SKULL OF THE CROCODILE. 107 



expanded, and smootlily excavated on their inner surface 

 to support tlie sides of the large prosencephalon; they 

 dismiss the great optic nerves by a notch. They show 

 the same tendency to a retrograde change of position as 

 the neighboring neurapophyses, 6 ; for though they sup- 

 port a greater proportion of their proper spine, 11, they 

 also support part of the parietal spine, 7, and rest, in 

 part, below upon the parietal centrum, 5: the neurapo- 

 physes, 10, 10, are called " orbitosphenoids." The neural 

 spine, 11, of the frontal vertebra retains its normal cha- 

 racter as a single symmetrical bone, like the parietal 

 spine which it partly overlaps; it also completes the 

 neural arch of its own segment, but is remarkably ex- 

 tended longitudinally forwards, where it is much thick- 

 ened, and assists in forming the cavities for the eyeballs ; 

 it is called the " frontal" bone. 



In contemplating in the skull itself, or such side view 

 as is given in Fig. 9, p. 22, of my work on the Archetype 

 Skeleton^ the relative position of the frontal, 11, to the 

 parietal, 7, and of this to the superoccipital, 3, which is 

 overlapped by the parietal, just as itself overlaps the flat- 

 tened spine of the atlas, we gain a conviction which can- 

 not be shaken by any difference in their mode of ossifi- 

 cation, by their median bipartition, or by their extreme 

 expansion in other animals, that the above-named single, 

 median, imbricated bones, each completing its neural 

 arch, and permanently distinct from the piers of such 

 arch, must repeat the same element in those successive 

 arches — in other words, must be " homotypes," or seri- 

 ally homologous. In like manner the serial homology 

 of those piers, called " neurapophyses," viz : the laminae 

 of the atlas, the exoccipitals, the alisphenoids, and the 

 orbitosphenoids, is equally unmistakable. Nor can we 



