Hi:

Sorry not to comment on HD8907 prior to this - I had
other things I needed to be doing, and I basically thought
that what Dave and Eric said sounded reasonable and I
likely wouldn't have anything different to say.

Lynne's email from a day or two ago
prompted me to at least spend a few minutes to see if I
had anything to contribute.   Attached are four postscript
plots, comparing HD8907's CMD position to stars in the
Pleiades and Praesepe (using our 133 pc for Pleiades, and
m-Mo = 6.05 for Praesepe) and providing isochrones from
D'Antona.    The B-V plots put HD8907 right
on the single star main sequence; the V-K plot puts it
a little above.   Based on the B-V plots, I agree with
Lynne's position - that is, the star does sit on the
main sequence, and there isn't much useful you can say
about its age other than it is not older than X from the
photometry.

The D'Antona plot shows isochrones for 100 Myr (lowest
curve), 1 Gyr and 3 Gyr.  The points are 1.4, 1.3, 1.2,
1.1, 1.0 etc. solar masses.  Acc. to Gray, a B-V = 0.49
star should have mass ~ 1.25 Msun on the main seq. -
or, alternatively B-V = 0.49 corresponds to Teff ~ 6300
or LogT = 3.80.

With Valenti's Fe/H ~ 0.00 for the star, I think the
direct comparisons to Pleiades should be valid; if
Praesepe is as metal rich as the OSU folks want, there is
probably a 0.1 mag or so shift that should be applied
when comparing its sequence to HD8907 (but the distance
to Praesepe is based on MS fitting and not astrometric
binaries as for Pleiades, so the "vertical" position of
the Praesepe stars in the CMD is at least than uncertain).
The distance modulus I used:

Anyway, to my eye, these figures suggest that 3 Gyr is
pretty much out of bounds, and < 1 Gyr is probably a
good conclusion.

John






> regarding the hr diagram age of HD8907:
>
> I guess I don't understand why we think we can learn anything about the
> age of a star that sits on the main sequence.  this star
> is older than 20 Myr and younger than 5-7 Gyr, but i don't see how  a
> probability distribution for the age can be derived.
>
> as i understand it, the valenti/fischer method of age determination
> samples the error distribution in the stellar paramters +/- 4-sigma
> and uses this to derive an age distribution.
>
> when one uses pre-ms tracks to interpolate, one finds that the
> solar-type main sequence stars pile up around 30 Myr of age, the age
> they would have if they were slightly pre-zams.  thus it is not
> surprising that the age derived using post-ms tracks is 1-3 Gyr, the
> age the star would have it it were slightly post-zams.  this behavior
> is caused by the fact that there are no solutions when sampling the
> observational error distribution  below the main sequence.  so all the
> weight in the probability distribution is towards errors which put the
> star above the main sequence, and hence slightly evolved.
>
> i can't quite parse the details of what jeff is doing from the text in
> the paper, but i suspect some of the above is going on.
>
> perhaps he is well aware of this effect and has tried to mitigate it. i
> would be curious for his comments.