It is very doubtful that the ancients, working without chemical reagents for enhancing the delicacy of the test, could have used the touchstone to detect a single grain of alloying metal in a stater of gold alloy, or 1 part in 144, as Theophrastus claims. It is even doubtful that they could have detected differences of 1 part in 144 in the composition of alloys that are less rich in gold, even though these are more easily assayed with the touchstone.
Professor Gowland,327 who investigated traditional methods used by the goldsmiths of Japan, shows that when gold is alloyed with silver only and no reagent is used, a skillful assayer can obtain results with the touchstone that do not differ by more than 1 part in 100 from the results obtained by exact modern methods. Even this degree of accuracy can probably be obtained only on
324 Loc. cit.
325 Pax, 1200.
326 S.v. κολλυβιστήί.
327 Journal of the Institute of Metals, IV (1910), 11.