Summary

Following on from my earlier thoughts on biodiversity based on information on West African ants, this page pulls together much wider findings on the ants of the whole of sub-Saharan Africa. It also incorporates records from the small series of recent revisions of genera, most having only a few member species, and sources such as www.Antweb.com.

The IUCN Red List of Endangered Species includes a few ants from sub-Saharan Africa [http://www.iucnredlist.org/static/introduction - accessed 25 August 2009]. These, Anoplolepis nuptialis (Santschi, 1917), Cardiocondyla zoserka Bolton, 1982, and Pheidole neokohli (Wasmann, 1915), are listed as vulnerable? Whilst it is true there are few records of these is that any reason for them to be regarded as "vulnerable". The queen of An. nuptialis was first described by Prins (1982) and he gave details of seven separate findings. The other two, both inquilines or social parasites, each have been recorded only once. C. zoserka was suggested as perhaps dependent on Cardiocondyla shuckardi but that, however, was described as the commonest and most widespread endemic species (of Cardiocondyla) in the Afrotropical Region (Bolton, 1982). Ph. neokohli was found as a parasite of Pheidole melancholica but again the host is quite widely recorded and apparently abundant.

Robertson (2002) put forward some estimates of the number of ant species likely to be found in Africa.

ABSTRACT: Forty-three of the 82 Afrotropical ant genera (52%) have been revised to modern standards resulting in a 50% increase in number of species. There are currently 1705 species of ants known from the Afrotropical region, a figure that could increase to over 2136 species if all ant genera receive a modern revision. To incorporate all undescribed species, total Afrotropical ant species richness was calculated by extrapolating from data on the proportion of undescribed species collected at Mkomazi Game Reserve in Tanzania and the Cape of Good Hope section of the Cape Peninsula National Park in South Africa. On this basis there are an estimated 3105 species of ants in the Afrotropical region, with 45% undescribed or currently occupying an infraspecific taxonomic rank. This extrapolation assumes that the average range sizes of described and undescribed species are similar, which in reality is unlikely because widely distributed species are more likely to have been collected and described. I present a method that distinguishes between widespread and localised species to correct for this problem, which extrapolates 4093 Afrotropical ants species, with 58% of species estimated to be undescribed or currently recognised only at infraspecific rank. It would take a highly productive systematist at least 21 years to complete revisions of the unrevised ant genera. A strategy is presented for improving specimen collection and taxonomy of Afrotropical ants.

[Robertson, H.G. 2002. Afrotropical Ants (Hymenoptera: Formicidae): Taxonomic Progress and Estimation of Species Richness. Journal of Hymenoptera Research, 9, 71-84]

The pentanomial system, of genus, subgenus, species, subspecies (or stirps) and variety, was described by Ward (2007) as "a peculiar and unwieldy" nomenclature. If peculiar is taken in its older, original, concept of unique then Ward was wrong. The mosquito systematists, or culicidologists, also used such a system although in, for instance, the bench mark work of J.N. Belkin, Mosquitoes of the South Pacific (Belkin, 1962), it was reduced to a trinomial, retaining the subgenus level. I find the subgenus is a useful separation when dealing with genera containing many member species. It seems ironic that the adherents of the strict binomial have resorted to use of the term "species-group". Ward (2007) cited Wilson (1955) as a leader in the modern taxonomy but the latter retained the subgenus. Wilson himself criticised separation of specimens into species on the basis of "a single character of pilosity" and referred to the statement in the joint paper of Wilson & Brown (1953), "that the species is only completely objective in terms of local faunas, i.e. where discrete populations co-exist at the same place at the same time". Brown (1958g) used subgenera as did Wilson & Taylor (1967). Suddenly, it appears, Brown (1973a) chose to abandon subgenera but create species-groups. At one point he wrote "the genus [Crematogaster] has been split into subgenera, but these are apparently only species groups, and some of them weak at that". Although widely cited, especially the appendix, as an authority for synonymization of genera and subgenera, Brown's paper is totally lacking in true taxonomy, presenting his opinion at that time. Brown wrote of "a good many more 'projected synonymies' that have never been proposed anywhere formally in print". A little later, he argued against the recognition of Stenomyrmex as a formal subgenus of Anochetus, referring to it as "(no) more than an arbitrarily-defined species group" (Brown, 1978). Neither Brown or Wilson contributed further to the matter but a consequence seems to be the use of notes such as a very variable species to encompass what had been viewed as subspecies. For example some Cataulacus species by Bolton (1974a), who later had to disentangle some of his synonymy (Bolton, 1982). Longino (2003a) appears to regard the subgenus as a reasonable nomenclatorial level.

[Ward, P.S. 2007. Edward O. Wilson and his contribution to ant systematics, pp 3-7. In Snelling, R.R., Fisher, B.L. & Ward, P.S. (eds). Advances in ant systematics (Hymenoptera: Formicidae): homage to E.O. Wilson - 50 years of contributions. Memoirs of the American Entomological Institute, 80.]

Sources of data

The distribution records collated here have been gleaned from 461 papers of an essentially taxonomic nature, the earliest being published in 1767 and the most recent in 2008. Additional distribution data came from 189 papers dealing with either the biology/ecology of species or providing the results of surveys. Unfortunately since about 1950, e.g. Weber (1943c and 1952a), very few perpetrators of surveys have followed up the reports with any novel taxonomic work. Moreover, in the 30 or so taxonomic revision papers published since 1970, even where the morphospecies collected in surveys have been among the specimens later recognised, in some cases as new species, it is rare to find any reference to the survey publications. For instance, Watt, A.D., Stork, N.E. and Bolton, B. (2002). The diversity and abundance of ants in relation to forest disturbance and plantation establishment in southern Cameroon. Journal of Applied Ecology 18: 39: 18-30.

Over the past six years a number of people have sought my help in identifying ants, mostly from sub-Saharan Africa. I have endeavoured to identify all the specimens. This has been facilitated by the work of Donat Agosti and Norman F Johnson in creating the www.antbase.org on-line access to almost all ant papers ever published. With that access, somewhat surprisingly, separation of taxa has proved possible to a remarkable extent. As a first step it was necessary to construct dichotomous keys for the many genera with no keys or only limited keys from the first half of the twentieth century.

Ward (2007) described the need to create "species pages" for all of the world's ants. Here on this website there are species pages for every one of the known taxa from sub-Saharan Africa.

Without undertaking the herculean task of sighting all the type specimens, some of which may no longer exist, I cannot claim certainty of identification in every instance. The advent of inexpensive and simple to use digital cameras, however, has enabled me to produce diagnostic photomontages of each of the taxa I have examined, thus providing an on-line voucher collection.

In that I continue my effort of 1976 in separating 201 species from Nigeria, almost all shown in accurate drawings (Taylor, 1976-80). Of those, 121 were first records for Nigeria. The specimens were deposited as a voucher collection with the British Museum of Natural History, with many duplicates left with the insect collection at the Cocoa Research Institute of Nigeria. Almost 100 of the taxa were designated by genus and code names. Some have since been identified in revisionary papers and I have endeavoured to recognise the remainder. Thus, I have compiled a webpage update to Taylor (1976-80).

Identification of ants collected in modern surveys and other findings

Most of these findings are otherwise unpublished.

• Ants collected in Benin by Sévérin Tchibozo, 2007; pitfall traps; 16 samples, 15 species; one new species, one new status (13.3% new).
• Ants collected in Cameroun by Gabriel Debout & Ambroise Dalecky, 2001; diverse collecting methods, some specifically from myrmecophyte trees; total 75 species, 32 new records for Cameroun; possibly two or three new species; three species recognised as new or revived status (8.0% new).
• Ants collected in the Central African Republic by Philippe Annoyer, 1998, 2005 & 2008; diverse collecting methods, including from platforms in the forest canopy; 176 species; 121 new records; 7 new species, 7 new status (7.9% new).
• Ants collected in the Congo, at Brazzaville, by Yves Braet & Eric Zassi, 2007; 18 pitfall traps; 46 species, two with revived status and two probable new species. (8.7% new)
• Ants from Congo, Réserve de Lésio-Louna collected by Eric Zassi, 2007; 180 pitfall traps; 69 species, 36 new records for Congo, one new species (1.4% new).
• Ants collected in Gabon by Yves Braet, 2006; diverse collecting methods; 144 species; 110 recorded for the first time in Gabon; one new genus and type species; possibly a second new genus and species; one new species; three new status species (3.5% new) [Braet, Y. & Taylor, B. 2008. Mission entomologique au Parc National de Pongara (Gabon). Bilan des Formicidae (Hymenoptera) reécoltés. Bulletin S.R.B.E./K.B.V.E., 144, 157-169].
• Ants collected in Ghana by S. Sky Stephens, 2005; a study of forest and plantations; diverse collecting methods; 70 species, 20 new records, five given new status (7.1% new).
• Ants collected in South Africa by Peter Hlavác, 2004, and others; mostly terrestrial hand collection; 34 species, three new species (8.8% new).
• Ants photographed in South Africa by Joan Young and others, 2008.
• Ants collected in the Usambara Mountains, Tanzania by Valery Grebennikov, 2002; mostly terrestrial collectiing by hand; 33 species, 25 new records, four new species (12.1% new).
• Ants from tree canopy sampling in the Mkomazi Game Reserve, Tanzania collected by George McGavin, 1995-1997; pyrethrum spraying of tree canopies; 49 species, 27 new records; two new species, five new or revived status (14.3% new)
• Ants collected in Sudan by Awatif Omer, 2005 & 2006; collecting by hand; 29 species, five with new or revived status (17% new).
• Ants collected in miscellaneous countries; most from hand collecting
  Sahelian
  Morocco - 1 species; collector A Lenoir.
  Tunisia - 2 species, collector T Lackner.
  Burkina Faso - 8 species, 6 new records, one new species; collectors A Lenoir & D M King.
  Niger - 5 species, 4 new records, one new status species; collector D M King.
  Chad - 3 species, 2 new records; collector D M King.
  Mali - 6 species, 12 new records, one new species; collector D M King.
  West African
  Senegal - 5 species, 3 new records; collector D M King.
  Guinea - 7 species, 5 new records, one new status species; ants associated with chimpanzee "ant dipping"; collector T Humle.
  Ivory Coast - one species, one new record; ants associated with chimpanzee "ant dipping"; collector T Humle.
  Nigeria - two species, 2 new records, one new status as species; found in imported yams in Israel; collector Z Klein.
  Cameroun - three species, 2 new records, 2 species new status; collector Hauke Koch.
  Congo Basin
  Central African Republic - 2 species, two new records, one new status; collector Degallier.
  East & Southern Africa
  Kenya - 7 species, 4 new records, two new status species collector D M King.
  Uganda - 4 species, 4 new records; ants associated with chimpanzee "ant dipping"; collectors S Uehara and, C Hasimoto; also one hand collection by Awatif Omer.
  Rwanda - five species, 5 new records; hand collecting in a coffee plantation; colletor Gabriel Bizimungu.
  Tanzania (south) - five species, 4 new records, one new species, three new status species; ants associated with chimpanzee "ant dipping"; collector S Uehara or a colleague.

The Ghana specimens are held by Sky Stephens.

The specimens from almost 1300 samples currently remain in my care but will be deposited in the Oxford University Museum of Natural History.

Analysis of findings

As at August 2009, I recognise 1946 species (a small number being as yet unnamed). A frequency score was derived for each species by allocating values as follows:
1 = known only from the holotype specimen; 2 = known from a single collection, including the holotype; 3 = known from 2-10 records; 4 = known from many records or known as widespread from survey work.
These were totalled up to give the score shown in the histogram, Figure 1, the maximum score being 168, i.e. 4 X 42 countries.

The scores were then entered in a standard Excel 2003 spreadsheet. The arrangement of countries is approximately in line with the ecological zones but, as will be seen in the analyses graphs, some parts, e.g. northern Angola and southern Sudan, have species in common with the Congo Basin forest.

General biodiversity

Figure 1

Arbitrary categorisation of species

Score 40+ = 36 species; 30-39 = 30 species; 20-29 = 84 species; 10-19 = 236 species (ca. five countries); 7-9 = 206 species (ca. four countries); 5-6 = 234 species (ca. three countries); 3-4 = 420 species (ca. two countries); 1-2 = 679 (single records); 6 species are dubious for Africa. This shows ubiquity and revels nothing about numerical abundance in terms of individuals.

An indication of the systematic range of occurrences can be gauged from Figure 2.

Figure 2

Most ubiquitous species

The top 36 species and their recorded distribution are shown in Figure 3.

Because of the Forelian scepticism of an apparently ever increasing numbers of species, for about 90 years from 1874 there was a confusing practice of denoting subspecies (or stirps) and even quadrinomial varieties (Forel, 1874; Arnold, 1962). This has made it impossible to be definite about the distribution of several of the commonest "species", among them "Camponotus maculatus", "Crematogaster castanea", "Crematogaster impressa", "Dorylus nigricans/burmeisteri", "Lepisiota capensis", "Oecophylla longinoda" and "Pheidole megacephala",. Some spurious results arise from the use of apparently similar coding for denotation of unidentified species, e.g Genus species A.

Figure 3

The analysis is distorted, however, by the simple fact that national boundaries bear no relation to ecological zones. The map below illustrates the true situation [Source - http://www.lib.berkeley.edu/EART/maps/africa-veg.gif].

Figure 4

Comparisons

• Graph 1 - Species occurrence sorted by Zaire as the main criterion, level 1 - most of the Pan-African taxa.
• Graph 2 - Species occurrence sorted by Zaire as the main criterion, level 2 - more Pan-African taxa and a cluster of eastern Africa species.
• Graph 3 - Species occurrence sorted by Zaire as the main criterion, level 3 - mainly forest zone species.
• Graph 4 - Species occurrence sorted by Zaire as the main criterion, level 4 - Zaire species scoring 3, more forest zone species.
• Graph 5 - Species occurrence sorted by Zaire as the main criterion, level 5 - most ubiquitous species without any Zaire species; effectively separated with the species to the left of the empty Zaire column. i.e. to the West, not being not found to the right, i.e. to the East.
• Graph 6 - Species occurrence sorted by Zaire as the main criterion, level 6 - without any Zaire species, two clusters - an upper set from southern Africa and a lower set of savannah and open woodland species.
• Graph 7 - Species occurrence sorted by Zaire as the main criterion, level 7 - almost all species are from the West African/western Congo Basin region; towards the right are a few purely East African species.
• Graph 8 - Species occurrence sorted by Mali as the main criterion and Sudan as a second separator - almost all species are from the Sahel zone and most of the species from the forest zone of south-western Sudan are excluded.
• Graph 9 - Species occurrence sorted by Mali as the main criterion and Sudan as a second separator - extract from the lower sector excluding the Mali species - almost all species are from the Congo Basin with the Sudan speeies being those from the south-western forests; towards the bottom are two small clusters of East African/Sahel species.

As the graphs show the same data sorted according to different criteria, thus there are some species appearing in two graphs.

There are no species in common between the two main Tanzania lists - the Mkomazi specimens all being arboreal (possibly a small number of those ascending from the ground) and in Graph 5, whereas all the Usambara species were collected by a terrestrial ecologist, Valery Grebennikov, with a main interest in Scarabaeidae, and in Graphs 1-4.

Abundance versus ubiquity

Small colonies.
Driver ants Anomma vast colonies but perhaps one colony per hectare (or even greater area) as the food needs for the wholly carnivorous colony must be vast.
Acacia dominants of the eastern African tree savannah occupy almost everyone of the millions of trees. for instance Crematogaster mimosae is known from only four countries and Cr. nigriventris from five, but the sheer numerical abundance of both has to be vast.
On the wholly man made cocoa monoculture of the Cameroun west through to Sierra Leone perhaps as many as 35% of the trees are occupied by Oecophylla longinoda as a dominant.
In the vast rain forests the canopy dominants may well be one of the Crematogaster (Atopogyne) species, also found on cocoa where there are large trees remaining within the cocoa farms.

Discussion

To return to Robertson (2002) on the number of ant species likely to be found in Africa. He noted there are currently 1705 species of ants known from the Afrotropical region, a figure that could increase to over 2136 species if all ant genera receive a modern revision. To incorporate all undescribed species, total Afrotropical ant species richness was calculated by extrapolating from data on the proportion of undescribed species collected at Mkomazi Game Reserve in Tanzania and the Cape of Good Hope section of the Cape Peninsula National Park in South Africa. On this basis there are an estimated 3105 species of ants in the Afrotropical region, with 45% undescribed or currently occupying an infraspecific taxonomic rank. This extrapolation assumes that the average range sizes of described and undescribed species are similar, which in reality is unlikely because widely distributed species are more likely to have been collected and described. Robertson gave a method that distinguishes between widespread and localised species to correct for this problem, which extrapolates 4093 Afrotropical ants species, with 58% of species estimated to be undescribed or currently recognised only at infraspecific rank.

My view is far more restrictive. I have 1300 samples in my collection comprising 565 species (all with photos original to this site). The collection contains what appear to be 35 new species (one new genus, one possible new genus); 125 species otherwise known solely from the original description; 54 species previously reported only twice; 45 species reported only two-three times previously; and, 339 common species. Most of the samples came from the quite intensive surveys reported above.

The surveys, which between them utilised all the standard methods of sampling invertebrate fauna, produced 35 taxa that I feel may be new species and I consider another 33 taxa merit resurrection or elevation from synonymy or other lower status. I separated 70 species from the Ghana survey by Sky Stephens. The survey results suggest about 12.5% new species (about half previously categorised as of subordinate status). Earlier, I separated 201 species from Nigeria, only 10 of which remain as possible new species, although 14 were named subsequent to my original publication. The Nigeria work of Bolton and myself yielded 11.9% new species. In total on this website I have a total of 100 species with new or revived status, those not from the survey coming from my perusal of all the original descriptions. That gives 5.1% of the grand total 1942 species.

The specimens from almost 1300 samples currently remain in my care but will be deposited in the Oxford University Museum of Natural History.

A large proportion of the species I have separated come from disentanglement of the subspecies and varieties of Dorylus species. Other revisions come in some quantity from battling with the Pheidole and Crematogaster, two major genera in all senses but apparently avoided by most modern taxonomists. Digital photography has made it much easier to compare specimens so, hopefully this battle will progress further.

The variance, I suggest, stems from Robertson having examined two rather isolated areas with a consequent higher level of endemicity than, possibly, anywhere else on the continent. The two Tanzania surveys I report, from the Mkomazi and Eastern Arc Mountains, yielded 14.3% and 12.1% new species respectively. Robertson perhaps under appreciated the quantity and depth of the many collections made by explorers and missionaries between about 1880 and 1940. For instance, of the 565 species in samples sent to me contain 224, i.e. almost 40%, that have been reported no more than three times. Nearly one-quarter of the species were known only from the holotype colllections.

Although many of the ecozones of the continent now have been subject to the activities of exhibitions historically, together with ecological or crop protection entomologists in modern times, there are lacunae. Most notable is the swathe of savannah running south through a vast area of central Tanzania and south through Zambia.