Scheffe is a very conservative multiple comparisons adjustment. It's goal isn't just to adjust for all pairwise comparisons but also to adjust for any possible contrast you might throw at the model post-hoc. These could be something like testing Ho: 7*Treatment1Mean + (1/3)*Treatment2Mean = 3 versus Ha: 7*Treatment1Mean + (1/3)*Treatment2Mean
eq 3
If you're only interested in pairwise comparisons then a Tukey adjustment is the route to go.