TROCHOPHORE OF HYDKOIDES UNClXATrs (eUPOMATUS). 575 



niesoblast bands liave not appeared. In the young troclio- 

 phore shown in tig. 1. they are just appearing as they grow 

 out from the gut-wall. As these cells just mentioned are 

 converted into the head-kidneys before the raesoblast bands 

 have appeared^ it is fair to assume that they do not represent 

 the coeh^mesoblast cells as Hatschek and Soulier claim. It 

 must also be recalled that both these investigators have not 

 followed the cell lineage, and therefore they have no definite 

 grounds of cytological importance on which to substantiate 

 their claim as to the nature of these cells. 



In the early stages of invaginatioii it is certain that the 

 coelomesoblast cells cannot be distinguished, as these investi- 

 gators state, by their conspicuous size and the manner in 

 Avhich they force their Avay into the segmentation-cavity. In 

 fact, I have been only able to distinguish them satisfactorily 

 in early stages by following their development backwards 

 from a stage when these are clearly recognisable in the anal 

 end of the gut to a stage towards the end of gastrulation ; 

 prior to this 1 cannot see that they differ from any of the 

 other endoderm cells. 



In fact, the condition in Eupomatus is simply a more 

 marked type of that found in Podarke. In this Annelid, 

 according to Treadwell (43), at the sixty-foui'-cell stage the 

 fourth group of micromeres have just formed. They are all 

 alike, but shortly one of them divides bilaterally, thus 

 aiding substantially in the establishment of the bilateral 

 symmetry. Then each buds off a small cell ventrally ; these 

 small cells form a part of the wall of the archenteron. 

 During the course of gastrulation the mesoblast cells lie in 

 the wall of the archenteron, Avith which they are carried 

 inwards, finally coming to lie in the anal region. They 

 protrude considerably, and in sections that pass (Text-fig. 6a) 

 a little to one side of the sagittal plane they seem to lie 

 actually free in the segmentation-cavity. 



The condition in Podarke and Eupomatus, again, is 

 only a more marked state of that found in Crepidula, 

 where the gi-eater portion of the primary mesoblast cell 



