Recent work has shown that most globular clusters have at least two chemically distinct components , as well as cluster-to-cluster differences in the mean [ O/Fe ] , [ Mg/Fe ] , and [ Si/Fe ] ratios at similar [ Fe/H ] values .
In order to investigate the implications of variations in the abundances of these and other metals for H-R diagrams and predicted ages , grids of evolutionary sequences have been computed for scaled solar and enhanced \alpha -element metal abundances , and for mixtures in which the assumed [ m /Fe ] value for each of the metals C , N , O , Ne , Na , Mg , Si , S , Ca , and Ti has been increased , in turn , by 0.4 dex at constant [ Fe/H ] .
These tracks , together with isochrones for ages from \approx 6 to 14 Gyr , have been computed for -3.0 \leq [ Fe/H ] \leq - 0.6 , with helium abundances Y = 0.25 , 0.29 , and 0.33 at each [ Fe/H ] value , using upgraded versions of the Victoria stellar structure program and the Regina interpolation code , respectively .
Turnoff luminosity versus age relations from isochrones are found to depend almost entirely on the importance of the CNO-cycle , and thereby mainly on the abundance of oxygen .
Since C , N , and O , as well as Ne and S , do not contribute significantly to the opacities at low temperatures and densities , variations in their abundances do not impact the predicted { T _ { eff } } scale of red giants .
The latter is a strong function of the abundances of only Mg and Si ( and Fe , possibly to a lesser extent ) , because they are so abundant and because they are strong sources of opacity at low temperatures .
For these reasons , Mg and Si also have important effects on the temperatures of MS stars .
Due to their low abundances , Na , Ca , are Ti are of little consequence for stellar models .
The effects of varying the adopted solar metals mix and the helium abundance at a fixed [ Fe/H ] are also briefly discussed .