CLASSIFICATION OF VARIATIONS 89 



be taken as a basis for discussion. The literature of horticulture 

 for example abounds in cases alleged, but I do not think anyone 

 can produce an illustration quite free from doubt. Such evidence 

 is usually open to the suspicion that the plant was either intro- 

 duced by some accident, or that it arose from a cross with a pre- 

 existing dominant, or that it owed its origin to the meeting of 

 complementary factors. In medical literature almost alone how- 

 ever, there are numerous records of the spontaneous origin of 

 various abnormal conditions in man which habitually behave 

 as dominants, and of the authenticity of some of these there 

 can be no doubt. 



When we know that such conditions as hereditary cataract 

 or various deformities of the fingers behave as dominants, we 

 recognize that those conditions must be due to the addition of 

 some element to the constitution of the normal man. In the 

 collections of pedigrees relating to such pathological dominants 

 there are usually to be found alleged instances of the origin of the 

 condition de novo. Not only do these records occur with such 

 frequency that they cannot be readily set aside as errors, but from 

 general considerations it must be obvious that as these mal- 

 formations are not common to normal humanity they must at 

 some moment of time have been introduced. The lay reader 

 may not be so much impressed with the difficulty as we are. He 

 is accustomed to regard the origin of any new character as equally 

 mysterious, but when once dominants are distinguished from 

 recessives the problem wears a new aspect. Thus the appearance 

 of high artistic gifts, whether as an attribute of a race or as a 

 sporadic event among the children of parents destitute of such 

 faculties, is not very surprising, for we feel fairly sure that the 

 faculty is a recessive, due to the loss of a controlling or inhibiting 

 factor; but the de novo origin of brachydactylous fingers in a 

 child of normal parents is of quite a different nature, and must 

 indicate the action of some new specific cause. 



Whether such evidence is applicable to the general problem 

 of evolution may with some plausibility be questioned; but 

 there is an obvious significance in the fact that it is among these 

 pathological occurrences that we meet with phenomena most 



