396 CARNEGIE INSTITUTION OF WASHINGTON. 



of ludoviciana had greatly increased the hairiness of the upper surface 

 and resembled gnaphalodes, but that the latter had undergone little 

 change. The taxonomic value of this lies not only in the demonstrated 

 change of one form into another, but especially in the fact that the 

 tomentum is shown to be more easily acquired than lost. Moreover, 

 the results indicate that ludoviciana is the primitive form and gnapha- 

 lodes the derived one, and suggest that this is true in similar paired 

 forms in Artemisia. 



The Phylogenetic Method in Taxonomy, hy F. E. Clements and H. M. Hall. 



In emphasizing phylogeny as the basis of taxonomy, it has been 

 found desirable to prepare monographic accounts of three complex 

 genera, Artemisia, Atriplex, and Chrysothammis, which are to be fol- 

 lowed by others as rapidly as possible. In addition to its value in in- 

 dicating methods of taxonomic research, each monograph will serve a 

 useful purpose by presenting a logical and concise classification of the 

 species, varieties, and minor variations comprised within the genus. 

 In the selection of genera to be treated attention has therefore been 

 given to economic importance as well as suitability for illustrating 

 taxonomic principles. The genera mentioned are of the first impor- 

 tance as indicators of soil and climatic conditions and therefore as 

 guides in agricultural practice. Many species of Artemisia and 

 Atriplex possess practical value to the stock-raising industry in the 

 West, and others are undergoing investigation as possible sources of 

 oils, potash, etc. Chrysothamnus is especially important as a source of 

 native rubber. An exact classification of the species of Artemisia and 

 Atriplex based upon relationship is coming to be demanded by physi- 

 cians and pharmacologists because of their importance as causes of 

 hay-fever and the value of their pollen extracts in treating this malady. 



The number of so-called species in each of these genera is so great 

 and their characterization so vague in many cases that even the 

 trained taxonomist finds difficulty in identifying them from floras 

 and manuals, while others, such as pathologists, foresters, grazing 

 experts, physicians, etc., either abandon the task or are led into serious 

 errors in identification. The need of thorough revisions is still more 

 urgent because of the almost total lack of information as to the rela- 

 tionships of the species as proposed by descriptive botanists. In the 

 attempt to arrange the numerous segregated forms into a system, it 

 has been found that they can be assembled into a comparatively small 

 number of species. Thus, the 160 so-called species of Artemisia can 

 be organized into 28 actual species, each of which is so natural and so 

 sharply defined from its neighbors that its recognition becomes possible 

 even for the non-taxonomist. Similarly, the 100 or more segregates 

 of Chrysothamnus comprise in reality but 12 species, and the genus 

 Atriplex, commonly supposed to include 120 to 200 species, yields 



