44 Prof. M'lutosh's yotsfrom (he 



nut ahvuNs easily observed, and this is apparently due to 

 the rells of tho liy[>odenn or to a supporting tissue witliin 

 it, tlio toriner interpretation being the more liki'l\ , as no 

 dirt'crentiation is observed in transverse seetion. 



Diverse views have been held with regard to the structure 

 of the filaments and pinnules ; thus Meyer described a 

 diverticulum of the coelom in each filament and pinnule, 

 whilst Orley insisted that only connective tissue occupied 

 tlie centre. It is by no means easy to decide, since in the 

 case of sections the parts arc considerably altered even in 

 good preparations. A ccelomic space occurs on both sides 

 at the level of the dumbbell-shaped region in front of the 

 brain (PI. IV. figs. 20 & 2l,cce.), and their walls are defined 

 by connective-tissue, and probably muscular, fibres, the 

 area surrounding them consisting of nucleated connective- 

 tissue cells. About this level the thoracic jacket or collar 

 has just become free or is only connected by a narrow 

 isthmus. As a rule, also, the two sides are asymmetrical in 

 section, the opercular half having no slits, but a considei'able 

 coelomic space, whilst the other side has only small apertures, 

 so that the area within (that is, ventral to) the slits is reticu- 

 lated, these reticulations in the succeeding sections becoming 

 less and less until only the branchial vessel is evident. The 

 ei)ithelium surrounding the slits becomes reguhirly arranged 

 and forms the hypoderm and cuticle of the filaments, each 

 side being attached to a separate filament. The elongated 

 centre of each filament in formation is almost wholly occupied 

 by nucleated connective tissue with the blood-vessel in the 

 centre, but two splits, one on each side of the mesentery, are 

 often seen at the distal end of the central area, occasional 

 strands of tissue crossing the spaces in some sections. The 

 definite median mesentery with its central blood-vessel and 

 the definite coelomic spaces at each side, and from end to 

 end in transverse section of a pinnule, as shown by Soulier 

 in Protula milliaci, have not been observed either in 

 filament or pinnule. In longitudinal sections of a filament, 

 the sides are formed of cylindrical nucleated epithelium, 

 whilst the centre is almost filled with nucleated connective- 

 tissue cells, a narrow split at one or other side being present, 

 and even this has a few strands with nuclei. 'J'lie pinnules 

 of this form ( PomatuCcruR) show only a central cavity in 

 which the blood-vessel is (PL IV. fig. 24), but the coelomic 

 fluid could readily rush to and fro in the space around it, 

 whether a special mesentery fixes it or not. On the whole, 

 therefore, the view that the coelomic spaces — carried forward 

 to the splits for the commencing branchial filaments — do not 

 blindly end there, but communicate with the filaments and 



