We propose a new strategy to search for dark matter axions in the mass range of 40--400 $\mu$eV by introducing dielectric haloscopes, which consist of dielectric disks placed in a magnetic field. The changing dielectric media cause discontinuities in the axion-induced electric field, leading to the generation of propagating electromagnetic waves to satisfy the continuity requirements at the interfaces. Large-area disks with adjustable distances boost the microwave signal (10--100 GHz) to an observable level and allow one to scan over a broad axion mass range. A sensitivity to QCD axion models is conceivable with 80 disks of 1 m$^2$ area contained in a $10$ Tesla field.

Stellar evolution models which include the effect of helium and heavy element diffusion have been calculated for initial iron abundances of [Fe/H] = -2.3, -2.1, -1.9 and -1.7. These models were calculated for a large variety of masses and three separate mixing lengths, \alpha = 1.50, 1.75 and 2.00 (with \alpha = 1.75 being the solar calibrated mixing length). The change in the surface iron abundance for stars of different masses was determined for the ages 11, 13 and 15 Gyr. Iron settles out of the surface convection zone on the main sequence; this iron is dredged back up when the convection zone deepens on the giant branch. In all cases, the surface [Fe/H] abundance in the turn-off stars was at least 0.28 dex lower than the surface [Fe/H] abundance in giant-branch stars of the same age. However, recent high dispersion spectra of stars in the globular cluster NGC 6397 found that the turn-off and giant branch stars had identical (within a few percent) iron abundances of [Fe/H] =-2.03 (Gratton et al. 2001). These observations prove that heavy element diffusion must be inhibited in the surface layers of metal-poor stars. When diffusion is inhibited in the outer layers of a stellar model, the predicted temperatures of the models are similar to models evolved without diffusion, while the predicted lifetimes are similar to stars in which diffusion is not inhibited. Isochrones constructed from the models in which diffusion is inhibited fall half-way between isochrones without diffusion, and isochrones with full diffusion. As a result, absolute globular cluster ages which are based upon the absolute magnitude of the turn-off are 4% larger than ages inferred from full diffusion isochrones, and 4% smaller than ages inferred from no diffusion isochrones.