26 THE CERATOPSIA 



between the fontanelle and the supratemporal fossae which at one time were confluent, and Hay was 

 partially correct. At all events, the fontanelle and its canals represent a vestigial structure the 

 function of which was apparently lost. Compare this area in Styracosaurus, PI. VIII. 



Another characteristic feature is the series of separate ossifications, one, the epijugal, at the 

 lower end of either jugal, and a series, to which the name epoccipitals is applied, around the periphery 

 of the crest. The Protoceratopsidae possessed the former, but apparently not the latter. In the first- 

 appearing genera of the Ceratopsidae the epoccipitals are also present, and while seemingly absent 

 in some later forms I seriously doubt whether they actually were missing during the lifetime of the 

 animal, except perhaps in Torosaurus. 



Of horns, the Protoceratopsidae possess none at all, for in the many skulls of Protoceratops 

 itself there is not the slightest indication of their presence. Leptoceratops, on the other hand, 

 while agreeing with Protoceratops in all essential details, except in the crest, is based upon so incom- 

 plete a skull that the areas where brow horns would naturally arise are missing, but the nasals are 

 present and show no trace of a nasal horn. The inference is that the two genera agreed in hornless- 

 ness as in other ways. In the Ceratopsidae, horns are a very characteristic feature and are invariably 

 present. The Belly River and Judith River genera had already developed the nasal horn; in fact, 

 Styracosaurus has the largest one on record, with Monoclonius or Centrosaurus a close second. 

 That of Chasmosaurus, however, is much smaller. Almost the entire progressive development of 

 the nasal horn, therefore, lies within the unrecorded part of ceratopsian evolution. Of the brow 

 horns, this is not true, for while Chasmosaurus kaiseni and Ceratops montanus possess fairly large 

 ones, they are as a rule in an extremely rudimentary state in most species of each of the several 

 genera which appear at the beginning of our record, especially in the Belly River series. This com- 

 pletes the tale of the missing chapters. 



The actual recorded evolution, aside from the increase in size together with the mechanical 

 adjustments of the skeleton necessary to carry the increasing weight, lies mainly in the head and 

 more particularly in the horns and crest. As to size, the recorded increase in the over-all length is 

 not very great. The length of the Yale Monoclonius (Centrosaurus) is 17 feet 4 inches, measured 

 on the curve, and of the Washington Triceratops, 23 feet 6 inches; the larger Chasmosaurus at 

 Ottawa is about 1 6 feet long, while Torosaurus, another long-crested form, probably did not exceed 

 21 feet. Of the skull alone, that of Centrosaurus is 4 feet 6 inches, Triceratops 8 feet 2 inches, 

 Chasmosaurus about 5 feet, and Torosaurus gladius, as restored, 8 feet 10 inches (although the 

 muzzle, which in the restoration looks disproportionally short, may have brought the actual length 

 to over 9 feet). The living weight of a horned dinosaur has never been accurately estimated, but 

 the percentage of increase can be roughly calculated, that of Triceratops over Centrosaurus being 

 about one third. 



The nasal horn seems to have reached its maximum size by Belly River time, for no subsequent 

 short squamosal form, such as Triceratops, has a proportionately longer horn than Centrosaurus, and 

 in several species it is much reduced, even to elimination. In this phylum, therefore, the recorded 

 evolution of this horn is entirely retrogressive. In the long squamosal phylum, however, the nasal 

 horn of Chasmosaurus is not as large as that of its Fruitland successor Pentaceratops, but in 

 Torosaurus the horn has diminished so that in this phylum — if we are correct in assuming this 

 generic sequence — the trend is first progressive and then retrogressive, the culmination occurring 

 at a later date than in the other phylum. 



Another interesting feature in connection with the nasal horn is its apparent shifting forward 

 in evolution, and here the rate of movement is approximately equal in both phyla, for in Belly River 

 genera it lies above the posterior limit of the nares, while in those from the Lance, it is over the 

 anterior limit; the position of those of the Edmonton Anchiceratops and the Fruitland Pentaceratops 

 is more or less intermediate. This shifting forward of the nasal horn results in a lengthening of the 

 face, especially in the long-crested phylum. 



The brow horns are always progressive in their evolution, regardless of phyla. In Belly River 

 time they are incipient, sometimes a mere rugosity above the orbit, again a short pyramidal horn 

 seemingly, sometimes at least, originating in a separate ossification which may be lost in macera- 



