Almanaque Náutico of the Observatorio de Marina de San Fernando (read in the volume for 1856, printed 1854)
The Spanish navy's annual astronomical ephemeris, computed at San Fernando for the meridian of its own observatory and printed two years ahead of the year it served. The volume for 1856 calls itself number 65 of the collection, sets out the formulae it was calculated from so the navigator can judge how far to trust each table, says the figures were checked either by differencing or by being computed twice independently, and is priced on its title page in silver for the overseas provinces.
Why this wins its question: Accounts of nautical ephemerides in English are effectively an account of the British Nautical Almanac, with the Connaissance des Temps as a footnote; that Spain ran a continuous annual ephemeris of its own from the 1790s, computed at the navy's observatory to a Spanish meridian and priced in silver for the American provinces, is close to invisible. This object works from one volume of the series and takes from it the things that decide whether such a book was any good: the formulae it printed so the navigator could judge each table, the German tables it admits standing on, and the checking rule - differencing where possible, otherwise computed twice by independent hands. It also keeps the volume honest about itself, noting that the series length is an arithmetic inference from its own volume number, that the claim to match foreign almanacs is the publisher's own, and that a title-page price for Ultramar is not a circulation figure.
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 title page states the authority, the place and the two-year lead. It reads "ALMANAQUE NÁUTICO PARA EL AÑO 1856, calculado de órden de S. M. en el Observatorio de Marina de la Ciudad de San Fernando", carries the crossed anchors and the seal of the "Real Observatorio Astronómico de S. Fernando", and is imprinted "SAN FERNANDO. Imprenta y librería Española, á cargo de D. Juan Alvarez, calle Real, número 47. 1854." The volume for 1856 was therefore printed in 1854.
The book was priced for the Americas on its own title page. The foot of the leaf reads "Precio de este Almanaque 25 rs. en la Península y 30 rs. vn. ó 12 rs. plata fuerte en Ultramar" - a peninsular price in reales, and an overseas price quoted in hard silver. The tariff is evidence that distribution to the overseas provinces was routine and priced in advance; it is not evidence of how many copies went, and none of the pages read says.
The prologue dates the series by numbering itself. It opens: "Continuando la publicacion del ALMANAQUE NAUTICO, ó Efemérides astronómicas acomodadas al uso y necesidades del navegante y viagero, damos el del año de 1856, volúmen 65 de esta coleccion, calculado para tiempo medio del meridiano del Observatorio de Marina, establecido en la ciudad de San Fernando." Counting 65 annual volumes back from the year 1856 inclusive puts the first at the year 1792; the volume itself does not print that date, and the inference is arithmetic, not testimony.
It measures itself against the foreign ephemerides openly, claiming coverage rather than priority: the almanac "comprende todo lo mas esencial que se halla en los que salen á luz en otros paises con relacion á los mismos objetos". The comparison is stated by the publisher about itself, and the volume names no foreign almanac and compares no tables.
The prologue's stated purpose is to let the user calibrate his own trust, which is an unusual thing for a table book to undertake: "Para que pueda formarse una idea del grado de confianza que merecen los varios artículos que presenta este almanaque, vamos á manifestar los fundamentos en que se apoyan, y los métodos de cálculo por los cuales se han deducido, entrando sobre esto en las mas indispensables esplicaciones." It then prints the expressions used, including the general precession referred to the mean equinox of 1750 and of 1800 and the mean obliquity of the ecliptic.
The foundations are credited to foreign fundamental astronomy, not claimed as Spanish. The mean obliquity expression is the one "dada por Bessel en su introduccion á las Tab. Regiom. pág. XXVII", and the nutations in longitude and obliquity "se han calculado por las fórmulas de la pág. XV de la misma obra, omitiendo los términos dependientes de 2C en atencion á su cortísimo período". What San Fernando is claiming is computation and verification to a national meridian, and it says which German tables it stands on.
The volume states a checking protocol, which is the part that decides whether an ephemeris can be trusted at sea. The last page of the prologue records that all the calculations, besides being made to types laid down beforehand, "han sido revisados y comprobados por las series de las diferencias de los resultados, siempre que estos se prestaban á tal clase de verificacion, ó llegando á ellos por doble mano y con toda independencia en caso contrario" - checked by differencing where the quantity allowed it, and otherwise computed twice by separate hands.
It also records what the work cost that year. The same page explains that publication was "ya demasiado retardada por la invasion del cólera-morbo que sufrió esta ciudad y el mismo establecimiento durante los meses de Agosto, Setiembre y Octubre del año actual", and is signed "Observatorio de San Fernando 30 de Diciembre de 1854. EL DIRECTOR DEL OBSERVATORIO, Saturnino Montojo". The epidemic reached the observatory itself; no deaths or numbers are given.
The San Fernando meridian was used as a reference from the American side as well as by ships. Writing from Santa Fe de Bogotá, Caldas gave the longitude of his own observatory as so many hours west "del Observatorio real de la Isla de León", the observatory at San Fernando, and a provincial survey of Antioquia reckoned its longitudes from Bogotá referred in turn to the same Spanish meridian. This claim rests on the Semanario, not on the almanac, which nowhere names an American user.
What one annual volume cannot establish should be said plainly. A single volume is its own witness for the length and continuity of the series, for the claim of completeness against foreign almanacs, and for the checking protocol - none of which is independently verified in anything read here. The volume shows a price for the overseas provinces but not a circulation, a stated verification method but not an error rate, and its own search layer indexes an errata table for 1856 whose leaf was not opened.