We present spectroscopic linestrength data for 4097 red-sequence galaxies in 93 low-redshift galaxy clusters , and use these to investigate variations in average stellar populations as a function of galaxy mass .
Our analysis includes an improved treatment of nebular emission contamination , which affects \sim { } 10 \% of the sample galaxies .
Using the stellar population models of D. Thomas and collaborators , we simultaneously fit twelve observed linestrength - \sigma relations in terms of common underlying trends of age , [ Z/H ] ( total metallicity ) and [ \alpha /Fe ] ( \alpha -element enhancement ) .
We find that the observed linestrength- \sigma relations can be explained only if higher-mass red-sequence galaxies are , on average , older , more metal rich , and more \alpha -enhanced than lower-mass galaxies .
Quantitatively , the scaling relations are age \propto \sigma ^ { 0.59 \pm 0.13 } , Z/H \propto \sigma ^ { 0.53 \pm 0.08 } and \alpha /Fe \propto \sigma ^ { 0.31 \pm 0.06 } , where the errors reflect the range obtained using different subsets of indices .
Our conclusions are not strongly dependent on which Balmer lines are used as age indicators .
The derived age - \sigma relation is such that if the largest ( \sigma \sim 400 km s ^ { -1 } ) galaxies formed their stars \sim 13 Gyr ago , then the mean age of low-mass ( \sigma \sim 50 km s ^ { -1 } ) objects is only \sim 4 Gyr .
The data also suggest a large spread in age at the low-mass end of the red sequence , with 68 % of the galaxies having ages between 2 and 8 Gyr .
We conclude that although the stars in giant red galaxies in clusters formed early , most of the galaxies at the faint end joined the red sequence only at recent epochs .
This “ down-sizing ” trend is in good qualitative agreement with observations of the red sequence at higher redshifts , but is not predicted by semi-analytic models of galaxy formation .