number of clusters for k-means

I am currently working on a clustering project and would like to obtain 
statistics for the number of clusters to include.  In SAS you get a 
pseudo-F statistic and a cubic clustering criterion.  Has anyone 
developed a function to get these values?

Sean