Dorylus

The Afrotropical ‘driver ants’ of this genus epitomize the army ant lifestyle, but they represent only a fraction of the diversity of Dorylus. Most species are much less commonly observed, and forage underground or in leaf litter. Hita Garcia, Wiesel and Fischer (2013) - The army ant genus Dorylus is mostly known for the spectacular swarm raids performed by some epigaeic species, mostly belonging to the subgenus Anomma, better known as "driver ants". These species perform huge swarm raids along the ground and lower vegetation with hundreds of thousands of blind, polymorphic workers to hunt down a great variety of prey taxa in large quantities, predominantly invertebrates (Gotwald, 1982, 1995). However, many more species within the genus live and hunt hypogaeicly and these army ants are generally less visible than their epigaeic relatives (Berghoff et al., 2002). Hypogaeic species hunt in columns and many species are known to be specialised predators of other social insects, such as termites or other ants (Darlington, 1985; Gotwald, 1982, 1995; Schöning & Moffett, 2007). Almost all species of Dorylus, like other army ant genera, live in monogynous colonies with dichthadiiform queens that have a massive egg-laying capacity, e.g. three to four million eggs per month in “driver ant” queens (Raignier & van Boven, 1955). In addition, Dorylus colonies migrate in irregular intervals to new nesting sites and new colonies emerge through colony fission (Gotwald, 1982, 1995).

Identification
Borowiec (2016) - Worker The workers of Dorylus are readily recognized by a combination of well-developed promesonotal suture, propodeal spiracle positioned high on the propodeum and lack of propodeal lobes, single waist segment, pygidium large and with a flattened surface and armed with two cuticular projections, and pretarsal claws simple. Other army ants of the Old World, Aenictus and Aenictogiton, are not easily confused with Dorylus as the former always has a well-differentiated second waist segment (postpetiole) and in Aenictogiton the gaster has more developed constrictions between gastral pre- and post-sclerites, resulting in apparent constriction between abdominal segments IV, V, and VI. Yunodorylus is superficially similar but is easily distinguished from all army ants by the propodeal spiracle situated low and presence of propodeal lobes. Among the New World army ants only Cheliomyrmex has one-segmented waist but Cheliomyrmex does not have a promesonotal suture, its pygidium is reduced and never armed with cuticular projections, and its pretarsal claws are armed with a tooth.

Male In general appearance Dorylus males are similar to other army ant genera but possess flattened femora that are much broader and more compressed than the tibiae and tarsi. This trait alone is sufficient to separate them from all other male dorylines, but a combination of single-segmented waist, M·f1 vein of fore wing arising from M+Cu at about 45° and situated near to cu-a, Rs·f2–3 lost, pterostigma narrow and inconspicuous can also be used to recognize Dorylus. The Old World army ant genera Aenictus and Aenictogiton have similar fore wing venation but both have a well-developed and broad pterostigma and the latter has a ‘free-hanging’ Rs·f3 vein. In the New World army ants M·f1 arises at a lower angle and is conspicuously proximal to cu-a, and Rs·f2–3 are present, forming two submarginal cells. Dorylus males also possess unique genital capsule morphology, where a tiny diamond-shaped structure is formed from a fragment of the basimeres and visible dorsally over the aedeagus (‘patella’ of Birket-Smith 1981; Brendon Boudinot pers. comm.). The telomeres in lateral view do not conceal inner valves of the genital capsule as in most dorylines but instead form a characteristic shape of a spiral arm folding first proximally and then projecting distally over the rest of genital capsule thus concealing it from above.

The subgeneric system within this genus is abandoned by Borowiec (2016): "Subgeneric classification is not currently adopted for any other doryline genus, and I propose informal species-groups to be recognized instead of the subgenera (for species known from the worker caste)." These species groups and their respective member species are listed here: Dorylus species groups. The subgenera key to workers, linked to below, includes both the subgeneric names and the species-groups they represent.

Garcia, Wiesel and Fischer (2013) - The taxonomic condition of Dorylus, especially for the African continent, can be classified as chaotic and useless for identification purposes. On a global basis, 59 species and 68 subspecies are recognised (Bolton, 2012), although the taxonomic validity of many of these taxa is highly questionable. The problem is that most descriptions were based on a single caste, and careful examination of taxa in order to find evidence for conspecificity among these is very rare (Schöning et al., 2008). Also, no modern taxonomic revision is available, which dramatically increases the difficulties to identify Dorylus to species level. Nevertheless, identification to subgenus level can be well performed with the keys provided in Gotwald (1982) (and see the previous paragraph).

Afrotropical species
List of Afrotropical species by caste, with images for the worker caste

Distribution
‘’Dorylus’’ ranges from Sub-Saharan Africa throughout North Africa and Asia Minor to Borneo in Southeast Asia. The Afrotropics harbor the highest number of species and are the home of the surface- and leaf litter-foraging species. (Borowiec 2016)

Biology
Borowiec (2016) - Because some species of this lineage are so conspicuous and are the most important arthropod predators of the Afrotropics, this group has attracted considerable attention.

The best studied species include the Afrotropical species that forage above ground (Raignier and Boven 1955, Raignier 1972, Gotwald 1995), but one subterranean species, Dorylus laevigatus has been the subject of some work (Berghoff et al. 2002a,b, 2003a,b, Weissflog et al. 2000). Good overviews of Dorylus biology can be found in Raignier and Boven (1955) and Gotwald (1995). The surface- and leaf litter-foraging species have been collectively referred to as ‘driver ants’ (Savage 1847), and traditionally classified in the polyphyletic subgenus Anomma (see Taxonomy and phylogeny above). Here I follow this convention and use the terms ‘driver ants’ and ‘surface-‘ or ’epigaeically-foraging species’ interchangeably.

The life cycle of Dorylus colony is similar to that of Eciton and many other army ants but there are no pronounced nomadic and statary phases. The brood production is not synchronized (Gotwald 1995, Schöning et al. 2005b), and the colonies move from old to new nesting sites at irregular intervals (Gotwald and Cunningham van Sommeren 1990, Schöning et al. 2005b). A mature colony will produce about a dozen virgin queens and eventually undergo fission. About half of the worker force will depart with the old, fertilized queen, while the other half will remain with the virgin queens. Ultimately, all except one of the new queens are cannibalized (Raignier 1972). The new colony does not produce sexual brood until the workers mothered by the old queen have died (Kronauer et al. 2004).

Reproduction
Copulation in Dorylus has been observed only once (Kronauer and Boomsma 2007a). Males collected at lights and two inseminated queens from established Dorylus moestus bivouacs were coupled under laboratory conditions. The male first uses his sickle-shaped mandibles to grasp the queen behind her petiole and performs bending movements, searching the tip of the queen’s abdomen. Once engaged, the pairs remained in copulation for five to ten hours. After this period, the male relaxes his grip on queen’s petiole but remains connected to the queen. Twenty hours after the copulations, the two pairs were killed and dissected, both males remaining attached to the queens. The males apparently succeeded in transferring sperm to the queens, and the dissections confirmed that the male accessory testes were empty after the copulations. Despite these observations, Kronauer and Boomsma (2007a) find little evidence for army ant queens re-mating later in life and point out that the males were not attracted to old queens in most trials.

The reproductive potential of Dorylus queens is impressive, at least in the surface-foraging species studied thus far. The queen mates between 15–20 times (Kronauer et al. 2004, 2006) early in her life and stores up to 880 million spermatozoa (Kronauer and Boomsma 2007a). A Dorylus wilverthi queen can produce an estimated 3–4 million eggs per month, for a total over 250 million eggs during her lifetime (Raignier and Boven 1955, Kronauer and Boomsma 2007b). This is even more than Eciton queens (see under Eciton; Schneirla 1971, Kronauer and Boomsma 2007b).

These army ants always occupy subterranean nests, either constructed by excavating large amounts of soil and/or taking advantage of a preexisting cavity (Schöning et al. 2005b, Boven and Lévieux 1968). Because of these underground habits, colony size estimates are rare. A single excavated colony of Dorylus laevigatus contained about 300,000 workers (Berghoff et al. 2002), and estimates of colony size for the surface foragers Dorylus nigricans and Dorylus wilverthi range from 1 million to over 20 million workers (Voessler 1905, Raignier and Boven 1955). The dry mass of Dorylus nigricans colonies has been estimated to be 9–15 kg (Leroux 1982). The underground nests of Dorylus are quite different from the above-ground bivouacs of Eciton (Gotwald 1995). Raignier and Boven (1955) categorized them as either occupying a single large chamber or dispersed among subterranean galleries and chambers. The first type is exemplified by D. wilverthi and the second by D. nigricans. Both nest types are often found among root systems of trees. These ants actively excavate soil and one estimate gives 20 kg of soil per day removed in the first week of a D. nigricans colony settling into a new site (Leroux 1977).

Dorylus emigrate irregularly and the colony often returns to the same nesting spot. Gotwald and Cunningham van Sommeren (1990) followed a single colony of Dorylus moestus for 432 days and observed 38 emigrations during that time, spanning an area of about 5 hectares. One colony of D. nigricans has been recorded to remain in one bivouac site for 125 consecutive days (Raignier and Boven 1955). The adaptive significance of the cycles in brood production and colony activity remains unclear, but it phasic species of Aenictus, Eciton, and other New World army ants rely heavily on brood of other social insects, Dorylus are more generalist (Gotwald 1995).

Dorylus gynes may or may not be able to move on their own during nest emigration. All queen specimens known so far are missing tarsal segments (Raignier 1972, Berghoff et al. 2002), so that they are assisted to a new site by the entourage of workers (Berghoff et al. 2002). Raignier (1972) observed missing tarsal segments in very young queens of D. nigricans, prior to their first emigration. The causes and significance of this tarsal mutilation are not known.

Foraging/Diet
A diversity of foraging habits and prey preferences has been documented for Dorylus (Gotwald 1995). According to the most popular classification (Schöning et al. 2005a, Kronauer et al. 2007), three major foraging strategies can be distinguished: subterranean, leaf litter, or surface foragers. The surface-swarming driver ants are generalist predators that will take any kind of prey, ranging from immatures of other insects to vertebrate carrion (Schöning and Moffett 2007). Seasonal, habitat, and intraspecific differences can be seen in prey composition and intake in these ants, but the proportion of social insect prey is small (Schöning et al. 2008). This is in contrast to Eciton burchellii, whose diet is general but it still relies heavily on this kind of prey (Rettenmeyer 1963). The few subterranean species of Dorylys that have been studied have also been recorded to be generalist predators but additionally often feeding on termites (Darlington 1985, Berghoff et al. 2002). Dorylus orientalis is recognized as a vegetable crop pest, apparently being mainly or exclusively herbivorous (Roonwal 1975). Variation in foraging can also be seen within the general foraging strategies. Schöning et al. (2008) reported that two surface-swarming species, Dorylus wilverthi and D. moestus, differ in their diets and raiding behavior. Dorylus moestus is often seen capturing earthworms and exhibits digging behavior, while earthworms are rarely a major component of the diet for D. wilverthi, whose workers have not been observed digging. Two sympatric, subterranean species of Dorylus from Asia have also been compared and shown to differ in their foraging behavior and prey preference (Berghoff et al. 2003).

Kronauer et al. (2007) used molecular phylogenetics and ancestral state reconstruction to address the evolution of the foraging niche in Dorylus. They categorized species as either subterranean, leaf litter, or surface foragers and inferred that subterranean foraging was the ancestral state for the genus. Both surface and leaf litter foraging strategies likely evolved once within Dorylus. The descendants of a leaf litter-dwelling ancestor gave rise to both surface foragers and species that reverted to subterranean foraging. An earlier study (Schöning et al. 2005a) showed how allometry in the worker caste is correlated with the foraging niche, although the authors did not examine this in a phylogenetic framework (Felsenstein 1985). These authors demonstrated that surface-adapted species possess appendages and mandibles that are longer relatively to their body size than in the leaf-litter and in the subterranean foragers. Kronauer et al. (2007) further assessed allometry in the context of Dorylus phylogeny and concluded that the species that reverted to underground foraging re-evolved morphology similar to the ancestral, short-limbed condition.

Associations with other Organisms
Similarly to New World army ants, Dorylus colonies have numerous invertebrate and vertebrate associates, although these companion faunas are not as well described (Gotwald 1995). Remarkably, the foragers of African driver ants are followed by several species of birds specializing on prey flushed by the ants, much like the swarms of Eciton burchellii in the New World (Peters et al. 2008). Other vertebrates, such as chimpanzees are known to rely on Dorylus for food (Kingdon 1997, Schöning et al. 2007, Sanz et al. 2010). Because the apes utilize sticks and plant stems to extract the ants, this is an important study system in the primate culture and tool use (Humle 2011).

A variety of other research has been carried out on Dorylus, but most of these studies are isolated in nature. Kronauer et al. (2011) documented significant amounts of hybridization between the driver ants D. wilverthi and D. moestus, and Barth et al. (2013) undertook a population genetics study on Dorylus fulvus.

Nomenclature

 *  DORYLUS [Dorylinae]
 * Dorylus Fabricius, 1793: 365. Type-species: Vespa helvola, by monotypy.

Borowiec (2016) - The long and confusing taxonomic history of the genus begins with a male ant from South Africa, described as Vespa helvola by Linnaeus in 1764. Later Fabricius (1793) created the genus Dorylus for that species. Similarly to Aenictus, for a time the males and females were known under different generic names, with Dorylus being applied to males and Anomma and Typhlopone to the workers. 85 years after the original description of Vespa helvola, T. S. Savage observed males and workers together in the field and recognized that they belonged to one species (Savage 1849). A very readable overview of the early taxonomic history of Dorylus can be found in Gotwald (1995:13). The modern subgeneric division of Dorylus was stabilized by Emery (1895b, 1910). This classification has come under scrutiny using molecular data in the recent decades, and two of the most speciose subgenera of Dorylus, Anomma and Dorylus s. str. were found to be not monophyletic (Kronauer et al. 2007). Because of these phylogenetic considerations, also backed up by morphological study (Caspar Schöning pers. comm.), I propose to abandon the traditional subgeneric classification. Although the surface-foraging (as opposed to leaf litter) species of Anomma species form a clade and it is even possible to differentiate it based on apomorphic morphological characters from other Dorylus (Schöning et al. in preparation), recognizing Anomma would likely leave the large Dorylus s. str. paraphyletic. Other Dorylus subgenera are likely monophyletic (Kronauer et al. 2007). Subgeneric classification is not currently adopted for any other doryline genus, and I propose informal species-groups to be recognized instead of the subgenera (for species known from the worker caste). Dorylus species groups

Dorylus is the sister taxon to Aenictogiton (Brady et al. 2006, 2014, Borowiec, in prep.). As explained above, the internal phylogeny of the genus (Kronauer et al. 2007) shows that the subgenera Anomma and Dorylus as they were traditionally defined are not monophyletic. The Asian species Dorylus laevigatus represents the earliest-branching lineage of the genus. The time-calibrated phylogeny of Kronauer et al. (2007) estimated crown group age of Dorylus to be between 30 and 64 million years, but more recent studies suggest much younger ages at about 22 million years (Brady et al. 2014) or even younger than 20 million years (Borowiec, in prep.).

Worker
Borowiec (2016) - Head: Antennae with 8, 9, 11, or 12 segments. Apical antennal segment not enlarged, not broader and longer than two preceding segments combined. Clypeus without cuticular apron. Lateroclypeal teeth absent. Parafrontal ridges absent. Torulo-posttorular complex vertical. Antennal scrobes absent. Labrum without median notch or concavity. Labrum with median notch or concavity. Proximal face of stipes projecting beyond inner margin of sclerite, concealing prementum when mouthparts fully closed. Maxillary palps 2- or 1-segmented. Labial palps 2-segmented. Mandibles elongately triangular to falcate, with teeth on elongated masticatory margin. Eyes absent. Ocelli absent. Head capsule with differentiated vertical posterior surface above occipital foramen. Ventrolateral margins of head without lamella or ridge extending towards mandibles and beyond carina surrounding occipital foramen. Posterior head corners dorsolaterally immarginate. Carina surrounding occipital foramen entirely absent, including ventrally. Mesosoma: Pronotal flange not separated from collar by distinct ridge. Promesonotal connection with suture conspicuous and complete, but immobile. Pronotomesopleural suture complete, continuous with promesonotal suture. Mesometapleural groove deeply impressed, conspicuous. Transverse groove dividing mesopleuron absent. Pleural endophragmal pit concavity present. Mesosoma dorsolaterally immarginate. Metanotal depression or groove on mesosoma absent. Propodeal spiracle situated high on sclerite. Propodeal declivity without distinct dorsal edge or margin and rectangular in posterior view. Metapleural gland with bulla visible through cuticle in smaller workers, mostly obscured in large workers. Propodeal lobes absent. Metasoma: Petiole anterodorsally immarginate, dorsolaterally immarginate, and laterally above spiracle immarginate. Helcium in relation to tergosternal suture placed at posttergite and axial. Prora forming a simple U-shaped margin. Spiracle openings of abdominal segments IV–VI circular. Abdominal segment III anterodorsally immarginate and dorsolaterally immarginate. Abdominal segment III more than half size of succeeding segment IV, which is weakly constricted at presegmental portion (uninodal waist). Girdling constriction of segment IV present, i.e. pre- and postsclerites distinct. Cinctus of abdominal segment IV gutter-like and sculptured but not cross-ribbed. Abdominal segment IV not conspicuously largest segment. Abdominal tergite IV not folding over sternite, and anterior portions of sternite and tergite equally well visible in lateral view. Girdling constriction between pre- and posttergites of abdominal segments V and VI present. Girdling constriction between pre- and poststernites of abdominal segments V and VI present. Pygidium large, with impressed medial field, and armed with cuticular spines. Hypopygium unarmed. Legs: Mid tibia with single pectinate spur. Hind tibia with single pectinate spur. Hind basitarsus not widening distally, circular in cross-section. Posterior flange of hind coxa not produced as raised lamella. Metatibial gland present as oval patch of whitish cuticle. Metabasitarsal gland absent. Hind pretarsal claws simple. Polymorphism: Highly polymorphic.

Queen
Dichthadiiform, blind, with median ocellus (see e.g. Barr et al. 1985).

Male
Head: Antennae with 13 segments. Clypeus without cuticular apron. Parafrontal ridges absent. Torulo-posttorular complex vertical, carinae separated by broad flat or convex area between exposed antennal sockets. Maxillary palps 2- or 1-segmented. Labial palps 1-segmented. Mandibles falcate. Ventrolateral margins of head without lamella or ridge extending towards mandibles and beyond carina surrounding occipital foramen. Carina surrounding occipital foramen ventrally absent. Mesosoma: Pronotal flange not separated from collar by distinct ridge. Notauli absent. Transverse groove dividing mesopleuron absent. Propodeal declivity reduced, without distinct dorsal edge or margin. Metapleural gland opening absent. Propodeal lobes absent. Metasoma: Petiole anterodorsally immarginate, dorsolaterally immarginate, and laterally above spiracle immarginate. Helcium in relation to tergosternal suture placed at posttergite and axial. Prora simple, not delimited by carina. Spiracle openings of abdominal segments IV–VI slit-shaped. Abdominal segment III more than half size of succeeding segment IV; latter weakly constricted at presegmental portion (uninodal waist). Girdling constriction of segment IV absent, i.e. pre- and postsclerites indistinct. Cinctus of abdominal segment IV absent, not impressed. Girdling constriction between pre- and postsclerites of abdominal segments V and VI absent. Abdominal segment IV not conspicuously largest segment. Abdominal sternite VII simple. Abdominal sternite IX distally armed with two spines, with lateral apodemes short, directed sideways. Genitalia: Cupula short relative to rest of genital capsule and shorter ventrally than dorsally. Basimere fused basally, with a fragment reduced to tiny, plate-like sclerite. Telomere folding backwards and then over rest of genital capsule, concealing it dorsally. Volsella gradually tapering toward apex. Penisvalva laterally compressed, rounded at apex. Legs: Mid tibia with single pectinate spur. Hind tibia with single pectinate spur. Posterior flange of hind coxa not produced as raised lamella. Metatibial gland absent. Metabasitarsal glands absent. Hind pretarsal claws simple. Wings: Tegula present, broad, demiovate in shape. Vein C in fore wing present. Pterostigma narrow. Abscissa R·f3 absent. Abscissae Rs·f2–3 absent. Cross-vein 2r-rs present, connected to Rs·f2–3&Rs·f4. Abscissae Rs·f4–5 differentiated into Rs·f4 and Rs·f5 by 2rs-m. Abscissa M·f2 in fore wing contiguous with Rs+M. Abscissa M·f4 in fore wing present, reaching wing margin. Cross-vein 1m-cu in fore wing present. Cross-vein cu-a in fore wing present, arising from Cu and distal to, at or near M·f1. Vein Cu in fore wing present, with only Cu1 branch prominent. Vein Cu in fore wing present, with both branches Cu1 and Cu2. Vein A in fore wing with abscissae A·f1 and A·f2 present. Vein C in hind wing unknown. Vein C in hind wing present. Vein R in hind wing present, extending past Sc+R but not reaching distal wing margin. Vein Sc+R in hind wing present. Abscissa Rs·f1 in hind wing present, shorter than 1rs-m. Abscissa Rs·f2 in hind wing present, not reaching wing margin. Cross-vein 1rs-m in hind wing present, shorter than M·f1. Vein M+Cu in hind wing present. Abscissa M·f1 in hind wing present. Abscissa M·f2 in hind wing present. Cross-vein cu-a in hind wing present. Vein Cu in hind wing present. Vein A in hind wing with abscissae A·f1 and A·f2 present.

Larvae
Larvae of Dorylus have been described in Wheeler (1943) and Wheeler and Wheeler (1984). Cocoons are absent.