No. 3-] THE VERTEBRAL COLUMN OF FISHES. 137 



the same transverse plane, but that the basiventrals lie in the 

 plane of the interdorsals. This I regard as a palpable error. 

 Fig. 10 of my paper already referred to was drawn under 

 the camera from a sagittal section near the mid-line. It shows 

 at the right hand that in the middle of the tail the homotype 

 pieces do lie quite accurately in the same transverse plane. 

 Sections of other specimens confirm this view. The authors 

 are correct in their statement that the basidorsals of the trunk 

 region rest on the summits of the interdorsals, but each basi- 

 dorsal rests on the interdorsal which was originally in front of 

 it, not on the one behind. I can discover no reason whatever 

 for supposing that the basiventrals have been pushed back- 

 ward. 



On page 202 of their paper Gadow and Abbott proceed to 

 explain the arrangement of the arcualia and the manner of their 

 consolidation into the definitive vertebrae. They say there is, 

 outside of the elastica, a thick zone of connective tissue, which 

 forms a layer of bone on its inner suface, and that, in this zone 

 of connective tissue, cartilage cells from the basal portions of 

 the arcualia grow round the chordal sheath preparatory to the 

 formation of the central discs. In the tail, all the arcualia 

 rest on the thin layer of bone which is just outside of the 

 elastica, and are themselves, for the most part, surrounded by 

 bone. In the trunk region, as already stated, the basidorsals 

 are out of contact with the notochord. The outgrowth from 

 the anterior half of the basiventral grows upward in front of the 

 downwardly directed outgrowth from the posterior half of 

 the interdorsal just above it ; while the interventral becomes 

 united in a similar way with the basidorsal just above it. 

 These contiguous semi-rings then fuse to produce " com- 

 plete rings of cartilage, hyaline, and perichondrally ossify- 

 ing in the arcualia, more fibro- cartilaginous in the newly 

 formed belt." In the tail, of course, there are two such rings 

 for each skleromere. But not all of the outgrowth from 

 each arcuale develops into cartilage, only the anterior half 

 of it. The hinder half becomes connective tissue and then 

 is changed directly into bone. It results from this that there 

 are for each skleromere of the tail two belts of cartilage and 



