Weight of Stone

Cushions.

Homogeneous stones in small cubes appear in all cases to break as shown in Fig. 1. The forms of the fragments a and b are, approximately, either conical or pyramidal. The more or less disk-shaped pieces c and d are detached from the four sides of the cube with' a slight explosion. In the angles e and f, the stone is generally found crushed and ground into powder. This general form of breakage occurs also in non-homogeneous stones when crushed on their beds, but in this case the modification which the grain of the stone produces must be taken into account.

The nature of the material in contact with the stone while under pressure is a matter of great moment. If the materials which press upon the top and bottom of the specimen are soft and yielding and press out sidewise, they introduce horizontal forces which materially diminish the apparent crushing strength of the stone. If the pressing surfaces are hard and unyielding, the resistance of these surfaces adds considerably to the apparent strength.

Formerly steel, wood, lead, and leather were much used as pressing surfaces. Under certain limitations, the relative crushing strengths of stones with these different pressing surfaces are 100, 89, 65, and 62 respectively.* Tests of the strength of blocks of stone are useful only in com paring different stones, and give no idea of the strength of structures built of such stone (see * 622) or of the crushing strength of stone in large masses in its native bed (see § 657). Then, since it is not possible to have the stone under the same conditions while being tested that it is in the actual structure, it is best to test the stone under conditions that can be accurately described and readily dup licated. Therefore it is rapidly coming to be the custom to test the stone between metal pressing surfaces. Under these conditions the strength of the specimen will vary greatly with the degree of smooth ness of its bed surfaces. Hence, to obtain definite and precise results, these surfaces should be rubbed or ground perfectly smooth; but as this is tedious and expensive, it is quite common to reduce the bed surfaces to planes by plastering them with a thin coat of plaster of paris, and inverting the cube on a sheet of plate glass or allowing the plaster to set under a small pressure between the metal pressing sur faces of the testing machine. With the stronger stones, specimens with

plastered beds will show less strength than those having rubbed beds, and this difference will vary also with the length of time the plaster is allowed to harden. With a stone having a strength of 5,000 to 6,000 pounds per square inch, allowing the plaster to attain its maximum strength, this difference varied from 5 to 20 per cent, the mean for ten trials being almost 10 per cent of the strength of the specimen with rubbed beds.

The testing machine should be provided with a plate having a ball-and-socket bearing, to secure a uniform distribution of the pres sure; and care must be taken to place the test piece accurately in the axis of the testing machine. If the specimen spalls off on only one side, it is almost certain that it was not well bedded or that it was not placed centrally in the machine. If the cube is well bedded and properly placed in the machine; it will fail suddenly with a considerable report and the pieces will fly in all directions.

Effect of Water.

The specimen should be reasonably dry, since a wet stone is not as strong as a dry one (see 79).

Data on Crushing Strength.

Table 2 shows the results of all the stones tested with the U. S. testing machine at the Watertown Arsenal from 1883 to 1905, except two stones noted below. The quantities in the columns headed "Min." represent the strength of the weakest stone of each particular kind, and are the mean of three or more tests; and similarly for the quantities in the columns headed "Max." Two samples of granite were tested which are not included in Table 2. The results in the table are for test specimens prepared in the usual way; but the two cubes referred to were dressed out to 24 inches on a side and then reduced on a rubbing bed to 2 inches. The average for six tests of this stone, granite from Salisbury, N. C., was 49,457 pounds per square inch, the maximum for a single cube being 51,990.* The results for this stone are the highest known to have been obtained for the crushing strength of any stone or brick, and are probably largely due to the manner of dressing the specimens and also to unusual care in selecting the sample (§ 13), in bedding the cubes, and in placing them in the testing machine; and hence are not to be taken as representative.