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

Testing different genetic groups of common carp by stocking them together into 'communal ponds', requires correcting observed weight gains for differences in initial weight among the tested groups. Reliable estimation of correction factors, by the multiple nursing technique, requires a minimum difference of 15 g in initial weight between multiply nursed samples. Empirical estimates of correction factors are biassed by errors in estimating initial weight differences. No consistent differences were isolated in the correction factors generated from different genetic groups. A multiple regression analysis based on data from ten tests generated the following prediction equation: b=4.543-0.113 X.+0.00671 Y. (R=0.716) where b=predicted correction factor; X.=mean initial weight; Y.=mean weight gain. Predicted correction factors, generated from the analysis of a large set of data, are at least as reliable as empirical estimates from a single experiment. Their use enables an economy in facilities, which would otherwise be required for producing and testing multiply nursed samples. The application of this technique to other fish species is discussed. © 1987.