312 DESCRIPTION OF THE PLATES. 



sucker-like feet arranged on either side of the ventrally placed ambu- 

 lacral furrows, are seen in two rows on either side of the middle 

 line occupied by the mesial articulations of the successive pairs of 

 ' vertebral ' or ' ambulacral ' ossicles, k i, k 2. 



If the results of Ludwig's researches on the development of Asterina gibbosa 

 are applicable to all Asteroidea, a Starfish ought to be placed as follows to make it 

 coincide with the position of the parts relative to the antero-posterior axis of the 

 larva. The madreporite g must be placed to the left; and the anus e (figured 

 too near the centre of the disc), with the interradius between III and IV, to which 

 it belongs, must be anterior. This interradius coincides on the ventral surface 

 of the adult with the position of the anterior appendage of the larva. Five 

 processes develope on the right side of the larva, i. e. the dorsal or abactinal surface 

 in the adult, and are arranged in a circle, open anteriorly. To form the arms 

 they fuse with five similarly arranged processes of the left side, i. e. the ventral or 

 actinal surface here. Both sets of processes may be numbered from 1-5, com- 

 mencing with the anterior ventral process of the larva, and ending with the anterior 

 dorsal, passing round the posterior extremity. In the fusion of the two sets of 

 processes, abactinal and actinal, a curious twist of the one set upon the other 

 takes place. The actinal process i coincides with III in the Plate, 3, which marks 

 the posterior extremity of the larva, with II, and 5 with IV. The abactinal 

 processes, which correspond at their first appearance with the actinal, shift so that 

 2 coincides with III, 4 with II, and i with IV. Consequently a line drawn through 

 the anal interradius down arm II marks a line in the adult which coincides 

 with the antero-posterior axis of the larva. The original antero-posterior axis 

 of the abactinal surface passes, however, through the madreporite and arm I ; 

 for while this axis retains its position with reference to the larval antero-posterior 

 axis on the actinal surface, it is shifted to the left on the abactinal. It may be 

 noted that the larval anus or gastrula mouth lies between the actinal and abactinal 

 processes 3, but closes before the twist takes place. The larval mouth lies similarly 

 between the two processes i, but closes, and a new mouth is formed in connection 

 with a new oesophagus which grows out towards the left side of the larva. 



It would be interesting to know whether the twist occurs in other Asteroidea, 

 and in Ophiuroidea as well. See Ludwig, Z.. W. Z. xxxvii. 1882. 



a. Pyloric division of stomach, communicating freely with the cardiac, 



and giving off a stem which bifurcates as it enters each ray. 

 b i. One of the arborescent divisions into which the radial diverticulum 



of radius I divides. It is only in the Asteroidea that the digestive 



tract has this radial arrangement of coeca. 

 b 2. Arborescent coecum of radius III, displaced, as is its fellow, into the 



interradial spaces. 

 b 3 and b 4. Terminations of coeca of radius V attached to the dorsal 



integumentary skeleton by a mesentery. 

 c. Cardiac division of stomach, bulging, but only for a short distance, into 



the cavity of the several rays at a lower level than the coeca, b. To 



