Connecting observational data to recent theoretical prescriptions for the luminosity of RR Lyrae, we present a new estimate of the bolometric magnitude of the first brightest red giant for a selected sample of galactic globular clusters. Comparison with theoretical prescriptions for the red giant luminosity at the He-flash discloses an excellent agreement. We also investigate the consistency of the non-canonical scenario for the luminosity of RR Lyrae suggested by Sandage, implying either an anticorrelation between Y and Z or nonstandard values for the RG core masses. We show that both these possibilities disagree with observational data producing a non-consistent scenario.