306 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



remain in the radial line, the ambulacra from the numerous subdivisions of the 

 anterior ray could not possibly join before reaching it. The shifting of the mouth 

 to the right appears to be an effort to minimize the number of groove trunks 

 reaching it and to equalize the amount of food delivered by each. 



The shifting of the mouth into the interradius A — B and never into the inter- 

 radius A — E is probably connected with the coiling of the digestive tube. 



In the early stages and until a considerable time after the resorption of the 

 orals the mouth is central. It first migrates anteriorly along the line of radius 

 A to a position near the base of the anterior arm and later moves to the right. 

 In the adults it may occupy a position anywhere along this course between the 

 center of the disk and the middle of the border of interradius A — B. 



P. H. Carpenter notes that in the Comasteridae the mouth occupies the center 

 of the peristome (fig. 687, p. 341) and is bounded by two lips — a large anterior and 

 a smaller posterior (fig. 1154, pi. 25) — so that its opening is very inconspicuous 

 and usually so much extended in a direction transverse to the antero-posterior 

 diameter of the disk that it presents the appearance of a simple slit. The circum- 

 oral portion of the peristome, the peristome proper, is a more or less oval depres- 

 sion in the ventral perisome of the disk which completely surrounds the oral open- 

 ing and gives off the ambulacral grooves. Beneath this depression lies the water 

 vascular ring, which gives off a trunk under each of the grooves radiating from 

 it. The nimiber and distribution of these are very variable and primarily depend 

 upon the way in which the grooves divide so as to give rise to the trunks corre- 

 sponding to the 10 primary arms. As a general rule, the two ambulacra corre- 

 sponding to the radii D and E unite into one large posterior trunk from which 

 the branches are distributed to the various arms into which these radii divide. In 

 other cases the left lateral ambulacrimi, E, leaves the peristome alone, while in 

 others it is partially united with the posterior ambulacrum, D, its anterior division, 

 E,, leaving the peristome by a separate trunk, while its posterior division, E^, 

 unit€s with the posterior ambulacrum. As a general rule, the right lateral ambu- 

 lacrum, C, leaves the peristome alone and supplies the arms of the corresponding 

 radius, but occasionally it is seen to unite with the posterior division, Bj, of the 

 right anterior ambulacrum, B. The mode of division of the two anterior ambu- 

 lacra (A and B) is excessively variable; as a general rule, there are no principal 

 trunks corresponding to the two radii A and B, and the primary divisions A,, Aj, 

 Bi, and Bj originate directly at the peristome (fig. 690, p. 341). In the specimens 

 with but few arms, however, each pair may be united for a longer or shorter 

 distance, as in Antedon. Not infrequently the posterior divisions A, and B„ of 

 these two anterior ambulacra unite for a longer or shorter distance with the two 

 large aboral groove trunks to form an open horseshoe-shaped cur^'e bounding the 

 anal area. The position of the anal tube in this area, and also in regard to the 

 whole surface of the disk, varies somewhat with the position of the mouth; it is 

 rarely, if ever, absolutely central. Its appearance differs very much according as 

 it is full or empty ; sometimes its aperture is so completely closed as to be scarcely 

 discernible though the tube below is widely distended, and sometimes the aperture 



