A review of the properties of the scattered starlight which contaminates the spectrum of reddened stars

When a star is observed behind an interstellar cloud of sufficient column density, we do not observe the direct light from the star, which is totally extinguished. Rather, we see only starlight scattered at small angles from the star. I use several papers published in New Astronomy to recount the different steps which permit understanding how, and under which conditions, scattered starlight can be more important than direct starlight in the spectrum of a reddened star. Associated problems -the fit of the extinction curve, the nature of the scatterers (small grains, or, atoms or molecules from the gas), the 2200A bump- are also, briefly, discussed.