The unrelenting economic inflation of A.D. 1515–1650 in Europe has been variously explained by the influx of silver from Mexico and the Viceroyalty of Peru (now Bolivia), the growth of the European population, and the decline of the silver market price. Silver, copper, and lead isotope analyses of A.D. 1550–1650 English coinage show a dominance of silver from Europe and Mexico, contrasting with a spectacularly small contribution from the Viceroyalty of Peru. This observation contrasts with the registration of metal production in the mines of the Spanish Americas. Hence the question: Where did Potosí silver go? This novel observation indicates that silver from Mexico was exported eastward, whereas Potosí silver flowed westward. However, aware of the Pacific route of the silver trade, scholars never agreed upon the volumes transported. Our work demonstrates that there was a Potosí-China route, and that it was largely disconnected from the Mexico-Europe routes.
The remarkable period of economic inflation in Europe, known as the Price Revolution, started in A.D. 1515 and lasted for ∼150 yr, with the prices of commodities, such as wheat and hides, increasing by a factor of five in southern England (Braudel and Spooner, 1967; Hamilton, 1934; Munro, 2003; Phelps Brown and Hopkins, 1956), the Duchy of Brabant (Van der Wee, 1978), and Spain (Hamilton, 1928) by 1630. Various factors have been invoked to account for the steep rise in prices. In 1566, the French royal councilor, Jean Cherruyt de Malestroit, assigned the all-too-evident rise in prices to coinage debasement (Munro, 2008). In his famous response to the “Malestroit paradox,” the 16th century French philosopher, Jean Bodin, most famously upheld by Hamilton (1934), formulated the “quantity theory of money,” which considers that the massive influx of silver from the Spanish Americas into the European economy was the prime cause of inflation. The advent of the 16th century coincided with the end of the bullion crisis (Munro, 1983) that had plagued the European economy for decades, and with the European discovery of the Spanish Americas. The mid-1500s were a period of bullion shortage, a major cause for which was the escalating Chinese demand for silver (Atwell, 1982; Flynn and Giraldez, 1995; von Glahn, 2003). The booming world demand for silver prompted the search for new sources in central Europe and in the Spanish Americas, and led to major discoveries, notably in Zacatecas (Mexico) and Potosí (Viceroyalty of Peru, now Bolivia). Demand also stimulated the improvement of metallurgical techniques, which shifted from smelting, an inefficient way of extracting silver from low-grade ores that required more lead than mines could produce (Blanchard, 1976), to the more efficient amalgamation, or patio, method. The modern paradigm revolves around the idea that the adoption by China of silver money by the turn of the 16th century (von Glahn, 2003) bolstered both silver exploration and a reduction of its extraction costs; a globally withering silver market is believed to be as much a cause of the Price Revolution as were the treasure fleets pouring silver into Europe (Doherty and Flynn, 1989).
Lead and copper are dominant minor elements of silver coinage. Lead isotope compositions are the best-known tracers of provenance studies (e.g., Stos-Gale et al., 1995), with isotope compositions of artifacts being compared with those of potential source ores. A more geochemically informed perspective can be obtained from lead isotope compositions recast as geological model ages, U/Pb, and Th/Pb ratios of the geological formations in which the ores are found (Albarede et al., 2012). For example, the vast majority of silver mines from central Europe indicate Hercynian tectonic episode ages (250 Ma) that contrast with the much younger ages of ores from southern Spain (20 Ma) and the Spanish Americas (20–130 Ma). Likewise, the apparent Th/U ratios of Potosí ores are substantially higher than those of Mexican ores. The stable isotopes of copper and silver have also been shown to allow discrimination between silver ores from central Europe, colonial Peru, and Mexico (Desaulty et al., 2011). If, as discussed elsewhere (Desaulty et al., 2011), the original isotopic signatures are left largely untainted by metallurgical processes and reminting, the combination of lead, silver, and copper isotopes therefore provides a powerful tool for breaking down the isotope compositions of any particular silver coin into its provenance components. In order to assess the contribution of different sources of silver to a strong European economy of the 16th century, we determined the lead, silver, and copper isotope compositions of 15 English silver coins of various denominations, minted between A.D. 1317 and 1640 under the reigns of Edward I, Edward II, Edward VI, Mary I, Elizabeth I, James I, and Charles I.
The analytical techniques used in this work include ion chromatographic separation and multicollector–inductively coupled plasma–mass spectrometry (MC-ICP-MS), and have been presented in detail elsewhere (Desaulty et al., 2011). The algorithms used to calculate lead model ages, U/Pb, and Th/U from the measured lead isotope compositions are those described by Albarede et al. (2012). Professional experts validated all the coin assignments.
RESULTS AND DISCUSSION
The Pb, Ag, and Cu isotope compositions of the 15 English coins are listed in Table DR1 in the GSA Data Repository1. The plot of Th/U (κ) versus model age (Fig. 1) is highly informative. All pre-Tudor (pre–Mary I) coins clearly fall in the field of European medieval coins (Desaulty et al., 2011). Their lead model ages (220 Ma and older) indicate a provenance of the ores in the old Hercynian basement, with sources most likely located either in central Europe (Dill et al., 2008) or in the Pennines and the Devon (England) (Moorbath, 1962). With one exception (Charles I coin, sample a), coins minted in the Tudor and the Stuart eras (Mary I–Charles I) contain very little of the high-Th/U component typical of Potosí silver, and plot in a position intermediate between Mexican lead (age <50 Ma and low Th/U) and Hercynian lead (age >220 Ma and low Th/U). Likewise, the copper and silver isotope compositions of the pre-Tudor coins fall in the field of medieval coins as reported by Desaulty et al. (2011) (Fig. 2). Copper and silver isotope compositions of most Tudor–Stuart coins are clearly distinct from those of Potosí silver, and fall in a position intermediate between Mexican and Hercynian European sources.
Two characteristics of the present data set are relevant to the issue of the sources of Tudor–Early Stuart silver coinage: (1) the contribution of typical Potosí silver is, in general, remarkably low, and (2) the contribution of Mexican silver is remarkably high. In order to facilitate the discussion of these two points, it is assumed that the lead in each coin may be broken down into a mixture of three idealized end-members from Hercynian Europe, Mexico, and Potosí (Table DR1; Fig. 1, inset). Proportions inferred from copper and silver isotope compositions are less reliable because the end-members may differ in both isotope compositions and copper/silver ratios.
With the exception of the Mary I sample, significant proportions (>∼30%) of Potosí lead show up only late in Charles I coinage. Nine out of the 12 silver coins from the 16th to early 17th centuries were minted from 50%–90% bullion from Mexico, or from foreign denominations that contained Mexican silver. If geologically old English lead was used during reminting, as is possibly the case for the Charles I sample c coin (ChIc) and the Edward VI sample a coin (EdVIa), these proportions represent a lower limit. The ChId sample contains a majority of Potosí-type lead, but with copper and silver falling in the field of medieval coins, which is most easily explained by the addition of European copper to Potosí silver.
A remarkable aspect of the microeconomic theory of Doherty and Flynn (1989) is that it deals with silver and gold in the same manner as it deals with any other commodity, such as wheat and silk. Even though silver and gold do not decay nearly as fast as other commodities, such a perspective importantly allows the cost of extracting these metals to be factored into economic models. In order to understand how lead isotope compositions of English silver coinage may switch from a Hercynian signature to a Mexican signature in a matter of years, we first considered that the flow of bullion is conservative, whereas, for reasons of debasement, inflation, and circulation of bills of exchange, cash fluxes do not conserve well. In geochemical parlance, the main dynamic control is the domestic silver residence or “flushing” time, τAg, which is the average time spent by silver coinage within the English economy before being exported. Put another way, 1/τAg is a measure of the probability that a unit of silver leaves the country per unit of time. τAg can be calculated by dividing MAg, the total silver content of the domestic stock of money, by the yearly export of silver, Q, and determining the lag of bullion export with respect to import. Mayhew (1995) used the compounded output of English mints to suggest that the money supply in England increased from 1.4 M£ (million pounds) in A.D. 1526 to 14 M£ in A.D. 1670. Meanwhile, imports changed from ∼600,000 £ in A.D. 1559–1561 (Stone, 1949) to ∼4 M£ in A.D. 1670 (Davis, 1954), with the last figure supported by data on taxes levied on imports (O’Brien, 1988). The flushing time on money in England therefore remained nearly constant at ∼3 yr for over a century. Given the prevalence of silver in foreign trade, and the rather narrow range of the bimetallic (silver/gold) ratio, we surmise that the English domestic stock of silver was also worth ∼3 yr of silver export. Such a short residence time is consistent with the apparently abrupt surge of Mexico silver in English silver coinage during the reign of Mary I, whereas European silver had dominated thus far. Because of the challenging character of the analytical procedure, the corpus size is small. Nevertheless, the large throughput of English mints (Challis, 1992) spurred by profits derived from the seigniorage (the difference between the value of money and the cost to produce it), i.e., the taxes added to the total price of a coin, acted to mix the isotopic signal and eradicate from the English silver supply most of the isotopic variability of metals inherited from different sources of metal. In geochemical parlance, the mixing time of the supply of silver being not much longer than its residence time, heterogeneities at any given time should be limited.
The short residence time of silver in the English economy therefore indicates that both the domestic silver supply and exports closely track silver imports, with a few years of delay, and therefore simply reflects a steady increase of foreign trade through time. The price curve (Phelps Brown and Hopkins, 1956) remains remarkably smooth over a scale of 130 yr (Fig. DR2 in the Data Repository); in spite of the Great Debasement of 1542–1551 (Challis, 1975) and of major windfalls, such as the capture of the Duke of Alva’s treasure ships in 1568 and Sir Francis Drake’s raids on silver shipments in Panama in 1572–1573, the inflation rate seems to be particularly insensitive to the military and economical vicissitudes of the moment (Fischer, 1996). Our results therefore demonstrate that inflation was not particularly endogenous to any European country, and was the result of global economic readjustment (Doherty and Flynn, 1989; Flynn, 1978; Munro, 2003). Sixteenth-century Spain and France were experiencing the same steep inflation as the rest of Europe (Hamilton, 1928). In contrast to England, however, an analysis of silver, lead, and copper isotopes in coinage from Spain, Mexico, and Potosí (Desaulty et al., 2011) suggests that neither Mexican nor Potosí silver significantly entered the Spanish monetary supply until the 18th century. Spain was clearly paying for imports with coinage minted outside of domestic mints, and was only reminting European coins or using silver from its own domestic mining. Likewise, the single 1551 French silver coin analyzed by Desaulty et al. (2011) contains only Hercynian lead. This indicates that, in spite of an apparent fast turnover of silver in England, the proportions of silver supplies from different sources were not in equilibrium from one European country to another.
Our data reveal the relative importance of different sources of bullion in the European economy. Expansion of mining in Saxony (Freiberg, Schneeberg), Bohemia (Joachimsthal), and Tyrol (Schwaz), among others, in the late 1400s and the early 1500s was substantial, and produced silver in quantities that were not exceeded by Spanish-American silver until the mid–15th century (Munro, 2003, 2008; Nef, 1941). Our observations on English trade, however, cannot easily be broken down into a sum of bilateral exchanges, as silver may have transited through several countries before it was finally reminted in England. For decades, scholars have documented silver registrations largely from the legajos and expedientes (files and correspondence) of the royal treasury accounts (TePaske and Klein, 1982; Sluiter, 1998). The reliability of the data has been extensively discussed and databases are easily available (e.g., http://www.insidemydesk.com/hdd.html, compiled by R.L. Garner). The unexpectedly modest penetration of Potosí silver into English coinage well into the 17th century contrasts with the prevalence of the Viceroyalty of Peru in silver registrations (Fig. 3). Given the similar isotopic observations made on Spanish reales (coins) from the same period (Desaulty et al., 2011), this leaves us with the nagging question: why did Potosí and Mexico silver travel different routes?
The answer to this question clearly lies in the geography of silver mining and transportation routes. The natural outlets of the New Spain and Zacatecas silver mines, which are separated from the Pacific Ocean by the Sierra Madre, were Mexico City and the harbor of Vera Cruz on the Gulf of Mexico (Bakewell, 1971). In contrast, silver from the Viceroyalty of Peru could not easily be transported eastward through what is now Brazil. Instead, with the help of the tradewinds, it was shipped from Lima to Acapulco (western coast of Mexico). Flynn and Giraldez (1995) assert that, by the 1570s, fiscal reforms in China and the founding of Manila (Philippines) triggered the rapid expansion of the Manila galleon fleet, and forcefully opened the floodgates of silver trade with China. The amount of silver flowing toward China across the Pacific Ocean is not agreed upon (Barrett, 1990; Flynn and Giraldez, 1994; Frank, 1998), but is thought to be huge because it made rather inexpensive silver available in Manila. If the predominance of Mexican silver observed in English coinage is relevant to the rest of Europe, even with large uncertainties, comparison of our data with silver registrations suggests that most Potosí silver was transported via the Pacific route, and that the share of the Pacific trade has been underestimated. The Pacific and Atlantic silver routes remained largely separate, and this pattern of global trade was to last for several centuries. Silver extraction and trade, however heavily taxed, was in private hands; in light of isotopic evidence, Mexico silver appears to have been used to buy commodities from Europe, notably mercury for amalgamation and slaves from Africa, which were both in high demand in silver mines. In contrast, silver extracted from the Viceroyalty of Peru, and in particular from Potosí, was essentially used for the purchase of Chinese goods, including gold, partly to be resold on European markets (Flynn and Giraldez, 1994). Profits realized on both routes combined to boost silver extraction in Latin America, and eventually led to the decline of the silver market, which silver-based European economies regarded as the 16th century Price Revolution.
The Institut National des Sciences de l’Univers, the Ecole Normale Supérieure, and a grant from the program CIBLE of the Région Rhône-Alpes supported this work. Janne Blichert-Toft edited the English of the manuscript. We thank John Munro and a reviewer of an earlier version for correcting substantial misunderstandings. We would like to also thank the reviewers of the present manuscript, and editor William Collins, for allowing such an odd topic into Geology. We dedicate this work to the memory of Chantal Rabourdin, a long-time friend and supportive colleague.