The Evidence from Variation. 149 



cultivated forms to be more prolific of gemmules than 

 wild species. The fact that bad variation, like ordinary 

 variation, is most common in cultivated forms, seems to 

 show that the tendency to vary is excited in buds, as it is 

 in fertilized ova, by the influence of gemmules which are 

 thrown off by the cells of the body under new or unnat- 

 ural conditions, and we can easily understand why it 

 should be more frequent where gemmules are abundant 

 than in a form with few gemmules, for the chance in 

 favor of the accidental transmission of a gem mule to a 

 growing or nascent bud will increase as the number of 

 gemmules increases. 



Changed Conditions do not act directly, but they cause 

 Subsequent Generations to vary. 



This strange and, as I hope to show, highly significant 

 law has been noted by many observers, and a long list of 

 illustrations might be quoted. 



As Darwin points out, it is certainly a remarkable fact 

 that changed conditions should at first produce, so far 

 as we can see, absolutely no effect, but that they should 

 subsequently cause the character of the species to 

 change. 



The late Dr. Jared P. Kirtland told me that for 

 more than forty years he tried in vain to obtain varieties 

 from the common red cherry, but that when at last va- 

 rieties began to appear the variability was very great: 

 that after it had once become established it continued 

 for many years with no diminution. 



It is well known that when new flowers are first intro- 

 duced into gardens they do not vary, although all, with 

 rarest exceptions, ultimately vary. 



Darwin, in his Variation, Vol. ii. p. 316, quotes the 

 following illustrations of this law : " Mr. Salter re- 



