Introduction The Ants of Africa
CHAPTER 3 - Mosaics - Evidence from Ghana - 6
Campbell; Belshaw & Bolton
Ghana - 1 Ghana - 2 Ghana - 3 Ghana - 4 Ghana - 5

Campbell (1979, 1983, 1984, 1990, 1994)

A co-entomologist with Bigger (see Mosaics - Ghana - 5) on the team studying swollen shoot virus. His studies centred even more firmly on mealybugs and the ant data adds to the information on populations rather than to the mosaic per se.

Between 1973 and 1977 (Campbell, 1983), he investigated four trees at each of 24 "sites" at CRIG. Sites 1-4 were in a 1 ha plot of Amelonado cocoa planted in 1952 and grown under dense forest shade. All the remainder were in a 13 ha 3 X 33 shade-fertiliser experiment, planted in 1959, with cocoa of open-pollinated Upper-Amazon parentage. The sampling for mealybugs was by the removal and subsequent examination of branches. Representative specimens of all ant species present on the tree at the time of sampling were collected for later identification.

Of the 14 species listed as dominants by Leston (1973), nine were represented. In terms of site presence these ants were - Oecophylla longinoda (13 sites); Tetramorium aculeatum (10); Pheidole species (9);Crematogaster castanea (9, of which 8 were with Oecophylla longinoda); Crematogaster africana (4); Crematogaster kneri (given as representing the Crematogaster wellmanni aggregate, 2, both with Oecophylla longinoda); Crematogaster stadelmanni (2); Lepisiota (as Acantholepis) capensis (1); and, Crematogaster clariventris (1). Tetramorium aculeatum and Pheidole species were variously found with Oecophylla longinoda, Crematogaster clariventris and Crematogaster africana. No direct relations between ants and Homoptera were testable under the conditions of the study other than the demonstration that Tylococcus westwoodi Strickland was found only on trees with Crematogaster stadelmanni.

During June and July 1974, Campbell (1984) examined all 14,000 cocoa trees in the 13 ha 3 X 33 shade-fertiliser experiment. All species of ant visible from the ground were recorded. Maps were prepared of the territories of the dominant ants and lists were compiled of the numbers of trees in every fertiliser subplot with each of the dominant ants. Only two species of ant, Lepisiota (as Acantholepis) capensis and Crematogaster clariventris were sufficiently widespread to permit assessment of the abundance of the Homoptera associated with them under a range of most of the shade and fertiliser treatments. Lepisiota capensis dominated one entire block (with no shade, light shade and heavy shade treatment plots) plus an adjacent heavily shaded area. Later surveys, in 1976 and 1978 (unpublished) showed that this species was eventually displaced almost everywhere by more typical dominants. It was not intimately associated with any honeydew producing Homoptera, but Toxoptera aurantii (apparently unaffected by shade levels) was found exclusively in its territory and Steatococcus species also occurred more frequently there. Crematogaster clariventris dominated the whole of the remainder of the site and Stictococcus sjostedti was almost the only species of Homoptera found in the carton tents of this ant. Mealybugs were not abundant with either dominant but Planococcoides njalensis, Planococcus citri and Pseudococcus concavocerarii were more common with Crematogaster clariventris.

The next paper (Campbell, 1990) has little fresh on ants, but the most recent (Campbell, 1994) reported studies on the 13 ha block which enabled the associations between Tetramorium aculeatum, Crematogaster clariventris and "Pheidole megacephala" (which was used as name of convenience for all closely related Pheidole species, as with Strickland, 1951a) to be closely examined. The combination was investigated particularly because of earlier evidence that Tetramorium aculeatum was frequently found as a co-dominant with one or both of the others. After the whole block survey (Campbell, 1984), ten spatially separated plots were selected with trees supporting all combinations of the three ant species and no other dominants. The plots were divided between heavy shade (by Terminalia ivorensis) and no shade. Sampling again was by branch removal from trees selected using a non-biased procedure. Mealybugs, coccids and psyllids were recorded but not aphids. Stictococcus sjostedti was the most abundant Homopteran but was almost totally absent when Crematogaster clariventris was not present and was negatively associated with "Pheidole megacephala". Planococcoides njalensis was the most numerous mealybug (47% of the total Pseudococcidae) but tended to be at its most common when "Pheidole megacephala" was present and was negatively associated with Tetramorium aculeatum. Planococcus citri (43% of the total Pseudococcidae) was more widespread but not obviously associated with any combination of the ants, although its numbers were higher when "Pheidole megacephala" ants were present, and it was negatively associated with Tetramorium aculeatum. A psyllid, Mesohomotoma tessmanni Aulmann (sometimes known as Tyora tessmani) also was widespread and abundant, no association was found with the ants although it was most abundant when all three ants were present.

Campbell concluded that his results confirmed those of Strickland (1951b). He differed from Leston (1973) and Majer (1976a) (who both had suggested that encouragement of Tetramorium aculeatum would help lower swollen shoot incidence by reducing vector numbers), arguing that it was the absence of mealybugs which made conditions unfavourable for the other saccarophilic dominants. In such conditions which tended also to be poorly growing trees, often resulting from heavy shade, Tetramorium aculeatum did not have to compete. The demonstration by Majer (1976a) that it spread on to trees from which other dominants had been removed added weight to the argument that the species is not strongly competitive.

Belshaw & Bolton (1993, 1994)

This study by Robert Belshaw and Barry Bolton was not a study of the mosaic but an investigation of the ant populations of the leaf litter layer in primary forest (14 areas), secondary forest (10 areas) and cocoa plantations (10 farms), during the period December 1991 and November 1992. Almost all the sampling was of leaf litter, with examination of soil samples at three sites. Because the important habitat, especially for nesting, of dead wood on the ground was not examined, it was not a complete study of the terrestrial ant fauna. In the first paper, it was found that the three habitats did not differ significantly either in species composition or in species richness. Similarly the geographic separation of the sites, scattered across the forest zone, did not appear to show any great effect on species distribution. The sites are shown in Map 1.

The second paper has the merit, however, of bringing some quantitative information on the leaf litter populations. At each site an area of about 1000m was defined, within which ten 1m quadrats were placed. On a single occasion, all the leaf litter in each quadrat was collected and ants extracted using a Winkler bag. The ants were pooled to form a single total for each site. From all sites and samples, a grand total of 43,824 ants in 197 species and 47 genera were recognised. All these are included in the Species list. Three species were found only in the soil samples and 18 species were felt to be "tourists", either being arboreal species foraging on the ground or Camponotus species, which forage on the surface rather than within the leaf litter.

Contents Mosaics - Introduction
1998, 2003 - Brian Taylor CBiol FSB FRES
11, Grazingfield, Wilford, Nottingham, NG11 7FN, U.K.