Left columns: luminosity functions for blue galaxies in high (green) and low σ (orange) clusters (top) and for low L_X (orange) and high L_X (green) clusters (bottom). Right columns: corresponding error ellipses. Again, care should be taken over the interpretation of the fitted parameters given the quality of the fits, though it is clear when simply looking at the data points that at least in the case of splitting by X-ray luminosity, there is a clear difference by the two subsamples. Continous and dashed lines and contours have the same meaning as in Fig. 6. The dashed line and contours for the split by X-ray luminosity show the (improved) fit for the high L_x subsample when only ranging over the same NUV₀ values as for the low L_x subsample. In this case, the faint end slopes between the two subsamples are similar and the main difference is in a ∼1 mag shift in M^*, it being fainter for the high L_x subsample.