A BIBLIOGRAPHICAL PARADOX