68 NILS HJ. ODHNER, STUDIES ON RECENT CHAMIDAE. 



the dorsal afferent vessels must be situated immediately beside each other in op- 

 posite legs of every filament. When the next step towards higher development is 

 taken by the coalescence of the reflected and the direct legs of each single filament, 

 this junction must be first effected between the shanks of these real afferent vessels. 

 Because the venous capillaries are directly exposed to the blood pressure, while the 

 arterial ones are exposed only indirectly, the result of the conditions mentioned is 

 that such a filament will soon assume an exclusively afferent function by having its 

 afferent capillary enlarged and the efferent vessel reduced in the same degree, being 

 retained only in those filaments which are not connected interlamellarly. If, låter 

 on, interfilamentar organic junctions appear between such a principal filament and 

 the adjacent ordinary ones on both sides of it, these too must become afferent. As 

 a compensation for this condition the remaining interjacent filaments situated inter- 

 mediarily between two afferent septa, i. e. in the apex of a plica, are reserved for 

 the action of carrying off the blood stream. In these, on the other hand, the prim- 

 ary afferent vessel is obliterated, so that, when the efferent principal filaments also 

 coalesce, the interlamellar septa become alternately afferent and efferent, as we find 

 them in Chamidae. 



As the posterior demi-branch in the Chamidae does not begin to appear until 

 after the anterior one, a fact that I have observed in small specimens and that one 

 must expect in accordance to the general rule, this demi-branch still remains in an 

 earlier stage of development than the anterior one. In such an argument we find 

 the explanation of the fact that, in the posterior demi-branch, only the afferent 

 vessels have coalesced into interlamellar septa. In Chama jukesi the act of coale- 

 scence seems not to have proceeded so far as in Echinochama, where the coalesced 

 parts of the veins are differentiated as short descending branches and even the ef- 

 ferent vessels show a similar prolongation in the points of reflection. In Chama 

 jukesi, on the contrary, a continuous lacunar formation of the veins and a simple 

 bending up of the arteriae represent an earlier stage of coalescence. 



With regard to the axio-marginal connections in the lower part of the anterior 

 demi-branch in some forms of Chamidae, these are certainly to be considered as erna- 

 nating from a communication that appears early between the marginal and the axial 

 vein, which, during the ontogenetical development, is split in a number of parallel 

 vessels. The conditions that appear in the efferent vessels in Chama jukesi may 

 possibly be explained as remains of an entire lacunar projection from the efferent 

 axial trunk into the filamentar rudiments just in the lower end of the gill. 

 Under the pressure of the local conditions these may probably, so to speak, take 

 a shorter cut in their development by a direct prolongation and a secondary separa- 

 tion throughout of their two arms, instead of a distal reflection and a subsequent 

 perfection of the reflected arm. 



The above explanations of the development of the gills in Chamidae are purely 

 hypothetical and only a close comparative investigation of the ontogeny will give a 

 solid basis for an opinion on this question. But several facts exist that seem to 

 verify this opinion. Thus in many Filibranchia and certain eleutherorhabdic Pseudo- 



