GILLS OF LAMELLIBRANCH MOLLUSCA. 49 



In Mytilus the movement of the blood, such as it is, pro- 

 bably proceeds from the lophophoral attachment down each 

 filament, up its ascending branch, and so into a transverse 

 canal,^ which unites all the free hook-like terminations of 

 the gill-plate ; the transverse or intcrfilamentary canal open- 

 ing at the anterior end of the series of gill-filaments into the 

 base of the foot for the inner of the mantle, for the outer gill- 

 plate near the mouth. On the other hand it is possible that 

 in Mytilus and Area there is no constant movement of the 

 blood in the filaments in any one direction. It should bo 

 possible to ascertain this in young transparent specimens of 

 Area, where, according to the observations of Mr. Lankester, 

 the blood-corpuscles are of about the same size as those of 

 the common Frog of the same shape, and like them, coloured 

 red by Ha?moglobin. 



So far the structure of the Lamellibranch gill as exempli- 

 fied by Mytilus differs little from that of the series of gill- 

 filaments of the other two classes of Mollusca Acephala the 

 Tentaculibranchia and Spirobranchia. 



We now have to point to a structure which occurs in the 

 case of Mytilus, Area, and its allies, which effects a connec- 

 tion between the adjacent gill-filaments. These are small 

 circular prominences on the sides of the gill-filaments carry- 

 ing elongated cilia (see PL IV, figs. 4, 6, 7), which interlock 

 tuith those of the neighbouring filament, and thus form inter- 

 filamentar junctions , or ciliated junctions as I shall call them. 

 These were described by Lacaze Duthiers in the young My- 

 tilus, and by Williams, but they have never been well figured 

 nor has their altogether exceptional character attracted the 

 attention which it deserves. The use of cilia in the manner 

 thus indicated is entirely without a parallel in the animal 

 kingdom. The knobs of interlocking cilia are disposed on 

 the series of gill-filaments along more or less parallel trans- 

 verse (antero-posterior) lines. They are the first indications 

 which we have of that matting and fusing together of the 

 constituent filaments of the Lamellibranch gill by transverse 

 growths, whereby its plate-like character becomes ultimately 

 developed to its full extent. 



Intimately connected with the adhesion of adjacent gill- 

 filaments by the interlocking of cilia is the further highly 

 characteristic process by means of which in the Lamelli- 

 branchia as in other MoUuscan classes, adjacent structures 



^ I have not been able to inquire into the existence of this transverse 

 canal by means of injections. Its existence appears probable from tlie ad- 

 hesion of adjacent hook-like terminations of the filanieuts of the free 

 laniellfE, and from the structure of those tetmiual processes. 



VOL. XVH. NEW SER. D 



