This paper is intended to outline the pattern of co-operativization practised in the co-operative farming experiments in Bangladesh. Results of the analysis gave rise to some hypotheses demanding micro-research. It has been pointed out that land reforms seem to be necessary for homogenizing the farm size and tenancy relations in order to ensure equity in returns to land and son-land factors in co-operative farming. Examples indicate that present forms of collectivization prior to mechanization can produce some "indirect" effects through better loan use, easy access input to and extension service, while the size effect on productivity may be non-spectacular or negative due to initial dis-economies. The paper concludes that goal conflicts between heterogeneous member groups can be tackled in the short run by limiting the level of collectivization.