16 TRANSACTIONS OF ROYAL SCOTTISH ARBORICULTURAL SOCIETY. 
them gave results within ro per cent. of the mean, while 96 per 
cent. of them came within 25 per cent. of the mean. 
It is well to state here that in all these researches with which 
I have been dealing, only absolutely bad stuff was rejected for 
testing ; no attempt was made to pick specially good specimens. 
One important result brought out in Bauschinger’s research, 
and confirmed in that of Johnson, was that the mechanical 
strength of timber is considerably affected by the ratio of the 
summer, or solid, growth to the spring, or open, growth in each 
annual ring; or, in other words, that the density or specific 
gravity is a very important factor. It was also proved, in this 
investigation, that there was a definite relation between the 
modulus of elasticity, as determined in the bending tests, and the 
crushing strength and bending strength, as determined in tests of 
the same timber. 
SHEAR TESTS. 
In testing timber beams, I frequently find that the beam gives 
way, not by tearing across, but by shearing along the neutral 
axis, and other experimenters have found exactly the same 
condition of things. 
For example, in Professor Lanza’s tests of beams, no less 
than eleven gave way in this fashion, the shear stress at the 
time when the rupture took place along the neutral plane being, 
in the case of spruce, only 191 lbs. per square inch, and in the 
case of yellow pine, 248 lbs. per square inch, and, by calculation, 
he showed that this was also about the intensity of the shearing 
stress, in a vertical plane, in the case of those beams which did | 
give way by actually tearing across. There has not been a 
very large number of direct shearing tests. 
The round tensile specimens which I prepared for this lecture, 
with intentionally small heads, act very satisfactorily as shear 
specimens, and the results given in Tables II. and III. give very 
accurately the shear strength of these different kinds of wood. 
In the Watertown experiments, the results obtained in direct 
shear tests were somewhat higher than those in the case of the 
beams which gave way by shear; this effect is due largely to the 
following cause—in a special shear test one fixes, by the form of 
the specimen, the particular plane at which shear shall take 
place; on the other hand, in the case of the beam, the shear 
takes place along the plane weakest to resist shear near to the 
neutral axis, and therefore one gets a lower shear strength. 
