To identify the properties of habitat mosaics produced by fire that enhance the persistence and status of invertebrates in eucalypt-dominated mallee habitats.

Rather than attempting to determine responses to fire by “invertebrates in general” from broad taxonomic surveys of all mallee invertebrates, this project chose to focus on a small number of what we believe are keystone groups; namely termites (Isoptera), psyllids (Hermiptera), scorpions (Scorpionida) and centipedes (Chilopoda). We believe these will be the most informative invertebrate subjects for study.

Termites and psyllids are food sources for insectivorous animals and abundances of these invertebrates may in turn influence vertebrate populations. Wood-nesting termites also generate hollows in trees and coarse woody debris which may be key habitat features for many other animals. There are approximately 16 species of termites found in the mallee areas of the eastern states and it is likely that they play important ecological roles. In addition, scorpions and centipedes are relatively large invertebrate predators that may potentially have a profound influence on structuring invertebrate communities.



There are approximately 16 species of termites likely to occur in the mallee habitats of Victoria, South Australia and New South Wales. The majority are subterranean, wood-feeding species, although some nest in trees and others forage upon grasses and litter. Therefore several different techniques for sampling termites were employed; active searches along transects and the use of baits. Baiting is less labour intensive than searches and allows data collection on a larger scale, as was required for this project. The disadvantage is that not all species may be attracted to the baits. Baits Toilet rolls have been found to representatively sample species richness and composition and relative abundance. The toilet rolls used were 400 sheet, two ply, bleached and unscented. Six toilet rolls, spaced 5 m apart, in 3 x 2 grid layout were placed at 20 sites within each of the 28 mosaics. The sites were selected to proportionally represent the fire ages present in the mosaic; therefore if a fire age patch covered 50% of the area of the mosaic then 10 of the 20 sites were placed within that fire age. In addition, sites were placed within all four quadrants of the mosaic, when possible, and in proportion of the topographic variation present (dunes and swales). Rolls were buried level with the soil surface. Each roll was tied with a length of yellow or blue flagging tape protruding above the soil surface so that the individual rolls could be re-located in the field. A total of 120 rolls were buried per mosaic, giving a total of 3360 rolls. Rolls were installed in July/August 2006 and checked in October/November 2006. During sampling, each roll was lifted and checked for termite presence. If present, any soldier and worker castes were collected into a vial of 70% ethanol for later identification. Each grid of six rolls was considered an independent sample and therefore termites collected from each roll in the grid were placed into the same labelled vial. Active searches Active searches for termites were carried out within each sample mosaic. Baits may attract only a particular set of species and active searches, while more time consuming, provided additional information about the termite species present. The active searches took place in October and November 2007. Active searches for termites were conducted using a previously established transect method with minor modifications for mallee conditions. Active search transects were placed 10 -15m parallel to pitfall lines, using the same transects as those used for habitat assessments. Five 50m x 4m transects were sampled in each mosaic (n = 140 transects across entire study), at pitfall sites that represented each fire age class present in the mosaic and in proportion to the area where possible. All possible termite habitats along the transect were searched by two people for up to one hour. This included soil cores (20 cores per transect, 10cm3), all dead wood lifted, broken up and searched, hollow and dead branches removed from trees and trees examined for nests to a height of 2m. Any termites found were recorded and specimens of soldier and worker castes collected into vials of 70% ethanol for later identification.


Each of 28 fire-induced mosaics was surveyed for psyllids at 10 sites (n = 280 sites). The location of the sites was stratified based on the proportion of each fire age class in the mosaic. Where possible, the sites were placed to include 1) topographic variation within each fire age class (dune/swale), 2) at least one site in each quadrant of the study mosaic and 3) all sites at least 100m from fire boundaries, at least 25m from a track and 200m apart. Multi-stemmed eucalypt species were identified at each site and four trees were selected for sampling. The selected trees were chosen to be representative of the proportion of each tree species in the local area, for example if all surrounding trees were Eucalyptus dumosa then the four selected trees were E. dumosa. From the selected trees, 20 leaves per tree (10 from the north-facing aspect and 10 from the south-facing aspect) were observed for lerp presence and the number of leaves with at least one lerp present was recorded. Lerp appearance was described to facilitate later identification. Voucher specimens of lerp observed were collected for identification. If possible, adult psyllids were collected for identification. Adults were collected by branch tapping over a white tray to dislodge them from the tree. They were then placed into vials of 70% ethanol.

Centipedes and scorpions

Each of 28 fire-induced mosaics was surveyed for scorpions and centipedes at 10 sites (n = 280 sites). Scorpions and centipedes were collected from pitfall traps at each site. The pitfall traps consisted of 10 x 20l plastic buckets buried to surface level along a 50m transect. The traps were placed 5m apart and had drift fencing along the transect length. The traps were opened for five consecutive nights and the scorpions and centipedes were collected each morning. They were placed into 70% ethanol for later identification and quantification. Sampling took place in late spring/summer 2006 (October – December).