398 THE AMERICAN NATURALIST. [Vor. XXXV. 
(1) “If the undeveloped end of the rst 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 fulfill these 
conditions, they are quite lost in the multitude 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.” 
These facts are in contradiction to Rosenberg’s theory. 
(2) The author points to the cases where * we see a tendency 
sometimes for the whole thorax to move forward (upward ?) by cervi- 
cal ribs associated with absence of the last thoracic ones or with 
their existence in a rudimentary condition.” “We also see cases in 
which, when the cervical rib on one side is distinctly larger than 
its fellow, the last rib on that side is either correspondingly smaller 
than its fellow, or even replaced by a pretty typical transverse 
process.” These concomitant variations **may extend even further, so 
as to include the sacralization of one side of the last lumbar, or even 
the absorption of one side of the atlas into the occiput.” 
The causes of the variation: The author confesses his inability to 
show the original cause of the phenomenon. “It is clear, however, 
that the vertebra at the junction of regions are particularly variable, 
and it seems hard to doubt that errors of segmentation may occut- 
The original error having occurred, there seems to be a tendency in 
the organism to reproduce the type as nearly as may be under the 
changed conditions; to make as normal a series of regions as CI 
cumstances will permit; and this tendency manifests itself to some 
extent independently in the two halves of the spine." For this 
tendency the author adopted the old and rather unsatisfactory theory 
of ** the vital principle." 
Professor Dwight closes his interesting work with the following 
additional deductions : 
1. Variations occur in two ways : (1) by irregular development of 
the costal elements at and near the ends of the regions of the spine, 
and (2) by irregular segmentation through which there are more OF 
fewer vertebrz: than normal | : 
2. Variations of both kinds are variations around a mean. It 1s 
not impossible that some of them may be reversive ; that any are 
progressive is mere assertion. : 
3. Assuming the correctness of Rosenberg's studies in ontogenesi 
