V IRTUALLY every textbook on international trade contains a discussion and/or derivation of the optimum tariff for a country with monopoly power in world trade. Invariably these are treatments of aggregate tariff policy. Little or no work has been done to explain the infra-structure of tariffs within a given country or to explain in what respects tariffs might be expected to be similar across countries. Some papers have explored the impact of tariffs and tariff changes on domestic welfare, labor's share of income and trade prospects for developing countries. Contributions to this literature include Anderson (1972), Balassa (1965, 1967a, 1967b, 1971); Ball (1967); Basevi (1966, 1968); Cohen and Sisler (1971); Kreinen (1961, 1967); Mitchell (1970); Travis (1968) and many others. Yet, none of these studies has offered a well specified analytical framework that would permit a systematic investigation of testable hypotheses concerning the relative protection of industries: within and across countries. Instead, many of these papers have relied upon intuitive but ad hoc explanations of the data they presented. The purpose of this paper is two-fold. First, we will begin a systematic inquiry into the determinants of tariff policy at the industry level by developing a simple model based upon profit maximizing behavior in section II. Secondly, we will use the derivations of section II, along with data from previous studies by Balassa (1965, 1971) to demonstrate that a number of the ad hoc explanations of tariff data by previous authors are consistent with our simple .model. In addition, we will develop and test new hypotheses based on the analytical developments of section II. For example, we will be able to explain the pyramiding of tariffs in developed and less developed countries, the pattern of the rank correlations between effective protective rates for countries at similar and different levels of economic development, the greater variation in tariffs on primary products than on manufactured goods in developing countries, the greater variation and average value of tariffs on manufactured goods in developing countries and, finally, the relationship between the skill intensity of production and tariff protection in manufacturing in both developed and developing countries. Perhaps, most important of all our results is the fact that, by demonstrating that ad hoc explanations of previous studies are consistent with our explicit formulation of tariff determination at the industry level, we facilitate the development of future advances in this area.