In order to define the quality of a brass required for any specific purpose, not only should the copper content be given, but also sufficient information concerning the physical properties required, as will enable the manufacturer to supply a suitable metal. If the user has not this information then he should submit a sample of satisfactory metal with his order, or indicate to the manufacturer the articles into which the metal is to be made up. Unless the metal manufacturer knows within fairly fine limits the properties of the metal required, he can hardly be expected to give satisfaction. The defining of the physical properties of a metal is most satisfactorily carried out by reference to recognised standards. Tensile strength and elongation are recognised standards, but they are not so readily applicable to thin metal strips as they are to wire or rod. These two tests require a specially-shaped test piece, the preparation of which is somewhat tedious, and the carrying out of the test requires expensive testing machinery. For these reasons there have been evolved other and more simple tests specially applicable to metal strip.

The best known of these is the Erichsen test, in which a small piece of metal about 21/2 in. square is gripped over the mouth of a die and tightly held whilst a round-nosed punch is forced into the metal until fracture takes place. The depth of the fractured cup is taken as a measure of the ductility of the metal tested, and the smoothness or roughness of the dome is a rough indication of grain size. The test which can perhaps be the most widely applied is the Brinell hardness test, in which a small steel ball of known dimensions is pressed into the surface of the metal under a known load. A small circular impression is thus made, and from its diameter the Brinell hardness number is obtained by reference to tables which gives the diameters in millimetres against the corresponding Brinell numbers. Brinell numbers as quoted are not always strictly comparable, owing to balls of different diameters being used with varying loads. A different Brinnel number will be found for soft brass if a 10 m.m. ball is used with a load of 500 k.g., than if the same ball were used with a 1,000 kg. load. It is therefore necessary when referring to the hardness of metals by Brinnel hardness numbers that the ball and load are also quoted. A necessary accessory to the Brinnell hardness machine is a microscope, for reading the diameter of the ball impressions.

The hardness of metal is also sometimes expressed in Shore scleroscope numbers, the Shore scleroscope being an instrument which allows a small hammer to fall on to the surface of the metal, and the height of rebound is taken as a measure of the hardness of the metal. The greater the rebound, the harder the metal.

In addition to these mechanical tests for the metals, the microscope may frequently be used with advantage in estimating the physical properties of a metal, in determining the grain size of an annealed metal and in judging the freedom of the metal from impurities.

Tests show that nothing is to be gained by annealing the metal at a higher temperature, or for a longer time than that necessary for complete recrystallization and growth of the grains to a size of .04 to .06 m.m.

The " tear-length " test is of recent introduction, and is not yet widely known. It is carried out by making two parallel cuts 1 cm. deep and 1 cm. apart in the edge of a thin metal strip. The tongue of metal thus formed is turned up at right angles to the surface of the strip and pulled in this direc-tion, whilst the strip is tightly held. This operation results in the tearing out from the strip of a triangular piece of metal. The length of this is taken as the " tear-length." For hard-rolled strip it is small, generally between 1 cm. and 2 cm. With soft annealed strip, the tear-length is longer, and, if taken in the direction of rolling, may be from 4 cm. to 6 cm., or more, long. If the tear, however, be taken across the strip at right angles to the direction of rolling, then its length, even in the annealed condition, remains comparatively short, thus indicating that the properties of an annealed metal strip are not the same in all directions. Brinell hardness, Erichsen value, and microscopic examination or grain size, do not reveal this difference in directional properties. The tear-length must not be taken as a standard in the same way as Brinell hardness, but in the absence of any other means of estimating the degree to which a metal has been annealed, it will be found useful, especially since no apparatus is required for its carrying out.

In discussing these various tests, it is not suggested that all users should adopt any or all of them. But we do suggest that all users should know and be interested in them, so as to see the problems which the mill must solve in producing its results, and the care and precision with which we are seeking to solve them.

The final operation on articles are made from brass strip or sheet, is frequently a press or a burnishing operation, which leaves the metal in a hard (meaning unfortunately often an unevenly strained) condition. The hardness is necessary in order that the product may withstand handling and use, in service, without liability to damage. On the other hand, the strained condition may carry with it a liability of the metal to crack spontaneously (or " season-crack ") if exposed to weathering or other corrosive conditions. Recent investigations, however, have put in the hands of the manufacturer means whereby this risk of season-cracking in hard cold-worked products may be removed. All that is required is a low temperature anneal of the product at a temperature in the region of 250° C. for say one hour. An anneal at this temperature actually produces a slight hardening of the metal, but internal stresses are lowered to such a point that liability to season-crack under adverse conditions of storage or use is removed. It is extremely useful and necessary to know when this liability to season-cracking exists in any product, and for this purpose a test known as the " mercury-test" serves as an infallible guide. The clean product is immersed for fifteen minutes in an aqueous solution of mercurous nitrate containing one gramme of the latter to every one hundred cubic centimetres of water, to which solution one c.c. of strong nitric acid is added. Immediately on the immersion of a brass in such a solution, it becomes covered with a film of metallic mercury. After fifteen minutes' immersion the product is taken out and examined for cracks. If none are visible, the liability to season-cracking is negligible. This liability to season-cracking of the brasses becomes less as the copper content increases, and with gilding metal, even in the hard-rolled or burnished condition, it does not exist. This points, of course, to the great importance, for special work, of selecting an alloy sufficiently high in copper.