EARLY DEVELOPMENT OF CRIBRELLA OCULATA. 377 



Like Sars, I have found that in confinement about eight to ten days may easily be 

 occupied by the embryonic period till hatching, whilst the larval period may perhaps 

 be put down as about six to ten days. Of the duration of the post-larval period I 

 cannot speak with any certainty, but the whole development is undoubtedly prolonged. 



Part I. Embryonic Period. 



If the forty or more eggs belonging to one parent be emptied out of her brood- 

 cavity, the first feature to be noticed is their great variation in size. They all appear 

 as bright orange spheres, and the majority have a diameter of about 1 mm., but 

 some may be as large as 1*2 mm., whilst others are clown to "8 mm. A similar variation 

 is found in the latest young stages and at all intermediate stages, so it is legitimate to 

 suppose that the abnormality in respect to size is not necessarily prohibitive of normal 

 development. 



Again, the eggs, as a rule, exhibit great diversity in their stage of development, 

 a fact observed previously by Sars. Some may be not yet segmented, whilst others 

 may have nearly completed their embryonic stages, and all intermediate conditions 

 between these two may be found. 1 have only very rarely found an embryonic egg 

 amongst the parents in which the majority are free larvae, and these clutches, in which 

 the young are more advanced (i.e., in second and third periods), usually show far 

 greater uniformity in structure. This is probably to be explained by the fact that 

 the early embryonic stages occupy a much shorter time than the late embryonic and 

 early larval, so that early inequalities are reduced to a minimum as development 

 proceeds. 



Each egg is enclosed in a very delicate egg-membrane, in which it remains till the 

 end of this period. 



Segmentation. — Taking into account the large quantity of yolk, one might perhaps 

 expect that there would be a modification of the total equal segmentation which is usually 

 regarded as typical of the Echinodermata, but one could hardly have anticipated 

 the actual facts of the case. 



These are so extraordinary that although the main features were known to me 

 four years ago, I have hesitated to publish them until I had gone over the fresh material 

 for several seasons, and even now the results, from this very peculiarity, leave a feeling 

 of dissatisfaction. To sum up, we may say that there is no definite type of segmentation 

 whatever. The eggs present the widest possible varieties, which only meet on common 

 ground at the end of segmentation when the egg is in all cases converted into a solid 

 morula, with approximately equal cells throughout. In one type which is by no means 

 uncommon, the first plane of segmentation is horizontal (fig. 7), and divides the whole 

 egg into two unequal segments. This is followed by two perpendicular planes producing 

 an eight-celled stage with a small blastoccele cavity (fig. 8). This type reminds one 

 forcibly of the egg of the frog, and is of the holoblastic unequal type. After this stage, 



