310 THOMAS DWIGHT ON 



extent each half may move iudependently of the other. As even theu the I'otary motion 

 is probably less than normal, tlie joint between the occiput and the atlas is less cup-shaped 

 and more adapted to a sliding motion than usual. It is interesting to observe a very 

 similar conformation in the sockets for the condyles on the atlas of H-3, which is fused 

 with the atlas. Thus rotary motion is possible to some extent in that joint, and the work 

 of flexion and extension was no doubt in part carried on by the other joints of the cer\'ical 

 vertebrae. To the action of the same cause we must turn for the modification in the 

 length of the thoracic vertebrae, by which 11 more than compensated for the ordinary 12. 

 In the same way is to be explained the mothfications of wliich we have had so many 

 examples, by which an approach to the normal dispositions of regions is so often obtained. 

 Still another instance is furnished by spine 202, in which a cervical rib which ends free 

 has a cartilage shooting out from the manubrium to meet it. Similar cases have been 

 described by Leboucq and others. This shows clearly that the production of the costal 

 arch depends on something more than the over-development of the costal element of the 

 vertebra. It begins at both ends in cases which approach a complete arch. Clearly there 

 is some force acting throughout the organism, not at one point alone, to inaugurate such 

 changes. 



If Rosenberg's views on the changes of position of the ilium, and the consequent 

 modification of vertebrae according to the position it finally assumes, which we have pro- 

 visionally adopted, be correct, there is the evident need of some explanation of the cause 

 of these modifications. I can see no other than the action of the vital principle. If it be 

 objected that this is only a name to cover our ignorance, then the objection must be made 

 against gravitation, magnetism, etc., for in no case do we know how the force works. 

 We see only results. 



Apart from other deductions, the following conclusions seem justified. 



I. Variations occur in two ways: (1) by irregular development of the costal ele- 

 ments at and near the ends of the regions of the spine, and (2) by irregular segmentation 

 through which there are more or fewer vertebrae than normal. 



II. Variations of both kinds are variations around a mean. It is not impossible that 

 some of them may be reversive ; that any are progressive is mere assertion. 



III. Assuming the correctness of Rosenberg's studies in ontogenesis, his ^iew may 

 account for some of the variations, but even in these cases something more is needed to 

 explain the concomitant changes. 



IV. Variation of the costal elements at one end of a region is often associated with 

 variation of an opposite nature of those at the other end. Several regions may be 

 involved, and the two sides may vary independently. 



