SOIL ORGANISMS - Soil Ecology - page 314

Australian wood feeding species Coptotermes acinaciformis is excluded are those formed
on deep sands (Gay and Calaby, 1970). The influence of fertility on species richness is
discussed below.
Species and generic richness
The most species-rich communities occur in certain humid tropical rainforests. Eggleton
et al. (1996) found 114 species in the moist pre-montane Mbalmayo Forest Reserve in
the Cameroon of which 48 species were recorded during the quantitative sampling of
two 1 ha plots in little-disturbed rainforest. However, as noted below, the greatest species
diversity occurred in a disturbed environment. In reviewing a number of studies from
Amazonia, Martius (1994) considered that species densities of 70 to 90 species were likely
to occur throughout the Amazonian rainforests reducing to 11-12 species in the seasonally-
flooded forests. Most other rainforests studied have lower species richness and Collins
(1983) found 43 species in South-East Asian rainforest in Sarawak. The mesophytic rain-
forests of northeastern Australia are particularly species poor, for zoogeographic reasons,
only four termite species occur regularly in these forests (Gay and Calaby, 1970).
Species richness in savannas is generally somewhat lower than in rainforests and Josens
(1983) recorded from 19 to 36 species at eight African locations studied, although slightly
more were found in some Australian savanna sites (Figure III.48a), as discussed below.
Plant community structure and certain soil properties clearly influence species rich-
ness and the relative balance of ecological categories. Of the 50 species recorded from
communities studied at Kakadu (tropical Australia), numbers varied from 16 to 36 over
five environments which ranged from monsoon forest (a dry rainforest facies) which had
23 species to a wet rockland which had 16. The most diverse environments were the open
forest and woodland savanna sites which had, respectively, 36 and 35 species followed
by the dry rockland which had 26 (Braithwaite et al., 1988).
Josens (1983) reported the existence of a positive relationship between species
richness and rainfall in African savanna environments. However, this was not confirmed
across the 30 tropical Australian sites studied by Braithwaite et al. (1988) and is unlikely
to hold generally across northern Australian environments because of the reduced
number of species in rainforest communities.
Termite species richness has been shown to be inversely related to soil fertility in
Australian environments (Braithwaite et al., 1988; Stafford-Smith and Morton, 1990)
and Goodland (1965) reported a negative correlation between soil fertility and the number
of termitaria in savanna environments in northern South America. Across 30 tropical
Australian sites representing five habitat types, Braithwaite et al. (1988) found strong
negative correlations between species richness and soil N and P concentrations. In contrast,
species richness in the live-wood feeding termites of these sites was positively correlated
with both these variables. Similarly, in the Malaysian rainforests studied by Salick and
Tho (1984) species richness, guild richness and population densities were all positively
correlated with a soil productivity index. Eggleton et al. (1996) found strong positive
correlations between termite abundance and the concentration of 'available' soil N in a
disturbance series of study plots in a rainforest environment in the Cameroon. Other soil-
based limitations to termite community structure are considered in Chapter IV.
Changes in land use from indigenous to man-modified landscapes normally result in