Large ants are not easy – the taxonomy of Dinoponera Roger (Hymenoptera: Formicidae: Ponerinae)

The taxonomy of the giant ants of the genus Dinoponera is revised based on female and male morphology. Eight species are recognized. Dinoponera nicinha sp. nov., from Amazonas and Rondônia, Brazil, is described and D. grandis (Guérin-Méneville, 1838) is revived. The species D. australis Emery, 1901 and D. snellingi Lenhart, Dash & Mackay, 2013, plus the subspecies D. australis bucki Borgmeier, 1937 and D. australis nigricolor Borgmeier, 1937 are synonymized under D. grandis sp. rev. An unnamed and unidentified male is reported. In general, male morphology has greater and more discrete variation than in females, but they are scarce in museum collections. Species distributions are updated and illustrated, the genus ranging from southern Colombia to northern Argentina, with no reliable records from the Guiana Shield and all nominal species occurring in Brazil. Intraspecific variation and natural history are discussed. New illustrated identification keys are provided for both sexes. Future studies should address the collection of fresh specimens for molecular work and to assess the conservation status of several species and populations.

Introduction to covering all surface or just anterior half; pilosity variable; hypostomal tooth usually longitudinally striate. Labrum smooth or weakly transversely striate-punctate; median longitudinal sulcus present or absent. Mandible subtriangular; usually with seven teeth; basal tooth separated from other teeth by diastema; dorsum weakly longitudinally striate only at inner base or along masticatory margin; long hairs present along masticatory margin; integument widely impressed at hair insertions. Mandible ventrally curved in lateral view. Mandible in ventral view with shallow longitudinal basal sulcus that gradually fades apicad; sulcus and row of long, flexuous hairs parallel masticatory margin. Discal area of stipes depressed, usually smooth, sometimes weakly rugulose; usually two long and flexible hairs visible at base. Submentum with weak sulcus at base; discal protuberance present; usually no microsculpturing visible. Palpal formula 4,4.
Mesosoma. In lateral view, dorsal margin slightly divided by promesonotal junction; pronotum slightly higher than rest of mesosoma. Dorsal margin of pronotum in lateral view convex. Pronotum with lateral dorsoposterior swelling of variable convexity; smooth and shining to opaque and microareolate and/or rugulose; lateral surface densely covered by appressed to decumbent pubescence. Anteroventral corner of pronotum rounded, angled or toothed. Propleuron densely punctulate and pubescent. Prosternal process longer than wide. Mesonotal and propodeal dorsal margins in lateral view continuous and widely convex, metanotal groove slightly impressed. Mesopleuron and mesosternum separated by welldeveloped carina. Anepisternum usually with mesopleural pit next to mesometapleural suture and to metathoracic spiracle. Mesometapleural suture straight and well developed. Metanotal spiracular sclerite ovoid. Dorsal and declivitous propodeal margins in lateral view rounded and continuous; propodeal spiracle opening vertical to oblique and slit-shaped; spiracle at same level as surrounding propodeal surface; metapleural-propodeal suture well-marked and straight, or weak and sinuous. Mesometapleuron and lateral surface of propodeum punctulate; appressed pubescence variable in length and density. Mesonotum and propodeal dorsum usually densely covered with punctulae bearing medially converging pubescence. Mesosoma usually with numerous long hairs, distance between each hair usually less than half length. Coxae, femora and tibiae microareolate or smooth, with long hairs throughout each surface, separated from each other by distance less than half its length. Protibial apex with robust seta and well-developed pectinate spur, spur lamellate on inner base. Protibia with pubescence concentrated on opposing side of robust seta. Meso-and metatibia each with row of robust setae around entire apex, each apex with simple spur and pectinate spur; pubescence more concentrated on opposite surface of pectinate spur. Posterior surface of probasitarsus with basal concavity containing row of dark brown setae and parallel row of fine, light brown hairs; surface covered by thick yellow hairs. Tarsi with dark brown setae concentrated ventrally on apex of each tarsomere. Tarsal claws bidentate, arolium absent.
Metasoma. Petiolar node in lateral view usually robust and sub-rectangular; taller than long; lateral surface smooth and shining to microareolate and opaque; anterior margin with short, whitish decumbent hairs. Node longer than broad in dorsal view, narrower than either propodeum or abdominal tergite III; lateral margins subparallel to anterodorsally converging, always with rounded corners; integument shining or opaque. Hairs present in variable density. Posterior surface usually with shallow longitudinal sulcus. Petiolar sternite height in lateral view always higher anteriorly than posteriorly; subpetiolar process anteroventrally projecting; anterior apex rounded or longitudinally carinate. Anterior margin of abdominal tergite III in lateral view broadly convex and vertical; dorsal margin broadly convex; prora discrete and slightly concave ventrally. Gastral spiracles rounded. Abdominal tergite III in dorsal view smooth and shining to microareolate and opaque, with hairs of variable length and density; pubescence either present on entire surface or concentrated only laterally. Abdominal pretergite IV with fine transverse striae, stridulatory file of variable size (usually well-developed in D. lucida Emery, 1901 andD. longipes Emery, 1901). Pygidium and hypopygidium each with lateral row of stout setae. Apex of hypopygium ending in two robust triangular spines which are usually covered dorsally by pygidium.
Head. In full-face view ovoid, wider than long. Anterior margin of clypeus with a transverse lamella, concave; surface of clypeal disc with slight swelling. Anterior tentorial pit large. Epistomal sulcus usually broadly convex. Supraclypeal area flat and triangular-shaped, located between antennal sockets. Distance between antennal sockets less than their maximum diameter. Torular arch simple, exposing antennal condyles. Frontal carina reduced to a short median longitudinal swelling between posterior margins of antennal sockets. Eye large and convex, occupying entire lateral cephalic margin; inner margin emarginate at height of antennal socket. Median and lateral ocellus present; lateral ocellus surpassing posterior head margin in full-face view, except in some specimens of D. grandis. Posterior margin of head convex. Head punctulate and slightly microareolate, varying from shining to silky sheen; with decumbent to suberect yellowish pubescence; variable density of hairs. Antenna with 13 segments; filiform. Scape wider than all other segments; at least two times as long as pedicel; length approximately half that of third antennal segment. Antenna with appressed pubescence; hairs, if present, only on funiculus. Ventral surface of head varies from microareolate and subopaque to smooth and shining. Mandible reduced, edentate, spoon-shaped in lateral view; weakly punctulate and shiny. Labrum reduced, narrower than labium; anterior margin concave, rarely rounded in D. quadriceps Kempf, 1971. Palpal formula 4,3, sometimes 5,3.
Mesosoma. Pronotum in dorsal view narrow, with concave posterior margin; posterolateral corner acute and bulging. Mesoscutum in dorsal view rounded; with anteromedian longitudinal carina and lateral parapsidal line; notaulus present or absent; in lateral view with convex dorsal margin. Mesopleural sulcus oblique and well-developed; mesopleural pit easily distinguible, opening anteroventrally. Spiracular sclerite approximately rounded, opening posterodorsally. Scutoscutellar sulcus easily distinguible and usually scrobiculate. Mesoscutellum in dorsal view with concave anterior margin and straight to convex posterior margin, usually with lateral strigulae; dorsal margin in lateral view strongly convex. Metanotum usually with silky sheen, narrow and sub-rectangular in dorsal view, convex in lateral view; metanotal disc slightly rugulose, laterally strigulate. Metapleuron with oblique median depression at posterior metapleural pit; metapleural pit opening usually facing anteroventrally, except in some specimens of D. grandis and D. lucida. Metapleuron with anteroventral carina that extends briefly posterodorsally. Metapleural-propodeal suture forms carina that fades dorsally to propodeal spiracle, becoming broad, sometimes scrobiculate, sulcus. Anterior margin of propodeum in dorsal view with median notch; dorsal and declivitous margins of propodeum slightly discontinuous, sometimes separated by weak carina; propodeal spiracle slit-shaped. Mesosoma varies from microareolate and subopaque to smooth and shining, becoming coarsely punctate on declivitous surface of propodeum, except in D. grandis; usually with decumbent pubescence; long hairs present or absent. Legs very slightly microareolate and shining; densely pubescent; long hairs present or absent. Protibial apex sometimes with one stout seta; protibial spur well-developed and pectinate with lamellate inner base. Meso-and metatibiae each with two pectinate spurs at apex: one well-developed and the other less developed. Posterior surface of probasitarsus with basal concavity containing row of short setae. Tarsi with decumbent setae, usually more concentrated on apex of each tarsomere; claws bidentate; arolium well-developed. Wings covered by short pubescence. Forewing with pterostigma and following longitudinal veins: C, Sc+R, R, Rs+M, M, M+Cu, Cu and A; crossveins 2r-rs, 2rs-m, 1m-cu and cu-a present. Hindwing with following longitudinal veins: R+Rs, Rs, M+Cu, M, Cu and A; crossveins 1rs-m and cu-a present.
Metasoma. Petiolar node in lateral view with rounded dorsal margin, anterior and posterior margins forming a continuous curve, usually with posterior hump; spiracle oval to rounded. Petiolar sternum continuous, lacking well-developed subpetiolar process, usually longitudinally carinate. In dorsal view longer than wide with rounded corners. Sculpturing varies from microareolate and subopaque to smooth and shining; pubescence and pilosity of variable density. Anterior margin of abdominal tergite III in lateral view broadly convex and curving posterad. Gastral spiracles oval to rounded. Abdominal pretergite IV with stridulatory file. Well-marked constriction between abdominal segments III and IV; abdominal tergite VIII (pygidium) triangular to spiniform. Cercus apically dilated, with long hairs; hypopygium elongate, posterior margin concave to straight. Sculpturing varies from microareolate and subopaque to smooth and shining; pubescence and pilosity of variable density.
Genitalia. Basal ring in dorsal view trapezoid, with lateral margins converging anteriorly; dorsoanterior margin thicker, forming lateral loop and median invagination; each loop surrounds a fenestra. Basal ring in lateral view with dorsal margin varying from straight to very concave, some species having anterior lobe; anteroventral process varying from rounded to subquadrate to subtriangular. Parameres notably divided in two regions: gonocoxite anteriorly and gonostylus posteriorly; gonostylus varying from broad and rounded to narrow and sharp. Cuspis volsellaris in lateral view finger-like and ventrally torulose, size variable; digitus volsellaris widening posteriorly, ventral margin varying from broadly to strongly concave; posterior region torulose. Penisvalva highly variable among species, with ventral margin serrated.
Color. Integument in shades of brown, with legs usually lighter.

Remarks
This genus is rarely mistaken with any other due to its enormous size. Dinoponera shares some synapomorphies with Pachycondyla Smith, 1858, the most notable being the presence of a row of stout setae on the hypopygium on each side of the sting (Schmidt & Shattuck 2014). However, females of Pachycondyla have no tooth on the anterior margin of the clypeus and rarely exceed 2 cm in body length. Neither females nor males of Pachycondyla have a stridulatory file on abdominal pretergite IV, a character present in both sexes of Dinoponera. A possible karyotypic apomorphy of Dinoponera is the presence of a pair of pseudo-acrocentric chromosomes, observed in D. grandis, D. gigantea, D. lucida and D. quadriceps (Santos et al. 2012).

Biology
Nests are always constructed in the soil, usually close to the base of trees or in areas shaded by vegetation (Araujo et al. 1990;Morgan 1993;Paiva & Brandão 1995;Fourcassié et al. 1999;Fourcassié & Oliveira 2002;Peixoto et al. 2010). Each nest may have one to thirty entrances, and some may be polydomic (Morgan 1993;Paiva & Brandão 1995;Fourcassié & Oliveira 2002). The entrances may contain branches and loose soil, forming mounds in D. longipes (Morgan 1993;Fourcassié et al. 1999;Peixoto et al. 2010). Nest depth varies from 10 to 143 cm, with species from more arid environments, such as D. quadriceps and D. grandis, usually having deeper nests. Nests are distributed in patches, and the spatial arrangement within each patch may be random or influenced by intraspecific territorial competition (Araujo et al. 1990;Paiva & Brandão 1995;Fourcassié & Oliveira 2002;Vasconcellos et al. 2004). Local density of nests can vary from 15 to 180 per hectare (Vasconcellos et al. 2004;Tillberg et al. 2014). In D. longipes, Morgan (1993) estimated an average distance of 35 m between nests.
In Dinoponera there has been loss of the queen; consequentially, reproduction is carried out by workers called gamergates (Haskins & Zahl 1971;Peeters & Crewe 1984;Araujo et al. 1990;Peeters 1991;Peixoto et al. 2008). All workers have reproductive potential, but a hierarchical system of dominance guarantees that the colonies are usually monogynous (Monnin & Peeters 1999;Monnin et al. 2003;Peixoto et al. 2008). Each colony has a cadre of workers of elevated hierarchical status that usually compete with each other for the reproductive role. These workers are often referred to as alpha, beta, gamma, and delta, being defined by the frequency and intensity of specific behaviours (Monnin & Peeters 1999;Peixoto et al. 2008). The gamergate is the alpha, the highest status in the colony (Monnin & Peeters 1999). Agonistic interactions performed by workers of higher status usually are blocking, gaster rubbing and gaster curling (Monnin & Peeters 1999;Peixoto et al. 2008). When beta, gamma or delta workers lay unfertilized eggs, the gamergate often recognizes them and destroys them by oophagy (Monnin & Peeters 1997). Dominance hierarchy is also determined by age (Araujo et al. 1990), as usually only young workers have high level positions (Monnin & Peeters 1999). As workers age their ovaries atrophy and they lose their reproductive capacity (Araujo et al. 1990). Thus, a newly emerged worker has a good chance to become a beta and possibly an alpha (Araujo et al. 1990;Monnin & Peeters 1999). Gamergates, however, may be relatively old, which suggests that fertilization is an important factor that prevents ovary atrophy (Araujo et al. 1990).
Reproduction occurs when a virgin alpha leaves the nest at night and waits for a male to fertilize it (Monnin & Peeters 1998). Male attraction probably occurs through pheromones produced only by alpha females (Monnin & Peeters 1998). After mating, the male genitalia remain attached to the gamergate, ensuring monandry (Monnin & Peeters 1998). A newly fertilized alpha can replace a dead gamergate or found a new colony by fission (Monnin & Peeters 1998). Fission can be gradual, as demonstrated for D. quadriceps and D. lucida (Peixoto et al. 2010;. Initially the new gamergate and other workers can migrate to a nearby nest and remain in contact with the parental nest. After this polydomous stage, that may continue for months, the two colonies gradually become independent from each other (Peixoto et al. 2010;. Several invertebrates can inhabit Dinoponera nests. The most common taxa are Termitidae, Gastropoda, Corinnidae, Zygentoma, Tenebrionidae, Histeridae, Phoridae, and Isopoda (Paiva & Brandão 1995;Vasconcellos et al. 2004). Ants of the genus Pheidole Westwood, 1839 are often found in nests of several species of Dinoponera (Paiva & Brandão 1995;Vasconcellos et al. 2004;Peixoto et al. 2010).

Distribution
Dinoponera has been recorded in most Brazilian states except Amapá, Roraima and Rio de Janeiro. The genus also occurs in other countries of South America, such as Colombia, Ecuador, Peru, Bolivia, Paraguay and Argentina (Kempf 1971;Lenhart et al. 2013). Lenhart et al. (2013) recorded the genus in Guiana, but this single record may be a labeling error, since no other records have been reported throughout the region between the Rio Negro in the Amazon and the Bartica district of Guiana.

Key to males of Dinoponera Roger, 1861
For detailed images of genitalia see Tozetto & Lattke (2020

Diagnosis
Female Malar area with weak longitudinal to oblique striae that do not reach anterior eye margin. Ventral head surface mostly strigulate. Anteroventral corner of pronotum in lateral view with tooth or acute angle. Dorsum of pronotum and abdominal tergite III strongly microareolate and opaque. Body length (BL) greater than 27 mm.

Male
Antenna with long suberect to erect hairs, clearly longer than maximum scape diameter. Mesoscutum with notaulus. Gastral tergites with long hairs. Dorsal margin of basal ring in lateral view broadly convex to straight.
Paralectotypes ( Head. Malar area without striae or with weak longitudinal to oblique striae that do not reach anterior eye margin. Gena microareolate and opaque, usually without rugulae. Apressed brownish pubescence present between eye and frontal lobe, extending posteriorly to frons. Frons microareolate and opaque; with flexuous, brownish and suberect hairs, longer than scape diameter; pubescence densest laterally. Occipital corner strongly microareolate and opaque. Antennal scape microareolate and silky, suberect and long hairs usually present on antennal segments 1-3. Ventral surface of head microareolate with arched strigulae covering entire surface or at least anterior half; posterior to postgenal suture with well-marked transverse strigulae. Hypostomal tooth with longitudinal strigulae. Labrum with median longitudinal sulcus weakly marked or absent; transverse rugulae absent. Mandibular dorsum weakly longitudinally strigulate on inner base, sculpture gradually fading apicad. Mesosoma. Dorsal margin of pronotum in lateral view broadly convex, usually with no pronounced dorsoposterior swelling; anteroventral corner of pronotum toothed or forming acute angle. Pronotal dorsum strongly microareolate, rugulose and opaque. Metapleural-propodeal suture weak and sinuous, with at least one curve ventral to position of propodeal spiracle. Metasoma. Petiolar node in lateral view elongate (usually DPI < 0.8); anterodorsal corner at same level as posterodorsal corner, sometimes slightly higher; anterior margin straight or slightly concave, forming blunt angle with dorsal margin; dorsal and posterior margin each broadly convex and forming blunt to rounded angle. Node lateral face microareolate and opaque (rarely silky). Node anterior margin in dorsal view slightly convex; posterior margin straight to broadly convex; lateral faces broadly convex. Abdominal tergite III strongly microareolate and opaque; densely punctulate on dorsal and lateral surfaces; distance between each punctulae approximately equal to their diameter; densely covered with brownish, flexuous, erect hairs and appressed pubescence on all surfaces. Head. Frontal carina forming short longitudinal swelling. Lateral ocellus clearly surpassing posterior head margin in full-face view. Head punctulate, weakly microareolate and shining; with yellowish decumbent pubescence and long suberect to erect hairs longer than ocellus height in full-face view. Antenna with appressed pubescence and long suberect to erect hairs clearly longer than maximum scape diameter; third to fourth most apical antennomeres with few sparse hairs. Ventral surface of head punctulate and slightly microareolate with silky sheen.

Male
Mesosoma. Mesoscutum with notaulus slightly impressed. Mesopleural sulcus punctate or slightly scrobiculate. Scutoscutellar sulcus scrobiculate. Mesoscutellum longitudinally strigulate laterally. Metapleural-propodeal suture with same microsculpture as rest of integument. Mesosoma mostly slightly microareolate and with silky sheen, becoming coarsely punctate on declivitous surface of propodeum; with decumbent to suberect pubescence and suberect hairs, distance between each hair usually less than half its length. Legs densely covered by decumbent pubescence; also with sparse suberect long hairs, greater than femur diameter, from coxa to tibia. Protibial apex with at least one robust seta.
Genitalia. Basal ring in dorsal view having slightly concave lateral margins, anteriorly narrower than posteriorly; fenestra rounded to slightly transversally directed; median invagination U-shaped; in lateral view, dorsal margin broadly convex to straight; anteroventral process rounded or slightly subquadrate. Gonostylus narrow and sharp. Dorsal margin of volsella in lateral view anteriorly straight to broadly convex and posteriorly concave; anteroventral corner with two anterior teeth; posteroventral margin strongly concave and forming sharp, triangular posterior lobe; digitus volsellaris with posterior margin straight or with a slight concavity. In lateral view, penisvalva with continuous dorsal and posterior margins, ending in rounded apex; ventral margin even and serrated; concave; anteroventral corner sharply pointed.
Color. Body mostly chestnut brown.

Remarks
A robust set of characters separate females of D. gigantea from the other species of the genus. It differs from D. hispida, D. longipes, D. mutica and D. nicinha sp. nov., all species with either an overlapping or close distribution, by the opaque and microareolate pronotal dorsum and abdominal tergite III. Additionally, D. gigantea usually has the anteroventral pronotal corner with a tooth or an acute angle in lateral view. The other species with a toothed pronotum are D. lucida and D. grandis. Dinoponera lucida has bluish iridescence on abdominal tergite III and striae on the malar area that reach the anterior eye margin. Dinoponera grandis has a body length (BL) less than 27 mm and a relatively shorter petiolar node in lateral view (DPI > 0.8), besides the presence of well-marked striae on the malar area.
Rarely, some specimens may either have a small tooth or obtuse angle on the anteroventral pronotal corner or the third abdominal tergite with very sparse piligerous punctulae, and occasionally both. In such cases, D. gigantea may be confused with D. quadriceps, but this may be avoided by examining the ventral surface of head, which has arched strigulae covering the entire surface or at least the anterior half in D. gigantea. These variants are more commonly found in Maranhão and Tocantins, where the distribution of these species may overlap. Dinoponera gigantea and D. quadriceps are nevertheless maintained separate, as argued in the remarks for D. quadriceps.
Additional support for this separation is provided by the huge differences between their habitats, as D. gigantea is typically from the Amazon and D. quadriceps is from the Caatinga (IBGE 2019). It is worth noting that Maranhão and Tocantins, where the 'problematic' specimens are found, are mostly within the Cerrado Biome, which is located between the Amazon and Caatinga Biomes (IBGE 2019).
The lectotype and two paralectotypes of D. gigantea from the ZSM were examined. Diller (1990) designated the only labelled specimen as the lectotype in the type series of six ants. One of the examined paralectotypes (ZSM HYM-25183) is unusual in having a body length (BL) less than 2.7 cm, a straight metapleural-propodeal suture, a relatively short petiolar node in lateral view (DPI = 0.79), the ventral cephalic surface weakly strigulate, abdominal tergite III with sparse piligerous punctulae and relatively short and thick body hairs, especially on the cephalic venter. Collectively these characters do not correspond with the typical characterization of any species in this genus. Nevertheless, recognizing a new species was not considered, as the exact collection locality is unknown. Perhaps this specimen is an atypical variant of D. gigantea. The worn integument and caked dirt may have abraded some hairs or made others appear more rigid than they should be, potentially complicating character interpretation. The specimen designated as lectotype by Diller (1990) is the one that better represents the forms described as D. gigantea and more resembles the specimens found in the North of Brazil. Kempf (1971: 372) gave a diagnostic description of the male based on specimens from Belém, and Lenhart et al. (2013) described the genitalia of a male from Belém. Tozetto & Lattke (2020) described the genitalia of 3 males, all from different sites in Pará. This last study found greater similarities between the basal ring, gonopod, volsella and penisvalva of D. gigantea and D. quadriceps than between other species. Lenhart et al. (2013), in their identification key to the males of the genus, separated D. gigantea and D. quadriceps from those of other species by the long and erect hairs of the funiculus. Kempf (1971) noted the similarities between these two species on account of eye and ocellus size. The aforementioned similarities between these two species suggest they are more closely related to each other than with any of the other known species. Nevertheless, males of D. gigantea can easily be separated from D. quadriceps by the presence of a notaulus on the mesoscutum and the abundant hairs on the body.

Biology
Within their known range these ants are found in areas where the average temperature of the coldest month is higher than 18°C and the annual precipitation is higher than 1000 mm (Alvares et al. 2013). They may be found at sites with different phytophysiognomies, ranging from dense forests to savannas with small woody plants ( IBGE 2012). Their nests, which may be polydomic, are usually built at the base of trees and may be up to 40 cm deep (Fourcassié et al. 1999;Fourcassié & Oliveira 2002). Each nest has from one to eight entrances surrounded by loose soil never piled into a mound (Fourcassié & Oliveira 2002). On average, each nest has some 41 adult workers (Monnin et al. 2003).
Workers forage alone searching for a broad variety of resources such as seeds, fruits, and dead or live arthropods (Fourcassié & Oliveira 2002). Foraging may be diurnal or nocturnal, with activity decreasing during the hottest hours of the day (Fourcassié et al. 1999;Fourcassié & Oliveira 2002). Workers exhibit general fidelity, with minor deviations, to a foraging route (Fourcassié et al. 1999). Spatial orientation during foraging apparently uses relatively large-scale reference points in the landscape, as minor changes, such as sweeping the leaf litter, do not change their path (Fourcassié et al. 1999). Workers from different colonies may find themselves at the edge of their foraging routes and exhibit territoriality, beating each other with their antennae (Fourcassié & Oliveira 2002). Hierarchical relationships within a colony are determined by the same agonistic interactions described for D. quadriceps, although with a lower frequency (Monnin et al. 2003).

Distribution (Fig. 28A)
Dinoponera gigantea has been recorded in northern Brazil and eastern Peru. In Brazil, the northernmost record is the type-locality along the Rio Negro (probably close to Barcelos, Amazonas, according to Kempf 1971). It is important to note that, except for the type-series, D. gigantea has only one record for Amazonas. All other records are mostly from Pará and Maranhão, with the southernmost record from Rio Tapirapé, Mato Grosso. The examined records from Peru date from 1966 and are close to the border with Brazil, more than 900 km from any other record. This may indicate a much more extensive distribution in the recent past. Lenhart et al. (2013) examined a specimen from Bartica, Cuyuni-Mazaruni, Guiana but this single record may be an artefact of a misplaced label, as it is 500 km from any other record for the genus. Despite numerous surveys and collecting, Dinoponera has yet to be recorded from the Guiana Shield (Wheeler 1916(Wheeler , 1918Lapolla et al. 2007;Alonso 2012;Alonso & Helms 2013;Helms et al. 2016;Groc et al. 2014;Franco et al. 2019). (Guérin-Méneville, 1838 Non Dinoponera grandis australis brevis -Santschi 1928: 416 (invalid name).

Diagnosis Female
Malar area with longitudinal to oblique striae that reach anterior ocular margin. Abdominal tergite III microareolate and opaque to slightly silky, never smooth and shining. Hind basitarsus length less than 6 mm long. Petiolar node short, DPI ≥ 0.8. BL ≤ 27 mm.

Male
Body, including antenna, without long hairs. Basal ring in lateral view with dorsal margin declining and extending anteriorly to form a broad lobe. Penisvalva in lateral view with a posterior rounded apex, projecting ventrolaterally. Head. Malar area with longitudinal to oblique striae that always reach anterior eye margin, sometimes extending to gena. Gena microareolate and opaque, sometimes silky, never smooth and shining, usually without rugulae. Apressed pubescence present between eye and frontal lobe, extending posteriorly to frons. Frons microareolate and opaque or with weak silky sheen; with flexuous, brownish and decumbent to suberect hairs, longer than scape diameter; pubescence short, appressed and scarce. Occipital corner varying from strongly microareolate and opaque to smooth and silky. Antennal scape microareolate and silky, suberect and long hairs usually present on antennal segments 1-3. Ventral surface of head microareolate with arched strigulae covering entire surface or just close to postgenal suture (rarely absent). Hypostomal tooth longitudinally strigulate. Labrum with median longitudinal sulcus weakly marked; with transverse rugulae. Mandibular dorsum weakly longitudinally strigulate on inner base, sculpture gradually fading apicad.

Lectotype of Ponera grandis
Mesosoma. Dorsal margin of pronotum in lateral view broadly convex, usually with no pronounced dorsoposterior swelling; anteroventral corner of pronotum toothed or forming acute angle, rarely obtuse. Pronotal dorsum strongly microareolate and opaque to weakly microareolate and silky. Metapleuralpropodeal suture sinuous, with at least one curve ventral to position of propodeal spiracle.
Metasoma. Petiolar node in lateral view relatively short (DPI ≥ 0.8); anterodorsal corner at same level as posterodorsal corner, sometimes slightly lower; anterior margin straight or slightly concave, forming blunt angle with dorsal margin; dorsal and posterior margin each varying from broadly convex to straight and forming blunt to slightly rounded angle. Node lateral face usually microareolate and opaque, rarely slightly smooth and silky. Node anterior margin in dorsal view convex; posterior margin straight to broadly convex; lateral margins broadly convex and converging anterodorsally. Abdominal tergite III strongly microareolate and opaque to weakly microareolate and silky; punctulae in variable density, always denser laterally than dorsally; densely covered with brownish, flexuous, decumbent to suberect hairs on entire surface; appressed pubescence very sparse on dorsum, denser laterally. Head. Frontal carina forming a short longitudinal line. Lateral ocellus usually reduced, sometimes not surpassing posterior head margin in full-face view. Head punctulate, very slightly microareolate and with silky sheen; with pallid decumbent pubescence and very few decumbent hairs only on clypeus and vertex. Antenna with very short appressed pubescence, without hairs. Ventral surface of head punctulate and slightly microareolate with silky sheen.
Metasoma. Petiolar node usually slightly microareolate and with silky sheen; densely covered by decumbent to suberect pubescence, without hairs. Abdominal tergite VIII triangular, sometimes with a very sharp apex. Gaster usually microareolate and subopaque; tergites densely covered by appressed to decumbent pubescence; without long hairs.
Genitalia. Basal ring in dorsal view with slightly concave lateral margins, anteriorly narrower than posteriorly; maximum diameter of fenestra longitudinally directed; median invagination usually V-shaped; in lateral view dorsal margin declining and extending anteriorly to form a broad lobe; anteroventral process subtriangular to trapezoid. Gonostylus broad and rounded. Dorsal margin of volsella in lateral view anteriorly straight to broadly convex and posteriorly concave; anteroventral corner projecting ventrally as a subtriangular, subquadrate or rounded lobe; rarely as a narrow and sharp process; posteroventral margin usually continuous, broadly concave and forming a discrete posterior rounded to acute lobe; digitus volsellaris with posterior margin straight to broadly convex. In lateral view, penisvalva with continuous dorsal and posterior margins, usually ending in a rounded apex which projects ventrolaterally; ventral margin uneven, usually with a concavity containing a short tooth or a spiniform projection; posteriorly with serrated rounded or triangular ventral lobe of variable size; anteroventral region concave or with a short triangular lobe.
Color. Body varying from completely light brown to black, sometimes with gaster much lightercolored.

Remarks
Dinoponera grandis, as discussed below, is strongly polymorphic. The paradox about this species is that females and males tell different histories when examined separately. While males of D. grandis will vary discretely in some characters, variation is more continuous in females. Because of this, even recognizing that D. grandis may represent a complex of several cryptic species, this is not tenable, based on the current information that separates these forms.
Females from southwestern Mato Grosso do Sul, Brazil, to Amambai, Paraguay, tend to have abdominal tergite III strongly microareolate and opaque and the anteroventral pronotal tooth long and acute, almost spiniform. Ants from Distrito Federal to the northeast of Goiás, Brazil, usually have the pronotal dorsum and abdominal tergite III weakly microareolate and with a silky sheen, and the anteroventral pronotal margin obtusely angular, edentate. Some specimens from southern Brazil have the pronotal dorsum and abdominal tergite III weakly microareolate and with a silky sheen. Some specimens from Paraná could be tentatively identified as D. lucida on account of their having the anterodorsal petiolar node angle slightly lower than the posterior angle, but in D. lucida this character is more pronounced plus the pubescence on the frons is abundant and long, quite different from D. grandis.
Besides the above-mentioned variation, all examined females of D. grandis have the malar striae reaching the eye, a relatively short petiolar node (DPI ≥ 0.8), a body length of less than 27 mm and a hind basitarsus less than 6 mm long. Occasionally, D. quadriceps and D. lucida may have some of these characters, but besides being usually larger, the following traits will also separate them. The petiolar node of D. lucida is anterodorsally lower than the posterodorsal corner and its abdominal tergite III is very smooth, with bluish iridescence. In D. quadriceps, the malar striae clearly fail to reach the eye and the petiolar node often has a higher anterodorsal corner. Santschi (1928) described a form from Paraguay as Dinoponera grandis australis brevis based on female characters such as a scape length as long as the head length and a short petiolar node, one third higher than long. Borgmeier (1937) considered it as a subspecies, making the name D. australis brevis valid. However, Kempf (1971) did not consider these differences to be significant and he synonymized brevis under D. australis. Scape length in the studied specimens of D. grandis ranges from being longer to shorter than the head length. The node length was found to be one third higher than long in several specimens, leading to the conclusion that such differences are intraspecific.
Guérin-Méneville (1838) apparently described D. grandis based on two specimens. In the original description, the author referred to the extremely smooth and shining integument and to the petiolar node shape, which is rounded "up and forward". These characters could refer to what is now considered D. longipes; however, Guérin-Méneville also mentioned the type locality, probably in Minas Gerais, Brazil. Roger (1861) analyzed Guérin-Méneville's specimens and recognized that one ant had somewhat sculptured and opaque integument, while the other was mostly smooth and shining. Roger also noted that one specimen had the petiolar node relatively narrower in lateral view, but despite this he synonymized D. grandis and D. gigantea. Dinoponera grandis, the more recent name, thus became a junior synonym of D. gigantea. In recent revisions, neither Kempf (1971) nor Lenhart et al. (2013) were able to study the types of these two species, so doubts lingered as to the status of D. grandis.
One of Guérin-Méneville's syntypes in the ZSM was designated as the lectotype by Diller (1990). It corresponds to the form Roger described as opaque and with a short petiolar node, collected in Minas Gerais, Brazil.
Examining this type, we saw that it is definitely a different species than D. gigantea, its current synonym. It thus became obvious we had to separate these two species. Based on the type locality, it would be possible that D. grandis was a synonym of D. australis, D. lucida or D. quadriceps. These last two species, however, are easily separated from D. grandis. In addition, the lectotype of D. grandis was revealed to have the same characters used by Emery (1901) to describe D. grandis australis and later used by Borgmeier (1937) to raise D. australis to species level: body size smaller than in other species of Dinoponera; anteroventral pronotal tooth present; integument less shining than in either D. longipes, D. lucida or D. mutica but more so than in D. gigantea; short petiolar node, slightly longer than broad; short legs and antennae; and the scape surpassing the posterior head margin.
Due to the similarities between the lectotype of D. grandis and the description and the known forms of D. australis, allied with the fact that both occur in sympatry, these two species are synonymized. Dinoponera grandis, which is the oldest name, thus becomes the valid name (senior synonym) and D. australis becomes the junior synonym. Although the type of D. australis has not been examined, choosing not to synonymize these names would mean that there would be two names with the same diagnosis.
In this study three other names are synonymized under D. grandis: the species D. snellingi Lenhart, Dash & Mackay, 2013, and the subspecies Dinoponera australis nigricolor Borgmeier, 1937 and Dinoponera australis bucki Borgmeier, 1937. This decision is discussed below. Borgmeier (1937) described D. australis nigricolor and D. australis bucki based on females and males taken from their respective nests, the former in central Goiás and the latter in Rio Grande do Sul, Brazil. He found D. a. nigricolor females to be identical to D. a. bucki females except for having the dorsal margin of the petiolar node margin slightly convex, a difference we consider as intraspecific. However, he considered male characters to be more reliable indicators, particularly gonostylus width. Other male differences he also considered diagnostic included the petiolar node shape, body color, shade of wing color and slight differences in the relative scape length. Unfortunately, he only had one male for each subspecies. Lenhart et al. (2013) described D. snellingi based on 3 males collected at a light in Campo Grande, Mato Grosso do Sul, Brazil. The ants were initially identified as D. australis because they were collected at the same locality and on the same day as some females of D. australis. They compared these males with the descriptions by Kempf (1971) and found similarities in the short pygidial spine and bicolored body, but they also discerned a number of differences. They recognized it as a different species on account of the following combination of characters: bicolored body; head with bulging compound eyes and ocelli; penisvalva with a large ventral lobe and finger-like serrated flange; a short and broad digitus volsellaris with a finely toothed basal lobe; and a distinctively shaped paramere.
The present study was able to gather the largest number of D. grandis male specimens with the widest geographic coverage until now (11 male specimens, contrasting with five and four in Lenhart et al. 2013 andKempf 1971, respectively). Below we discuss variation in the diagnostic characters given by Borgmeier (1937) and Lenhart et al. (2013), finding many of them to be ineffective for species delimitation and thus weakening the validity of these names.
Node shape, in general, is convex as seen laterally, but it can vary from evenly convex to unevenly convex, with the highest point shifting from mid-petiolar distance to points just anterior or posterior to it. Variability was found even in the relative height of the node. Given such variation, the differences in node shape used by Borgmeier to distinguish D. a. bucki from D. a. nigricolor becomes ineffective when more specimens are compared.
The male specimen of the D. a. bucki type series is designated here as the lectotype, given the apparent greater discreteness of male characters. It is strikingly bicolored with the gaster light chestnut and the rest of the body dark brown, but the bicolor pattern is also present in a male from Goiás (Jataí) and in another from São Paulo (Itirapina), the latter with a more yellowish gaster and darker body. Lenhart et al. (2013) described D. snellingi as having the same colors, but a male from Mato Grosso do Sul (Itaum), with genitalia very similar to those of D. snellingi, is entirely chestnut. The lectotype male of D. a. nigricolor is entirely black, with the gaster slightly lighter. This pattern was also found in males from Paraná (Guarapuava and Laranjeiras do Sul) and Mato Grosso do Sul (Itaum). Another male collected at the type locality of D. a. nigricolor is entirely light chestnut, as well as a specimen from Mato Grosso (Chapada dos Guimarães).
The relative scape length was found to be slightly less in D. nigricolor than in D. a. bucki by Borgmeier (1937). The ratio of scape width divided by scape length varies from 0.73 to 0.93 in six males, including the types, so this variation becomes muddled when studying more specimens.
In general, all males studied here have reduced eyes and ocelli compared with other species of Dinoponera, but only in the lectotype of D. a. bucki are they as small as those of the D. grandis male of fig. 4e in Lenhart et al. 2013. The lectotype of D. a. nigricolor, a topotypic male, and a specimen from Itirapina, São Paulo, all have the lateral ocelli surpassing the posterior head margin in dorsal view, as in the description of D. snellingi and fig. 4d in Lenhart et al. 2013. Therefore, this character is not exclusive to D. snellingi and its usefulness in separating species becomes doubtful. Borgmeier (1937) found the gonostylus in D. a. nigricolor to be relatively wider than in D. a. bucki. Lenhart et al. (2013) described the gonostylus shape in D. snellingi as distinct from that in D. grandis. We detected differences in the shape and relative width of the gonostylus of other D. grandis males that are similar to those in D. a. bucki and D. a. nigricolor. Thus, these observed differences prove continuous among specimens, rendering them useless for diagnosis.
The anteroventral corner of the volsella in lateral view usually varies from rounded to subtriangular between specimens, with no evident pattern. An exception is provided by the lectotype of D. a. nigricolor and a topotype male, which have their anteroventral process narrow and sharp (spiniform). The posteroventral margin of the volsella of these specimens also differs by having a small subquadrate lobe close to the apex. All other males have a continuous margin. In the Itirapina male this margin is more strongly convex, giving the whole volsella an arched appearance.
All the observed males have a large anteroventral lobe on the penisvalva. The anteroventral corner, however, has two different states: either a concavity or a short triangular lobe. The latter state was found only in the Itaum and Itirapina males. The penisvalva illustration of D. snellingi in Lenhart et al. (2013) does not show this triangular lobe; instead it shows a more rounded and convex corner, which is also different from the concave corner found in other D. grandis males.
Penisvalva shape is the most variable character among examined males. In D. snellingi it is described by Lenhart et al. (2013) as having a large triangular ventral lobe with a vertical ridge running through the middle. This ventral lobe is longer than wide and is strikingly different from that of other Dinoponera males. Additionally, the ventral margin of the penisvalva has a small tooth. Males from Itaum and Itirapina have a very similar penisvalva shape, although the vertical ridge is inconspicuous. The penisvalva in the type of D. a. nigricolor and the topotypic specimen have a ventrally pronounced posteroventral lobe that forms a blunt angle, but it is never longer than wide, as in the illustration of the type of D. snellingi (Lenhart et al. 2013: fig. 11b). The penisvalva of D. a. bucki has a rounded posteroventral lobe and a long, acute ventral spine, as do the other D. grandis males. It is very similar to the description of males from Argentina (Lenhart et al. 2013).
Most of the characters mentioned above are not solid enough to justify the validity of D. snellingi, D. a. nigricolor or D. a. bucki. Even though the shape of the anteroventral corner in the volsella and penisvalva seems to vary discretely in D. a. nigricolor and D. snellingi, there are more gaps than certainties about the delimitation of these entities.
These males were taken during the same collection events as D. grandis females, but these few males do not allow us to know whether these seemingly distinct characters simply represent extremes of continuous variations.
A last consideration is that the validity of D. snellingi, D. a. nigricolor and D. a. bucki would imply we know the form of the D. grandis male, a tenuous assumption. The first description of a D. grandis male (Santschi 1921: 85) was based on specimens from Argentina, very far from the type locality in Minas Gerais, Brazil. It is much more likely that the male of D. grandis corresponds to one of the specimens described as D. snellingi, D. a. nigricolor or D. a. bucki, all of which occur much closer to the type locality.
Based on all of this evidence, we argue for the synonymy of these names; however, we strongly encourage future studies on D. grandis, as the high variability of this species can be a sign of the existence of a complex of cryptic species. For future studies, we recommend that the first question addressed be: what is the form of the D. grandis male? The next step must be to gather fresh specimens by extensive fieldwork in all distribution areas, the results of which would also enable molecular and phylogeographic studies. In the light of human-induced climate change, and continuing deforestation and forest fires, we consider such fieldwork to be of utmost importance. Not only would it help elucidate the questions surrounding the situation of D. grandis, but it would also inform conservation measures that could help preserve the remaining populations of these giant ants.

Biology
Dinoponera grandis is found in a wide variety of environments, with predominance in savannas, at least for Brazilian records. Nests have a random spatial distribution and may have a density of up to 180 nests per hectare (Tillberg et al. 2014). Each nest has an entrance without a surrounding mound and may reach a depth of 143 cm (Paiva & Brandão 1995). Despite the relatively large nest size, colonies have on average only 14 workers, which indicates that nests are constantly reused and not completely constructed every generation (Paiva & Brandão 1995;Monnin et al. 2003). Pheidole dinophila Wilson, 2003 has been recorded as nesting within nests of D. grandis at more than one locality (Wilson 2003).
They are omnivorous, but favor hunting other invertebrates (Tillberg et al. 2014). Foraging strategies involve route fidelity for individual workers and the distribution of different routes among workers, favoring greater efficiency in the area being covered (Tillberg et al. 2014). Organization inside the nest reflects division of tasks, with foraging workers occupying chambers closer to the surface and higher status workers occupying deeper chambers (Paiva & Brandão 1995). This nest occupation dynamic was confirmed by Smith et al. (2011), demonstrating that workers of deeper chambers have up to 39% of their dry body mass composed of fat, while workers found in shallow chambers may have less than 1% body fat. Dominance hierarchy is determined by agonistic conflicts similar to those in D. quadriceps, guaranteeing colony monogyny (Monnin et al. 2003). Head. Malar area with longitudinal to oblique striae which gradually fade posteriorly, sometimes not reaching anterior eye margin. Gena smooth or weakly microareolate, with bluish iridescence, usually without rugulae. Appressed pubescence present between eye and frontal lobe, extending posteriorly to frons. Frons smooth to weakly microareolate, with bluish iridescence; integument slightly irregular close to hair insertions; decumbent to suberect thick, stiff hairs; pubescence densest laterally. Occipital corner smooth and with bluish iridescence. Antennal scape weakly microareolate and shining; antenna with suberect, thick, stiff hairs except for 4-5 apical segments. Ventral surface of head weakly microareolate to smooth; arched, weakly impressed strigulae present on anterior half; well-marked transverse strigulae posteriorly. Hypostomal tooth with longitudinal strigulae. Labrum with weak median longitudinal sulcus that extends from anterior margin to half its length; with transverse rugulae. Mandibular dorsum with longitudinal strigulae along inner margin, gradually fading apicad.
Metasoma. Petiolar node in lateral view elongate (DPI < 0.8); anterodorsal corner usually higher than posterodorsal corner; anterior margin slightly concave, forming blunt angle with dorsal margin; dorsal and posterior margins each broadly convex and forming rounded angle. Node lateral face smooth or weakly microareolate with bluish iridescence. Node anterior margin in dorsal view strongly convex; posterior margin broadly convex; lateral margins broadly convex, slightly converging anterodorsally. Abdominal tergite III smooth or weakly microareolate, with bluish iridescence; punctulae sparse, slightly denser laterally than dorsally; covered by suberect thick, stiff hairs on entire surface; appressed pubescence very sparse on dorsum, denser laterally.

Male
Unknown.

Remarks
The diagnostic characters easily separate D. hispida from any other Dinoponera. This species occurs in sympatry with D. gigantea, which differs by having a microareolate and opaque integument with slender, flexous hairs. It may be confused with either D. longipes, D. mutica, or D. nicinha sp. nov., which also have a shining integument and occur in northern Brazil. Dinoponera mutica and D. nicinha sp. nov. can be separated by not having the petiolar node with a blunt anterodorsal angle or a rounded posterodorsal angle in lateral view. Dinoponera longipes may have the node slightly similar to that of D. hispida, but it lacks strigulae on the head venter and has flexible golden hairs throughout the body. Another species with smooth and shining integument is D. lucida, but it differs by having the anterodorsal node angle lower than the posterodorsal angle in lateral view. Dinoponera lucida only occurs in the Atlantic forest of eastern Brazil.
The holotype of D. hispida was examined by images available in the online database of the MCZC (https://mczbase.mcz.harvard.edu, Identifiers: Entomology, 36191) and additional images. Besides that, the description of Lenhart et al. (2013) is very complete and detailed, permitting a clear recognition of the species.

Biology
The life history of this species is unknown. It is restricted to the Amazon, characterized by humid forests and large trees (IBGE 2012). According to the Köppen classification system, the predominant climate in southeastern Pará is Tropical Monsoon, with the average temperature of the coldest month over 18°C and high annual precipitation (Alvares et al. 2013).
Distribution (Fig. 28B) Dinoponera hispida has a relatively restricted distribution, known only from a few sites in Pará that are being affected by large-scale mining, a large hydroelectric plant (Tucuruí Hydroelectric Power Plant) and extensive deforestation. This combination of habitat disturbance makes the species a likely candidate for being considered as threatened. However, its distribution could be greater if there are more specimens in collections presently misidentified as the sympatric D. gigantea. Emery, 1901 Figs 20, 29A

Diagnosis
Female Anteroventral corner of pronotum in lateral view unarmed, forming obtuse angle. Petiolar node in lateral view with anterodorsal corner clearly higher than posterodorsal corner. Abdominal tergite III smooth and with bluish iridescence; covered with sparse golden pubescence. Body covered with flexible golden hairs.
Head. Malar area with weak longitudinal to oblique striae that do not reach anterior eye margin. Gena weakly microareolate with bluish iridescence, usually without rugulae. Golden appressed pubescence present between eye and frontal lobe, extending posteriorly to frons. Frons smooth and with bluish iridescence; with flexuous, golden and decumbent to suberect hairs, longer than scape diameter; pubescence densest laterally. Occipital corner smooth and with bluish iridescence. Antennal scape weakly microareolate and silky; antenna with decumbent to suberect and long hairs, except for 4-5 apical segments. Ventral surface of head smooth, with silky sheen; longitudinal strigulae mostly absent or, if present, only near hypostomal tooth. Labrum without median longitudinal sulcus, transverse rugulae absent. Mandibular dorsum weakly longitudinally strigulate on inner base, sculpture gradually fading apicad.
Mesosoma. Dorsal margin of pronotum broadly convex in lateral view, usually with no pronounced dorsoposterior swelling; anteroventral corner of pronotum forming obtuse angle, never toothed. Pronotal dorsum smooth, with bluish iridescence. Metapleural-propodeal suture well-marked and slightly sinuous, with at least one curve ventral to position of propodeal spiracle.
Metasoma. Petiolar node in lateral view elongate (DPI < 0.8); anterodorsal corner higher than posterodorsal corner; anterior margin slightly concave, forming blunt angle with dorsal margin; dorsal and posterior margins broadly convex, forming rounded angle. Node lateral face smooth with bluish iridescence. Node anterior margin in dorsal view strongly convex, posterior margin straight to broadly convex, lateral margins broadly convex and converging anterodorsally. Abdominal tergite III smooth with bluish iridescence; punctulae in variable density, evenly distributed laterally and dorsally; covered by flexuous, golden suberect hairs and appressed pubescence evenly distributed on dorsal and lateral surfaces, ranging from dense to sparse.

Male
Not examined. and a toothless anteroventral pronotal margin in lateral view. However, the node shape of D. longipes separates it from these two species. Dinoponera quadriceps may have a similar node shape, but differs by having abdominal tergite III microareolate and opaque with brownish, not golden, hairs.

Remarks
Images of the holotype (AntWeb, CASENT 0903861) were examined. The original description includes some characters that are at odds for all the examined specimens, such as the length of the legs and the antennae being longer in comparison with other Dinoponera. Despite this, the node shape and golden pilosity were considered sufficient to maintain the validity of this species. The density of the pilosity and pubescence on abdominal tergite III varies greatly. A specimen from Rio Branco, Acre State, has the body completely covered by dense long and appressed pubescence, but specimens from Colombia have sparse pubescence. The availability in Brazil of D. longipes in collections is meagre and several female specimens identified as D. longipes actually belong to D. nicinha sp. nov.
The male was described by Lenhart et al. (2013) based on 2 specimens from Tingo María, Peru. According to these authors, D. longipes can be separated from other species by the thick decumbent antennal setae and by the shape of abdominal tergite VIII, which is shorter than in D. gigantea and D. quadriceps and longer than in D. grandis and D. snellingi. Comparing the description and illustrations of D. longipes with the males of D. lucida, it seems that they are similar, but the shape of the penisvalva is very different, not forming a rounded ventral lobe in D. longipes.

Biology
Dinoponera longipes occurs in regions with dense tropical rainforests, but Morgan (1993) found most nests in areas with sparse understory vegetation. Each nest can have one to thirty entrances with soil particles forming mounds around each (Morgan 1993). Most nests are close to roots or in areas shaded by vegetation and the distance between them averages 35 m (Morgan 1993). In two nests observed by Morgan (1993), the number of workers was seven in a young colony and 120 in a mature colony.
The diet is predominantly carnivorous and foraging is solitary and mainly nocturnal (Morgan 1993). The ants interact, displaying agonistic behaviors similar to those described in other species, such as immobilization and antennal beating (Morgan 1993;Monnin et al. 2003).

Distribution (Fig. 29A)
Dinoponera longipes is present in Brazil, Colombia and Peru. Lenhart et al. (2013) also examined specimens from Ecuador. In this study, the northernmost record is from Florencia, Caquetá, Colombia, and the southernmost record is from Porto Acre, Acre State. This species has not been collected in Brazil in over 25 years. Emery, 1901 Figs 21-22, 29A Dinoponera grandis lucida Emery, 1901: 48 (☿).

Diagnosis
Female Malar area with longitudinal to oblique striae that always reach entire anterior ocular margin and extend to gena until at least half of eye length. Anteroventral corner of pronotum in lateral view with tooth or acute angle. Dorsum of pronotum and abdominal tergite III smooth, with bluish iridescence. Anterodorsal corner of petiolar node in lateral view lower than posterodorsal corner.

Male
Antenna with stiff decumbent hairs shorter than maximum scape diameter. Gastral tergites with long hairs. Abdominal tergite VIII triangular, with a very sharp apex. Penisvalva ending in a rounded ventral lobe.

Material examined
Non-type specimens (
Head. Malar area with longitudinal to oblique striae that always reach entire anterior eye margin, striae extend to gena until at least half of eye length. Gena weakly microareolate and silky, usually without rugulae. Longitudinal to oblique striae and brownish appressed pubescence present between eye and frontal lobe, pubescence extends posteriorly to frons. Frons slightly microareolate and silky; with flexuous, brownish, suberect hairs, longer than scape diameter. Occipital corner slightly microareolate and silky. Antennal scape microareolate and silky, antenna with long suberect hairs, except for 4-5 apical segments. Ventral surface of head microareolate with arched strigulae that may cover entire surface or just anterior half. Hypostomal tooth with longitudinal strigulae. Labrum with weak transverse rugulae, without median longitudinal sulcus. Mandibular dorsum longitudinally strigulate, irregularly impressed along inner margin, gradually fading apicad.
Mesosoma. Dorsal margin of pronotum in lateral view sharply convex, usually with very pronounced dorsoposterior swelling; anteroventral corner of pronotum toothed or forming acute angle. Pronotal dorsum smooth, with bluish iridescence. Metapleural-propodeal suture weak and sinuous, with at least one curve ventral to position of propodeal spiracle.
Metasoma. Petiolar node in lateral view elongate (usually DPI < 0.8), anterodorsal corner lower than posterodorsal corner; anterior margin straight to slightly concave and forming blunt angle with dorsal margin, dorsal and posterior margins broadly convex to straight and forming blunt angle. Node lateral face slightly microareolate and silky. Node anterior margin in dorsal view convex, posterior margin broadly convex to straight, lateral margins broadly convex, slightly converging anterodorsally. Abdominal tergite III mostly smooth with bluish iridescence, usually with anterior rugulosity; punctulae denser laterally than dorsally; densely covered by brownish, flexuous, suberect hairs on entire surface; decumbent pubescence very sparse on dorsum, denser laterally. pubescence and long suberect to erect hairs, longer than ocellus height in full-face view. Antenna with appressed pubescence and stiff decumbent hairs shorter than maximum scape diameter. Ventral surface of head punctulate and slightly microareolate with silky sheen.
Metasoma. Petiolar node smooth and shining; with sparse suberect pubescence and abundant suberect hairs. Abdominal tergite VIII triangular, with a very sharp apex; mostly smooth and shining. Tergites with sparse suberect pubescence and abundant suberect hairs.
Genitalia. Basal ring in dorsal view with lateral margins slightly convex, anteriorly narrower than posteriorly; maximum diameter of fenestra longitudinally directed; median invagination V-shaped; in lateral view, dorsal margin broadly concave and extending anteriorly to form a lobe; anteroventral process subquadrate to rounded. Gonostylus broad and rounded. In lateral view, dorsal margin of volsella anteriorly convex and posteriorly slightly concave; anteroventral corner projecting anteriorly as a rounded to subquadrate lobe; posteroventral margin concave, sometimes forming a posterior subtriangular lobe; digitus volsellaris usually with posterior margin broadly convex. In lateral view, penisvalva with continuous dorsal and posterior margins, ending in a rounded ventral lobe; ventral margin uneven and serrated, with a short median spine placed in a concavity; anteroventral corner rounded to subquadrate.
Color. Body mostly chestnut brown.

Remarks
Females of Dinoponera lucida are distinct from other Dinoponera by having the anterodorsal corner of the petiolar node in lateral view lower than the posterodorsal corner. Additionally, D. lucida also has longitudinal to oblique striae on the malar area that reach the entire anterior ocular margin, extending on to the gena until at least half of eye length. Rarely, D. grandis may exhibit a similar node shape, but the anterodorsal corner is not as low as in D. lucida. Dinoponera lucida may also be separated from D. grandis by the bluish iridescence of the dorsum of abdominal tergite III, with no visible microsculpturing. Besides that, D. lucida and D. grandis are allopatric. Dinoponera mutica may have the anterodorsal node corner slightly lower than the posterodorsal corner, but it does not have wellmarked striae on the malar area that reach the entire anterior ocular margin and no anteroventral pronotal tooth. Dinoponera lucida is currently considered endangered according to the red book of threatened Brazilian fauna (ICMBio 2018). Escárraga et al. (2017) described the male of D. lucida. It differs from D. gigantea and D. quadriceps mainly by not having long antennal hairs. Besides that, in these two species and in Dinoponera morphospecies 1, the dorsal margin of the genital basal ring in lateral view does not form an anterior lobe as in D. lucida. D. longipes may be similar to D. lucida, but they can be separated by the penisvalva shape, which forms a ventral rounded lobe only in D. lucida. In D. grandis and D. snellingi, the ventral penisvalva lobe is curved laterally and not continuously rounded; neither does the antenna have stiff hairs.

Biology
Dinoponera lucida occurs mainly in dense ombrophilous forest, characterized by high humidity, medium to large trees, and large numbers of epiphytes (Campanili & Schaffer 2010). These forests are predominantly hot and humid, with average temperatures varying from 20 to 26°C (Alvares et al. 2013). Nests of D. lucida are found always in the soil, usually less than 1 m away from trees and in shaded areas (Peixoto et al. 2010). Each nest can contain one to four entrances with small sticks, leaves and loose soil all around (Peixoto et al. 2010). The distribution pattern of the nests is aggregated, with local densities varying from 20 to 52 nests per hectare (Peixoto et al. 2010). On average, each nest is 35 cm deep, with four to five chambers (Paiva & Brandão 1995;Peixoto et al. 2010). The number of adult workers per colony can vary from 22 to 106, but rarely exceeds 50 (Paiva & Brandão 1995;Peixoto et al. 2008Peixoto et al. , 2010. Peixoto et al. (2010) suggested that the number of workers tends to be higher before and at the beginning of summer, the period before nest fission. The number of males, however, does not seem to have seasonal variation, since a gamergate can be substituted at any time of the year (Peixoto et al. 2010). Dinoponera lucida is omnivorous, feeding mainly on small dead invertebrates or angiosperm seeds (Peixoto et al. 2010). The workers forage alone during the day, even at warmer hours (Peixoto et al. 2010). During foraging, a worker may forage up to 20 m from the nest entrance (Paiva & Brandão 1995;Peixoto et al. 2010). However, the distance traveled may be higher when local nest density is smaller, suggesting intraspecific competition for territory (Peixoto et al. 2010).
Founding of new colonies occurs by fission, i.e., some workers (among them a newly fertilized one) migrate to a nearby nest probably doing tandem running (Hölldobler & Wilson 1990;Peixoto et al. 2008Peixoto et al. , 2010. This process can happen gradually, since the older colony goes through a polydomic stage before the total separation (Peixoto et al. 2010). Newly founded colonies are monogynous and with age exhibit polyethism, the older ants foraging and younger ones taking care of the brood (Hölldobler & Wilson 1990;Peixoto et al. 2008Peixoto et al. , 2010. In contrast to D. quadriceps, agonistic interactions among workers are uncommon in colonies with gamergates and when they occur the gamergate does not participate in the interactions (Peixoto et al. 2008). The existence of a dominance hierarchy within the colonies of D. lucida has not yet been clearly determined by researchers (Peixoto et al. 2008).
Many arthropods live in D. lucida nests, including a species of Pheidole Westwood, 1839 which occurs only in this association (Paiva & Brandão 1995;Peixoto et al. 2010). Buys et al. (2010) recorded a wasp of the genus Kapala Cameron, 1884 (Eucharitidae) parasitizing a pupa of D. lucida, which had about 70 to 90% of the body consumed. Distribution (Fig. 29A) Dinoponera lucida is endemic to the Atlantic Forest biome of eastern Brazil, occurring throughout Espírito Santo State, south of Bahia, the eastern boundary of Minas Gerais and north of São Paulo. The northernmost record for this study is Jacobina, Bahia, and the southernmost is in Cruzeiro, São Paulo dating from 1933. This is the only known record for São Paulo and could imply that D. lucida had a wider distribution or that this is a labeling error. The state of Rio de Janeiro, which is between the states of São Paulo and Espírito Santo, does not have records of Dinoponera. A specimen was received with a label from Rio das Ostras, Rio de Janeiro, but it is a probable labeling error for two reasons: (1) no published records of Dinoponera were found for the state, despite being a heavily collected region for insects in general; and (2) before the fire at the National Museum of Rio de Janeiro, consultation with researchers there found that the collection of ants did not have any Dinoponera from the state. For these reasons, this record was not included in the distribution map. Emery, 1901 Figs 23, 28B
Head. Malar area punctulate and weakly striate; striae longitudinal to oblique, usually not reaching anterior eye margin. Gena smooth (rarely weakly microareolate) with bluish iridescence; usually transversely rugulose in median area. Appressed brownish pubescence present between eye and frontal lobe, extending posteriorly to frons. Frons smooth, with bluish iridescence; with flexuous, brownish, suberect to erect hairs, longer than scape diameter; pubescence densest laterally. Occipital corner smooth and with bluish iridescence. Antennal scape weakly microareolate and silky; antenna with long decumbent to suberect hairs, except for 5-6 apical segments. Ventral surface of head mostly smooth and silky; longitudinal strigulae present on anterolateral cephalic margin, including hypostomal tooth. Labrum usually without median longitudinal sulcus; transverse rugulae absent. Mandibular dorsum with well-marked longitudinal strigulae on inner base that gradually fade apicad. Mesosoma. Dorsal margin of pronotum in lateral view convex, usually with pronounced dorsoposterior swelling; anteroventral corner of pronotum forming obtuse angle, never toothed. Pronotal dorsum smooth, with bluish iridescence. Metapleural-propodeal suture well-marked, straight to slightly sinuous.
Metasoma. Petiolar node in lateral view elongate (DPI < 0.8); anterodorsal corner at same level or slightly lower than posterodorsal corner; anterior, dorsal and posterior margins straight; antero and posterodorsal corners forming blunt angle. Node lateral face smooth with bluish iridescence. Node anterior posterior and lateral margins in dorsal view broadly convex. Abdominal tergite III smooth with bluish iridescence; punctulae denser laterally than dorsally; covered by flexuous, brownish, decumbent to suberect hairs on all surfaces; brownish appressed pubescence sparse on dorsum, denser laterally.

Male
Unknown.

Remarks
Dinoponera mutica may resemble the sympatric D. nicinha sp. nov. The most useful diagnostic traits are the golden pilosity, the sparse pubescence on the dorsum of abdominal tergite III and the petiolar node in lateral view with its convex dorsal margin in D. nicinha sp. nov. Dinoponera grandis may also occur in sympatry, but has DPI > 0.8, while D. mutica usually has DPI < 0.8. Besides that, D. mutica has no visible microsculpturing on abdominal tergite III, unlike D. grandis. The petiolar node of D. mutica in lateral view can present an anterodorsal angle slightly lower than the posterodorsal angle, sometimes resembling D. lucida. However, D. lucida has an anteroventral pronotal tooth and strigulae on the cephalic ventral surface. Dinoponera quadriceps and D. gigantea are different on account of their predominantly microareolate and opaque integument.

Biology
Little is known about the habits of D. mutica. Although apparently occurring in sympatry with D. grandis, a species more common in savanna environments, Meurer et al. (2015) suggested that D. mutica may be restricted to forest environments.

Distribution (Fig. 28B)
This species is distributed in the midwest and part of northern Brazil, as well as Bolivia and Paraguay. The northernmost record is in Jacareacanga, Pará, Brazil, and

Diagnosis
Female Anteroventral corner of pronotum in lateral view forming obtuse angle. Petiolar node in lateral view with convex dorsal margin, anterodorsal corner at same level as posterodorsal corner. Abdominal tergite III smooth and with bluish iridescence, covered with sparse golden pubescence. Body covered by golden hairs.

Material examined
Head. Malar area with weak longitudinal to oblique striae that do not reach anterior eye margin. Gena smooth, with bluish iridescence; usually without rugulae. Golden appressed pubescence present between eye and frontal lobe, extending posteriorly to frons. Frons smooth, with bluish iridescence; with flexuous, golden, decumbent to suberect hairs, longer than scape diameter; pubescence densest laterally. Occipital corner smooth and with bluish iridescence. Antennal scape weakly microareolate and silky; antenna with decumbent to suberect hairs, except for 5-6 apical segments. Ventral surface of head mostly smooth, with bluish iridescence; longitudinal strigulae present on anterolateral cephalic margin, including hypostomal tooth. Labrum without median longitudinal sulcus, transverse rugulae absent. Mandibular dorsum weakly longitudinally strigulate on inner base, sculpture gradually fading apicad.
Metasoma. Petiolar node in lateral view elongate (DPI < 0.8), antero-and posterodorsal corners at same level, anterior and posterior margins broadly convex, dorsal margin convex, antero-and posterodorsal corners rounded. Node lateral face smooth, with bluish iridescence. Node lateral margin in dorsal view broadly convex, anterior and posterior margins convex. Abdominal tergite III mostly smooth, with bluish iridescence; laterally and dorsally with evenly distributed punctulae; densely covered by golden, flexuous, suberect hairs and sparse pubescence evenly distributed on dorsal and lateral surfaces.

Male
Unknown.

Remarks
The main difference between D. nicinha sp. nov. and other species of Dinoponera is in the shape of the petiolar node. This character has proven to be useful for separating species in this genus, as its shape is generally constant within the species. Specimens of this species were found in collections identified as D. longipes, but besides the node shape they can also be separated by the ratio of the length of the hind basitarsus to the head width, which in D. longipes is greater (always > 1.3 in D. longipes and < 1.3 in D. nicinha sp. nov.). Dinoponera nicinha sp. nov. can also be confused with the sympatric D. mutica, but the pilosity is not golden in D. mutica and abdominal tergite III has dense pubescence on its side (sparse in Dinoponera nicinha sp. nov.). The node in D. mutica has the anterior and dorsal margins in lateral view meeting at a blunt angle, slightly lower than the posterodorsal angle. Dinoponera nicinha sp. nov. has rounded antero-and posterodorsal node angles, both at the same level.

Biology
Dinoponera nicinha sp. nov. occurs in forest environments with high rainfall and dense coverage of medium to large trees, with a maximum of six months of drought during the year (IBGE 2012). Its natural history is unknown, except for the fact that most specimens were collected in pitfall soil traps.

Diagnosis
Female Malar area without striae or with weak longitudinal to oblique striae that do not reach anterior eye margin. Ventral head surface without strigulae. Head with short to inconspicuous pubescence. Anteroventral corner of pronotum in lateral view forms an obtuse angle. Abdominal tergite III microareolate and opaque, with sparse pubescence on dorsum.

Male
Antenna with long suberect to erect hairs, clearly longer than maximum scape diameter. Mesoscutum without notaulus. Gastral tergites without long hairs. In lateral view, dorsal margin of basal ring broadly concave to straight. Mesosoma. Dorsal margin of pronotum in lateral view broadly convex, with no pronounced dorsoposterior swelling; anteroventral corner of pronotum forms obtuse angle. Pronotal dorsum weakly microareolate, opaque to silky. Metapleural-propodeal suture well-marked and usually straight.

Material examined
Metasoma. Petiolar node in lateral view elongate (usually DPI < 0.8); anterodorsal corner usually higher than posterodorsal corner; anterior margin slightly concave to straight, forming blunt angle with dorsal margin; dorsal and posterior margins broadly convex and forming rounded angle. Node lateral face microareolate and usually opaque, rarely silky. Node anterior margin in dorsal view strongly convex, posterior margin broadly convex to straight, lateral margins broadly convex and converging anterodorsally. Abdominal tergite III microareolate and opaque; punctulae in variable density, usually denser laterally than dorsally; covered by brownish, flexuous, suberect hairs on entire surface; appressed pubescence usually sparse on dorsum, denser laterally.

Male
Measurements. Head. Frontal carina usually forming short longitudinal swelling. Lateral ocellus clearly surpassing posterior head margin in full-face view. Head punctulate, weakly microareolate and with silky sheen; with yellowish decumbent pubescence and scarce long suberect to erect hairs, longer than ocellus height in full-face view. Antenna with appressed pubescence and long suberect to erect hairs clearly longer than maximum scape diameter; third to fourth most apical antennomeres without hairs. Ventral surface of head punctulate and slightly microareolate; with subopaque to silky sheen. Mesosoma. Mesoscutum without notaulus. Mesopleural sulcus punctate or slightly scrobiculate. Scutoscutellar sulcus scrobiculate. Mesoscutellum longitudinally strigulate laterally. Metapleuralpropodeal suture with same microsculpture as rest of integument. Mesosoma mostly slightly microareolate and with silky sheen, becoming coarsely punctate on declivitous surface of propodeum; with decumbent to suberect pubescence and suberect hairs, distance between each hair usually greater than half its length. Legs densely covered by decumbent pubescence; coxae with suberect long hairs, greater than femur diameter. Protibial apex with at least one stout seta.
Genitalia. Basal ring in dorsal view with slightly concave lateral margins, anteriorly slightly narrower than posteriorly; fenestra rounded to slightly transversally directed; median invagination U-shaped; dorsal margin in lateral view broadly concave to straight; anteroventral process subquadrate. Gonostylus narrow and sharp. Dorsal margin of volsella in lateral view anteriorly straight and posteriorly concave; anteroventral corner with two anterior teeth; posteroventral margin strongly concave and forming sharp triangular posterior lobe; digitus volsellaris with posterior margin straight or with a slight concavity. In lateral view, penisvalva with continuous dorsal posterior margins, ending in a subacute apex; ventral margin even and serrated, with slight concavity; anteroventral corner sharply pointed.
Color. Body mostly chestnut brown.

Remarks
Females of Dinoponera quadriceps have a robust set of characters that usually permit easy separation from those of other species of the genus. However, D. quadriceps and D. gigantea, geographically close species, can sometimes cause confusion. The main characters dividing the two species are: (1) in D. quadriceps the anteroventral pronotal margin forms an obtuse angle, but in D. gigantea it is toothed or forms an acute angle; (2) the metapleural-propodeal suture in D. quadriceps is well-marked and straight, but in D. gigantea it is weaker and sinuous; (3) the cephalic venter is mostly strigulate in D. gigantea but in D. quadriceps the strigulae are absent or sometimes present just anterolaterally; (4) the dorsal region of abdominal tergite III in D. quadriceps is sparsely punctulate, but densely so in D. gigantea; and (5) the anterodorsal corner of the petiolar node is higher in D. quadriceps than in D gigantea. Some specimens of D. quadriceps can present one or two D. gigantea characters, and viceversa, so the use of all these characters affords the most reliable identifications. It is worth noting that the difficult cases are more common among specimens from Maranhão and Tocantins, as in D. gigantea. Notwithstanding these complications, we consider the aforementioned set of diagnostic characters to provide stronger support for maintaining D. quadriceps and D gigantea as separate species than for synonymizing them. The alternative of describing a new species from Tocantins and Maranhão is not possible because the characters seem to vary randomly. Determining the degree of expression of each character is also important. Species differences are frequently a degree of difference one from another, as in the case of the anteroventral pronotal corner in D. quadriceps, which may be strongly angular but not toothed or forming an acute angle.
A specimen from Itabaianinha (Sergipe) has a short and robust peciolar node (DPI > 0.8), characteristic of D. grandis. However, these two species differ in the absence of: (1) the pronotal anteroventral tooth; (2) striae in the malar area and (3) strigulae on the ventral head surface in D. quadriceps. The other species of Dinoponera can be easily distinguished by having abdominal tergite III smooth and shining.
It was not possible to examine the holotype of D. opaca. It is kept as a junior synonym of D. quadriceps, on account of the examination of the holotype by C. Baroni Urbani in 1974 at the request of W. Kempf (1975: 344). Dr Baroni Urbani found that the type agreed with the description of D. quadriceps. Lenhart et al. (2013) examined the type and found total coincidence with D. quadriceps. The Rio de Janeiro record for D. opaca is undoubtedly a labelling error, as there are no other records of Dinoponera for Rio de Janeiro, a heavily collected area for ants, and if there were any Dinoponera, the expected species would either be D. lucida or D. australis.
Males of D. quadriceps are different from those of D. lucida, D. longipes, Dinoponera morphospecies 1 and D. grandis by having hairs on the antennae that are clearly longer than the scape diameter and the gonostylus narrow, with a sharp apex. As mentioned before, males of D. gigantea and D. quadriceps are similar in many characters, but the former is longer, more pilose and has a visible notaulus on the mesoscutum.

Biology
The species occurs mainly in arid or savanna areas with the temperature of the coldest month greater than 18 °C and the precipitation of the driest month less than 60 mm (Alvares et al. 2013). Nests are usually built in the ground, near roots and often in direct contact with termite mounds (Araujo et al. 1990;Paiva & Brandão 1995). However, Vasconcellos et al. (2004) found nests more than 3 m from trees in a remnant of Atlantic Forest, indicating nesting habits may vary with the environment. Each nest can have two to ten entrances surrounded by branches and some loose earth (Araujo et al. 1990;Paiva & Brandão 1995). Nest depth varies from 10 to 120 cm and the number of chambers can reach 16 (Araujo et al. 1990;Paiva & Brandão 1995;Vasconcellos et al. 2004). Assis et al. (2017) recorded D. quadriceps reusing nests of Atta sexdens (Linnaeus, 1758), which are deep enough to ensure protection against hot and dry weather. The density of nests varies from 15 to 40 per hectare (Vasconcellos et al. 2004).
The size of the colony can vary from 12 to 141 workers (Paiva & Brandão 1995;Monnin & Peeters 1997, 1998Vasconcellos et al. 2004;Nascimento et al. 2012;Asher et al. 2013). Vasconcellos et al. (2004) found males almost every month of the year, suggesting male production is not seasonal.
Their diet is omnivorous, with a preference for arthropods found dead on the soil (Araújo & Rodrigues 2006). Foraging is solitary and occurs during the day, avoiding the hottest hours (Medeiros et al. 2012). The size of food items varies, but regardless of its weight there is no recruitment of other workers (Araújo & Rodrigues 2006). Before finding food, the worker is usually slow and meanders about its usual route, moving up to 35 m from the nest. Foraging tends to be faster during the return trip and follows a relatively straight line (Araújo & Rodrigues 2006;Azevedo et al. 2014). Araújo & Rodrigues (2006) suggested that spatial orientation during foraging may follow chemical and visual cues. (2014) divided the life cycle of the colony into stages. The first stage is monodomic, with the ants occupying a single nest. Subsequently the colony becomes polydomic, until gradually the ants stop visiting the parental nest and the colonies become isolated, completing the fission. The newly separated colonies go through a period of territorial conflict that results in the extinction or migration of the old colony . Vasconcellos et al. (2004) also pointed out that intraspecific territorial competition may be one of the main factors responsible for the spatial arrangement of nests.

Medeiros & Araújo
Colonies of D. quadriceps are predominantly monogynous (Monnin & Peeters 1997, 1998, 1999, though Araujo et al. (1990) found colonies with more than one gamergate. The dynamics that determine the hierarchy of dominance in this species is well studied, mainly by Monnin & Peeters (1999). Hierarchical position is determined by age and agonistic interactions involving mainly the ants with high status in the colony (Monnin & Peeters 1999). Newly emerged workers usually replace older ones in the highest positions of the hierarchy. Agonistic interactions are usually related to the stability of monogyny in the colony, and can be performed by the gamergate itself, inhibiting other workers (Monnin & Peeters 1999). After copulation, gamergates chew away the end of the male abdomen, which remains trapped in the female genitalia, preventing further copulation (Monnin & Peeters 1998). Distribution (Fig. 29B) Dinoponera quadriceps is endemic to northeastern Brazil, occurring predominantly in the Caatinga. The northernmost record is in Fortaleza, Ceará and the southernmost record is in Viçosa, Minas Gerais. However, Viçosa is more than 500 km away from the next southernmost record (Janaúba, Minas Gerais), leaving doubts as to the reliability of the former record. Another unexpected record is from Brasília, which is also more than 500 km distant from any other record. These records may reflect a wider distribution in a recent past (both date from the 1970s) or just labeling errors. Lenhart et al. (2013) cited records of D. quadriceps in the cities of Santarém and Óbidos, as well as along the Tapajós River, all in the state of Pará. However, we found no records of this species within at least 1000 km of that region. This discrepancy may be due to either wrong labeling or differences in the criteria used for species identification in each study. The latter is most probably the case, given the similarity between D. quadriceps and D. gigantea.

Preliminary considerations
The morphology of this single male is described because it is significantly different from that of all other known males of Dinoponera. However, the lack of an associated female does not allow us to identify this specimen. This male may belong to either one of the sympatric species, D. mutica or D. nicinha sp. nov., both of which do not have described males. Given the present impossibility of associating this specimen with an available name, we present it here with the intention of increasing our knowledge about Dinoponera males, but avoiding a wrong association. We also feel compelled to present it due to the recent significantly increased rate of deforestation and bush fires that have ravaged vast swathes of potential distribution range for Dinoponera, including the species represented by this male. In view of this calamity, it is not unreasonable to expect dwindling opportunities to collect additional specimens of Dinoponera in many parts of the Brazilian Amazon. This male is evidence of the wonderful biodiversity that is literally turning into ashes before our very eyes, and we have no certainty that more specimens will come to light.

Diagnosis
Male Antenna with thick and stiff decumbent hairs, as long as or slightly longer than maximum scape diameter. Abdominal tergite VIII spiniform. Penisvalva ending in a rounded apex, without a ventral lobe. Anteroventral corner of penisvalva in lateral view with long and very sharp spine. Head. Frontal carina forming short longitudinal swelling. Lateral ocellus clearly surpassing posterior head margin in dorsal view. Head punctulate, weakly microareolate and shining; with yellowish suberect pubescence and long decumbent to suberect hairs, longer than ocellus height in dorsal view. Antenna with appressed pubescence and thick, stiff decumbent hairs as long as or slightly longer than maximum scape diameter. Ventral surface of head punctulate and slightly microareolate with silky sheen.
Genitalia. Basal ring in dorsal view with slightly concave lateral margins, anteriorly much narrower than posteriorly; maximum diameter of fenestra longitudinally directed, median invagination V-shaped; dorsal margin of ring in lateral view anteriorly concave, anteroventral process subquadrate. Gonostylus broad and rounded. Dorsal margin of volsella in lateral view straight; anteroventral corner rounded; posteroventral margin strongly concave and forming a posterior triangular lobe; digitus volsellaris with posterior margin with a slight concavity. In lateral view, penisvalva with continuous dorsal and posterior margins, ending in rounded apex; ventral margin even and serrated, concave; anteroventral corner with long, very sharp spine.
Color. Body mostly brown, with head and gaster apex slightly clear.

Remarks
Dinoponera morphospecies 1 is different mainly because of the shape of the anteroventral corner of the penisvalva, which ends in a long, sharp spine. It also differs from D. gigantea and D. quadriceps by the antennal hairs, which are shorter and thicker. An additional character that distinguishes this male from D. lucida and the D. grandis is the concavity of the basal ring in lateral view, much weaker in Dinoponera morphospecies 1. In D. longipes the penisvalva is wedge-shaped, not rounded, in lateral view, and the spine is shorter (observations made from the illustrations and description in Lenhart et al. 2013). Hopefully, additional specimens will be collected in the future, permitting the association of this male with its conspecific females. This could be done directly by collecting nest series of both sexes of the same species, or indirectly by collecting those of another species, making apparent which species it does not belong to.

Biology
Unknown.

Discussion
We have redescribed all known species of Dinoponera and described a new species using the largest amount yet available of specimens for a study of this group, particularly from Brazilian institutional collections. Due to the large size of these ants, a trait that would make them attractive for even generalist insect collections, we expected to find more of these ants in collections. Especially surprising and worrying is the paucity of recent collecting events for several of these species.
It was possible to note that males are morphologically more variable than females. In D. grandis, males seem to be discreetly variable, in sharp contrast with the continuous variability of females. This offers the hope of permitting less challenging identifications than when using the females, but in practice this is cold comfort as males are harder to come by, and the apparent greater discreteness of male characters may be an artefact of the few males examined.
While we believe that we have been able to significantly improve the alpha taxonomy of these species, there is still much to be done. The application of molecular methods could give us valuable insights into the phylogenetic relationships amongst the different lineages and provide additional evidence for taxonomic decisions. During the course of this study we unsuccessfully tried to extract and amplify DNA from several specimens. Given the paucity of museum specimens, particularly from recent collecting events, we advocate for field work specifically targeting the known populations and searching localities where some populations may still remain; this is especially urgent for D. grandis. The long history of deforestation, fragmentation and conversion of natural landscapes into agricultural landscapes has probably complicated matters even further, both for the survival of these ants and the opportunities for studying them. Particularly worrying is the dramatic increase in bush fires in recent times. Even though one species, D. lucida, figures in the Brazilian Red list of endangered species (ICMBio 2018), it is possible that other species and populations, such as D. hispida, are arguably in even more dire straits.  (Perty, 1833) and D. grandis Guérin-Méneville, 1838. B. D. hispida Lenhart, Dash & Mackay, 2013, D. mutica Emery, 1901 and Dinoponera morphospecies 1.  Emery, 1901and D. lucida Emery, 1901. B. D. nicinha sp. nov. and D. quadriceps Kempf, 1971.