522 G. N. CALKINS AND L. H. GREGORY 



5. Conjugation tests are made once a month. Variations in 

 conjugating power are even more marked than are size varia- 

 tions. There is also a difference in conjugating power in prog- 

 enies from different ex-conjugants, series H' and L for example, 

 are weak series, G. M. Q and T' are strong. In these series no 

 attempt was made to compare more than four pure lines from 

 each ex-conjugant, but of these four there are clear differences in 

 the conjugating power. In series H' only two of the four quad- 

 rants comprise conjugating forms. In the T' series one quadrant 

 comprises no conjugating forms, the other three give epidemics 

 at each test; in the M series two quadrants give few pairs, the 

 other two quadrants give epidemics of conjugation at each test. 

 Series E' is also a weak conjugating series. 



The most marked differences in conjugating power are shown 

 by the J series. Beginning with the November test and running 

 through until April, every Hne of quadrant A gave large numbers 

 of pairs while not one pan- appeared in any pure line from the 

 three other quadrants. 



In the April test some lines of the other quadrants began to 

 develop the power to conjugate, but the number of pairs obtained 

 was small, in marked contrast to the epidemics in all lines of 

 quadrant A (diagram 3). 



The conjugation tests are made in all cases with individuals 

 from permanent cultures in which conjugations had not taken 

 place. The tests therefore are made each month with individ- 

 uals from the same pure lines but of different ages with respect 

 to the ancestral ex-conjugants. The power to conjugate develops 

 in different lines at different ages, this again showing variability in 

 the progeny of an ex-conjugant. Thus in series G conjugations 

 began in one line (quadrant D) three weeks after isolation of the 

 ancestral ex-conjugant; in the other quadrants it did not begin 

 until two months after isolation (diagram 6). In the J series con- 

 jugations in Jl, 7, 8, and 21 (quadrant A) began in November, in 

 lines J 15, 16, 17, 20, 25, 29, and 30 conjugations did not appear 

 until April, five months later (diagram 3). 



The power to conjugate of the conjugating lines of the J se- 

 ries (quadrant A) is correlated with (a) a lower vitality as measured 



