The flat taped bundle test as applied to the measurement of the strength of cotton fibers has been modified during the course of several years. The method has proved satisfactory for both routine and research purposes. The techniques and apparatus are described in detail. Some causes of error and factors other than the fiber strength which introduce variations in test results have been investigated and are discussed. The effects of varying bundle width and weight) direction of fibers, jaw pressure and separation, humidity, and type of tape used have been studied. The observed precision of test results and their statistical reliability are presented. A calibration method is described which greatly reduces the effects of several operators. THE increasingly widespread application of fiber tests by breeders and users of cotton endows the subject of fiber strength measurement with considerable practical significance.Fibers may be measured singly or in bundles. The techniques for measuring the strength of single fibers have been thoroughly covered by several writers [1, 2, 3] . Three techniques have been developed. using bundles of fibers. These are the round wrapped, or modified Chandler bundle [4,5 ] , the flat untaped bundle developed by Pressley [6], and the flat taped bundle [7,8,9,10]. This paper deals with the flat taped bundle method of measuring fiber strength as applied to cotton.The method is not widely used in this application although it is specified as a standard strength test for staple rayon fibers by the A.S.T.M. Committee D-13 [10]. However, the method has proved satisfactory in this laboratory for both routine and research purposes. During the time that it has been used, many experiments have been made to appraise it.The work reported here does not represent an exhaustive exploration of the method, but it has been sufficient to indicate the apparent relative influence of some of the variables on the strength of the fibers.The indicated trends become of increased interest when compared with corresponding phases of the Pressley test as described recently by Williams and Painter in this JOURNAL [11].Although the method used at the Kendall Mills Laboratories conforms in general with the methods reported by others [7, 8, 9, 10~ , there have developed enough differences in technique to make a description of the test desirable. A feature of the method is the use of a reference cotton to which all test results are related. By this means results obtained by different operators are in effect calibrated, and diff erences between them are eliminated.