343 



than the others. He cites Murie to the effect that the reduction of 

 the cervical vertebrae to six in the Manatus arises from anfalraost 

 complete regression of the 3d. Putting aside the question of inter- 

 calation, the idea that this is a critical point receives additional con- 

 firmation from the four cases in this series. 



Irregular Segmentation. There are cases in this series 

 both of the addition and of the want of a vertebra above the sacrum, 

 in which the regions are well defined and otherwise practically normal. 

 Two striking examples are 764 and 478. The formula of the former is 

 C. 7, T. 13, L. 5, S. 5, C. probably 5; that of the latter is C. 7, T. 11, 

 L. 5, S and C. 9. In both, the cervical, lumbar and sacral regions 

 are very nearly typical. The fact that the 1st coccygeal is fused with 

 the sacrum in both is interesting as showing that it is probably quite 

 an indifferent occurrence. There seems to be no other way of account- 

 ing for these spines, and for others nearly as good, but by admitting 

 a departure from the usual segmentation. 



Concomitant Variations. Having discussed most of the 

 variations, considered separately, we have now to consider the relation 

 of a variation at one part of a column to a variation at another part. 



This brings us to the central point in the discussion of Rosen- 

 beeg's theory. It has been remarked that he has admitted that it 

 would be natural to expect some relation between the condition of 

 the spine at one end of the thorax and that at the other, but that 

 such a relation is not to be found. Thus if the undeveloped condition 

 of the 1st thoracic rib is a step towards the future, it would be 

 reasonable to expect in the same spine a corresponding advance below 

 the thorax. Conversely, if there is an archaic condition below the 

 thorax there should be an analogous condition above it. While there 

 are cases that fulfil these conditions, they are quite lost in the multi- 

 tude which do not, and which even present contradictory conditions 

 at the opposite ends of the spine, being retrogressive at one end and 

 progressive at the other. 



Let us take first the ten complete spines of this series with 

 cervical ribs (including H-3). In no case is there an increase in number 

 of the praesacrals, which should be expected if this were an archaic 

 manifestation. In half the cases the number of praesacrals is 

 normal, in three the last is partially sacralized, and in two, the 

 most perfect cases (267 and 202), there is a praesacral too few. 

 Turning to the literature of cervical ribs, as given by Gruber to 

 1869 and continued by Pilling to 1894, I find but a single case 

 of unmistakable cervical rib with an increase in the number of 



