Trapezium cluster massive stars
stellar types listed in Table 1 of Ferland+12
 http://adsabs.harvard.edu/abs/2012ApJ...757...79F

That paper used TLusty stellar atmospheres.
Trapezium_Tlusty shows this

Wagle+ in preparation claim that WMbasic gives better fit to [Ne III] lines.
Trapezium_WMbasic shows this

Veusz file and pdf compare these two.

***************************

The original form came from Will Henney in email of
2012 Dec 5.  His notes follow:

    Stellar parameters for C, A, D are from Simón-Díaz \cite{2006A&A…448..351S}
    I originally had th1B as a B2V star, which should be cooler than what Gary had (28kK and 10**4.4 Lsun), but I can't find the provenance of this
        actually th1B (BM Ori) is a quintuple system: see SIMBAD record
        SIMBAD says primary is B1V citing \cite{Close:2003}, but that paper does not mention spectral types at all. I suspect this is a confusion with B1, B2, etc being the designations of the quintuple components
        \cite{Weigelt:1999} find a B3 spectral type for the primary (B1) with log L = 3.25 and T = 18000
        B2, B3, and B4 are all lower mass, but B5 has mass similar to B1 accoding to \cite{Weigelt:1999}. However, this must be a mistake: according to \cite{Abt:1991}, who detect B1 and B5 as an eclipsing spectroscopic binary, the mass ratio is 0.3, implying that B5 has a mass of only 2 Msun..
    th1C is a triple system \cite{Lehmann:2010a}
        \cite{Schertl:2003} have broad constraints on the parameters of C2 (dark gray line in their Fig 4). If we take the middle of this range, we get a giantish B3IV star: log L = 3.8, T = 17000, M = 9 Msun, age much less than 1 Myr.
        We can work out the surface gravity:
            g = G M / R2 = G M 4 pi sigma T4 / L = 6.673e-8 9 1.989e33 4 pi 5.6703e-5 17000**4 / 10**3.8 3.82e33 = 2950 cm/s2 => log g = 3.5
        [X] We should check these parameters against \cite{Lehmann:2010a}, who say 12 Msun in their abstract. The article itself is not freely available -grab it when I get back to work
            If we use the same evo tracks as in \cite{Schertl:2003}, then this would imply B2IV with log L = 4.0, T = 21000
            So, after checking out the Lehman article, it seems the 12 Msun is robust, since it is based on orbital solution. The translation to luminosity and Teff is rather more speculative.
            In fact, according to \cite{Kraus:2007}, the companion is O9.5V with Teff = 32000, log L = 4.68. This is also consistent with the Schertl tracks, but it is assuming 15.5 Msun.
            If we instead use 12 Msun as the best mass estimate \cite{Lehmann:2010a}, then we would get log L = 4.2, Teff = 25,000
                Recalculate surface gravity for new parameters
                    g = G M / R2 = G M 4 pi sigma T4 / L = 6.673e-8 12 1.989e33 4 pi 5.6703e-5 25000**4 / 10**4.2 3.82e33 = 7322 cm/s2 => log g = 3.86

2014 May 06, changed some tlusty stars to WMBasic - had used tlusty for all stars originally.  WMBasic produces better agreement with H II region optical emission, especially [Ne III]
            Using this new solution for th1C2 changes the broadband ratios by about 5%.
    th1A is a double system

stars = dict(
    th1C = dict(T=39000., g=4.1, L=5.31),
    th1A = dict(T=30000., g=4.0, L=4.45),
    th1D = dict(T=32000., g=4.2, L=4.47),
    th1B = dict(T=18000., g=4.1, L=3.25),
    th1C2 = dict(T=25000., g=3.86, L=4.2),
    )
