Fish communities from four catchments in the Murray-Darling river system were analysed in relation to climate, hydrology and river regulation. Using the annual proportional flow deviation as a measure of river regulation, the Paroo River catchment was assessed as unregulated, the Darling River catchment as mildly regulated and the Murrumbidgee River and River Murray catchments as highly regulated. A total of 11 010 fish, representing nine native and three alien species, was caught during high and low flow seasons in the four catchments. Native species, such as golden perch Macquuriu ambigua (Percichthyidae), bony herring Nematalosa erebi (Clupeidae) and spangled perch Leiopotherapon unicolor (Teraponidae), dominated fish communities in the Paroo and Darling catchments, but alien species, mostly carp, Cyprinus carpio (Cyprinidae), were also abundant. Both native and alien species were more abundant in these catchments after flooding, but there was little change in species composition between high and low flow seasons at the catchment level. Carp dominated communities in the Murray and Murrumbidgee catchments. There was a significant trend for reduced species diversity in increasingly regulated catchments. River regulation may alter the relative abundance of native and alien fish by desynchronizing environmental cycles and the reproductive cycles of native species. Ordination of species abundances showed discrete fish communities that reflect the geographical separation between catchments. Differences between communities are related to opportunities for dispersal, the environmental tolerances of dominant species and the modifying effects of river regulation. Fish communities in lakes exhibited less seasonal variation than riverine communities within the same catchment, indicating the greater seasonal stability of lakes compared with regulated and unregulated river reaches. Management of fish resources needs to include catchment-specific strategies within current State and basin-wide management programmes.

Microhabitat use of instream wood habitat by Trout Cod Maccullochella macquariensis (Cuvier), a critically endangered species with a restricted distribution, was examined in the Murrumbidgee River in New South Wales, Australia. Habitat variables were scored or measured at 100 m intervals along the river or wherever Trout Cod were captured using electrofishing. The occurrence of Trout Cod was significantly dependent on the presence of instream woody habitat and 95% of samples where trout cod were caught were associated with the presence of woody habitat. Trout Cod were more likely to be found on simply-structured woody habitats, away from the river-bank and their abundance showed no relationship with water velocity. The low abundance of other fish species at the sampling sites suggests that the relationships demonstrated are not due to inter-species interactions. The results will assist with specific management actions to restore and protect populations of this endangered species.

Rapid bioassessment (RBA) techniques for evaluating river health are now commonplace and there is much debate on the best methods that should be used. One of the important features of RBA is subsampling of large qualitative or semi-quantitative samples to reduce the costs associated with handling and identifying animals. In Australia, the Australian River Assessment System (known as "AusRivAS") has been widely used since 1994 to monitor and assess river health. To test the efficacy of AusRivAS protocols, four live-sorting protocols, the standard Australian River Assessment Scheme (AusRivAS) and three suggested improvements, were evaluated in three habitat types and in clear and turbid rivers. The suggested improvements included using magnification during the live-sort process, separate sorting of coarse and fine fractions and increasing the amount of time or animals collected. There was no statistically significant difference between any of the trialed live-sort protocols in terms of the number of taxa collected compared to the number remaining, the community composition, the abundances of individual families collected, or the AusRivAS Observed/Expected taxa ratios. The lack of differences between the live-sort protocols suggests that technicians using the current standard AusRivAS protocols are able to effectively obtain a representative subsample of animals from the whole kick or sweep net qualitative sample. This has the advantage of cost savings because no retraining will be required, field procedures will remain uncomplicated and previous river health assessments will remain valid.