The U. C. Berkeley Infrared Spatial Interferometer has measured the mid-infrared visibilities of the carbon star IRC+10216 and the red supergiant VY CMa. The dust shells around these sources have been previously shown to be time-variable, and these new data are used to probe the evolution of the dust shells on a decade time-scale, complementing contemporaneous studies at other wavelengths. Self-consistent, spherically-symmetric models at maximum and minimum light both show the inner radius of the IRC+10216 dust shell to be much larger (150 mas) than that expected from the dust condensation temperature, implying that dust production has slowed or stopped in recent years. Apparently, dust does not form every pulsational cycle (638 days), and these mid-infrared results are consistent with recent near-IR imaging which indicates little or no new dust production in the last three years (Tuthill et al 2000). Spherically symmetric models failed to fit recent VY CMa data, implying that emission from the inner dust shell is highly asymmetric and/or time-variable.

Measurements by the U.C. Berkeley Infrared Spatial Interferometer at 11.15 micron have yielded strong evidence for multiple dust shells and/or significant asymmetric dust emission around NML Cyg. New observations reported also include multiple 8-13 micron spectra taken from 1994-1995 and N band (10.2 micron) photometry from 1980-1992. These and past measurements are analyzed and fitted to a model of the dust distribution around NML Cyg. No spherically symmetric single dust shell model is found consistent with both near- and mid-infrared observations. However, a circularly symmetric maximum entropy reconstruction of the 11 micron brightness distribution suggests a double shell model for the dust distribution. Such a model, consisting of a geometrically thin shell of intermediate optical depth ($\tau_{11 micron} \sim 1.9$) plus an outer shell ($\tau_{11 micron} \sim 0.33$), is consistent not only with the 11 micron visibility data, but also with near-infrared speckle measurements, the broadband spectrum, and the 9.7 micron silicate feature. The outer shell, or large scale structure, is revealed only by long-baseline interferometry at 11 micron, being too cold ($\sim$ 400 K) to contribute in the near-infrared and having no unambiguous spectral signature in the mid-infrared. The optical constants of Ossenkopf, Henning, & Mathis (1992) proved superior to the Draine & Lee (1984) constants in fitting the detailed shape of the silicate feature and broadband spectrum for this object. Recent observations of H$_2$O maser emission around NML Cyg by Richards, Yates, & Cohen (1996) are consistent with the location of the two dust shells and provide further evidence for the two-shell model.