Findings presented in this paper answer, inconsiderable part such questions as: How much of the yarn strength variance is explainable by 8 separate and variously combined fiber properties? How closely can yarn strength be estimated or predicted from a knowledge of such fiber properties only? Which combination of fiber properties gives the best results? And which and how many fiber tests may be omitted from a series and still give sufficient data for satisfactorily predicting yarn strength? The data used in these analyses represent 766 individually tested cottons covering a wide range of American Upland types as well as a wide range of growth conditions. Samples representing 16 commercial varieties, grown in duplicate at each of 8 locations across the rainfall part of the American Cotton Belt, during crop years 1935, 1936, and 1937, have served as a basis of this study. All fiber, spinning, and yarn tests were preformed by standardized methods under controlled atmospheric conditions.