10 



the eggs were exposed to the salt-solutions for a very long period 

 (3 — 5 hours or even longer). In one of the cases figured (Fig. 9), three 

 asters have formed after 5 hours in KCl-solution. Another (Fig. 21, 

 stated to be of the same stage as Fig. 17; cf, p. 41), is the "stage" 

 with four asters, which under his interpretation must represent the 

 product of two divisions of the primary centrosome. This "stage" is 

 from an egg treated for 41/2 hours with Na Cl-solution. 



From these cases the mathematical reader may readily estimate 

 the length of time that M'ould be required at this rate to produce such 

 a stage as that shown in Fig. 20, where the egg is filled with »echte 

 Strahlungen (f and centrosomes. Now, in Toxopneustes the exposure 

 to the salt-solution was rarely as long as two hours, and often less 

 than one hour; yet in this short time the eggs frequently developed 

 hundreds of asters, and scores of these, as shown in sections, contained 

 perfectly distinct centrosomes. That so great a number could arise in 

 so short a time by the division of a single primary center is a priori 

 improbable, and utterly impossible if the observed rate of division 

 observed at a later period is assumed to be even doubled or trebled in 

 the early stages. 



It must, I think, be evident from the foregoing on what doubtful 

 evidence the conclusions of Petrunkewitsch regarding progressive 

 division are based. It is indeed possible, as I pointed out (op. cit. 

 p. 555), that a division of the cytasters may take place at the time the 

 nuclear bipolar figure forms; but my evidence of this was of the same 

 inconclusive character as that produced by Petrunkewitsch, and 

 must be supplemented by renewed observations on the living eggs. 



The second point of Petrunkewitsch, namely, his failure to 

 observe the formation of asters and centrosomes in enucleated frag- 

 ments, is not open to criticism; though there is a rather puzzling 

 inconsistency between the statement at p. 37 that such fragments 

 »bleiben unveränderter and the one on the preceding page that radia- 

 tions do appear in the fragments, though only in rare cases and in 

 small numbers. Accepting the latter statement as corresponding to 

 the facts, it is not surprising that so few asters were observed in non- 

 nucleated fragments, or that these showed no central bodies; for, in 

 Toxopneustes too, the asters frequently fail to form, and as stated 

 in my paper (p. 562) they vary greatly in development, often showing 

 no central bodies. Against this stands the fact that the asters of the 

 enucleated fragments not infrequently do show as characteristic and 

 definite a radiation, and as clearly marked a central body as in the 

 whole eggs. It is not admissible to suppose that these asters are "pri- 

 mary radiations", or are otherwise not comparable with "true asters"; 



