490 CORRELATION OF DISTRIBUTIONAL FEATURES. 



In order to make a thoroughgoing and entirely satisfactory investi- 

 gation into the nature of the conditions which Hmit a plant at the 

 various portions of its distributional edge, we should ascertain the 

 entire complex of conditions for numerous localities along this edge. 

 We should know at the outset the amplitude of each condition for 

 the entire range of the plant or vegetation, and should know the 

 localities (usually on the edge of the geographic range) at which the 

 extreme values are encountered. For each station at which a maximum 

 or minimum value of any condition was found we should then ascertain 

 the enth'e constellation of other conditions. Since no extreme condi- 

 tion operates in any other way than in conjunction with the associated 

 conditions of the same locality, we might thus be able to ascertain the 

 controlling complex of conditions for the given locality, and, in turn, 

 for all other localities in the edge of the distributional area. But the 

 problems with which we are dealing are too new, and the available 

 quantitative data and precise information pertaining to them are too 

 limited at present, to warrant serious attempts to pass beyond the 

 limits of very general considerations. 



II. COMPARATIVE CLIMATIC FEATURES OF THE NINE GENERAL 



VEGETATIONAL AREAS. 



Before considering the complexes or constellations of climatic condi- 

 tions which characterize each of our various vegetational areas, it will 

 be instructive to see how the extreme values of several of the leading 

 features of the climate compare in these areas. Our later discussions 

 will concern the whole climatic character of each botanical area, 

 whereas we now wish to compare each climatic element singly and as 

 differing in intensity among the nine general vegetational areas. For 

 this purpose 12 of the climatic charts have been selected (plates 34, 

 35, 36, 40, 43, 46, 52, 53, 59, 65, 69, and 72), and their climatic dimen- 

 sions are presented in figures 21 to 26. These charts will now be con- 

 sidered in order. 



Number of days in normal frostless season (plate 34, fig. 21). — A com- 

 parison of the blocks in this graph shows that the longest frostless 

 seasons are found in the Desert and the western section of the Northern 

 Mesophytic Evergreen Forest, and that nearly as short a season is 

 found in the Grassland. The first two of these have maximum values 

 which nearly coincide, meaning that there is almost exactly the same 

 amplitude^ in the length of the frostless season in these two very dis- 



* The word "amplitude" is used throughout the succeeding pages to express the degree of dis- 

 similarity between the values of a climatic condition in the different parts of a botanical area. 

 Numerically it is the difference between the maximum and the minimum value for the area. This 

 is done in order to avoid the use of the word "variation," which might be taken to indicate the 

 seasonal differences, or the differences from year to year, at the same climatological station. 

 We are not here concerned with the seasonal or annual march of any of the climatic condi- 

 tions, nor with any of their other variations, but solely with the differences which their index 

 values exhibit from place to place, and with the broad or narrow amplitude of these differences in 

 given areas. 



