1084 REPORT OF COMMISSIONER OF FISH AND FISHERIES. [10 I] 



such an inequality in the relative width of the unpaired epural aud hy- 

 pural fin system as could lead to the production of heterocercy ? We see 

 that such inequality has not arisen and that i)erfect diphycercy has been 

 maintained in tbe Dipnoi, from which it follows that it is safe to assert 

 that there must have been some dift'erence between the forces which 

 acted upon the ancestry of the latter and that of the existing TeUosteL 

 In this way only can we conceive archaic characters to have been pre- 

 served aud handed down to the present. , 



Huxley's statement,* that "in all Teleostean fishes the extremity of 

 the spinal column bends up, aud a far greater number of the caudal fin- 

 rays lie below than above it," must now be qualified so far as to admit 

 that in some Teleosts there is no such upbending of the end of the axial 

 skeleton, as in Mol/i, Fierasfer, aud Gastrostomus, for example; nor is 

 it invariably true that a greater number of fin-rays lie below than above 

 the termination of the axial skeleton. The result of these investigations 

 has accordingly been to modify somewhat the accepted views of the 

 evolution of the caudal fin from a prototype which was essentially 

 orthaxial i)osteriorly, or one in which the posterior terminus of the axial 

 skeleton was straight and in line with the thoracic portion. 



The parallelism existing between the tails of the orthaxial lophocer- 

 cal larvae of existing fishes and the orthaxial caudal of CceJacanthus, 

 Coccosteus, Glyptolwmus, and Gyroptychius is not exact, because these 

 extinct forms had more or less clearly marked rays or apophyses devel- 

 oped. Coccosteus seems to have had no caudal fin, but had apophyses 

 developed above and below its notocliordal axis posteriorly, thus attain- 

 ing a specialization almost as marked in this respect as that observed 

 in the existing genus Hippocampus. These are facts which njust not be 

 lost sight of, for we may assume that the development of a form reca- 

 pitulates the development of the phylum to which it belongs, when in 

 fact it does so very inexactly, as the examples above cited show. An- 

 other illustration may be cited, viz, that of the protopterygian stage of 

 the Salmon embryo when its embryonic rays, cartilaginous apophyses, 

 and notochord recall the permanent condition found in the Dipnoi. 

 But in the Salmon embryo we find has no perichondrial ossifications de- 

 veloped at this time about the epaxial and hypaxial apophyses of its 

 axial skeleton, as in the Dipnoi^ nor is the end of the chorda any longer 

 orthaxial at its termination, as in the latter. 



These accelerated departures from the primitive type are evidently 

 dependent for their manifestation u{)on heredity, since structures char- 

 acteristic of classes, orders, families, genera, and species seem to ap- 

 pear in most embryos in about tiie order of rank in which they stand 

 above, though in many cases so pronounced is the direct influence of 

 heredity that a family or, at most, an ordinal, character will appear at 

 the end of the first day, as in the case of the development of the barbels 

 of Amiurus. 



* Auat. of Vcrtebratetl Animals, j). 131. 



