BOOK XI. 509 



let down from the pulley of the crane arm ; the inside diameter of this ring 

 is six digits, and it is about a digit and a half thick ; the ring is then engaged 

 in the hook whose straight stem is in the cake, and thus the cake is raised from 

 the mould and put into its place. 



The copper and lead, when thus melted, yield a small amount of " slag "^* 

 and much litharge. The Utharge does not cohere, but falls to pieces Uke the 

 residues from malt from which beer is made. Pompholyx adheres to the walls 

 in white ashes, and to the sides of the furnace adheres sfodos. 



In this practical manner lead is alloyed with copper in which there is but 

 a moderate portion of silver. If, however, there is much silver in it, as, for 

 instance, two librae, or two librae and a bes, to the centumpondium, — which 

 weighs one hundred and thirty-three and a third librae, or one hundred and 

 forty-six librae and a bes,'^^ — then the foreman of the works adds to a centum- 

 pondium of such copper three centumpondia of lead, in each centumpondium 

 of which there is a third of a libra of silver, or a third of a libra and a semi- 

 uncia. In this manner three liquation cakes are made, which contain 

 altogether three centumpondia of copper and nine centumpondia of lead.^* The 

 lead, when it has been liquated from the copper, weighs seven centumpondia ; 

 and in each centumpondium — if the centumpondium of copper contain two 

 librae of silver, and the lead contain a third of a libra — there will be a libra 

 and a sixth and more than a semi-uncia of silver; while in the exhausted 

 Uquation cakes, and in the Uquation thorns, there remains a third of a libra. 



"An analysis of this " slag " by Karsten [Archiv. ist Series IX, p. 24) showed 63.2% 

 lead oxide, 5.1% cuprous oxide, 20.1% silica (from the fuel and furnace linings), together 

 with some iron alumina, etc. The pompholyx and spodos were largely zinc oxide (see note, 



P- 394)- 



**This description of a centumpondium which weighed either 133I librae, or 146I librae, 

 adds confusion to an already much mixed subject (see Appendix C). Assuming the 

 German pfundt to weigh 7,219 troy grains, and the Roman libra 4,946 grains, then a centner 

 would weigh 145.95 librae, which checks up fairly well with the second case ; but under what 

 circumstances a centner can weigh 133J librae we are unable to record. At first sight it might 

 appear from this statement that where Agricola uses the word centumpondium he means the 

 German centner. On the other hand, in the previous five or six pages the expressions one-third, 

 five-sixths, ten-twelfths of a libra are used, which are even divisions of the Roman 12 unciae 

 to one libra, and are used where they manifestly mean divisions of 12 units. If Agricola 

 had in mind the German scale, and were using the libra for a pfundt of 16 untzen, these divisions 

 would amount to fractions, and would not total the sicilicus and drachma quantities given, 

 nor would they total any of the possibly synonymous divisions of the German untzen (see 

 also page 254). 



^*If we assume Roman weights, the charge in the first case can be tabulated as follows, 

 and for convenience will be called the fifth charge : — 



5TH Charge (3 cakes). 

 Amount of copper .. .., ,. .. .. .. .. 211. 8 lbs. 



Amount of lead 

 Weight of each cake 

 Average value of charge 

 Per cent, of copper 



635.4 lbs. 



. . 282.4 lbs. 



. . 218 ozs. 18 dwts. 



25% 



Average value of original copper per ton . . . . . . . . 583 ozs. 6 dwts. 16 grs. 



Weight of argentiferous lead liquated out . . . . . . . . 494.2 lbs. 



Average value of liquated lead per ton 352 ozs. 8 dwts. 



Weight of residues . . . . . . . . . . . . . . . . 353 lbs. 



Average value of residues per ton . . . . . . . . . . 20 ozs. (about). 



Extraction of silver into the argentiferous lead . . . . . . 94% 



The results given in the second case where the copper contains 2 librae and a bes per 



centumpondium do not tie together at all, for each liquation cake should contain 3 librae 

 gi unciae, instead of i^ librae and J uncia of silver. 



35 



