63 R. HOLMAN PECK. 



(PI. VI, fig". 14) to be the transverse sections of rod-like 

 bodies. These rod-like bodies — the well-known " Stabchen " 

 of German writers — appear to be simply condensed or firmer 

 portions of the larger tnass of chitinous substance. It is 

 interesting to compare these sections with those of Dreissena 

 (PI. VII, figs. 25, 26) in which the rods are absent. 



Looking now at the longitudinal section (Plate VI, fig. 14), 

 we see that where the transverse fibrous junctions are deve- 

 loped, the rods bend a little outwards and become very thin 

 and small so as almost to cease. They do not, however, quite 

 disappear at the level of the transverse junctions, but are 

 continuous from end to end of each filament, merely becom- 

 ing small at every transverse interval. At the base of the 

 gill — the line of origin — the rods of neighbouring filaments 

 join, forming an arch, the anterior of one filament joining the 

 posterior of the next in order, as we pass forward from anus 

 to mouth. I have not figured this arrangement which is 

 drawn by Posner, who calls attention to the similar arrange- 

 ment in the chitinous rods of the gill bars of Amphioxus.l 



Now, if we take a section across the filaments at the level 

 of the transverse junctions we shall find the relationship of 

 the fibrous transverse bands to the chitinous half-wall of the 

 filament. Such a section is exhibited in PI. V, fig. 13. The 

 chitinous rods are seen reduced to a minimum size, the semi- 

 lunar masses of paler chitinous substance are broader and 

 larger than in the other section and break up deeply into 

 fibrous substance, the fibres being attached to a central row 

 of corpuscles. 



There is no question whatever that Jtere at least there is 

 complete obliteration of the primitive cavity of the gill-fila- 

 ment. Even though we may admit that the space between 

 the semilunar masses in fig. 12 represents the primitive 

 cavity, it cannot be regarded as Posner regards it, namely, as 

 a distinct canal. It is widely open to the general cavity of 

 the lacunar tissue and at every transverse junction is entirely 

 obliterated. Compare with this the condition of Dreissena 

 (Plate VII), which also furnishes interesting evidence fur 

 the relationship of the chitinous deposit to the cell-elements 

 of the lacunar tissue. The chitinous deposit is merely fo 

 be regarded as a peculiar and largely developed intercellular 

 product. 



The Epithelium of the Gill-filaments — I have devoted special 

 care to this part of the subject, since my results differ from 



' The similarity between tlio structure of the |)liaryiix of Anipliioxus and the 

 Lainellibrauch gill is an iustructive examj)le of what Mr. Lankester has called 

 liomoplusy as opposed to homogeuy i^' Annals and Mag. Nat. Hist.,' 1870). 



