. 17, No. 4.) It is
characteristic of a brittle wood which gives way suddenly
without warning, like a piece of chalk. In this case the surface
of fracture is described as brash.

~Compression failure~ (see Fig. 17, No. 5) has few variations
except that it appears at various distances from the neutral
plane of the beam. It is very common in green timbers. The
compressive stress parallel to the fibres causes them to buckle
or bend as in an endwise compressive test. This action usually
begins on the top side shortly after the elastic limit is
reached and extends downward, sometimes almost reaching the
neutral plane before complete failure occurs. Frequently two or
more failures develop at about the same time.

~Horizontal shear failure,~ in which the upper and lower
portions of the beam slide along each other for a portion of
their length either at one or at both ends (see Fig. 17, No. 6),
is fairly common in air-dry material and in green material when
the ratio of the height of the beam to the span is relatively
large. It is not common in small clear specimens. It is often
due to shake or season checks, common in large timbers, which
reduce the actual area resisting the shearing action
considerably below the calculated area used in the formulae for
horizontal shear. (See page 98 for this formulae.) For this
reason it is unsafe, in designing large timber beams, to use
shearing stresses higher than those calculated for beams that
failed in horizontal shear. The effect of a failure in
horizontal shear is to divide the beam into two or more beams
the combined strength of which is much less than that of the
original beam. Fig. 18 shows a large beam in which two failures
in horizontal shear occurred at the same end. That the parts
behave independently is shown by the compression failure below
the original location of the neutral plane.

[Illustration: FIG. 18.--Failure of a large beam by horizontal
shear. _Photo by U. S, Forest Service._]

Table XI gives an analysis of the causes of first failure in 840
large timber beams of nine different species of conifers. Of the
total number tested 165 were air-seasoned, the remainder green.
The failure occurring first signifies the point of greatest
weakness in the specimen under the particular conditions of
loading employed (in this case, third-point static loading).

|-----------------------------------------------------------|
| TABLE XI