The universe's first minutes produced a specific cocktail of light elements: hydrogen, helium, deuterium, and lithium-7. Big Bang nucleosynthesis calculations predict the primordial lithium abundance with precision tied to the cosmic baryon density measured independently from the cosmic microwave background. The prediction: A(Li) = 2.72 dex.
The observation: A(Li) ≈ 2.0–2.4 dex. Ancient halo stars — metal-poor survivors from the Galaxy's first billion years — show lithium abundances on the Spite plateau, a remarkably flat distribution across a wide range of temperatures and ages. The plateau's uniformity suggests something fundamental. Its value, roughly three times below the primordial prediction, is the cosmological lithium problem. For four decades, it has motivated proposals ranging from new particle physics to non-standard nucleosynthesis to decaying dark matter particles that would selectively destroy lithium.
Yang, Dou, Meng, Wu, and Bi (arXiv 2602.22516, February 2026) show that the problem disappears when stellar models include rotation. Stars are not passive containers for primordial material. They have convection zones whose depth depends on mass and age. They settle gravitationally — heavier elements sink. They diffuse. They rotate, and rotation drives meridional circulation that mixes material between the surface and the hot interior where lithium burns.
The models start with the full primordial abundance of 2.72 dex and evolve stars for 8 to 13 billion years with rotation, magnetic fields, gravitational settling, and radiative diffusion. The result: lithium depletes to the observed plateau range of 2.0–2.4 dex, naturally, through known stellar processes. The flatness of the plateau arises because the competing effects — convective mixing, gravitational settling, rotational transport — approximately balance across the relevant temperature range (5900–6400 K). Stars hotter than 6400 K have shallower convection zones and less depletion; stars cooler than 5900 K have deeper convection and more. The plateau is the band where the processes compensate.
The lithium was there. The stars ate it. The initial abundances from the first three minutes were correct. The discrepancy was in the assumption that stellar surfaces preserve what they inherited.