March 31, 1922] 



SCIENCE 



337 



tion by other factors than the plants are, and 

 it may be that the smaller divisions of eco- 

 logic communities will be different in ani- 

 mals and plants; and perhaps these smaller 

 communities will differ even in different groups 

 of animals. It will be well, therefore, for 

 field workers not to depend too rigidly on the 

 plants or on any other one factor in describ- 

 ing distribution. 



Indeed, it may sometimes happen that even 

 the faunal and floral areas, or the principal 

 ecologic associations, for the various groups 

 of organisms will differ. It certainly will not 

 be best to force unruly facts of distribution 

 to conform to any rigid system of description. 



On the other hand, a classification of habi- 

 tats and biotic areas will be of the greatest 

 use when it is applicable, so far as possible, 

 to all groups of animals and plants. For this 

 reason the ecologic communities recognized for 

 all organisms should coiTespond as nearly as 

 possible without obscuring the facts. 



To form a universal classification will re- 

 quire the establishment of more divisions than 

 would a classification for one group of or- 

 ganisms alone. For mammals, for instance, 

 the grouping of all the fresh-water environ- 

 ments of a faunal area into one nominal habi- 

 tat, the aquatic, would probably suffice; but 

 if fresh-water fishes, invertebrates, and plants 

 are to be considered, a number of habitats in 

 the water must be recognized. Even when 

 mammals alone are considered it can do no 

 karm to describe more than one aquatic habi- 

 tat, and it is of great advantage to have a 

 classification of wide application. 



One of the great advantages of using eco- 

 logic habitats and biotic areas for the state- 

 ment of distribution is that these units are not 

 founded on the assumption that any one parti- 

 cular factor of the environment is most im- 

 portant in the limitation of distribution. Units 

 of distributional classification based on a bias 

 for some one particular factor, such as tem- 

 perature, as being most important in the con- 

 trol of distribution can not have the confidence 

 of persons who consider the basis of classifica- 

 tion unsound, or at least unproved. But the 

 facts of distribution can be described by the 

 use of biotas and ecologic communities with- 



out an assumption that any one factor is all 

 important. However, if one factor is actually 

 the most important one in the control of dis- 

 tribution over any area, this relation is not 

 obscured by the employment of the units of 

 description suggested. 



It is my opinion that we are not as yet 

 sufficiently informed as to the exact distribu- 

 tion of any group of animals or plants to 

 render possible anything more than a pre- 

 liminary classification of distribution in any 

 part of the world. I would emphasize, there- 

 fore, the need for the piecise statement of dis- 

 tribution in terms of units which are capable 

 of combination into a number of possible sys- 

 tems of classification, rather than to describe 

 distribution in terms of large and relatively 

 unstable biogeographical regions, life-zones, 

 or ecologic formations. 



With a knowledge of the biotas and eco- 

 logic communities of the world it will be 

 an easy matter to compare floras and faunas 

 of different geographic regions; or the com- 

 munities of similar habitats of different biotic 

 areas may be compared as desired. Zoogeog- 

 raphers and phytogeographers may, if they 

 wish, combine biotic areas to form provinces, 

 regions, or life-zones; and communities may 

 be combined at pleasure by the ecologists to 

 make fonnations or other large divisions. 



The time has come in the study of the factors 

 limiting distribution when little more progress 

 can be made by statistical methods, the attempt 

 to correlate the distribution of climate or other 

 barriers to distribution and of groups of ani- 

 mals and plants in the mass. Eather we must 

 critically determine the factors concerned in 

 the distribution of individual species. To do 

 this will require carefully controlled experi- 

 ments in the laboratory, coiTelated with long 

 continued measurement and observation of the 

 physical and biological factors of the natural 

 environments. 



Before laboratory experiments can be ef- 

 ficiently carried out, however, we must know 

 the exact distribution in nature of the species 

 of a nim als and plants and their environments. 

 This is the greatest need of biogeography at 

 the present time: to describe biotic areas and 

 habitats and to determine the precise habitat 



