82 PROCEEDINGS OF THE ACADEMY OF [1893. 



digits. The regenerative powers therefore seem to become dimin- 

 ished with the specialization of a type, and to be manifested last of 

 all in the distal ends of the extremities. The most marked changes 

 may occur in the way of duplication, triplication, etc., of the axis, 

 even in the highest types, provided the development or processes of 

 segmentation are disturbed early enough, and in some definite way. 

 The conclusion which is warranted from the whole of the foregoing 

 is that the regenerative power of organisms disappears as we rise in 

 the scale of organization, last of all in the peripheral extremital parts. 

 A further observation is justified, which is, that the power toprodvce 

 monstrosities or congenital aberrations of development due to external 

 disturbances of segmentation, during growth, diminishes in the higher 

 forms pari passu with the advance in development. Finally, it may 

 be affirmed with much show of probability that aberrations of 

 development produced by disturbances of the processes of segmen- 

 tation during growth, may become hereditary, as illustrated by the 

 probable inheritance of partially duplicated axes in the tails of 

 Japanese gold-fishes, or of supernumerary digits in many mammals 

 and in Man. 



How such new parts, originated as here supposed, can transmit to 

 the ova of the parent body, of which they form a part, a tendency 

 to cause them ( the new characters ) to reappear in the offspring 

 developed from such ova or germs, cannot be explained without 

 supposing that the new part of the body of the parent influences at 

 a distance the characters and potentialities of the germinal matter- 

 produced by it in its gonads. The difficulties in the case of the 

 partly double body of the Japanese Gold-fishes are peculiar, in that 

 we have to assume that hereditary tendencies are transmitted from 

 a parent body and part of another of each sex, making in all, ten- 

 dencies transmitted from two bodies and parts of two others, in the 

 course of the ordinary sexual reproduction of the double-tailed 

 species, since there cannot be the slightest doubt that, in this case, 

 we have to deal with partially double monsters (as is proved by 

 their morphology), with normally developed reproductive powers. I 

 confess to an utter inability to see how this can be done by means of 

 gemmules, the oldest hypothesis of heredity, first formulated by 

 Democritus, then restated in a more modern form by Buffon, then 

 by Erasmus Darwin in the Zoonomia, and lastly by Charles Darwin. 

 This doctrine therefore has a very respectable pedigree, but it is 

 strange that none of the later writers credit their predecessors with 



