METHODS <>F SAMPLING AND TESTING HIGHWAY MATERIALS 53 
bearing. The rile shall be placed upon the fillet while the plaster of Paris is 
still somewhat plastic. The load shall be applied through an upper knife bear- 
ing of the same size and length as the lower bearing. A plaster-of -Paris fillet 
1 inch wide shall be east along the length <>f the crown of the tile to equalize 
the lower bearing before the upper one is brought into contact. 
Both of the bearings shall be sufficiently rigid to transmit and receive uni- 
form loads throughout their lengths without deflection, and shall be so attached 
to the machine as to transmit and receive the maximum stresses produced by 
the test without lost motion, vibration, or sudden shock. 
20. When three-edge bearings (fig. 21 » are used, the ends of each specimen of 
tile shall be accurately marked in halves of the circumference prior to the test. 
The lower bearings shall consist of two wooden strips with vertical sides, 
each strip having its interior top corner rounded to a radius of approximately 
one-half inch. They shall be straight, and shall be securely fastened to a rigid 
block with their interior vertical sides 1 inch apart. 
The upper bearing shall be a wooden block straight and true from end to end. 
The test load shall be applied through the upper bearing block in such a way 
as to leave the bearing free to move in a vertical plane passing midway between 
the lower bearings. 
In testing a tile which is " out of straight.'* the lines of the bearings chosen 
shall be from those which appear to give most favorable conditions for fair 
bearings. 
(B) ABSORPTION TESTS OF DRAIN TILE 
21. Not less than three separate test specimens from each of five separate 
tiles shall be taken as a " standard sample " for the absorption test. Of the 
three specimens from each tile, one shall be taken from one end, another from 
the opposite end. and the third shall be taken from the middle portion of the 
tile. Each specimen shall be of from 12 to 20 square inches in area, measured 
upon the exterior or convex side, and shall be as nearly square as the nature of 
the material will readily permit. The specimens shall be obtained by breaking 
the tile, and shall be apparently sound, solid pieces of the wall of the tile, and 
shall not show cracks or fissures or shattered edges due to the shock of breaking 
or cutting. The specimens may lie obtained from the broken pieces of the tiles 
used in the strength test, if the restrictions as to the size and location of the 
specimens can be duly observed. The specimens shall be so marked as to 
permit the identity of each one to be ascertained at any stage of the test. 
22. Preparatory to the absorption test, all specimens shall be first weighed 
and then dried in a drier or oven, at a temperature of not less than 110° C. 
(230° F. ) for not less than three hours. After removal from the drier, the 
specimens shall be allowed to cool to a temperature of 20° to 25° C. (68° to 
77° F.) and reweighed. If the specimens were apparently dry when taken, and 
the second weight closely checks the first, the specimens shall be considered dry. 
If the specimens were known to be wet when taken, they shall be placed in the 
drier for a further drying treatment of two hours, and reweighed. If the third 
weight checks the second, the specimens shall be considered dry. In case of 
any doubt, the specimens mnst be redried for two-hour periods until check 
weights are obtained. • 
23. The balance used shall be sensitive to 0.5 gram when loaded with 1 kilo- 
gram, and weighings shall be read at least to the nearest gram. Where other 
than metric weights are used, the same order of accuracy must be obtained. 
In reweighing after immersion, the specimens shall be removed from the 
water, not allowed to drain for more than one minute, the superficial water 
removed by towel or blotting paper, and the specimens at once put upon the 
balance. 
24. Specimens after weighing shall be placed in a suitable woven- wire recep- 
tacle, packed tightly enough to prevent jostling, covered with distilled water or 
rainwater, raised to the boiling point and boiled for five hours, and then cooled 
in water to a final temperature of 10° to 15° C. i 50° to 59° F.). 
25. The test results shall be calculated as percentages of the initial dry 
weight, carried to the nearest first decimal place. The results shall be reported 
separately for each individual specimen, together with the mean of the 15 or 
more specimens comprising the standard sample, the maximum and the mini- 
mum single observations entering into the mean, and the variation between the 
maximum and the minimum of the three specimens of each tile represented in 
the standard sample. 
