3.3 Using a Sightline-Specific Curve
Without question, the best of all possible dereddening choices is to use an extinction curve derived explicitly for the sightline of interest. This advice is not so unhelpful as it may sound, because the archives of the IUE satellite may well contain normal stars in the direction (and at the distance) of the object of interest, from which extinction curves may be derived (e.g., halo stars in the case of extragalactic sightlines). The accuracy of the dereddened energy distribution may thus be limited chiefly by the uncertainties in the ``pair method'' of constructing the extinction curve (see, e.g., Massa et al. 1983) and by the determination of E (B - V) for the object of interest. For a well-determined curve, ``ironing out'' the bump - in the cases when E (B - V) is not known explicitly - should yield the same degree of accuracy as for the cases when E (B - V) is known a priori. An example of such an approach is the determination of the intrinsic energy distributions of rare or exotic objects located in clusters with more mundane objects. Massa & Savage (1985) utilize this technique, using extinction curves derived from main sequence B stars in open clusters to determine the continuum shapes of cluster O stars.