Brodifacoum is a highly effective anticoagulant rodenticide that presents a secondary hazard to some non-target wildlife. The high acute toxicity of brodifacoum to mammals and birds, and its prolonged persistence in liver predicates secondary risk to predators and scavengers of poisoned rodents. Hence there is a need to improve ability to monitor and predict hazards of brodifacoum to non-targets, and optimise use patterns accordingly. Use of a less persistent anticoagulant rodenticide, diphacinone, is an alternative approach currently under investigation in New Zealand. This thesis describes a series of laboratory and pen studies that address information gaps relevant to the assessment of non-target hazards in continued use of brodifacoum, and of using diphacinone as an alternative. Non-lethal techniques for determining sublethal brodifacoum exposure in birds was investigated in chickens. Elevation of prothrombin time was a less reliable index than residual concentrations in tissues. Samples requiring less invasive procedures, such as dried blood spots or faeces, have potential to detect recent sublethal brodifacoum exposure and refinement of these indices could be useful in proactive monitoring of avian wildlife. Residual brodifacoum in eggs of sublethally-exposed hens raised further questions regarding wider non-target hazard and adverse effects on development of fertile eggs or chicks. A laboratory trial with rats found a positive correlation between residual brodifacoum concentrations in liver and the amount of brodifacoum ingested as bait. An estimated 14-22% of ingested brodifacoum was excreted in rat faeces in the period between ingestion of a lethal dose and death, indicating another potentially significant environmental pathway for brodifacoum transfer. In considering diphacinone as a less persistent alternative rodenticide to brodifacoum, evaluation of residual concentrations and persistence in pig tissues was required to estimate secondary hazard to human consumers and adequate with-holding periods for hunting feral pigs in areas where diphacinone was applied. A pen trial showed that domestic pigs were more susceptible to diphacinone toxicity, and thus primary poisoning risk, than previously estimated. Hepatic half-life of diphacinone in pigs was approximately 14 days, indicating reduced persistence in comparison to brodifacoum and enabling estimates of with-holding periods for hunting feral pigs from areas where diphacinone baits were applied. To investigate potential hazards of diphacinone use to invertebrates a trial using tree weta, a native New Zealand invertebrate, was undertaken. Weta readily ate diphacinone wax block baits with no mortality or weight loss evident, indicating low susceptibility. Residual whole-body diphacinone concentrations did not increase with the amount of diphacinone bait eaten. A simple, deterministic risk assessment suggested that, as a single secondary exposure, the maximum diphacinone concentration measured in weta would present a low risk to non-target birds. Given international recognition of the high secondary hazard and corresponding restrictions on use of brodifacoum, continued availability of brodifacoum to non-licensed users and sustained field applications for possum and rodent control in New Zealand is an exceptional use pattern. New data in this thesis suggest that baiting strategies that minimise the amount of brodifacoum available in the environment are important and regulatory review of some New Zealand brodifacoum applications should address this. In parallel, development of diphacinone as an alternative to brodifacoum should continue, as new data here confirms lower persistence in mammalian liver than brodifacoum, and also indicates low toxicity to invertebrates. However further investigation of multiple-exposure hazard and potential sublethal effects of diphacinone on non-target mammals and birds is warranted before extensive and sustained field applications of diphacinone are undertaken.