520 BOOK XI. 



the other on the left. Finally, outside the hearth is the receiving-pit, which 

 is a foot wide and three palms deep ; when this is worn away it is restored 

 with lute alone, which easily retains the lead aUoy. 



If four liquation cakes are placed on the plates of each furnace, then the 

 iron blocks are laid under them ; but if the cakes are made from copper 

 " bottoms," or from liquation thorns, or from the accretions or " slags," of 

 which I have partly written above and will further describe a Httle later, 

 there are five of them, and because they are not so large and heavy, no blocks 

 are placed under them. Pieces of charcoal six digits long are laid between the 

 cakes, lest they should faU one against the other, or lest the last one should 

 faU against the wall which protects the third long waU from injury by fire. In 

 the middle empty spaces, long and large pieces of charcoal are hkewise laid. 

 Then when the panels have been set up, and the bar has been closed, the 

 furnace is fiUed with small charcoal, and a wicker basket full of charcoal is 

 thrown into the receiving-pit, and over that are thrown live coals ; soon 

 afterward the burning coal, hfted up in a shovel, is spread over aU parts of 

 the furnace, so that the charcoal in it may be kindled ; any charcoal which 

 remains in the receiving-pit is thrown into the passage, so that it may likewise 

 be heated. If this has not been done, the silver-lead aUoy liquated frum the 

 cakes is frozen by the coldness of the passage, and does not run down into the 

 receiving-pit. 



After a quarter of an hour the cakes begin to drip silver-lead alloy,^* 

 which runs down through the openings between the copper plates into the 

 passage. When the long pieces of charcoal have burned up, if the cakes 

 lean toward the wall, they are placed upright again with a hooked bar, but 

 if they lean toward the front bar they are propped up by charcoal ; more- 

 over, if some cakes shrink more than the rest, charcoal is added to the former 

 and not to the others. The silver drips together with the lead, for both melt 

 more rapidly than copper. The liquation thorns do not flow away, but remain 

 in the passage, and should be turned over frequently with a hooked bar, in 

 order that the silver-lead may hquate away from them and flow down into 

 the receiving pit ; that which remains is again melted in the blast furnace, 

 while that which flows into the receiving pit is at once carried with the remain- 



i^For the liquation it was necessary to maintain a reducing atmosphere, otherwise the 

 lead would o.xidize ; this was secured by keeping the cakes well covered with charcoal and 

 by preventing the entrance of air as much as possible. Moreover, it was necessary to preserve 

 a fairly even temperature. The proportions of copper and lead in the three liquation products 

 vary considerably, depending upon the method of conducting the process and the original 

 proportions. From the authors consulted (see note p. 492) an average would be about as 

 follows : — The residual copper — exhausted liquation cakes — ran from 25 to 33% lead ; the 

 liquated lead from 2 to 3 % copper ; and the liquation thorns, which were largely oxidized, 

 contained about 15% copper oxides, 80% lead oxides, together with impurities, such as 

 antimony, arsenic, etc. The proportions of the various products would obviously depend upon 

 the care in conducting the op'eration ; too high temperature and the admission of air would 

 increase the copper melted and oxidize more lead, and thus increase the liquation thorns. 

 There are insufficient data in Agricola to adduce conclusions as to the actual ratios produced. 

 The results given for the 6th charge (note 17, p. 512) would indicate about 30% lead in the 

 residual copper, and would indicate that the original charge was divided into about 24% of 

 residual copper, 18% of liquation thorns, and 57% of liquated lead. This, however, was 

 an unusually large proportion of liquation thorns, some of the authors giving instances of as 

 low as 5%. 



