2014MNRAS.445.3839N

We compare model predictions to observations of star counts in the red giant branch bump (RGBB) relative to the number density of first-ascent red giant branch at the magnitude of the RGBB, EWRGBB. The predictions are shown to exceed the data by (5.2±4.3) percent for the BaSTI models and by (17.1±4.3) percent for the Dartmouth models, where the listed errors are purely statistical. These two offsets are brought to zero if the Galactic globular cluster metallicity scale is assumed to be overestimated by a linear shift of ∼ 0.11 and ∼ 0.36dex, respectively. This inference based on RGBB star counts goes in the opposite direction to the increase in metallicities of Δ[M/H] ~ 0.20dex that would be required to fix the offset between predicted and observed RGBB luminosities. This comparison is a constraint on `deep-mixing' models of stellar interiors, which predict decreased RGBB star counts. We tabulate the predictions for RGBB star counts as a function of [Fe/H], [α/Fe], CNONa, initial helium abundance, and age. Though our study suggests a small zero-point calibration issue, RGBB star counts should nonetheless be an actionable parameter with which to constrain stellar populations in the differential sense. The most significant outliers are towards the clusters NGC 5024 (M53), NGC 6723, and NGC 7089 (M2), each of which shows a ∼ 2σ deficit in their RGBB star counts.