16 BR. J. F. GEMMILL ON THE 



The view put forward by Semon (6) that the first Echino- 

 derms were Pentactida-like forms (whose nearest representatives 

 are now to be found among the Syna/ptidce) has proved unsatis- 

 factory on many grounds. If we follow the Pelmatozoon theory, 

 of which Bather (1) is a foremost supporter, we must figure the 

 Proto-echinoderm as a Cystid-like animal. But from the em- 

 bryological point of view, as was first brought out by MacBricle, 

 there are strong objections to deriving the Asterids from any 

 form which was or had been attached by the (larval) right or 

 the (adult) aboral aspect, and in his memoir on the development 

 of Asterina, MacBricle (5, p. 398) put forward the view that the 

 Asteroids and Crinoids separated off from each other as early as 

 the fixed stage of the ancestor. 



It seems to me that recent embryological evidence and, in 

 particular, the data from Asterias rubens L., strongly support 

 this view, or, at any rate, that part of it which derives the 

 Asterids directly from the fixed ancestor. The larva of A. rubens 

 is a feeding bipinnaria, and conforms almost exactly to the con- 

 ventional Divleurxda. It has, as I believe, a primitive circulatory 

 centre (4, p. 273) resembling that found in the Enteropneusts 

 and Pterobranchs, the phyla most nearly allied to the Echino- 

 derms. It develops an attaching organ in the middle line 

 anteriorly, fixes itself, bends leftwards, and gradually acquires 

 radial symmetry. During this process, the larval cesophagus, 

 stomach, and intestine are retained, while the changes undergone 

 by the ccelomic cavities are perfectly simple and direct. The 

 attaching stalk, now connected with the oral side of the disc, 

 becomes separated off after the sucker-feet have acquired 

 adhesive and locomotor functions. On the other hand, the only 

 Crinoid larva we are acquainted with (that of Antedon) is quite 

 unlike a Dipleurula. It does not feed for itself, and the 

 ontogenetic development of its alimentary canal and coelomic 

 cavities is very different from ' any course that could have been 

 followed in evolution. In view of these facts, it seems to me 

 that we are compelled to derive the Asterids from the fixed 

 ancestor directly and not through the intermediary of a Pelma- 

 tozoic form. But the acquirement of radial symmetry could 

 only take place during an evolutionary period of great, if un- 

 known, duration, and throughout the later portion of this period 

 the attaching stalk must have been connected with the oral 

 aspect, and this aspect must have been turned towards the sea- 

 bottom or the surface of attachment. The difficulty now presents 

 itself that the characteristic Asterid mode of obtaining food 

 seems incompatible with this kind of fixation. MacBricle's 

 suggestion (5, p. 394) that the evolutionary change took place 

 at parts of the sea-bottom, where currents were continually 

 bringing along objects which might be seized upon by the 

 adhesive tentacles, is of too "ad hoc" a nature to be entirely 

 satisfactory without supplemental data. But if we find that 



