SAMPLING AXD TESTING HIGHWAY MATERIALS. 
51 
eter calibrated at full immersion by the United States Bureau of 
Standards. 
4. The container in which the sample is to be tested shall be of Container. 
tin, cylindrical in shape, and shall have a flat bottom. Its inside 
dimensions shall be substantially as follows: Diameter, 55 milli- 
meters (2.17 inchesi ; depth. 35 millimeters (1.38 inches). (A 
3-ounce gill style ointment box, deep pattern, fulfills these require- 
ments.) For cold-application road oils the container shall be a 
2-ounce can approximately 60 millimeters in diameter and 20 
millimeters deep. 
II. PREPARATION OF SAMPLE. 
5. The sample as received shall be thoroughly stirred and agi- 
tated, warming, if necessary, to insure a complete mixture before 
the portion for analysis is removed. 
III. PROCEDURE. 
6. Weigh 50 grams of the water-free material to be tested into Method of test 
a tared container conforming to the requirements of section 4. 
For cold-application road oils use 20 grams of material. Bring 
the oven to a temperature of 163° C. (325° F.) and place the tin 
box containing the sample in one of the recesses of the revolving 
shelf. The thermometer shall be immersed for the depth of its 
bulb in a separate 50-gram sample of the material under test. 
placed in a similar container, and shall be conveniently suspended 
from the vertical shaft. This sample shall rest in one of the 
recesses upon the same shelf and revolve with the sample or 
samples under test. Then close the oven and rotate the shelf 
five to six revolutions per minute during the entire test. Main- 
tain the temperature at 103° C. (325° F.) for five hours, then 
remove the sample from the oven, cool, and weigh, and calculate 
the loss due to volatilization. 
7. During the five-hour period the temperature shall not vary 
more than 1° G. All tests showing a greater variation in tem- 
perature shall be rejected. 
8. Under ordinary circumstances a number of samples having 
about the same degree of volatility may be tested at the same 
time. Samples varying greatly in volatility should be tested 
separately. Where extreme accuracy is required, not more than 
one material should be tested at one time, and duplicate samples 
of it should be placed simultaneously in the oven. Such dupli- 
cates shall check within the limits of accuracy given below. 
Results obtained on samples showing evidence of foaming during 
the test shall be rejected. 
IV. ACCURACY. 
9. Up to 5 per cent loss in weight, the results obtained may be 
considered as correct within 0.5. Above 5 per cent loss in weight. 
the numerical limit of error increases 0.01 for every 0.5 per cent 
increase in loss by volatilization, as follows: 
Volatiliza- 
Numerical 
True volatili- 
tion loss. 
correction. 
zation loss. 
Per cmt. 
Per cent. 
5.0 
±0.50 
4. 50 to 5.50 
± .51 
4.91 to 6.01 
6.0 
± .52 
5. 48 to 6.52 
10.0 
± .60 
9. 40 to 10. 60 
15. 
± .70 
14.30 to 15. 70 
25.0 
± .90 
24. 10 to 25. 90 
40.0 
±1.20 
38. 80 to 41. 20 
Note. — If additional period-; of heating are desired, it i^ recom- 
mended that they be made in successive increments of rive hours 
each. When other tests ol the sample after heating are required, 
melt the residue in the container at the lowesl possible temperature 
and thoroughly mix by stirring, taking care to avoid incorporating 
air bubbles in the mass. 
