A number of clay tablets survive from southern Mesopotamia carrying architectural plans. Walls are drawn as parallel lines, gaps in them mark doorways, and cuneiform numbers are squeezed into the rooms giving dimensions. Most date to the Ur III period, roughly 2100 to 2000 BC, though the practice is older than that.
A Sumerian building plan on clay from the Ur III period, with walls drawn as parallel lines and dimensions written into the rooms in cubits. The strips around the central image are the tablet's four edges, photographed separately. Credit: CDLI contributors. 2026. Cuneiform Digital Library Initiative. August 26, 2026.
Measurements in cubits, fists and open hands
The unit is the kuš, the cubit, about fifty centimeters. Above it sits the ninda or GAR, twelve cubits, so around six meters. Below it are fractions with names taken from the hand. A fist is a third of a cubit. An open hand is a half.
One plan in the Ashmolean gives a room as one and a half ninda less a fist in length and eleven cubits in width. Another gives one ninda, two cubits and two fists. This is a working notation, precise enough to build from and awkward enough that somebody had to be trained in it.
Seven of the eight checkable plans were drawn to scale
The claim that Mesopotamians designed to scale is testable, because some plans give enough dimensions to compare against the drawing.
Eight of them do. Seven turn out to be drawn to scale. The exception is RTC 145, where the top-to-bottom dimensions are compressed relative to the sides, and not by any consistent factor. Rooms drawn at two, three and three times the wall thickness are given as six, twelve and eight cubits.
The likeliest explanation for that one is mundane. The scribe was fitting cuneiform text into the rooms, and the space the writing needed shaped the drawing more than the building did.
The likeliest use was working out how many bricks you needed
Mesopotamia built in mud brick, and mud brick has to be counted before it is made.
Administrative tablets from the end of the third millennium record this side of the work in detail, with estimates of brick quantities and labour. For the first terrace of the ziggurat at Ur, the figure is a thousand workers over five months producing 7.6 million bricks. A plan with dimensions on it is the document that makes an estimate like that possible.
Very few of these survive, and there may be a reason for that
Given how much was built, the number of surviving plans is small.
Two explanations get offered. One is that plans were destroyed once the building was finished, having served their purpose. The other is that most were drawn on something impermanent, wax tablets being the usual suggestion, and that the clay examples we have are the exceptions rather than the rule.
Whichever is right, the survivors are not obviously special buildings. Most have fairly ordinary layouts, which argues against the idea that only prestige projects got planned on clay.
Gudea had himself carved holding one around 2120 BC
The practice was well established before the Ur III tablets. In the Louvre there is a headless diorite statue of Gudea, ruler of Lagash, seated with a tablet across his lap. On the tablet is the ground plan of Eninnu, the temple of Ningirsu at Girsu, with buttressed walls and towers flanking the doors. Beside it lie a stylus and a graduated measuring rule.
The statue is known as the Architect with a Plan. A second one, Statue F, shows the same arrangement with the rule and stylus present and the tablet blank.
What is worth noticing is that Gudea chose to be portrayed this way. Not with weapons, not receiving tribute, but holding a drawing and the instruments for making it.
What the plans show is a house built around a courtyard
The layouts are consistent. Rooms sit around the edges of an open central court, left unroofed for light and air, with the rooms facing inward. The same arrangement runs through southern Mesopotamian domestic architecture for the better part of two thousand years.
Wall thickness is readable from the drawings and it matters. A wall of one cubit, fifty centimeters, will carry a single storey. The plan published from the Rylands collection indicates walls of around a meter and a half, which is enough for a second floor and for keeping the heat out.
Why subscribe. Most archaeology coverage is a press release with a headline on it. Ancient Content goes to the excavation report, the museum record and the published paper, and says plainly when a claim is thinner than it looks.
That takes time, and paid subscriptions are what buy it.
A researched piece every day
Short documentaries, and full access to the video archive
The complete written archive
Corrections and caveats the wire copy leaves out
Sources. Cuneiform Digital Library Initiative on Ashm 1911-238 and related plans. Battini, L., on RTC 145, Revue d’Assyriologie 2016. Heisel, J.P. (1993), Antike Bauzeichnungen. Bagg, A. (2011). Louvre, statues of Gudea AO 2 and AO 3293. University of Hamburg, Centre for the Study of Manuscript Cultures, on architectural plans on clay.