32 VICTOR E. SHELFORD. 



The study of factors governing distribution of fish and other 

 animals has never been reduced to an adequate working basis. 

 The problems are indeed complex, but the difficulty has arisen 

 in part from two causes, namely, (A) the lack of knowledge of 

 the activity which takes place within the- narrowest limits (Shel- 

 ford, 'ii 3 ), and (B} lack of recognition of the important factors 

 and features of the environment. 



The conclusions of workers on distribution often seem to have 

 been to the effect that the food relations of fishes should stand 

 as first in importance, as factors of distribution. Hankinson 

 ('io) states that the pond weed zone, the living and feeding place 

 of the fish of \Yalnu t Lake, is probably the most important 

 habitat. Our evidence on the same species points clearly to 

 the breeding grounds. Indeed much careful work must be done 

 before broad generalization should follow, but it is evident that 

 here as in birds (Merriam, '90; Adams, '08) and in the ti^'-r 

 beetles (Shelford, '07, 'ii 3 ) the breeding place and the breeding 

 activities are the most important. (Reighard, 'io, and cita- 

 tions.) Is variation in nest building real or only apparent 

 because we do not know the most important factors and seize 

 upon details wholly unessential to fish? What are the la\\> 

 governing the mores of species? Experimental work correlated 

 with field observations can answer these questions, and it is at 

 this point that contributions of lasting value can be made. The 

 first step in the necessary work of raising natural history from its 

 present state of vagary is to determine what activity takes place 

 within narrowest limits and which is least modifiable in as many 

 groups of animals as possible. 



The second difficulty lack of recognition of the important and 

 unimportant in en\ inmments is one which we have emphasized 

 before. 



The ecologist often uses vegetation as an index of conditions. 

 There is objection to this. Investigators have seen that the same 

 species of animals are not always associated \\itli a given species 

 of plant. Indeed, species of plants c.innot often and perhap> 

 usually be taken as an index, of the environmental conditions of 

 animals, especially in water, because species ol plants an- not 

 necessarily an index of conditions. Tin- physiological condition 



