Desiccation, cone-and-dish, carbon tetrachloride distillation, and Karl Fischer procedures for the estimation of moisture in ballistitetype powders have been compared. Desiccation over concentrated sulfuric acid has been found satisfactory with respect to accuracy and also the simplest and most reproducible of these procedures, but under certain conditions the other methods are feasible for more rapid determinations.MOST solventless double-base powder manufactured at the present time is a mixture of nitrocellulose and nitroglycerin to which a few per cent of stabilizers, plasticizers, and inorganic salts have been added. Powders of this type are not notably hygroscopic, but finished samples in equilibrium with t heir surroundings contain a few tenths of a per cent of moisture. Precise control of the concentration of this constituent is essential to ensure satisfactory behavior of the finished powder. Several procedures for the estimation of moisture in smokeless powders have been reported (2, 3, 4), many of which are tedious and involved or are not capable of the necessary accuracy. The authors have selected the four procedures that appear most promising and have compared them carefully with regard to accuracy, reproducibility, and ease of operation, using several samples of ballistite-type powder containing approximately 0.2% moisture. The results of these comparisons are presented below, together with specific recommendations based upon them.EXPERIMENTAL PROCEDURES Desiccation Procedure. The desiccation procedure for the estimation of moisture involves storing the powder sample in a desiccator until the W'eight is constant; the moisture content is calculated from the total loss in weight of the sample.