In our effort to protect the human subject of medical research we have failed to ensure that the subject will be treated justly, specifically, we have failed to ensure that the subject will not be exploited. I argue that we fail in this regard because we do not understand the concept of exploitation well enough to identify and address all instances of the problem. This dissertation builds a model of exploitation, complete with a principle of justice in transactions, and applies this model to human subjects research. The model of exploitation shows that A exploits B just in case A gains through his transaction with B, and A's gain is unjust with respect to B. Drawing from both economic theory and social and political philosophy I show that A's gain is unjust with respect to B just in case A's gain is larger than the fair market price and A can make no entitlement claim to the difference. Applying this model to human subjects research shows that whether a subject is exploited depends on a number of variables, including what the subject can offer to the researcher, what the subject seeks to gain through his participation, and the alternative ways in which the subject can pursue his desired goods. Each variable can be instantiated in a number of ways, and various combinations of these instantiations can characterize several different subject populations. For each population we can construct a "fair research protocol" and determine whether a given clinical trial is exploitative to subjects recruited from this population. Using this method I show that at least three current research practices, namely, offering money for participation according to an expense-reimbursement model, recruiting from the clinic, and advertising clinical trials as treatment options, are exploitative to at least some patient populations. I also show that these cases of exploitation, and all other cases of exploitation in human subjects research, can be prevented in three ways: first, by modifying protocols to exclude exploitative elements; second, by changing recruiting methods to exclude subject populations vulnerable to exploitation; and, finally, by prohibiting the research.