For the past ten years or so, librarians have been discussing the way reference service is provided, the most conspicuous recent installments being the “Re‐thinking Reference in Academic Libraries” conference and the widely publicized changes at Brandeis University. No one has heard every statement in this debate because it has been so extensive in time, space, and medium. Nevertheless, it seems safe to say that mathematical queuing theory has not played a significant role in it. At first, this lacuna is puzzling, since queuing problems are one of the fundamental types treated in operations research, which in “those thrilling days of yesteryear,” the fifties, sixties, and early seventies, was taught in some library and information science programs. By 1978 articles that applied the theory specifically to reference work appeared. Yet, the second edition of a text on operations research for librarians that appeared in 1991 says not a word about reference queues. Perhaps this theory has been neglected in recent discussions because it is usually applied to telephone calls or local area networks and other configurations of computers, while its use in analyzing queues of people in banks, supermarkets, and the like was passed over as too mundane. Perhaps, too, the failure of banks to take the theory to its logical limits in arranging the queues for bank tellers has led to caution.