associated with the use of uncooked or partly cooked food. A selective action is often manifested, those persons who have eaten the incriminated food substance raw or imperfectly cooked being most seriously affected, while those who have partaken of the same food after cooking remain exempt. [Illustration: FIG. 6.--_Bacillus enteritidis_, Gärtner; pure culture; Van Ermengem preparation. (Kolle and Wassermann.)] The discovery of the connection of paratyphoid bacilli with meat poisoning dates from the investigation by Gärtner,[67] in 1888, of a meat poisoning outbreak in Frankenhausen, a small town in Germany. This epidemic was traced to the use of meat from a cow that was slaughtered because she was ill with a severe enteritis. Fifty-eight persons were affected in varying grades of severity; the attack resulted fatally in one young workman who ate about eight hundred grams of raw meat. Gärtner isolated from the spleen of the fatal case and also from the flesh and intestines of the cow a bacillus to which he gave the name _B. enteritidis_. Inoculation experiments showed it to be pathogenic for a number of animal species. Bacilli with similar characters have since been isolated in a number of other meat poisoning epidemics in Germany, Belgium, France, and England. One well-studied instance of food poisoning due to the paratyphoid bacillus has been reported in the United States.[68] The bacteria of the paratyphoid group are closely related to the true typhoid bacillus, but differ from the latter organism in being able to ferment glucose with gas production. They are more highly pathogenic for the lower animals than is the typhoid bacillus, but apparently somewhat less pathogenic for man. Most types of paratyphoid bacilli found in food poisoning produce more or less rapidly a considerable amount of alkali, and, if they are inoculated into milk containing a few drops of litmus, the milk after a time becomes a deep blue color. Several distinct varieties of paratyphoid bacilli have been discovered. The main differences shown by these varieties are agglutinative differences. That is, the blood serum of an animal that has been inoculated with a particular culture or strain will agglutinate that strain and also other strains isolated from certain other meat poisoning epidemics, but will not agglutinate certain culturally similar paratyphoid bacteria found in connection with yet other outbreaks. Except in this single matter of agglutination reaction, no constant distinction between these varieties has been demonstrated. The clinical features of the infections produced in man and in the higher animals by the different varieties seem to be very similar if not identical.