TYPHOID FEVER. 383 The thorough investigations of Wyatt Johnston and his associates in Montreal have shown that reliable reac- tions can only be secured when the cultures employed are of an ordinarily virulent typhoid bacillus, and are grown in an alkaline medium for about twenty-four hours. I prefer fresh agar-agar cultures, distributed throughout sterile clean water, rather than bouillon cultures, because of the larger number of bacteria in the former, the con- sequently greater number of agglutinations formed, and the readiness with which they are found upon micro- scopic examination. It is necessary, however, to make a microscopic examination of the diluted culture before adding the serum or blood, in order to be sure that there are no natural clumps of bacteria present to simulate the specific agglutinations. This is of great importance. The natural clumps of bacilli are more apt to occur in cultures grown upon fresh, moist agar-agar than upon that kept for a short time until the surface has become partially dried. The chief difficulty experienced in making the test seems, at present, to reside in the prepa- ration of the blood in accurate dilution—i. e. securing it in measured amounts. The original method of Widal, to collect about 5 c.cm. of blood in a test-tube by the introduction of a hypo- dermic needle into a vein, is a rather more serious and disturbing operation than most patients care to undergo for purposes of diagnosis. Blood dried upon paper, as suggested by Johnston, or upon glass, while extremely convenient for transporta- tion, is not susceptible of accurate dilution for quantita- tive estimation. Cabot has successfully made dilutions with a medicine- dropper, by using one drop of blood and as many drops of culture, dropped from the same instrument, as were necessary for the desired dilution. This method seems to be very practical, but can only be employed at the bedside, or where it is not necessary to keep or trans- port the blood.