524 



naturally be supposed that though some bones are so rarely 

 developed as not to seem worth including, and though some that 

 are included are very apt to be absent , yet that certain others are 

 invariable : forming, as it were, the basis of the ideal type. Let 

 us see whether the facts bear out this supposition. In his "summary 

 of modifications of corporal vertebra? " (p. 96), Professor Owen 

 aays " The haemal spine is much less constant as to its existence, 

 and is subject to a much greater range of variety, when present, 

 than its vertical homotype above, which completes the neural arch." 

 Again he says " The hcemapophyses, as osseous elements of a 

 vertebra, are less constant than the pleurapophyses." And again 

 " The pleurapophyses are less constant elements than the neurapo- 

 physes." And again " Amongst air-breathing vertebrates the 

 pleurapophyses of the trunk segments are present only in those species 

 in which the septum of the heart's ventricle is complete and imper- 

 forate, and here they are exogenous and confined to the cervical 

 and anterior thoracic vertebrae." And once more, both the.newra- 

 pophyses and the neural spine " are absent under both histological 

 conditions, at the end of the tail in most air-breathing vertebrates, 

 where the segments are reduced to their central elements." That 

 is to say, of all the peripheral elements of the " ideal typical 

 vertebra," there is not one which is always present. It will be ex- 

 pected, however, that at any rate the centrum is constant : the bone 

 which " forms the axis of the vertebral column, and commonly the 

 central bond of union of the peripheral elements of the vertebra 

 (p. 97), is of course an invariable element. No : not even this is 

 essential. 



"The centrums do not pass beyond the primitive stage of the notochord 

 (undivided column) in the existing lepidosiren, and they retained the like 

 rudimental state in every fish whose remains have been found in strata 

 earlier than the permian sera in Geology, though the number of vertebrae is 

 frequently indicated in Devonian and Silurian ichthyolites by the fossilized 

 neur- and heemapophyses and their spines " (p. 96). 



Indeed, Professor Owen himself remarks that "the nenrapo- 

 physes are more constant as osseous or cartilaginous elements of the 

 vertebra? than the centrums " (p. 97). Thus, then, it appears that 

 the several elements included in the " ideal typical vertebra '' have 

 various degrees of constancy, and that no one of them is essential. 

 There is no one part of a vertebra which invariably answers to its 

 exemplar in the pattern-group. How does this fact consist with the 

 hypothesis ? If the Creator saw fit to make the vertebrate skeleton 

 out of a series of segments, all formed on essentially the same model 

 if, for the maintenance of the type, one of these bony segments is 

 in many cases formed out of a coalesced group of pieces, where, as 

 Professor Owen argues, a single piece would have served as well or 

 better ; then we ought to find this typical repetition of parts uni- 



