Comorbidity - the clustered occurrence of two traits or disorders - may be studied in genetically informative designs such as the classical twin study, to test whether genetic and/or environmental factors underlying the two disorders are correlated. When a genetic correlation is found, this can be explained by several mechanisms, including pleiotropy (the same genes influencing multiple traits), and causality (one trait causing the other). With a cotwin control design, it can be investigated which scenario is most plausible. In this design, monozygotic twin pairs discordant for the first trait (i.e., one twin is affected, the other is not) are compared in terms of their risk for the second trait: under a causal model, only the twins affected for the first trait will be at increased risk for the second trait. Under genetic pleiotropy, this risk will be increased in both twins because they share the same risk genes. We first discuss the cotwin control design and then illustrate its application with data on migraine and neuroticism that were collected in 5,200 Dutch twins, including 1,648 complete twin pairs (981 monozygotic and 667 dizygotic pairs). There was a significant association between migraine and neuroticism, which could be attributed to genetic and environmental correlations (rG = .27 and rE = .19). In monozygotic and dizygotic twin pairs discordant for neuroticism, the risk of migraine was significantly higher in the twins with a high neuroticism score. This pattern of results is consistent with a causal relationship, suggesting that neuroticism increases the risk of migraine.