Using new narrowband color observations of early-type galaxies in clusters , we reconstruct the color-magnitude relation ( CMR ) with a higher degree of accuracy than previous work .
We then use the spectroscopically determined ages and metallicities from three samples ( Trager et al . 2008 , Thomas et al . 2005 , Gallazzi et al . 2006 ) , combined with multi-metallicity SED models , to compare predicted colors for galaxies with young ages ( less than 8 Gyr ) with the known CMR .
We find that the CMR can not by reproduced by the spectroscopically determined ages and metallicities in any of the samples despite the high internal accuracies to the spectroscopic indices .
In contrast , using only the < Fe > index to determine [ Fe/H ] , and assuming a mean age of 12 Gyr for a galaxy ’ s stellar population , we derive colors that exactly match not only the color zeropoint of the CMR but also its slope .
We consider the source of young age estimates , the H \beta index , and examine the conflict between red continuum colors and large H \beta values in galaxy spectra .
We conclude that our current understanding of stellar populations is insufficient to correctly interpret H \beta values and that the sum of our galaxy observations supports an old and monolithic scenario of galaxy formation . This result has a devastating impact on every study that has used the H \beta index to calculate galaxy age , as the use of the H \beta versus MgFe diagram will result in incorrectly deduced young ages .