1910.] NATURAL SCIENCES OF PHILADELPHIA. 661 



If, however, we examine Traquair's 10 restoration of Coccostcus decipiens 

 Ag\, we find that in this region occur two well-defined transverse 

 elements which are terminated at their outer extremities by processes 

 which are directed backward and upward. In the specimen here 

 considered it can be proved b} r a careful examination of impressions 

 on the matrix that the bone originally believed to be an antero- 

 superognathal is not a short compact element, but is in reality a frag- 

 ment once connected with a long transverse bone which borders the 

 anterior margin of the left antero-ventrolateral. The part believed 

 to be a gnathal then becomes the upwardly and backwardly directed 

 fork of the interlateral shown in Traquair's restoration of Coccosteus 

 decipiens. The evidence for this interpretation is twofold — (1) the 

 shape of the bone corresponds to the Coccostean interlateral and (2) 

 the bone is in exactly the position where we might expect to find such 

 an element. 



In Coccosteus decipiens this interlateral plate has a rodlike ventral 

 portion and a lateral .portion with its two processes, one directed 

 dorsally, the other posteriorly. Between these two processes fits 

 the lowest plate of the lateral armoring, namely, the anterolateral. 

 This anterolateral articulates behind with the small posterolateral 

 and above it overlaps the antero-dorsolateral. It therefore serves 

 to link the ventral to the dorsal shield. 



If, now, we turn to the higher species of Dinichthys, such as D. inter- 

 medins Newb., we find that apparently the Coccostean interlateral, 

 the Coccostean anterolateral and perhaps even the Coccostean postero- 

 lateral have fused together into one large bone, the so-called "clavicu- 

 lar." We say apparently, for though some specimens exhibit the 

 condition shown in Newberry's classic illustration, 11 others occur with 

 the two components (the lower rodlike portion and the upper platelike 

 portion) separate. Eastman 12 has homologized the broad portion of 

 the "clavicular" with the Coccostean anterolateral and its lower 

 portion with the Coccostean interlateral. 



Though not wushing to express too hasty an opinion, the author 

 believes that the restoration of Dinichthys would present fewer diffi- 

 culties if we could regard the lower (interlateral) portion of the " clavic- 

 ular" as an element distinct from the upper (anterolateral) portion. 



10 R. H. Traquair, "On the Structure of Coccosteus decipiens, Agassiz," 

 Ann. and Mag. Nat. Hist., 6th ser., vol. V, 1S90, p. 125. 



11 U. S. G. S., Monographs XVI, pi. XLVIII. See also Hussakof, in Mem. 

 Am. Mus. Nat. Hist., vol. IX, part III, p. 133. 



12 New York State Museum Memoir 10, p. 119. 



