ARTIFICIAL PARTHENOGENESIS IN ANNELIDS 663 



next morning, after they had been in these solutions for 

 twenty-four hours, the first solution contained many eggs 

 in the beginning stages of segmentation but not one swim- 

 ming larva could be discovered. The second and third 

 solutions contained a large amount of swimming larvae; in 

 the second they were more numerous than in the third. In 

 the control material only a few eggs began to segment ; no 

 swimming larvae were to be found. This confirms our former 

 observation that an addition of i| c.c. 2^n KC1 to 99^ c.c. 

 sea-water is insufficient to produce parthenogenesis, while 

 the addition of 1 KC1 is sufficient. 



Eighteenth series. There is something paradoxical in 

 the fact that the addition of 2 c.c. 2^ n KC1 to 98 c.c. sea- 

 water can produce parthenogenesis in three minutes, while 

 the addition of ^c.c. 2^w KC1 to 99^ c.c. sea-water cannot 

 accomplish the same result in twenty-four hours. Before I 

 accepted this as a fact I wished to see it confirmed once 

 more. The same solutions were applied as before: 



(1) | c.c. 2|n KC1 + 99| c.c. sea-water 



(2) 1 :< " +99 



(3) li " +98i 



(4) Normal sea-water (control) 



Part of the eggs were put back into normal (sterilized) 

 sea-water after thirty minutes, while the others remained in 

 the solution during the next twenty-four hours. As far as 

 the latter are concerned the results were exactly like those 

 described in the preceding experiment. The eggs that had 

 remained in solution 1 over night had not developed beyond 

 the early cleavage stages. No egg had reached the trocho- 

 phore stage. In the second solution a large number of swim- 

 ming larvae were found, and in the third solution they were 

 numerous. About 75 per cent, of all the eggs were in the 

 trochophore stage, and many of these were swimming about. 



The eggs that had remained in these solutions only thirty 



