A published scientific paper is the end-result of a complex interaction between authors, referees, editors and publishers. Each brings to the process a different agenda, and a widely disparate adherence to standards of competence and integrity. Recent internationally recognized instances of major scientific fraud represent the tip of a potentially large iceberg formed as a consequence of the ineptness or delinquency of one or other elements in the chain of surveillance designed to exercise control over the process of scientific publishing. This subjective analysis attempts to explain why and where the regulatory mechanisms that ought to detect and eliminate the publication or the dissemination by other means of poor, erroneous, or frankly fraudulent scientific finds have broken down, and what can be done to fix them.