156 OVE PAULSEN 



tion of more or less xerophytic herbs, especially hemicryptophytes 

 growing in a soil containing humus — but no sulphides or chlo- 

 rides, and where the moisture is greater in the upper layers. 



This definition which is nearly in agreement with that of 

 Schimper and of Tanfiljew, applies to the Great Plains in Amer- 

 ica, as we saw them at Akron, Colorado. 



As deserts I would term plant formations the soil of which 

 is devoid of humus or very poor in it and contains often sul- 

 phates or chlorides. The formations are very open and they 

 frequently include trees and shrubs, which together with long- 

 lived herbs are strongly adapted to drought. Spring-plants 

 are numerous. 



Desert thus defined includes, presumably, all that in Amer- 

 ica is called desert, from the sage brush plains to salt flats, sand 

 deserts and the shrub and cactus desert in Arizona. 



It is indeed a wide range for one term to cover, too wide for 

 a scientific term. It might be regarded as a provisional term, 

 and, as the different forms of desert become successively prop- 

 erly known and described, experience and comparison might 

 teach one how to distinguish and how to term them scientifically. 



The comparison could be made by aid of the system of growth- 

 forms established by Raunkiaer. This distinction of growth- 

 forms is based upon the degree of protection afforded to the 

 buds which survive the unfavourable season of the year, this 

 being the season which limits the distribution of plants. 



Raunkiaer's system enables us to express statistical data 

 for the different regions which can then be numerically com- 

 pared by their "biological spectra." By this term is meant 

 the percentage of the different growth-forms (biological types) 

 of the region in question. By the making of a biological spec- 

 trum all species, rare or common, are alike in importance. On 

 the other hand, in studying plant formations, Raunkiaer (1909) 

 has made up a method by which firstly the species composing the 

 formation are given a character in numerals indicating their 

 frequency, and secondly the growth-forms get their statistics 

 representing the role they play in the plant-society. 



Had there been time enough the writer could perhaps have 



