DEVELOPMENT OF SOME STARFISHES. 555 



appear that the change towards loss has advanced further on the 

 west coast of Scotland than on the east and in Norway. 



(2) Segmentation. — Segmentation proceeded regularly in most 

 of my specimens, equal holoblastic cleavage being the rule and not 

 the exception. The eggs float, the pole which'keeps undermost 

 being usually a little lighter in colour than the upper pole, but 

 there is not the same marked difference either as regards specific 

 gravity or colour between the upper and the under poles as in 

 the eggs of Solaste7' endeca. Masterman did not find that there 

 was a.ny definite or consistent type of segmentation in his ova of 

 Criby^Ua, and he does not refer to differences in colour or specific 

 gravity between the upper and the under sides (9, p. 377). 



(3) Blasiula formation. — The surfa.ce-furrows along which 

 egression takes place during blastula-formation almost certainly 

 exhibit reversion to the segmentation patterns. In the earlv 

 blastula the lines ai-e very numerous, simulating an advanced 

 stage in segmentation ; then they become fewer, simpler, and 

 deeper, so that sometimes we cannot tell with the naked eye 

 whether we are looking at an advanced blastula or at an 8-celled, 

 a i-cehed, and a 2-celled stage in segmentation. Peculiarities in 

 the arrangement of the early segmentation-furrows (e. g. obliquity 

 of the fii-st cleavage plane, or marked inequality in the first and 

 second divisions) tend to emerge again in the late blastula, as may 

 be seen if one watches the development of isolated ova which 

 exhibited the peculiarities in question. I have had the oppor- 

 tunity of noting a similar phenomenon in the case of Solaster 

 papposus, unci in Solaster endeca (1, p. 12) I called attention to 

 the similarity between the cell-egression furrows and the segmen- 

 tation-farrows. ■ Masterman's illustration (9, pi. i. fig. 6)'^of an. 

 early gastrula of Crihrella strikingly recalls a 4-celled stage in 

 cleavage. 



(4) Hypencliyme. — In my gastrulae and early larvae there was 

 only a small quantity of cellular material within the archenteron, 

 less than one-fourth of this cavity being filled with the material 

 in question (cf. Solaster, 1, p. 45). Masterman describes the 

 archenteron during these stages as being entirely filled with 

 hypenchyme (9, p. 381). 



(5) Pharyngeal or perioral coilom. — I can confirm the obser- 

 vation of Masterman that this ccelom arises in the form of 

 interradial pouchings from the posterior coelom, and can add 

 further that such pouchings occur in interradii I.-II. and I.-V.* 

 as well as in the others, a point regarding which Masterman did 

 not speak with certainty (9, p. 416). 



The pharyngeal coelom arises by a single outgrowth in 

 Asterina (7, p. 358) and Asterias ruhens (2, p. 259), as also in 

 Ophioihrix (8, p. 497) and Synapta (8, p. 536). On the other 



* The inirabeving- of raj^s adopted in this papev is that emploj-ed by MacBride in 

 his account of the development oi Asterina, and by mj'self in the case o^ Solaster, 

 Asterias, and Forania. The madreporic radius is I.-Il., ray II. being on its dextral 

 or watch-hand side as viewed from the oral aspect. 



39* 



