Variations in Anatotnical Structure 669 



ences between Coniferae and Dicotyledons. Furthermore, the 

 fact that certain primitive types of arborescent Ranales, which do 

 not possess vessels, resemble the conifers in the structure, size, and 

 general variability of their tracheids, suggests that the shortening 

 of the xylem elements in Dicotyledons has been associated in some 

 manner with the evolution of vessels. 



The evidence at hand indicates, also, that the size of tracheids, 

 fiber-tracheids, libriform-fibers, and vessel-segments fluctuates 

 more or less during the later stages of the development of a tree, 

 and that the size of the elements in any given annual ring is not 

 constant, but varies at different heights in the stem. 



Associated with such variations in the size of the xylem elements 

 are concomitant changes in the shape, structure, and arrange- 

 ment of the elements, and their mechanical properties. 



B. The significance of these variations in the identification of the 

 timbers of commerce. 



The fact that xylem or wood is not a homogeneous material, 

 comparable to iron or steel, but is extremely variable even in dif- 

 ferent parts of a single tree, has not been fully appreciated by 

 most investigators who have interested themselves in the prob- 

 lems of the classification and identification of plant tissues. In 

 the majority of cases, a very limited amount of material, often a 

 single small specimen of each species, has been studied in detail 

 and used in constructing keys for distinguishing the woods of 

 different plants. Such keys are obviously subject to grave errors, 

 since the material of each species examined may give no indica- 

 tion of the structural variations which occur within the species. 

 For example, in the endeavor to secure authentically identified 

 Specimens, the investigator has frequently been led to refer to 

 herbarium material. This is a particularly dangerous proceed- 

 ing, since the wood of twigs is usually very different from that of 

 the outer layers of the bole of a large tree. 



It is evident, accordingly, that, in endeavoring to secure satis- 

 factory criteria for distinguishing different woods, the investigator 

 must examine sufficient material of each species to be certain of the 

 limits of variability of each diagnostic character. 



There appears to be good reason for believing that the study of 

 the variability of the different elements of the xylem in different 

 groups of plants, and of the factors which produce or control this 



