When the sample was removed for standardization it contained not only a thick suspension of Pfeiffer bacilli, but also bits of agar and blood-stained debris. It was necessary to rid the suspension of the gross contamination, and this was done at first by filtering it through sterile glass wool filters, and later by centrifuging it at slow speed for about 10 minutes. The suspension then contained little but the Pfeiffer bacilli, and was placed in the Hopkins tube and centrifuged for ½ hour on the sixth contact of the rheostat. This gave the per cent. of Pfeiffer bacilli in the suspension, and the necessary dilutions to make 1,000,000,000 per cubic centimeter were readily determined. The Hopkins tube consists of a centrifuge tube, with a capillary tube sealed on at the smaller end. The centrifuge tube is graduated in 10 c.c., 5 c.c. and 1 c.c. amounts, and the capillary portion is graduated in 0.01, 0.02, 0.03, 0.04 and 0.05 c.c. amounts. To standardize the vaccine, 10 c.c. of the sample was centrifuged in the tube and the amount of sediment read on the capillary scale. If the amount of bacilli fell between the graduations, an additional amount of sample was added, so that the sediment reached one of the graduated lines, the exact amount of sample added being noted. The percentage of the suspension could thus be determined by dividing the number of c.c. of sample used into the amount of the sediment obtained, and the number of bacteria calculated according to Hopkins table. The table available to us did not list the Pfeiffer bacillus, but according to it a 1 per cent. suspension of staphylococcus contains 10 billion organisms to the cubic centimeter, and we estimated that Pfeiffer bacilli were about half the size of staphylococci. This assumption was borne out by a number of Wright’s method counts on standardized suspension of bacilli. We, therefore, calculated that a 1 per cent. suspension of Pfeiffer bacilli should contain about 20 million organisms. Then, if 10 c.c. contain 0.02 c.c. of bacterial sediment, the per cent. was calculated by taking 0.02/10 = 0.2 per cent., the strength of the suspension. If 1 per cent. contains 20 billion, then 0.2 per cent. contains 4 billion per c.c. In order to get a 100 million per c.c. suspension, it would be necessary to dilute the original suspension 40 times. Every method of standardization is more or less inaccurate, but the above described method gave a fairly uniform product. Drying and weighing is claimed by many to be more accurate, but even with this procedure a fair amount of non-bacterial sediment is present in the material to be weighed.