(1) Reaction changes and production of gas in organic salt media are not sufficiently constant to form a reliable diagnostic criterion for the differentiation of bacterial types.

(2) The enhancement or inhibition of bacterial growth in such media in the majority of cases bears a direct relationship to the utilisation of the salt by the organism. This furnishes a useful differential test for certain organisms when citrates are used, but cannot be applied in the cases of all salts.

(3) The bacterial decomposition of the salts of those organic acids which form insoluble lead salts can be clearly demonstrated by the addition of suitable quantities of a solution of lead acetate to the culture.

(4) By the use of six organic salts, seven different groupings of the common Salmonella types can be obtained, whereas the sugar reactions have, up to the present, yielded only four different groupings.

(5) Regarding other groups of bacteria, the organic salts form an easy means of distinguishing between pathogenic and certain non-pathogenic vibrios, and between certain of the members of the coli-aërogenes group, and also between B. diphtheriae and Hofmann's bacillus, as well as between B. mallei and B. whitmori.

(6) The six organic salts employed in this test are relatively inexpensive, will stand sterilising by autoclave, and can be obtained with certainty in a state of purity much more readily than the rarer “sugars.”

(7) The nature of the decomposition products of citric acid has been fully examined in the case of Bacillus suipestifer; it has been shown that the products are acetic acid, carbon dioxide and succinic acid, and a simple explanation of the mechanism of this reaction is put forward. In the case of fumaric acid a preliminary examination shows that the acid is converted into succinic acid probably by direct reduction. Maleic acid appears to behave in an analogous manner to fumaric acid. Further work on these acids is in progress.

(8) A large number of organic acid salts have been tried, but only the six suggested have given useful results. It appears that simple aliphatic monobasic and dibasic acids, with the exception of formic acid, are not decomposed readily by the bacteria investigated, and this is also true of monohydroxycarboxylic acids. Readiness of decomposition is first shown by the dihydroxydicarboxylic acids (tartaric acids), and appears to be at its best in the hydroxytricarboxylic acid (citric acid).

(9) While organic salt fermentation tests have been found particularly useful in the cases of the bacterial groups dealt with in this paper, they cannot be substituted for the “sugar reactions” in general use.