Tenochtitlan

The distinction matters because almost everything memorable about Tenochtitlan — the vast population, the great market, the Templo Mayor's final bulk, the causeways crowded with canoes — belongs to the period after 1428. What existed in 1377, the year this map shows, was a settlement of perhaps a few thousand people on a mud island, with a modest temple of earth and wood, plots of maize staked out into the marsh, and a dynasty one generation old. It was not yet a capital of anything.

The Mexica, as they called themselves, are usually known today as the Aztecs, a name coined in the nineteenth century from Aztlan, their legendary homeland. They spoke Nahuatl, as did most of the peoples who had moved into central Mexico in the centuries after the collapse of Teotihuacan and then of Tollan, and they told their history in painted annals that were read aloud, year by year, in the presence of the ruler. Those annals are why we know anything at all about the town's early years, and they are also why the early dates cannot be taken at face value.

At a glance

Where did the Mexica come from?

The Mexica told a migration story, and they told it in pictures. The Codex Boturini, also called the Tira de la Peregrinación, opens with a figure standing on an island among reeds, beside a temple and a divided tree: Aztlan, the place of herons, somewhere in the north. The god Huitzilopochtli speaks to the people, and they set out. The strip then runs year by year through a long sequence of halts, each marked by a place glyph and a small scene of arrival, quarrel, or sacrifice, until the wanderers reach the Basin of Mexico.

The chronology is native, and its conversion to Christian years depends on which correlation of the 52-year calendar a modern scholar adopts. Chimalpahin, writing in the early seventeenth century from older annals, dates the departure to the year 1 Flint, which he equates with 1064. The migration then took generations, not months: the Mexica stopped at Tollan, at Chapultepec on the western shore of the lake, and at Tizaapan on land belonging to Culhuacan, and at each stop they were tolerated, employed, and eventually expelled.

Michael Smith's comparison of the migration accounts, published in 1984, found something more orderly than the mess of variant dates suggests. The chronicles agree on the relative order in which three Nahuatl-speaking contingents arrived, and the dates attached to each cluster: 1195 for the first group into the basin itself, 1220 for the group that settled the surrounding valleys, and 1248 for the Mexica, with a standard deviation of only 5.2 years for the last. Smith's conclusion was that the migration narratives preserve real population movements, and that their accuracy declines the further back they reach. That is a useful rule for everything that follows: the closer to 1500 an event is, the more the sources can be trusted.

What the story explains is a political predicament. The Mexica had no territory of their own, no recognised royal line, and no claim on the settled order of the basin. They were useful as fighters — the annals have them hired as mercenaries by the more established towns — and dangerous as neighbours. In about 1323, according to the annals, they asked the ruler of Culhuacan for his daughter, ostensibly to marry her into their chiefly line, and then sacrificed her and dressed a priest in her skin. The story is told with relish and it is not a flattering one; what it does is justify the expulsion that followed, and the flight to an island nobody else wanted.

Was the city founded in 1325?

1325 is the date on the plaque, the date in the schoolbooks, and the date most scholars accept. It is not, however, the date the federal government commemorated in 2021 as the seven-hundredth anniversary, which was 13 May 1321. Historians pointed out at the time that 1321 appears in no source and that it had apparently been chosen to make the anniversary run into the quincentenary of the city's fall in 1521; as one of them put it, the date was invented to match the defeat.

The problem is not that 1325 is wrong. It is that it is one of several. A census of 39 pictorial and alphabetic sources made by Elizabeth Hill Boone in 1992 — the count is hers, and López Luján and Matos summarise it — found 15 that give an exact year for the portents that marked the site. They fall across 48 years of the fourteenth century. The year 2 House, which converts to 1325, appears nine times, in the Codex Mendoza, the Codex Mexicanus, the Anales de Tlatelolco, the Anales de Cuauhtitlan, the Leyenda de los Soles, two of Chimalpahin's relations and the Crónica Mexicayotl. The year 2 Flint, or 1364, appears twice, in the Codex Aubin. Single sources give 8 Rabbit (1318), 12 Rabbit (1322), 1 Flint (1324) and about 4 Rabbit (1366). A recently published codex, the Tira de Tetepilco, adds 4 House, or 1301.

There is even a stone. The Teocalli de la Guerra Sagrada, carved around 1507 under Moctezuma II, shows on its roof a cartouche containing the sign 2 House, and on its rear face the eagle on the cactus growing out of a stone, with the war glyph of water and burnt earth beneath its beak. The Mexica themselves, at the height of their power, chose 1325 as the year to carve in stone. That is about as close to an official answer as the evidence gets, and it was made nearly two centuries after the event.

Archaeology complicates the picture further, and in a useful way. Deep shafts sunk under Mexico City's cathedral in the 1990s, to correct the building's tilt, cut through the whole sequence of the city's occupation. The layers from 12.3 to 13 metres down contain Aztec II pottery, the ware of the Middle Postclassic; obsidian hydration gave dates of 1198 to 1294 for it, but the newer radiocarbon chronology for the basin puts the production of that pottery at 1331 to 1447, with a median of 1358. Below it, from 13 to 13.7 metres, came a surprising concentration of Tollan-phase and Aztec I sherds, including Mazapa ware imported from Tula itself, dated to the centuries between AD 900 and 1100. Beneath that lay culturally sterile soil.

So there was a small Toltec-era hamlet on or beside the island long before the Mexica arrived. Leonardo López Luján and Eduardo Matos Moctezuma, who published this sequence in 2025, draw the sensible conclusion: the city did not appear in a single year, the archaeological layers agree with the sources that place the Mexica arrival in the 1300s, and the search for a precise day is futile. The date on the plaque is a convention, adopted in 1925 for the six-hundredth anniversary, and it is a good convention — but it is not a record.

Why build a city on a swamp?

The Basin of Mexico was an inland sea, or rather five shallow lakes in a chain, covering some 1,500 square kilometres before the conquest. The northern three — Zumpango, Xaltocan and Texcoco — were saline; the southern two, Xochimilco and Chalco, were fed by springs and stayed fresh. Texcoco was the largest and the lowest, so the whole system drained towards it, and it was brackish, windswept, and ringed with salt flats and halophytic scrub. In drought years the lakes separated; in flood years they joined.

Tenochtitlan occupied a small island on Texcoco's western side, near the shore and near the springs that seeped from the foot of Chapultepec. The choice looks perverse and was in fact shrewd. An island was defensible in a basin where warfare was endemic and where the strongest power, Azcapotzalco, was a few kilometres away on the mainland. It was also worthless: nobody else wanted it, which is precisely why the Mexica were allowed to have it, in exchange for tribute. And it sat at the meeting point of the salt-water lake and the fresh-water margins, which gave access to both.

The marsh could be made into farmland, and this is the achievement on which everything else rested. A chinampa was built by driving stakes into the lake bed and weaving reeds between them to make an underwater fence, then piling in mud, aquatic plants and refuse until the enclosed plot stood above the water line. Willows, especially the ahuejote, were planted along the edges; their roots bound the plot and their shade limited evaporation. Ditches between the plots were dredged for mud and used as canals, so that the same water carried the canoe and fertilised the field. The Spanish called these jardines flotantes, floating gardens, which they were not: the plots were staked to the bottom, and the water around them was shallow enough to stand in.

This is how the city grew. Every generation added plots, and the plots became ground; the island was not so much built on as accumulated. By 1519 the urban area covered between 8 and 13.5 square kilometres, several times its original size, and the canals that crossed it were both streets and irrigation channels. The lake also fed the town directly: fish, waterfowl and their eggs, frogs, the larvae and algae that were harvested and sold as food, salt from the eastern shore, reeds for mats and roofs. The Mexica were not farmers who happened to live on water. They were specialists in a wetland economy, and the tribute they paid their overlords was drawn from it.

What was an altepetl?

The unit of political life in central Mexico was not the empire and not the tribe. It was the altepetl, a word meaning "water, mountain", which is to say a place with a settled population, a ruler, a patron deity, a temple, a market and a landed estate. An altepetl could be a great city or a town of a few thousand; what made it one was that it had its own tlatoani, its own god, and its own land, and that it governed itself while owing tribute or service to whoever was stronger.

The ruler was elected, not simply inherited. The choice lay with the leading men of the altepetl, and it normally fell on a son or brother of the previous ruler, but the principle was election by the great men and the assent of the people, and the sources say so repeatedly. Diego Durán, who had access to Nahuatl informants in the sixteenth century, put it plainly: dominion never passed by right of inheritance, but by election, and a candidate had only to be of the right lineage and acceptable to the council. This is why the Mexica could take a ruler from Culhuacan in the 1370s and why, fifty years later, they could pass over more legitimate candidates for a bastard son with a good war record.

Below the ruler, the altepetl was divided into wards. The sources call them calpulli, "big houses", and the word carried a double sense: a kin group with a migration history, and a territorial district with its own temple, school, market and allotment of land. In the colonial Nahuatl documents the everyday word for the ward is more often tlaxilacalli, while calpulli carries the flavour of origin and status; the two were used interchangeably often enough that the distinction is one of tone rather than of institution. Each ward held land corporately and distributed it to its households; each supplied labour and tribute on a rotating basis; each had its own leaders, who were the conservatives of local tradition and, as will be seen, the people with the most to lose when a ruler decided to rewrite the past.

Tenochtitlan had four such quarters — Moyotlan, Teopan, Atzacualco and Cuepopan — arranged around the sacred precinct, each subdivided into wards, each with its own small temple and its own stretch of chinampa. The northern half of the island was a separate altepetl entirely, Tlatelolco, founded by a dissident Mexica group and ruled by its own dynasty; it was absorbed only in 1473. This is worth holding on to, because the great market that every account of Tenochtitlan describes was in Tlatelolco, not in Tenochtitlan.

Who ruled, and who chose?

The Mexica elected their first tlatoani in the 1370s. He was Acamapichtli, "handful of reeds", the son of a Mexica leader named Opochtli Iztahuatzin and of Atotoztli, a daughter of the Culhua king Coxcoxtli. The match was the whole point: it gave a landless immigrant group a genealogy that reached back through Culhuacan to the Toltecs, and it gave the Culhua a client on the island. The Mexica had been asking for a ruler from a legitimate line; what they got was a half-Culhua prince with a Mexica father, which was exactly the compromise the situation allowed.

The dates of his reign are a warning about the sources. Chimalpahin gives 1367 to 1387; the Codex Aubin gives 1376 to 1395; one modern list gives 1350 to 1403. The disagreements are not small, and they propagate: every subsequent reign shifts with them. What is not in doubt is the sequence — Acamapichtli, Huitzilihuitl, Chimalpopoca, Itzcoatl — and the fact that each was elected.

Huitzilihuitl's reign is where the Mexica's position began to improve, and the improvement came through marriage rather than war. He took as wife Ayauhcihuatl, a daughter of Tezozomoc, the Tepanec king, and the annals say the tribute was reduced as a result. He also campaigned, in the basin and beyond it, and the Mexica began to acquire subject towns of their own; the line between serving the Tepanecs and profiting from their wars was thin, and the Mexica learned to work it.

The office these men held was less absolute than the word king suggests. A tlatoani judged disputes, allocated land through the wards, led the army in person, represented the altepetl to its overlord, and spoke for it in the council of the basin's rulers; but he ruled with a council of the leading men, and the ward heads retained control of land and labour within their own districts. When the Mexica elected a ruler they were not choosing a despot but appointing a manager of a corporate estate, one whose authority rested on consent as much as descent — which is why a disputed election, as in 1427, was a crisis for the whole polity and not merely for a family.

Chimalpopoca, his son, was Tezozomoc's grandson, and his reign of about ten years was entirely framed by that relationship. When Tezozomoc went to war with Ixtlilxochitl of Texcoco in 1418, the Mexica fought on the Tepanec side and were rewarded with a share of the conquered city's tribute. Chimalpopoca also interceded for the displaced Texcocan prince Nezahualcóyotl, who was allowed to live in Tenochtitlan under Tepanec protection — an act of clemency that would turn out to be the most consequential thing he did, since Nezahualcóyotl would later be the ally who helped destroy Azcapotzalco.

What did tribute to Azcapotzalco actually mean?

Azcapotzalco was the great power of the basin in the late fourteenth century. Under Tezozomoc, who reigned from about 1371 to 1426, it became something like an empire: the sources credit it with authority over some 40 altepetl, and Tezozomoc installed his sons on thrones around the lake, including at Tlatelolco, where his son Cuacuapitzahuac ruled until 1407. It collected tribute in goods, in labour and in soldiers, and it adjudicated between its clients.

For the Mexica, tribute was not an abstraction. Durán's account of it is one of the most vivid passages in the whole corpus, and it repays attention because it describes a relationship rather than a tax table. When the Mexica elected a king of their own, Tezozomoc called his lords together and pointed out the danger: the Mexica had already occupied Tepanec land, and now they had made a king and a court of their own, and their children might one day claim to rule over their neighbours. Better, he said, to double the tribute — twice what the Mexica were used to giving, of the things and vegetables they gave in token of recognition and subjection.

The messengers went to Tenochtitlan. Besides the fish, frogs and vegetables they already owed, the Mexica were now to bring full-grown savins and willows to plant in Azcapotzalco, and to build a raft on the water planted with every crop of the land — maize, chilli, beans, squash, amaranth. The Mexica wept. That night Huitzilopochtli spoke to one of their elders and told them to accept, and to do it. They carried the trees, planted them where the king ordered, and floated the raft into Azcapotzalco sown with maize in ear, chilli, tomatoes, amaranth, beans, squash and flowers. Tezozomoc, astonished, then raised the stakes again: the raft must also carry a duck sitting on eggs and a heron sitting on eggs, and the eggs must hatch on the day it arrived.

Whether the story is literally true is beside the point; what it encodes is exact. The tribute was aquatic produce, horticultural skill and conspicuous submission. The Mexica were being asked to demonstrate, in public, that they could make the lake yield whatever the overlord demanded — which was both a humiliation and, in the long run, an advertisement. Durán adds that they kept to this arrangement for fifty years, which fits the span between Acamapichtli's election and the war.

Alongside the goods went men. Mexica contingents fought in Tepanec campaigns, including the war against Texcoco in 1418, and military service of that kind was the most valuable thing a client could offer and the most dangerous thing to be good at. It also brought rewards: lands, a share of tribute, and the experience of campaigning far from the island. The marriage alliance with Tezozomoc's daughter did the rest. By the 1420s Tenochtitlan was a substantial town with a royal line, a claim to Culhua ancestry, a temple of its own, and a growing reputation for soldiers — and it was still, in law and in fact, a subject of Azcapotzalco.

What did the sacred precinct look like before 1428?

The centre of the town was a walled enclosure at the meeting point of the four quarters, raised on a platform and reached by causeways and canals. Inside it stood the Templo Mayor, the double pyramid that the Mexica called the Hueyi Teocalli, and around it the temples, schools, ball court and skull rack that made the precinct a model of the cosmos rather than a collection of buildings.

The Templo Mayor was a mountain. Its two shrines, side by side at the top of a single platform, were dedicated to Tlaloc, the god of rain and agricultural fertility, and to Huitzilopochtli, the Mexica patron and god of war and sun. The pairing was a statement about what the city lived on: water and maize on one side, conquest and blood on the other. The whole structure represented Coatepec, the serpent mountain where Huitzilopochtli was born and where he killed and dismembered his sister Coyolxauhqui — which is why, at the foot of the stairs, the Mexica eventually laid a great disc carved with her broken body.

Seven successive temples were built on the site, each enclosing the last. The first, according to the sources, was of earth and perishable wood; it lies below the water table and has never been excavated, and may not survive. The second was built during the reigns of Acamapichtli, Huitzilihuitl and Chimalpopoca — that is, between about 1375 and 1427 — and its upper part has been found, with two stucco-covered stone shrines on the north side, a chacmool, and a sacrificial stone carved with a face. This is the Templo Mayor of the tributary period: not the great pyramid of the accounts, but a solid, decorated building of the standard central Mexican type, enlarged as the town grew.

The third temple was built between 1427 and 1440, under Itzcoatl, and it is the first whose decoration proclaims the new order: a staircase flanked by eight stone standard-bearers, one of them bearing the date 4 Reed, or 1431. From that rebuilding onwards the precinct expanded with each reign until it covered an area of a few hundred metres on a side and contained, by one conquistador's count, 78 structures.

The precinct was also a school and an archive. The calmecac, the house of the lineage, trained the sons of nobles for the priesthood and the higher offices; the telpochcalli trained commoners for war. The temple kept the offerings that commemorated each stage of construction — shells, fish, jaguar bones, the remains of sacrificed animals — and the priests kept the books that recorded the days. Before 1428 that apparatus was modest in scale and local in ambition. It became something else when the city became an empire, and much of what visitors now see at the site belongs to the later stages.

How did the island connect to the world?

By water, mostly. Canoes were the vehicles of ordinary life: dugouts of cedar and mahogany, some large enough to carry a family and its produce, punted or paddled through the canals that served as streets. Bernal Díaz, seeing the city in 1519, remembered countless canoes plying the lake, and the traffic was already old by then. Everything heavy — maize, timber, stone, pots, people — moved by canoe unless it was going to the mainland.

Three causeways tied the island to the shore. One ran north to Tepeyac, one west to Tlacopan, one south to Ixtapalapa; a fourth, shorter, linked Tenochtitlan to Tlatelolco. They were built of earth and stone, wide enough, the chroniclers said, for ten horsemen abreast, and they were broken at intervals by wooden bridges that could be lifted at night to let canoes pass and, in an emergency, to isolate the city. The western causeway is credited to Chimalpopoca, which would make it a work of the tributary period, built with Tepanec acquiescence.

Fresh water was the harder problem. The lake was brackish, and the town's drinking water came from springs at Chapultepec on the mainland, carried along a raised channel. The first aqueduct was a makeshift affair of stakes, reeds and mud; under Chimalpopoca, in about 1426, a better one was built with Tezozomoc's assistance, and it is described as a wooden channel running from the hill to the island. Its construction is also the point at which the relationship with Azcapotzalco broke. In Durán's account the Tepanec lords were furious that the Mexica should presume to bring water across their land, when they had after all allowed them to live there and given them the water they asked for; that the Mexica now wanted a channel cut through Tepanec territory was, to them, a claim of mastery. The water question and the sovereignty question were the same question.

Every quarter had its market, but the great market of the island was at Tlatelolco, the sister city to the north, and it was there that the Mexica state's commercial life concentrated. It was a daily market, held in a walled square surrounded by arcades, with officials who inspected goods and punished fraud, and it dealt in everything from maize, beans, salt and chilli to cotton cloth, cacao, obsidian, feathers, jade and slaves. The pochteca, the long-distance merchants, were organised in their own corporations and carried goods from the Gulf coast and the Pacific; they appear in the sources as a distinct and privileged group, and their trade was one of the things the Mexica would later use imperial power to protect. But the market's scale in the tributary period is impossible to measure, and the figures usually quoted — Cortés thought 60,000 people traded there daily — describe the city after a century of empire.

What did people eat?

The diet was maize, and the maize came from the lake. A chinampa plot could be cropped continuously, and the standard combination — maize with beans climbing its stalks and squash spreading at its feet — spread the risk of failure and improved the soil. Amaranth, chilli, tomatoes and a dozen kinds of greens filled the gaps. The plots were worked from canoes, harvested into canoes, and fertilised with mud dredged from the canals and with human waste collected and sold for the purpose.

To this the lake added protein. Fish were taken in nets and weirs; waterfowl were netted, snared and hunted with throwing sticks, and their eggs collected; frogs, tadpoles and the axolotl were eaten; insects, larvae and the blue-green algae that the Mexica called tecuitlatl were harvested from the surface and sold in cakes. The Spanish found the algae unappetising and the Mexica found it ordinary. Salt came from the eastern shore of Texcoco, where it was evaporated from the brine and sold in cakes and blocks; it was one of the basin's most widely traded commodities.

There were no large domestic animals. Turkeys and Muscovy ducks were kept, and so were dogs, bred for meat as well as for company; deer, rabbits and wild birds were hunted on the mainland. Tools were of obsidian and chert, and obsidian from the Pachuca and Otumba sources moved through the basin in quantity — the cathedral excavations found it in every layer. Some of it travelled a long way: turquoise, marine shell and feathers came up from the south and the coasts, and the basin had been tied into a northern exchange network since the twelfth century, in what archaeologists sometimes call the Pueblo–Aztec regional system.

Maize was eaten as tortillas, griddled on a clay comal, and as atole, a gruel, and it was prepared by soaking the grain in water with lime or wood ash, which loosens the hull and makes the amino acids in the grain available. The nixtamal process is one of the technical foundations of Mesoamerican life, and it is invisible in the chronicles, which record the doings of lords and not the work of kitchens. What the chronicles do record is the storage: maize kept in granaries and sealed jars, tribute grain counted in loads, and the constant anxiety about a failed harvest that runs through every account of the basin's politics.

The vulnerability of this system was water. Too much and the lake rose into the streets; too little and the chinampas dried and the harvest failed. Both happened, and both were remembered as omens as much as disasters. The great flood and the famine that followed in the 1440s and 1450s came after the period covered here, but they belong to the same story: a city built on a lake lived by managing the lake, and the greatest engineering works of the fifteenth century — the dikes that separated fresh water from salt — were attempts to solve a problem the Mexica had created by settling there in the first place.

What changed in 1428?

Everything, and quickly. Tezozomoc died in 1426, and the succession broke down. The sources disagree about who succeeded him and even about which city Maxtla, the claimant who emerges as the villain of the story, actually ruled — Durán calls him lord of Cuyuacan, not of Azcapotzalco. What is consistent is the outcome: the Mexica ruler Chimalpopoca was murdered, either by Tepanec agents who entered the palace at night or, in other accounts, taken prisoner and killed; and the Mexica, who had been his subjects, chose a new king and a new policy.

The new king was Itzcoatl, a son of Acamapichtli by a woman of Azcapotzalco, and therefore a man with a doubtful claim to legitimacy and an excellent claim to military competence. He found allies where the Tepanecs had made enemies: Nezahualcóyotl, the Texcocan prince who had lived in Tenochtitlan as a protected exile, and the lords of Tlacopan and other towns that had suffered under Azcapotzalco. The war that followed ended with the storming of Azcapotzalco in 1428 and the division of its dominion among the victors — the arrangement that historians call the Triple Alliance.

The most consequential act of the new regime was an act of historiography. María Castañeda de la Paz, tracing the episode through the sixteenth-century sources, notes that Sahagún records an order by Itzcoatl to burn the old books, on the grounds that they contained falsehoods and that the people should not be misled by them. The passage is short and unambiguous, and modern scholars have read it as the moment when the Mexica state created an official history: a single account of the migration, the founding and the dynasty, tailored to the new ruling house and to its claim on the Toltec inheritance. It did not entirely succeed. A colonial document concerning two nobles of Axapusco records that they were enemies of Moctezuma II precisely because they had never burned their old paintings as Itzcoatl had commanded.

Alongside the new history came a new apparatus. Tlacaelel, a half-brother of Chimalpopoca, took the office of cihuacoatl, a position second only to the ruler, and held it for half a century; he is credited in the sources with elevating Huitzilopochtli above the other gods, with tightening the sumptuary laws that separated nobles from commoners, and with shaping the ideology of sacrifice that justified conquest. The tribute system was reorganised around the triple alliance, with two fifths of receipts going to Tenochtitlan, two fifths to Texcoco and one fifth to Tlacopan. Expansion ran south and east into Morelos and the Huasteca; westwards it met a limit, the Tarascan state of Michoacán, whose armies defeated the alliance in 1479 and were never subdued. The Templo Mayor was rebuilt from 1427, and each subsequent ruler rebuilt it again, larger.

And the city began to fill. The Mexica now drew labour and materials from a dominion that stretched beyond the basin, and they used them on the island: causeways, canals, palaces, temples, and the elaborate waterworks that made the lake habitable. Within two generations Tenochtitlan was the largest city in Mesoamerica, and the town of the tributary period had become a foundation myth — remembered, carved, and in part invented by the state that replaced it.

How do we know any of this?

The evidence is of three kinds, and none of them is straightforward. The first is pictorial: painted annals in which a year sign and a scene carry the whole record, read aloud with an oral commentary that was never written down. None of the pre-conquest books that survived the burning is a history of the Mexica; what we have are colonial copies and new paintings made under Spanish eyes — the Codex Mendoza, compiled around 1541 for the viceroy, the Codex Aubin, the Codex Azcatitlan, the Tira de la Peregrinación — and they differ from one another.

The second is alphabetic: Nahuatl and Spanish histories written in the sixteenth and seventeenth centuries by men who had access to the annals and to living memory. Hernando Alvarado Tezozómoc, a grandson of Moctezuma, wrote the Crónica Mexicáyotl in Nahuatl. Domingo de San Antón Muñón Chimalpahin, of Chalco, wrote a series of relations that are the backbone of any modern chronology. The friars — Durán, Sahagún, Tovar, Torquemada — collected and arranged the same material with their own theological purposes. Every one of these writers had a patron, a community and an argument.

The third is archaeological, and it is the only evidence not mediated by a sixteenth-century interpreter. It is also uneven: the water table under Mexico City has preserved organic material beautifully and made deep excavation almost impossible, so the earliest levels of the Templo Mayor cannot be reached, and much of what we know about the town's beginnings comes from salvage work under colonial buildings. The cathedral shafts that produced the Aztec II sequence were dug to stop a building from sinking, not to answer a historical question.

Two cautions follow. The first is chronological: the Mexica counted years in a 52-year cycle, and converting a year sign into a Christian year requires choosing among several correlations, which is why the same event can appear as 1325 in one source and 1364 in another. The second is political: after 1428 the state had an interest in a particular version of its own past, and the version it promoted is the one most of our sources transmit. That does not make the story false. It does mean that the further back the story goes, the more it resembles a charter, and the less it resembles a record.

Common myths about Tenochtitlan

Timeline of Tenochtitlan

What the map shows next

At 1390 on the map the island is still drawn as a small territory on the western shore of the lake, with Texcoco to the east and the Tepanec lands around it; the shape does not change much across the century, because the transformation of Tenochtitlan was not territorial. What changed was the amount of labour and tribute the town could command.

At 1428 on the map the tributary period ends and the map hands over to the Triple Alliance, which is drawn from 1429. The city itself would go on to dominate the basin, then Mesoamerica; the empire it built fell to Castile in 1521, and the island's ruins lie under the modern Mexican capital, whose flag still carries the eagle on the cactus that the Mexica carved on the Teocalli of the Sacred War — the device the First Mexican Empire adopted in 1821.