126 J. Scully — On, the 'Effects produced hy small quantities of [No. 2, 



bismuth carbonate. Cadmiuni is very unlikely to be found in silver 

 bullion and its consideration may be neglected, but if tbe presence of 

 lead is suspected tbe carbonate filtered from tbe silver solution is dissolv- 

 ed in nitric acid, evaporated down with the addition of sulphuric acid, 

 and the lead sulphate formed (if any) collected and weighed in the usual 

 way. The bismuth is again precipitated as carbonate and treated as be- 

 fore directed. Many experiments have been made with synthetically 

 prepared mixtures of silver, copper, lead, and bismuth, the latter two 

 metals being in very small proportion to the silver, so as to imitate the 

 composition of some refined bars. Ullgreen's plan for the separation of 

 the carbonates of lead and bismuth, by dissolving them in acetic acid 

 and then precipitating the bismuth by means of a lead rod, does not work 

 satisfactorily and requires too long a time for the precipitation. 



II. As it seemed likely that a large number of experiments would 

 be required to determine accurately the smallest amount of bismuth 

 that would injure the ductility of our coinage alloy, and the still smaller 

 proportion that would not sensibly affect this ductility, it was deter- 

 mined to begin the enquiry by a number of laboratory experiments on 

 small bars of silver ; before trying the effects of bismuth on ordinary 

 coinage bars and with the procedure for lamination carried out in 

 the Mint. These laboratory experiments were made in the following 

 way : Pure silver prepared for assay check purposes, or an alloy of 

 silver and copper of which the exact composition had been determined 

 by assay, was melted in a clean plumbago crucible under charcoal. 

 When the metal was in fusion the necessary amount of bismuth was 

 rolled in a piece of paper, carried down at once to the bottom 

 of the silver bath, and then thoroughly mixed with the silver by 

 stirring. The calculated composition was confirmed by assay of the 

 silver. When this mixture had been accomplished, the contents of the 

 crucible were poured into an open iron ingot mould, and after cooling, 

 either quickly by plunging the casting into water or slowly in contact 

 with the mould, the bar so cast was tested for brittleness by hammering 

 and rolling. The bars cast were of two sizes, one set being 3' 75 inches 

 long, 1-125 broad, 0*375 thick and weighing about 6*2 troy oz. ; and 

 another set 2*69 inches long, 1*125 broad, 0*25 thick and weighing 

 about 4*1 troy oz. When reduced to the fullest extent by rolling, these 

 bars were converted into straps about 0*015 in thickness. In laminat- 

 ing them they were twice annealed, first after having undergone four 

 pinches in the rollers, and again after the tenth pinch from the beginning 

 Similarly shaped bars of silver, without bismuth, were occasionally lami- 

 nated in the same way to obtain a sure means of comparison. Before 

 any result was accepted as to brittleness or its absence, the bar under 



