2o 4 / ARTHUR THOMSON [march 



experiments on hybridisation, made by botanists, zoologists, and prac- 

 tical people, have led us to expect one of three results when a 

 crossing has a successful issue. (1) The hybrid may be intermediate 

 between its parents, sometimes so exactly that we may liken the 

 blending to warp and woof; or (2) the hybrid may show an exaggera- 

 tion of the characters of one parent, often with little apparent 

 realisation of the peculiarities of the other ; or (3) the hybrid may be 

 very different from either parent, showing features at first sight novel, 

 but which on closer investigation are sometimes interpretable as the 

 reassertion of the characters of a remoter ancestor. But the extra- 

 ordinary thing is that at least two of these three different results may 

 be illustrated in one brood or litter. 



According to the theory of reversion, confessedly a somewhat 

 unfortunate term, the resemblance which an offspring often exhibits to 

 a more or less distant ancestor, is due to the realisation of characters 

 which were throughout part of the inheritance, but remained latent or 

 unexpressed for one or more generations. As to the fact of resem- 

 blance to ancestors there is no more doubt than there is as to 

 resemblance to parents ; the theoretical element is simply in the idea 

 of latent characters. If we do not accept the idea that resemblance to 

 ancestors is due to the reassertion of latent elements in the inheritance, 

 we must find some other explanation. And there seem to be two 

 possibilities — (1) that the ancestral resemblance may be due to the 

 fresh and independent occurrence of the same permutations and 

 combinations of germinal material as took place when the ancestral 

 character had its origin ; or (2) that the character of resemblance may 

 be an individually acquired " modification," reproduced, apart from 

 inheritance, by a recurrence of suitable external conditions. 



It seems impossible to read the literature on the subject without 

 becoming convinced that many phenomena are labelled reversions on 

 the flimsiest of evidence. Thus the occurrence of a Cyclopean human 

 monster with a median eye has been called a reversion to the sea- 

 squirt, and gout has been called a reversion to the reptilian condition 

 of liver and kidneys. Often there has not been the slightest attempt 

 made to discriminate between true reversion (i.e. the re-expression of 

 latent ancestral characters) and the phenomena of arrested develop- 

 ment, or of abnormalities which have plainly been induced from 

 without. Often, too, there has been no scruple in naming or inventing 

 the ancestor to whom the reversion is supposed to occur, although 

 evidence of the pedigree is awanting ; and the vicious circle is not 

 unknown of arguing to the supposed ancestor from the supposed 

 reversion, and then justifying the term reversion from its resemblance 

 to the supposed ancestor. Little allowance has been made for 

 coincidence, and the postulate of characters remaining latent for 

 millions of years is made as glibly as if it were just as simple as a 

 throw-back to a rn-eat-orandfather. 



