Abstract
Species assemblages in the Hauraki Gulf, New Zealand, and the factors affecting them, were determined from abundance data collected from 1381 trawl tows by two research vessels during 1964–97. Correspondence Analysis and Ward's Cluster Analysis were used to identify assemblages, and Canonical Correspondence Analysis (CCA) to explore the relationships between the assemblages and depth, sediment type, latitude, longitude, and year. Twenty‐six species that occurred in more than 5% of the stations were assigned to four assemblages, and a further four species showed no affinities with other species. For four commercial species, there was no evidence of ontogenetic variation in assemblage membership. The CCA models explained only 11–12% of the variance in the data. Year was the most important explanatory variable, suggesting that between‐year variation in species abundance affected assemblage composition. Sediment type was also important for many species. Assemblage A species were most abundant at shallow ( 50 m) mud sites, and Assemblage D species at deep mud sites. Assemblage C species had varied habitat preferences. The commercially important Pagrus auratus and Nemadactylus macropterus preferred mud, and Chelidonichthys kumu preferred sand. Zeus faber was most abundant at muddy sites, but also occurred (sometimes abundantly) at sandy sites.