U-B and B-V extinction Brian Skiff Your logic is fine for relatively bright (say V < 9) and unreddened stars. If the star is right on the galactic plane (this one at only 0.5 deg latitude) and faint (like this one), then start assuming some reddening. For not-too-faint stars, a common value for the color excess is around 0.5 magnitudes in B-V, that is B-V is redder by that much. (The mag. 10-11 A0V stars common in the Milky Way spectral surveys typically have B-V ~0.5.) If you suspect the star is early type (A0 or earlier), then just form the UBV reddening-free parameter Q = -0.72(B-V) + U-B. For the values you sent, you get -0.50. Next look at the Q-to-MK type conversion in my colors calibration file: ftp://ftp.lowell.edu/pub/bas/starcats/uvby.calib ...down toward the bottom, and you see that Q of -0.5 corresponds to a mid-B type main-sequence star---entirely reasonable for a Herbig Ae/Be star, given all the uncertainties involved. Q in fact works for all spectral types, but doesn't yield unique values. But usually you have some other information to guess where a star is on the HR diagram, not UBV colors alone, so you can use that to restrict the guess. Just looking at the colors, if B-V = 0.8, that notionally corresponds to a K0V, so U-B _must_ be something like 0.35 or so. But you have U-B = 0.09. If you were looking at high galactic latitude, where there's no reddening, then you could argue that this was a very metal-poor K dwarf. And at high latitude you would not expect to see a B-type star except extremely rarely, and without reddening U-B would always be strongly negative, so the colors rule that out. Since you're in the plane, then *bing* you think "reddened B star much more likely than metal-poor K dwarf". And in this case the reddening appears to be about 0.9 in B-V (intrinsic B-V for mid-B star is about -0.1), not so surprising for a faint star at 0.5 deg latitude. The large V-I is puzzling, but could be caused by peculiarities in the shell around the star, etc. So it's not terribly disturbing.