g test.]

One size of this machine will handle specimens for transverse
tests 9 inches wide and 6-foot span; the other, 12 inches wide
and 8-foot span. For compression tests a free fall of about 6.5
feet may be obtained. For transverse tests the fall is a little
less, depending upon the size of the specimen.

The machine is calibrated by dropping the hammer upon a copper
cylinder. The axial compression of the plug is noted. The energy
used in static tests to produce this axial compression under
stress in a like piece of metal is determined. The external
energy of the blow (_i.e._, the weight of the hammer X the
height of drop) is compared with the energy used in static tests
at equal amounts of compression. For instance:

Energy delivered, impact test 35,000 inch-pounds
Energy computed from static test .26,400 " "
Efficiency of blow of hammer .75.3 per cent.

_Preparing the material_: The material used in making impact
tests is of the same size and prepared in the same way as for
static bending and compression tests. Bending in impact tests is
more commonly used than compression, and small beams with
28-inch span are usually employed.

_Method_: In making an impact bending test the hammer is allowed
to rest upon the specimen and a zero or datum line is drawn. The
hammer is then dropped from increasing heights and drum records
taken until first failure. The first drop is one inch and the
increase is by increments of one inch until a height of ten
inches is reached, after which increments of two inches are used
until complete failure occurs or 6-inch deflection is secured.

The 50-pound hammer is used when with drops up to 68 inches it
is reasonably certain it will produce complete failure or 6-inch
deflection in the case of all specimens of a species; for all
other species a 100-pound hammer is used.

_Results_: The tracing on the drum (see Fig. 41) represents the
actual deflection of the stick and the subsequent rebounds for
each drop. The distance from the lowest point in each case to
the datum line is measured and its square in tenths of a square
inch entered as an abscissa on cross-section paper, with the
height of drop in inches as the ordinate. The elastic limit is
that point on the diagram where the square of the deflection
begins to increase more rapidly than the height of drop. The
difference between the datum line and the final resting point
after each drop represent