176 UPPER CRETACEOUS TELEOSTS 
does not appear to represent as efficient a modification for predatory feeding as does 
the more mobile and protrusile acanthopterygian jaw. The rise of the acanthoptery- 
gians probably accounted for the disappearance of the ichthyotringoids in the Upper 
Cretaceous. The more generalized Apateodus, however, existed until the extreme 
end of the Cretaceous and outlasted its more specialized descendants. The only 
recent teleosts which possess a long snout and an unelongated body seem to be the 
Xiphiidae, possibly strengthening the argument that this rostral elongation 
represents a less successful way of life. 
The Ichthyotringoidei must have arisen as a very early offshoot from a basal 
salmoniform stock, a stock not too far removed from the elopiforms. 
(B) Suborder CIMOLICHTHYOIDEI 
This new suborder has been erected to contain several quite closely related genera, 
all of which are difficult to link with any Recent forms or with any so far described 
fossil forms. The two families Cimolichthyidae and Dercetidae show great similari- 
ties in overall skull and body structure, in particular the rostral region and the 
reduction of the body scaling to two or three isolated scute rows on each flank (Text- 
fig. 82). Both families occur in the Cenomanian and extend through much of the 
Upper Cretaceous. Many incomplete and fragmentary specimens of large Dercetis 
species from the English Chalk (Woodward, 1903), e.g. Dercetis maximus and Dercetis 
latiscutatus, are closely related to Cimolichthys levesiensis from the same horizons. 
These large species of Dercetis have not been considered in the systematic account 
since they are represented by very imperfect material. 
The structurally interesting features shown by the cimolichthyids and dercetids 
are mainly concerned with the extension of the head or body or both. These 
elongations of the head and body have occurred to a greater or lesser extent in all of 
the genera. Czmolichthys shows slight snout and body elongation ; Dercetis has a 
variably extended snout with a long body (Text-figs. 23-26) ; Rhynchodercetis has 
an enormously elongated snout and a very long body (Text-figs. 28 and 29), and 
Pelargorhynchus has some snout elongation but much body lengthening. 
Snout elongation is always brought about by the extension of the preorbital region 
and includes the vomer, mesethmoid, palatines, premaxillae, maxillae and also the 
frontals to a marked extent where they extend forwards between the postero- 
lateral arms of the mesethmoid (Text-figs. 16, 17 and 31). Rhynchodercetis differs 
very slightly from the others in that the anterior rostral region is formed solely by 
the premaxillae and there is a tendency for the dentary to be shorter than the 
rostrum (Text-fig. 29), the latter being a feature seen in several other unrelated 
groups, e.g. Aspidorhynchus and Belonostomus (Gardiner, 1960 : 362) among fossil 
halecostomes, and Jstiophorus among Recent fish. There is obviously some correla- 
tion between the mode of life, particularly trophic habits, and the greater extension 
of the upper jaw, but any obvious correlation is difficult to envisage. Possibly the 
rostrum acted as a cutwater, reducing the amount of turbulence during forward 
motion and thus enabling the fish to approach its prey undetected. All the other 
