308 THE FORMATION 
tions of water-content and light are readily produced in the field by drain- 
age, irrigation, shading, clearing, etc. In fact, all the changes of habitat 
indicated under experimental evolution serve equally well to initiate experi- 
ments in experimental vegetation; indeed, the same experiment covers both 
fields. It is impracticable, however, fo modify the temperature-of a habitat 
without changing its water-content or light, and consequently the influence 
of temperature can not be determined through experiment by modification. 
The extent of the area modified should be as large as convenience will per- 
mit, in order that the number of individuals may be large enough to indicate 
clearly the resulting adjustment in position and arrangement. The best re- 
sults can be obtained where a small separate area of a formation can be 
modified, e. g., where a small swamp can be drained, or a depression flooded. 
In the case of light, however, it is usually impossible to’clear or to shade a 
large area, and the study must be restricted to a relatively small group of 
plants. In regions where lumbering is actively carried on, the consequent 
clearing initiates invaluable experiments over large areas, and this is like- 
wise true of forest plantations. Modification of a large area has decided ad- 
vantages in bringing out the changes in the more prominent structural fea- 
tures, but the causes and the details of the adjustment can be worked out 
much more satisfactorily in a small area. 
359. Denuding. The modification of the habitat by denuding is the sole 
method of initiating succession by experiment. It is consequently of the 
most fundamental importance in investigating aggregation, ecesis, and com- 
petition, as well as the reactions exerted by the invaders of the different 
stages. The possibilities of denuding an entire habitat or an extensive 
area are not great, and the investigator must content himself with denuded 
quadrats, transects, and migration circles, which are small enough to permit 
a critical study of all the factors in succession. It is of course unnecessary 
that the denuding be done by the ecologist himself, provided he is able 
to follow the succession from the very beginning. Accordingly, it becomes 
possible for him to make the very best use of all those changes wrought 
by man in which the vegetation is destroyed over considerable areas. These 
are essentially natural experiments, and at this point the methods of natural, 
and artificial habitats merge. 
The manner of denuding depends in a degree upon the nature of vege- 
tation, but, when time, convenience, and safety are all taken into account, 
the actual removal of the vegetation as indicated under the denuded quadrat 
is by far the most satisfactory. Under certain conditions, flooding or burn- 
ing can be used to advantage, but cases of this kind are infrequent. The 
purpose of the experiment determines the kind of area to be denuded. 
