2 
A. N. ANDERSEN, B. A. MYERS AND K. M. BUCKINGHAM 
METHODS 
Ants were sampled between 1983 and 1985 by 
collecting them directly from the ground and 
from vegetation, and by pitfall trapping. Pitfall 
traps were 35 mm diameter plastic vials, partly 
filled with 70% ethanol, which were buried with 
their rims flush to the soil surface. Traps were 
similar to those used at Wilsons Promontory 
and Wyperfeld, and the catches are likely to pro¬ 
vide a good indication of the relative abundance 
of species on the ground (Andersen 1983, 1986a, 
1991b). The traps were arranged in a 5 x 6 grid 
with 5 m spacing (area of grid 500 m 2 ), and oper¬ 
ated for 4-day periods during January 1983, 
October 1983 and May 1985. Hand collections 
were conducted during these periods and also on 
other occasions covering all seasons. Ant species 
captured in traps were scored according to a five 
point abundance scale (1 = 1 ant; 2 = 2-5 ants; 
3 = 6-20 ants; 4 — 21-50 ants; 5 = > 50 ants) 
in order to reduce distortions caused by large 
numbers of ants falling into a few traps (see 
Andersen 1991b). These abundance scores were 
used directly as counts when calculating relative 
abundances. 
Most of the ants collected could not be ident¬ 
ified to species with certainty because of our 
generally poor taxonomic knowledge of Aus¬ 
tralian ants. Where possible, these species were 
assigned to informal species-groups (denoted by 
inverted commas, eg. Camponotus ,, claripes ,r ) 
derived by the senior author from type speci¬ 
mens held in the Museum of Victoria and in the 
Australian National Insect Collection, CSIRO 
Division of Entomology, Canberra (see Ander¬ 
sen 1991 a). A complete set of voucher specimens 
is held by the senior author. 
In the absence of a published biogeographic 
treatment of the Australian ant fauna, each 
species was judged to belong to groups with 
Bassian, Eyrean or widespread distributions ac¬ 
cording to the senior author’s understanding of 
Australian ants. The pattern of community or¬ 
ganisation was investigated by classifying spe¬ 
cies into functional groups according to their 
presumed habitat requirements and competitive 
interactions. This classification is modified 
from Greenslade (1978), and has been used ex¬ 
tensively in studies of Australian ant com¬ 
munities (see Greenslade & Greenslade 1989, 
Andersen 1990, 1991c). 
The biogeographic profile and pattern of com¬ 
munity organisation of the Long Forest mallee 
fauna was compared with those of Wyperfeld 
and Wilsons Promontory. The species lists from 
these other localities were obtained from de 
tailed studies of small plots, as at Long Forest 
mallee. The Wilsons Promontory data are from 
two woodland sites each of approximately 0 25 
ha (Andersen 1986a, 1988), and the Wyperfeld 
data are from adjacent heath and mallee sites 
each of 0.13 ha (Andersen 1983,1984, Andersen 
& Yen 1985). sen 
RESULTS 
A total of 77 species from 21 genera were col 
lected (see appendix), with 44 of these caught in 
traps. The species accumulation curve (F^e. n 
suggests that the number of species occurring in 
the 500 m 2 trapping grid was in excessof 50 The 
richest genera were Camponotus (\ 2 species) 
Monomorium (11), lridomyrmex(Z\M m eck 
(7), Melophorus (5) and Pheidole ( 5). The mean 
number of species per genus was 3.7, midway 
between that at Wilsons Promontory (2.9) and at 
Wyperfeld (4.6). Iridomyrmex, Camponotus inti 
Melophorus together contributed 32% of total 
species at Long Forest mallee, again midway 
between Wilsons Promontory (23%) and Wvner- 
field (42%). 
The biogeographic profile of the Long Forest 
mallee fauna resembles that at Wyperfeld far 
Fig. 1. Accumulation of ant species in pitfall traps(500 
m 2 grid) at Long Forest mallee. Each point represents 
the running total of ten traps (circles = January 1983; 
triangles = October 1983; squares = May 1985). 
Number of species records is the sum of the total num¬ 
ber of species recorded in each trap, regardless of 
species turnover across traps. 
