This thesis sets out to examine the problems that have been associated with paraquat, both in relation to its occupational use as a herbicide and its misuse in cases of accidental or deliberate human poisoning. In order to provide a frame of reference for the later discussion, the general properties of paraquat are reviewed, together with its general toxic effects and possible mode of action in mammalian systems. The degree of risk associated with paraquat use in normal agricultural practice is examined. The available published literature is reviewed and the results of studies on both formulation workers and spraymen are discussed. It is concluded that the use of paraquat does not constitute a significant risk to health when sprayed at concentrations of up to 0.5% paraquat ion. Studies of situations in which low volume/high concentration application methods have been used would indicate that they are likely to produce an unacceptable level of risk and constitute dangerous agricultural practice. The problem of paraquat misuse is examined, both in regard to its extent and the effectiveness of treatment in cases of human poisoning. Treatment measures currently advocated are reviewed and their effectiveness discussed in relation to the series of 108 poisoning cases which is presented. On the basis of the data set out it is concluded that the treatment of paraquat poisoning is only likely to be effective in those cases who have, (a) ingested less than 5g of paraquat ion, (b) had treatment instituted within 2 hours of ingestion, and (c) show plasma paraquat levels in the order of 1-2mg/litre 2-4 hours after ingestion. It is also concluded that the only effective form of treatment is the vigorous and rapid removal of paraquat from the gut using Fuller's Earth and purgation or gut lavage. There is little clinical evidence to suggest that measures designed to remove paraquat from the circulation after absorption or block its action in the body have any effect on the clinical course of poisoning. Some general conclusions are drawn and a select bibliography is appended.