336 STUDIES IN EVOLUTION 



tion from the features of the normal adult. On Plate XVII 

 is given a figure of the youngest example found, which has 

 a length and width of 3 mm., while the usual adult is about 

 16 X 16 mm., varying in relative proportions with the in- 

 crease of senile obesity. The change in outline during 

 growth is from sub-circular to sub-triangular, and in earlier 

 stages the ventral fold and sinus are very ill defined. The 

 peculiar triangular exfoliation of the shell on the umbo of 

 the ventral valve is evidently a constant feature in every 

 stage of growth after the shell becomes attached. The 

 nature of this peculiarity was indicated by Billings in the 

 original diagnosis of the genus Eichwaldia (^Ann. Rept. Cana- 

 dian Greol. Survey, 1857-58), and was demonstrated more 

 fully in Dictyonella by Professor Hall, in the Twentieth 

 Report on the Condition of the New York State Cabinet of 

 Natural History (pp. 274-278, 1867). This area is underlaid 

 by an internal shelf or diaphragm attached along its lateral 

 margins, and having fully or rather more than the width of 

 the median sinus. Through the space thus left between the 

 shell and the internal diaphragm, communication is afforded 

 with the outside world. Mr. John Young has called atten- 

 tion to the fact that in D. Capewelli the margins of the 

 external reticulated layer of the shell about the umbonal bare 

 spot are rough and ragged, the superficial hexagonal cells 

 being without finish along these edges, suggesting therefrom 

 that the animal was attached to marine objects by the sub- 

 stance of the shell, and afterward broken away from its 

 attachment. (See Davidson, General Summary, pp. 355, 

 356.) It is true that the anterior edge of this area may be 

 rough and uneven, but the lateral edges appear invariably 

 straight and diverge at an essentially constant angle. The 

 latter represent the lines of attachment of the internal plate 

 to the interior of the valve, and if the shell has been broken 

 in detachment from foreign bodies, the fracture in these 

 directions has been guided by these lines, but on the unsup- 

 ported anterior margin it has been rough and irregular. 

 Upon the hinge-line of the ventral valve there exists no 



