444 



LORANDE LOSS WOODRUFF. 



figure shows that A underwent endomixis at periods Nos. 4, 9, 

 15, 21 and 28. AE showed endomixis at period 4 and then was 

 lost by an accident. A new AE was started by isolating line 

 by line from AEt (which had been branched from it 25 days 

 before and subjected to the constant culture conditions). 



AE 



FIG. 4. 



AE then showed endomixis at the periods 9 and 15 and died 

 out in period 25 without again undergoing endomixis. Still 

 another AE was isolated immediately from AEt and had endo- 

 mixis during the 28th period. A similar survey of the graph of 

 shows that endomixis took place at periods 4, 9, 15 and 20. 

 Subculture died at period 25 from causes unknown. It was 

 restarted from the main culture of this race at once and had 

 endomixis at period 28. Subculture B was begun in period 

 18, that is immediately on the isolation of the race from 'wild' 

 material. Endomixis occurred at periods 21 and 28. 



Tabulation of these data shows that endomixis occurred in 

 these subcultures as follows: 



