Reddening of microlensed large Magellanic Cloud stars versus the location of the lenses

We propose an observational test that can break the indeterminacy of two main classes of microlensing models of the Magellanic Clouds: (a) the lenses are located in the Galactic halo, and (b) the lenses are located in the LMC disk. The source stars in the latter (self-lensing) models tend to be at the far side or behind the LMC disk, thus experiencing more reddening and extinction by dust in the LMC disk than ordinary stars in a nearby line of sight. Clearly, such bias would not occur in the MACHO halo lensing models. We show that this reddening effect is at a level readily observable for the present 30 or so microlensing alerts fields, either with multiband photometry from a good seeing site, or more definitively with ultraviolet (UV) spectroscopy with Space Telescope Imaging Spectrograph. Stars behind the LMC dust layer should stand out as UV-faint objects (by more than 1 mag than average stars in the LMC). HST can also resolve numerous faint neighboring stars within a few arcseconds of a lensed source, hence removing blending in these crowded regions and building a reddening map to control the patchiness of dust.