grams--is now weighed into as many bottles of about 300 c.c. capacity as determinations are to be made, 100 c.c. of the filtered liquors, prepared as before described, are introduced into each bottle, and the bottles are then each shaken for 10 minutes (Mr. Alsop states that in his experience 5 minutes is sufficient). The contents of the bottles are now filtered through funnels, the stems of which are plugged with pure cotton-wool, and the liquor is returned till a clear filtrate is obtained, of which 50 c.c. is evaporated as in the International method. It is now necessary to accurately correct the residue obtained, for the amount of water carried in by the wet powder. The loss of weight of the powder which has been dried, divided by its wet weight, gives the water contained in each gram of wet powder, and this multiplied by the weight of wet powder added to the liquor, gives the weight in grams (or volume in c.c.) of water which has been added to each 100 c.c. of liquor. Consequently, if the residues found be multiplied by this weight plus 100, and the product divided by 100, the weight will be obtained which should have been given by 50 c.c. of undiluted but detannised liquor; and from this the non-tannins are calculated exactly as in the case of the residues from the filter process. Of course, in practice, a factor is found, by which it is simply necessary to multiply all the residues, to correct them to undiluted weight. The process sounds somewhat complicated, but in reality, where a large number of determinations have to be made, is quite as quick, if not quicker than the filter method; which it is quite possible it may ultimately supersede, as much attention is being devoted to its improvement.