definite diameters, on wliicli to form speeies, and it is only by com- 

 paring together very large collections of specimens, that we are 

 enabled to determine specific relations. 



Davidson is inclined to acknowledge bnt one species of Streptor- 

 hynchus from the Devonian and Carboniferons of Enrope, and Prof. 

 Hall has also nnited the nnmerous Devonian forms of America 

 under one specific name. But from the descriptions given by these 

 two noted Palaeontologists, and from the specimens at command 

 for comparison, there seems to be as much difference between the 

 Erere forms, and 8. crenistria or C/ieiminc/cnsis, as exists between 

 the latter two species themselves. 



In S. Agassizii the raised lines are iilways very fine, regular, and, 

 almost without exception, arranged closely together. The beak is 

 never much extended, twisted or turned to the side, and the hinge 

 area retains about the same width in all the specimens, never being 

 wide, while altogether the test is never very unsymmetrical. 



It is trne that these characters may seem to be varietal, but they 

 obtain through all the specimens collected at Erere, which amount 

 to over 500 in number, more or less perfectly preserved, all of which 

 have been carefully compared with one another. 



Devonian sandstone of Erere. (Morgan Expeditions 1S70 and '71.) 



[This species, the most common of the Erere fossils, I have ded- 

 icated to my honored teacher. Prof. Agassiz. — 0. F. H.] 



Chonetes Comstockii, Uartt, sp. iiov., Plate IX, figs. 5, 14, 18, 19 and 31. 



Test rather above the medium size, depressed-concavo-coiivex, transverse, 

 with the breadth one and one-half to one and three-fourths the length, and 

 greatest along the hinge which is straight. The cardinal extremities are 

 acutely angular, and, though always defective, are sometimes probably 

 slightly produced, as is shown by the lines of growth on one or two speci- 

 mens. The outline is somewhat sub-quadrate ; the lateral margins, slightly 

 rounded, extend forward, nearly parallel with one another, for about one-half 

 the length of valve, when they bend rapidly round to unite with the anterior 

 margin, the outline of which is more or less gently convex. 



Ventral valve very slightly convex, generally most prominent just posterior 

 to the center, whence it slopes with slight curvature to the front, the curve 

 from the same point to the beak being more rapid. Across the valve, from 

 side to side, the curvature is gentle and more or less regular, the sides towards 

 the cardinal angles being usually more or less, sometimes very much, flattened. 



