428 



BULLETIN OF THE UNITED STATES FISH COMMISSION. 



of this membrane now grows and turns upward. There is no mechanical hciuUiKj of 

 the filaments, but a {/roivth upward of their fased ends. After the outer edge of this 

 membraue formed by tlie fused ends of the fihiments lias grown upward for some dis- 

 tance, the inner portion — that first formed — divides so as to mark out the ascending 

 filaments as continuations of the descending ones. Tt is possible that this fusion of 

 the ends of the descending iilamrnts may not take jjlaie in the developing gills of 

 forms like Arca^ where in the adnlt the contiguous filaments are nowhere fused Avith 

 one another. 



Pelseneer (No. 17) now regards Nucula as the most ])riniitive form. He believes 

 A)iomia aiul also the Arcidce to be directly descended from forms with plate gills. As 

 I have said, these relations may be the true ones, and yet it seems difiicult to exi>lain 

 the structure of the gills of the former in terms of those of the latter. One might 

 much more easily suppose, a priori, that tiie latter had developed from the filamentous 

 tyi)e, perhaps by the degeneration of the ascending portion of its filament. 



The point which I wish to make is shown in the accompanying cuts. Fig. 1 

 represents two gill plates of one side of the body oi Xuciila. The shaded portion rep- 

 resents the chitinous ventral edges. Tlie emls of the ])lates here point doAvnward and 

 outward. Fig. 2 represents Pelseneer's hypothetical type, in which the plates are 

 much elongated ventrally and have become more like filaments. In such a series 

 Pelseneer could not place Soloiomya or Yohlia, but had to throw tbem out l)ecause 

 their plates, instead of extending their outer ends downwai'd, extende*! them in quite 

 the opposite direction. 



Now come the sim])lest filamentous gills, those of Area, Fig. 3. (T have figured 

 the gills of Area 2)e,vata.) In this the ascending filanu'ut (a/l) is fully formed, and the 

 diiference between this condition and that of liiieiil a is seen to l)e very grent, notwith- 

 standing this hjq^othetical form. 



Fig. 1. 



Fig. 



It is possible that the ascending limb of the filament is a new structure which 

 has suddenly developed in those forms most closely connected with the forms with 

 plate gills ; if it is, however, merely a continuation outward of the descending filament, 

 it seems as if we ought to regard the gill i>late of Kueula as being homologous, not alone 

 to the descending lind) of a filament, as Mitsukuri has done, but to both descending 

 and ascending limbs. 



