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



however, the progeny of one quadrant (B) died out on January'- 



12, leaving only three pure lines. In these the February tests 

 showed conjugations in all three; 49 pairs from quadrant A; 1016 

 pairs from quadrant C, and 1890 pairs from quadrant D. Here, 

 therefore, two lines of the four are strong conjugating lines and 

 these were derived from the same parent cell in the 2-cell stage. 



A third set of rich cultures was started iVIarch 15 and conjuga- 

 tion tests were made on the 19th and 22nd. These two tests 

 gave 4610 pairs in quadrant A, 2760 paii's in quadrant C and 

 2540 paii'S in quadrant D. The series was then discarded since 

 one quadrant had died out and the others were all capable of 

 giving high numbers of pairs. The progeny from this ex-con- 

 jugant were particularly fertile j'ielding conjugations in less than 

 two weeks after isolation of the ancestral ex-conjugant, and the 

 epidemics of conjugation at each succeeding test. 



/. Conjugations in the M series (diagram 6) 



In this series all four cells of the original 4-cell stage gave rise 

 to conjugating lines, in which however, the power to conjugate 

 is quite different. The series was started as before on December 



13, and the first rich cultures were made on January 7. The 

 first conjugation tests on Janurj^ 9, 11, and 24, yielded 4 pairs 

 from quadrant B; 57 pairs from quadrant C; 4 from quadrant D, 

 but none from quadrant A. The second set of rich cultures were 

 made on February 13 and conjugation tests on February 17 and 

 20. These tests yielded 1420 pairs in quadi'ant A; 105 pairs in 

 quadrant C; 99 in quadrant B; and 2 in quadrant D. 



These tests show that if the conjugating ability is present it 

 may be weak at one time (e.g., quadrant A in January) but may 

 blossom out in an epidemic at another time. 



A third set of rich cultures was started March 15th and one 

 test was started on the 19th. This gave no conjugations in quad- 

 rant C, 218 pairs in quadrant B, and 1802 pairs in quadrant D, 

 showing that the earlier sterility in the last was onl}^ temporary. 

 Quadrant A died out before this thii'd test was started. The 

 enth'e series was then discarded. 



