THE DYSENTERY BACILLUS. 217 



Three of these types (Shiga, Plexner, and Stroag) correspond exactly 

 to the agghitination reactions, while the other one (Bacillus Y) is only 

 iiTegrilarly differentiated from that of Flexner by that reaction. Fur- 

 thermore, the action on mannite corresponds to a difference in toxic- 

 ity; the Shiga-Kruse t3'pe which does not -produce acid in mannite, is 

 very toxic for animals, while the other types which produce acid in 

 mannite are not nearly as toxic. This difference in toxicity was clearly 

 shown by Martini and Lentz at the time they differentiated their two 

 groups on the basis of the agglutination reactions. 



A number of other types have been added at times, and the above have been 

 separated into a number of others by various worliers. Thus Shiga (8) made 

 a fifth type which differed from the "Flexner" in that it produced acid in 

 mannite in the first twenty-four hours, and then alkali. It was very close or 

 similar to the Flexner type in all other properties, including agglutinability. 

 Ohno(9) grew a large number of strains of the dysentery bacillus in the sugars 

 for fourteen days and in this way divided the dysentery bacillus into 15 varieties 

 because of their action on the sugars; but when he tested the agglutinability of 

 his various strains he did not observe these varieties. 



Hetseli(lO) already had pointed out that the reaction differed somewhat ac- 

 cording to the percentage of sugar in the medium. In addition to sugar, the 

 media contain peptone and albumoses, and while the decomposition products of 

 the sugars are mainly acid, those of the peptone and albumoses are mainly 

 alkaline. The bacteria attack both classes of substances and upon the amount 

 of one or the other present, as well as the avidity with which one or the other 

 is attacked, depends the acidity or alkalinity of the sum total of the products 

 of decomposition. Again, an organism may change in its relative avidity for 

 the carbohydrate and the proteid content of the medium. Hiss and Russell (6) 

 and a number of others have shown that an organism may change in its action 

 on sugars after it has been on artificial media for some time. 



Finally, the absorption experiment of Castellani has been used for the separa- 

 tion of the dysentery bacillus into types, but it gives so many that it almost 

 hopelessly complicates the placing of any particular strain. 



During an epidemic of bacillary dysentery in the Philippines occur- 

 ring in the summer of 1909, I was able to isolate dysenteiy and dysen- 

 tery-like bacilli from the stools of a number of cases, and I proceeded to 

 study the organisms along the lines indicated in the foregoing review of 

 the literature. In isolating the organisms, I tried the various lactose- 

 agar media, but did not find any of them to be superior to the litmus- 

 lactose-agar for practical work on the dysentery bacillus. 



I prepared my plates in the laboratory in Manila, took them with me 

 into the provinces where the dysentery was most active, streaked the 

 plates there and then brought them back to the laboratory with me, or 

 else, when I remained in the provinces for a few days, I picked colonies 

 and transplanted them into tubes before coming back. In this way I 

 was able to isolate a bacilkis of the Shiga-Kruse type from 12 out of 40 

 cases of severe, acute dysentery in natives of two tovms of Batangas 

 Province where there was an epidemic of acute dysentery with a high 



