HISTORICAL DEVELOPMENT 5 
a comparative study of certain polydemic species common to the arctic 
islands, Jan Meyen and Spitzenberg, and to the Alps. Both of these meth- 
ods are fundamental to field experiment, but the results are inconclusive, 
inasmuch as altitude is a complex of factors. As no careful study was 
made of the latter, it was manifestly impossible to refer changes and dif- 
ferences of structure to the definite cause. In a paper that has just ap- 
peared, E. S. Clements (1905) has applied the method of polydemic com- 
parison to nearly a hundred species of the Rocky mountains. In this work, 
the all-important advance has been made of determining accurately the de- 
cisive differences between the two or more habitats of the same species in 
terms of direct factors, water-content, humidity, and light. In his own 
investigations of Colorado mountain vegetation, the author has applied the 
method of field cultures by planting seeds of somewhat plastic species in 
habitats of measured value, and has thought to initiate a new line of re- 
search by applying experimental methods to the study of vegetation as an 
organism. In connection with this, there has also been developed a method 
of control experiment in the plant house under definitely measured differ- 
ences of water and light. 
8. Ecology of the habitat. Since the.time of Humboldt, there have been 
desultory attempts to determine the physical factors of habitats with some 
degree of accuracy. The first real achievement in this line was in the 
measurement of light values by Wiesner in 1896. In 1808 the writer first 
began to study the structure of habitats by the determination of water- 
content, light, humidity, temperature, wind, etc., by means of instruments. 
These methods were used by one of his pupils, Thornber (1901), in the 
study of a particular formation, and by another, Hedgcock (1902), in a 
critical investigation of water-content. Two years later, similar methods 
ef measuring physical factors were put into operation in connection with 
experimental evolution under control in the plant house. E. S. Clements 
(1905), as already indicated, has made the use of factor instruments the 
foundation of a detailed study of polydemic species, i. e., those which grow 
in two or more habitats, and which are, indeed, the most perfect of all ex- 
periments in the production of new forms. In a volume in preparation 
upon the mountain vegetation of Colorado, the writer has brought the use 
of physical factor instruments to a logical conclusion, and has made the 
study of the habitat the basis of the whole work. Out of this investigation 
has come a new concept of vegetation (Clements 1904), namely, that it is 
to be regarded as a complex organism with structures and with functions 
susceptible of exact methods of study. 
