Texcoco

The city stood where the modern town of Texcoco de Mora stands, about 25 km east of the Zócalo in Mexico City, at the foot of the hills that close the Basin of Mexico on the east. In the fifteenth century it faced a shallow, saline lake across a few kilometres of marsh and chinampa plots, with the freshwater lakes of Chalco and Xochimilco beyond to the south-west, and the snow-covered volcanoes of the eastern rim behind it. That position — between the salt water and the mountains, where springs come out of the rock — is the whole explanation of the city's wealth, and of its water engineering.

What follows is the story of a polity that was already old when the Aztec empire was new, that lost its independence twice, and that outlived the empire by a century. It is also a story about evidence. Nearly everything written about Texcoco before the Spanish conquest was written down afterwards, by men who were either its rulers' descendants or their rivals, and the two groups disagreed about almost everything that mattered. The reader should hold that in mind from the start, because it explains why the dates in this entry are usually given with a hedge, and why the most famous episodes — a boy watching his father killed from a tree, a king praying to an unknown god — are the episodes we can least check.

At a glance

Why did a city grow on the eastern shore?

The Basin of Mexico is a closed bowl of about 7,000 square kilometres, ringed by volcanoes and hills, with no outlet for its water. Rain that falls on the surrounding slopes collects in a chain of five shallow lakes at the bottom, and because there is no river to carry it away, it leaves only by evaporation. The lakes were therefore chemically different from one another. Those fed by springs at the southern end, Chalco and Xochimilco, were fresh; those fed by thermal and mineral springs in the north and centre, Xaltocan, Zumpango and Texcoco, were saline. In the wet season they merged into one sheet of water at about 2,242 m; in the dry months they separated again into pools ringed by marsh.

That chemistry set the pattern of settlement. Fresh water meant chinampas — the artificial plots built up from lake mud between stakes, which could produce several maize crops a year — and chinampa country was the densest farmland in Mesoamerica. Salt water meant salt, which was taxable, and waterfowl, fish and the larvae and eggs of the lake insects that everybody ate. The eastern shore had something else as well: the only substantial run of springs close to the lakeshore, at the foot of the Sierra, and a slope behind the town that could be cut into terraces and watered by gravitation. Texcoco's later fame as a garden city was not a whim of its kings. It was the geology of the place, used well.

The site also controlled the roads. From the eastern shore, tracks climbed through the pass at Tepetlaoztoc towards the plains of the Gulf and the obsidian deposits around Pachuca, and north-east towards Tulancingo and the Huastec lowlands. An altepetl that held Texcoco taxed the trade of the whole eastern corridor. When the Acolhua kings later claimed a domain that stretched to Tollantzinco and Cuauhchinanco, they were describing the hinterland of that road, not a random set of towns.

What the town looked like before the fifteenth century is hard to recover. The Spaniards levelled much of it, the modern town covers the rest, and the lake has been drained. Archaeology in the Texcoco region has concentrated on settlement patterns rather than on the ceremonial centre; Jeffrey Parsons' survey of the 1960s and 1970s remains the basic work for the countryside, and it showed a dense scatter of villages along the shore terraces. Of the pre-conquest city itself, the visible remains are the hillside works at Texcotzingo, a few platforms and walls in the town, and the great wall at Huexotla a few kilometres south — one of the last standing structures of its kind in the valley, and a reminder that the Acolhua built in stone on a scale the lake towns did not.

Who were the Acolhua?

The Acolhua were one of the Nahua-speaking peoples who moved into the Basin of Mexico in the centuries after the collapse of Tula, the Toltec capital, in the twelfth century. Their name is attached to the whole eastern province, Acolhuacan, and their language and gods were the same as those of the Mexica who built Tenochtitlan; the difference between them was political, not cultural. Sixteenth-century accounts put their arrival at around 1200, a date modern scholars reach by averaging the figures given in several annals — which is another way of saying that the annals do not agree.

Their own story of where they came from is told in pictures. The Codex Xolotl, painted at Texcoco around 1541 and now in the Bibliothèque nationale de France, is a cartographic history: a map of the valley across which a sequence of events is laid out, with place glyphs, footprints, rulers seated on mats, marriages drawn as lines between houses, and battles marked by shields and arrows. It opens with the arrival of Xolotl, the Chichimec chief, and follows his descendants as they settle, marry into local dynasties and fight over the shore. Two companion documents, the Mapa Quinatzin and the Tlohtzin map, cover the same ground and were probably painted in the same workshop in the 1540s. Together they are the foundation of everything written about Texcoco, and they were made by the family that had the strongest interest in the story coming out a particular way.

Those documents present the Acolhua as heirs to the Chichimecs — hunter-gatherers from the north who came down into the settled valley, took the lands of the survivors of Tula, and gradually adopted the civilised ways of the people they conquered. It is a familiar Mesoamerican formula, and historians read it as genealogy dressed as migration. The names, the sequence of rulers and the marriage alliances are probably genuine; the dates are not. Xolotl is given a reign of more than a century in some accounts, which is the annals' way of saying that several rulers have been compressed into one founding figure. What can be said with more confidence is that the Acolhua were established on the eastern shore by the thirteenth century, that their earliest important centre was at Tenayuca on the western side of the basin, and that their principal town on the eastern shore before Texcoco was Coatlinchan.

Their language story is interesting too. The ruling family of the Acolhua may have spoken Otomi rather than Nahuatl at first; Otomi is still spoken in the villages east of Texcoco, and the sixteenth-century sources say it was the Acolhua king Techotlalatzin who ordered his court to speak Nahuatl. Whether that is a memory of a real language shift or a later justification for Nahua dominance, the direction of the change is not in doubt: by 1500 the Acolhua court was as Nahuatl as Tenochtitlan's, and the Texcocan variety of the language was regarded as the most elegant in the valley. Diego Muñoz Camargo, writing in the 1580s, called it "courtly and refined".

How did Texcoco become a capital?

The move came with a ruler called Quinatzin, whom the annals place in the late thirteenth and early fourteenth centuries. His father Tlotzin had ruled at Tenayuca; Quinatzin, according to the Codex Xolotl and the histories that depend on it, transferred the seat of Chichimec power eastwards to the shore and reduced Tenayuca to a secondary town. The date usually given is around 1337. That is late for a founding, and it fits a pattern visible across the basin: the fourteenth century was when the lake towns sorted themselves into a few dominant centres — Azcapotzalco in the west, Culhuacan in the south, Texcoco in the east — each absorbing its neighbours' tribute and manpower.

Quinatzin's Texcoco was not yet a great city. It was the seat of a regional lord who claimed the Chichimec title and commanded the loyalty of a cluster of shore towns, with a palace and a temple platform and a market that drew canoes across the lake. His son Techotlalatzin, who reigned from 1357 or 1377 — the sources differ by twenty years, and the difference has never been resolved — did more to make it a capital. He is credited with the Nahuatl language reform, with the first substantial building at the ceremonial centre, and with a diplomatic marriage to a daughter of the lord of Coatlinchan. He also had to live with the Tepanecs. Azcapotzalco, on the western shore, was the strongest power in the valley under its long-lived king Tezozomoc, and Techotlalatzin's domain, whatever its own propaganda said, was within its sphere.

How far that sphere reached is a live question. Jerome Offner's reassessment of the extent and structure of Techotlalatzin's "empire" is the standard starting point, and the drift of modern work has been to shrink it: the Acolhua of 1400 were one of several substantial powers around the lake, not the masters of the eastern half of Mesoamerica that their descendants' histories describe. The Texcocan account, written down in the sixteenth century, has an interest in making the pre-Tepanec past look like a golden age, because that makes the Tepanec conquest a catastrophe and its reversal a restoration. Some of that is surely true; all of it is surely not.

By 1400, then, Texcoco was a real capital with a dynasty, a monumental centre and a claim to the Chichimec legacy. It was also, in the phrase the sources use, a place that had to send tribute to Azcapotzalco. The two facts together explain the next fifty years.

What did the Tepanec war of 1418 destroy?

Ixtlilxochitl, who came to the Acolhua throne in 1409, is the first Texcocan ruler about whom we can write a connected narrative, and the narrative is a tragedy. He was the son of Techotlalatzin and, in the annals, a man determined to stop paying tribute. The immediate provocation is given variously: he married a Mexica princess, Matlalcihuatzin, daughter of the king of Tenochtitlan, instead of a daughter of Tezozomoc, which was an insult; he took the title Chichimeca Tecuhtli, Lord of the Chichimecs, which was a claim to precedence; and he began building alliances among the shore towns. In 1415 the quarrel became a war, when Tepanec forces attacked Iztapaluca on the lake's southern shore. Ixtlilxochitl marched to its relief with a reported 4,000 men, and then laid siege to Azcapotzalco itself.

The siege failed. Texcoco's lake squadron was intercepted by the ships of Tlatelolco, and after four years of inconclusive fighting Tezozomoc resorted to treachery: he offered peace, and set an ambush at a hunting preserve at Chiconauhtla. Warned in time, Ixtlilxochitl sent a kinsman dressed in his robes, who was seized and tortured to death in his place. Then the Tepanec army came on the city in earnest. Most of Ixtlilxochitl's subject towns changed sides; only Huexotla, Iztapaluca and Coatepec came to his aid. The city was invested, and the king fled into the hills above it with his household.

The date of his death is 24 September 1418 in the chronology worked out by his descendant and namesake, the historian Fernando de Alva Ixtlilxochitl, and 1419 in others. The manner is agreed in outline: he was cornered in the foothills of Mount Tlaloc and killed, and his son Nezahualcoyotl, then sixteen, watched from the branches of a tree. The scene is one of the most reproduced images in Aztec historiography, and it comes from the Codex Xolotl, painted more than a century later by the same family. It may be true. It is certainly art.

What the war destroyed was not just a king. Texcoco was partitioned and handed to the victors: the sources say Tezozomoc installed two governors over the city, and that its tribute was assigned to Tenochtitlan, which had fought on the Tepanec side and was rewarded with its neighbour's revenues. The Acolhua state ceased to exist as a political unit for a decade. Its ruling family survived in exile, its subject towns were redistributed, and its palace — the building that held whatever records the chancery kept — passed into other hands. When the histories talk about the loss of Texcoco's memory, this is the first of the two catastrophes they mean.

How did Nezahualcoyotl get his city back?

The years between 1418 and 1428 are the part of the story that turned a defeated prince into a national hero, and they are also the part with the least independent documentation. What follows is the Texcocan version, which is the only connected one.

Nezahualcoyotl spent the first months hunted. He hid in ravines and maize fields, was sheltered by his father's former servants, and crossed to Tenochtitlan by canoe after a scouting party sent by Itzcoatl, his mother's uncle, found him. His aunts then negotiated with Tezozomoc for his life, and the old king agreed to let him live in Tenochtitlan under supervision rather than have him killed. He stayed there for several years, probably educated at the calmecac, the school for noble youths, and was eventually allowed to live in a family palace at Texcoco — a prisoner on a long leash. In 1427 — 1426 in some accounts — Tezozomoc died, and everything shifted.

Tezozomoc had designated one son, Tayatzin, to succeed him at Azcapotzalco, and left another, Maxtla, at Coyohuacan. Maxtla took the throne by force, killed his brother, and imprisoned Chimalpopoca, the king of Tenochtitlan, who had been party to a plot against him. Chimalpopoca died in captivity — murdered according to the official version, a suicide according to a minority of sources, and according to a recent argument the victim of a coup inside Tenochtitlan itself. His successor Itzcoatl turned his death into a declaration of war. Nezahualcoyotl, who had by then taken refuge with the lords of Huexotzinco and Tlaxcala, joined the coalition with Acolhua exiles and dissident Tepanecs from Coyohuacan.

The campaign of 1428 was fought on both sides of the lake. The coalition — Tenochtitlan, Texcoco, Tlacopan, Huexotzinco, Chalco and Tlaxcala, in the Texcocan telling — took Acolman and Otumba in the north, then divided: one army destroyed Azcapotzalco, the other cleared the Acolhua towns and entered Texcoco. Maxtla was besieged and killed. The Tepanec empire, which had dominated the valley for two generations, disappeared in a season, and the three victors who remained after Huexotzinco withdrew — Tenochtitlan, Texcoco and Tlacopan — signed the agreement that historians call the Triple Alliance.

Nezahualcoyotl was not crowned at once. The sources put his enthronement in 1431, three years after the war, and the interlude matters: during it the Acolhua towns were reoccupied, their lords reinstated or replaced, and the new king's authority negotiated town by town with the backing of Itzcoatl. When he finally took the throne he was, in a formal sense, Tenochtitlan's nominee. That fact shaped Texcoco's position for the rest of its existence.

What did the Triple Alliance give Texcoco?

The Triple Alliance was a division of the world into three shares, and Texcoco got the second-largest. Tribute from conquered provinces was split in fifths: two to Tenochtitlan, two to Texcoco, one to Tlacopan. Land taken in war was held in common, and the three kings took the title huetlatoani, "elder speaker", in rotation. Each of the three also had a sphere of its own inside the valley: Tenochtitlan the southern lake towns, Tlacopan the Tepanec west, and Texcoco the Acolhua east, with its cluster of subject cities — Otumba, Huexotla, Coatlinchan, Chimalhuacan, Tepetlaoztoc, Chiautla, Tezoyuca, Teotihuacan, Acolman, Tepechpan, Chiconauhtlan — and, beyond the mountains, Xicotepec, Cuauhchinanco and Tollantzinco.

On paper that is parity with Tenochtitlan. In practice the alliance became an empire with one capital, and Texcoco's share of the glory was larger than its share of the power. The two cities were 25 km apart across the water and their kings were kinsmen, which made the partnership work and made it unequal: Tenochtitlan grew to 212,500 people by 1519 on Michael Smith's estimate, Texcoco to 24,100. The Mexica supplied the ideology of conquest, the greater part of the army and the final say in war. Texcoco supplied legitimacy — the Chichimec title, the legal learning, the genealogies that connected the alliance to the Toltec past — and a second court where a Mexica king's rivals could be sheltered and a second opinion could be found.

Texcoco also fought its own wars. Nezahualcoyotl's campaigns are recorded in his own annals and in the painted histories of other towns: Coatepec, Acolman, Teotihuacan and Zempoala in the north, where the Relaciones geográficas of the 1580s remembered Acolhua rule lasting into his reign; Tulancingo in 1450, according to Chimalpahin; the Huastec country around Tuxpan and Xicotepec in the 1440s, if the Códice de Xicotepec is read that way; and Chalco in 1454, an ugly conquest that a Chalca war song remembers with the line "you destroy the realm: you ruin Chalco here on earth", addressed to Moctezuma and Nezahualcoyotl together. Acolhua armies marched with the Mexica everywhere, and Acolhua captains took captives for the temples of both cities.

The alliance held, with friction, until 1521. It survived the disastrous famine of the 1450s, the floods of the 1440s and 1490s, and the succession crises of three cities. Its fault lines were visible to everyone: Tlaxcala in the east and the Tarascan state of Michoacán in the west were never conquered, and the flower wars fought against them were as much a political ritual as a military campaign. Texcoco's own loyalty was tested in 1515, when its king Nezahualpilli died and the Mexica king Moctezuma II intervened in the succession, splitting the Acolhua realm between two of the dead king's sons. That intervention is the reason Texcoco was divided when Cortés arrived, and the reason one of those sons was willing to ally with him.

How was the Acolhua state run?

An altepetl was not a territory so much as a people with a place: a population organised into calpolli, ward-like corporations that held land, paid tribute, supplied labour and soldiers, and ran their own schools and temples. Texcoco was a confederation of such wards under a king, and the wards were not equal. The city's craftsmen were organised by trade into more than thirty categories, each living in its own quarter — featherworkers, goldsmiths, lapidaries, painters, potters — a resettlement policy attributed to Nezahualcoyotl and probably copied from Tenochtitlan. The palace fed and housed them, and their products went out as gifts, diplomatic payments and tribute to the other two allies.

Above the wards sat the king, and beside him four councils whose existence we know from the Texcocan chroniclers. There was a council of war, presided by the heir and by the ruler of Teotihuacan, which judged military offences. There was a council of government and justice, presided by two of the king's brothers, which heard cases involving nobles and commoners. There was a council of finance, which managed tribute. And there was the Council of Music, which supervised the poets, the historians, the astrologers and the painters — the closest thing in the Americas to an academy, and the institution that most impressed the Spaniards who heard about it.

The province was administered by calpixque, tribute collectors of commoner origin, each responsible for a district. Charles Gibson's reconstruction, followed by Frederic Hicks, counts seven or eight such districts around the capital — Texcoco itself, Atenco, Tepepolco, Axapochco, Quauhtlatzinco, Tetitlan, Tecpilpan — each with about a dozen subject villages, the largest with twenty-seven. The lists of what they owed survive in the Texcocan histories and are staggering in their detail: the districts of Texcoco and Atenco each owed, for seventy days a year, up to 400,000 tortillas, 32,000 cacao beans, 100 turkeys and measured quantities of maize, all to feed the palace and the councils. José Luis Martínez, who edited the modern edition of the chronicles, called such figures "excessive if not imaginary"; Hicks' point is that they are not records of delivery but assessments, and that they show a state capable of thinking in hundreds of thousands.

The palace itself, as the Mapa Quinatzin draws it, was a compound of courts and halls on several levels, with the king's apartments above and the council rooms below, and a courtyard in which the lords of the subject towns were shown seated in order of precedence. Ixtlilxochitl, who had the plan in front of him, counted more than 300 rooms. Excavation cannot confirm the number; the terraces under the modern town have never been systematically dug. What the drawing does confirm is the political geography: a kingdom arranged as a building, with the king at the top, the councils in the middle, and the subject lords arranged around a court according to rank.

What did Nezahualcoyotl's laws actually do?

Texcoco's reputation as a state of law rests on a code of eighty ordinances attributed to Nezahualcoyotl and described in the sixteenth and seventeenth centuries by Motolinía, Torquemada, Pomar and Ixtlilxochitl. The code covered treason, theft, adultery, homicide, drunkenness, the misuse of inheritances and misconduct in war. Punishments were severe and public: hanging for a housebreaker, death for adultery, death for the nobleman who drank in public. The third leaf of the Mapa Quinatzin paints a grid of crimes and their penalties, which is the nearest thing we have to an illustrated statute book from the Americas, and it agrees with the written descriptions closely enough that the code is not simply a chronicler's invention.

What is less certain is how much of it was Texcocan. The Texcocan writers — Pomar and Ixtlilxochitl, both descendants of the royal line — insisted that Tenochtitlan copied Texcoco's law, and that the Mexica sent difficult cases to Texcocan judges. Jongsoo Lee has shown that this claim appears only in those two writers. Motolinía and Torquemada, who knew both cities, described a legal culture shared across the valley and saw little distinction between them. Lee's conclusion is that Nezahualcoyotl rebuilt a legal system after 1428 using Tenochtitlan as his model, where he had spent his adolescence, and that his achievement was that of a restorer rather than a Solon. What is not in doubt is the reputation: Texcocan judges were respected across the valley, and Motolinía says cases were sent to them from other cities, except those concerning war.

The stories about the code's enforcement have the flavour of founding legend. Motolinía says the king had four of his own sons executed for adultery with his concubines, and other sources add sons put to death for treason. Such stories are the common property of Mesoamerican dynastic history — the ruler who kills his kin to show that law applies to the great — and they are unfalsifiable. The Texcocan kings were certainly capable of killing their relatives; the question is whether a specific execution happened because a statute demanded it or because a throne was contested. What can be said is that the alliance's other cities accepted Texcocan legal pre-eminence in civil matters, which implies a real body of learning, and that the code survived the conquest long enough for Spanish friars to describe it in detail a century later.

What did Texcoco build?

The engineering is the part of Texcoco's history that can still be walked. The hill of Texcotzingo, about 6 km east of the modern town, was cut into terraces and plumbed with canals in the middle of the fifteenth century, and enough of the rock cutting survives to reconstruct the system. Water came from springs on the flanks of Mount Tlaloc, the nearest of them at Yeloxochitl about 8 km south; the reconstructed channel from that spring runs for 19.5 km around the contours of the hills, dropping gradually so that it arrives at the gardens about 55 m below the summit. Where a ravine crossed the line, the builders filled it with a stone conduit: the largest surviving aqueduct on the site is 170 m long, 3 m wide and up to 3 m high, with two channels that branch at the western end to ring the hill.

On the summit and the terraces were the pools. The largest, cut from a single block of rock and known since the nineteenth century as the Queen's Bath, holds about 12,600 litres; a smaller one with a seat carved inside, the King's Bath, holds about 1,000. Whether they were baths, ritual basins or header tanks for irrigation is argued — the excavators who have measured the flow rates think the system needed a control structure to slow water arriving at 3.8 metres a second, and the terraced pools would serve that purpose. Around them were planted the flowers and trees that the tribute rolls brought in from the hot country: magnolias, plumeria, cacao, the hand-flower macpalxochitl, and the white magnolia the Nahua called eloxochitl, one of the plants Prescott's readers knew from his description of the gardens.

The same engineering was applied to Tenochtitlan. After the floods of the 1440s — the first serious inundation in the city's recorded history — Nezahualcoyotl designed the great dike that ran from Atzacoalco to Iztapalapa and cut the lake in two, separating the saline northern water from the fresher southern body that surrounded the island capital. The date is given as 1445 or 1449. A second project, the aqueduct that brought spring water from Chapultepec into the city along two parallel stone channels, was begun in 1454 according to Chimalpahin and finished in 1466; the Annals of Cuauhtitlan place the start nine years later. It was Texcocan labour and Texcocan design, and it worked: the Spaniards found it still running in 1519, and Cortés cut it during the siege of 1521 to deny the defenders drinking water.

The dike and the aqueduct are usually presented as gifts from a philosopher-king to his uncle, and there is no reason to doubt the collaboration — Moctezuma I was Nezahualcoyotl's kinsman and ally, and the works served both cities. But they are better understood as the routine business of two states that had learned to manage a closed basin, and their failure mode is instructive: when a later king, Ahuitzotl, tried a similar aqueduct from Coyoacán in 1499, the works broke the following year and flooded the city, and the king died soon afterwards. The Texcocan works were not magic. They were the product of a long institutional memory of floods, and of a court that kept engineers.

Did Nezahualcoyotl write the poems?

The poems attributed to Nezahualcoyotl are the best-known body of Nahuatl verse in the world, and their authorship is the most contested question in Aztec literary studies. They survive in two colonial manuscripts: the Cantares Mexicanos, a collection copied between about 1585 and 1597 from earlier originals, and the Romances de los señores de la Nueva España, which was preserved with Juan Bautista Pomar's Relación de Tezcoco of 1582. Some poems carry the heading "Song of Nezahualcoyotl" or name him as the singer; others are clearly by other hands. A poem labelled "Song of Nezahualpilli" is explicitly said to have been composed by a singer called Tececepouhqui, which shows that a royal name in a heading need not be the author's.

The nineteenth-century argument set the terms. Alfredo Chavero doubted in the 1880s that any of the poems were pre-conquest; Daniel Brinton replied in 1890 that they were genuine, written down soon after Nahuatl was alphabetised. Ángel María Garibay and later Miguel León-Portilla treated the first-person poems as Nezahualcoyotl's own; John Bierhorst, who translated the Cantares in 1985, pointed out that sixteenth-century historians mostly did not describe the king as a poet, that the attribution becomes firm only with Torquemada and Ixtlilxochitl in the seventeenth century, and that the conventions of Nahuatl song make authorship a slippery category. The poems themselves are certainly pre-conquest in idiom, whatever their authorship: they are built on the paired metaphor of in xochitl in cuicatl, "flower and song", and they circle the same subjects — the transience of life, the fragility of the world, the drum and the dance — that Nahua sages taught in the calmecac.

One poem in particular has done the king's reputation great damage. In 1778 José Granados y Gálvez printed a Spanish verse dialogue attributed to Nezahualcoyotl containing the line "the round world is but a sepulchre". It is a forgery, adapted from an Otomi folk song, and it was picked up by Prescott and by half the nineteenth century as proof that the king had reached a European philosophy by himself. León-Portilla's verdict is blunt: Nezahualcoyotl would not have used such metaphors, which are foreign to the thought of ancient Mexico. The real poems are stranger and better than the forgery — less consoling, more interested in the moment when the flowers are cut.

The same caution applies to the king's religion. Ixtlilxochitl and Pomar both describe him as a sceptic who doubted the idols, sacrificed to them unwillingly, and finally built a temple at Texcotzingo to "the unknown God, the Cause of Causes", with no image inside and no blood sacrifice. Prescott turned this into a set-piece, complete with a speech in quotation marks. Modern scholarship treats the story as a late construction with obvious Christian purposes, produced by a family that needed its ancestor to look like a proto-Christian nobleman. Lee's argument is that a Mexica king's duties required him to maintain the sacrificial order, not to oppose it, and that Nezahualcoyotl did exactly that: he went to war, took captives, and dedicated temples. The philosophical king and the killing king are the same man, and the sources that separate them are all colonial.

What was in the archive, and who burned it?

Texcoco kept books the way a cathedral keeps relics: as a claim on the past. The Nahuatl word is amoxtli, a painted and folded screen of deer hide or fig-bark paper, and the room that held them was the amoxcalli. Ixtlilxochitl, writing in the 1620s, says the palace at Texcoco held "the royal archives of all the things recounted", and explains why: the city had been "the metropolis of all the sciences, usages and good customs", and its kings "were the legislators of this New World". The claim is family pride, but it is supported by the use later writers made of Texcocan materials, and by the survival of three cartographic histories from a single Texcocan workshop in the 1540s.

We do not know how many books there were, and the numbers in circulation are worthless. Popular accounts speak of thousands, or hundreds of thousands, of codices; no inventory, no excavation and no contemporary description supports a figure. What can be said is that the archive was a working collection of genealogies, land records, tribute assessments, calendars, ritual manuals, songs and histories, that it was consulted by the next generation of Texcocan writers, and that a handful of pre-conquest manuscripts survive from the whole of Mesoamerica — a few dozen at most, none of them from the Texcocan palace itself.

The destruction has a documented date and a documented agent, and neither is the one usually given. On 31 December 1520 — a Monday, in the Julian calendar the Spaniards still used — Cortés entered Texcoco with his Tlaxcalan and Huexotzincan allies. The city's king, Coanacoch, had left by canoe for Tenochtitlan the same afternoon, and the population began to scatter after him. Cortés, suspecting a trap, wanted to sack the city and was dissuaded by Ixtlilxochitl II, who pleaded for the townspeople. He could not hold his allies. Ixtlilxochitl's account is precise: the Tlaxcalans and the other friends Cortés brought "looted some of the principal houses of the city, and set fire to the most important part of the palaces of King Nezahualpiltzintli, so that all the royal archives of the whole of New Spain were burned". He adds that this was one of the greatest losses the land ever suffered, because with them the memory of its antiquities perished.

That is the primary account, written by a man who had the surviving fragments in his house. It is corroborated in substance by Pomar, writing in 1582, who says the descendants of Nezahualpilli still wept over the loss of "their paintings in which they had their histories", having been left "as if in the dark, without notice or memory of the deeds of their ancestors". Neither writer blames a bishop. The attribution to Juan de Zumárraga, the first archbishop of Mexico, appears later and is a compound of two separate things: the inquisitorial trial of Don Carlos Ometochtzin, a Texcocan noble burned for idolatry in 1539, after which Pomar says the remaining family papers were burned by their owners for fear of the archbishop; and the general campaign against native books that the friars conducted elsewhere. Joaquín García Icazbalceta demonstrated in the nineteenth century that the archbishop was not responsible for the palace archive, and the most recent study of Texcocan writing, by Daniel Martínez González, treats the "Zumárraga burned the library" story as a later accretion.

The loss was not total. Some books survived in private hands, and the generation born before the conquest — Ixtlilxochitl's grandparents' generation — rebuilt the record in new forms: the Codex Xolotl, the Mapa Quinatzin, the Tlohtzin map, the Códice de Xicotepec, the Codex en Cruz, all painted in the 1540s to 1570s by scribes who had learned the old conventions and were now working for Spanish courts as well as native ones. Texcoco's memory survived because its aristocracy needed it to survive, to prove rank, land and services to the crown. That is a thinner thread than a library, but it held.

What did the conquest do to Texcoco?

The last years of independence were a civil war. When Nezahualpilli died in 1515 — 1516 in some accounts — several of his sons claimed the throne. Moctezuma II backed Cacamatzin, and the dispute ended in a partition of the kingdom into three parts, one of them — the north, with its own court at Otumba — going to Ixtlilxochitl, who never forgave the Mexica king. Cacamatzin was deposed by Cortés in 1519 and killed the following year, in circumstances the sources describe differently. His brother Coanacoch succeeded him and governed until the Spaniards arrived on 31 December 1520, when he left for Tenochtitlan and the city was sacked. Cortés installed a third brother, Tecocoltzin, who died in 1521; Ixtlilxochitl II took the throne that year.

His alliance with Cortés was the single most consequential decision any Aztec noble made. Texcoco became the Spanish base for the siege of Tenochtitlan: the brigantines that gave Cortés command of the lake were built there in the winter of 1520–21, from timber brought over the mountains from Tlaxcala and assembled by Texcocan carpenters, and thirteen of them were launched in April 1521. Texcocan warriors marched with the besieging army, and Texcocan labour built the causeway camps. When the city fell in August, Ixtlilxochitl was baptised Fernando Cortés Ixtlilxochitl, with the conqueror as his godfather, and ruled Texcoco under Spanish authority until his death in 1531.

The dynasty survived him, which is unusual. Three of his brothers followed in turn, then a nephew, Hernando Pimentel Nezahualcoyotl, who ruled as tlatoani and governor for two decades — Matthew Restall's point being that the years of civil war and conquest were a short interruption in a dynastic run that lasted a century and a half. Texcoco was designated one of four urban centres of New Spain and later raised to the rank of ciudad, lost the rank, became the seat of an alcaldía mayor, and recovered the title in 1543 and again in 1551. Its Franciscan convent went up on the edge of the old ceremonial centre, and the town kept a native cabildo and a school of Nahuatl scribes who wrote some of the best sixteenth-century documents in the language.

Decline came slowly, from the 1580s, as the Spanish economy reoriented towards the mines and Mexico City, and as epidemic disease hollowed out the wards. The lake did the rest. The dikes were destroyed in the siege and never properly rebuilt; the city flooded repeatedly, and from 1607 the Spanish crown began cutting the drainage channel that would eventually empty the basin. By the nineteenth century Texcoco was a market town in a dry plain, and the lake that had made it was a salt flat with a railway across it. The modern municipality is part of Greater Mexico City's sprawl, and the site of the palace is under the streets around the plaza. The name survives in the modern municipality, in a Mexico City borough named after its most famous king, and in the lakebed that still floods in the rainy season.

Common myths about Texcoco

Timeline of Texcoco

What the map shows next

At 1325 on the map the map draws Texcoco as a single small territory on the eastern side of the basin, with Tenochtitlan appearing on its island the same year and the Tepanec capital still unmarked. The shape is a rough polygon of some 2,000 square kilometres, and it does not sit where the city was: it covers the freshwater lakes of Chalco and Xochimilco and the plain running south towards Amecameca, and it leaves out Texcoco itself, the hill of Texcotzingo and the Acolhua town of Coatlinchan. Treat it as a marker for "an Acolhua polity in the eastern basin", not as a boundary.

At 1418 on the map the same territory is drawn as an independent state, which it was not: Texcoco had been paying tribute to Azcapotzalco since the 1370s, and in September 1418 it ceased to exist as a political unit altogether. At 1428 on the map the map reaches the end of the entity's drawn range, on the year the Tepanec empire fell. Everything that makes Texcoco worth an entry — the alliance, the law code, the aqueducts, the gardens, the archive and its burning — happens after the map stops, so the reader who wants the rest of the story has to go to the Triple Alliance and Tenochtitlan entries for the years the border data does not cover.