ECHINODERMATA. 467 



a rudimentary tube foot, and on each side of the middle lobe two 

 fresh ones next spring out, and so on in succession. The terminal 

 median lobe forms the tentacle at the end of the arm, and the eye is 

 developed at its base. The growth of the water-vascular canals keeps 

 pace with that of the arms, and the tube feet become supported at 

 their base by an ingrowth of calcareous matter. The whole of the 

 calcareous skeleton of the larva passes directly into that of the adult, 

 and spines are very soon formed on the plates of the abactinal surface. 

 The original radial plates, together with the spines which they have, 

 are gradually pushed outwards with the growth of the arms by the 

 continual addition of fresh rows of spines between the terminal plate 

 and the centre of the disc. It thus comes about that the original 

 radial plates persist at the end of the arms, in connection with the 

 unpaired tentacles which form the apex of the radial water-vascular 

 tubes. 



It has already been mentioned that according to Metschnikoff (No. 560) 

 a new oesophagus is formed which perforates the water- vascular ring, and 

 connects the original stomach with the original mouth. Agassiz (No. 543) 

 maintains that the water- vascular ring grows round the primitive oesophagus. 

 He says "During the shrinking of the larva the long oesophagus becomes 

 "shortened and contracted, bringing the opening of the mouth of the larva 

 "to the level of the opening of the oesophagus, which eventually becomes 

 "the true mouth of the starfish." The primitive anus is believed by 

 Metschnikoff to disappear, but by Agassiz to remain. This discrepancy 

 very possibly depends upon these investigators having worked at different 

 species. 



There is no doubt that the whole of the larval organs, with the 

 possible exception of the oesophagus, and anus (where absent in the 

 adult), pass directly into the corresponding organs of the starfish and 

 that the prse-oral part of the body and arms of the larva are absorbed 

 and not cast off. 



In addition to the Bipinnarian type of Asteroid larva a series of 

 other forms has been described by Miiller (No. 561), Sars, Koren, and 

 Danielssen (No. 554) and other investigators, which are however very 

 imperfectly known. The best-known form is one first of all discovered 

 by Sars in Echinaster Sarsii, and the more or less similar larvae subse- 

 quently investigated by Agassiz, Busch, Miiller, Wyville Thomson, etc. 

 of another species of Echinaster and of Asteracanthion. These larvse on 

 leaving the egg have an oval form, and are uniformly covered by cilia. Four 

 processes (or in Agassiz' type one process) grow out from the body ; by these 

 the larva? fix themselves. In the case of Echinaster the larvae are fixed in the 

 ventral concavity of the disc of the mother, between the five arms, where a 

 temporary brood-pouch is established. The main part of the body is con- 

 verted directly into the disc of the young starfish, while the four processes 

 come to spring from the ventral surface, and are attached to the water- 

 vascular ring. Eventually they atrophy completely. Of the internal 

 structure but little is known ; till the permanent mouth is formed, after 

 the development of the young starfish is pretty well advanced, the stomach 

 has no communication with the exterior. 



30-2 



