That, however, some risk attends the placing of too implicit a reliance on this method of diagnosis alone is evident from the fact that a _negative_ reaction, or in other words, absence of all agglutinising phenomena, is sometimes associated with blood throughout what is beyond all question a well-defined case of typhoid fever, whilst in the first week of this disease the test is frequently negative in character. Rouget, who has made a very careful inquiry into the value to be attached to the sero-diagnosis of typhoid fever, states that he has found in a large number of examinations of blood derived from undoubted typhoid patients the agglutination phenomena fail altogether; it is, therefore, not surprising that the sero-diagnosis of this disease is still the subject of much discussion and investigation. An interesting example of how particular serums may be employed for the detection of particular poisons has been furnished by Dr. Calmette. In some districts of India the natives have an ugly custom of wreaking their vengeance on their enemies by poisoning their cattle, and to effect this both expeditiously and secretly they employ subtle poisons which they know can only be detected with great difficulty. Serpent venom is a favourite substance, whilst abrine, a highly toxic vegetable poison, is another. The method adopted for the application of this abrine is highly original, and consists in taking small bits of wood shaped like miniature clubs, so diminutive in size that they can be concealed in the hand. In the head of the club small holes are bored, and tiny pointed rodlets of a hard greyish substance are fitted into them. Armed with these crude instruments, the natives scratch the cattle in several places, and, although but little external sign of injury is to be seen, the rod-points penetrate the skin and are broken off, and the poison is left to work its lethal way through the animals' system. Mr. Hankin forwarded some of these broken-off rod-points to Dr. Calmette for the identification of their composition, and he diagnosed the material employed as abrine in the following original manner. He first introduced some of this rod material into animals, and found that their symptoms were suggestive of abrine poisoning. To confirm his suspicions, however, he took some more of this rod material, and, before inoculating it into animals, he mixed it with serum derived from animals which had been artificially rendered immune to abrine poison. Instead of the animals into which this mixture of serum and "rod material" had been introduced dying like the previous ones, they remained alive. Had the "rod material" consisted of some poison other than abrine, the abrine serum would not, according to Dr. Calmette, have negatived its action, and it has thus been indicated how protective serums may be successfully employed for the detection of poisons.