390 HARRIS HAWTHORNE WILDER ON 



fishes is the direct attachment of the pelvic girdle, which by the medium of a specialized 

 pair of ribs, becomes articulated to a single vertebra, usually the 19th. This vertebra, 

 the sacral, lies naturally in the cloacal region and it seems a matter of douljt whether, as 

 in higher forms, to consider it the boundary between the trunk and the tail, or whether 

 as in fishes, to limit the latter region to those vertebrae which bear closed haemal arches, 

 the first of which is usually the fourth vertebra posterior to the sacral one, /. e., the 23d. 

 The first of these alternatives seems the more natural and open to the fewest objections, 

 for, while it places the first few caudal vertebrae in the same list structurally with those 

 of tlie trunk, from which, indeed, they are practically indistinguishable, it avoids the 

 equally great embarrassment of lea\dug the same number as a nameless and anomalous 

 group intervening between the sacral and caiidal regions, a relationship unlike anything 

 occurring elsewhere among vertebrates. 



The statements given above, that the two vertebrae which have a definite distinction, 

 the sacrum and the "first haemal arch vertebra," are usually the 19th and 23d respec- 

 tively, suggest variation in tliis order, a matter which has been made the subject of 

 papers by G. H. Parker ('96) and Bumpus ('97), and has been referred to by Waite (97). 



Regarding the sacral vertel^ra, according to the first two authors it was the 19th in 

 65% of 127 specimens examined, the 20tli in 27% and in the remaining 8% it was placed 

 " obliquelj-,'' that is, A\ith au attachment to the 19th vertebra upon one side and the 20th 

 upon the other. This obliquity is usually sinistro-dextral, i. e., with the left sacral rib 

 in advance of the right (left, 19tli; right, 20th). Out of eight such oblique specimens 

 examined b^' Bumpus, seveu were sinistro-dextral, and only one dextro-sinistral. 

 Parker's two specimens were both sinistro-dextral. 



Waite describes three oblique specimens, which show for the most part additional 

 abnormalities. In one of these the right sacral rib was on the 19tli vertebra and the left 

 on the 20th, thus making the direction dextro-sinistral. In the other two cases the right 

 rilj was on the 18th, a case not found by the other authors named, and the left on the 

 19th, again dextro-sinistral. 



Waite also figures a case with two normal sacral ribs on the 19th vertebra, and an 

 additional smaller rib upon the right side of the 20th, also attached to the ilium. 



A similar variation occurs in the position of the 1st haemal arch, although a careful 

 comparison by Bumpus has shown that variation here is entirely independent of that of 

 the sacral vertebra. Out of 98 specimens examined with reference to this, the 1st haemal 

 arch was borne on the 22d vertebra in 11 specimens, on the 23d in 82, and on the 21:th 

 in 5. 



The variation in the total number of vertebrae, which is considerable, is seen to be 

 mainly that of the caudal region, as it has been shown that there is a variation of but two 



