The Ants of Africa CHAPTER 3 - Mosaics - Evidence from Cameroun - 3 Watt, Stork & Bolton (2002)
Introduction

The diversity and abundance of ants in relation to forest disturbance and plantation establishment in southern Cameroon
by Allan D. Watt, Nigel E. Stork & Barry Bolton (2002), Journal of Applied Ecology, 35, 18-30.

The study was carried out in plantations of the timber tree Terminalia ivorensis established at Bilik and Ebogo within the Mbalmayo Forest Reserve (3° 23' - 3° 31' N; 11° 25' - 11° 31' E), 650 m above sea level, 9200 ha in size and about 50 km south of Yaoundé (see Map 7). Annual rainfall averages 1520 mm, most falling in two wet seasons and monthly average temperatures range between 23° and 26° C. The area and the different ways of forest management examined to compare the impact on the subsequent plantation were more fully described by Watt et al. (1997). The tree is indigenous to West Africa, including western Cameroun, but does not occur naturally in the Mbalmayo area. A related tree, Terminalia superba, was the fifth most common tree at two of the three project study sites.

Experimental Plots

Ebogo - four 1 ha plots established in 1987; Bilik - five 1 ha plots established in 1988.

• Uncleared forest control - untouched except for minimal clearance of ground vegetation for access lines.
• Partial manual clearance - ground vegetation and some small trees removed by machete, other small trees felled by chainsaw. Caused minimal soil compaction.
• Partial mechanical clearance - most of undergrowth and many large trees removed with a bulldozer, resulting in 50% loss of canopy cover and some soil compaction.
• Complete clearance - all large trees felled by chainsaw, then whole site cleared by bulldozer; leaving no above ground plant material and causing considerable soil compaction.

Following the clearance there was substantial weed growth, primarily of Eupatorium odoratum (Compositae), Musanga cecrepoides (Moraceae) and Scleria barteri (bush razor, Cyperaceae).

Sampling techniques

Insecticide knockdown sampling was done in November 1992, February 1993 and November 1993 at both sites. The ground vegetation was cleared from around a number of trees and 1 m² collecting trays were suspended from ropes tied between the trees. A 250 ml collecting bottle, with a small amount of 70% alcohol, was attached to each tray. The trees were then fogged with permethrin insecticide, using an SN11 Swingfog fogger, and the insects left to fall from the canopy for 2 hours. The trays were then sprayed with 70% alcohol to wash the insects into the collecting bottles.

In November 1992, 15 trays were used in each plot. This was increased to 25 trays per plot for the later surveys, except in the complete clearance plot at Ebogo in February 1993 when 40 trays were used. Before fogging, the ground was partially cleared in the central seven (out of 20) lines of Terminalia in each plot. Five lines were chosen at random and then a sampling area was selected at random from the middle half of each line. The sampling areas contained three, occasionally two, trees and one or more trays were hung under the canopy of each. Although fogging was targetted at the Terminalia trees, it was felt that the sampling reflected the abundance of arthropods on the target trees, on other older trees and "tourists".

Leaf litter - the sampling of leaf litter for ants was carried out in November 1993 in all three plots at Ebogo; a complete clearance plot at a third site Eboufek; and in uncleared forest plots at Ebogo, Bilik and Eboufek. Ten 1 m² samples of leaf litter and top soil were collected from a 50 m transect across the Terminalia rows, with half the samples from tree rows and half from between the rows. The samples were sieved and then placed in "Winkler bags", to dry for two days, causing the insects to drop down into an alcohol pot at the bottom of the bag.

Ant findings

Ants made up 63% of the overall total of arthropods, with a mean of 136 ants per m². At Bilik there was no significant difference in the number of ants from each of the silvicultural treatments, but at Ebogo the partial manual clearance plot had considerably more ants than the other plots (12-fold in 1992 and 5-fold in 1993). This hyperabundance distorted the findings so that the average number of ants was much greater at Ebogo than Bilik (343 and 45 per m² respectively); without the partial manual plot the Ebogo average was 46 ants per m².

Mean number of individual ants per m² tray in the knockdown samples
(weighted to allow for there being 15 collecting trays per plot in Nov 1992 and 25 thereafter) - from Table 5 in Watt et al. (1997).

Number of ant species in the knockdown samples
15 collecting trays per most plots in Nov 1992 and 21-35 in February 1993 - from Table 2 in Watt, Stork & Bolton (2002).

I have combined the information in Watts, Stork & Bolton (2002) to show the abundance of the full list of canopy ants from all the samples in the graph below and have included the overall ant species and forms (morphospecies) in the linked Cameroun Summary Table.

Watt, Stork & Bolton (2002) provided species accumulation curves for the canopy and leaf litter ants. Observing that the curves for almost all the plots were still showing an upward trend, rather than flattening out, they made jack-knife estimates for each plot.

The most frequent species, the undescribed Technomyrmex 2, numbered 80% of all individuals in the knockdown collecting trays. It, however, was numerically predominant in the "Manual clearance" plot at Ebogo, and the authors suggested that might have had a negative effect on the number of species in that plot. Assuming Bolton has used the same denominations for undetermined forms in all his papers, this Technomyrmex species 2, which also was recorded in the Mblamayo leaf litter, was previously known solely from Cubitermes (termite) nests (Campo and other forests - Dejean, Bolton & Durand, 1997) and not on vegetation or canopy in Cameroun. Technomyrmex species (indet.) (2) was described as widespread, 168 workers from eleven sites, in Ghana semi-deciduous forest leaf litter by Belshaw & Bolton (1994b).

What these authors have not done is to examine their results in terms of the mosaic, i.e. which species were associated in the individual sample trays. A useful comparison could then have been made with the primary forest canopy study at Campo - see Dejean Campo studies. At Campo, 167 climax trees yielded up only 26 species of ants, primarily due to the overwhelming dominance of Crematogaster depressa, which tolerated only nine other species in its territories.

As the oldest plots were less than six years old, my feeling is that what Watt and his colleagues found was the result of opportunistic occupation, on immature trees. Their findings have relevance in helping to understand the problems of new plantations but do not tell us much more than that. Comparing their findings to my synthesis of the ant mosaic on cocoa, the Terminalia plots were probably a "NO DOMINANT" situation at the early stage of an open (unjoined) canopy.

© 2003, 2012 - Brian Taylor CBiol FSB FRES 11, Grazingfield, Wilford, Nottingham, NG11 7FN, U.K.

href="mosaic11b"