Summary: In this study, we measured the relationship between plasma and cerebrospinal fluid (CSF) apomorphine levels and their clinical effects in two patients with Parkinson's disease (PD). After subcutaneous injection of apomorphine, serial samples of plasma and lumbal CSF were taken and serial scoring of motor responses was done using the Webster Rating Scale. The ratio of the highest level of apomorphine in CSF and plasma was 0.036 for patient A and 0.025 for patient B. The time lag between the highest level of apomorphine in plasma and CSF was 20 min for patient A and 10 min for patient B. Plasma levels of apomorphine correlated weakly with clinical motor responses. However, we could establish a highly strong correlation between apomorphine CSF levels and clinical motor responses: 0.93 and 0.89 for patients A and B, respectively. We conclude that a two-compartment pharmacokinetic model explains the clinical effects of apomorphine better than does a one-compartment model. In a two-compartment model, clinical effect can clearly be correlated to apomorphine levels in the central compartment.