40 BASES AND CRITERIA. 



the controlling factors of the two during a term of years, but this difference 

 is often less than that shown by the grassland climate in the dry and wet 

 phases of the same climatic cycle. The rainfall of the wet phase if continued 

 for a century or two under natural conditions would turn the prairie into 

 forest, that of the driest period would under the same conditions convert it 

 into desert. Similarly the distribution of rainfall is so erratic that two con- 

 tiguous locaUties may show striking differences amounting to the success or 

 failure of a particular crop. Progressive changes of rainfall, temperature, and 

 evaporation occur with increasing altitude, latitude, and longitude. Further, 

 each climate shades imperceptibly into the next, often through wide stretches. 

 These are all elementary facts and the climatologist might well say that they 

 are taken account of in the ordinary way of determining means or normals. 

 As a matter of climatology this is true, but from the standpoint of indicator 

 vegetation it is not. It is a simple matter to trace the line of 20 inches of 

 rainfall, or of the 60 per cent ratio of rainfall to evaporation and to assume 

 that it marks the line between prairies and plains. Such an assumption 

 reverses the proper procedure, in which the associations themselves must be 

 permitted to indicate their respective climates. When this has been done 

 and the limits of the various communities established, it will be possible to 

 determine the correlated factors. 



The real importance of climatic variations within a chmax habitat lies in 

 the fact that the correlations of vegetation and climate must be studied on 

 the spot year by year. No single station can be typical of the whole habitat, 

 and no year of the whole cycle. Yet for each station and for each year the 

 indicator evidences of the vegetation should correspond closely if not exactly 

 with the controlling factors. As a result, the study of representative localities 

 for each year throughout a climatic cycle should disclose the range of fluctua- 

 tion in both cUmax habitat and vegetation, and establish all the indicator 

 values of the latter upon a secure basis. 



The minute study of habitats reveals differences which are reflected in the 

 behavior of plant and community, and hence cause the latter to serve as 

 indicators. It is probable that every square foot of a habitat differs in some 

 degree from every other one. Moreover, when the reactions of competing 

 plants are taken into account, the differences are often more minute. In 

 natural studies of competition made in Colorado and in Cahfornia, as well as 

 in competition cultures, differences of height and flowering have been found 

 for each inch or two. Corresponding differences of density are of even more 

 frequent occurence in herbaceous communities. These indications have been 

 checked by factor determinations only in a few cases as yet, but there can be 

 little question that many more habitats show the most minute differences, 

 each with the corresponding indication in terms of density, height, reproduc- 

 tion, etc. In short, the indicator correlation of plants and habitats exemplifies 

 a universal principle which applies from the relation between climax formation 

 and habitat through units of diminishing rank to the relation between the 

 individual plant and its miniature habitat. 



Inversion of factors. One of the early puzzles encountered in indicator 

 studies, especially in connection with succession, was the occurrence of the 

 same dominant in adjacent but diverse areas. This was first noted for Andro- 

 pogon scopariiLS and Calamovilfa longifolia in sandhill and badland regions. 

 These were found in rough areas and in blowouts on the one hand and in 



