creating groups based upon correlations

I have 64 managers and a time series of their returns. I've run the
correlations for all the managers and now i'm looking to group them
based upon their correlations. So, for example, any managers that is
more htan 65% correlated with manager 1 goes in group1, any manger more
than 65% correlated with manager 2 goes in group 2, etc--but then you
get all these managers that fit into several different groups so this
obviously isnt the ideal way to do this. Any suggestions on how to do
this more effectively? thanks!