The possibility that there might be a neurohumoral cerebrospinal fluid (CSF) link in motion sickness was directly tested in cats by blocking the flow of CSF from the third into the fourth ventricle. Evidence obtained thus far is consistent with the hypothesis. Cats with demonstrably sound blocks did not vomit in response to an accelerative motion sickness stimulus, whereas cats with imperfect ''leaky'' blocks vomited with little or no increase in latency. Although there are several putative candidates, the identification of a humoral motion sickness substance is a matter of conjecture.