76 REVERSION. [introd. 



to interpret Homology on this view of Heredity, I have already 

 spoken in Section VI. 



Secondly, the metaphor of Heredity misrepresents the essential 

 phenomenon of reproduction. In the light of modern investiga- 

 tions, and especially those of Weismann on the continuity of the 

 germ-cells, it is likely that the relation of parent to offspring, 

 if it has any analogy with the succession of property, is rather 

 that of trustee than of testator. 



Hereafter, perhaps, it may be found possible to replace this 

 false metaphor by some more correct expression, but for our 

 present purpose this is not yet necessary. In the first exami- 

 nation of the facts of Variation, I believe it is best to 

 attempt no particular consideration of the working of Heredity. 

 The phenomena of Variation and the origin of a variety must 

 necessarily be studied first, while the question of the perpetua- 

 tion of the variety properly forms a distinct subject. Whenever 

 in the cases given, observations respecting inheritance are forth- 

 coming they will be of course mentioned. But speaking of dis- 

 continuous Variation in general, the recurrence of a variation 

 in offspring, either in the original form or in some modification 

 of it, has been seen in so many cases, that we shall not go far 

 wrong in at least assuming the possibility that it may reappear 

 in the offspring. At the present moment, indeed, to this state- 

 ment there is little to add. So long as systematic experiments 

 in breeding are wanting, and so long as the attention of naturalists 

 is limited to the study of normal forms, in this part of biology 

 which is perhaps of greater theoretical and even practical im- 

 portance than any other, there can be no progress. 



2. Reversion. 



Around the term Reversion a singular set of false ideas have 

 gathered themselves. "^On the hypothesis that all perfection and 

 completeness of form or of correlation of parts is the work of 

 Selection it is difficult to explain the discontinuous occurrence 

 of new forms possessing such perfection and completeness!^ To 

 account for these, the hypothesis of Reversion to an ancestral 

 form is proposed, and with some has found fayour. That this 

 suggestion is inadmissible is shewn at once by the frequent occur- 

 rence by discontinuous Variation, of forms which though equally 

 perfect, cannot all be ancestral. In the case of Veronica and 

 Linaria, for example, a host of symmetrical forms of the floral 

 organs may be seen occurring suddenly as sports, and of these 

 though any one may conceivably have been ancestral, the same 

 cannot be supposed of all, for their forms are mutually exclusive. 

 On Veronica buasbaumii, for instance, are many symmetrical 

 flowers, having two posterior petals, like those of other Scrophu- 

 larinese : these may reasonably be supposed to be ancestral, but 



