Nine rescued chimpanzees near Madrid picked quartz and calcite out of pebble piles within seconds, then held them up to the light for minutes at a time. That much is measured. The claim that it explains 780,000 years of human mineral-collecting rests on five sites and a date the paper does not source.
Over eleven days in autumn 2021, a team led by the crystallographer Juan Manuel García-Ruiz ran object-choice trials on nine adult chimpanzees at the Rainfer sanctuary near Madrid — two social groups of five and four, mean age 32.1 The results appeared in Frontiers in Psychology in March 2026.1
In the first trial, a 3.3 kg transparent quartz crystal and a 2.1 kg sandstone rock of almost identical height were fixed to matching pedestals in one group's yard.1 All five chimpanzees in that group engaged with the crystal; none gave the rock more than two minutes. The rock was pulled down, abandoned for 38 minutes, dragged about ten metres and left. The crystal went into the dormitories and stayed two days, and staff recovered it only by trading bananas and yogurt. Interaction time favoured the crystal at p = 0.0052.1
In a second trial, crystals were seeded into piles of pebbles. Across four piles holding twelve crystals between them, the chimpanzees took eleven, mostly leaving the pebbles.1 One animal, Sandy, carried pieces to a platform in her mouth and sorted quartz, calcite and pyrite into a group of their own — three minerals that differ in colour, lustre and symmetry, and share only the fact of being crystals.1
The study's framing rests on one number: that hominins have collected crystals as far back as 780,000 years ago. It leads the abstract, the institutional press release and the coverage that followed.18 It carries no citation.
The paper's own introduction never uses it. It gives Zhoukoudian in China — about twenty quartz pieces reported by Pei Wenzhong in 1931 — as at least 600,000 and possibly more than 800,000 years old, crediting a 2001 uranium-series study.17 The roughly 770,000-year figure for that site's lower layers comes from a different study altogether: a 2009 cosmogenic burial dating in Nature, which reports 0.77 ± 0.08 million years and is absent from the reference list.2
The site that is actually dated to 780,000 years is Gesher Benot Ya'aqov in Israel, fixed to the last magnetic reversal by Naama Goren-Inbar and colleagues in 2000, and reported to contain small unworn quartz crystals.36 The study does not mention it.
Desk analysis. The headline number is real archaeology. It is simply not this paper's archaeology — and at Zhoukoudian, the dating uncertainty alone (±80,000 years) spans about a quarter of the whole record being described.
The numbers behind the paper's own charts. Mean interaction times (Figure 2) and mean crystal counts (Figure 4) are plotted with standard deviations but never printed, and no dataset is deposited.1 This desk could not reproduce either figure.
Whether these animals stand for any others. All nine are enculturated; most came from circuses. The paper states they are already familiar with human-made objects that are transparent, flat-faced and straight-edged — precisely the properties it identifies as the attractors.1 No wild ape has been tested; the authors call that the priority.1
Replication. The crystal-versus-rock trial was quantified once, in one group. The second group's run could not be measured at all, because a chimpanzee carried both objects indoors before recording began.1 The authors state their results cannot be fully interpreted statistically.1
Whether an aesthetic drive predates tool use. Nothing here tests it. A preference for transparent, angular objects is not the same as an aesthetic sense, and no experiment separates either from novelty.
Two errors we could not reconcile. The paper places the Ga-Mohana crystals in Namibia; the site is in South Africa's Northern Cape.4 Its Results say all five chimpanzees engaged with the crystal, while the same paragraph refers to four individuals.1 We have used five, the figure stated with an explicit N. No published critique of the paper exists yet; it is four months old.
This is a live dispute in Palaeolithic archaeology, and the chimpanzee study inherits it whole. It is a dispute about evidence, not values — better excavation could in principle settle it.
Well-formed crystals turn up in hominin-occupied shelters unworked, unperforated and, where anyone has checked, geologically foreign to the site. At Ga-Mohana Hill, Jayne Wilkins and colleagues recovered 22 calcite crystals from 105,000-year-old deposits, showed they could not have arrived by natural processes, and traced the nearest source to about 2.5 km away.4 Objects carried in for no practical reason are, by definition, collected.
The taphonomic test Wilkins applied has never been applied to the older finds. Zhoukoudian's crystals were reported in 1931 in a paper about quartz artefacts, at a site where quartz was the worked toolstone; absence of retouch is not evidence of aesthetic motive. Robert Bednarik's review — the study's principal authority — calls the Zhoukoudian material only possibly non-utilitarian.5 The study itself concedes the record may reflect what excavators looked for.1