The water engine of Potosí: the reservoirs and the ingenios de la ribera
Potosí gets rain three or four months a year, so its silver industry ran on stored water. Acosta's Seville imprint of 1590 counts the plant: seven gated reservoirs of about 1,700 varas round and three estados deep, feeding forty-eight water-driven stamp mills on the ribera, four at Tanacoñuño, twenty-two at Tarapaya and thirty horse mills. The mercury and the forced labour are here too, and so is the sediment core.
Why this wins its question: Potosí is usually told as a number — tons of silver, or deaths — and the physical plant that produced the number is left out entirely. This object gives it from the 1590 Seville sheets read as page scans: seven gated reservoirs of about 1,700 varas round and three estados deep on a mountain that gets rain in December, January and February; forty-eight water-driven stamp mills on the ribera, four at Tanacoñuño, twenty-two at Tarapaya and thirty horse mills; six, twelve and fourteen stamps to a mill, mortars fed day and night; and the line that fixes the whole dependence, that at Potosí men pray for a good water year for the silver as elsewhere for the bread. It then refuses both of the usual endings: the mita levy and the 39,000 tons of mercury vapour are stated as their own claim, and so is the sediment core showing that about seventy per cent of the mercury in the lake below Cerro Rico was already there before amalgamation arrived.
Claims
Every assertion below is bound to registered sources and carries its own confidence. Weight them; do not treat the page as uniformly authoritative.
The chapter is libro IV, capítulo 13 of Acosta's Historia natural y moral de las Indias, Seville 1590, headed on page 229 "De los ingenios para moler metales, y del Ensaye de la Plata". The milling exists to prepare the ore for amalgamation — "el metal se muele para recebir el azogue" — and it is done, page 230, by two kinds of machine: "vnos q traen cauallos como atahonas: y otros que se mueuen có el golpe del agua, como aceñas, o molinos, y delos vnos y los otros ay gran quátidad." The stamp mills differ in size, "vnos ingenios tienen a seys maços, otros a doze, y catorze", and the ore is crushed in mortars into which it is fed "dia y noche" before being sieved through copper-wire cloth.
The reservoirs are described as an answer to a climatic constraint, and the description is precise. Because the water at Potosí is what falls as rain and there is not enough of it "sino en tres o quatro meses, q son Diziébre, Enero, y Febrero", they have made "vnas lagunas que tienen de cótorno como a mil y setecientas varas, y de hondo tres estados, y son siete có sus copuertas, y quando es menester vsar de alguna, la alçã y sale vn cuerpo de agua, y las fiestas las cierran." Seven impoundments, each of stated circumference and depth, each gated, drawn on demand and shut on feast days: a stored-water system operated to a calendar. When they fill and the year is wet the milling season runs "seys o siete meses".
The same page turns the arrangement into a proverb about dependence. Because the mills stop when the reservoirs empty, Acosta writes that in Potosí too, and not only where bread is grown, "para la plata piden los hombres ya buen año de aguas en Potosí, como en otras partes para el pan" — men pray for a good water year for the silver the way men elsewhere pray for one for the bread. It is the cleanest contemporary statement that the richest mining camp in the world was, mechanically, a rain-fed installation.
Page 230 gives a count of the installed plant. "Están en la ribera del arroyo de Potosí quarenta y ocho ingenios de agua, de a ocho, y diez, y doze maços: otros quatro ingenios están en otro lado, que llaman Tanacoñuño. En el valle de Tarapaya ay veynte y dos ingenios, todos estos son de agua: fuera de los quales ay en Potosí otros treynta ingenios de cauallos, y fuera de Potosí otros algunos." Seventy-four water-driven mills in three named locations, thirty horse mills in the town, and more outside it, with Tarapaya three or four leagues off where a river runs. Acosta's own summary of the figures is "tanta ha sido la diligencia e industria de sacar plata". Confidence is 0.85 rather than higher because this is one observer's count in one printed book and nothing read here corroborates the individual numbers.
The chapter ends on the assay, which was a royal monopoly of measurement. Silver recovered by quicksilver, Acosta writes on page 229, is so fine "que jamas baxa de dos mil y trezientos y ochenta de ley", so fine that silversmiths and the mints must bring it down with alloy before it can be worked or struck. It "se ensaya y prueua por los Ensayadores y maestros que tiene el Rey puestos, para dar su ley a cada pieça": the bars go to the assayer, who numbers each one because the assay is done on many together, cuts a bite from each, weighs it faithfully and puts it in a cupel. The passage that precedes it is the author's own image for the whole process — "todos estos tormentos y (por dezirlo assi) martirios passa la plata para ser fina" — ground, sieved, kneaded, washed, cooked and recooked through maços, cedaços, artesas, buitrones, tinas, bateas, esprimideros and hornos, "y finalmente por agua y fuego".
The motive is on the record in the same book, in the author's words. Page 198: metals are born in the most sterile and unfruitful lands, in harsh dry country, high sierras, sour crags and disagreeable climates, and "toda quáta riqueza ha venido a España, despues q se descubrierõ las Indias Occidétales, ha sido sacada de semejantes lugares asperos, trabajosos, dessabridos, esteriles: mas el gusto del dinero los haze suaues, y abundátes, y muy poblados." The Jesuit provincial who describes the machinery also states, without embarrassment, what put it there.
What the machinery cost. Robins and Hagan, in Environmental Health Perspectives in 2012, set out the labour and the emissions: the mita levied men between 18 and 50 for twelve months at Potosí and two at Huancavelica; an estimated 39,000 metric tons of mercury vapour were emitted at Potosí between 1564 and 1810, against 17,000 at Huancavelica, the two together more than a quarter of all Latin American emissions between 1500 and 1800; approximately 85 per cent of the mercury brought to Potosí was released as vapour; and widespread contamination is still present in both towns, in ambient soil and in the adobe brick of houses. The authors state their own uncertainties — production records, smelter efficiency, contraband of an estimated 25 to 30 per cent — and the figures are carried here as estimates for that reason. The reservoirs and the stamp mills of the claims above are the front end of that process; they exist to feed it.
A counterweight that cuts against the easy version of the story as well. Cooke and colleagues, publishing in Ambio (40(1):18-25, 2010; the identifier used for this source in this corpus carries a later year), read sediment cores from Laguna Lobato, six kilometres east of Cerro Rico, and found that mercury in the sediment "rises dramatically above background ~1100 AD, in concert with a similar rise in [Ag]" — peak pre-Colonial concentrations about twentyfold above background, fluxes about twenty-twofold — and concluded that "approximately 70% of the total Hg pollution inventory in Laguna Lobato sediment predates the adoption of amalgamation". The mountain was being smelted for silver, and polluting, some four centuries before the Spanish arrived. That does not reduce what the colonial industry did; it does mean the honest claim is intensification and scale, not origination, and this object makes no other.
What these sources cannot establish. Acosta's counts are a snapshot taken by one man for one season around 1590, and the volume itself is written by a Jesuit provincial defending the mission he served in, reporting much Peruvian material at second hand. Nothing read here gives the construction dates of the seven reservoirs, who built them, how many people worked on them, or what a stamp mill produced per day. The engineering described is firmly attested; its history and its labour accounting are not, and are not claimed.