Yet again on Vega in 2MASS

First of all, sorry to start another topic, but from google group was
not possible to resume the older one.

Anyhow, I'm still a bit puzzled about Vega magnitudes in the 2MASS. I
resume here part of the oder topic I was involved

That's what I thought before, but further research (and some
consultation with Martin Cohen) shows I was wrong. See Cohen et al.
2003, AJ, 126, 109. Near the end, you will find that if Vega could
have been measured by 2MASS, its JHK magnitudes would have been -0.001,
+0.019, -0.017. I think the original intention was to have Vega=0, but
the final result didn't quite come out that way.

Cohen says that the zero-points (zp) offset for Vega are +0.001,-0.019
and +0.017 to be used when computing absolute fluxes for a star of
magnitude m using the following formula
F(0)*10^(-0.4*(m+zp)).

I assume that zero-points relate with the Vega magnitudes (as
determined from a post facto check) and the following magnitudes for
Vega are to be used: J=-0.001, H=+0.019 and K=-0.017.
That agree with the following way of computing absolute quantities
F(0)*10^(-0.4(m-m_vega))
obviously giving right the same result as the previous formulation.

Just wondering because in the recent paper by Apellaniz
(astro-ph/0510785) he reports the Vega zero-point for V J H K and write
down 0.026, +0.001, -0.019, +0.017. At first check sounds ok cos he
uses right the same zp given by Cohen and that in the visible
determined from Bohlin. However, since the one in the visible
correspond to the observed magnitude of Vega in V band, to be
consistent he should write the value with opposite signs in the IR,
right?

I know it might sounds silly question, but just to be sure things are
done properly...

Still another question: when generating synthetic magnitudes, I would
think to procede this way: in the Johnson-Cousins bands, I use the most
updated synthetic model for Vega (with the absolute calibration of
Megessier 1995) since one expect the model to improve and refine over
the years matching better and better the observations.
However, when doing synthetic photometry in the 2MASS system, the best
bet is to stick to the absolute fluxes given by Cohen (that are
ultimately based on Kurucz 1992 Vega model), right? Even if current
Kurucz model is better (well, probably most of the changes are in U
band) for consistency one should keep the same model (with the Hayes
1985 absolute calibration) as used to define the 2MASS photometric
system, right?

Thanks in advance for the reply and if you wanna contact my directly
feel free to reply at the following luccas at utu.fi

Best Regards!