102 The Mechanistic Conception of Life 



different places from those they would have held m the case of 

 normal segmentation. Still, normal embryos resulted. One 

 might object again that the preformation of the germ-regions 

 existed in the protoplasm, and not in the nucleus, I have made 

 a series of experiments to the results of which these objections 

 cannot be made. I shall describe these experiments somewhat 

 fully, as they have not yet been published, though I cannot 

 enter into details at this place. 



I brought eggs of a sea-urchin, within ten to twenty minutes 

 after impregnation, into sea-water that had been diluted by the 



Fig. 34 Fig. 35 



Fig. 34. — Fertilized egg of a sea-urcliin (Arbacia) put into dilute sea-water. 

 The protoplasm swells until the membrane m bursts and part of the protoplasm 

 6 flows out; each of the two droplets may develop into a blastula, so that from 

 such an egg two larvae may arise, as Indicated m Fig. 35. 



addition of about 100 per cent distilled water. In this solution 

 the eggs took up so much water that the membrane (m. Fig. 34) 

 burst and part of the protoplasm escaped in the form of a drop 

 (6, Fig. 34), which often, however, remained in connection 

 with the protoplasm inside the membrane after the eggs were 

 brought back into normal sea-water. These eggs gave rise 

 to adherent twins, the ejected part h of the protoplasm, as well 

 as the part remaining inside the membrane, developing into a 

 normal and perfectly complete embryo. The part of the proto- 

 plasm, which at first had connected the two drops, formed the 

 part where the twins remained gro\Mi together. Of course, it 

 often happened that, by accident or rapid movement, the twins 

 were separated, and they then developed into perfectly normal 

 single embryos. Since we caimot assume that in every case the 



