THE DEVELOPMENT OF OPHIOTHEIX FBAaiLIS. 561 



pair of discs lying at the sides of the stomachy and concluded 

 that they had been segmented off from the oesophageal cavities, 

 which he could still make out. This is the first assertion of 

 a trace of metameric segmentation in any Echinoderm larva. 

 In subsequent stages he discovered that the left anterior 

 cavity assumed the five-lobed form characteristic of the 

 hydrocoele, whilst the right grew smaller and ultimately dis- 

 appeared. In the next stage he showed that the hydrocoelic 

 rudiment had by growth changed its longitudinal position for 

 a transverse one^ and was, in fact, beginning to surround the 

 larval oesophagus. Miiller had supposed that, as in Echiuoidea, 

 the larval oesophagus disappeared, but Metschnikoff shows 

 that it persists. He further corrected Miiller's error about 

 the " palmate organs/' and showed that each of these corre- 

 sponded to one lobe of the original palmate organ or hydro- 

 coele. He showed also that the primary madreporic pore, 

 although formed on the left side, rotates during metamor- 

 phosis to the right along with the left pra)-oral arm of the 

 Pluteus. In a word, Metschnikoff made out almost everything 

 that it is possible to discover in a Pluteus without the use of 

 sections. 



The viviparous species, Amphiura squamata, offered an 

 opportunity of studying the embryonic type of development 

 in an Ophiurid. Its viviparity was discovered by Quatrefages 

 in 1842. He communicated his results to Milne Edwards, by 

 whom they were published. Krohn published in 1851 a paper 

 on the subject (14), dealing with the relation of the embryo 

 to the maternal tissues. In 1852 Schultze (30) discovered a 

 transitory larval skeleton, which he compared to the skeleton 

 of the Pluteus. 



In the paper referred to above Metschnikoff records most 

 valuable observations on the development of Amphiura 

 squamata, although the younger embryos are difficult to 

 obtain. He found a thick-walled blastula, and then a later 

 stage, in which the coelom was represented by two thick- 

 walled bodies lying at the sides of the archenteron. In the 

 next staofe each of these bodies had divided into anterior and 



