outline…; pygidium with a very short spine, not projecting beyond the long cerci; hypopygidium apically broadly truncate, the truncation either straight, or convex, or concave."Description of the male genitalia.Basal ring with thick dorso-anterior loop structures, reduced; parameres short, broad, rounded, small lobe on dorsal edge, emarginated ventro-basal edge (Fig. 9D); cuspis volsellaris with few bumps or teeth, digitus volsellaris with numerous small circular bumps at distal lateral face, tuft of setae on ventro-distal side of broad cusp, large tear-shaped lobe on basal ventral corner, covered in minute teeth (Fig. 10D); penis valve of aedeagus with lateral arm of apodeme at anterior border, no ventral concavity under ridge at base of apodeme, dorsal edge rounded, sloping posteriorly, ventro-anterior triangular projection followed by circular notch, ventral projecting tooth, smaller hemispherical notch with sclerotized border, thin, finely serrated distal edge, noticeable lateral apical fold with slight serration ending ventrally in serrated ridge, rounded un-serrated lobe at distal apex of valve.Distribution.Dinoponera australis has the widest known range of the Dinoponera.This species is found in the department of Santa Cruz in Bolivia, southern Brazil in the states of Mato Grosso, Goias, Minas Gerais, São Paulo, Mato Grosso do Sul, Paraná, Santa Catarina and Rio Grande do Sul, eastern Paraguay in the departments of Itapúa, Alto Paraná and Guairá, as well as the province of Misiones in Argentina (Fig. 13).
Discussion.Dinoponera australis is the most aberrant of the Dinoponera species because of its relatively small size, sparse non-flagellate pubescence, as well as the male characters stated above which distinguish this species.The male coloration difference is the basis for the designation of the subspecies D. australis bucki and D. a. nigricolor.These may be separate species or the product of intra-specific variation, but this cannot be diagnosed here as the types designated by Borgmeier (1937) were not available to us and the extent of intra-specific variation could not be determined from the limited sample size available.
Material Worker diagnosis.Dinoponera gigantea can be distinguished from other Dinoponera species by the following combination of character states: antero-inferior corner of pronotum with distinct tooth-like process (Fig. 1D); integument finely micro-sculptured and not shiny (Fig. 12B); drab pilosity notably dense, long and flagellate; scape length longer than head width; total body length over 30 mm.Dinoponera gigantea is the largest species in the genus reaching up to 3.6 cm total body length.Dinoponera gigantea can be separated from all but two species by the presence of a tooth-like process on the antero-inferior corner of the pronotum.Dinoponera lucida and D. australis have a tooth-like process on the pronotum, but are smaller (usually less than 30 mm).In addition D. lucida has a shiny integument and D. australis lacks the long, flagellate pubescence.Description of the worker.Measurements (mm) (n=15) TBL: 31.62-36.02(34.34);MDL: 4.59-5.35(4.92); HL: 5.89-6.65 (6.31); HW: 5.74-6.27(6.00); SL: 5.95-6.43(6.30); WL: 8.71-9.94(9.35); PL: 2.72-3.06(2.81); PH: 3.08-3.67(3.59); PW: 1.85-2.07(1.98); GL: 9.43-12.24(10.94);HFL: 8.10-9.3(8.74).A description of the external morphology of the worker is given in Kempf (1971): "Length of scape exceeding maximum width of head.Pubescence on front of head quite dense yet inconspicuous, not concealing the integument.Gular (ventral) surface of head reticulate-punctate throughout, very finely striate in front, the striae strongly converging mesad toward the anterior border.Sides of head reticulate-punctate, subopaque.Antero-inferior corner of pronotum dentate.Pronotal disc reticulate-punctate, subopaque, the paired swellings rather inconspicuous, but the median impression between swellings distinct, integument irregularly wrinkled.Tarsus I of hind leg longer than maximum head length.Petiole reticulate-punctate and subopaque, rectangular in profile, the anterior surface straight to slightly concave; the anterior upper corner more narrowly, the posterior corner more broadly rounded; posterior surface with the vertical sulcus always distinct; in dorsal view the petiole is decidedly longer than broad, width-length proportion below 0.80.Terga I and II of gaster opaque, sharply reticulate-punctate, densely foveolate (from each foveola arises a short decumbent hair), with scattered, bristle bearing, larger pints.The appressed pubescence, although inconspicuous, is densely spread over the entire terga, stridulatory file on acrotergite (portion of tergum that is normally concealed under the overlapping preceding tergum) of tergum II short, narrow, inconspicuous, not crossing beyond anterior half of acrotergite (hence not easily seen if entire acrotergite is not exposed)." Male diagnosis.Males of this species are easily recognized by their funiculus which is covered in long standing setae (Fig. 4F), shiny dark reddish brown integument and the combination of a long pygidial spine (Fig. 4K), volsella with two basal teeth, lobed end of digitus volsellaris (Fig. 10C) and deep concavity on the side of the penis valve of the aedeagus (Fig. 11C).
Description of the male.A description of the external morphology of the male is given in Kempf (1971).Measurements done by Kempf (1971) are included as only one male D. gigantea was examined by us while the measurements of Kempf (1971)

Standing hairs long and abundant on body, lacking on mid-dorsum of terga II-V of gaster; long hairs on scapes rather numerous, length not much longer than twice the diameter of scape…Anterior border of labrum rounded, not visibly excised… Pygidial spine long and well developed. Parameres (gonostyli) of genitalia in side-view narrow and spear-pointed… Hypopygidium (subgenital plate of subandrium) apically rounded…"
Description of male genitalia.Basal ring with wide dorso-anterior loop structures, dorsal depression on basal ring posterior to cleft between dorso-anterior loops, ridge extending from anterior of depression to center; parameres long, narrow, rounded spade-shape, emarginated ventro-basal edge (Fig. 9C); volsella with finger-like cuspis volsellaris and broad cusp-like digitus volsellaris, cuspis volsellaris with pointed end, flange extending on dorsal edge, digitus volsellaris with numerous small circular bumps, lobe on postero-dorsal edge, 2 teeth on medial ventro-basal corner of volsella, posterior tooth with lobe on posterior edge; penis valve of aedeagus with lateral arm of apodeme at anterior border, deep, wide, ventral concavity under ridge at base of apodeme, distal edge of valve wedge-shaped, proximal ventral edge of valve ending in tooth descending at roughly 45°, ventral edge with large laterally curved lip, serrated edge (Fig. 11C), serrations facing laterally on either side of aedeagus in dorsal view (similar to penis valve in Fig. 8A).
Distribution.Dinoponera gigantea has been found on the coast of Guyana, in the Brazilian states of Amazonas, Pará including Marajo Island, Mato Grosso and Maranhão as well as the Loreto Province in Perú.Dinoponera gigantea is reported to be common in un-flooded forests in the vicinity of Belém, Pará (Kempf 1971, Overal 1980) (Fig. 13).It is probable that D. gigantea is found in French Guyana, Surinam, Venezuela and southeastern Colombia because these regions are adjacent to known D. gigantea localities and have similar lowland rainforest habitat.However, no specimens from these nations are known to us, perhaps as a result of a lack of sampling or the range is absent from Colombia and southwestern Venezuela.
A record from Rio de Janeiro (from the CASC) is puzzling as it is disjunct from the known range of D. gigantea.The most southeastern locality for D. gigantea is over 1,480 km to the nearest portion of the state of Rio de Janeiro.In addition, Rio de Janeiro is in a well collected area where no other Dinoponera have been found.The label reads 'R.d.Janeiro, Brazil, D. Davis' and the specimen agrees in all morphological characters with D. gigantea.This locality is omitted from the species' range map (Fig. 13) because of its questionable nature.If other collections can validate this locality it would mean a significant range extension for D. gigantea.
Discussion.Dinoponera gigantea is a valid species with a distinct suite of morphological characters listed in the diagnosis above.The validity of D. gigantea is strengthened by range overlaps with D. longipes and actual sympatry with D. hispida, both with no integration of morphological characters.
Description of the worker.Measurements (mm) (n=5) ; EL: 0.72-0.97(0.84); WL: 7. 89-8.71 (8.36); PL: 2.5-2.56 (2.52); PH: 2.87-3.18(3.05); PW: 1. .(See Fig. 2) Entire body with short, thick, stiff, subdecumbent to erect setae (Fig. 1C); integument black, smooth, shiny, appearing polished with bluish luster (Fig. 12A).Head: mandibles long, linear, 7-toothed, large diastema between basal tooth and six apical teeth; clypeus with two laterally projecting teeth on anterior edge, bulges medially, extends posteriorly between frontal lobes, sparse appressed setae from distal edges to disc of clypeus, short stiff setae on anterior edge; large bilobed labrum; ventral surface of head with varying amounts of fine striation, erect bristle-like setae; gena with fine striations running from eye into clypeus; median furrow running from termination of clypeus between frontal lobes to shallow pit in middle of frons (with ocelli in two individuals of type series); frontal lobes raised and conspicuous, with striations at posterior constriction; antennae all with erect bristlelike setae, funiculus covered in minute appressed pubescence; scape long, extending past posterior border of the head, covered in erect bristle-like setae; frons with sparse pads of short appressed setae; entire head covered with erect to subdecumbent bristlelike setae.Mesosoma: antero-inferior corner of pronotum rounded, without toothlike process; pronotal disc with slight bulges; mesonotum fused with propodeum and episternum, separated by slight furrows; basilar sclerite large, ovaloid; propodeum with broadly rounded dorsal outline; propodeal spiracle nearly vertical slit; sulcus running from center of propodeum along lower edge of propodeal spiracle to posterior edge of propodeum at dorsal edge of bulla; mesosoma and coxae with white pubescence, especially dense on basilar sclerite, appressed white pubescence along dorsal surface facing medially, middle posterior dorsum of pronotum lacking appressed pubescence found on mesonotum and propodeum.Legs: long, femur and tibia with sparse erect bristlelike setae.One well-developed, antennae cleaning, comb-like spur on fore leg; spinelike and less developed denticular comb on mesothoracic leg; spine and comb-like spur on hind leg, posterior side of fore leg basitarsus with dense pads of golden setae; tarsus or mesothoracic and hind leg with short, stiff setae, tarsal claws bidentate.Petiole: large and tabular with narrow attachments at base to the propodeum and gaster, narrow in dorsal view; fine erect setae on anterior surface above articulation with mesosoma; bulging at antero-dorsal corner as in D. longipes; integument nitid, papillate at insertion of setae on dorsal surface; keel-like subpetiolar process, anterior triangular projection (Fig. 1G).Gaster: small protuberance at articulation of gastral sternite III and the petiole, covered in erect setae; stridulatory file of varying size on acrotergite of gastric tergum II; covered in bristle-like setae, sparse short appressed setae on terga I and II; polished integument; posterior edges of the pygidium and hypopygidium with characteristic rows of spines.

Male. Unknown.
Distribution.Known only from the type locality (Fig. 13).Discussion.Dinoponera hispida is considered a valid species based on the above mentioned morphological characters, as well as its sympatry with D. gigantea without any visible character integration.Based on the morphological character states D. hispida is most similar to D. longipes and D. mutica.Dinoponera longipes differs greatly in the setae which are flagellate and golden, completely lacks gular striations and lacks the raised insertions of setae on the petiole.Dinoponera mutica also differs in pilosity, possessing drab-colored flagellate pubescence and lacks the bulging antero-dorsal corner and papillate dorsum of the petiole.
Etymology.hispida, from the Latin hispidus: bristle, referring to the conspicuous bristle-like setae covering the friend and body.Worker diagnosis.This species can easily be recognized by the golden luster of its conspicuous long, flagellate hairs especially on the frons.In addition this species has the following combination of character states: pronotal corner rounded without tooth-like process (Fig. 1E), no gular striations, a reflective, smooth and shiny integument (Fig. 12A).All specimens have a petiole which bulges on the dorso-anterior edge except for those from the Rio Madeira and Rio Negro in Brazil.Description of the worker.Measurements (mm) (n=16) TBL: 30.85-34.75 (32.83);MDL: 4.61-5.33(4.89); HL: 5.48-6.87(6.12); HW: 5.23-5.84(5.57); SL: 5.54-6.56(6.23); WL: 7.84-9.33(8.51); PL: 2.46-2.82(2.64); PH: 2.77-3.59(3.28); PW: 1.44-1.85(1.67); GL: 9. .A description of the external morphology of the worker is given in Kempf (1971): "Antennal scape from slightly shorter to slightly longer than maximum head width (index: scape L/head W × 100 = 94-103).Pubescence on front of head (as well as on thorax and dorsum of gaster) golden brown, very dense and rather long.Gular face of head smooth and shining, without vestiges of striae antero-laterally.Antero-inferior corner of pronotum obtusely angulate, not dentate.Pronotal disc smooth and shining, but densely covered with piligerous punctulae; paired swellings from faint to distinct; integument not wrinkled.Tarsus I of hind leg decidedly longer than maximum length of head capsule.Petiole smooth and shining; shape resembling that of quadriceps…, dorsal surface faintly to distinctly slanted backwards; width-length proportion distinctly lower than 0.80: vertical sulcus on posterior surface present in specimens from Acre Territory, Brasil, absent in specimens seen from Perú.Terga I and II of gaster smooth and shining but densely covered with punctulae from which arises the long and dense pubescence that covers the entire segments.Stridulatory file on acrotergite of Tergum II of gaster very short but broadly triangular, not extending backwards beyond the anterior half of acrotergite." Male diagnosis.Distinguished from other Dinoponera by the following combination of character states: funiculus of antennae with short, thick decumbent setae (Fig. 4H); pygidial spine (Fig. 4M) shorter than in D. gigantea and D. quadriceps but longer and narrower than in D. australis and D. snellingi, volsella with broad basal lobe covered in minute teeth (Fig. 10E).
Distribution.Dinoponera longipes have been collected in eastern Perú in the departments of Loreto, Amazonas, Huánuco, San Martin and Pasco, as well as Ecuador in the province of Pastaza.In Colombia it has been recorded near the Peruvian border in the department of Amazonas.In Brazil, D. longipes has been found in Acre, Amazonas as far east as Manaus, as well as along the Rio Madeira in Rondônia (Fig. 13).
Discussion.Doubt was raised by Kempf (1971) as to whether D. longipes was a valid species.Since few specimens have been collected from western Brazil a clinal variation in character form with D. gigantea or D. mutica was a possibility (Kempf 1971).The specimens we have examined from Brazil show no such integration.Additionally, evidence of species validity comes from the unique morphology of the male.Among the holdings at the CASC, males were located with 20 worker specimens from Tingo Maria, Departamento de Huánuco, Perú; all D. longipes.The nearest known locality of another species is D. gigantea 550 km away at Estirón Rio Ampiacu in the Departamento de Loreto, Perú.There is a possibility that these could be males of a yet undiscovered species.However, relatively intensive collecting of Dinoponera in the area by numerous collectors has not revealed any other form.

Dinoponera lucida
"Antennal scape remarkably longer than head width.Pubescence on front and vertex generally longer and denser than in gigantea, but lacking the golden luster on longipes.Gular face smooth and shiny, with fine, more or less distinct striation antero-laterally and antero-mesially (sometimes nearly effaced).Sides of head smooth and shining in spite of the very fine, superficial microsculpture which is reticulate-punctate.Antero-inferior corner of pronotum obtusely angulate or rounded.Pronotal disc smooth and shiny, lacking irregular fossae and wrinkles; the paired swellings rather weakly expressed.Tarsus I of hind leg decidedly longer than head width.Petiole of distinctive shape…, shorter than that of gigantea and longipes, but width-length proportion still under 0.08; anterior and posterior upper corners subequally rounded; smooth and shining; vertical sulcus on posterior face usually obsolete, present only in one Bolivian specimen.Terga I and II of gaster very indistinctly, superficially and finely reticulate-punctate yet quite smooth and shining, lacking the dense foviolae of longipes on disc where the pubescence is likewise scarce.Stridulatory file well-developed, triangular but short, visible only when acrotergite of tergum II is fully exposed." Male.Unknown.Distribution.Dinoponera mutica is found in central South America in the Brazilian states of Rondônia, Mato Grosso, Goias and Mato Grosso do Sul, in eastern Bolivia and northwest Paraguay (Fig. 12).
Discussion.Dinoponera mutica is a valid species based on our study.Dinoponera quadriceps is the closest to D. mutica in terms of morphological characters and is not synonymized in this work because of the differences stated in the diagnosis above.Males of D. mutica may provide further support for separation from D. quadriceps.Dinoponera quadriceps has a finely micro-sculptured integument which is not shiny (Fig. 12B), lacks gular striations and has a petiole which bulges on the dorso-anterior edge.Dinoponera longipes and D. hispida may also be confused with D. mutica but this species lacks the dense golden pubescence of the former, or the short, stiff setae and forward bulging petiole of the latter.
We also agree with the synonymy of Dinoponera opaca by Kempf (1975) after examination of the type.Dinoponera quadriceps and D. mutica differ in micro-sculpturing, gular striations and petiole shape.Distribution records show a distance of over 900 km between the two species, but if specimens are found with an integration of characters in the area of Tocantins and northern Goias than these species should be synonymized.
Material Male diagnosis.Specimens of this species are distinct in several respects.The combination of a bicolored body and head possessing bulging compound eyes and ocelli (Fig. 4D) is unique to this species.More definitive is the shape of the aedeagus which possesses a large ventral lobe and finger-like serrated flange (Fig. 11B).The short broad digitus volsellaris with finely toothed basal lobe (Fig. 10B) is distinctive, as well as the paramere shape (Fig. 9B).
Discussion.Dinoponera snellingi is a new species based on the suite of morphological characters presented in the diagnosis above.Most important are the shape of the aedeagus, volsella and parameres all of which we consider apomorphic characters.The type specimen males were unassociated with workers.Initially D. snellingi specimens were considered males of D. australis; as workers of this species were collected at the same location and at the same date (see D. australis materials examined).Additionally the specimens shared the same character states of bicoloration and short pygidial spine that Kempf (1971) used to designate D. australis.However, the size of the compound  eyes (compare Fig. 4D and 4E), bulging ocelli at the posterior of the head (compare Fig. 4D and 4E), short broad volsella with large tear drop-shaped basal lobe (Fig. 10B) and penis valve of the aedeagus with disto-lateral process, disto-ventral lobe and serrated flange on the ventral edge (Fig. 11B) provide strong evidence supporting that these male specimens represent a novel species.
We have compared male specimens of D. snellingi with those of D. australis collected in nest series and found they differ in the characters listed above.Campo Grande is within the range of Dinoponera mutica and there is a possibility that these specimens represent the currently unknown males of D. mutica.However, the males of D. snellingi  (Kempf 1971, 1975, Araujo et al. 1990, Peeters et al. 1999, Monnin and Peeters 1999, Fourcassié and Oliviera 2002, Monnin et al. 2003, Mariano et al. 2004, Araújo and Rodriques 2006, Marques-Silva et al. 2006).
are closest in character states to the male of D. australis, the worker caste of which differs greatly in many characters from the other known Dinoponera workers including D. mutica (see the Dinoponera australis discussion).Therefore we hypothesize that the male of D. mutica will most likely be similar to D. quadriceps or D. longipes, based on the similar worker morphology, and the unknown worker of D. snellingi will be similar to the worker of D. australis.Species groupings based on worker and male character states overlap; leaving D. australis with D. snellingi allied and separate from the other Dinoponera species.Until associated workers are discovered, we contend that it is better to describe these unique males rather than allow them to remain misidentified and unstudied or describe them as males of D. mutica with only anecdotal evidence as justification.
Etymology.Named in honor of the late Roy Snelling who made considerable contributions to the field and spirit of myrmecology.

Discussion
A synthesis of our understanding of Dinoponera morphological characters and geographic distribution supports the six species designations of Kempf (1971Kempf ( , 1975) ) as well as establishes the two new species Dinoponera hispida and Dinoponera snellingi.The collection records from material loaned to us, in conjunction with those found in previous works (Kempf 1971, 1975, Araujo et al. 1990, Peeters et al. 1999, Monnin and Peeters 1999, Fourcassié and Oliviera 2002, Monnin et al. 2003, Mariano et al. 2004, Araújo and Rodriques 2006, Marques-Silva et al. 2006) establishes range overlaps between 7 species, including sympatry between D. longipes-D. gigantea, D. gigantea-D. quadriceps-D. hispida and D. australis-D. snellingi-D. mutica.Dinoponera lucida is the only species which has no sympatric records nevertheless, a record in the state of Bahia, Brazil comes within 62 km of D. quadriceps.In all cases there is no perceptible integration of characters.
Worker characters, though seemingly indistinct upon first inspection, allow relatively easy identification of Dinoponera species.The most important characters are the tooth on the antero-ventral corner of the pronotum, in conjunction with pilosity, microsculpturing and body size.In males the pilosity of the funiculus, paramere shape, lobes of the volsella and shape of penis lobe of the aedeagus are the differentiating characters.Kempf (1971) questioned the stability of characters between species' ranges then thought of as being mutually exclusive.The areas between D. longipes, D. mutica, D. quadriceps and D. gigantea were specifically questioned.With the exception of D. mutica, our study has examined material that shows overlap between these ranges with no integration of characters used to define these species.
Despite the work presented here, many questions still remain in terms of Dinoponera taxonomy.The male caste is still undescribed in D. mutica, D. lucida and D. hispida and the worker caste is unknown for D. snellingi.These unknown castes likely exist in the collections of Brazil and elsewhere.Ranges are roughly defined but our study revealed several vast range extensions and country records.As evident in Fig. 13, large areas of South America remain uncollected.

Figure 1 .
Figure 1.Features of Dinoponera workers.A Head, frontal view B-C Occiput of head, oblique anterolateral view D-e Pronotum, lateral view F-H Petiole, lateral view.A-B D. longipes C D. hispida D D. gigantea.e-F D. mutica G D. hispida H D. lucida.

Figure 2 .
Figure 2. Dinoponera hispida worker.Head in full frontal view; body in lateral view.

Figure 3 .
Figure 3. Dinoponera longipes male.Head in full frontal view; body in lateral view with wings not shown.

Figure 4 .
Figure 4. Features of Dinoponera males.A-e Head, frontal view F-J Right scape, first and second funicular segments, frontal view K-O Pygidial spine, dorsal view A, F, K D. gigantea B, G, l D. quadriceps C, H, M D. longipes D, i, N D. snellingi e, J, O D. australis.

Figure 5 .
Figure 5. Wings of known males.A D. quadriceps B D. snellingi C D. gigantea D D. australis e D. longipes.

Figure 7 .
Figure 7. Dinoponera snellingi male genitalia.A dorsal view B lateral view C ventral view.

Figure 8 .
Figure 8. Dinoponera longipes male genitalia.A dorsal view B lateral view C ventral view.

Figure 10 .
Figure 10.Dinoponera right volsella of known males, lateral view.A D. quadriceps B D. snellingi C D. gigantea D D. australis e D. longipes.

Figure 11 .
Figure 11.Dinoponera right penis valves from the aedeagus of known males.A D. quadriceps B D. snellingi C D. gigantea D D. australis e D. longipes.

Figure 12 .
Figure 12.Worker head, oblique antero-lateral view illustrating microsculpturing difference.A Dinoponera mutica (this smooth integument type is also found in D. lucida, D. longipes and D. hispida) B D. quadriceps (this rough integument type is also found in D. gigantea and D. australis).