Try and get ahold of Norm Campbell formerly of CSIRO Mathematics and statistics, WA. Norm is the greatest authority on your problem. Norm took early retirement last year but you can still find him through the CSIRO net.
Darroch is in SA, stats Dept. Ask him for the paper from 1985 - at leat he was there when I met him in 1986. I seem to have missed earlier correspondence but I suspect that you are trying to standardize distance measures by some kind of a division. This means you should use appropriate methods for compositional data. Check back with me if that is indeed so. CVA means Canonical Variate Analysis - the subject of Norm's PhD in Mathematical statistics at Imperial College, London. Richard A. Reyment FD (Stockholm), D.Sc. (Melbourne). == Replies will be sent to list. For more information see http://life.bio.sunysb.edu/morph/morphmet.html.
