HABITAT STRUCTURE AND INVERTEBRATE ASSEMBLAGES ON STREAM STONES - A MULTIVARIATE VIEW FROM THE RIFFLE

Highly structured habitats, those with complex or heterogeneous physic al structure, commonly contain more species than simply structured one s. However, tests of this relationship have been hampered by habitat-s pecific definitions and measurements of habitat structure and, in stre am studies, by insufficient information on spatial variation in faunal abundances. We sampled the fauna of 90 stones collected from an order 4 section of the Taggerty and Steavenson Rivers, southeastern Austral ia. The spatially nested sampling design encompassed three spatial sca les (sites within the same stream order, riffles at the same site and groups of stones within the same riffle) but no temporal replication, as we were only able to sample once. Habitat structure of stones was q uantified by measuring the amount and type of epilithon on stones, tog ether with stone shapes, sizes, textures and surface complexities. For the latter two attributes, we used a new method, involving image anal ysis, which can be used to quantify habitat structure in comparable wa ys in other systems. Species richness (S) was related to some measures of habitat structure, being higher on stones with a rough texture and weakly related to the amount of epilithon present. Total numbers of i ndividuals (N) were also higher on rough stones, and faunal compositio n (as quantified by hybrid multidimensional scaling) was related to ab undance of epilithon. However, flow environments, as quantified by wat er velocities and depths, were also related to faunal composition. Rif fles varied in both faunal densities and near-bed flow environment, bu t not in any consistent way. Such variability means that individual ri ffles cannot be reasonably used to 'represent' stream sections or orde rs, an assumption commonly made in stream studies. Neglect of such iss ues has also resulted in poor conceptual integration between large- an d small-scale studies in stream ecology.