PRINCIPLES OF CLASSIFICATION. 1» 



these climatic factors varies enough from one place to 

 another to overshadow the more obvious effects of soil, 

 ground-water, etc. (The differences between the vege- 

 tation of the northeastern and southwestern parts of the 

 state are indeed doubtless due in part to temprature, but 

 there are no differences in vegetation in Alabama that 

 can be reasonably ascribed to differences in average an- 

 nual rainfall. 



Worse still, climatic factors, except in a few special 

 localities like the summits of high mountain ranges, vary 

 gradually from place to place, so that the location of lines 

 based on any one of their numerous functions, such as 

 average, maximum and minimum temperature, length of 

 growing season or period between frosts, and seasonal 

 variations of rainfall, is in general wholly arbitrary, and 

 likely to be influenced largely by the scale used (whether 

 Fahrenheit or Centigrade in the case of temperature, or 

 inches or millimeters in the case of rainfall).* 



Almost equally worthless for our purposes are those 

 environmental factors which vary greatly in short dis- 

 tances, such as light (governed by slope of ground, den- 

 sity of forests, etc.), evaporation (governed largely by 

 the same factors), and soil moisture (governed largely 

 by topography). Such factors are verj' useful for dis- 

 tinguishing local forest types, such as swamps, ham- 



*But for the difficulty \iientioned in this paragraph it might be 

 worth while to give some consideration to the seasonal distribu- 

 tion of rainfall, which varies perceptibly in different parts of our 

 state. This is well illustrated by the three rain maps between 

 pages 176 and 177 of Dr. Smith's report on the agricultural feat- 

 ures of Alabama (Smith 7 in bibliography). The lines on the last 

 map, showing annual rainfall, do not correspond very closely with 

 the known distribution of any trees (or anything else, appar- 

 ently), but the other two maps show that there is a general corre- 

 spondence between hardwood forests and regions of heavy winter 

 rains (December to February), and between the principal long-leaf 

 pine area and heavy summer rains (June to August). But the 

 distribution of forest types can be correlated much more staisfac- 

 toriiy with soil, which seems to be much more closely connected 

 with geology than with climate; and it is possible that the rela- 

 tive proportion of pines and hardwoods in the forests has some 

 influence itself on the seasonal distribution of rainfall, whose 

 irregular distribution on the map would be difficult to account for 

 otherwise. (In this connection see footnotes by the writer in Bull. 

 Torrey Bot. Club 37:415-416. 1910; Torreya 12:146. 1912.) 



