The temperatures and colors of stellar models are much less secure than predicted luminosities because they depend sensitively on the still very uncertain physics of stellar envelopes and atmospheres .
Consequently , it is important to ensure that the models which are used to interpret stellar populations data satisfy existing observational constraints on these properties .
As shown in this study , the available T _ { eff } contraints do not pose a problem for evolutionary calculations in which the free parameter ( \alpha _ { MLT } ) in the Mixing-Length Theory of convection ( which continues to be widely used ) , is set to the value required by a Standard Solar Model .
That is , it is still not possible to say whether some dependence of \alpha _ { MLT } on mass , metallicity , or evolutionary state should be invoked .
On the other hand , the evidence seems compelling that adjustments of the color transformations from model atmospheres are needed to achieve consistency with observations of cool stars .
Stellar models that have been normalized to the Sun and transformed to the observed plane using recent empirical color– T _ { eff } relations are able to provide superb fits to the M 67 and Hyades color-magnitude diagrams ( CMDs ) .
Because the properties of metal-poor stars and star clusters ( metallicities , distances , etc . )
are more uncertain , it is much harder to constrain the T _ { eff } and color scales at low metallicities than at [ Fe/H ] \approx 0.0 .
These difficulties are highlighted in the following examination of the constraints provided by Population II subdwafs and several well-observed globular clusters ( M 3 , M 5 , M 68 , M 92 , and 47 Tuc ) .
It is not clear that significant improvements to current semi-empirical color– T _ { eff } relations for metal-deficient stars will be possible until much tighter observational constraints become available .
Brief comparisons of the evolutionary tracks computed by different workers are also included in this investigation .