Empirical evidence is presented consistent with the hypothesis that instructional expenditures are distributed unequally, and that less is spent on non-white and poor students than on others in large American cities. The most experienced teachers are generally to be found in schools attended by the less poor white children. More important, the verbal ability of teachers, an important predictor of teacher effectiveness (although not highly correlated with salary or experience) is also higher in these schools. An analysis of the physical plant--as measured by age of building, size of grounds, and presence of special facilities--further suggests that the allocation of physical resources may also be influenced by the economic characteristics of the neighborhood. The results of such discrimination are, of course, that less-favored groups receive a lower quality education, on the average, than the average middle-class or white child. This unequal treatment might have been predicted on the basis that poor and non-white groups have least political power and social status; an analysis of the tensions that would be created by a system of equal expenditures in cities heterogeneous in social and economic makeup would further suggest the existence of such discriminatory education. (Author/JM)