Wohlfarth, G.W., Agricultural Research Organization, Fish and Aquaculture Research Station, Dor, D.N. Hof Hacarmel 30820, Israel
Moav, R., Agricultural Research Organization, Fish and Aquaculture Research Station, Dor, D.N. Hof Hacarmel 30820, Israel

Communal testing consists of stocking the different genetic groups to be tested into the same (communal) pond. In this system, genetic differences between groups of fish and differences in pond productivity are not confounded, as they are in separate testing. Communal testing reduces the number of replications required and enables larger numbers of groups to be tested in a smaller number of ponds. Growth testing in communal ponds means that the testing environment and the commercial production environment in which the results of genetic investigations are to be applied are not identical. This difference is due to the mutual competitive interactions between test groups in communal ponds, which cannot occur when they are stocked separately. Interactions may be due to differences in initial weight between the groups, or to differences in competitive ability not associated with differences in initial weight. The methodological investigations which allow these sources of bias to be removed are described. The effect of differences in initial weight is corrected with the aid of the environmentally generated coefficient of regression of weight gain on initial weight in a communal pond. This regression requires empirical estimation in each test, by the multiple nursing technique, specifically constructed for this purpose. Relative growth rankings of paris of test progenies were identical when each group was stocked separately into a series of replicated ponds and when both were stocked together into communal ponds. Thus growth estimates from communal testing are reliable predictors of expected genetic differences in separate ponds. Communal testing allows any pond or other growth vessel to be used as a testing facility, irrespective of its area or aquacultural management. Alternatively, it enables breeding studies to be integrated with other unrelated investigations. This results in a more rational utilization of facilities and the possibility of isolating and estimating interactions between experimental variables. © 1985.