46 CARNEGIE INSTITUTION OF WASHINGTON. 



heart that I congratulate you on the good fortune that made a'ou 

 find this combination in the appointed director. 



Fluctuating variability, however, has been the chief line of study 

 for Mr. Davenport, and he would be a bold man who would try to 

 show the way where such a guide is at hand. 



For my part I prefer confining myself to such questions as are 

 more obviously touched by my own line of work. The experience 

 of agriculturists and horticulturists has long since established the 

 fact that new forms of animals and plants from time to time arise. 

 How they originate is another question, which it is not the task of 

 practice, but of science, to answer. The fact, however, is undeni- 

 able, and all observations point to sudden changes or so-called sports 

 as the first beginning. Especially in the dominion of horticulture 

 Korshinsky has shown, by an ample critical survey of the historical 

 evidence, that sudden sports are the prevailing rule and probably 

 even the exclusive manner of originating of new varieties. 



Such considerations have led to the conviction that what occurs 

 in horticulture must also occur in the experimental garden. If the 

 conditions are the same, why should not the phenomena be the same, 

 too ? If mutations are rare in horticulture, the experimenter has 

 only to arrange his work so as to be able to detect rare occurrences 

 in his cultures, too. In doing this I have succeeded in observing 

 mutations quite analogous to the horticultural instances, and col- 

 lecting all the evidence concerning their ancestry and their descend- 

 ants as well as concerning the mode of their appearance. 



Moreover, I have had the good fortune of discovering a wild 

 plant which is even yet in a condition of mutability. Yearly it is 

 observed to produce new species. It is the large-flowered evening 

 primrose, which bears the name of Lamarck, the founder of the 

 theor)^ of evolution. It clearly shows how new species arise from 

 an old stock, not by continuous and slow changes, but all of a sud- 

 den. The stock itself is not altered by the process nor even notice- 

 ably diminished. The new species which it produces arise on all 

 sides. Some of them are in a higher, others in a lesser degree fit 

 for their life conditions ; some persist during years, while others 

 disappear nearly as soon as they arise. 



This instance of experimental mutations is found largely to agree 

 with the experience of breeders, especially in horticulture, and like- 

 wise with the conclusions that have been drawn from comparative' 

 studies. The assumption that those species and genera which now 

 consist of large groups of closely allied forms have originated in 



