Introduction The Ants of Africa
CHAPTER 3 - Mosaics - Evidence from Ghana - 5
Ghana - 1 Ghana - 2 Ghana - 3 Ghana - 4 Ghana - 6

Mike Bigger was based at CRIG between 1969 and 1976, where like Campbell (see Mosaics - Ghana - 6), he was entomologist with a team investigating the biology and population dynamics of mealybug vectors of cocoa swollen shoot virus.

Bigger (1975)

In an early report, he described investigations of 14 outbreaks of cocoa swollen shoot virus (transmitted by mealybugs), the results of which were interpreted by use of the Canonical Variate Analysis technique (Bigger, 1975). His brief summary was that four conventional dominants, Crematogaster africana, Crematogaster clariventris, Crematogaster striatula and Oecophylla longinoda formed the major groups in the analysis, with a fifth group being non-dominated, although Pheidole megacephala was notable. Ecological factors, tree girth (maturity), canopy area, shade and other insects seemed not to be major sources of variance. Perhaps this is not surprising as the investigations were of specific circumstance, i.e. those which were favourable for swollen shoot vectors. Interestingly, Bigger noted how "C. striatula trees are small, with poor canopy and are low in most associated insects", and that this was in marked contrast to Majer's (1974, cyclostyled report) description of Crematogaster striatula as being found in high yielding dense canopy cocoa, with more associated insect species than any other dominant.

Bigger also described how the predominant mealybug species at CRIG was then Planococcus citri and not Planococcoides njalensis as had been the case at when Strickland (1951a) surveyed the CRIG cocoa. This Bigger attributed to the general decrease in overhead shade and an increase in the use of Amazon cocoa (replacing Amelonado cocoa).

Bigger (1981a)

Bigger's main study was to follow the pattern of development of the insect fauna of newly planted cocoa (see Bigger, 1993b, below). During the preparatory stage for that study (Bigger, 1981a), he took the opportunity to thoroughly investigate a block of 25-year old Amelonado cocoa at CRIG. This originally had been planted in 12 square blocks each of 10 X 10 cocoa trees, and it is interesting to note (Bigger, 1981b) that the site probably was that used by Strickland (1951b). Six blocks were shaded (with planted Gliricidia septum), three were unshaded and three were semi-shaded. The blocks were separated by wider inter-row spaces but where the trees were in good condition the canopies were linked across the gap. Only in the shaded area, however, was the canopy in good condition, much of the other cocoa was degraded. Three sampling methods were used - the ants foraging on leaf litter were manually collected from five 54 m plots by a collector crawling on hands and knees for one hour (between 0900h and 1000h on bright mornings) in each area; canopy ants were sampled by pkd, with trunks encircled by boards to give a total catch area of 10 m in each of the five plots; and, lastly, immediately after the pkd, the cocoa trees in each plot were cut down and examined for sedentary insects. The survey was completed by a visual examination of the more important ants present on all 1200 trees. This revealed that almost the entire block was dominated by two species, Crematogaster clariventris on 84% of cocoa trees and Oecophylla longinoda on 17%. Oecophylla longinoda occupied two distinct territories and Bigger suggested that this comprised two separate colonies. He was unable to determine whether there was more than one colony of Crematogaster clariventris, but felt that this was likely because of the large area it occupied. There was some overlap of the species along one edge of the larger Oecophylla longinoda colony and in the smaller colony. Only 5% of the trees had neither species. In the areas of better canopy and mostly where Oecophylla longinoda was absent, other common ants were Tetramorium aculeatum, Cataulacus guineensis and Polyrhachis decemdentata. Crematogaster castanea was evenly distributed throughout the block.

The pkd samples included a total of 34 species of ant. Of these, only three were present in large numbers. Two plots were from Crematogaster clariventris territories, a shaded plot (I) with 1572 workers and an unshaded plot (III) with 803 workers. In those plots, Tetramorium aculeatum numbers were 176 (I) and 222 (III). Worker numbers from the three Oecophylla longinoda plots were 6018 from a shaded plot (II), 2074 from an unshaded plot (IV) and 1021 from an unshaded plot with a degraded canopy (V). On (V), Tetramorium aculeatum workers numbered 105, but these may all have come from a single tree without Oecophylla longinoda. Crematogaster castanea was not found in the Crematogaster clariventris plots, suggesting that the associations found by Room (1971) were true here also, even though numbers in the three Oecophylla longinoda plots were only 13, 8 and 9. More Polyrhachis decemdentata were found with Oecophylla longinoda (48, 15 and 5) than with Crematogaster clariventris (4 and 3). Cataulacus guineensis appeared to be unaffected by the two dominants with 16 (I), 17 (III), 0 (II), 25 (IV) and 10 (V). Of the other 28 species, numbers were mostly in single figures and the findings are listed in the species section. In one Oecophylla longinoda plot (IV), 77 workers of Pheidole megacephala were found (I suspect this was a single tree colony).

The leaf litter collections gave a total of 45 species of ant. Apart from ground moving or foraging workers of the two dominants, Crematogaster clariventris with 75 workers (I) and 5 (III), and Oecophylla longinoda with 143 (II), 23 (IV) and 60 (V), moderate numbers were collected of

Lepisiota (as Acantholepiscapensis, 19 (V) Lepisiota (as Acantholepis) spinosior, 24 (III)
Lepisiota (as Acantholepis) species A, 11 (I) Lepisiota (as Acantholepis) species B, 16 (I)
Camponotus acvapimensis, 9 (I), 1 (II), 12 (IV) and 19 (V) Camponotus flavomarginatus, 16 (III)
Cataulacus guineensis, 3 (III) and 14 (V) Diplomorium longipenne, 13 (II)
Oligomyrmex species, 18 (I) Paratrechina species 3, 11 (I)
Pheidole species A, 27 (I) and 60 (III) Pheidole species K, 26 (V)
Pheidole species M, 18 (II) Technomyrmex species A, 17 (I) and 2 (IV)
Tetramorium species A, 4(I), 12 (III), 14 (IV) and 2 (V) Tetramorium species L, 57 (II) and 3 (III)
Tetramorium (former Xiphomyrmex species B), 3 (I), 14 (II), 9 (III) and 6 (IV) Tetramorium species (as Triglyphothrix species K, 42 (II) and 5 (III)

The ants could be separated into three groups, 17 species were found on both canopy and leaf litter, 17 species on the canopy only, and 28 on leaf litter only. With hindsight, Bigger noted that the results of the leaf litter and canopy sampling may have been influenced rather more by shade than by the dominant ants. He noted that Crematogaster striatula was found only by the leaf litter sampling, apparently foraging from outside the area and from some distance away.

He also tabulated results of the incidence of eight of the "more abundant ant species in surveys of cocoa carried out in West Africa" - Oecophylla longinoda, Tetramorium aculeatum, Crematogaster africana, Crematogaster buchneri, Crematogaster castanea, Crematogaster clariventris, Crematogaster depressa and Crematogaster striatula. The merit of his analysis, however, was diminished by some misinterpretation of others' findings. For instance, he cites Room (1971) as collecting from "168 scattered sites", but these were canopy samples from 30 sites, and were biased deliberately by Room to include samples from 30 trees dominated by each of five dominants. The work of Strickland (1951a) was all at CRIG and not "S. Ghana". The results from Nigeria (Taylor, 1977; Taylor & Adedoyin, 1978) also were incorrectly cited, especially with regard to data on Crematogaster africana which was separately counted in two of the five instances, and lumped with Crematogaster depressa only in the two large scale surveys. Bigger's own results are curious in that nine of the listed samples have less than 100% of trees occupied by the eight selected "abundant" species - in one sample that total was only 13.7%, but Bigger made no comment on this. He did remark that Crematogaster clariventris was far more common at CRIG, including the specific study which featured in this paper, than elsewhere. In part, he attributed this to the clearance of large shade trees and the ability of Crematogaster clariventris to nest on shorter trees. That remark, however, highlighted a shortcoming of the distribution of his "29 scattered sites" in the virus outbreak survey - he noted that 11 of the 13 sites occupied by Crematogaster clariventris were at Tafo, many in the vicinity of the Amelonado block. Thus, his use of "sites" seems to be misleading when compared to Room (1971) and Taylor & Adedoyin (1978), who surveyed farms scattered across wide areas of southern Ghana and Western Nigeria respectively. Unlike most other authors, he doubted that Tetramorium aculeatum is a true dominant, particularly because he regarded it as frequently sharing territory with other dominants.

Bigger (1981b, 1993a, b)

The remainder of Bigger's reports all stem from observations made on the Amazon cocoa planted, initially under temporary shade, in the cleared Amelonado area.

Perhaps the most interesting report, as far as the ant mosaic is concerned, is the most recently published (Bigger, 1993b). This describes the evolution of ant and Homopteran occupation of cocoa over a period of five years. The presence/absence of 23 species of insects was recorded for every tree of each of four adjacent blocks of Amazon cocoa - with 238, 204, 200 and 224 trees respectively, and only 18-months old at the start of the study. Bigger (1981b, 1993a) describes making 306 consecutive weekly census counts in the block of 238 trees between 1971 and 1976. Presumably, the same regularity and frequency applied to the recording in the other three blocks.

Ants do not feature greatly in the first paper (Bigger, 1981b) which concentrated on the mealybugs. Planococcoides njalensis was the more common mealybug early in the study but declined towards the end. This he attributed to its reliance on attendant ants, especially Crematogaster castanea, which was displaced by Crematogaster clariventris; thus, confirming the earlier finding by Strickland (1951b).

The later paper (Bigger, 1993a) reports a time series analysis of the variation in the studied insects across the nearly six years. An interesting feature is that he reports results not only for relatively well-studied species, Oecophylla longinoda, Tetramorium aculeatum and Crematogaster (only at genus level), but also Polyrhachis laboriosa, Pheidole megacephala and Camponotus (also only at genus level). The only marked seasonal variations in abundance, however, were recorded for Crematogaster species, which followed the annual variation in morning relative humidity. The aphid, Toxoptera aurantii, showed a variation in abundance which followed or paralleled the number of new shoots per tree, a feature of this species being its capacity for very quick reaction to increased availability of food, due to its short generation time.

Bigger (1993b) examined Ant-Homopteran associations and the changing pattern of infestation as the cocoa matured. As before, the ants were treated as being in five major groups -
Camponotus species (Camponotus acvapimensis and Camponotus flavomarginatus), both regarded as sub-dominants;
Polyrhachis laboriosa, sub-dominant;
Oecophylla longinoda, dominant;
Tetramorium aculeatum, dominant;
Crematogaster species (Crematogaster africana, dominant; Crematogaster castanea, sub-dominant; Crematogaster clariventris, dominant; Crematogaster species near striatula, dominant; and Crematogaster (Sphaerocrema) species, sub-dominant)
and four minor species, all excluded from the association analyses - Pheidole megacephala (described as important only "because it is an important mealybug tender"); Lepisiota (as Acantholepis) species; Tapinoma species; and Monomorium species - the latter three were described as being rare.

The ant-homoptera associations fell into three groups -
Crematogaster species with Planococcoides njalensis;
Oecophylla longinoda with Stictococcus species;
and, Camponotus species and Polyrhachis laboriosa with Planococcus citri and Toxoptera aurantii.
Why the Crematogaster and Camponotus species were not recorded individually was not explained, and is curious as they are quite easy to separate by eye.

The changes in ant species with time was as follows:
Camponotus species were found on at least 50% of the trees, although in Block I there was a steady decline from > 80% to about 40% of trees. Bigger regarded the overall situation as "unusual" in having a large number of trees unoccupied by the usual dominants and it was the lacunae which were filled by the Camponotus.
Polyrhachis laboriosa was low in numbers until the trees were some 4 > years old, after which there was a steady increase.
Oecophylla longinoda was present through years 3, 4 and late 5 but then its numbers declined to nil.
Tetramorium aculeatum was on 25-30% of trees through years 2, 3 and 4 but then dropped to < 10%.
Crematogaster species occupied < 20% of trees until year 5 when numbers increased to about 60%.
Pheidole megacephala was present on 5-10% of trees throughout, with some seasonal variation, mainly a drop in winter, when temperatures were 1-2 C below the annual mean.

Although, as mentioned above, he stated the Crematogaster species were not separated, comments were made of various differences between the five named species. Crematogaster africana was more or less constantly present in Block III, primarily (or wholly) because there was a colony centred on a large nest on a Ficus tree in the Block. He described how Crematogaster clariventris was not dependent on shade trees for nests and so can become dominant in unshaded cocoa. Moreover, it and Oecophylla longinoda both moved centres of occupation around the other Blocks. Crematogaster castanea was described as being entirely an inhabitant of the canopy. Crematogaster species near striatula was described as having very similar behaviour to that known for Crematogaster striatula but not so aggressive. The Crematogaster (Sphaerocrema) species had its nests under bark. In his discussion, Bigger acknowledged that a factor affecting the ants was probably the increasing tree maturity and increasing overshade from the growing "shade" trees, Gliricidia septum, and tree cassava. The overshade increased from 12-45% in February 1972 to some 90% by April 1975.

Contents Mosaics - Introduction Mosaics - Ghana - 6
1998, 2003 - Brian Taylor CBiol FSB FRES
11, Grazingfield, Wilford, Nottingham, NG11 7FN, U.K.