150 THE THEORY OF EVOLUTION 



ing and selection has been by a great series of experi- 

 ments, but, unfortunately, the conditions under 

 which these were carried out are too inexact and the 

 records, which are usually lacking altogether, are, 

 even when preserved, too vague to be of much service 

 in a modern investigation. Domestication, as was 

 pointed out at some length in a previous lecture 

 (p. 36), shows how surprisingly animals and plants 

 may be changed by human agency, but tells us little 

 of what we are now seeking to learn. 



The changes which are experimentally produced 

 must be transmissible to the offspring, if they are 

 to have any evolutionary significance. Plants, in 

 particular, are very susceptible to extensive modifica- 

 tion through external influences, such as the chemical 

 constitution of the soil, heat, light, moisture, eleva- 

 tion above sea-level, and the like, and these modifica- 

 tions will continue as long as the external conditions 

 remain the same, but disappear when the plants 

 are once more placed under the original circum- 

 stances. The French botanist Bouvier took shoots 

 or layers from several kinds of common plants and 

 set out, from the same individual, some shoots in 

 the lowlands, others high up in the mountains. The 

 differences in the results were remarkable; the moun- 

 tain plants had more vigorous roots, but the part 

 above ground was much smaller, with more delicate 

 stem and leaves; the flowers were fewer in number, 

 but individually larger and more intensely coloured. 

 When taken back to the lowlands, the plants re- 



