METHODS OF STUDY 323 
control is left undisturbed. The study of reactions to contact has been 
carried on by the use of pans described in connection with light and tem- 
perature and with the use of mica chips, leaves, etc. 
In all experiments the containers are divided into several divisions 
and the number of animals noted in each division counted at each 
reading. About ten readings are taken, the number being determined 
by the number of animals used, which is determined by the number that 
can be observed before they can move any considerable distance. This 
is a function of the speed of movement, which also determines the fre- 
quency of reading. Readings should be taken at such intervals as to 
enable the animals to completely adjust their positions with reference 
to the conditions in the interim. 
The most effective method of study is that of mixing animals of differ- 
ent habitats; this removes the necessity of accurate measurement for 
rough comparison. The degree of accuracy of such experiments is 
determined almost entirely by the ingenuity and care exercised by the 
experimentor. Accuracy of measurement can be acquired, but in the 
case of some factors, such as light, with some difficulty. Such accuracy 
should, however, be the constant aim of the worker. 
While a high degree of accuracy may be attained in the field in the 
case of some factors and reactions, it is, in other cases, necessary to 
perform experiments in the laboratory also. As a rule all experiments 
should be performed in both field and laboratory. 
(b) To determine the most important activities: The first step in 
field observation is the continuous watching of animals throughout a 
number of life cycles. Experimentation is almost always necessary also. 
It is only under unusually favorable conditions that the relative impor- 
tance of the various periods of the life history of an animal can be 
ascertained without experimentation. On the other hand, experimen- 
tation must be correlated with field observation. Simple experimenta- 
tion on the behavior of animals in the laboratory does not illuminate 
this matter to any appreciable extent. 
To determine the habitat preference of animals, they should be placed 
in cages, In which they find several different sets of natural conditions, 
and the selection made by the animal noted. ; 
METHODS OF TAKING A CENSUS 
Species are of importance because each usually has a physiological 
makeup and habitat preference differing from other species. To make a 
census of the animals present in a given habitat it is necessary to visit 
