240 



THE AMERICAN NATURALIST [Vol. XLVII 



ORDOVICIAN ( ?) FISH REMAINS IN COLORADO 

 During the past year Mr. P. G. Worcester, of the Colorado 

 Geological Survey, has been investigating certain strata near 

 Ohio City, Colorado, supposed to be of Ordovician age. The 

 particular horizon under discussion contains Beccptaculifcs 

 oweni Hall (southeast of Fairview Mt.), Haly sites catenulatus 

 (L.) (basin east of north end of Fossil Ridge), Platystro- 

 phia (?) sp. (Fossil Ridge), and Heliolites (?) sp. with Haly- 

 sites catenulatus at head of Alder Creek, west of Fossil Ridge. 

 The Heliolites (?) is the same as that in the Canon City Ordovi- 

 cian. These fossils were identified by Professor J. Henderson, 

 and so far as it is possible to determine from them, the rocks 

 should certainly be Ordovician. The first Devonian fossils were 

 found about 100 feet above this horizon. 



However, closely associated with the invertebrates cited, and 

 certainly of the same age, are rather numerous fragmentary 

 remains of fishes. These may be briefly described as follows : 



1. A fragment of a plate exhibiting fine grooves with deep 

 pits; resembling, so far as it goes, the plate of Coccosteus dis- 

 jectus from the Old Red Sandstone, figured by A. S. Woodward, 

 Cat. Fossil Fishes Brit. Mus., Part II, pi. VIII, fig. 1. The 

 structure is also nearly identical with that of Astraspis deside- 

 rata, from the Ordovician of Canon City, as figured by Walcott, 

 Bull. Geol. Soc. Amer., Vol. 3 (1891), pi. 3, f. 7. Some of the 

 other figures of Asfraspis mi.uht well belong to Coeeostean fishes. 



2. A large fragment, having a diameter of over 30 mm., is 

 covered with irregular obtuse vermiform ridges, and is exactly 

 like the opercular plate of Rhizodus ornatus (Woodward, t. 

 c, pi. xii, f. 5). so far as the sculpture goes. This particular 

 "species is lower Carboniferous, but Khizodontid fishes also occur 

 in the Devonian. 



3. Numerous fragments of striated spines, some short, conical 

 and straight ; others more slender and curved. These appear to 

 exactly correspond, so far as they go, with the spines of Dipla- 

 canthus, from the lower Old Red Sandstone. One of the sup- 

 posedly Coccosteoid plates. 5 mm. thick, with the surface finely 

 striate, with punctate more or less brandling stria? or grooves, 

 occurs in the same piece of rock as a supposed Dipla-canthu* 

 spine, the two almost touching. 



According to the available evidence, we seem therefore to 



