32 KANSAS UNIVERSITY SCIENCE BULLETIN. 



Iq order to confirm previous results and to ascertain more 

 exactly the sort of variation leading to these new races and 

 the conditions under which they arise, a new series of experi- 

 ments was undertaken during the summer of 1905. It soon 

 became apparent that variations leading to permanent races 

 are not so common as the early success of the experiments of 

 1904 had led me to suppose. For nearly six weeks experi- 

 ments were conducted before a second permanent race was 

 obtained. Under a great variety of conditions, long filaments 

 such as had formed the starting-point of race A were selected, 

 but, on being isolated, they either failed to grow or reverted 

 to the type. Various media, principally plain and glucose 

 broth, were used, in all of which the long filaments were 

 found, and isolations were generally made from young test- 

 tube or hanging drop cultures. 



During the period mentioned about 140 filaments of various 

 character were isolated from the original type of B. coli, which 

 was the parent of race A (B. coli, type I ) , all except one with- 

 out success. From another stock of B. coli which I isolated 

 from feces [B. coli, type II) about fifty more such isolations 

 were made, principally under conditions which had proved 

 successful in obtaining race A, and of these only one new 

 race was obtained. In all, over 190 isolations in long fila- 

 ments were made, from which only two new races were ob- 

 tained. 



In a number of the above isolations I recorded only failure 

 or success in obtaining a new race, but in 95 I kept a com- 

 plete record, and the results were as follows : Out of 52 isola- 

 tions from the original type, B. coli I, 36 failed to grow at all, 

 6 partially developed and failed to grow further, and 9 grew 

 well, but reverted to the parent type, and 1 formed a per- 

 manent new race. Of 43 isolations from B. coli, type II, 29 

 failed to grow, 5 showed limited growth, and 8 reverted to 

 the type, and 1 formed a permanent new race. The 95 iso- 

 lations fairly represent the whole number of attempts made 

 under the best conditions of temperature and medium. The 

 two filaments which grew into permanent new races showed 

 no characteristics differing from the majority found in un- 

 successful attempts. 



