N. B. M. A. standard, as the preparation of a charge for rattling requires much more care and discretion; and the supplies con sumed are more expensive, as there are many more bolts under continual wear and many more small accessories to provide, such as wedges, gaskets, washers, clips, etc. The iron abrasives worn out are less expensive than those of the standard process. The first cost of the Talbot-Jones machine is greater and the workman ship is distinctly of a better grade than is required for the stand ard process. The cost of a test by the Talbot-Jones process. how ever, is somewhat greater in money, time, energy, and skill than the N. B. M. A. standard.
The rattler test should be regarded as a mean (1) of determining whether any particular brand of brick meets the specifications and (2) of comparing any new brick with those that have been tested by service in the pavement; but (3) it should not be used to discriminate between two grades of brick differing only slightly in rattler losses, since owing to variations in the quality of different brick from the same lot and to accidental variations in making the test, the rattler can not give mathematical accuracy. Further, the loss in the rattler is the combined loss due to impact and abrasion, and consequently the loss of a brittle brick may be chiefly due to chipping, while that of a soft tough brick may be due principally to abrasion. It is not known that the relative losses due to impact and abrasion in the rattler is the same as in the pave ment ; and hence for this reason also the rattler should not be used to discriminate between two grades of high-quality paving brick.
The first of these experimental sections was laid in May, 1898, in Detroit, Mich., on Franklin Street, between Beallbien Street and the Grand Trunk freight depot. Transverse strips of fourteen kinds of brick were laid in a distance of 222 feet. The blocks were tested
by the former N. B. M. A. standard rattler test (§ 742). The foundation was 6 inches of concrete. The sand cushion was 1 inch thick, and the joints were filled with coal-tar distillate No. 6. For a comparison between the results of the rattler tests and a general observation of the effect of three years' wear in the pave ment, see Municipal Engineering, Vol. 22, pages and 363 65. These comparisons fail to show any close agreement be tween the former N. B. M. A. rattler test and service in the pavement.
The second experiment was inaugurated Oct. 16, 1900, on the driveway leading to the Chicago Avenue pumping station at Chicago, Ill. Nine kinds of paving blocks were laid in a distance of 392 feet. The time is too short and the traffic too light to have yet determined any valuable results.
The third experiment was begun Nov. 2,1901, on Holiday Street between Fayette and Baltimore Street, in Baltimore, Md. Seven kinds of clay blocks, two kinds of artificial sheet asphalt, one piece of natural rock asphalt, and one sample of creosoted pine blocks, were laid in a distance of 221 feet, the abutting street at each end being paved with Belgian blocks (§ 808). For details concern ing the tests of the brick and the construction of the pavement, see Annual Report of the City Engineer of Baltimore for 1901, pages 73-74 and 112. The bricks were tested by the N. B. M. A. standard rattler test, and were laid on a 14-inch sand cushion resting on new cobble-stone pavement. The joints were filled with sand. After eight months' wear, the fourth best brick block by the rattler test seems to have worn best in the pavement, and the best brick block in the rattler test seems to be among the poorest in the pavement.
The non-agreement between the results of the rattler test and service in the pavement indicates that the rattler test is not an infallible guide in selecting paving brick; but this disagreement does not prove that this test should be abandoned. The rattler test is the best method known for determining the qualities of paving brick, and possibly it may be ultimately modified so as more closely to agree with results of actual service.