The megalithic structure at Puma Punku is one of those sites that Bad Archaeologists really love. They claim that it can’t be explained by what we know about the technical accomplishments of early Andean civilisations, that it is unimaginably old, that it was built as a port… the list of claims is almost endless. It gets plenty of comments on this site, but one of my favourites is that it is certainly a headache for any mainstream archeologist, or th eone that calls me a disinformation agent. It is not a headache at all: it is actually a well studied and reasonably well understood structure. As for disinformation agent, well, if I am one, then I’ve never had the massive payments I must be due by now.

The site

Jumbled masonry at Puma Punku

Chaotically jumbled masonry at Puma Punku (Source)

Puma Punku (“Puma Gate”) is a megalithic platform that lies around a kilometre to the south-west of the main complex at Tiwanaku; the zone between has been shown by geophysics to contain numerous structures. It was thus clearly part of the complex of monuments that makes up the city and consists of a terraced mound resembling other pyramids in the complex, but with only three steps. It is composed of soil with a facing of dressed megalithic stones composed of andesite and red sandstone. The largest block measures 7.81×5.17×1.07 m and weighs around 131 tonnes. The blocks are held together partly through gravity and partly through I-shaped cramps. The cramps are made from a copper-arsenic-nickel bronze alloy known from artefacts dated 600-900 CE in the area between Tiwanaku and San Pedro de Atacama in northern Chile.

Puma Punku: the claims

Bad Archaeologists love to quote Arthur Posnansky (1873-1946), an amateur archaeologist who dated Tiwanaku to c 15,000 BCE. His dating was based on what he considered to be astronomical alignments between the buildings in the city. This was based largely on an assessment that the structure known as the Kalasasaya, for which the alignments did not match those of today. Rather than abandoning his hypothesis that the building incorporated a solar alignment, he sought out a date at which the supposed alignment would make sense. Currently, the obliquity of the cliptic is around 23.4°, while Posnansky’s calculations require it to be 23.15°. Checking the tables available to him, he discovered that the last time the axial tilt corresponded to this angle was around 15,000 BCE.

According to Posnansky, Puma Punku consists of “a true and magnificent pier or wharf… where hundreds of ships could at the same time take on and unload their heavy burdens”. He attributed its ruined condition to cataclysmic flooding from Lake Titicaca, a “catastrophe… caused by seismic movements which resulted in an overflow of the waters of Lake Titicaca and in volcanic eruptions”.

It has become a favourite among those for whom Graham Hancock is an authority on ancient civilisations. Given his acolytes’s fondness for the place, it is surprising just how little space he devotes to a discussion of it, which amounts to under three pages (including an almost half-page illustration).

Hancock dismisses orthodox archaeological opinion, stating that “not a single orthodox historian or archaeologist was prepared to accept such an early origin for Tiahuanaco preferring… to agree on the safe estimate of AD 500”. Hancock is not going to tell his readers that this is not a “safe estimate” at all but a rigorously determined date based on radiocarbon determinations of the stratigraphy of the site. Posnansky’s calculations, on the other hand, depend on establishing alignments between elements of a structure that was ruinous even before he started to survey it. The “high-powered team” of scientists that Hancock claims Posnansky co-opted to check his calculations (Hans Ludendorff (1873-1941), director of the Astronomical Observatory of Potsdam), Friedrich Becker (1900-1981) of the Vatican Observatory, Arnold Kohlschutter (1883-1969) of the University of Bonn and Rolf Müller (1898-1981) of the Ludwig-Maximilians-Universität in München) may indeed have verified the alignments. However, they were all astronomers, who only checked the calculations. They were not archaeologists and were not in a position to evaluate whether or not the supposed alignments were intended by the builders or were just something in Posnansk’s imagination.

Puma Punku nonsense debunked

The level of the lake was not high enough at any time when Tiwanaku was occupied to have reached Puma Punku. The port of Tiwanaku lay 18 km to the west-north-west, at Iwawi (or Iwawe), on the current shoreline of Lake Titicaca, with which it was connected by road. The port was discovered in the 1960s, so there can be no excuse for Bad Archaeologists continuing to claim that Puma Punku, at a higher elevation, was the city’s port. Iwawi was the port by which stones were brought to Tiwanaku for the construction of its monuments.

A reconstruction of Puma Punku

A reconstruction of Puma Punku, based on excavations by the University of Wisconsin in 2002 (Source)

In spite of the speculations of Bad Archaeologists about the site’s construction in remote antiquity, Puma Punku is well dated by radiocarbon. A high precision date has been obtained from the soil fill of the lowest terrace of the structure: 1500 ± 25 bp, which calibrates to 569 ± 21 CE (527-611 CE at 2σ, meaning that there is a 98% chance that the sample falls in that date range). This gives us a terminus post quem for the construction of the lowest and earliest part of the structure. The terminus post quem is an important archaeological principle by which a deposit cannot be older than the date of the youngest object in it: in other words, there is a 98% probability that Puma Punku dates from after 527 CE. Other evidence—such as the radiocarbon chronology of Iwawi—shows that the level of Lake Titicaca at this time was much the same as it is today. Its massive building blocks, weighing up to 131 tonnes, came from a quarry close to the current shore of Lake Titicaca, 10 km away, yet more evidence that the water level was close to that of today. So Puma Punku is not a port structure.
The ruined state of the monument can be attributed to two causes, neither of them related to Posnansk’s unattested floods. Firstly, the site is unfinished: we do not know why the site was abandoned, but it was left in an apparently unfinished state. Secondly, Tiwanaku in general has suffered a great deal from later quarrying, particularly after the Spanish conquest of the Inka Empire, when we know that the site was raided for building stone. Other elements were broken up for hardcore during the construction of the local railway in the nineteenth century.
E Fred Legner has suggested that the Tiwanaku culture had a three-tiered cosmology, with the Templete representing the underworld of fertility gods and the ancestors (Ucu Pacha), while the Akapana and Puma Punku represent the Hanan Pacha (upper world) where the celestial gods reside; humans occupy the central world (Cay Pacha). This is the sort of hypothesis that is impossible to refute: it’s an interesting idea, but nothing more.


Puma Punku is part of a fascinating pre-Columbian city in the Andes. Its age is tolerably well established and it fits a pattern of construction seen throughout the region controlled by Tiwanaku. Claims that it is a harbour construction from remote antiquity do not stand up to scrutiny. Because the place is remote and the product of a civilisation that is not regularly covered in school curricula outside Bolívia, it can appear mysterious. Like so many such places, though, putting it into a cultural context helps to remove much of the mystery: mystery-mongers such as Graham Hancock never do this and pretend to their readers that archaeologists’ accounts of it are nothing more than guesswork. They are wrong and, perhaps, deliberately mislead their readers.