Codex Vergara · Tepetlaoztoc, near Texcoco · c. 1540
The Codex Vergara: 386 Aztec fields and the largest area their sides allow
Around 1540, Acolhua scribes near Texcoco drew each household’s fields twice over: in one register with the sides measured in rods, in another with the areas. Four sides do not fix a shape, but they do fix the most area it can hold. Flex a field below and compare that ceiling with the area the scribe wrote down.
- Sides
- Area now
- Ceiling
- Recorded
A survey of households
The Codex Vergara comes from Tepetlaoztoc, an Acolhua town near Texcoco in the Valley of Mexico, a generation after the Spanish conquest. A pictorial census records who lived in each household, and the household’s fields are drawn in two registers: outlines with their side lengths written in glyphs, and again with their areas. Lengths are in a rod, the T of this page, about 2.5 metres; shorter lengths were written as a heart, an arrow and a hand, worth 0.4, 0.5 and 0.6 of a rod.
In 2011 María del Carmen Jorge y Jorge, Barbara Williams, Clara Garza-Hume and Arturo Olvera published 386 four-sided fields, sides and areas together, with the fractions already turned into decimals; they had set aside 22 of 408 records. They then asked how the scribes’ areas compare with what the sides allow.
The most four sides can hold
Hinge four rods into a loop and push one corner: the loop flexes, and its area changes. The area is largest when all four corners sit on one circle. For sides a, b, c and d that largest area is √((s−a)(s−b)(s−c)(s−d)), where s is half the perimeter: Brahmagupta’s formula, from 7th-century India. It is a yardstick, not a claim about how the scribes worked.
A recorded area above the ceiling fits no shape with those sides. Every recorded area under it in this set can be drawn with them. Neither makes a scribe dishonest: the scribes had rules of their own for working out an area, and a rule that is not exact can land on either side. One widespread rule, multiplying the averages of opposite sides, can overshoot the ceiling.
How close the scribes came
Of the 386 fields, the paper counts 250 whose recorded area is at or under the ceiling, 122 of them exactly on it, and 136 above it. Three quarters sit within 5 % of the ceiling, either way; 85 % within 10 %. In 130 fields the opposite sides match, 90 of them squares. Drawn as rectangles, such sides enclose length times width, their largest possible area, and all the exact hits are among them.
This page recomputes every field from its four sides, and checks each against the paper’s two tables: the maximum printed beside the sides, and the gap printed in the second table. 384 fields agree with both. The chart shows the recomputed counts, with Table 1’s in brackets where they differ.
Two fields that do not add up
For two fields the published figures disagree with each other. Field 04-06-002 is listed with sides of 41, 21, 22 and 11 rods and a recorded area of 715. Those sides allow at most 400.4 square rods, and the second table says so, a gap of −78.6 %. But the first table prints a maximum of 498.5, a gap of −43.4 %, and the appendix’s drawing of the field gives the fourth side as 16 rods, which yields exactly that. Table 1’s counts add up only with the field at −43.4 %. Which length the manuscript has, the appendix alone cannot say; this page uses the listed 11.
Field 05-36-004 is a rectangle, 22 by 10 rods, recorded at 220 square rods: exactly the ceiling. The second table gives its recorded area as 22 and its gap as 99 % (22 against 220 would be 90 %); Table 1’s counts are consistent with it among the fields more than 10 % under.
Recomputed from the sides, Table 1 changes by one field in four bins: 123 exact instead of 122, 18 more than 10 % under instead of 19, and one field moves from the −20 to −50 % bin to the −50 to −100 % bin. With the two published figures put back, the recomputation gives Table 1 exactly. The number above the ceiling stays at 136, and no bin moves by more than one field.
What this does not test
The sides come already converted from glyphs to decimals, so the reading of the heart, arrow and hand is not tested here. Nor are the scribes’ own area rules, which Williams and Jorge y Jorge reconstructed in 2008: this page measures the record against a geometric ceiling, not against those rules. For the fields far above the ceiling the 2011 authors suggest mismatched drawings, recording errors and approximate rules; the cause of each stays open.
The page about khipus on this site checks another kind of record, Inka knotted cords, against the totals they carry.
Sources
Every figure on this page is traced in the project’s source notes.
- María del Carmen Jorge, Barbara J. Williams, C. E. Garza-Hume, Arturo Olvera, “Mathematical accuracy of Aztec land surveys assessed from records in the Codex Vergara”, PNAS 108 (37), 15053–15057, 2011, full text, Table 1 and the appendix (Tables S1 and S2, field drawings).
- Barbara J. Williams, María del Carmen Jorge y Jorge, “Aztec arithmetic revisited: land-area algorithms and Acolhua congruence arithmetic”, Science 320 (5872), 72–77, 2008.
- Khipu, on this site.