What freeing the metallicity did to this fit. ← all stars
| fit | Teff | log g | log n(He)/n(H) | Z | reduced χ² |
|---|---|---|---|---|---|
| three parameters | 32149 | 6.081 | -2.47 | held 0 | 2.61 |
| four parameters | 32155 | 6.065 | -2.49 | -0.72 ± 0.31 | 2.41 |
W(He II 4686) = -0.08 Å measured on the spectrum. The temperature moved +6 K.
| pair | correlation |
|---|---|
| Teff ↔ Z | -0.734 |
| log g ↔ Z | +0.327 |
| helium ↔ Z | +0.052 |
A correlation past about 0.5 means the two cannot be separated by these lines at this signal-to-noise, and the parameter it is paired with should be read as the pair rather than on its own.
Offset from the line centre in Ångström across, normalised flux up; the observed spectrum is the thin black trace and each model is drawn over it. The strip underneath is observed minus model for whichever fit the buttons select — every panel switches together, and they share a vertical scale. That strip is where a deep Balmer core stops hiding the difference: two models that look identical on the profile separate plainly in the residual. The figure in the top right is how far the models part company at their widest, in per cent of the continuum.