JUXE 7, 1895.] 



SCIENCE. 



021 



backward to the clotted area a. If the 

 plane of section is along the line II the tail 

 ^\•ill regenerate upward so as to be restored 

 over the area indicated bj' the dotted line 

 enclosing h. If the plane of section of the 

 tail be along III the tail will be regenerated 

 downward to the dotted line enclosing the 

 area c. It is therefore e\4dent that Bar- 

 furth's law determines the inclination of the 

 axis of the regenerated part to the body- 

 axis, through the diflerent conditions of 

 surface tension that must be set up over re- 

 generating surfaces, whenever the inclina- 

 tion of these to the axis of the whole or- 

 ganism is changed. 



Xew equilibria of surface tension estab- 

 lished reciprocallj' between the cohering 

 but independentlj' developing segments of 

 the oosperm of the sea-urchin, that have 

 been imperfectlj- separated by mechanical 

 or other means, also cause changes to be 

 produced in the forms of the single larv;e 

 of such coherent groups, and in the spicular 

 skeleton, for the same reason, as is proved 

 by Figs. 23 to 2.5 given bj' Professor Loeb.* 

 Those ligures also illustrate the thesis that 

 the seolotropy of the distinctly developing 

 segments of the egg must be nearly the 

 same, and that component or resultant equi- 

 poteutial surfaces are developed by the in- 

 teracting molecular machinery of such co- 

 herentlj- developing or compound larva?. 



The angular divergence of duplicated tails 

 and toes as well as the axes of monstrous 

 embryos is explained by Barfurth's discov- 

 ery, taken together with the principle that 

 division of a germ does not change the 

 leolotropy of its segments. If this inter- 

 pretation is the correct one, the origin of 

 superuumerarj- digits must be traced back 

 to mechanical disturbances of the processes 

 of ontogeny. The rationale of the manner 

 in which divergent supernumerary toes 

 may be produced is shown in Fig. 3, repre- 



*Bi(>h(jical Lccluie.i (Xo. III. ). Delivered at Woods 

 HoU, Mass., in 1893. Giun & Co., Boston. 



senting the regenerating toes of the foot of 

 a salamander. 



If the toes were cut straight across at 

 the points I., II., III., IV., the toes would 

 regenerate normally. If, however, the re- 

 generating surfaces were divided into two 

 areas in each case by a line along which re- 

 generation were prevented, two toes \^•ould 

 arise from each surface. The angular 

 divergence of the pairs of supernumerary 

 toes thus produced would be measured by 

 the angular inclination to one another of 

 the two areas at the end of each original 

 toe that was thus doubly regenerated. In 

 other words, supernumerary digits are the 

 results directly or indirectly of something 

 akin to mutilations. That such duplica- 

 tions may be produced by mutilations there 

 can be no doubt, and of their transmission 

 by inheritance to otl'spriug there is also no 

 doubt. These facts make it probable at 

 auj' rate that regeneration of distal parts, 

 and the likelihood with which they reappear 

 in duplicate, is due to causes similar or 

 identical in character with those that lead 

 to the production of double monstei-s, by 

 shaking, mutilation or other physical inter- 

 ference with the normal development of the 

 oosperm. The question of the inheritance 

 of mutilations is consequently far from be- 

 ing concluded as viewed from this new 

 standpoint. Much evidence might be ad- 

 duced in support of my contention did space 

 allow. The hereditary transmission of 

 such monstrosities as supernumerary digits 



