The male "sausage fly" form of the genus are remarkable and those of Anomma driver ants are reputed to be the largest of all known ant morphs. Similarly, the special large "dichthadiigyne" special form of queens, which are wingless, without ocelli, and capable of periodic egg production on a very large scale are particularly characteristic.

Wheeler (1922) gave a key to the subgenera, the validity of which was confirmed by Gotwald & Barr (1980). The largest workers show the most distinctive development and it is these which have been drawn. Barr et al. (1985) and Gotwald (1985) recognized three groupings within the genus - Dorylus (including Anomma, Dichthadia and Typhlopone), Alaopone and Rhogmus - but felt unable to resolve whether the groupings should be regarded as full genera or simply subgenera. Thus, Wheeler's (1922) subgenera have been retained here, a decision supported, with reservations, by Bolton (1995).

Modern work of a high quality and using the latest techniques such as DNA sequencing have led to the development of a molecular phylogeny for the "Dorylus army ants" (Kronauer et al., 2007). The study appears to confirm the status of the six subgenera. The separation of Dorylus (D.) and D. (Anomma) is lower than for the other four subgenera.

Within the subgenera, the species lists follow Bolton (1995), although I found his listing confusing with respect to the subgeneric status of some species and curious in having instances denoted as "new combination" without any obvious revisionary evidence. Soldiers, workers, queen and male forms are denoted as being in the descriptions listed by Bolton (1995). As Bernard (1952) noted, there is a need for revision to decide the association of male-only species with worker-only species - a good example of this was his revision of Dorylus (Anomma) stanleyi where the worker of molesta was recognised as with the male stanleyi. As the earlier name is molesta, I have adopted that as the species name (correctly Latinised to molestus) and not stanleyi.

The subgenus Dichthadia is monotypic with a single species from South East Asia - a modern review of the South East Asian Dorylus, including laevigatus, by Stefanie Berghoff (2002, doctoral dissertation) can be found at - http://deposit.ddb.de/cgi-bin/dokserv?idn=967765773&dok_var=d1&dok_ext=pdf&filename=967765773.pdf.

In June 2002, I attempted a fresh review of the subgenus Anomma. Over a long period, ending in 1933, the description of a plethora of subspecies and varieties led to taxonomic confusion and gave little idea of how many distinctly separable species could be reliably recognised. Bolton (1995), for instance, lists sixteen species and some 28 names of such as subspecies, varieties and "unavailable names". After receiving a number of samples, from field collections in Cameroun, Guinea and Tanzania, and each with a range of sizes of specimens, I made a start on the slow process of reading the original publications. From the summary literature, such as Bolton (1995) and Hölldobler & Wilson (1990), I was aware of the bionomics research undertaken by Raignier & van Boven (1955) in the former Belgian Congo and knew that they had made some, apparently minor, taxonomic findings. To my surprise, however, I found that their treatise contained an extensive taxonomic analysis of fresh specimens they collected during their field studies. Even better, they had made a detailed review of all the historic literature and of the considerable material, including many type specimens, held in the Royal Museum of the Belgian Congo, at Tervuren in Belgium. Their taxonomic analysis shed much light on the situation but, curiously and despite their derision of the use of "subspecies" and "varieties", they refrained from attempting a formal revision. The only reason seems to have been their feeling that there is a state of natural variation between individuals in the enormous colonies, numbering a million or so workers, and this makes precise descriptions of "species" impossible. I guess also that they held back because their material and that in Tervuren came only from the Congo Basin. With their work and the fresh material sent to me from such widely separated locations, I felt a little more secure in trying to bring some sense to bear. With further specimens and better images, I now recognise twenty species known from, at least, the worker stage and a further seven species known only from descriptions of males or queens.

The hypogaeic (subterranean) forms are thought to feed mainly on termites, other ants and, in some cases, earthworms. The epigaeic forms, however, are generalized feeders (Hölldobler & Wilson, 1990, page 595), and are nearly all Anomma species (Gotwald, 1974). Hölldobler & Wilson (1995) give a somewhat poetic but graphic summary of driver ant activity, describing the colonies as "superorganisms". The similarity between the Dorylus driver ants of Africa and the Eciton army ants of the Americas now is accepted as being an example of parallel evolution, with no close relation between the two groups.

I have drawn up a key to separate the males into subgenera, this is at DORYLINAE males.

Key to workers, major and media, of African subgenera - after Wheeler (1922) and J van Boven (1975: 197, see , which includes genus level separation of males and queens).

1	Antennal funiculus with a maximum of 8 segments	Subgenus Alaopone
--	Antennal funiculus with a maximum of 9, 10 or 11 segments	2
2	Impressed semi-circular area of pygidium without distinct margins	3
--	Pygidium with a sharply margined, semicircular impression; antennae 11-jointed	4
3	Frontal carinae of head diverging posteriorly, the minimum separation being no more than the width of the antennal scape; alitrunk with promesonotal suture hardly visible; petiole longer than wide and with lateral spiracles distinctly raised; subapical tooth of mandibles simple, antennae 11-jointed; major worker around TL 13 mm	Subgenus Typhlopone
--	Frontal carinae near parallel, minimum separation much smaller than width of scape; promesonotal suture deep; petiole as wide as long and lateral spiracles not raised; subapical tooth of mandible double or truncate; major worker TL around 8 mm	Subgenus Rhogmus
4	Frontal carinae only slightly divergent and often armed with a spine, minimum separation of carinae never more than the maximum width of the scape; funiculus segments; antennae short and thick; all joints of the funiculus, except the last, as wide, if not much wider, than long; in major mesonotum continuous with propodeum	Subgenus Dorylus
--	Frontal carinae widely divergent and never armed with spines, always separated by more than the maximum width of the scape; antennae elongate, at least some funicular joints longer than wide; major (at least) with distinct mesonotal saddle	Subgenus Anomma