EARLY DEVELOPMENT OF CRIBRELLA OCULATA. 375 



itself to the side of a glass vessel, the interior of the brood-cavity is visible and the 

 progress of development can be watched. After hatching, the uniformly ciliated larvae 

 can be seen slowly moving about or rotating in the chamber, and occasionally they may 

 attach themselves in the later stages to the glass or to the buccal membrane of the 

 mother by their sucker. When the eggs are set free early in development all the 

 larvge sooner or later adhere to the glass by their suckers, until the young starfish can 

 creep about by their tube-feet. Whether any of the larvae leave the brood-chamber of 

 the mother in natural conditions I do not know, but these facts appear to indicate that 

 in the event of a parent either accidentally or intentionally discontinuing her protective 

 functions, each larva would be quite capable of proceeding in its development independ- 

 ently of its parent. Through the courtesy of Dr Appelloff of the Bergen Museum I 

 have been enabled to see a collection of embryos belonging to two species of Solaster, the 

 larval forms of which appear to be closely similar to those of Cribrella, though I under- 

 stand that in these instances there is no ' brooding.' Asterina gibbosa and Ophiura 

 brevispina appear to be other instances of a similar type of development. In all these 

 species there is a large amount of yolk which serves for the nourishment of the larva, 

 in some instances throughout the whole metamorphosis ; the larva is either uniformly 

 ciliated or banded, and is not pelagic, but lives at or near the bottom, in many cases 

 actually attaching itself to a foreign body. This type of development we may term the 

 ' demersal,' as it appears to have the same relationship to the ' pelagic ' type with its 

 bipinnaria or brachiolaria as the demersal types of development have to the pelagic. 

 Similar divergent types can be noticed in the Enter op neusta (Tornaria and Bateson's 

 larva). Other instances of the demersal type have been described in the Echinodermata, 

 and it appears to be of sufficiently frequent occurrence to forbid our regarding the 

 bipinnaria and other pelagic larvae as being by any means the rule amongst asterids. 



It is not uncommon to find at the present day certain statements regarding these 

 1 brooding ' types which would imply that the parent contributes nourishment to the 

 young during the progress of development. This view has undoubtedly been held by 

 several writers in the past, and the pre-oral lobe of Asterias mulleri has even been 

 described as a ' placenta ' ( Wyville Thomson). Neither in Asterias mulleri nor in 

 Cribrella oculata is there the slightest evidence that the connection between parent 

 and young is anything more intimate than a purely mechanical one, from which a 

 certain amount of protection is obtained. It has already been stated above that in 

 Cribrella (and the same applies to A. mulleri) the development proceeds perfectly 

 normally after the eggs are removed from the parent, which could hardly be the case 

 if they were thereby deprived of even a part of their nutriment. 



Methods. 



The eggs and larvae were preserved immediately on being taken from the mother. 

 One series was fixed in corrosive-acetic and passed gradationally through alcohols : 



