We use a laboratory experiment, calibrated to a representative New Zealand fishery, to assess three sealed-bid, multi-unit demand auction mechanisms under consideration for allocating quota for species being introduced into the Quota Management System. These auctions allocate the K units to the K highest bids, but prices are determined according to discriminative, Kth price, and K + 1st price rules. In our experiment, the auctions are equally efficient, but revenue is highest in the discriminative auction, and lowest in the K + 1st price auction. Results are robust to industry subjects, and subjects are responsive to the across-unit bidding incentives present in multi-unit demand auctions.