THE PRODUCTION AND TROPHIC ECOLOGY OF SHALLOW-WATER FISH ASSEMBLAGES IN SOUTHERN AUSTRALIA .3. GENERAL RELATIONSHIPS BETWEEN SEDIMENTS, SEAGRASSES, INVERTEBRATES AND FISHES

Fishes and benthic invertebrates were sampled in seagrass and unvegeta ted habitats at 14 localities across southern Australia in order to de termine any consistent relationships between animal production, fish c onsumption and environmental parameters over a large spatial scale. Mo st of the features identified in a related study at Western Port were confirmed over the extended geographic range; however, an exception wa s that the production of fishes in the smallest size-classes (i.e. <1 g wet weight) was not consistently greater in seagrass habitat than un vegetated habitat. The more important characteristics of seagrass and unvegetated habitat types identified in the extended study were (1) mo re fish species were associated with seagrass habitat than unvegetated habitat, (2) the majority of small-fish species in both habitat types fed on crustaceans, with relatively few species capable of utilising algae and virtually no species utilising seagrass, (3) the few species that did ingest algae often occurred in high abundance, (4) the size of prey eaten by fishes increased consistently with fish size, with pr ey length averaging approximate to 6% of fish length, (5) abundant lar ge fish species generally consumed smaller prey than rarer large fish species at the same body size, (6) greater benthic invertebrate and de mersal fish production occurred in seagrass habitat than in unvegetate d habitat, (7) most of the production of crustaceans >1-mm sieve size was ingested by fish predators while only a small proportion of non-cr ustacean benthic production was consumed. Fish production was highly c orrelated with crustacean production and seagrass biomass, and was neg atively correlated with wave exposure (measured as fetch) across the r ange of sites. The estimated production of crustaceans was highly corr elated with the biomass of seagrass material and also with the proport ion of particles <63 mu m in the sediments. The overall relationships between macrofaunal production (M: mg . m(-2). day(-1)), macrocrustace an production (C; mg . m(-2). day(-1)). demersal fish production (D; m g . m(-2). day(-1)), fetch (F; km), seagrass biomass (L; g . m(-2)), p roportion of particles <63 mu m (S;%) and temperature (T;degrees C) we re: In M = 1.41 + 0.088 In (L + 1) + 0.38 In S + 0.89 In T. In C = -0. 93 + 0.27 in (L + 1) + 0.22 In S + 0.89 In T. In D = -1.23 - 0.62 In F + 0.36 In (L + 1) + 1.04 In T. These regression equations can be used as models to predict the production of macrofauna, crustaceans and sm all fishes at unexamined sites. When predictions were compared with es timates of annual production at the eight sites previously examined in Western Port, most predictions lay between 50 and 200% of measured va lues. Additional work in Australia and overseas should allow these mod els to be refined.