The primary aim of this thesis is to identify a coherent legal response to the particular causal problem of the 'evidentiary gap'. In order to do this, it is necessary to understand how the 'evidentiary gap' relates to causation in negligence more generally, so the thesis addresses both the nature and function of the tort of negligence as well as the role played by causation within that tort. It argues that negligence is best understood as a system of corrective justice-based interpersonal responsibility. In this account, causation has a vital role so the test of causation must be philosophically sound. Causation, however, also occupies only a limited role so analysis must draw fully on the doctrines of damage and breach which bracket the causation inquiry, as well as notions of quantification of loss. The NESS test for causation is shown to be preferable to the but-for test because it is conceptually more adequate and therefore able to address causal problems that the but-for test cannot. This thesis rejects claims for proportionate recovery based on the notion of loss of a chance of avoiding physical harm in medical negligence, but proposes limited recovery for loss of a chance as an independent form of damage arising because of unique considerations of interpersonal responsibility in the doctor-patient relationship in cases of misdiagnosis/ mistreatment of existing illness. It is argued that the Fairchild test of material contribution to risk of harm in cases of evidentiary gap is not consistent with corrective justice, and that this cannot be resolved by reconceptualising the gist of the action as the risk of harm. The Fairchild exception lacks coherence because of its instrumentalist basis, so should not be applied outside of the mesothelioma context.