all-malt products of the first two breweries under consideration. This, however, can be readily explained when it is considered that in the case of brewery No. 2 in the all-malt beers (sample No. 22017-D) only 58 pounds of malt were used in the preparation of a barrel of beer containing 31 gallons; while in the case of sample No. 29512-B there were used, in preparing a barrel of similar capacity, 68 pounds of malt and 23 pounds of cerealin. That is, in the second product there is, in the same volume of liquid, the extractive material from 68 pounds of malt and 23 pounds of cerealin, while in the first product there is present the extractive material from only 58 pounds of malt. Since the analysis is made upon the finished liquid it is evident that the percentage composition of any particular ingredient should be very much larger in the second product because of the very much larger amount of material used in its preparation. It is apparent, therefore, that no direct comparison can be made between the percentage composition of these different brews in order to determine the effects of the raw materials upon their composition. The most satisfactory way to have tested this question of the effect of raw materials on the finished product would have been to make a series of worts with exactly the same percentage of solids, some of pure malt and others of mixtures of pure malt and corn, rice, and cerealin; then a direct comparison between the results would have shown the effects of these various materials. This method was impracticable because it was necessary to take the brews as actually made under varying commercial conditions. The object sought can be accomplished, however, by calculating the results of these analyses either to the basis of dry material in the original wort or by calculating them to the basis of a wort with constant water content. It was decided to calculate all of the results to the basis of a wort containing 15 per cent of solids, as this would give a uniform basis for comparison and would be approximately an average wort. The method employed in calculating the various beers and ales to this uniform basis was as follows: