Just a guess.
Perhaps if you expand the experiment with a range of heads for each orifice it might produce a fuller picture of what is happening.
Plot a curve of head against velocity with five or six different heads and see what you get.
I am guessing, from the limited results you have posted, that the relationship between head and velocity for a given orifice size might be nonlinear.
There is possibly also a nonlinear relationship between velocity and orifice size for a given head.
If this is the case, then you might find there is an optimum head for any given orifice size (and visa versa an optimum orifice size for any head).
__________________
You have GOT to Laugh !
