Radick: We have 3 or 4 minutes for questions...Tom?

Ayres: It'll take longer than that [laughter]. David, I'd really like you to comment on the HIPPARCOS results. At Cool Stars 10, there was a lot of discussion about the dichotomy of the main sequences of Praesepe and the Pleiades, and John Stauffer argued very persuasively that the differences couldn't be explained by our current understanding of stellar evolution. The HIPPARCOS folks had gone down a list of systematic effects that they'd studied and eliminated in determining the cluster distances, and what it comes down to is, nobody knows what's going on.

Soderblom: Right; there are no knobs in there you can twist far enough to make the Hipparcos observations agree with models, and this is a fundamental problem. It comes down to being us, spectroscopists who study stars, versus the astrometrists, people who have devoted their careers to understanding the nitty-gritty details of analyzing HIPPARCOS observations. There's no way I can second-guess them on that stuff, nor do I want to. At the same time, there's something funny going on, and several ideas come to mind. First, suppose their distance to the Pleiades is correct, and those stars really are three tenths of a magnitude fainter than we thought up to now. . Then surely the Pleiades is not totally unique in this Galaxy we live in, and there must be other stars around that are like them, and there must be a few of them in the solar neighborhood. So I go out and look at that HR diagram I showed you, and look at solar-type stars that are below the apparent ZAMS, and I find any young ones that happen to be near the Sun. I don't really have an answer for that yet, but I have a paper on this in preparation. Another possiblity was that if HIPPARCOS is wrong by just about exactly one millarcsecond, that is, change the HIPPARCOS Pleiades parallax by one milliarcsecond then the Pleiades are bang on, the distance modulus is 5.6, and everybody's happy. Curiously, the Hyades, which is in the same general part of the sky, is also ``off'' by about one milliarcsecond compared to the distance you would have thought prevailed the ground-based observations.

Ayres: But what about Praesepe? It's the same distance as the Pleiades, and the HIPPARCOS parallax there is right on.

Soderblom: Well, what I'm getting at is, the HIPPARCOS measurements are correlated in regions of the sky, so if there are ``zonal errors,'' this is the kind of thing it might affect. So suppose, for example, that you look at bright, early A stars, and their apparent density with distance. You know that they're all in the HIPPARCOS catalog because they're bright. The density ought to look funny as you get out to the distance where the error is comparable to the parallax itself.

Garrison: Just to answer your question about whether spectral types are still useful: I never would have said that they were more useful than anything else. But I think they are useful as a complementary check, and if a good-quality, moderate-dispersion spectrum of a star doesn't look like the Sun, then no matter how quantitative you get otherwise, it still doesn't look like the Sun. So you have to use all the data in every way you possibly can.