R help - Feb 2006 - Graybill-Bowden confidence band

Hi R folks,

Anyone know how I can compute the alpha significance point for a  
distribution described as "the maximum absolute value of k Student t- 
variables, each based on v degrees of freedom and having a common  
pairwise correlation coefficient rho"? In my case rho is known to be 0.

Or, more to the point, anyone know how to compute a Graybill-Bowden  
confidence band around my OLS regression line?

Thanks a bunch!
Will Terry