The cluster at the metal-poor end
NGC 5897 is the most distant and metal-poor globular cluster the Kepler Space Telescope observed through the K2 mission, and until now its stars had not been weighed by listening to them. Solar-like oscillations were detected in 20 red giant branch stars and 6 early asymptotic giant branch stars, and the frequency of maximum power excess was derived for each. Asteroseismic scaling relations turned those frequencies into mean masses: 0.74 plus or minus 0.01 solar masses on the red giant branch, 0.65 plus or minus 0.03 solar masses on the early asymptotic branch.[1]
The quantity the work is after is the difference between those two averages: an integrated mass loss of 0.08 plus or minus 0.03 solar masses between the two phases. The number matters less for its size than for where it was taken. This is a measurement at a metallicity of [Fe/H] = -2.04, in the very metal-poor regime where the observational record has stayed thin.[1]
The trend that runs the other way
Placed on the mass loss-metallicity relation for Type I globular clusters, this point extends the relation into the very metal-poor regime and supports a picture in which integrated red giant branch mass loss falls as metallicity falls. The same measurement has another reading: if asteroseismic scaling relations behave a little differently in the two phases, or if only a biased subset of stars shows detectable oscillations, the gap could come from the method rather than from mass loss. The study itself states plainly that observational work on this dependence has so far produced conflicting trends.[1]
The prosaic reading deserves its turn first. The K2 time series are not long enough to fully resolve the large separation pattern, and the study carries that as a systematic uncertainty of between 8 per cent and 9 per cent. The authors' note matters: deviations of this kind primarily influence the individual seismic masses rather than the inferred mass loss. The error budget therefore stays tighter around the difference than around either average on its own.[1]
The measurement that would move it
One cluster is one point, and the updated relation stands or falls on the next very metal-poor cluster measured the same way. If such a cluster is observed with a longer, uninterrupted photometric series, its published integrated mass loss will either fall inside the 0.08 plus or minus 0.03 solar-mass interval or outside it, and either outcome carries information. What the telescope named after Kepler lets us hear is a distribution of frequencies; drawing a mass out of it means carrying the error bar along.[1]