T< l889 n '] PROCEEDINGS OF THE NATIONAL MUSEUM. 237 



With the lengthening and angulation of the cardinal margin the lig- 

 ament gradually shifted to a point where it became posterior to the 

 beaks. Perhaps it would be better to say that the portion in front of 

 the beaks either became segmented off as a cartilage, or became obso- 

 lete and vanished, while the portion on tbe posterior side gradually 

 elongated, as the elongation of the posterior hinge-margin rendered a 

 longer ligameut more useful. It has already been pointed out that a 

 curved ligament would involve stresses leading to its own destruction. 

 The curvature of the cardinal margin, now the common property of a 

 vast majority of bivalves, was inevitable with increase in size and 

 asymmetrical development of the anterior and posterior ends of the body. 

 Consequently, that the ligament should be shifted was a mechanical 

 necessity unless the evolution of the group was to be confined within 

 extremely narrow limits as regards hinge characters. 



The infolding of the ligament, and the development of a cartilage and 

 its supports, would be especially likely to occur in forms with a thin 

 edentulous hinge, where the least shifting would be necessary (Soleno- 

 mya, Anatina), rather than in those with a broad, flat hinge margin. 

 In harmony with this proposition, we find the archaic forms, with inter- 

 nal cartilage, have generally a narrow edentulous cardinal border, the 

 exceptions belonging to the more recently specialized types (Mactra, 

 Spondylus), while the groups without an internal cartilage contain the 

 broadest and heaviest types of hinge (Pectunculus, Veneridce). 



The infolding of a cartilage which arose by longitudinal segmenta- 

 tion would leave a line of weakness in the arch of the umbones. In 

 thin shells with strong adductors there would be a tendency to fracture 

 here. This singular feature has been perpetuated in what may be 

 termed the normal umboual fissure of Solenomya, Periplonta, and simi- 

 lar forms. Traces of it are evident in TJiracia, while the un fractured 

 suture itself is visible in Isocardia, Pachyrisina, aud Pecckiolia, 



In the thin-shelled Cuspidariidw a special buttress is often developed 

 to support the shell at this weak point. In the Isocardiidce an inde- 

 pendent cartilage was possibly never developed, but the infolding of 

 the anterior part of the ligament went far enough to leave permanent 

 traces on the shell. That it did not result in a cartilage, if this was 

 the case, may possibly be due to the fact that, owing to the great 

 size and spiral character of the umbones, the anterior part of the liga- 

 ment was turned up instead of downward, and therefore did not tend 

 to shift toward the interior. If it is not clear how the thickening or 

 vertical extension of the ligament below the cardinal axis should cause 

 its separation iuto two parts', I need only recall the familiar experience 

 of every one in breaking off a wire or piece of tin by bending it back- 

 ward and forward on the line of the desired fracture. The mechanical 

 principles and results iu the two cases are precisely similar. 



When finally developed iu the same individual the ligament and carti- 

 lage work in identically the same manner but in different directions. 



