STRENGTH AND TESTING OF TIMBER 317 
Curly-grained timber, as a rule, gives much higher results 
than straight-grained timber in shearing tests parallel 
to the grain. 
When the load is applied to a portion oi the tested 
specimen, as is usual in practice, the strength in cross 
compression is, on the average, 12 per cent, higher than 
when the load is applied over the whole surface. 
In the Western Australian timber tests it was found that 
the strength of beams cut on the " quarter," that is, radial 
to the circumference, was 12 per cent, less than that of 
those cut in the ordinary manner. 
As a general rule, the weij^ht and density of seasoned 
timber is the measure of its strength, the heaviest timbers, 
even those of the same species, being the strongest in com- 
pression and bending tests ; but density is no criterion of 
tensile strength, and some comparatively light timbers have 
great tensile stress, as, for instance, ash and hickory. 
The weight of timber is very uncertain and very puzzling, 
and doubtless the great variation in the weights of timber, 
as given by different people, is mainly due to the greater or 
lesser amount of moisture in the timber ; and to estimate 
weights by small pieces is very uncertain, each piece of a 
log or tree being of a different weight to an adjoining 
piece, yet one often finds the weight of timber given to two 
places of decimals. 
The weights per cubic foot given in this book are for 
well-seasoned wood. 
The number of rings per inch have no bearing on the 
weight as a rule, nor do they influence strength. Although 
a piece of pine or fir timber with, say, sixteen rings to the 
inch might weigh more than one with, say, only six rings, 
yet, in quite a number of cases, one finds the reverse, and 
in the author's experience one with thirty-four rings to the 
inch weighed less per cubic foot than one with twenty, 
