INTRODUCTION 7 



the study of a particular group of woods was not sufficiently 

 exhaustive to permit of final conclusions, and that in many 

 cases the diagnoses were not drawn with sufficient attention 

 to strict accuracy of statement and that regard for exhaustive 

 detail which would render them of the greatest value. In deal- 

 ing with the structure of the wood for taxonomic purposes it 

 has been found that the diagnosis must take cognizance of a 

 wide range of detail, and that it must be most searching in 

 its character. This is a necessary result of the high degree of 

 development of the organisms and the consequent differentia- 

 tion of the structure along several lines of development. It is a 

 neglect of this fact in the past which appears to explain why 

 previous investigators have failed to construct a system of classi- 

 fication which would not only give some additional information 

 respecting phylogeny, but which would at the same time permit 

 of a satisfactory recognition of species and genera. Under these 

 circumstances it appeared desirable to commence de novo and, 

 in the first instance, make an exhaustive study of the anatomy 

 of the wood, utilizing for purposes of classification such facts 

 as might be obtained in this way. As a convenient starting 

 point for the discussion of relationships it was considered that 

 none could be secured which would be better adapted to the 

 purpose in view than the classification generally employed as 

 based upon the external morphology of the vegetative organs 

 and inflorescence. For this purpose I at first selected the then 

 most recent and authoritative compilation relating to the North 

 American Coniferales, as embodied in Professor Sargent's valu- 

 able work on the woods of North America in the Tenth Census 

 of the United States. To this were later added Sargent's Silva 

 of North America and the sequence of Engler and Prantl con- 

 tained in the Natiirlichen Pflanzenfamilien, as expressing the 

 latest views on the subject. 



Assuming the typical character of the trees discussed by 

 Sargent, it was held that any well-authenticated specimen of 

 wood from any such tree would also be typical, and in this way 

 it would be possible to form a type series the structure of which 



