Baik'y.] J-^O [jlay 1, 



many or several species and to have passed out with exceeding slow- 

 ness. 



To my mind, all this is but the normal result of the divergence of 

 character, or the survival of the unlike. A new type finds places of least 

 conflict, it spreads rapidly and widely, and thereby varies immensely. It 

 is a generalized type, and therefore adapts itself at once to many and 

 changing conditions. A virile plant is introduced into a country in which 

 the same or similar plants are unknown, and immediately it finds its 

 opportunity and becomes a weed, by which we mean that it spreads and 

 thrives everywhere. Darwin and Gray long ago elucidated this fact. 

 The trilobites, spirifers, conifers, ginkgos, were weed-types of their time, 

 the same as the composites are to-day. They were stronger than their 

 contemporaries, the same as our own weeds are stronger than the culti- 

 vated plants with which they grow. After a time, the new types outran 

 their opportunity, the remorseless struggle for existence tightened in upon 

 them, the intermediate unlikenesses had been blotted out, and finally only 

 one or two tj^pes remained, struggling on through the ages, but doomed 

 to perish with the continuing changes of the earth. They became spe- 

 cialized and inelastic ; and the highlj^ specialized is necessarily doomed to 

 extinction. Such remnants of a vanquished host remain to us in our 

 single liriodendron, the single ginkgo and sassafras, and the depleted 

 ranks of the conifers. 



My attention was first called to this line of thought by contemplating 

 upon the fact that cultivated plants difter widely in variability, and I was 

 struck by the fact that many of our most inextricably variable groups — as 

 the cucurbits, maize, citrus and the great tribes of composites — are still 

 unknown in a fossil state, presumably because of their recent origin. 

 Many other variable genera, to be sure, are well represented in fossil 

 species, as roses (although these are as late as the Eocene), pyrus, pru- 

 nus and musa ; but absolute age is not so significant as the comparative 

 age of the type, for types which originated very far back may be yet in 

 the comparative youth of their development. The summary conclusions 

 of a discussion of this subject were presented to the American Associa- 

 tion for the Advancement of Science two j'ears ago.* A modification of 

 these points, as I now understand them, would run something as follows : 



1. There is a wide difference in variability in cultivated plants. Some 

 Bpecies vary enormously, and others very little. 



3. This variability is not correlated with age of cultivation, degree of 

 cultivation, or geographical distribution. 



3. Variability of cultivated plants must be largely influenced and 

 directed, therefore, by some antecedent causes. 



4. The chief antecedent factor in directing this variability is probably 

 the age of the type. New types, in geologic time, are polymorphous; 

 old types are monomorphous and are tending towards extinction. The 

 most flexible types of cultivated plants are such as have probablj' not yet 



* Proc. A.A.A.S. 189-1, 255 ; Botanical Gazette, xix. 3S1. 



