From the foregoing it will appear to the reader that average results are of very little value to the architect or engineer, unless--(1) the brickyard is mentioned from which the bricks experimented with came; (2) the particular class of brick from that yard; (3) the method of experimenting, as to whether any substance was placed between the dies of the press and the brick to be crushed, and if so, what; (4) if recessed or initialled; (5) whether machine or hand made, and (6) as to whether the surfaces of the bricks were concave, convex, or flat. Results on bricks not localised are not of much value, and it is absolutely useless for working purposes to give in one figure the strength of “London Stocks,” “Staffordshire blues,” “Leicestershire reds,” and the like. In a general way, of course, it will be admitted that the “Staffordshire blue” is a stronger brick than the “London Stock,” and so forth; but that is as much as can be permitted--it is of no practical use to give relative figures in general terms. It frequently happens that the capacity of the machine used for testing the strength of bricks is not enough for those bricks having a very high resistance to crushing. In the recent experiments by Professor Unwin, more than once alluded to in this article, it was found necessary to experiment with half-bricks only, and he ascertained that bricks tested as half-bricks shew about 25 per cent. less resistance per square foot than when tested as whole bricks. Further observations on strength are made under the next heading in connexion with other forms of testing the value and physical properties of bricks. CHAPTER XV. ABRASION, SPECIFIC GRAVITY. _Abrasion._--In this country it is not customary to test bricks and stone by means of the abrasion process, though many English materials have been dealt with in this manner on the continent. Abrasion tests are of special value in regard to paving bricks, and this mode of experiment is largely carried out in the United States. As Mr. H. Ries remarks,[21] the abrasion test approximates closely the conditions under which the paving brick is used, and is, therefore, an important one. The usual method of conducting this test is to put the bricks in an ordinary “foundry rattler,” filling it about one-third full. It is then rotated at the rate of about 30 revolutions per minute, and about 1,000 turns are sufficient. The bricks are weighed before and after to determine loss by abrasion. A more recent modification is to line the “rattler” with the bricks to be tested and then put in loose scrap iron. This is claimed to give more accurate results, and avoids loss by chipping due to the bricks knocking against each other, as in the previous method, although that has been somewhat obviated by Professor Orton, jun., by the introduction of a few billets of wood into the rattler.