I recommend sand of .2 to .5 diameter grain size, which is to say that it's been meshed such that no grains outside that range are included. This leaves an excellent balance of surface area to interstial space (open space between grains) allowing good rates of diffusion up and down through the substrate, which will keep the lower reaches from becoming supersaturated with H2S and other toxic products of decay.
Phone building supply suppliers - sand and gravel quarries, that sort of thing. Ask if they can double mesh sand for you and what it will cost. They'll quote by the TON! Won't be as much as you might expect. Several people going in on it together, take what you want and maybe the supplier can just take back the rest and re -sell it if you don't a a large bin to store it in. Sounds like a chore, but you can re sell the sand at 50 cents a pound if you can find folks interested.
My LFS started carrying this sand soon after I specified it in this article here:
http://www.wetwebmedia.com/ca/volume_7/volume_7_1/dsb.html
I can't prove anything but I do feel proud when I see those bags of sand in the LFS. Many deep sand bed, planted freshwater tank folk in this city are using this sand and I've heard nothing other than good things. The LFS owner told me he thinks they cycle in about half the time of un-meshed sad.
I agree with Wes, UGFs and RUGFs pro'ly won't work with it and would eliminate the anaerobic and anoxic regions so would result in less fertile substrate and much less denitrification.