R help - Jul 2010 - help me on discriminant analysis and Bayesian discriminant analysis in R

Dear R-listers,

I have a set of plant multivariate morphological data (including
thousands of individuals from several species) which may be identified
as several units of morphological variation.

Firstly, I want to defined the optimal number of such units
Then, I want to choose proper discriminant analysis tool to determine
which of the unit each individual is belonging to. It was said that
Bayesian discriminant analysis is better. I prefer Bayesian tools.

I have searched the R-related websites and read several books. But, I
still have not find any helpful information.

Could you please give any directions on my problem? That is vary valuable to me.

Best

Sincerely,
Mao J-F