Date of Award


Degree Type


Degree Name

Master of Science


Faculty of Communications, Health and Science

First Advisor

Dr Pierre Horwitz

Second Advisor

Dr Mark Harvey


Long term studies encompassing seasonal variation in abundance and species present, forming base-line phenological data, arc required to understand the responses of the carabid (Ground Beetle) fauna to major habitat changes. In addition, few habitats unaffected by human disturbance remain in Australia, and evaluation of these fauna in these areas is required before further habitat loss occurs. Generally an unknown group of terrestrial predatory invertebrates in Australia, carabids have been the subjects of few ecological studies in this country. This restricts the utility of the Carabidae in environmental or conservation assessment programs in Australia. A study was initiated to assess the effects of habitat fragmentation and to provide base-line phenological data on the carabid fauna occurring in remnant bushland areas within part of the Perth Metropolitan Area, encompassing the Quindalup, Spearwood and Bassendean Dunes Systems, and Ridge Hill Shelf geological landforms of the Swan Coastal Plain. An intensive study, documenting and comparing seasonal occurrence of carabids was also carried out in bushland remnants within the Quindalup Dune System. Pitfall traps from a Western Australian Museum survey and from the author's fieldwork, accessed 39 sites in 14 remnant areas across the four geological landforms of the Swan Coastal Plain concurrently in 1993 through to 1997. A total of 3049 specimens of 37 species representing 26 genera and 11 subfamilies were collected. The richest geological landform was found to be the Bassendean Dune System with 17 species, followed closely by the Quindalup Dune System (12) and the Ridge Hill Shelf (14 species). Relationships were found between the carabid fauna and the size of the remnant areas. The r values of the regressions between the total number of carabid species (r = 0.3782, p< 0.05), and the number of volant species (r = 0.3776, p< 0.05) and the Jog of remnant area were statistica11y significant. However, these r values are very low and indicate that only about 14% of the variation in total and volant species richness is accounted for by remnant area. The non-volant species richness and log of remnant area correlation was not significant (r = 0.1912, p>0.05). In this case, the variable remnant area accounted for less than 4% of the variation in the non-volant species richness. The distribution of volant to non-volant species across the Plain was highly irregular, with volant species represented usually by few individuals in either the Quindalup or Bassendean Dune Systems whereas the non-volant species were generally more common and widespread. This is probably attributable to the collection method rather than a reflection of real distributional patterns. A series of environmental parameters were generated by the climate program Bioclim. Most of the carabid species present on the Plain are at the extremity of their ranges and the physical and environmental parameters were found to have limited influence on species richness, individual species' abundances or distributions spatially across the Plain. Environmental parameters were also scored for each sampling period for the Quindalup Dune System sites between 1996 and 1997 to determine their influence on the species distribution temporally and spatially in that dune system. As previously, these parameters had little apparent influence on either the non-volant assemblage species richness or individual species abundances. Evidence of seasonality of several species was observed, along with possible spatial and temporal partitioning between two species, Scaraphites lucidus and S. silenus. The former was only found on Quindalup and associated soils, and the latter found in most other remnants across the Plain. Similarity classifications revealed that it is possible to discriminate the broader geological features of the Swan Coastal Plain on the basis of the entire carabid assemblage. However, finer-grained discrimination is possible if only the non-volant assemblage is used. Within the Quindalup Dune System discrimination between the remnants is not as clear, with sites tending to cluster in three main groups, beach associated, heath associated and older areas. Distinct seasonal activity levels (within a six week window) were documented for the first time for several common species, with only two species active at some level through all seasons. Most species were found to be active in the adult form during spring and summer months, activity tapering off in autumn. Slight variations between remnants were observed. Redescriptions of eight non-volant species and Gnathoxys pannuceus sp. nov. (with details of gross male genitalia) are presented. An identification key to the carabid species encountered on the Quindalup Dune System is also presented. While this study does not provide conclusive evidence of habitat fragmentation directly affecting the carabid assemblage structure it indicates that the species are not uniformly distributed across the Swan Coastal Plain either temporarily or spatially. Also, the presence of both rare and undescribed forms within the remnants underline the desirability of further surveys of these localities. Research into the relationships between the various taxa in these localities is required before fully informed conservation decisions for either the carabid fauna or the remnants themselves can be made.