INSTITUTE OF SCIENCE, PHILADELPHIA. 487 



chial marsupium, the original prodissoconch is followed by a rounded-tri- 

 angular bivalve shell. The young in this stage, termed a glochidium, may un- 

 dergo considerable developmental change, and remain encysted for a lengthy 

 period on some suitable host, without material change in size or form of its 

 valves. In perfectly preserved specimens the glochidial valves may be dis- 

 tinctly seen at the apex of the umbones. The base of the triangle forms the 

 line of the hinge, the rounded apex opposite corresponds to the ventral border, 

 within it is furnished with a hook or claw-like, sometimes denticulate, process, 

 by means of which it attaches itself to the gills or fins of fishes, where, by 

 proliferation from the epithelium of the host, the glochidium becomes encys- 

 ted. This stage of development was, until recently, supposed to be peculiar 

 to the Naiades, but the writer has recently discovered that a genus of marine 

 pelecypods, probably allied to Avicula and named by Carpenter Philobrya *, 

 also passes through a Glochidium stage, thus adding another link to those 

 which connect the Aviculidce with the Naiades. That this particular stage of 

 growth is due to the habit of encysting itself rather than to the incubation by the 

 parent, is evident from the fact that many other bivalves which protect their 

 young in this manner (Lepton, Parastarte, Thecalia, Milneria, Ostrea) neverthe- 

 less exhibit only a normal prodissoconch. 



The surface of the glochidial valve is conspicuously punctate, due to a 

 poriferous structure perhaps connected with the larval respiration. 



I have already shown f that it is absolutely required by the laws of 

 mechanics that the axis of revolution must be a right line and that, therefore, 

 that part of the hinge-line functioning in connection with the ligament in the 

 opening and closing of the valves must be straight. If, as in some forms 

 superficially appears, the ligament seems to be curved, or to extend over a part 

 of the hinge-line which is curved, an examination will show that either that 

 portion of the ligament, which departs from a straight line, is not functional, or 

 that it possesses sufficient elasticity to fall into line when exercising its func- 

 tions. 



It is obvious, also, that the longer the line covered by a functional liga- 

 ment, the more rigidly will the opening and shutting of the valves be con- 

 trolled ; and, conversely, that the shorter and higher up the ligament the more 

 easily a rocking motion of the valves (which would prevent their exact closing, 

 margin upon margin) may be brought about. To obtain a long straight line 

 on the border of a rounded body the hinge-line must descend, or the outline 

 of the body must be modified. The former, for obvious reasons, is the easier 

 process and through it, as promoted by natural selection, the hinge-line, from 

 being a minute segment of the arc described by the dorsal margin is lowered 

 until it forms a chord truncating part of the arch between the beaks and form- 



* Originally Bryophila, which proved to be already in use. 



f Am. Joura. Science, xxjtviii, Dec. 1889, pp. 445-462, also Proc. U. S. Nat. Mus., xll, pp. 232-248, 1889. 



