sample variance-covariance matrices

["gulay sahinsev" <sgulay@hotmail.com>]

Dear Denizens,
I have been researching about multivariate control charts, and have recently 
read a paper about a comparison of multivariate control charts for 
individual observations by Sullivan and Woodall (1996).
I have 76 observations obtained from an aircraft component production line, 
and I am currently trying to apply some different approaches for their 
multivariate analysis.
In Sullivan and Woodall's paper, they state that the sample covariance 
matrix suggested by Holmes and Mergen which uses one half the differences Vi 
= X(i-1) - Xi   i= 1,2,..., m-1

S5 = 1/2 V V' /(m-1)
is approximately Wishart distributed. What is its degrees of freedom then?

Do you suggest using this sample variance-covariance matrix?
Thank you very much for your suggestions in advance.Any further papers and 
researches referred in this paper were unfortunately not available.
If anybody is interested in the results of the study, I can give further 
information later on as my study proceeds.