24 MASS. EXPERIMENT STATION BULLETIN 319 



well-defined types. Also by frequent transfer of such variants and by selection 

 of colonies which reveal the greatest degree of variation, certain definite forms 

 can be established. 



Experiment V. Two old strains which revealed some degree of variation in 

 colony type were selected for this experiment. These strains were the same as 

 Strains I and II employed in Experiment I. Both strains produced smooth 

 and rough colonies as well as intermediate colonies. An attempt was made to 

 select and isolate both the roughest and the smoothest types of colonies. At each 

 transfer each type was selected, and by this procedure it was possible to isolate 

 a rough type which appeared stable and a smooth type which at times would 

 exhibit a certain degree of roughness. At first transfers were made every 48 

 hours, but later the interval was extended to from four to five days. 



The rough and smooth colony types isolated from Strain I were designated 

 A and A-l, respectively. The two types, rough and smooth, isolated from Strain 

 II were designated B and B-l, respectively. These different colony types will be 

 described in another part of this section. The rough types resembled each other 

 very closely on plain meat extract agar, although slight differences could be 

 detected at times. The rough forms from both Strains I and II appeared to be 

 stable, since little variation was observed in the numerous transfers of these types. 

 The smooth types did not resemble a typical smooth 5. pullorum colony in every 

 respect, as will be brought out later. Also type B-l very infrequently produced 

 a markedly rough type of colony. On one occasion during the period of 14 months 

 of frequent transfer a papillated colony was detected in this culture. The appear- 

 ance of this type was rather sudden and it disappeared in the same manner. An 

 attempt to propagate and stabilize the papillated variant met with failure. The 

 cellular morphology was not as striking as the colony forms, and little variation 

 could be detected in the cell types during the period under observation. (See 

 Table 4.) 



Experiment VI. The purpose of this experiment was to determine the influence 

 on variants and on normal 5. pullorum strains of frequent transfer on liver infusion 

 agar. Since it is well recognized that liver infusion agar yields a heavier bac- 

 terial growth than plain meat extract medium, the possibility suggested itself 

 that the former might be more suitable for the isolation and stabilization of 

 variants. Liver infusion agar is employed in some laboratories for the cultivation 

 of strains for antigen production, and by some even for the maintaining of stock 

 cultures. Hence, it seemed all the more important to employ this medium to 

 determine its influence on variants and normal strains. The strains used in this 

 experiment were: 



Strain I — Same as Strain I in Experiments I and V 

 Strain IV-a — Variant isolated from Strain IV in Experiment I 

 Strain XII — Same as Strain XII in Experiments II, III, and IV 

 Strain XIV — Same as Strain XIV in Experiments II, III, and IV 

 The liver infusion agar was prepared in the same manner as in previous experi- 

 ments. The strains were transferred on plates at intervals of 48 hours. Later 

 in the experiment the interval between transfers was increased to four or five 

 days. The cultures were incubated at 37° C. When more than one type of colony 

 appeared on the plate, only the roughest were selected for transfer. The colonial 

 response of Strains I, IV-a, and XIV was impressive. A striking difference was 

 observed between these three strains and the so-called normal Strain XII. The 

 last appeared as typical S. pullorum on liver infusion agar. These strains were 

 cultivated on this medium and transferred at frequent intervals for a period of a 



