The printed annual returns of Mexico City, 1796: marriages, births, deaths and a hospital-by-hospital account
Anyone who bought the Mexico City almanac for 1797 got three pages of the city's own vital statistics: marriages, births and deaths in all fourteen parishes, a twelve-house hospital return with the dead in a column of their own, and a resumen adding them up. A viceregal decree of 1791 had made the returns due every December so the volume could appear in January. Every printed total balances but one row.
Why this wins its question: Accounts of statistics in the Spanish empire go to the Revillagigedo enumeration of 1793 and stop, leaving the impression of a single heroic count in a century that otherwise measured nothing. What actually existed alongside it was duller and, for a reader who wants to know whether a state watched its own population, better: an annual return of births, marriages and burials by parish, an annual hospital return with the dead in a column of their own, and a viceregal decree fixing the fifteen days in December by which every body had to send its list. This object prints those three pages as they stand, re-adds every column instead of quoting the totals, names the one row in twelve that does not balance and by how much, and points out that the resumen everyone would cite adds a twelve-month parish year to a calendar hospital year and calls the result 1796.
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 almanac printed the city's births, marriages and burials by parish. Its heading is "Razon del numero de los Matrimonios que ha habido, y de las Personas que han nacido, y fallecido en las catorce Parroquias de esta Corte, desde 1 de Diciembre de 1795, hasta fin de Noviembre de 1796", and it lists all fourteen — Sagrario, San Miguel, Santa Catalina, Santa Veracruz, San Joseph, Santa Ana, Santa Cruz, San Sebastián, Santa María, San Pablo, Acatlán, Salto del Agua, La Palma and San Antonio — in three columns, ending 1,316 marriages, 6,206 born and 3,518 dead (p. 196).
The next page turns to the hospitals, and its column key is more careful than the table looks. "La 1 columna asigna los q̃ hubo en el año, inclusa la existencia del dia último del anterior: la 2 los que salieron: la 3 los que murieron; y la 4 los que existian en 31 de Diciembre" — so the first column is everyone present during the year, not admissions, and the last is a count taken on 31 December. Twelve houses are named: Jesús Nazareno, San Juan de Dios, Real de Indios, General de San Andrés, San Pedro, San Hipólito, Espíritu Santo, Divino Salvador, San Lázaro, San Antonio Abad, Terceros de San Francisco and Belemitas, totalling 14,655 present, 12,655 out, 1,169 dead and 931 remaining (p. 197).
A third page adds the deaths up across the institutions that recorded them: "RESUMEN de los que han fallecido el año de 1796 — En las Parroquias 3518, En Conventos de Religiosos 33, En los de Religiosas 43, En los Hospitales 1169. Son por todo 4763" (p. 198). The figure a reader carried away was therefore not a parish figure but a compiled one, with the hospital dead counted as their own class.
The tables were checked for this object rather than quoted on trust, and they hold. All three columns of the parish table re-add exactly to 1,316, 6,206 and 3,518; all four columns of the hospital table re-add exactly to 14,655, 12,655, 1,169 and 931; the resumen's four lines re-add exactly to 4,763. Eleven of the twelve hospital rows also balance across the page, the year's total equalling those who left plus those who died plus those still there on 31 December.
One row does not balance, and it is the second one. San Juan de Dios is printed as 1,938 present in the year against 1,786 who left, 164 who died and 88 remaining, which come to 2,038. The error cannot be a slip of the totting-up, because every column total on the page is consistent with the figures as printed, including the 1,938: whatever went wrong went wrong in the return before the compositor added the columns. It is a hundred exactly, which is what a digit does.
The returns were not a printer's initiative but an order with a deadline. The nota under the resumen records that "El Exmô. Señor Virrey Conde de Revilla Gigedo por Decreto de 12 de Enero del año de 1791 resolvió, que para que esta Guia de Forasteros salga siempre con el debido arreglo, todos los Tribunales y Cuerpos de que se compone dirijan á su Autor sus respectivas Listas en los quince primeros dias del mes de Diciembre de cada año, para que pudiendose coordinar en los restantes, se verifique su publicacion en principios de Enero" (p. 198).
The hospital column that names the dead also names a house where almost nobody left alive. The Divino Salvador, the city's house for insane women, returns 85 present in the year, 11 who left, 15 who died and 59 still in it on 31 December; San Lázaro returns 106, 6, 21 and 79 on the same four heads. Against them the Belemitas return 1,033 present, 1,000 out and none dead. The table does not explain any of this, and it does not have to: it prints the numbers that let a reader see it.
The two tables do not cover the same year, which matters for the total everyone would quote. The parish table runs from 1 December 1795 to the end of November 1796; the hospital table counts to 31 December 1796. The resumen adds the one to the other and calls the sum "el año de 1796". It is a month out on one of its two components, and nothing on the page acknowledges it.
What this evidence is, and what it cannot settle. These are three pages of an almanac sold for profit, printing what parishes and hospitals said of themselves, with no audit beside them and no method stated; nothing here shows who compiled a parish return or how a house decided that a patient had "left". The resumen adds hospital deaths to parish deaths as though no one who died in a hospital was buried from a parish, and this object cannot show whether that is a real separation or a double count. There is no age, no sex, no cause of death and no population to divide by, so none of it yields a rate; the figures cover the parishes of the capital and say nothing of the valley around it; and the undercount that Humboldt found in the enumeration of 1793 — households shrinking their lists for fear of a tax — has no test here at all.