The genetic basis of bipolar affective disorder remains opaque despite years of intensive investigation. One of the most serious difficulties for genetic research is the enormous phenotypic heterogeneity of psychiatric illnesses. As a response to this problem, geneticists have searched for alternative strategies to identify those individuals at genetic risk for developing the disorder. One approach is to use endophenotypes or intermediate traits. Gottesman and Gould (2003), in their discussion of endophenotypes, suggest five criteria that should be characteristic of a trait in order for it to qualify as an endophenotype. These five criteria are used in order to assess the viability of using measures of neuropsychological dysfunction as endophenotypes for genetic studies of bipolar disorder. A review of the literature suggests that executive dysfunction is characteristic of people with bipolar disorder in both the acute and chronic stages of the illness, that neurocognitive function is influenced by genetic factors and that neuropsychological deficits have been reported in the nonaffected relatives of bipolar probands. Nevertheless, it is unclear whether neuropsychological dysfunction co-segregates with affectively ill individuals. We conclude that the use of neurocognitive markers of bipolar illness suffers from a number of serious drawbacks but given the absence of more appropriate endophenotypes, the neuropsychological profiling of probands and their relatives may nevertheless prove to be a worthwhile exercise.