finely divided mineral substances which exert a toughening effect. [43] _Journal Soc. Chem. Ind._, 1916, p. 872. Sheet rubber gives results in some ways inferior to compact crepe rubber when subjected to physical tests. Tensile strength seldom exceeds 15 kilos, but the elongation is usually higher--up to 600 or 700 per cent. That is to say, it stretches more, but breaks more easily. If, however, we take into consideration the diminution in sectional area of the test piece during stretching, it will be seen that crepe and sheet rubber have compensating properties. As this matter of sectional area reduction during stretching is important, both for raw and vulcanised rubber, it may be briefly referred to here. When rubber is stretched, the volume does not appreciably alter--at any rate, as regards uncompounded rubber. Hence the reduction of sectional area on stretching bears a simple relationship to the amount of stretching. If we double the length of the test piece, we halve the sectional area; if we treble the length, we reduce it to one-third, and so forth. Hence, if we multiply the breaking strain by the final length (_i.e._, length at break, taking the original length = 1), we obtain a figure, the "tensile product," which embodies both breaking strain and stretching capacity. In effect it gives us the breaking strain calculated on the sectional area at the _moment of rupture_ of the test piece. Adopting this formula, we obtain for crepe-- _Tensile _Final Length--i.e., _Tensile Strength._ Elongation + 1._ Product._ 30 × 4 = 120 and for smoked sheet 15 × 8 = 120 The difference in properties between crepe and sheet may probably be attributed to the heavier rolling of the crepe; which compacts the rubber. But if the crepe is rolled too much, the tensile strength falls, and there is no increased elongation to compensate. For the same reason, crepe which has been rerolled in this country is inferior to crepe as received direct from the plantation. At the most it is permissible to unite two or three layers of thin crepe to a thicker one by a single passage through even speed rollers, if the physical properties of the original rubber are to be conserved.[44] [44] Bulletin R.G.A., February, 1922, p. 64. Attempts to prepare crepe for use in a raw state, by rerolling uneven or irregular surfaced crepe in this country, only result in a rubber with inferior physical properties. Nor can sheet be rerolled to give crepe of good physical properties. The power required to break down the sheet and the heat developed, even on cold rollers, are an indication of physical properties destroyed. For subsequent vulcanisation this is not a matter of importance, because the vulcanising process restores to the rubber the properties which are lost in the process of rolling and milling or mastication.