seems clear from these descriptions that the ancients applied this test in only a qualitative way. It could also have been used for rough quantitative measurements, since a similar procedure was once used in the French mints to estimate the composition of silver alloys,313 and a somewhat similar method was used to estimate the percentage of copper in gold alloys in the old mints of Japan.314 Possibly the ancients also used the fire test for rough assaying, though the available evidence certainly gives us no grounds for believing this. It is very unlikely, as some have conjectured, that in testing gold by fire they ever employed anything like the accurate modern method of fire assaying in which the gold is isolated by the fusion of a weighed sample of metal with chemical reagents, the silver is removed from the gold by acid treatment after this fusion, and finally the pure metal is carefully weighed.
45. the stone wor\s by friction. When gold or gold alloys are tested by being rubbed on a touchstone, a plainly visible streak of metal remains on the black surface of the stone. The intensity of the yellow color of this streak is directly related to the gold content of the metal.315 In modern practice, streaks made by gold alloys of known composition are placed alongside a streak made by the metal that is being tested; this is done with touch needles, which are a graded series of heavy flat needles tipped with gold alloys. The color of the streak left by the metal is then compared with the colors of the standard streaks, and a match in color indicates that the metal has the same gold content as the corresponding standard alloy. In the modern use of the touchstone, the streaks left by the standard alloys and the unknown metal are nearly always tested by means of chemical reagents. WitJi this refinement the method becomes more delicate and more reliable. There is no mention in ancient literature of the use of touch needles or of standard alloys in any form, but, unless standards of some sort had been used, the method would have produced only rough results. However, the statements of Theo-
313 W. C. Roberts, Journal of the Society of Arts, XXXII (1884), 882.
314 W. Gowland, Journal of the Institute of Metals, IV (1910), 11.
315 No black streak is formed by gold alloys, even those of low fineness, as is erroneously stated by G. Thomson, Classical Review, LVIII (1944), 36. This misconception has been discussed by D. E. Eichholz, Classical Review, LIX (1945), 52.
