by Earle R. Caley
The following article originally appeared in Stack's Numismatic Review, 1946, #2 (April). We thank Stack's for contributing this article to the public domain in 1946.
Some time ago Mr. Harry J. Stein sent to me for examination an Alexandrian tetradrachm of abnormally low weight. This billon coin was of the common type issued in the second year of the reign of this emperor. There was no possibility that it was an electrotype and very little possibility that it was a forgery of any kind. The original observation of Mr. Stein was fully confirmed by me for the weight of this coin as determined on an accurate chemical balance was found to be only 7.16 grams. An ordinary specimen of the same type submitted by Mr. Stein for purpose of comparison was found to have a weight of 12.26 grams. The normal weight of Alexandrian tetradrachms of the second year of the reign of Vespasian, as indicated by the average of the weights of those listed in the work of Milne,  is 13.19 grams. Milne lists 15 such coins and their weights range from a low of 11.73 grams to a high of 14.22 grams. Hence the weight of the specimen in question is not only very much lower than the average weight of those in this list but is also much lower than even the lightest one.
It might be supposed that the abnormally low weight of this coin is merely the consequence of unusually small dimensions. However, this is not the explanation. The mean diameter of the coin was found to be about 24 millimeters and its mean thickness about 2.5 millimeters. The average diameter of those listed by Milne is 26.2 millimeters with a low of 24.0 millimeters. Milne does not give any data on the thickness of the coins in his list, but the thickness of this coin as compared to that of other Alexandrian tetradrachms in my own collection is only a little below normal. Since this unusual coin is about normal in size but abnormally low in weight its specific gravity must therefore be much below normal. A careful determination of its specific gravity by the method of Archimedes gave the result 5.80. Pure silver has a specific gravity of 10.50 and pure copper 8.90. Furthermore, from their usual silver content it can be readily calculated that the specific gravity of Alexandrian tetradrachms of the time of Vespasian should not fall much below 9.25. Actually, the apparent specific gravity of the metal of this unusual coin is below that of any normal coinage metal, ancient or modern, with the exception of aluminum and its alloys, and the specific gravities of these modern metals all fall much below 5.80. It might also be supposed that the low weight and specific gravity of the coin is a consequence of the presence of a heavy layer of corrosion products or patina of low density, but actually the coin had a fairly bright metallic appearance and there were indications that it had been cleaned rather recently.
On examining the metal of this coin I first observed that the edge of the flan was so soft in most places that moderate pressure with the finger nail caused easily visible indentations. On viewing scraped parts of the edge under sufficient magnifications I saw at once that the metal had a spongy structure. The faces of the coin appeared to be relatively free from this spongy structure, probably because of destruction of it by polishing, but the metal immediately below the surface of the faces was also generally spongy. In fact this sponginess extended deep into the body of the coin. The explanation for both the abnormally low eight and the abnormally low specific gravity observed in the coin is thus obvious for the metal of which it was composed was for the most part spongy and porous, in other words full of minute cavities.
After examining this coin and obtaining the facts above given, I recalled that I had in my own collection a late Ptolemaic tetradrachm  that had always appeared to be to be exceptionally light in weight though fully of normal size. The actual weight was now found to be only 7.98 grams. A careful determination of the specific gravity of this coin gave the result 5.66 which is close to that of the coin submitted by Mr. Stein, though actually even a little lower. This coin had also been cleaned. Examination of the metal showed that it had the same sort of extensive spongy structure.
The question that now remains to be answered is this: Was the spongy structure observed in these two coins produced at the time of minting or was it a subsequent development? Aside from the small likelihood that coins of such low weight would have been accepted in circulation there is the strong probability on technical grounds that the spongy structure was a subsequent development. Is is not generally realized how deep the processes of corrosion often extend into the metal of ancient coins of billon or bronze that have been buried in the earth for centuries. Though such coins may have only a very thin patina or layer of corrosion products visible on their surface, and thus appear to have been but little affected, the corrosion may in fact extend deep into the metal. This internal corrosion always tends to occur in the boundaries between the grains of crystals of the metal, and frequently the metal of an ancient coin or other metal object is so affected by what is known technically as intergranular corrosion that the actual metal that remains is all or nearly all in the form of numerous isolated grains or crystals wholly surrounded by corrosion products.  When a coin extensively corroded in this way is cleaned chemically with solvents, these solvents dissolve out the intergranular corrosion products to a greater or less degree, depending upon the nature of the solvent and the duration of treatment. If the solvent is active and the time extended nearly all the corrosion products may be dissolved out, thus leaving the whole coin in a spongy condition. The action is similar when electrolytic methods of cleaning are used except that some of the intergranular corrosion products will be reduced to new metal that will in part be plated back on the uncorroded metal in the coin. However, the final result is similar for the cleaned coin will seldom be composed of compact metal. It will be remembered that both the coin submitted by Mr. Stein and the one in my own collection had been cleaned. The methods of cleaning are unknown, but it is probably that one of the usual methods of cleaning silver or billon coins was used, such as the use of ammonia water as a chemical solvent or the simple electrolytic method involving the use of lemon juice and iron tacks.
My explanation for the low weight and low specific gravity observed both in the coin submitted by Mr. Stein and in the one in my own collection may now be summarized. I believe that both these coins were originally full weight and proper specific gravity, that they became extensively corroded internally in the course of long burial in the ground, and that on being cleaned thoroughly by some chemical or electrolytic method the intergranular corrosion products were largely dissolved out, thus leaving behind coins composed of spongy metal abnormally low in both weight and specific gravity.
Certain conclusions of general numismatic interested follow from the results of this special inquiry. One is that the weights of cleaned ancient silver coins of low fineness, which were originally much corroded, should be used with caution in any studies of the ranges of weight of such coins, or in obtaining their average weights, especially when these average weights are to be used for the determination of monetary standards or relationships. I suspect that in the past all too many such coins of abnormally low weight have been included in obtaining the average weight of a particular type of coin and that the generalizations based upon such an average weight are not too reliable. There seems to be a tendency in studies based upon the weights of coins to include the weights of all published specimens. This is justifiable when the coins are of gold, electrum, or even of fine silver, but will billon coins it is evident that a considerable degree of critical selection should be exercised. Another general conclusion is that the specific gravity method for estimating the fineness of ancient billon coins may be grossly misleading when applied to cleaned coins which were originally much corroded. It is also misleading when applied to uncleaned coins of this sort. For example, the coin of normal weight submitted by Mr. Stein along with one of abnormal weight was an uncleaned coin in good condition covered with a thin layer of light green patina. The specific gravity of this coin was found to be only 8.15, which is considerably below that of pure copper, to say nothing of being much below what it should be for an alloy of silver and copper. The low apparent specific gravity of the metal of corroded billon coins is caused by the relatively low specific gravities of the various corrosion products such as silver chloride, copper oxide, and copper carbonate. These general conclusions apply with much less force to ancient silver coins of a high degree of fineness, not only because such coins are less likely to undergo corrosion in the first place but also because they are less likely to be corroded internally. Nevertheless, for the purpose of obtaining highly reliable information based upon the weights or specific gravities of ancient silver coins of any sort it is always safer to include only those coins that show little indication of ever having been corroded.
1. J. G. Milne, Catalog of Alexandrian Coins, Oxford, 1933, pp. 11-12.
2. According to the classification of Svoronos this was a coin of Ptolemy XIII struck around 71 B.C.
3. For a detailed study of the process of intergranular corrosion as it occurs in ancient bronze the reader should consult an article by the author entitled, "The Corroded Bronze of Corinth," published in Proceedings of the American Philosophical Society, Vol. 84, pp. 689-791. Accompanying this article are photomicrographs of the internal structure of ancient bronze which has been seriously affected by intergranular corrosion.