REQUEST: distribution of the T square chart

["Jim Clauson, Breakthrough Systems" <jim@jclauson.com>]

DENizens,

The following is relayed for  Gulay Sahinsev sgulay@hotmail.com - Hotmail 
rendered his message in html-only and I had to reformat for him.

Hi Denizens,

I was lately confused about the distribution of the T square:

In their book named "Applied Multivariate Statistical Analysis", Johnson 
R.A. and Wichern D.W. (1992) state that T square is distributed as (n-1)p / 
(n-p) F, where the degrees of freedom of F distribution are p, n-p.

Here p is the number of correlated variables and
n is the number of observations.

Coming to my question, where does the constant (n-1)p / (n-p) come from? If 
I want to change the covariance matrix in the T square chart, do you think 
that the constant of this distribution changes?

Gulay Sahinsev