See also libbra, but most of the units that have been spelled either way are on this page.


In Rome and the Roman Empire, a unit of mass, from the mid-19th century taken to be about 327.45 grams. link to a table showing relationships between Roman units of mass This magnitude comes from Böckh's Metrologische Untersuchungen, repeated in the Preface of Theodor Mommsen's influential Römische Geschichte (1854-56) and then by Hultsch.


I have always been interested in the weight of the Roman/Byzantine pound, and in particular in the fact that the best known estimate, namely the one used by Mommsen (327.45 grams), is always quoted to two decimal places. What was so special about this particular value, which many people take as gospel, I thought, and so I finally decided to find out just where this figure came from.

In fact Mommsen got his figure from Böckh's Metrologische Untersuchungen über Gewichte, etc…of 1838 (see Mommsen's Geschichte des römischer Münzwesens (History of Roman Coinage - available on-line), p. xix).

So where did Böckh get his figure? Well, it turns out to be basically an average value derived from a rather mixed bag of just 54 gold coins, as detailed on pp. 164 & 165 of his book.

Actually, Böckh first quotes five heavy stone weights from Herculaneum with 10 lb equivalent weights ranging from 3232 to 3285 grams (there were four 10-lb and one 2-lb weights). The average weight for 1 lb was then 325.88 grams (= 6135.39 Paris Grains), with a standard error (my calculation) of 1.08 grams (I'll keep the 2 decimal places for the moment).

In fact, although Böckh first says these weights were all well preserved, he then says two of them were worn – leaving these out I find a mean 1-pound weight of 326.77 gram with a standard error of 0.88 gram.

He then turns to estimates by others based on gold coins. He quotes a value for the pound of 6154 Paris Grains calculated by Letronne from data on some unspecified gold coins, and also another somewhat different figure, apparently derived from an analysis by Paucker of the same data, of 6165 Paris Grains (= 327.45 grams).

It seems that Böckh himself then calculates, using 27 varying gold pieces of the “Free States”(the same data as Letronne's I think), a pound weight of 6167.16 Paris Grains.

Finally, he calculates a value of 6162.93 Paris Grains from 27 “Constantinian”solidi, and then averages this figure with the previous one to give 6165.04 Paris Grains. (One might well quibble with averaging figures from seemingly very different periods, but there we have it).

At this point Böckh settles on a round figure of 6165 Paris Grains for the pound, i.e, 327.45 grams.

No error estimate is given but from my own experience with solidi and other gold coins the standard deviation in the distribution of the weights in the sample would be around 1%; for the equivalent pound weight, this would amount to circa 3 grams. For a combined sample of 54 items, as used by Böckh, the standard error of the calculated mean would then be circa 0.4 gram.

Thus we have Mommsen's famous estimate of the pound, and now also a reasonable idea of its relative accuracy.

I must also say though that my calculations with gold coins always give rather lower figures for the pound – circa 322 grams for Mars Eagle types, and 319-321 grams for solidi. Why Böckh got such high values I can't say.

Curiously, we note that although Mommsen gives 327.45 grams in his main work (The History of Rome), in his History of Roman Coinage he gives 327.43 grams (p. xix & p. 900) — why this tiny difference I don't know.

I don't know what Böckh included in his 27 gold coins of the “Free States”, but I have ten Oath Scene types of various sizes in my files, with equivalent pound weights ranging from 321.1 to 331.2 grams. Working from these coins gives a mean weight for the pound of 327.0 grams, with a standard deviation of 2.7 grams and a standard error in the mean of 0.9 gram.

It would seem that these types at least were struck on a fairly heavy weight standard. But how Böckh got his heavy estimate of 6162.9 Paris Grains (= 327.3 grams) for solidi I can't understand - a quick survey of the weights of solidi in the CNG archives shows almost no coins weighing more than 4.50 grams, equivalent to a pound of 324.0 grams.

J. R. Glanfield
A post on the blog Moneta-L, 24 and 26 January 2014, reprinted by permission.
In correspondence, Mr. Glanfield pointed out that
“In respect of the stone weights, it's also worth noting that Boeckh chose to work with only five selected examples with a narrow range of 1 lb equivalent weights, although he also lists dozens of weights (in both stone and metal, and from various periods), which show a much wider range of individual 1 lb equivalent weights.
“Note also that Boeckh's calculations of the weight of the pound from gold coins presumably assume no seigniorage, as do my own estimates.”
Google has digitized both of the books cited:


…Sp. Maelium, cui tribunatus plebis magis optandus quam sperandus fuerit, frumentarium divitem, bilibris farris sperasse libertatem se civium suorum emisse, ciboque obiciendo ratum victorem finitimorum omnium populum in servitutem perlici posse…

Titus Livy.
History of Rome, Book 4, 15.

[6] that Spurius Maelius, to whom a tribuneship of the commons should rather be an object of wishes than of hope, a wealthy corn-merchant, had conceived the hope to purchase the liberty of his countrymen for two pounds of corn; had supposed that a people victorious over all their neighbours could be cajoled into servitude by throwing them a morsel of food;

Translation by D. Spillan.
London: Bohn, 1857.
To ilustrate the use of the libra with prefixes to indicate multiples and submultiples.


In Spain, ? – 20th centuries, a unit of mass, varying locally, but the libra most commonly used was the Castillian libra (libra de Castilla), about 460.093 grams, (approximately 1.014 pounds avoirdupois). This is the origin of the value of the libra, 20th century, in Chile, Cuba, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, and Peru.

The table below shows values of the libra in Spain in the mid-19th century.

Province Name of unit Size,¹
Alava libra 460.093
Albacete libra 458
Alicante libra 533
Almeria libra 460.093
Ávila libra 460.093
Badajoz libra 460.093
Baleares (Palma) libra 407
Barcelona libra 400
libra medicinal 300
Burgos libra 460.093
Cáceres libra 456
Cádiz libra 460.093
Canarias libra 460.093
Castellón libra 358
Ciudad-Real libra 460.093
Córdoba libra 460.093
Coruña libra 575
Cuenca libra 460.093
Gerona libra 400
Granada libra 460.093
Guadalajara libra 460.093
Guipúzcoa libra 492
Huelva libra 460.093
Huesca libra 351
Jaén libra 460.093
León libra 460.093
Lérida libra 401
Logroño libra 460.093
Lugo libra 573
Madrid libra 460.093
Málaga libra 460.093
Murcia libra 460.093
Navarra libra 372
Orense libra 574
Oviedo libra 460.093
Palencia libra 460.094
Pamplona libra 372
Pontevedra libra 579
Salamanca libra 460.093
Santander libra 460.093
Segóvia libra 460.093
Sevilla libra 460.093
Soria libra 460.093
Tarragona libra de Gerona 400
Teruel libra 367
Toledo libra 460.093
Valencia libra (but see below) 355
Valladolid libra 460.093
Vizcaya (Bilbao) libra 488
Zamora libra 460.093
Zaragoza libra 350

Even this oversimplifies the situation. Consider just Valencia. According to Doursther², there were four libra in use in Valencia:

1. Gaceta. Madrid, 28 December 1852. Official figures.

Dirección General del Instituto Geográfico y Estadístico.
Equivalencias entre las Pesas y Medidas Usadas Antiguamente en las Diversas Provincias de España y las Legales del Sistema Métrico-Decimal.
Publicadas de Real Orden.
Madrid: Imprenta de la Dirección General del Instituto Geográfica y Estadístico, 1886.

2. Doursther, page 233.


In Colombia, 20th century, a unit of mass = 16 onzas = 500 grams (approximately 1.1012 pounds). An example of seeking round values in the course of metrication.

United Nations, 1966.


In the Dominican Republic and El Salvador, 20th century, = 1 pound avoirdupois (approximately 453.592 grams). An example of the name of an old unit taking on a value from a new trading system.

United Nations, 1966.


In Portugal, ? – 19th century, a unit of mass, approximately 459 grams (approximately 1.012 pounds avoirdupois). Also called an arratel.

From Portugal the unit spread to Brazil A map showing the location of Brazil., 16 – 19th centuries, and survived to the 20th century¹ in the Cape Verde Islands and Paraguay.

1. United Nations, 1966.


In Argentina, a unit of mass, = 16 onza, about 459.35 grams (about 1.0127 pounds avoirdupois). On 18 December 1835, the city of Buenos Aires defined the libra as the weight of 33 cubic pulgadas of distilled water at its maximum density.¹

The libra de boticario, however, was different; according to the Treasury Ministry in 1879, 344.55 grams, which is 12 of the standard onza of Buenos Aires.

1. Juan Alvarez, page 154.


In Patras in the Peloponnese, Greece, 19th century, a unit of capacity determined by mass, used for oil and honey, a nineteenth of a barile, about 3.064 kilograms (6.775 pounds avoidupois).

Doursther page 208.


In Costa Rica, = 460.0627 grams, divided into 16 onzas. (In apothecary’s weight, however, the libra was 345.04704 grams and divided into 12 onzas. Apothecary’s weight was already out of use in Costa Rica in the 19th century.) In modern times, = 460 grams.

[Costa Rica. Ministerio de formento.]
Medidas y pesas del sistema métrico, y tablas de equivalencia con las antiguas.
San José de Costa Rica: Imprenta nacional, 1885.

Technical Factors..., 1972, page 137.


In Chile A map showing the location of Chile., a unit of mass, about 460 grams.

Technical Factors..., 1972, page 132.


In Cuba and El Salvador, a unit of mass, = 16 onzas = 460.00 grams.

Technical Factors..., 1972, pages 140, 155.


In Ecuador, a unit of mass, = 460.00 grams. chart symbol

Technical Factors..., 1972, page 150.


In the province of Alicante, Spain, 19th century, a unit of liquid capacity for oil = 600 milliliters.¹ In concept, this amount of oil would weigh 1 libra.

In the province of Huesca, two units of liquid capacity, one for oil, about 370 milliliters, and one for spirits, about 360 mL.¹ In Navarra, 410 mL.²

Doursther reports (page 236) that the libraof oil in Spain = 1/25 arroba = 4 panillas = 16 onzas, about 502.6 milliliters.

1. Dirección General del Instituto Geográfico y Estadístico.
Equivalencias entre las Pesas y Medidas Usadas Antiguamente en las Diversas Provincias de España y las Legales del Sistema Métrico-Decimal.
Publicadas de Real Orden.
Madrid: Imprenta de la Dirección General del Instituto Geográfica y Estadístico, 1886.

2. Juan Alvarez, page 175.


In ancient Rome, a unit of capacity for oil.


Mesure de capacité employée en particulier pour l'huile. Suétone parle quelque part d'une distribution de dix librae d'huile par personne que fit faire César¹⁷. Il pourrait, sans doute, s'agir de dix livres en poids, et il est bien certain que le nom de libra, donné à la mesure de capacité, vint du rapport établi avec la livre; mais, d'autre part, Horace, dans une de ses satires, fait allusion à la coutume où l'on était à Rome de vendre l'huile dans des mesures en corne¹⁸, et un passage d'un traité de Galenus nous apprend précisément que ce sont elles qui constituaient la λίτρα, équivalent du latin libra, qu'elles portaient tracée une division en douze parties du nom d'onces, et il ajoute qu'il a voulu savoir quel en était le poids.¹⁹ La libra d'huile était équivalente en volume à l'hemina.²⁰ E. Michon.

17. Suet. Caes. 38.

18. Horat. Sat. II, 2, 61.

19. Galenus, De compos. medic. II, 13, VI 13 (Metro. script. t I, p 213); cf. Hultsch, p 111, n. 1.

20. Ibid, p 120.

A measure of capacity especially used for oil. Suetonius speaks somewhere of a distribution of 10 libra of oil per person which was made by Caesar. It could undoubtedly be a matter of 10 pounds weight, and it is certain that the name libra, given to a measure of capacity, came from the established relationship with the libra weight. But on the other hand Horace, in one of his satires, alludes to the custom at Rome of selling oil in corn measures, and a passage in a treatise of Galen tells us precisely that this constitutes the λίτρα, equivalent to the latin libra, which bears traces of being divided into 12 parts, and he added that he wanted to know what the weight was. The libra of oil is equivalent in volume to the hemina.

Ch. Daremberg, Edm. Saglio and Edm. Pottier, editors.
Dictionnaire des Antiquités Grecques et Romaines…, volume 3, part 2, page 1231.
Paris: Librairie Hachette et Cie, 1904.

home | units index | search |  contact drawing of envelope |  contributors | 
help | privacy | terms of use