GILLS OF LAMELLIBRVNCH MOLLUSCA. 61 



transverse fibrous junctions coincides with the external 

 oi)ening or stoma of one water-passage (PL VI, fig. 14 w). 



Thus it becomes probable that in the process of the evolu- 

 tion of the Anodon's gill from such a simpler form as that 

 of My tikis, the gill-filaments themselves first of all developed 

 the horizontal transverse attachments which now present 

 themselves as complete fibrous continuations of the con- 

 nective substance of the gill-filament's wall. Then the deep 

 or internal face (that forming the interlamellar surface) 

 of the gill-filaments ceased to develope chitin and produced 

 loose bulging growths of lacunar tissue, which extended not 

 only along the vertical line of the filament, but across the 

 transverse junctions, uniting to neighbouring outgrowths by- 

 concrescence. The concrescence has, however, always failed 

 to block up a portion of the interfilamentar water-way, 

 namely, a cylindrical passage running more or less obliquely 

 upwards and backwards from each rectangular division of 

 the inter-filamental space. 



Thus the undulating horizontal arrangement of the inter- 

 nal stomata of the water-passages is explained. 



It should be especially pointed out that sometimes, as an 

 exception here and there, we find that concrescence has 

 actually overstepped the limits thus laid down, and the water- 

 passage corresponding to a rectangular interfilamentar area 

 has disappeared, having been obliterated by the growing- 

 together of its own walls. 



The fibrous interfilamentar junctions consist of horizontal 

 and obliquely crossed fibres (PI, VI, fig. 14, ^^^and h), which I 

 cannot agree with Posner in considering to be muscular. The 

 relation of these fibres to the substance of the gill-filaments 

 will be best understood by reference to a series of transverse 

 sections of the filaments. 



In Plate V, fig. 12, a transverse section of a filament is 

 represented taken at a part where the sub-filamentar out- 

 growth or bulging of the deep surface is as small as can be found. 



Beneath the epithelium to the front part are seen two 

 semilunar masses [ch.) : these are the greatly thickend cbi- 

 tinous Avails of the filaments, but onlv correspond to one 

 half of that wall as developed in Area or Mytilus. Small nuclei 

 (protoplasmic residue) are seen in the chitinous substance. 



Passing to the lower part of the section we fitid there is no 

 further chitinous deposit, but a loose trabecular tissue (lac.) 

 amongt the mesh-work of which float blood-corpuscles (be). 

 The semilunar chitinous masses show in the section each 

 a smaller more highly-refracting and yellow-coloured body, 

 which we find on comparison with the longitudinal section 



