Long ago I knew a fellow whose mother was a "goldbeater's bookie." That's the person who takes the sheets of beaten gold out of the frame in which they are hammered and puts them into a book, the form in which the buyer sees it. The process is done with static electricity, rather than tongs or spatulas, to avoid damaging the leaf, which I believe is only 200 atoms thick. The family joke was that after she died she was to be put into the shop's recovery furnace, where the floor sweepings were fired to recover lost gold.

Thus, recovery is possible, and the correct method is to fire your you-know-what in a porcelain crucible. You can calculate the weight of gold you would recover from perhaps 4 square cm of leaf at 200 atoms thick. You can then calculate the amount of you-know-what you would have to fire to get any particular amount of gold. The next step would be to run the experiment, weight the gold recovered, compare it with the calculated gold eaten, and determine the percent of recovery.

Any amateur scientists out there? Analytical balances ready? Crucibles ready to heat?