THE STAEFISH SOLASTEE ENDECA. 27 



cavity, now isolated, grows rapidly smaller and finally disappears. In this process its 

 lumen becomes filled up with wandering cells. 



As Macbride rightly stated in connection with the development of Asterina (15), 

 it is a point of great importance to determine whether, when the hydrocoele ring com- 

 pletes itself, the stalk of the preoral coelom is encircled by this ring or remains outside. 

 According to Macbride, the former condition occurs quite definitely in the case of 

 Asterma. In most of my Solaster specimens, owing probably to the relatively later 

 period at which the hydrocoele ring completes itself, the line of the stalk-cavity was 

 obliterated at the time of closure of the hydrocoele ring. It happens, however, that 

 variations in the degree of development of diff"erent parts occur in difiierent examples, 

 and in one or two instances I was able to ascertain with certainty the existence, for a 

 short time at least, of a communication on the oral side of the completed hydrocoele 

 ring, between the cavity of the preoral lobe and that of the axial sinus. 



As the relative lateness of closure of the hydrocoele ring in Solaster is readily 

 explicable on the ground that other four arms are, so to speak, intercalated after the 

 five primary ones, we may conclude that the stalk of the preoral lobe has the same 

 morphological relation to the hydrocoele ring in Solaster as in Asterina. 



Meantime the sucker and the tips of the larval arms are actively absorbed, but 

 normally a large part of the preoral-lobe epiderm is incorporated with that of the 

 oral surface (p. 19). This brings the cells of the larval apical field into the position 

 where the ring-nerve is developing, and makes it possible that the nervous system 

 of the larva should take some part in the formation of that of the adult (see further, 

 pp. 46-47). 



Sometimes the preoral lobe fails to apply itself so closely to the left body-wall as to 

 allow the incorporation of the epiderm with the left body-wall to be fully carried on as 

 described above. Such cases, if partial, may in the end develop all right, the redun- 

 dant tissue being ultimately absorbed. For a time the area (oral field) enclosed by the 

 hydrocoele is smaller than normal, but this gradually rights itself, though there is 

 some delay in the formation of the mouth-opening. 



Right Lateral Diverticulum of the Anterior Coelom. — The right lateral diverticulum 

 becomes separated from its parent cavity on the third or fourth day after fixation, 

 giving rise to the so-called epigastric coelom of the Starfish. Two chief factors are at 

 work in causing the separation. One is the gradual extension of the dorsal and ventral 

 horns of the posterior coelom in the larval sagittal plane (pp. 32-33), until in the end 

 they meet back to back at the madreporic interradius, enclosing between their layers 

 the various structures of the axial complex to be afterwards described (pp. 29-31). 

 The other factor is the sharp flexion and torsion of the preoral lobe and the subsequent 

 atrophy of the stalk-cavity. The notch produced externally on the disc (pp. 17, 22) by 

 these processes cuts deep into the neck between right lateral diverticulum and anterior 

 coelom, reducing this to a V-shaped cleft (PL IV. fig. 49 ; PI. V. fig. 55). The cleft is 



E 2 



