Macroinvertebratecommunity structure was compared in benthic samples taken by Surber and kick methods from a lotic system in south-western Australia. Eleven sites were sampled concurrently in winter, spring and summer 1987.Surber samples contained fewer individuals and more taxa, particularly those with a low frequency of occurrence. This was attributed to the lower surface area, but greater intensity of Surber sampling. It is proposed that the Surber method is more suited to taking cryptic and closely adherent taxa in sites with a highly heterogeneous substratum.Percentage similarity between paired Surber and kick samples was determined by Sorensen's and Czekanowski's coefficients, with mean values of 66% and 60% for June, 61% and 49% for September and 66% and 49% for December respectively. Ordination demonstrated a division of upland from lowland sites on axis 1, with a separation of paired-samples on axis 2. This pattern held across qualitative and quantitative datasets, with and without a downweighting on rare taxa. At each level of classification fewer paired-samples separated in qualitative than quantitative datasets.Kick sampling provided a substantial saving in costs over Surber sampling, particularly when qualitative data were utilised, making the method suitable for routine, biological monitoring. However, the initial use of replicated Surber sampling, particularly in areas that have not been previously sampled is recommended for environmental impact studies to detect rare taxa, that may be endangered.

Macroinvertebrate assemblages downstream from the Canning Dam, on the Canning River, Western Australia, were sampled to assess the impact of long-term impoundment and the role of a major tributary in community recovery.Kick samples and associated physical measurements were taken from riffle zones in three reaches in March and July 1989. The lower reach was located immediately downstream of the confluence with Stinton Creek, the first major tributary below the dam, with the middle and upper reaches positioned between the tributary and the dam.Reduced concentrations of dissolved oxygen in the upper and middle reaches in March, compared to the lower reach were attributed to significantly higher levels of organics, high summer water temperatures, and a proposed increase in microbial activity. The build-up of organics was probably related to reduced flushing, as a result of impoundment. Stinton Creek increased the discharge of the Canning River below the confluence, particularly in winter, which presumably prevented the build-up of organics in the lower reach.Significant differences in the aquatic macroinvertebrate fauna between reaches were detected. A total of 68 taxa was recorded from the lower reach, 88 per cent of which were also present in the middle and upper reaches. However, the middle and upper reaches contained a greater number of taxa (1 12 and 90 respectively), approximately 50 per cent of which were not recorded from the lower reach. The additional taxa were more typical of lowland rivers or lentic (standing water) systems, suggesting that physical conditions in the middle and upper reaches were more like a lowland river than an upland stream.More collectors and shredders occurred in the upper and middle reaches, associated with the accumulation of particulate organic matter. Ordination and classification procedures based on macroinvertebrate assemblages clearly separated samples from the upper and middle reaches from the lower reach. There was also a distinct seasonal separation.These observations support a hypothesis that while the reduced flow below Canning Dam had an impact on the macroinvertebrate fauna, confluence with a major tributary (Stinton Creek) allowed recovery of the macroinvertebrate community, through the tributary acting as a source of increased discharge. The implications for the management of impounded rivers in southwestern Australia are discussed.