238 THOMAS DWIGHT- ON 



vertebrae of the lumbar region are modifications of thoracic vertebrae, and that both 

 phylogenetically and ontogenetically the ilium travels upward on the spine, so that 

 originally the sacrum is composed of a group of vertebrae further from the head than it 

 is later. As it advances to include vertebrae above, it releases j';/aH passu others below 

 it. If the ilium does not advance in any particular instance so far as usual, the spine is 

 archaic; if it advances too far, the spine is of the future. In the same way the appear- 

 ance of a rib on the 20th vertebra, which is particularly likely to happen when the 

 sacrum does not include the 25th, is a return to a previous condition, while the non- 

 appearance or imperfect development of the costal element of the 19th is a foreshadowing 

 of what is to come. In his first paper Rosenberg put the cervical region aside ; which 

 must be borne in mind in considering his rule for the comparison of the columns of pri- 

 mates, namely that the last vertebra of each region at a certain stage of development is 

 at a higher stage transformed into the first vertebra of the region below it. Evidently 

 this cannot apply to the cervical region, for otherwise the 7th cervical would change into 

 a first thoracic. In his last work Rosenberg has stated the matter more comprehensively. 

 He considers (as he did before) the middle part of the thoracic region of the spine as the 

 most primitive, and holds that the division of the column into regions (not orAy in man, 

 but also in other mammals, and indeed in other vertebrates,) is the work of two chief 

 factors which simultaneously, and to a certain extent in the same manner, exert an influ- 

 ence on the column in opposite directions. " These factors are both processes of trans- 

 formation, of which the one, acting on the smaller proximal (i. e., cervical) division of the 

 column, works distally ; while the other, affecting the greater distal (i. e., lumbar, sacral 

 and caudal divisions) works proximally." In other words, the cervical region on the one 

 side and the lumbar on the other tend to absorb into themselves the thoracic vertebrae 

 nearest to them. Such a change at either one end or the other is to be considered pro- 

 gressive, and a change in the opposite direction retrogressive. Rosenberg recognizes that 

 if these two opposed factors stood to one another in a fixed relation it would be possible 

 from the condition of one end of the thoracic region to predicate that of the other ; but he 

 admits that, so far from this being the case, the most various combinations occur. In his 

 first paper Rosenberg enunciated the principle of concomitant variations, as for instance 

 of the 20th vertebra frequently bearing a rib when the 25th is free, which is undoubtedly 

 true, though not in all cases; for which it seems to me he has hardly received the credit 

 he deserves. He never, so far as I know, has given any explanation of the cause of this 

 phenomenon, beyond tacitly implying that when the variation was in one direction it 

 was to be accounted for by reversion and in the other by progressive development. 



I hope to show that there is a principle to account for these changes and that, looked 

 at from another point of view, the contradictory nature of the changes at the two ends 



