CONDITIONS FOR CONJUGATION 361 



Comparing this race, no. 3, with race 1, it will be seen that in 

 both conjugation occurred naturally under the more or less 

 uniform conditions afforded in large stock cultures. In culture 

 3a there is at first apparently a periodicity marked by ' epidemics' 

 in the occurrence of the process at intervals of about four weeks. 

 In culture 3b, started from an exconjugant, the first epidemic 

 appeared after about five weeks, and in 3d, similarly related to 

 3b, after about four weeks. Conjugation in race lb, while not 

 occurring at periodic intervals in the stock culture, was readily 

 evoked in isolation experiments during the entire period when 

 such experiments were regularly conducted (March 18th to 

 April 18th). The same irregularity tends to manifest itself 

 in the culture 3a after the onset of the second period of conjuga- 

 tion (March 12th). 



This irregularity in the occurrence of conjugation during 

 the later stages of growth in a culture in which it has already 

 taken place, in the form of a preceding 'epidemic,' makes it 

 impossible to measure accurately the interval between conjuga- 

 tion periods except with cultures started from exconjugants, 

 and in which conjugation has not previously occurred. Even 

 the first epidemic of conjugation in a newly developed strain 

 may show two separate phases (see observations on races 

 lie and llf, p. 363) suggesting that diverse lines, showing 

 acquired differences in susceptibility as regards conjugation 

 may have arisen among the offspring of one individual. Later 

 irregularities in conjugation — after a previous epidemic in the 

 same culture — may perhaps be accounted for in part, also, as 

 due to hereditary differences established through conjugation, 

 among the exconjugants of the first epidemic. Thus diverse 

 strains, with different degrees of susceptibility, each having 

 perhaps a somew^hat different periodicity, might be perpetuated 

 in the same culture coming originally from the same race. If 

 this be true, a new clon derived from a single exconjugant would 

 presumably give conjugants in a more sharply defined epi- 

 demic than a culture containing the offspring of several exconju- 

 gants, although originally from the same strain. 



