By a staggering co-incidence I'm reading a book on the subject at the moment. It goes by the name of "Aircraft Performance - theory and practice" by a chap called Martin Eselby. The publishers sent me a copy for my comments, which so far are fairly favourable, although I can't seem to digest this stuff as fast as I once did.
Consulting it (which to my relief roughly co-incides with my hazy memory), best glide (i.e. best endurance) co-incides with Vmd. However, best range in the glide is a little faster. Specifically it's the speed at which Cd / Cl^(3/2) is a minimum, which comes out at Vmd x 1.316. (1.316 = 3^0.25).
But this relationship assumes still air, and is the speed at which the tangent from the origin meets the drag curve. For a particular case, you should shift the origin left or right a bit to account for the headwind or tailwind component to give the airspeed for best range on the ground.
Ouch, that's too much thinking for a Saturday morning.
G