R help - Mar 2011 - EM and Mixtools

I have 2 questions concerning the EM algorithm. Is it true that the
EM algorithm gives unique answers for the means and variances of a mixture
of 2 normals? I am using mixtools and I am surprised that it works better
than a Bayesian program I wrote.
If so can someone say why the mixing probabilities are so good?

-- 
Thanks,
Jim.

	[[alternative HTML version deleted]]