660 



PROCEEDINGS OF THE ACADEMY OF 



[Dec, 



more careful examination of these bones has led the author to recon- 

 sider these provisional determinations and to come to the conclusion 

 that the elements in question are in no way connected with the head- 

 shield, but are part of the "clavicular" and lateral armoring. 



Fig. 2. — A, Dinichthys halmodeus (Clarke)? Marcellus shale near Syracuse, 

 N. Y. Black lines an- traced from a photograph. Dotted lines represent 

 restored outlines. In the case of the interlateral the dotted lines follow 

 impressions in the matrix. Length of ventral shield about 195 mm. 



B, Coccosteus decipiens Ag. Bestoration of the ventral armor (modified 

 from Traquair). 



C, Coccosteus decipiens Ag. Bestoration of abdominal armor side-view (mod- 

 ified from Traquair). 



A.V.M., antero-ventromedian ; P.V.M., postero-ventromedian; A.Y.L., antero- 

 ventrolateral; P.V.L., postero-ventrolateral; B.D.L., postero-dorsolateral; 

 A.D.L., antero-dorsolateral; I.L., interlateral; A.L., anterolateral; P.L., 

 posterolateral; M.D., dorsomedian. 



Fig. 2 A represents a tracing taken from the photograph of the 

 original specimen. The plates of the ventral shield are all capable 

 of certain identification, as are also the two antero-dorsolaterals and 

 the postero-dorsolateral. Immediately forward of the ventral shield 

 is a group of bones which at first appear to be hopelessly confused. 



