TESTING WOOD. 131 



holding and bringing strains to bear upon large scantlings. 

 Pieces less than a quarter of an inch square were often used. 

 Such pieces might give an unduly unfavourable result from the 

 cutting across of individual fibres ; or, on the other hand, being 

 freer from knots or other defects, more readily seasoned through- 

 out and more homogeneous, they are rather picked than average 

 samples, and may give an unduly favourable result. 



Density. We can only give here a bare outline of the prin- 

 ciples, methods, and results of testing. Much, as we have seen, 

 depends upon density ; and, admitting that, owing to air or 

 moisture in the wood, the results are not as satisfactory as could 

 be wished, we have two simple methods, described by Professor 

 Unwin, for determining this character, viz. (i) by measuring and 

 weighing planed rectangular blocks, and (ii) by weighing the 

 block and the water it displaces. In the former method, if 

 b = the breadth, t = the thickness, and h = the height of the block in 



inches, its volume will = *wj~ cubic feet ; and if W = its weight 



in pounds, the heaviness of the wood per cubic foot will 

 1728W 



= -M~ lbs - 



In the second method, if W = the weight of the block in 

 pounds, W' = the weight in pounds of the water it displaces, 

 and G = the weight of a cubic foot of water, i.e. 62'4 Ibs. at the 



normal temperature, then -^- = the volume of the block in 



cubic feet, and the heaviness of the wood per cubic foot will 



, GW 

 **^' 



Mr. Stephen P. Sharpies, who made the examinations of North 

 American timbers incorporated by Prof. Sargent in the Ninth 

 Census of the United States, vol. ix. (1880), making at least two 

 determinations for each species, calculated the specific gravity by 

 measurement with micrometer calipers and weighing. The 

 specimens tested were 100 millimetres long and about 35 milli- 

 metres square, and were dried at 100 C. until they ceased to lose 



