HABITATS 41 



accumulation of such data, it is possible to build up 

 circumstantial evidence as to the types of environ- 

 ment in which each species can live, and the impor- 

 tant limiting factors controlling the numbers and 

 distribution. For many species we have vague data, 

 such as that they hve in a dry habitat or are abundant 

 in dry years and scarce in wet years. Ecological 

 work tries to estabUsh exact answers to the questions : 

 at what season does dryness act, does it act directly, 

 or indirectly through vegetation, or is dryness just 

 an indicator of some other environmental factor such 

 as temperature, how dry or how hot and for how 

 long ? The circumstantial evidence obtained by 

 correlation of animal communities or species with 

 descriptions of their habitats is of great value in 

 pointing the way, narrowing the field of inquiry 

 for experimental work, which now has to be con- 

 sidered. 



Most experimental work on physical factors has 

 been concerned with methods and technique — still 

 in an embryonic stage in many instances — and with 

 one-factor experiments. This work, probably owing 

 to the technical difficulties which have still to be 

 solved, but also to other difficulties which will be 

 mentioned, has as yet produced few^ theoretical 

 generahzations of great value in ecologj^, although 

 the practical uses of such work are often considerable. 

 Earher experimenters concerned themselves with dis- 

 covering which factors in the environment were of 

 importance in animal life, and with methods for their 

 measurement. On land we have to measure and 

 control for experimental purposes hght (intensity, 

 quality, length of day), heat (the temperature, and 

 the total ' heat budget ' in a given time), moisture 

 (the relative humidity, saturation deficiency, rain- 

 fall, etc.), other physical factors such as soil, sub- 

 stratum, shelter, etc. In fresh water and the sea 

 there are parallel but often different methods of 

 measuring light and heat, also methods for sampling 



