138 



OF WOOD IN GENERAL. 



grown, the age of the tree, the part of the tree from which the 

 timber is taken, and the extent to which it has been seasoned, 

 will all modify the results. Thus Bauschinger showed that 

 strength varies according to the proportion of summer to spring 

 wood, and that the centre of a tree is therefore weaker; whilst 

 the following table of the range of variation in 26 trees of Pinus 

 paliistris, quoted by Professor Unwin from a Report of the U.S. 

 Department of Agriculture, shows how butt, middle, and top 

 logs differ in strength, largely no doubt for the reason, which we 

 explained in a previous chapter, that the annual increments of 

 wood forming cones do not extend uniformly from end to end of 

 a log. [See pp. 79-80 and Fig. 41.] 



As to seasoning, since timber loses from \ to i or, when per- 

 fectly dried, J of its weight in the process, and strength and the 

 co-efficient of elasticity vary directly with density, its effect is 

 obvious. 



Unfortunately the systems employed for stating the results of 

 experiments vary almost as much as the dimensions of the 

 specimens tested, so that it is a matter of considerable calculation 

 to compare the records of different experimenters. Mr. Sharpies, 

 for instance, defines the co-efficient of elasticity, or rather of 

 longitudinal extensibility, as the weight in kilograms sufficient to 

 elongate a stick one centimetre square to double its original 

 length, were that possible, and states results ranging from 25,699 

 in Ficus aurea to 165,810 in the Western Tamarack (Ldrix 

 occidentdlis). To translate his results into the tons per square 



