This paper asks whether a particular way of measuring the Hubble constant H0 with gravitational waves is biased if the intrinsic mass distribution of binary black holes changes with redshift. The method, called spectral‑siren cosmology, uses the fact that gravitational‑wave detectors measure the masses stretched by cosmic expansion. Features in the true, source‑frame mass distribution—peaks or drops in where black holes cluster—act like an internal ruler that ties observed masses to redshift. If that intrinsic mass spectrum itself evolves over cosmic time, it could in principle shift the inferred H0.

The authors revisit measurements using the GWTC‑4.0 catalog from the LIGO–Virgo–KAGRA network. They adopt a standard parametric mass model used in previous population studies: a broken power law plus two Gaussian peaks (called BPL+2P). They allow the main mass‑scale parameters (for example the two power‑law slopes, the break mass, and the two peak locations and widths) to change smoothly with redshift. The smooth change is described by a transition function controlled by a characteristic redshift and a transition width. The full analysis fits cosmology (H0 and matter density), merger‑rate evolution, and a flexible 45‑parameter population model to the data, with weak priors.

Their main result is that, at current detector sensitivity and with the GWTC‑4.0 catalog, they find no compelling evidence that the binary‑black‑hole mass spectrum evolves with redshift. Allowing evolution does move the posterior for H0 slightly toward lower values, but the shift is modest and not statistically significant. The authors also show with diagnostics at the population and event level that this small change is consistent with an over‑flexible population model: when they analyze simulated signals drawn from a non‑evolving population with the evolving model, the same mild shift appears. In other words, the extra freedom can nudge H0 when the data do not strongly constrain the extra parameters.