This paper measures thorium (Th) — one of the heaviest observable elements — in 47 metal‑poor stars. Thorium is made by the rapid neutron‑capture process (the r‑process). By analysing a large, homogeneously treated sample, the authors test how consistently the r‑process produces very heavy elements and how those yields change over time and metallicity.

The team used high‑resolution stellar spectra taken between 2016 and 2024. They focused on stars already identified as r‑process‑enhanced (RPE). Instruments included the MIKE spectrograph on the Magellan II telescope, the duPont echelle, and the TS23 spectrograph at McDonald Observatory. To ensure reliable results they kept spectra with signal‑to‑noise ratios above about 30 at 4000 Ångström, resolving power between 40,000 and 80,000, and only report thorium detections at three‑sigma significance or better.

Their main findings are twofold. First, thorium abundances become less spread out as stars become less metal‑poor: the scatter in [Th/H] and [Th/Fe] falls from about 0.6 dex at the lowest metallicities to about 0.2 dex at higher metallicities. (A “dex” is a logarithmic step: 1 dex is a tenfold change.) Second, Th and the lanthanide elements europium (Eu) and dysprosium (Dy) are usually produced together. The average [Th/Eu] is close to zero across a wide metallicity range (roughly [Fe/H] = −3.0 to −1.5) and across a large range of r‑process enhancement. Still, the absolute logε(Th/Eu) range in the sample is 1.02 dex, with an observed scatter of ±0.20 dex and an estimated intrinsic scatter of ±0.11 dex at the lowest metallicities. From this the authors infer that about 68% of r‑process events yield Th/Eu ratios that vary only by roughly ±30%, while about 5% of events show much larger variation, changing yields by factors greater than about 3 and approaching tenfold differences.