78 PROCEEDINGS OF THE ACADEMY OF [1893. 



from the stump of the old ray in a divergent fashion. This fact I 

 have repeatedly observed, as well as the regeneration, in situ, of the 

 scales, where they have been forcibly, removed from the fish by 

 accident. A case of this kind I traced and described some years 

 ago, the subject being a fine specimen of the mirror-carp that had 

 lost one of the large scales of the lateral line. In this case, how- 

 ever, there is no duplication of the scale during its regeneration. 



It is also a notorious fact that when extra toes or fingers are 

 developed in the human subject, or in a mammal, they tend to 

 become hereditary. And the same seems to be the case with birds, 

 in the instance of the Dorking fowl amongst domestic breeds. It is 

 my belief from the evidence presented above that such supernumerary 

 digits have been in the first place produced by disturbances during 

 segmentation or karyokinesis. While the embryo was developing 

 its dibits in such higher forms and that such disturbances have been 

 registered, so to speak, as part of the hereditary tendencies of the 

 organism, and have thus been handed on to offspring as develop- 

 mental tendencies. The tendency to reproduce supernumerary toes 

 in the cat is well-known. The same tendency is shown in the dispo- 

 sition to reproduce extra thumbs or toes in certain human families. 

 The most extraordinary of these duplications is however presented 

 by the double or fan-tailed races of Japanese Gold-fishes, where the 

 duplication sometimes extends even to the anal fin, so that I have 

 elsewhere been tempted, upon the evidence presented by Dr. Watase 

 as to the development of these fishes, to put in a plea for the actual 

 realization of an eight-limbed vertebrate, a thing against which the 

 learned Professor Haughton, of Dublin, has argued with his wonted 

 acumen, in a still earlier paper touching the utility of the four- 

 limbed condition of the higher vertebrates. 



I believe, however, now that the key to the mystery of the pro- 

 duction of the double-tailed gold-fishes is within our reach in the 

 light of the evidence presented above, that we can guess how 

 the Japanese went to work to produce their singular breeds of 

 Carassius auratus. I believe, in short, that their practice was very 

 simple, and that they have probably antedated the practical appli- 

 cation of the facts of experimental embryology, as known in Europe 

 and our still more western civilization, by several centuries. Since 

 the Japanese records of some of these breeds are very old, as well 

 as the fact that the work of the Marquis de la Savigny, figuring the 

 breeds of these fishes, was published in 1790, we can infer for them 

 an antiquity of, at the very least, a couple of centuries. 



