An analysis of the allocation of new scientists between teaching and research. The study uses an analytic framework built on specified policy objectives and a model for the allocation of new Ph.D.s in science between the two areas. Some simplifying assumptions lead to a general welfare function for this aspect of a national science policy. Special cases of the general welfare function, including minimization of the time required to attain a given scientific capability, are treated explicitly. The production of new scientists is determined in the model by a production function, dependent on teaching scientists and on time.