INTERNAL CAUSES OF VARIATION 187 



The best students, however, insist that very little real evidence 

 has been produced on the subject, and none at all tending to 

 prove the existence of this influence. 1 



With a view to testing somewhat the real extent of this be- 

 lief, the author addressed letters to the best-known dog fan- 

 ciers of the United States. Of thirty-seven answers received, 

 one writer is a believer in telegony ; six somewhat mildly 

 express uncertainty ; two are non-committal ; and twenty-eight 

 are outspoken against the theory. The most outspoken of them 

 all is a well-known fancier of long experience. Judging from 

 this small number, it would seem that this belief among dog 

 fanciers has been overrated. 



Proof by the method of instance. Without a reasonable doubt 

 belief in telegenic influence rests upon stray instances, difficult 

 of understanding by those who happened to be the observers, 

 and hastily accepted as evidence. Now nobody should be more 

 careful than the breeder to judge accurately the nature and 

 value of evidence. A single instance may be good negative 

 testimony, but it is seldom worth much as positive evidence. 

 The products of breeding are so many and so various, and the 

 causes of variation are so numerous and so complicated, that a 

 particular result can seldom be assigned to the operation of any 

 single cause. It is more likely the mixed or composite result 

 of many influences, both internal and external ; and in order to 

 know the effect of a single cause it is necessary to isolate the 

 case if possible, or, if not, to resort to the examination of large 

 numbers of cases, subject to varying degrees of influence, and 

 thus indirectly to estimate the effect of any special cause of vari- 

 ation. For example, stripes and bars were once common color 

 markings of horses, as they are now of asses, especially zebras 

 and quaggas. Consequently a certain proportion of colts, what- 

 ever the parentage, will be born with traces of shoulder and leg 

 markings. Now, under the laws of chance, a certain portion of 

 these will be the direct offspring of striped or barred sires, and 

 will attract no attention, the markings being considered heredi- 

 tary. By the same law of chance a certain (smaller) portion will 

 be the offspring of parents not barred, and a still smaller number 



1 Proceedings of the Royal Society, LX, 273. 



