VOL. 



!"'] PROCEEDINGS OF THE NATIONAL MUSEUM. 235 



of the hinge structure is discernible. It is highly probable that the 

 evolution of hinge teeth closely followed the differentiation of the 

 Pelecypod class. 



The first bivalves are all minute, as far as known, when compared 

 with a majority of their descendants. They are usually Dimyarian, as 

 I assume the architype to have been. It is highly probable that they 

 possessed a developed foot and that their gills were either lamellifonn 

 on either side of an arterial stem, as in Nucula, Solcnoniya, and many 

 Gastropods, or filiform, as in Dimya and certain Pectens. The siphons 

 were probably little developed and the lobes of the mantle rather widely 

 separated, or perhaps entirely free. 



As long as the shell remained small and subglobular the ligament 

 short and wholly internal, the imperfect character of the hinge was of 

 less importance. With the essential difference between the anterior 

 and the posterior halves of the animal, and especially with any material 

 increase in the magnitude of the adult, more or less discrepancy would 

 develop itself between the two ends of the shell, the subglobular form 

 would disappear, and certain other consequences would follow. Either 

 the ligament must increase with the size of the shell and become longer 

 or its power would become inadequate for the proper performance of 

 its functions. 



Here I will turn aside for a moment from the direct line of argument 

 to describe the mechanical relations of ligament and shell, a proper 

 understanding' of which is very necessary to the comprehension of the 

 whole question. 



With a wholly external ligament the operation of the valves is that 

 of two appendages to the free ends of a C-shaped spring. The action 

 of the muscles in pulling the valves together includes the bringing 

 nearer to each other of the two extremities of the ligament, which the 

 latter by its elasticity resists; consequently the operation of the liga- 

 ment is in the direction of opening the valves to a certain distance. 

 Beyond this distance the separation of the valves tends to compress the 

 ligament, which again resists, and therefore beyond the normal distance 

 of separation the action of the ligament tends to prevent the valves 

 from opening. This very simple matter may be observed by any one 

 who* will examine an ordinary clam with the ligament in fresh condi- 

 tion and whose adductor muscles have been severed, 



When the ligament, in harmony with the elongation of the cardinal 

 margin, becomes elongated it must be either straight or angulated. 

 For obvious reasons a ligament forming a curve or the arc of a circle is 

 mechanically impossible. This any one can prove to their own satisfac- 

 tion by putting two light wooden saucers edge to edge, convexity out- 

 ward, and attaching a leather or paper ligament by cement. A curved 

 ligament, when the valves open, will tear or break at once either itself 

 or the edge to which it is fixed. In other words, the axis of motion of 

 the hinge must be a straight line. If any part of the ligament diverges 



