50 
BULLETIN 691, U. S. DEPARTMENT OF AGRICULTURE. 
in a sample of some fluid, nonvolatile bitumen, while the other is kept in air 
at the same level. The first thermometer serves to show the temperature of 
the samples during the test, while the latter gives prompt warning of any 
sudden changes in temperature due to irregularities in the gas pressure, etc. 
Before making the test the interior of the oven should show a temperature 
of 163° C. as registered by the thermometer in air. A round tin box 1 6 
centimeters in diameter and 2 centimeters deep is accurately weighed after 
carefully wiping with a towel to remove any grease or dirt. About 20 grams 
of the material to be tested is then placed in the box. The material may then 
be weighed on a rough balance, if one is at hand, after which the accurate 
weight, which should not vary more than 0.2 gram from the specified amount 
is obtained. It may be necessary to warm some of the material in order to 
handle it conveniently, after which it must be allowed to cool before deter- 
mining the accurate weight. 
The sample should now be placed in the oven, where it is allowed to remain 
for a period of five hours, during which time the temperature as shown by the 
thermometer in bitumen 
&> should not vary at any 
time more than 2° C. 
from 163° C. The sam- 
ple is then removed from 
the oven, allowed to cool, 
and reweighed. From 
the difference between 
this weight and the total 
weight before heating, 
the percentage of loss on 
the amount of material 
taken is calculated. 
Highly volatile and 
nonvolatile materials 
should not be subjected 
to this test at the same 
time in the same oven, 
owing to a tendency on 
the part of the latter to absorb some of the volatile products of the former. 
Some relative idea of the amount of hardening which has taken place may 
be obtained from the results of a float or penetration test made on the residue, 
as compared with the results of the same tests on the original sample. Before 
any tests are made on the residue, it should be melted and thoroughly stirred 
while cooling. 
FLOAT TEST. 
Fig. 4. — N'ew York Testing Laboratory oven. 
[Bulletin 314, p. 9.] 
NEW YOEK TESTING LABORATORY FLOAT APPARATUS. 
The float apparatus consists of two parts, an aluminum float or saucer and a 
conical brass collar (fig. 5). The two parts are made separately, so that one 
float may be used with a number of brass collars. 
In making the test the brass collar & is placed with the small end down on 
a brass plate, which has been previously amalgamated with mercury by first 
1 The box for a 50-gram sample should be 
meters deep. 
centimeters in diameter and 35 centi- 
