66 K. HOLMAN PECK. 



deposit in the filament of Dreissena shows us the presence of 

 rings of deposit or varying tension, which woukl no doubt 

 make it an interesting object for the polariscope (see Plate 

 VII, figs. 25, 26). 



If the section be taken near the horizon of a transverse 

 fibrous junction, we find (see upper part of fig. 25 and 

 fig. 26) that the chitinous mass does not retain its open 

 semi-canalicular character, but becomes a closed solid mass, 

 with a central tract of corpuscular elements {nch.). More- 

 over, at its deeper end we find connections with the fibrous 

 or trabecular elements of the lacunar tissue, and also curious 

 lobe-like additions of chitinous matter added to the chief 

 mass, and penetrated by obscure fibrous septa. This ex- 

 tension and thickening of the chitinous skeleton of the 

 Dreissena gill-filament only takes place at the horizons of 

 interfilamentar transverse fibrous junctions. Between these 

 horizons the chitinous axis of the filament is a half-canal, 

 and in section of a horse-shoe shape, or rather like a letter 

 u Avith its sides thickened internally. The lacunar tissue of 

 Dreissena exhibits fewer hand-like trabeculee than does that 

 of Anodon ; the processes of its cells being rather more 

 fibrous and filamentar in character. The interlamellar junc- 

 tions are of the condensed variety of lacunar tissue, with 

 fusiform cells (fig. 30, Plate VI), but in parts exhibit a 

 looser structure (compare the two in fig. 26, Plate VII). 

 Pigmented cells with dark brown granules, besides blood- 

 corpuscles, are found ijoating in the blood-lymph, or attached 

 to the processes of the lacunar tissue. 



Measurements. — The filaments of Dreissena have a short 

 diameter of only -j-^,j^th of an inch as compared with ^i i.th 

 inch in Anodon. The epithelial cells are of about the same size 

 in the two genera ; the larger latero-frontal cells of Dreissena 

 being somewhat shorter relatively to breadth as compared 

 with Anodon. 



The object of the prefsent paper has been to explain the 

 structures illustrated in the accompanying plates, on the 

 hypothesis that the Lamellibranch gill is not primitively a 

 membranous plate, but a series of independent filaments. 

 This hypothesis, availing itself of the principle of con- 

 crescence — of the existence of which we have independent 

 proof — does satisfactorily explain all the details of structure 

 which I have submitted to the reader's notice. If I shall 

 have proved that — at any rate so far as the four genera. 

 Area, Mytilus, Anodon, and Dreissena are concerned — the 

 filamentar hypothesis is the correct one, I shall feel satisfied. 

 I reserve the discussiou of other types for a future occasion. 



