264 



FOREST TREES OF NORTH AMERICA. 



RELATION BETWEEN TRANSVEESE STRENGTH AND SPECIFIC GRAVITY, ETC. Continued. 



GENERAL EEMAEKS. 



An examination of the results obtained from the various tests made upon the woods of North America indicate 

 at least the important fact that within the limits of anj- species the weight and strength of any specimen of wood 

 depends upon the actual proportion of the space occupied in the layers of annual growth with open ducts to the 

 space occupied with compact, woody tissue, and to the size of these ducts; or in the case of the wood of 

 Coniferae, the proportion of space occupied with cells formed early in the season to that occupied with the smaller 

 cells of the summer growth. The proportion between these two kinds of growth varies not only in every individual 

 tree, but in different parts of the same tree. The causes which thus affect the growth of wood are not very 

 apparent. It is not soil, nor age, nor general climatic conditions, it appears, which produce the different proportion 

 between the solid and the light portions of the annual growth in any species, because in the same individual this 

 proportion is found to vary from year to year. It varies very irregularly ; nor does the rapidity of growth, as has 

 been supposed, greatly affect the strength of wood, because the proportion of open to compact growth is little 

 affected by rapid or slow increase of the tree's diameter. How far annual climatic variations affect the nature of 

 the annual layers of growth has not been demonstrated, although it is not impossible that in years in which 

 conditions favorable to rapid growth are extended late into the season, the proportion of the annual layer occupied 

 by open, weak growth to the growth of the whole year would be greater than that formed in a year during which 

 the season favorable for rapid growth was less extended. 



It follows that while such experiments as those conducted by Mr. Sharpies are necessary to establish 

 maximum and relative values for any species, these being established, actual values of any given specimen of 



