The March 1975 issue of this Review contained a controversy over the welfare implications of tariff imposition, resulting from the analysis in Section III of a paper by Raveendra Batra (1973a). The issue was raised by Murray Kemp and Edward Tower. Although this controversy is very interesting, I do not propose to join it, but rather point out that the analysis used in Section II of Batra's paper is misleading. Under the assumptions of gross substitutes, dD3/dP2 > 0, and dX2/dP2 > 0 in his notation, he derives the result that an effect of a change in the terms of trade on social welfare mainly depends on factor-market distortions. My basic objection to his analysis is the treatment of d and X in his paper. He claims that the size of both d and X depend solely on factor market distortions, independent of the factor market orderings. This argument crucially relies on the hypothesis that changes in capital, dKi, and labor, dLi, due to a change in the terms of trade, have the same signs. For the case of Section I, the terms of trade are fixed, and theretore the capital-labor ratio does not change at all. However, as soon as the terms of trade are allowed to vary in Section II, it is no longer necessarily true that capital and labor will change in the same direction in any industry. I construct a simple example below for the home good industry to illustrate this point. Then using Figure 1, I show how a change in the terms of trade affects the capital-labor ratio in each industry as well as allocations of labor and capital to each industry. Suppose that the social welfare function is given by the following Bergson family type, 3