150 THE PLANT 
graphic record of the course of variation. When the latter is made evident 
in structural feature also, histological slides are an invaluable part of the 
record. 
Polydemic species are by far the best and most frequent of all natural 
experiments. In addition to plants that are strictly polydemic, i. e., -grow 
in two or more distinct habitats, there are a large number which occur in 
physically different parts of the same habitat. The recognition of polydemics 
is the simplest of.tasks. As a rule, it requires merely a careful examination 
of contiguous formations in order to ascertain the species common to two 
or more of them. The latter are naturally most abundant along the eco- 
tones between the habitats, and, as a result, transition areas and mixed 
formations are almost inexhaustible sources of ecads. Many adaptable 
species are found throughout several formations, however, and such are 
experiments of the greatest possible value. Not infrequently species of thé 
manuals are seen to be ecads, in spite of their systematic treatment, and to 
constitute natural experiments that can be readily followed. Finally, it 
must be kept in mind that some polydemics are stable, and,do not give rise 
to ecads by structural adaptation. They not only constitute extremely in- 
teresting experiments in themselves, but they should also be very carefully 
followed year by year, since it seems probable that the responses are merely 
latent, and that they will appear suddenly in the form of mutants. In 
natural experiments it is sometimes difficult to distinguish which form is the 
ecad and which the original form of the species. As a rule, however, this 
point can be determined by the relative abundance and the distribution, but 
in cases of serious doubt, it is necessary to appeal to experimental cultures. 
Although habitats differ more or less with respect to all their factors, the 
study of polydemics needs to take into account only the direct factors, 
water-ccntent, humidity, and light. Humidity as a highly variable factor 
plays a secondary part, and in consequence the search for ecads may be 
entirely confined to those habitats that show efficient differences in the 
amount of water-content or of light. Temperature, wind, etc., do not pro- 
duce ecads, and may be ignored, except in so far as they affect tle direct 
factors. Complexes of factors, such as altitude, slope, and exposure, are 
likewise effective oniy through the action of the component simple factors 
upon water and light. The influence of biotic factors is so remote as to 
be negligible, especially in view of the fact that ecads are necessarily 
favorable adaptations, and are in consequence little subject to selective 
agencies. The essential test of a habitat is the production of a distinguish- 
able ecad, byt a knowledge of the water-content and light values of the 
habitats under examination is a material aid, since a minute search of each 
formation is necessary: to reveal all the ecads. It is evident that habitats or 
