Mystery spot in supernova 1987A - Reflection or fluorescence by an interstellar cloud?

This paper explores fluorescence and reflection models of the companion to SN 1987A obseved by speckle interferometry, recalling a 1901 precedent. The apparent small angular size of the companion is a severe constraint. A fluorescence model cannot reach the observed brightness unless the ultraviolet burst from the supernova contained as many as 2 x 10 to the 58th ionizing photons. This is about 25 times stronger than generous current models. Even then, the expected line ratios and widths do not fit the observations. The absence of narrow H-alpha and H-beta lines in the supernova spectrum, the ratio of fluxes of the companion in H-alpha and forbidden N II line filters, the invisibility of the companion at 4861 (H-beta), and its detection at 5330 fail to agree with theory. A dust-reflection model is more promising, and the color can be reddened by the evaporation of small grains, but the model still falls more than about 1 mag short in brightness. Furthermore, a dust reflection should have increased in relative brightness in May-June 1987, rather than disappearing as the mystery spot did. If all the observations are correct, neither model is likely to work.