[3] _Aqua valens_, literally strong, potent, or powerful water. It will appear later, from the method of manufacture, that hydrochloric, nitric, and sulphuric acids and _aqua regia_ were more or less all produced and all included in this term. We have, therefore, used either the term _aqua valens_ or simply _aqua_ as it occurs in the text. The terms _aqua fortis_ and _aqua regia_ had come into use prior to Agricola, but he does not use them; the Alchemists used various terms, often _aqua dissolvia_. It is apparent from the uses to which this reagent was put in separating gold and silver, from the method of clarifying it with silver and from the red fumes, that Agricola could have had practical contact only with nitric acid. It is probable that he has copied part of the recipes for the compounds to be distilled from the Alchemists and from such works as the _Probierbuchlein_. In any event he could not have had experience with them all, for in some cases the necessary ingredients for making nitric acid are not all present, and therefore could be of no use for gold and silver separation. The essential ingredients for the production of this acid by distillation, were saltpetre, water, and either vitriol or alum. The other substances mentioned were unnecessary, and any speculation as to the combinations which would result, forms a useful exercise in chemistry, but of little purpose here. The first recipe would no doubt produce hydrochloric acid.

[4] Agricola, in the _Interpretatio_, gives the German equivalent for the Latin _aerugo_ as _Spanschgrun_--"because it was first brought to Germany from Spain; foreigners call it _viride aeris_ (copper green)."

The English "verdigris" is a corruption of _vert de grice_. Both verdigris and white lead were very ancient products, and they naturally find mention together among the ancient authors. The earliest description of the method of making is from the 3rd Century B.C., by Theophrastus, who says (101-2): "But these are works of art, as is also Ceruse (_psimythion_) to make which, lead is placed in earthen vessels over sharp vinegar, and after it has acquired some thickness of a kind of rust, which it commonly does in about ten days, they open the vessels and sc.r.a.pe off, as it were, a kind of foulness; they then place the lead over the vinegar again, repeating over and over again the same method of sc.r.a.ping it till it is wholly dissolved; what has been sc.r.a.ped off they then beat to powder and boil for a long time; and what at last subsides to the bottom of the vessel is the white lead.... Also in a manner somewhat resembling this, verdigris (_ios_) is made, for copper is placed over lees of wine (grape refuse?), and the rust which it acquires by this means is taken off for use. And it is by this means that the rust which appears is produced." (Based on Hill's translation.) Vitruvius (VII, 12), Dioscorides (V, 51), and Pliny (x.x.xIV, 26 and 54), all describe the method of making somewhat more elaborately.

[5] _Amiantus_ (_Interpretatio_ gives _federwis_, _pliant_, _salamanderhar_). From Agricola's elaborate description in _De Natura Fossilium_ (p. 252) there can be no doubt that he means asbestos. This mineral was well-known to the Ancients, and is probably earliest referred to (3rd Century B.C.) by Theophrastus in the following pa.s.sage (29): "There is also found in the mines of Scaptesylae a stone, in its external appearance somewhat resembling wood, on which, if oil be poured, it burns; but when the oil is burnt away, the burning of the stone ceases, as if it were in itself not liable to such accidents."

There can be no doubt that Strabo (X, 1) describes the mineral: "At Carystus there is found in the earth a stone, which is combed like wool, and woven, so that napkins are made of this substance, which, when soiled, are thrown into the fire and cleaned, as in the washing of linen." It is also described by Dioscorides (V, 113) and Pliny (XIX, 4).

Asbestos cloth has been found in Pre-Augustinian Roman tombs.

[6] This list of four recipes is even more obscure than the previous list. If they were distilled, the first and second mixtures would not produce nitric acid, although possibly some sulphuric would result. The third might yield nitric, and the fourth _aqua regia_. In view of the water, they were certainly not used as cements, and the first and second are deficient in the vital ingredients.

[7] _Distillation_, at least in crude form, is very old. Aristotle (_Meteorologica_, IV.) states that sweet water can be made by evaporating salt-water and condensing the steam. Dioscorides and Pliny both describe the production of mercury by distillation (note 58, p.

432). The Alchemists of the Alexandrian School, from the 1st to the 6th Centuries, mention forms of imperfect apparatus--an ample discussion of which may be found in Kopp, _Beitrage zur Geschichte der Chemie_, Braunschweig, 1869, p. 217.

[8] It is desirable to note the contents of the residues in the retort, for it is our belief that these are the materials to which the author refers as "lees of the water which separates gold from silver," in many places in Book VII. They would be strange mixtures of sodium, pota.s.sium, aluminium sulphates, with silica, brickdust, asbestos, and various proportions of undigested vitriol, salt, saltpetre, alum, iron oxides, etc. Their effect must have been uncertain. Many old German metallurgies also refer to the _Todenkopf der Scheidwa.s.ser_, among them the _Probierbuchlein_ before Agricola, and after him Lazarus Ercker (_Beschreibung Allerfurnemsten_, etc., Prague, 1574). See also note 16, p. 234.

[9] This use of silver could apply to one purpose only, that is, the elimination of minor amounts of hydrochloric from the nitric acid, the former originating no doubt from the use of salt among the ingredients.

The silver was thus converted into a chloride and precipitated. This use of a small amount of silver to purify the nitric acid was made by metallurgists down to fairly recent times. Biringuccio (IV, 2) and Lazarus Ercker (p. 71) both recommend that the silver be dissolved first in a small amount of acid, and the solution poured into the newly-manufactured supply. They both recommend preserving this precipitate and its cupellation after melting with lead--which Agricola apparently overlooked.

[10] In this description of parting by nitric acid, the author digresses from his main theme on pages 444 and 445, to explain a method apparently for small quant.i.ties where the silver was precipitated by copper, and to describe another cryptic method of precipitation. These subjects are referred to in notes 11 and 12 below. The method of parting set out here falls into six stages: _a_--cupellation, _b_--granulation, _c_--solution in acid, _d_--treatment of the gold residues, _e_--evaporation of the solution, _f_--reduction of the silver nitrate. For nitric acid parting, bullion must be free from impurities, which cupellation would ensure; if copper were left in, it would have the effect he mentions if we understand "the silver separated from the gold soon unites with it again," to mean that the silver unites with the copper, for the copper would go into solution and come down with the silver on evaporation.

Agricola does not specifically mention the necessity of an excess of silver in this description, although he does so elsewhere, and states that the ratio must be at least three parts silver to one part gold. The first description of the solution of the silver is clear enough, but that on p. 445 is somewhat difficult to follow, for the author states that the bullion is placed in a retort with the acid, and that distillation is carried on between each additional charge of acid. So far as the arrangement of a receiver might relate to the saving of any acid that came over accidentally in the boiling, it can be understood, but to distill off much acid would soon result in the crystallization of the silver nitrate, which would greatly impede the action of subsequent acid additions, and finally the gold could not be separated from such nitrate in the way described. The explanation may be (apart from incidental evaporation when heating) that the acids used were very weak, and that by the evaporation of a certain amount of water, not only was the acid concentrated, but room was provided for the further charges.

The acid in the gold wash-water, mentioned in the following paragraph, was apparently thus concentrated. The "gla.s.s" mentioned as being melted with litharge, argols, nitre, etc., was no doubt the silver nitrate. The precipitation of the silver from the solution as a chloride, by the use of salt, so generally used during the 18th and 19th Centuries, was known in Agricola's time, although he does not mention it. It is mentioned in Geber and the _Probierbuchlein_. The clarity of the latter on the subject is of some interest (p. 34a): "How to pulverise silver and again make it into silver. Take the silver and dissolve it in water with the _starckenwa.s.ser_, _aqua fort_, and when that is done, take the silver water and pour it into warm salty water, and immediately the silver settles to the bottom and becomes powder. Let it stand awhile until it has well settled, then pour away the water from it and dry the settlings, which will become a powder like ashes. Afterward one can again make it into silver. Take the powder and put it on a _test_, and add thereto the powder from the settlings from which the _aqua forte_ has been made, and add lead. Then if there is a great deal, blow on it until the lead has incorporated itself ... blow it until it _plickt_ (_blickens_). Then you will have as much silver as before."

[11] The silver is apparently precipitated by the copper of the bowl. It would seem that this method was in considerable use for small amounts of silver nitrate in the 16th Century. Lazarus Ercker gives elaborate directions for this method (_Beschreibung Allerfurnemsten_, etc., Prague, 1574, p. 77).

[12] We confess to a lack of understanding of this operation with leaves of lead and copper.

[13] We do not understand this "appearance of black." If the nitrate came into contact with organic matter it would, of course, turn black by reduction of the silver, and sunlight would have the same effect.

[14] This would be equal to from 62 to 94 parts of copper in 1,000.

[15] As 144 _siliquae_ are 1 _uncia_, then 1/4 _siliqua_ in 8 _unciae_ would equal one part silver in 4,608 parts gold, or about 999.8 fine.

[16] The object of this treatment with sulphur and copper is to separate a considerable portion of silver from low-grade bullion (_i.e._, silver containing some gold), in preparation for final treatment of the richer gold-silver alloy with nitric acid. Silver sulphide is created by adding sulphur, and is drawn off in a silver-copper regulus. After the first sentence, the author uses silver alone where he obviously means silver "containing some gold," and further he speaks of the "gold lump"

(_ma.s.sula_) where he likewise means a b.u.t.ton containing a great deal of silver. For clarity we introduced the term "regulus" for the Latin _mistura_. The operation falls into six stages: _a_, granulation; _b_, sulphurization of the granulated bullion; _c_, melting to form a combination of the silver sulphide with copper into a regulus, an alloy of gold and silver settling out; _d_, repet.i.tion of the treatment to abstract further silver from the "lump;" _e_, refining the "lump" with nitric acid; _f_, recovery of the silver from the regulus by addition of lead, liquation and cupellation.

The use of a "circle of fire" secures a low temperature that would neither volatilize the sulphur nor melt the bullion. The amount of sulphur given is equal to a ratio of 48 parts bullion and 9 parts sulphur. We are not certain about the translation of the paragraph in relation to the proportion of copper added to the granulated bullion; because in giving definite quant.i.ties of copper to be added in the contingencies of various original copper contents in the bullion, it would be expected that they were intended to produce some positive ratio of copper and silver. However, the ratio as we understand the text in various cases works out to irregular amounts, _i.e._, 48 parts of silver to 16, 12.6, 24, 20.5, 20.8, 17.8, or 18 parts of copper. In order to obtain complete separation there should be sufficient sulphur to have formed a sulphide of the copper as well as of the silver, or else some of the copper and silver would come down metallic with the "lump". The above ratio of copper added to the sulphurized silver, in the first instance would give about 18 parts of copper and 9 parts of sulphur to 48 parts of silver. The copper would require 4.5 parts of sulphur to convert it into sulphide, and the silver about 7 parts, or a total of 11.5 parts required against 9 parts furnished. It is plain, therefore, that insufficient sulphur is given. Further, the litharge would probably take up some sulphur and throw down metallic lead into the "lump".

However, it is necessary that there should be some free metallics to collect the gold, and, therefore, the separation could not be complete in one operation. In any event, on the above ratios the "gold lump" from the first operation was pretty coppery, and contained some lead and probably a good deal of silver, because the copper would tend to desulphurize the latter. The "powder" of gla.s.s-galls, salt, and litharge would render the ma.s.s more liquid and a.s.sist the "gold lump" to separate out.

The Roman silver _sesterce_, worth about 2-1/8 pence or 4.2 American cents, was no doubt used by Agricola merely to indicate an infinitesimal quant.i.ty. The test to be applied to the regulus by way of cupellation and parting of a sample with nitric acid, requires no explanation. The truth of the description as to determining whether the gold had settled out, by using a chalked iron rod, can only be tested by actual experiment. It is probable, however, that the sulphur in the regulus would attack the iron and make it black. The re-melting of the regulus, if some gold remains in it, with copper and "powder" without more sulphur, would provide again free metallics to gather the remaining gold, and by desulphurizing some silver this b.u.t.ton would probably not be very pure.

From the necessity for some free metallics besides the gold in the first treatment, it will be seen that a repet.i.tion of the sulphur addition and re-melting is essential gradually to enrich the "lump". Why more copper is added is not clear. In the second melting, the ratio is 48 parts of the "gold lump", 12 parts of sulphur and 12 parts copper. In this case the added copper would require about 3 parts sulphur, and if we consider the deficiency of sulphur in the first operations pertained entirely to the copper, then about 2.5 parts would be required to make good the shortage, or in other words the second addition of sulphur is sufficient. In the final parting of the "lump" it will be noticed that the author states that the silver ratio must be arranged as three of silver to one of gold. As to the recovery of the silver from the regulus, he states that 66 _librae_ of silver give 132 _librae_ of _regulus_. To this, 500 _librae_ of lead are added, and it is melted in the "second" furnace, and the litharge and hearth-lead made are re-melted in the "first" furnace, the cakes made being again treated in the "third" furnace to separate the copper and lead. The "first" is usually the blast furnace, the "second" furnace is the cupellation furnace, and the "third" the liquation furnace. It is difficult to understand this procedure. The charge sent to the cupellation furnace would contain between 3% and 5% copper, and between 3% and 5% sulphur.

However, possibly the sulphur and copper could be largely abstracted in the skimmings from the cupellation furnace, these being subsequently liquated in the "third" furnace. It may be noted that two whole lines from this paragraph are omitted in the editions of _De Re Metallica_ after 1600. For historical note on sulphur separation see page 461.

[17] There can be no doubt that in most instances Agricola's _stibium_ is antimony sulphide, but it does not follow that it was the mineral _stibnite_, nor have we considered it desirable to introduce the precision of either of these modern terms, and have therefore retained the Latin term where the sulphide is apparently intended. The use of antimony sulphide to part silver from gold is based upon the greater affinity of silver than antimony for sulphur. Thus the silver, as in the last process, is converted into a sulphide, and is absorbed in the regulus, while the metallic antimony alloys with the gold and settles to the bottom of the pot. This process has several advantages over the sulphurization with crude sulphur; antimony is a more convenient vehicle of sulphur, for it saves the preliminary sulphurization with its attendant difficulties of volatilization of the sulphur; it also saves the granulation necessary in the former method; and the treatment of the subsequent products is simpler. However, it is possible that the sulphur-copper process was better adapted to bullion where the proportion of gold was low, because the fineness of the bullion mentioned in connection with the antimonial process was apparently much higher than the previous process. For instance, a _bes_ of gold, containing 5, 6, or 7 double _s.e.xtulae_ of silver would be .792, .750 or .708 fine. The antimonial method would have an advantage over nitric acid separation, in that high-grade bullion could be treated direct without artificial decrease of fineness required by inquartation to about .250 fine, with the consequent incidental losses of silver involved.

The process in this description falls into six operations: _a_, sulphurization of the silver by melting with antimony sulphide; _b_, separation of the gold "lump" (_ma.s.sula_) by jogging; _c_, re-melting the regulus (_mistura_) three or four times for recovery of further "lumps"; _d_, re-melting of the "lump" four times, with further additions of antimony sulphide; _e_, cupellation of the regulus to recover the silver; _f_, cupellation of the antimony from the "lump" to recover the gold. Percy seems to think it difficult to understand the insistence upon the addition of copper. Biringuccio (IV, 6) states, among other things, that copper makes the ingredients more liquid. The later metallurgists, however, such as Ercker, Lohneys, and Schluter, do not mention this addition; they do mention the "swelling and frothing,"

and recommend that the crucible should be only partly filled. As to the copper, we suggest that it would desulphurize part of the antimony and thus free some of that metal to collect the gold. If we a.s.sume bullion of the medium fineness mentioned and containing no copper, then the proportions in the first charge would be about 36 parts gold, 12 parts silver, 41 parts sulphur, 103 parts antimony, and 9 parts copper. The silver and copper would take up 4.25 parts of sulphur, and thus free about 10.6 parts of antimony as metallics. It would thus appear that the amount of metallics provided to a.s.sist the collection of the gold was little enough, and that the copper in freeing 5.6 parts of the antimony was useful. It appears to have been necessary to have a large excess of antimony sulphide; for even with the great surplus in the first charge, the reaction was only partial, as is indicated by the necessity for repeated melting with further antimony.

The later metallurgists all describe the separation of the metallic antimony from the gold as being carried out by oxidation of the antimony, induced by a jet of air into the crucible, this being continued until the ma.s.s appears limpid and no cloud forms in the surface in cooling. Agricola describes the separation of the silver from the regulus by preliminary melting with argols, gla.s.s-gall, and some lead, and subsequent cupellation of the lead-silver alloy. The statement that unless this preliminary melting is done, the cupel will absorb silver, might be consonant with an attempt at cupellation of sulphides, and it is difficult to see that much desulphurizing could take place with the above fluxes. In fact, in the later descriptions of the process, iron is used in this melting, and we are under the impression that Agricola had omitted this item for a desulphurizing reagent. At the Dresden Mint, in the methods described by Percy (Metallurgy Silver and Gold, p. 373) the gold lumps were tested for fineness, and from this the amount of gold retained in the regulus was computed. It is not clear from Agricola's account whether the test with nitric acid was applied to the regulus or to the "lumps". For historical notes see p. 461.

[18] As will be shown in the historical note, this process of separating gold and silver is of great antiquity--in all probability the only process known prior to the Middle Ages, and in any event, the first one used. In general the process was performed by "cementing" the disintegrated bullion with a paste and subjecting the ma.s.s to long-continued heat at a temperature under the melting point of the bullion. The cement (_compositio_) is of two different species; in the first species saltpetre and vitriol and some aluminous or silicious medium are the essential ingredients, and through them the silver is converted into nitrate and absorbed by the ma.s.s; in the second species, common salt and the same sort of medium are the essentials, and in this case the silver is converted into a chloride. Agricola does not distinguish between these two species, for, as shown by the text, his ingredients are badly mixed.

The process as here described falls into five operations: _a_, granulation of the bullion or preparation of leaves; _b_, heating alternate layers of cement and bullion in pots; _c_, washing the gold to free it of cement; _d_, melting the gold with borax or soda; _e_, treatment of the cement by way of melting with lead and cupellation to recover the silver. Investigation by Boussingault (_Ann. De Chimie_, 1833, p. 253-6), D'Elhuyar (_Bergbaukunde_, Leipzig, 1790, Vol. II, p.

200), and Percy (Metallurgy of Silver and Gold, p. 395), of the action of common salt upon silver under cementation conditions, fairly well demonstrated the reactions involved in the use of this species of cement. Certain factors are essential besides salt: _a_, the admission of air, which is possible through the porous pots used; _b_, the presence of some moisture to furnish hydrogen; _c_, the addition of alumina or silica. The first would be provided by Agricola in the use of new pots, the second possibly by use of wood fuel in a closed furnace, the third by the inclusion of brickdust. The alumina or silica at high temperatures decomposes the salt, setting free hydrochloric acid and probably also free chlorine. The result of the addition of vitriol in Agricola's ingredients is not discussed by those investigators, but inasmuch as vitriol decomposes into sulphuric acid under high temperatures, this acid would react upon the salt to free hydrochloric acid, and thus a.s.sist to overcome deficiencies in the other factors. It is possible also that sulphuric acid under such conditions would react directly upon the silver to form silver sulphates, which would be absorbed into the cement. As nitric acid is formed by vitriol and saltpetre at high temperatures, the use of these two substances as a cementing compound would produce nitric acid, which would at once attack the silver to form silver nitrate, which would be absorbed into the melted cement. In this case the brickdust probably acted merely as a vehicle for the absorption, and to lower the melting point of the ma.s.s and prevent fusion of the metal. While nitric acid will only part gold and silver when the latter is in great excess, yet when applied as fumes under cementation conditions it appears to react upon a minor ratio of silver. While the reactions of the two above species of compounds can be accounted for in a general way, the problem furnished by Agricola's statements is by no means simple, for only two of his compounds are simply salt cements, the others being salt and nitre mixtures. An inspection of these compounds produces at once a sense of confusion.

Salt is present in every compound, saltpetre in all but two, vitriol in all but three. Lewis (_Traite Singulier de Metallique_, Paris, 1743, II, pp. 48-60), in discussing these processes, states that salt and saltpetre must never be used together, as he a.s.serts that in this case _aqua regia_ would be formed and the gold dissolved. Agricola, however, apparently found no such difficulty. As to the other ingredients, apart from nitre, salt, vitriol, and brickdust, they can have been of no use.

Agricola himself points out that ingredients of "metallic origin"

corrupt the gold and that brickdust and common salt are sufficient. In a description of this process in the _Probierbuchlein_ (p. 58), no nitre is mentioned. This booklet does mention the recovery of the silver from the cement by amalgamation with mercury--the earliest mention of silver amalgamation.

[19] While a substance which we now know to be natural zinc sulphate was known to Agricola (see note 11, p. 572), it is hardly possible that it is referred to here. If green vitriol be dehydrated and powdered, it is white.

[20] The processes involved by these "other" compounds are difficult to understand, because of the lack of information given as to the method of operation. It might be thought that these were five additional recipes for cementing pastes, but an inspection of their internal composition soon dissipates any such a.s.sumption, because, apart from the lack of brickdust or some other similar necessary ingredient, they all contain more or less sulphur. After describing a preliminary treatment of the bullion by cupellation, the author says: "Then the silver is sprinkled with two _unciae_ of that powdered compound and is stirred. Afterward it is poured into another crucible ... and violently shaken. The rest is performed according to the process I have already explained." As he has already explained four or five parting processes, it is not very clear to which one this refers. In fact, the whole of this discussion reads as if he were reporting hearsay, for it lacks in every respect the infinite detail of his usual descriptions. In any event, if the powder was introduced into the molten bullion, the effect would be to form some silver sulphides in a regulus of different composition depending upon the varied ingredients of different compounds. The enriched bullion was settled out in a "lump" and treated "as I have explained," which is not clear.

[21] HISTORICAL NOTE ON PARTING GOLD AND SILVER. Although the earlier Cla.s.sics contain innumerable references to refining gold and silver, there is little that is tangible in them, upon which to hinge the metallurgy of parting the precious metals. It appears to us, however, that some ability to part the metals is implied in the use of the touchstone, for we fail to see what use a knowledge of the ratio of gold and silver in bullion could have been without the power to separate them. The touchstone was known to the Greeks at least as early as the 5th Century B.C. (see note 37, p. 252), and a part of Theophrastus'

statement (LXXVIII.) on this subject bears repet.i.tion in this connection: "The nature of the stone which tries gold is also very wonderful, as it seems to have the same power as fire; which is also a test of that metal.... The trial by fire is by the colour and the quant.i.ty lost by it, but that of the stone is made only by rubbing,"

etc. This trial by fire certainly implies a parting of the metals. It has been argued from the common use of _electrum_--a gold-silver alloy--by the Ancients, that they did not know how to part the two metals or they would not have wasted gold in such a manner, but it seems to us that the very fact that _electrum_ was a positive alloy (20% gold, 80% silver), and that it was deliberately made (Pliny x.x.xIII, 23) and held of value for its supposed superior brilliancy to silver and the belief that goblets made of it detected poison, is sufficient answer to this.

To arrive by a process of elimination, we may say that in the Middle Ages, between 1100 and 1500 A.D., there were known four methods of parting these metals: _a_, parting by solution in nitric acid; _b_, sulphurization of the silver in finely-divided bullion by heating it with sulphur, and the subsequent removal of the silver sulphide in a regulus by melting with copper, iron, or lead; _c_, melting with an excess of antimony sulphide, and the direct conversion of the silver to sulphide and its removal in a regulus; _d_, cementation of the finely-divided bullion with salt, and certain necessary collateral re-agents, and the separation of the silver by absorption into the cement as silver chloride. Inasmuch as it can be clearly established that mineral acids were unknown to the Ancients, we can eliminate that method. Further, we may say at once that there is not, so far as has yet been found, even a remote statement that could be applied to the sulphide processes. As to cementation with salt, however, we have some data at about the beginning of the Christian Era.

Before entering into a more detailed discussion of the history of various processes, it may be useful, in a word, to fix in the mind of the reader our view of the first authority on various processes, and his period.

(1) Separation by cementation with salt, Strabo (?) 63 B.C.-24 A.D.; Pliny 23-79 A.D.

(2) Separation by sulphur, Theophilus, 1150-1200 A.D.

(3) Separation by nitric acid, Geber, prior to 14th Century.

(4) Separation by antimony sulphide, Basil Valentine, end 14th Century, or _Probierbuchlein_, beginning 15th Century.

(5) Separation by antimony sulphide and copper, or sulphur and copper, _Probierbuchlein_, beginning 15th Century.

(6) Separation by cementation with saltpetre, Agricola, 1556.

(7) Separation by sulphur and iron, Schluter, 1738.

(8) Separation by sulphuric acid, D'Arcet, 1802.

(9) Separation by chloride gas, Thompson, 1833.

(10) Separation electrolytically, latter part 19th Century.

PARTING BY CEMENTATION. The following pa.s.sage from Strabo is of prime interest as the first definite statement on parting of any kind (III, 2, 8): "That when they have melted the gold and purified it by means of a kind of aluminous earth, the residue left is _electrum_. This, which contains a mixture of silver and gold, being again subjected to the fire, the silver is separated and the gold left (pure); for this metal is easily dissipated and fat, and on this account gold is most easily molten by straw, the flame of which is soft, and bearing a similarity (to the gold) causes it easily to dissolve, whereas coal, besides wasting a great deal, melts it too much, by reason of its vehemence, and carries it off (in vapour)." This statement has provoked the liveliest discussion, not only on account of the metallurgical interest and obscurity, but also because of differences of view as to its translation; we have given that of Mr. H. C. Hamilton (London, 1903). A review of this discussion will be found in Percy's Metallurgy of Gold and Silver, p. 399. That it refers to cementation at all hangs by a slender thread, but it seems more nearly this than anything else.

Pliny (x.x.xIII, 25) is a little more ample: "(The gold) is heated with double its weight of salt and thrice its weight of _misy_, and again with two portions of salt and one of a stone which they call _schistos_.

The _virus_ is drawn out when these things are burnt together in an earthen crucible, itself remaining pure and incorrupt, the remaining ash being preserved in an earthen pot and mixed with water as a lotion for _lichen_ (ring-worm) on the face." Percy (Metallurgy Silver and Gold, p.

398) rightly considers that this undoubtedly refers to the parting of silver and gold by cementation with common salt. Especially as Pliny further on states that with regard to _misy_, "In purifying gold they mix it with this substance." There can be no doubt from the explanations of Pliny and Dioscorides that _misy_ was an oxidized pyrite, mostly iron sulphate. a.s.suming the latter case, then all of the necessary elements of cementation, _i.e._, vitriol, salt, and an aluminous or silicious element, are present.

The first entirely satisfactory evidence on parting is to be found in Theophilus (12th Century), and we quote the following from Hendrie's translation (p. 245): "Of Heating the Gold. Take gold, of whatsoever sort it may be, and beat it until thin leaves are made in breadth three fingers, and as long as you can. Then cut out pieces that are equally long and wide and join them together equally, and perforate through all with a fine cutting iron. Afterwards take two earthen pots proved in the fire, of such size that the gold can lie flat in them, and break a tile very small, or clay of the furnace burned and red, weigh it, powdered, into two equal parts, and add to it a third part salt for the same weight; which things being slightly sprinkled with urine, are mixed together so that they may not adhere together, but are scarcely wetted, and put a little of it upon a pot about the breadth of the gold, then a piece of the gold itself, and again the composition, and again the gold, which in the digestion is thus always covered, that gold may not be in contact with gold; and thus fill the pot to the top and cover it above with another pot, which you carefully lute round with clay, mixed and beaten, and you place it over the fire, that it may be dried. In the meantime compose a furnace from stones and clay, two feet in height, and a foot and a half in breadth, wide at the bottom, but narrow at the top, where there is an opening in the middle, in which project three long and hard stones, which may be able to sustain the flame for a long time, upon which you place the pots with the gold, and cover them with other tiles in abundance. Then supply fire and wood, and take care that a copious fire is not wanting for the s.p.a.ce of a day and night. In the morning taking out the gold, again melt, beat and place it in the furnace as before. Again also, after a day and night, take it away and mixing a little copper with it, melt it as before, and replace it upon the furnace. And when you have taken it away a third time, wash and dry it carefully, and so weighing it, see how much is wanting, then fold it up and keep it."