Luminosity functions and best fits left column) and associated error ellipse (right column) for low (orange) and high (green) σ clusters (top row), low (orange) and high (green) L_X clusters (middle row) and sparse (orange) and dense (green) clusters (bottom row). The circle in each figure indicates the M^* and α values for the total LF. Clearly, for the subsamples split by X-ray luminosity and surface density, the luminosity distributions as sampled here are not well fitted by Schechter functions, the location and size of the error ellipses in these cases should be used with caution. Any interpretation should take note of the actual numbers rather than the fits to the data. The continuous lines and contours are the fits to the entire range of NUV₀ values available for a given subsample. The dashed lines and contours are for Schechter fits restricted in range in NUV₀ where both subsamples have data. Clearly, in the case of splitting by L_x, even this does not result in a good fit for the high L_x subsample, though obviously the data imply a lower ratio of high to low L_UV galaxies in the high L_x subsample relative to that of the low L_x subsample. For the split by surface density, restricting the fitting range has no real effect.