France had begun testing atomic bombs above ground near Reggane in 1960, but fallout from open-air detonations was impossible to control. So the program moved underground and south, to In Ekker, about 150 kilometres north of Tamanrasset. There, engineers tunneled horizontally into the Taourirt Tan Afella — a granite peak of the Hoggar — carving each shaft into a spiral so the curve would crush and seal itself when the bomb went off, choking back gas and dust. Concrete plugs and four heavy steel doors were meant to finish the job. The whole elaborate design was a promise of safety, and it was that promise that justified bringing high-ranking spectators to stand and watch. Béryl was the second such test fired into this mountain.

The spiral did not collapse fast enough. The concrete plug was pulverized, the end door blasted free, and a fraction of the explosion's radioactivity escaped with the gas, molten slag, and lava. The lava cooled and hardened on the tunnel floor, but the cloud rose to roughly 2,600 metres and drifted east on a shifting wind, leaving fallout detectable for hundreds of kilometres downwind. Pierre Messmer, the French Defence Minister standing among the witnesses, later described seeing "a kind of gigantic blowtorch flame" leap horizontally toward them before the dark cloud bloomed. The official instruments that measured the radioactivity were immediately classified and locked away under military secrecy. What should have been a contained underground shot had become an open-air contamination event, witnessed by the very officials it endangered.

Nine soldiers of the 621st Groupe d'Armes Spéciales bore the worst of it. Stationed at an isolated post, they crossed the contamination zone after removing their masks, and absorbed doses estimated at around 600 millisieverts — many times the annual limit now permitted for nuclear workers. They were flown to a military hospital outside Paris for monitoring. As many as a hundred more personnel took lower doses, near 50 millisieverts, when the cloud rolled over the command post on that unexpected wind. Among the exposed were the ministers themselves: Defence Minister Messmer, who left that same evening after being decontaminated, and Gaston Palewski, the Minister of Scientific Research, who would die of leukaemia in 1984. And there were others the records barely mention — Tuareg villagers of the surrounding desert, exposed to the fallout, whose health was never systematically followed.

Official accounts insisted the nine irradiated soldiers showed no specific illness in the monitoring that followed, and the affair was buried in classified files for decades. But the Béryl incident did not stay buried. It surfaced in French documentaries and books — Vive la bombe!, Gerboise bleue, the testimony of survivors — and became a touchstone in Algeria's long demand for an honest accounting of the bombs detonated on its soil. The deepest injustice lies in the asymmetry of the record: the ministers' exposure is documented to the millisievert, while the Tuareg families who lived beneath the same cloud remain, in the official files, a single passing line. The granite of Tan Afella still holds vitrified rock and buried contamination, a sealed scar in the Hoggar that the desert has not been able to erase.