ECOLOGICAL SUCCESSION. 363 



full accord and we might conclude that there are no relations 

 between primary ecological characters and taxonomic characters. 

 Such however can hardly be strictly true, but we cannot see what 

 the real relations may be. If our point of view of ecology is 

 correct the ecological characters of a race experimentally segre- 

 gated, or experimentally produced must in practice consist 

 primarily of reaction to physical factors or combinations of physical 

 factors or to entire environmental complexes, secondly of a definite 

 rate of metabolism, time of appearance or the like, thirdly of 

 specificity of behavior, and fourthly of structural characters 

 modifying behavior (see p. 339, order that of expediency). Rela- 

 tively fixed taxonomic integumentary characters have no bearing 

 on ecological matters, not even according to the broadest defini- 

 tions of the subject. The characters which are not related to 

 the environment, and which are of no ecological value, are the 

 ones quite generally used in breeding work, specificity of behavior 

 standing second, and plastic structure third, primary ecological 

 matters usually receiving no adequate attention or only such attention 

 as comes incidentally with the handling of the material; the results 

 consisting of noted differences in reaction to light of doubtful 

 intensity and quality, or similar temperature differences, etc. 

 The testing of primary ecological characters can be adequately 

 conducted in three ways. First, by the measurement of reaction 

 to all or several of the chief environmental factors under rigidly 

 controlled conditions, with each factor accurately measured 

 qualitatively and quantitatively, and with the measurements 

 of the speed or intensity of the reaction repeatedly determined. 

 Second, by testing the reaction of the animals to a graded environ- 

 mental complex of known constitution, and third, by putting the 

 animals out into a graded series of natural environments selected 

 with due reference to the species in question. Being easily open 

 to experiment, the question of the relation of taxonomic and 

 ecological characters should be left for experimental studies to 

 answer. 



The relation of habitat preference to the so-called structural 

 adaptations and to their origin is, as we have seen, not intimate, 

 and the method of experimental attack less obvious than in the 



