267 



but other characters so much resembling those of the Hemichor- 

 data that Actinotrocha appears to bear the same relation to Balano- 

 glossus and to Phoronis respectively as the tailed Ascidian larva does 

 to the Amphioxus and to the adult Ascidian. In other words, the 

 development of Phoronis reaches its highest morphological stage in 

 the fully formed Actinotrocha and subsequent stages leading to the 

 adult condition involve processes of degeneration similar to those 

 found in Ascidian ontogeny. 



In the «stomach« region (ii Actinotrocha a pair of diverticula of 

 its w^alls run forwards on either side of the oesophagus for a short 

 distance and end blindly. These are readily seen in a view of the 

 larva intact and, have been figured and labelled by Wilson (Quarterly 

 Journ. Micr. Science XXI) and others as »glandular part of sto- 

 mach«. In sections, however, they present no trace of glandular 

 structure but shew all the characteristics of notochordal tissue. These 

 two hypoblastic diverticula I would compare to the notochord of 

 higher forms: they arise from the region which, as will be seen, is 

 comparable to the collar of the Hemichordata. 



The large hoodlike pre-oral lobe is filled internally by a large 

 impaired coelomic cavity definitely lined by mesoblast and shut off 

 posteriorly from the two tentacular cavities by a mesentery. Under the 

 apical plate there is a haemocoele space between the epiblast and the 

 mesoblast. The posterior part of the pre-oral coelome comes in con- 

 tact with the front swollen termination of the dorsal aorta, which is 

 itself merely a haemocoelic (blastocoelic) space dorsal to the gut. Here 

 the coelomic wall is glandular and thickened. These relations prompt 

 an almost irresistible comparison with the , proboscis gland*^ and ,heart'^ 

 of B alano glossus. 



Immediately behind the pre-oral coelome, lie the tentacular coe- 

 lomic cavities which are separated by mesenteries from the pre-oral 

 coelome in front and the trunk coelomes behind. These are produced 

 into the tentacles on eiter side. I have not yet determined how far 

 they are separated from each other by a dorsal mesentery. It is in this 

 region that the notochordal diverticvila arise and the two cavities are 

 obviously homologous with the »collar cavities« oftheH emichordata. 



The two trunk cavities lie on either side of the intestine and seem 

 to fuse dorsally; there is a circular haemocoele space lying immedia- 

 tely under the thickened peri-anal ciliated ring. 



I am inclined to believe that the ciliated nephridial tubules open 

 into the collar cavities, but cannot speak with certainty. 



There is a ciliated area along the free border of the pre-oral lobe, 



