The Ants of Africa
CHAPTER 3 - Mosaics - Evidence from Nigeria - 2
Taylor; Taylor & Griffin; Taylor & Adedoyin
Mosaics Introduction Mosaics Nigeria 1 Mosaics Nigeria 3

The location of all the study areas at CRIN can be seen on the CRIN Plan as it was in 1975.

A list of all the species we found at some time on cocoa in Nigeria is on the linked list, Taylor 1981 list, which is updated from that published in 1981 in the Final report of The International Black Pod Project. I now (2004) have further updated the nomenclature on that list.

Taylor (1975)

The motive for the programme of work, based at CRIN, was an investigation of the role on ants in the dissemination of the fungal disease of cocoa, known as black pod disease, caused by species of Phytophthora. Although the specific ant-disease studies will be outlined separately, there were aspects of the results which added to the knowledge of ant mosaics. A preliminary experiment was undertaken in 1974 to examine whether exclusion of insects, including ants, would reduce the incidence of the disease (Taylor, 1975). The key result in terms of mosaics was that the exclusion effects were confused by the differences in ant populations between the three treatments and the impact of invasion by species into the lacuna caused by elimination of resident ants. For instance, it was felt that the presence of Oecophylla longinoda in the untreated control plot might have had a negative effect on the disease by excluding other ants suspected by being disease vectors. It was realised that relatively small plots, in this instance with 110 trees each, in a single block of 500 trees, were probably inadequate to overcome the influence of dominant ants with large colonies or foraging areas.

Taylor (1977)

Thus, in 1975, large scale studies of the ant mosaic were undertaken at CRIN. These were a season-long study of 17 plots, each of 25 trees, within a large block of cocoa, Onipe 1/1; a repeated survey of all trees in a small, 400 trees, block of isolated cocoa, S1/1; a selective survey in six other blocks of cocoa; simple "rapid" surveys of 13 other cocoa blocks; a survey of cocoa under full forest shade; surveys, of 100 trees each, in blocks of kola, cashew, coffee, oilpalm and plantain; and a survey of the ants on forest trees, when the trees were felled during a land-clearance (Taylor, 1977).

The S1/1 study enabled the ant mosaic to be plotted two-dimensionally, showing how some of the common dominants form clear, separated domains. The illustration shows the results of a survey on 7 July 1975. From the season-long regular examinations, part of the black pod disease study, the dominants were Oecophylla longinoda (on 138 of 276 trees with ants) and Crematogaster clariventris (on 28 trees) at the canopy level and Lepisiota guineensis (as Acantholepis) ?capensis) (110 trees) at the off-ground level. The plotting of other species gave an indication of how these might be influenced by the dominants. The main species were Odontomachus troglodytes (20 trees), Pheidole megacephala (41 trees), Crematogaster gambiensis, Crematogaster painei (species A); and Camponotus acvapimensis. Comparison with the existing reports from Ghana showed that this was the first time detailed mapping of species other than the accepted dominants had been done.

At the Onipe 1/1 block, the scale of which was bigger than any of the single-site work in Ghana, the study of all ants observed on cocoa trees gave results which were mapped so as to show regular and occasional occurrences of each species and uniquely (using isometric graph paper) to portray the mosaic in a true three-dimensional manner. The major arboreal dominants encountered were Oecophylla longinoda (on 40.4% of cocoa trees), Tetramorium aculeatum (11.3%), Crematogaster africana (12.8%) and Crematogaster depressa (uncommon). Two other species, Lepisiota cacozela (as Acantholepis T); (10.8%) and Pheidole megacephala (15.2%) were common dominants at the ground to lower trunk level. The variations between each of the plots were quite remarkable, as was the complexity of an undominated situation.

This is illustrated by three of the plots, click here. On a plot-by-plot basis, it was on only 7 of the 17 plots that the majority of the trees had a single dominant, six of which involved Oecophylla longinoda. The occurrence of the two large Crematogaster species, Crematogaster africana and Crematogaster depressa, was determined by the presence of mature, big indigenous trees on which they built their large carton nests. The mapping also showed the apparent close association of Camponotus vividus with Crematogaster africana, and of various small Crematogaster species with Oecophylla longinoda. Odontomachus troglodytes was fairly commonly found ascending trees, especially in non-dominated situations.

The selective surveys of six other cocoa blocks at CRIN were biased to include four sets of 50 trees, dominated by Oecophylla longinoda, Tetramorium aculeatum, Crematogaster africana and Crematogaster depressa respectively, plus 100 supposedly non-dominated trees. In fact the latter turned out to be dominated in the main by Lepisiota guineensis (Acantholepis ?capensis) (on 71 trees) and/or Pheidole megacephala (25 trees), other common species were Camponotus acvapimensis (25 trees) and Plagiolepis brunni (23 trees). The results, combined with the Onipe 1/1 and S1/1 data, gave enough values for the common species for association analyses to be performed. In these a total of 1426 possible occurrences of ants were considered and 56 species of ant were recorded. Table 1 summarises the results of the association analyses. If species never occurred together the association could not be tested, this was the case for Oecophylla longinoda and Crematogaster africana. Paratrechina grisoni (as Paratrechina species 3) was found to have no association with either a dominated or a non-dominated situation.

The associations between species, both positive and negative, are shown visually in the linkage diagrams. Note: - updated nomenclature is given on the pages.

In the "rapid" survey identifications were made in the field with the naked eye, separation of Crematogaster africana and Crematogaster depressa, of Lepisiota (as Acantholepis) species and of Crematogaster gambiensis / zavattarii (nigeriensis) was not possible, similarly several small species were lumped as Pheidole megacephala. In total, 4898 cocoa trees were examined. The results together with results of surveys of other tree crops and of indigenous forest trees are shown in Table 2. The different levels of abundance of ant species on the other tree crops was related to the habitat provided by the crop. For instance, Oecophylla longinoda which constructs its nests by binding leaves together cannot utilise the large leaves of plantain, the narrow leaves of oil palm or the small, leathery leaves of cashew. The leaf axils of oil palm are usually debris filled and afford excellent nesting sites for Pheidole megacephala, Lepisiota guineensis (as Acantholepis ? capensis) and, near the ground, for Odontomachus troglodytes. The main curiosity was the presence of Crematogaster striatula on coffee and kola, it was among the commonest cocoa dominants reported from Ghana but was rarely found on cocoa at CRIN. Other species with a measurable level of occurrence on cocoa plots were Tapinoma lugubre (formerly Technomyrmex detorquens) (>5% of trees at Onipe 1/1), Crematogaster (Orthocrema) muralti (>1% of trees in biased surveys), Platythyrea conradti (>1% of trees); also Myrmicaria fumata (earlier reported as striata), Monomorium vaguum (as Diplomorium lujae) and Camponotus (Myrmacrhaphe) species T were found on 0.1-1.0% of trees.

Taylor & Griffin (1981)

Other season-long information came from a major epidemiological study in 1975, the E5/1 experiment (Taylor & Griffin, 1981), which included twice-weekly recording of the ants seen building tents over Homoptera, etc., on 1139 cocoa trees. The trial site was defined within what originally had been a spacing trial, with the trees planted in several configurations. Canopy surveys revealed that most of the canopy was dominated by Tetramorium aculeatum, which perhaps was the reason for there being an interesting pattern of soil or lower trunk dwelling ants tending the Homoptera, mostly aphids, and constructing tents. Of the arboreal dominants the only other species present was Crematogaster zavattarii (nigeriensis / ? gabonensis) with a single large colony which occupied much of one corner of the site.

The results of the ant distribution records are shown in the series of site plans - overall ant pattern | Camponotus acvapimensis details | and the photograph shows the E5/1 central area.

The trial protocol included counts of all pods and this gave a measure of tree fecundity or productivity. In the area of densest planting the productivity was very low, with relatively few flower cushions and little in the way of Homoptera and little ant activity. In contrast productivity was highest where the spacing was greatest, such as that shown in the view of the central area. Homopteran populations were high and ant activity was high). The level of soil insolation also was highest in the more open spacing, this provided nest sites for the most abundant tent-building ants, Camponotus acvapimensis and, less, Myrmicaria fumata (earlier reported as striata). Both built soil tents which were sources of black pod disease. Camponotus acvapimensis appeared to have about 15 colonies from which the ants built soil tents on 462 trees, on 97 of which the tents were sources of black pod disease, i.e. the soil contained spores of the fungus. Myrmicaria fumata had six colonies and built soil tents on only 17 trees, on 11 of which Camponotus acvapimensis also was active. In the medium shade areas, where productivity and Homopterans were quite high, the commonest tent-builders were Pheidole megacephala, with three distinct colonies of 4-10 trees, and Odontomachus troglodytes, with 20-25 nests (among the roots of a cocoa tree) from each of which foragers visited 2-5 trees. Paratrechina grisoni (species 3) was the only other builder of tents associated with black pod disease, being found as 9 small colonies, from most of which the ants visited only 1-2 trees. Records of the heights of tent building activity showed that Odontomachus troglodytes rarely climbed above 140 cm, whereas the others were commonly found at 2.5-3.0 m high. Crematogaster zavattarii was truly arboreal, nesting under flaky bark on the trunks, and not being found below about 1.0 m and otherwise to about 4.0 m. Its carton tents were not associated with black pod disease. In terms of the ant mosaic the E5/1 study revealed a pattern of habitat determined colonisation, with the major influence being canopy density and consequent soil insolation.

Taylor & Adedoyin (1978)

The ant mosaic findings from CRIN were extended in early 1976 with a survey, led by Sampson Adedoyin, of 76 cocoa farms spread throughout the cocoa growing area of western Nigeria see the Map (Taylor & Adedoyin, 1978). Fifty trees were examined at each farm and notes made of the canopy quality. The overall occurrence of species is included in Table 2 . The result for Crematogaster painei (Crem. species A); includes Crematogaster bequaerti. The results were subjected to a principal components analysis (PCA) and this confirmed the CRIN findings with regard to the pattern of species which were both numerically abundant and mutually exclusive. It additionally confirmed the associations, both positive and negative, which occur between co-dominants and sub-dominants. For most of the common species the frequency, expressed as a % of the total 3800 trees, was of the same order as that found at CRIN. Most notable differences were the higher frequency for Crematogaster clariventris, Plagiolepis brunni and some of the lesser species.

The ant associations as determined by the PCA are illustrated in the diagram; the separation of the dominants matches the results obtained from the work at CRIN (see the linkage diagrams).

{Development of the ant mosaic in Nigerian cocoa}

Development of the ant mosaic

The combined analyses enabled the development of a description of how the changing habitat presented by a cocoa plantation - as it passes from saplings, with an unclosed canopy; through maturity, with a closed interlocked canopy; into the final stage, of ageing and canopy degradation - profoundly affects the potential occupancy by the main species of dominant ants.
This is summarised in the illustration (right) - L. capensis now regarded as Lepisiota guineensis..

The main extraneous factor was the presence of the indigenous shade trees which were obligate nesting sites for Crematogaster africana and Crematogaster depressa (presumably why they were not found at over half the 76 farms). Interestingly, Camponotus acvapimensis usually is described as a savannah species but its frequency was rather higher in the wetter, eastern, area of the 76-farm survey.

Contents Mosaics Introduction Mosaics Nigeria 1 Mosaics Nigeria 3
1998, 2001, 2004, 2005, 2011 - Brian Taylor CBiol FSB FRES
11, Grazingfield, Wilford, Nottingham, NG11 7FN, U.K.