302 
TIMBEE 
do with the testing of this material know that the smaller 
the specimen the higher the result. Small sample pieces of 
timber are obtained in more perfect condition, have fewer 
flaws and knots, often no knots at all, while these imperfec- 
tions cannot be avoided in the case of timber of the sizes 
used in constructional work ; hence in most cases it has 
been assumed, daring recent years, that the results obtained 
from small test specimens should, when put into practice, 
be reduced by 50 per cent, or even more. 
It is a curious thing that in the case of timber, although 
one of the oldest constructive materials, and the one of 
which more is used than any other, we have been content 
to depend upon what are more or less doubtful tests, whilst 
in the case of steel, a material only a few years in use, we 
can rely upon accurate tests. It is not satisfactory, con- 
sidering the number of powerful testing appliances now in 
use, that the tests made on timber of large scantling are so 
very meagre. Although the movement in the direction of 
remedying this state of things is slow, it is gradually 
coming to be considered necessary that tests made on 
timber, to be of any value, should be upon pieces such as 
are used in actual work, and not upon small or specially 
prepared specimens. The late Professor J. B. Johnson 
initiated a valuable series of tests on timber for the United 
States Government. Professor Lanza also carried out a 
great number of useful tests on large-sized timber, chiefly 
white and yellow^ pine (pitch pine) and spruce, at the 
Watertown Arsenal, and the Massachusetts Institute of 
Technology, United States, which have thrown a good deal 
of practical light on the subject, and the United States 
Government are continuing these tests. Quite recently the 
Government of Western Australia has, under the direction 
of Mr. G. A. Julius, B.Sc, M.E., tested thousands of 
different-sized pieces of the hardwoods of this and other 
