160 BULLETIN OF THE UNITED STATES FISH COMMISSION. 



cranium are lettered so tliat their names may easily be made out from 

 the list of reference letters. I have attemjited to show the arrange- 

 ment of the pavement of choroid cells, cc, in the eye as may be shown 

 in preparations mounted in balsam. I would also direct especial atten- 

 tion to the conical teeth represented on the epithelium of the lower 

 jaw. These seem to be developed in epithelial pits and are not in direct 

 connection with the skeleton of the jaw, which is, moreover, not yet bony. 

 They surmount little epithelial papilloB and grow by the addition of ma- 

 terial from below ; their composition does not appear to be calcareous, 

 but corneous, since I find them to resist the action of acids. Kathke 

 has described teeth of a somewhat similar character in the embryos of 

 the viviparous bleuuy (Zoarces). 



In embryos of this age the branchial leaflets are also developed. 

 They at first ai)pear on the posterior border of the gill arches as small 

 papillie, which, as they elongate, throw out processes from their edges^ 

 so that they eventually acquire a bipinnate structure. In these bipin- 

 uate fleshy processes capillary loops are formed, which communicate 

 between the branchial arteries and veins. The leaflets with their capil- 

 laries are the agents directly concerned in the aeration of the blood of 

 the young fish. 



The coraco-scapular rods esc, although apparently cartilaginous, have 

 an histological composition different from the cartilages of the head^ 

 being much more hyaline. It is also embedded in a vertical fold which 

 extends some way beyond the upper and lower borders of the breast fiu. 

 This may be called the pectoral fold. It is not at all imi^robable that 

 we will yet find embryo Teleosts in which there are continuous lateral 

 folds, for we already know species in which the x)riuiiti\ e natatory fold 

 is discontinuous at a very early age. Such is the case with Hippocampus 

 and Syngnathus, according to my own observations this season. Hippo- 

 campns never develops a caudal fin, so that we would naturally not 

 expect to find the natatory fold prolonged over the end of the tail ; but 

 the posterior position of the early rudiments of the pectorals in CyMum 

 and ParepMppuSj it appears to me, is a very strong argument against 

 their origin from a ijosterior branchial arch j indeed, it is the strongest 

 3^et offered against that doctrine by anj^ data derived from a study of 

 the development of the paired fins of Teleosts. In other words, since 

 we now know that the natatory fold, from which the unpaired median 

 fins are developed, is sometimes discontinuous, I see no reason why we 

 should not expect to find the lateral fin-folds discontinuous, as there are 

 more reasons why they should be so in the Teleost than in the Elasmo- 

 branch embryo. In fact, it would appear that the greater the longi- 

 tudinal extent of the unpaired fins in proportion to the length of the 

 body of the adult the more likelihood there is of finding a continuous 

 dorsal and ventral natatory fold developed in the larva, and vice verm. 

 The longitudinal extent of the paired fins of Teleost fishes is less, vastly 

 less, in resj)ect to the number of rays, than those of the Elasmobranchs, 



