How & Dell: Reptile assemblages of the northeast Pilbara 
the 180 sampling days of James (1994). An exceptionally 
high lizard richness of 57 species has also been 
documented on Lorna Glen Station in the eastern 
Gascoyne bioregion (Cowan pers. comm.). The duration 
of sampling thus plays a very significant role in defining 
species richness and the use of species accumulation 
curves has been proposed to define the probable richness 
of sites sampled over a limited temporal span (Thompson 
& Withers 2003; Thompson et al. 2003b). 
The diversity of the herpetofauna in the study area 
probably reflects the juxtaposition of diverse habitats, 
such as the extensive areas of Triodia spp. grasslands on 
sandy soils, with isolated rockpiles and breakaways, and 
also the loamy soils and fringing woodlands of creeks 
and drainage lines. Forty-one lizard species were 
recorded from the eight main sampling sites and an 
examination of relationships between lizard assemblages 
on sites (Fig 2) provides some support for this 
proposition. The assemblage on the loams associated 
with riparian vegetation at WS1 is quite dissimilar to the 
remaining sites that are located on sands dominated by 
Triodia. Additionally, the site close to a rocky breakaway, 
WS8, also has a distinct assemblage from those on the 
surrounding sandy plain. James (1994) recorded between 
23 and 32 lizard species at the subsites in central 
Australia, considerably more than the 17-25 recorded in 
this study. However, similarity between subsites was 
high at around 58% (James 1994), such that only 40 
species were recorded in total compared to the 41 in this 
study. Further differences in the two studies occurred in 
the period of peak activity of reptiles. In central 
Australia, activity was greater in spring than autumn 
Games 1994), while on the Abydos Plain activity was 
greater in late summer than in spring (Table 3). 
Thompson et al. (2003a) contrast species diversity and 
evenness across 12 subsites (biotopes) at Bungalbin and 
10 at Ora Banda, both in the Eastern Goldfields of 
Western Australia, and indicate that species richness 
varies from 17 to 26 in the former and 14-28 in the latter. 
They concluded that 'These data indicate considerable 
variation in reptile assemblages for closely related sites 
both with and without similar soils and vegetation". 
Pianka (1986) suggested that the richness of Australia's 
arid herpetofauna resulted from the influence of several 
environmental factors, such as the interdigitating of 
several habitats, presence of the ubiquitous and unique 
Triodia spp. grasses, a variable and unpredictable rainfall 
and the replacement of the role of mammalian predators 
by reptilian taxa. Morton and James (1988) proposed a 
multi-causal scheme to explain the diversity and 
abundance of Australian desert lizards that invoked an 
interaction between climatic, edaphic and (a complex of) 
biotic factors. Their synthesis was re-examined by Pianka 
(1989) who evaluated eleven causal factors that made a 
contribution to the richness of desert lizards, including 
the role of fire and the biogeographic history of regions. 
Pianka concluded that many links in Morton and James' 
causal network were well established, but for others the 
evidence was weak. Certainly, the role of fire in creating 
a complex of serai community stages is significant in 
enhancing lizard diversity (see below). 
The reproductive activity of reptiles from the 
Woodstock/Abydos study area indicated that most 
species were reproductively active during the spring and 
summer (How et al. 1991) and that reproductive activity 
may be curtailed when environmental conditions had 
been unfavourable for many months. How et al. (1991) 
also concluded that the lower incidence of reproductive 
activity seen in March 1988, compared with subsequent 
February-March samplings, was the result of a long 
drought that preceded that sampling. James (1991b) 
stated that most species of lizard in the central Australian 
study site reproduced only once during the three year 
study with most species failing to reproduce in a 'dry' 
year. It is highly likely that many reptile species in the 
Pilbara reproduce opportunistically following improved 
environmental conditions after episodic rainfall events 
and thus show similar stochastic population responses 
following rain to those reported by James (1994) in 
central Australia. 
The above average rainfall during the first two 
summers of this study promoted rapid growth and 
seeding in Triodia spp. and other grasses, while 
numerous shrubs and trees produced vegetative growth 
and flowers. The impact of this was to provide visibly 
improved trophic resources for vertebrates in the form of 
new plant growth, grass seeds and insect prey, and 
contrasted with conditions observed at the very 
beginning and the end of our study period. Amphibians 
responded almost immediately to the March 1988 deluge 
and the breeding chorus was extensive over the study 
area. 
The diets of three sympatric Ctenotus skinks at 
Woodstock were examined (Twigg et al. 1996) and 
indicated that, despite a large dietary overlap, two of the 
species showed dietary separation on a seasonal basis. 
James (1991a) also showed a seasonal change in diet 
within species of Ctetiotus in central Australia that 
responded to changes in abundance of their main dietary 
prey. These seasonal shifts in dietary pattern in selected 
abundant species of lizard could conceivably be 
extrapolated to less abundant taxa and, when coupled 
with improved conditions for reproduction, could 
account for many of the long-term changes in abundance 
of species in arid study sites. The long-term appraisal of 
the lizard assemblages in the Great Victoria Desert 
showed only minor changes in use of resource states by 
lizards when examined nearly 12 years apart, and 
although some taxa increased in number between these 
periods, others declined but the general use of space, diet 
and time remained similar (Pianka 1986). 
Fire has played an important part in structuring arid 
zone assemblages (Pianka 1989), and at Abydos/ 
Woodstock fire has previously been the focus of 
considerable research (Suijendorp 1967). However, with 
knowledge of the regeneration rate in areas burnt in 
January 1989 and January 1990, it was apparent that at 
least ten years had elapsed since the major sampling sites 
were previously burnt. Other studies of the impact of fire 
on arid zone reptiles (Caughley 1985; Fyfe 1980; Masters 
1996) have relied on examining the assemblages in areas 
burnt at various times prior to the study. In our study 
the same major sampling sites were monitored for two 
years before and one year after burning (Table 4). 
Pianka (1989) has reviewed the available data on the 
impact of fire on desert reptile assemblages and 
concluded that it played a key role in contributing to the 
richness of lizards. The impact of fire on the lizard fauna 
93 
