186 JOURNAL OF FORESTRY 



That somewhat similar conditions prevail in the case of the 

 tracheids, fiber-tracheids, and libriform-fibers of arborescent and large 

 fruiticose Dicotyledons, is indicated by the investigations of Sanio/' 

 and Prichard and Bailey^° (see curve B, fig. 24). Furthermore, the 

 curve in figure 25 suggests that the height of the rays varies in the 

 conifers much as does the size of the tracheary elements. 



These curves illustrate, in a diagrammatic and quantitative manner, 

 what has been emphasized earlier in this article, that the anatomical 

 structure of wood is not constant even in different parts of a single 

 tree. Furthermore, it indicates very clearly that diagnostic criteria, 

 secured from a study of herbarium specimens or the wood of small 

 stems, are lakely to prove to be unreliable in classifying and dis- 

 tinguishing the mature wood of large trees. 



SUMMARY AND CONCLUSIONS 



1. There has been a marked tendency among those who have 

 advocated the use of minute anatomical characters in the classification 

 and identification of wood, on the one hand, to overestimate the possible 

 economic applications of such diagnostic criteria, and, on the other 

 hand, to greatly underestimate the variability of anatomical structures. 



2. The fact that the average lumberman and tradesman has to 

 handle a large amount of material in a comparatively limited space 

 of time, eliminates the use of any except the most obvious anatomical 

 characters. 



3. There are, however, certain important, although somewhat re- 

 stricted, economic fields of usefulness for very accurate and reliable 

 keys in the hands of technically trained experts. 



4. The inaccuracies in existing systems of classifying woods are 

 largely due to the fact that investigators have not studied the limits 

 of variability of anatomical characters, but have assumed that their 

 diagnostic criteria are constant and comparatively invariable. 



5. A careful study of some of the supposedly more reliable diag- 

 nostic criteria, such as the distribution of wood-parenchyma, form and 

 structure of the rays, type of pitting, etc., indicates very clearly that 

 these characters may fluctuate considerably, not only in certain families, 

 genera, and species, but also in different parts of a single tree. 



'Sanio, Karl. Anatomic der gemeinen Kiefer (Pinus silvestris L) II. Tahrb. 

 wiss. Bot., Vol. IX, pp. 50-126. 



10 Prichard, R. P., and Bailey, I. W. The significance of certain variations in the 

 anatomical structure of wood. For. Quar. Ined. 



