34 KANSAS UNIVERSITY SCIENCE BULLETIN. 



be an almost complete series of gradations between those fila- 

 ments which showed no growth, and those which immediately- 

 reverted to the type. There were, first, filaments which ex- 

 hibited no growth at all, though in many cases they showed 

 considerable motility after isolation ; second, those which 

 merely increased somewhat in length ; third, those which 

 produced several filaments or short rods, then ceased to grow 

 further ; fourth, those which grew at once into vigorous cul- 

 tures, but during the first generation showed an abnormally 

 large number of long filaments ; and finally those which 

 grew into normal cultures without any peculiarities. The 

 variations which produced permanent new races seem to lie 

 between the third and fourth gradations. 



Some of the filaments which partially developed showed 

 some interesting phenomena. The motile filament would 

 divide and produce perhaps one or two long motile filaments, 

 one or two non-motile ones, and perhaps two short, very 

 motile rods. I have found these short rods motile twenty- 

 four hours after the original isolation was made. Their fail- 

 ure to grow cannot be attributed to conditions of culture, 

 since, close beside them, in the same medium and under the 

 same cover- glass, single normal cells of apparently no greater 

 motility rapidly developed numerous offspring. 



The new race which I originated during the summer of 

 1905 from B. coli, type I (new race H ) , exhibits characteristics 

 similar to those of race A, but more closely approaches the 

 type. It has been cultivated nineteen months, on various 

 media, but neither reverts to the type nor approaches race A. 

 I have made three tests of its power of fermentation, using 

 fermentation tubes of the type No. 2. In the first test, made 

 soon after its origin, race H produced a lower maximum than 

 the type. It was then passed through four subcultures, and 

 a second test made five days after the first. It then showed 

 a higher maximum than the type. About four months later 

 it was tested again, and this time showed a maximum 13.7 

 per cent, greater than the type. 



The third new race (race Y), which I obtained in the sum- 

 mer of 1905 from B. coli, type II, shows characteristics simi- 

 lar to the other two, but differs less from the type than A and 



