INITIATION OF DEVELOPMENT IN EGG OF ARBACIA. 389 



every single mature egg sJwu'cd a membrane. Since, moreover, 

 experiments were made throughout the season, the results can not 

 be interpreted as mere incidental findings based on insufficient data. 

 These membranes induced by hypertonic sea-water separate 

 more slowly than membranes lifted from the eggs following nor- 

 mal insemination. These membranes are, nevertheless, as clear 

 and as distinct and possess as wide a perivitelline space as normally 

 fertilized eggs. In the hypertonic sea-water the egg shrinks, its 

 periphery retaining a smooth contour. One gains the impression 

 that the perivitelline space arises in part as the result of this 

 shrinkage. That this is not wholly correct seems to be indicated 

 by those eggs that undergo an equal amount of shrinkage without 

 forming membranes. Moreover, on return to sea-water the eesr, 



oo 



though it increases in size, does not obliterate the perivitelline 

 space. 



If the intensity of the membrane separation process be too great, 

 the membrane formed is eccentric ; the perivitelline space is not of 

 the same width in all zones of the egg. In such cases the egg, as 

 seen in optical section, is flattened in that zone above which the 

 membrane is at its greatest distance from the egg. On return of 

 the egg to normal sea-water this eccentricity of the membrane 

 persists. The cortex of that zone, in these cases, from which the 

 membrane has separated least, is apt to be swollen. This seems 

 to indicate that the reaction underlying membrane must be of a 

 certain order to insure best results. 



The membrane does not always arise in the manner described. 

 In some cases the egg presents a crenated surface beneath the 

 membrane. This crenation may quickly disappear, leaving the egg 

 cortex below the membrane perfectly smooth. If the crenation 

 persist, on return to normal sea-water the perivitelline space is 

 very narrow ; indeed, it may be absent, in which case the membrane 

 is closely stuck to the swollen cortex. 



Finally, in some cases the membrane may be extremely thin, 

 though otherwise the egg and perivitelline space are about as found 

 in the normally fertilized egg. 



These observations on the effect of hypertonic sea-water in 

 bringing about membrane separation, fortunately, do not stand 

 alone. I find that Loeb almost twenty years ago made a similar 



