PALEONTOLOGY. 9 1 



COLUMNARIA STELLATA, Hall. 

 Synon., Favistella stellata, Hall. 



Convex masses of similar structure with the former species ; the 

 tubes are somewhat less unequal in the same specimens, varying 

 in size between three and six millimeters. Radial lamellae from 

 twenty to thirty, alternately larger and smaller, the larger ones 

 extending to the centre. Diaphragms flat, closely set, intersected 

 by the radial lamellae, which only in the peripheral circumference 

 form continuous vertical leaves. The crests continued to the centre 

 of the diaphragms are merely superficial. Found abundantly on 

 the north side of Drummond's Island, in the shales of the Hudson 

 River group. The equivalent exposures in Bay de Noquets, oppo- 

 site Escanaba, do not contain any. It is likewise found in the up- 

 per part of the Hudson River, or Cincinnati group, near Madison, 

 Indiana, and in other localities along the Ohio River. 



During the progress of the survey, I found specimens which I 

 consider as identical with this species in the Niagara group of 

 Point Detour, in which formation it was not known to occur. 



Plate XXXIV. — Fig. 3 gives a surface view of a specimen from 

 Drummond's Island (Hudson River group). 



Plate XXVIII. — Fig. i is a silicified specimen from the Niagara 

 group of Point Detour. 



COLUMNARIA HERZERI, N. Sp. 



Colonies of tubes, partially in close contiguity, of polygonal form, 

 and intimately united with their walls ; partially free, circular, 

 laterally joining into chain-like rows, not unlike Halysites, or open- 

 ing singly on the surface. Diameter of tubes three millimeters. 

 Structure otherwise entirely corresponding with the associated 

 form, Columnaria stellata. Is described by Nicholson under the 

 name Favistella calycina. 



The specimens were found by Rev. H. Herzer, of Louisville, in 

 the Cincinnati group of Kentucky. It may only be a modification 

 in the growth of the usual form C. stellata. Not figured. The 

 specimens are too imperfect for photographic delineation. 



