J. K. Thdcher — Mediati and Paired Fins. 283 



tlu' lartio second dors.al, and sliows one of the longest rays. Where 

 they are shorter we may liave only one bifurcation, or one of the two 

 l)riniary branches, that toward the longer rays, may again divide, 

 while the other toward the shorter remains uncleft. Farther forward 

 at the beginning of the second dorsal, where the rays are still shorter, 

 they do not divide at all, but end somewhat bluntly though com- 

 pressed from side to side. 



Each ray is largest in the middle and here lies quite close to its 

 adjacent rays ; below they grow more slender, and therefore are some- 

 what spaced, but expand somewhat to a foot resting on the myelonal 

 canal. 



I have seen no cases of concrescence between adjacent rays. With 

 the exception of the variation in the branching and length in different 

 parts of the fin, before alluded to, the rays are all similar and parallel 

 one to another. 



This branching is plainly a true dichotomy and not the product of 

 concrescence, as is evidenced by the total absence of anything else 

 resembling concrescence, by the similarity in size between two adja- 

 cent differently branched rays, and by the regularity of the branching. 



On each side of the row of skeletal elements are muscular bundles 

 of a somewhat blacker color than the two great masses of lateral 

 muscles. The muscles of the median fin wedge themselves into the 

 angle between the lateral muscles of the two sides along the 

 median line. They ai-e sharjdy distinguished from these. The fibers 

 of the lateral muscles run longitudinally, while those of the fin mus- 

 cles are parallel to the pi'iniordial median fin-rays. There is abso- 

 lutely no continuity between the two in any part. Moreover the 

 bundles of the fin-muscles show no relations to the segments of the 

 lateral muscles. A cross section, PI. XLIX, fig. 3, shows the relation 

 between the fin-muscle bundles and fin-rays. 



The numerical relation between the fin-rays and the segments of 

 the lateral muscle is shown in PI. XLIX, fig. 2, where we have a little 

 less than four of the former to one of the latter. 



The relation between the tin-rays and the neural arches is shown 

 in figure 1, where we have 35 rods and 23 arches. These neural 

 arches rise from the sheath of the notochord, to stiffen the fibrous 

 sides of the myelonal canal and to apply themselves to its fatty- 

 fibrous ridge-pole. The fin-iays abut on the same ridge-pole in the 

 mid-dorsal line, but they are in no way connected with the neural 

 arches. 1 have met with no cases even of concrescence between the 

 two. 



