Columbarium

A Roman columbarium is a collective tomb lined with tiers of arched niches for cremation urns, taking its name directly from the Latin word for a dovecote. In 1898 the Leipzig writer Rudolf Kleinpaul asked his reader to look closely at an ordinary pigeon house to understand the form: with its many cells it is an artificial rock cave, containing as many openings as nests, each screened off so brooding doves do not disturb one another. That arrangement, Kleinpaul explained, is precisely what the Romans called a Columbarium, “ein Begriff, der auf die gleichfalls Nischenreihen enthaltenden Grabkammern übertragen worden ist,” transferring the name to subterranean chambers built with the identical grid of recesses.

Take the birds out and put the urns in, and the building does not change. A columbarium tomb is a room lined from floor to vault with small openings, set one above another and side by side, each wide enough for a pair of pigeons or for the ashes of one or two people.

Appearance

The chambers are sunk several metres down and roofed with barrel vaults, and every wall inside is filled with niches from the floor up. On the Via Appia the openings are arched, with a marble slab cut to fit beneath many of them and the name of the dead person incised on the slab. Inside sits the urn: pottery, glass, or a small marble chest. Most of the chambers have lost their painted decoration, and the first of the Vigna Codini rooms kept four frieze panels on the narrow faces of its central pier.

The three chambers dug under the Vigna Codini, between the Via Appia and the Via Latina near the Aurelian Walls, are the fullest surviving version of the form. The first was opened in 1840 and measures about 5.65 by 7.50 metres, with a central pier that carries four barrel vaults and gives access to the upper rows. Its narrow faces carry four frieze panels with Dionysian scenes, and the niches all round the walls have been counted at close to five hundred burials, about two urns to a niche. The second chamber, found in 1847, keeps painted stucco, pediments over the niches, floral work and pictures of musical instruments, and a set of portrait busts. The third, found in 1852, is laid out as three corridors around a small garden and holds seven rows of rectangular niches, four urns in each. Estimates of the numbers buried in these rooms run from three hundred to nine hundred and fifty, which is worth knowing before quoting any single figure.

Kleinpaul noticed one further detail that ties the two buildings tighter than the architecture does. Egyptian dovecotes, he wrote, were assembled out of pots, stacked into pyramids that take up a whole upper floor of a house. The excavators at Karanis in the Fayum found exactly that: pigeon nests built of pottery jars set into mud brick, wall after wall of them, with the jar mouth facing into the tower.

Function

A columbarium was a cremation tomb for people outside the elite. The ashes in these walls belong to freedmen and slaves, to the staff of the imperial household, and to traders and tax collectors. A few of the dead were rich enough to add a marble panel or a personalised niche, and the great majority were not: the rows are uniform because uniformity was the point, and the opening and its plaque were the whole monument.

Dorian Borbonus defines the type as “closed, collective funerary monuments that deposit cremation ashes in urns and niches on their interior walls,” and argues in Columbarium Tombs and Collective Identity in Augustan Rome that the form was not handed down by patrons. Burial clubs, the collegia of the Roman lower classes, bought the plots and organised the burials, which is why the interior order is so strict and the epitaphs so brief. Borbonus reads that brevity as an economy of its own: in a room packed with inscriptions, a short and clear text carried further than a long one cut into a roadside tomb.

The type belongs to a narrow window. It appears in the Augustan age, flourishes through the first century CE, and fades in the second, when Rome turned back to inhumation and the niche tomb lost its market. An epitaph in the first of the Vigna Codini chambers gives a terminus ante quem of 10 CE, which puts that room before the end of Augustus’s life. Borbonus reads the later, above-ground columbaria as a change in mood rather than in architecture: once the door was above ground and the niche could be decorated more openly, commemoration turned competitive and the collective form dissolved. He also notes where the habit went next. The Christian catacomb took up the same work of gathering a community’s dead into one ordered place.

The name that covered everything

Latin gave the word to at least three things that look alike, and the dictionary keeps them in one place. Lewis and Short cite the dovecote in Varro, Columella, Pliny and the Digest, and Smith’s dictionary opens by saying that the word “is used to express a variety of objects, all of which however derive their name from their resemblance to a dove-cote.” These include the sepulchral chamber, the beam sockets of a temple architrave, and the channels cut into a water-lifting wheel through which scoops emptied into a trough. Ancient writers named by shape, and a small hole in a grid of small holes was a pigeon hole wherever it appeared.

What the word does not tell us is what a Roman called the tomb on the day of the funeral. Smith’s dictionary records the sepulchral sense and stops there, and modern handbooks use columbarium as the standard name for the type because there is no better one. Whether a freedman standing in the Via Appia chamber would have understood the word as a description of his own grave, or only as a word about birds, remains open. The building is documented in detail; the sentence he said over it is not.

Cross-Cultural Connections

The strongest confirmation that the dovecote and the tomb are one thought comes from a building that is both. The Roman mausoleum at Mazor in Israel, dated by radiocarbon on its mortar to the third or fourth century CE, carries an integrated pigeon installation of about sixty rectangular niches that went up with the tomb rather than being added later. Boaz Zissu, who published the re-examination in 2025, calls it the only documented structure in Israel where a pigeon installation is definitely connected to a burial, and he puts three functions in the same wall: dove breeding for fertiliser, devotion to the Syrian goddess Atargatis-Derketo, and commemoration of the dead. Palmyrene funerary portraits of children holding doves point the same way, to a Syrian tradition in which the bird and the grave belong to each other.

The agricultural dovecote is the commoner half of the family, and it dwarfs the tombs in size. The towers at Karanis were about four and a half metres square at the base, with grain bins in the vaulted basement and at least a thousand nesting places above; Husselman judged an operation of that scale a commercial one rather than a household one. In the Negev, the Byzantine farmers of Shivta built pigeon towers to keep their vineyards in manure. Kleinpaul’s observation about the rock dove explains why the towers look the way they do: the bird nests in caves and holes in cliffs by nature, so a cliff of holes is what a human builder has to imitate, and a wall of holes is what he gets.

The sequence moved from natural habitat to stone: Roman builders imitated the cliff hollows of the wild rock dove for their dovecotes, then adapted that exact lattice of recesses for cremation urns. Two thousand years later the same grid returns in the Catacombs of Paris, where the bones of six million people are stacked in patterns, and in the Vatican Necropolis, where cremation urns and inhumation graves share one Roman alley under the floor of St Peter’s. The Ħal Saflieni Hypogeum in Malta is the older and stranger cousin, a rock-cut tomb for a whole community, cut by hand rather than bricked up in rows.

Modern Survival

The word outlived the practice by a long way. Cremation came back in the nineteenth century and the niche wall came back with it, and the Latin name came back too: a municipal cemetery brochure and a diocesan rulebook both say columbarium, and the structure they mean is a wall of openings about the size of a dovecote’s. The San Francisco Columbarium, built in 1898 by Bernard J. S. Cahill to go with a crematorium he had designed three years earlier, is the surviving monument to that revival. It is the only piece of the Odd Fellows cemetery left after the city stopped burials, and the Neptune Society restored it after decades of neglect; it carries San Francisco Landmark number 209 as well as a place on the National Register.

The numbers behind the revival are plain. The Cremation Association of North America put the United States cremation rate at 61.8 per cent for 2024. In 2016 the Congregation for the Doctrine of the Faith issued Ad resurgendum cum Christo, which requires that cremated remains be kept in a sacred place, and rules out scattering them or dividing them among heirs. Buddhist Asia arrived at the same structure from its own direction: Japan keeps its ashes in a nokotsudo, a bone-receiving hall inside the temple grounds, which is the niche wall under another name.

English took the word as a burial term first. The columbarium that appears in the 1540s is the Roman tomb and nothing else, and the literal sense, a pigeon house, is only recorded in English from 1881. For more than three hundred years the language kept the far half of the metaphor and let the birds go.

Sources

• Marcus Terentius Varro, Rerum rusticarum libri tres III.7, on the building of a dovecote. Translated by Lloyd Storr-Best as Varro on Farming. London: G. Bell and Sons, 1912; text at LacusCurtius, University of Chicago.
• Lucius Junius Moderatus Columella, De re rustica VIII.8, on dovecotes and the vessels the nests are made from.
• Rudolf Kleinpaul, Die Lebendigen und die Toten in Volksglauben, Religion und Sage. Leipzig, 1898. Chapter IV.4, ‘Die heiligen Tauben’, pp. 226-229.
• William Smith, ed., A Dictionary of Greek and Roman Antiquities, 3rd ed. London: John Murray, 1875, ‘Columbarium’, p. 323 (article by Anthony Rich); online at LacusCurtius with a note by Bill Thayer.
• Online Etymology Dictionary, ‘columbarium’, for the 1540s English use of the tomb sense and the 1881 first record of the dovecote sense.
• ‘Columbaria at Vigna Codini’, in Religions of Greece and Rome: Site Reports. New London: Connecticut College (open textbook), for the chamber dimensions, the urns to a niche and the excavation dates, after Coarelli and Borbonus.
• Harry Thurston Peck, Harper’s Dictionary of Classical Antiquities. New York: Harper and Brothers, 1898, ‘Columbarium’.
• Charlton T. Lewis and Charles Short, A Latin Dictionary. Oxford: Clarendon Press, 1879, ‘columbarius’ and ‘columbarium’ for the ancient citations.
• Dorian Borbonus, Columbarium Tombs and Collective Identity in Augustan Rome. Cambridge: Cambridge University Press, 2014.
• Elinor M. Husselman, ‘The Dovecotes of Karanis’, Transactions and Proceedings of the American Philological Association 84 (1953), pp. 81-91.
• Sophia Germanidou, ‘Dovecotes from the Roman and Byzantine Periods: An Overview’, HEROM: Journal on Hellenistic and Roman Material Culture 4.1 (2016).
• Boaz Zissu, ‘The Roman Mausoleum at Mazor: New Perspectives on an Outstanding Monument’, Cercetări Arheologice 32.2 (2025).
• Congregation for the Doctrine of the Faith, Ad resurgendum cum Christo: Instruction on the Burial of the Deceased and the Conservation of Ashes in the Case of Cremation, 15 August 2016.
• Cremation Association of North America, 2024 cremation statistics (United States rate 61.8 per cent).
• Japanese Wiki Corpus, ‘Nokotsudo (A Japanese Ossuary)’, and the columbarium pages of the Tendai Buddhist Institute, for the temple bone-hall.
• Rijksmuseum, Amsterdam: photogravure plate ‘Columbaria, Plate II: Tomb of the Officers of Caesar’s Household’, ca. 1867-1877, object RP-F-2001-7-164-2. Public domain.