UV extinction as a fundamentally better measure of dust

Using a physically grounded measure of dust and dust-to-gas ratio is of fundamental astrophysical importance. The gas-to-dust ratio of reddened stars in the Milky Way and Magellanic Clouds (and of galaxies in general) is commonly expressed as a linear relation between the hydrogen column density, N(H), and the reddening, E(B-V) (or more recently A(V), the extinction at V band). However, because extinction curves to individual stars vary, there is no reason why (except by chance) either E(B-V) or A(V) would be the best measure of dust as opposed to extinction at some other wavelength. We have utilized for the first time full extinction curves in the MW, and have found that UV extinction around 2900 Å is better correlated with gas, and that the relation is linear. Scatter and nonlinearity increase for A(V) and for E(B-V), and are especially poor for near-IR extinction. We also find that the very large discrepancy between MW and SMC gas-to-dust ratios of 1.2 dex in N(H)/E(B-V) is reduced to ~0.7 dex for far-UV extinction at A(1500), which matches the difference in cosmic abundances of carbon.