An analysis has been presented of the results obtained in complement fixation tests employing both the soluble, or S, and virus, or V, antigens in 85 cases of mumps meningoencephalitis. Thirty-eight of 71 cases for which clinical data were complete had failed to reveal preceding, concurrent or subsequent involvement of the salivary glands. Only 25 (35%) had such involvement prior to onset of meningitis. In 62 of the patients, definite serologic evidence of infection with mumps virus was obtained in one of two ways: (1) The presumptive test was positive in 43 cases, i.e., the first serum showed a high antibody level against the S antigen and distinctly lower titers against V. (2) In 19, a positive diagnosis was based on the demonstration of a rise in both antibodies. The remaining 23 patients gave evidence suggestive of mumps infection by virtue of high antibody levels in the first available serum. The anti-S > anti-V pattern occurs most frequently when the meningoencephalitis is not preceded by salivary gland involvement and when the serum is obtained before the end of the first week of illness. Delay in withdrawing the blood and occurrence of overt or possibly inapparent glandular involvement prior to the meningitis is associated with the presence of antibodies in relatively high titer to both the S and V antigens. Such results can only be regarded as suggestive of recent infection. Even when serum from patients with no salivary gland involvement is obtained in the first week of illness, about one third will not show the anti-S > anti-V pattern. At present, under proper conditions, the early serologic diagnosis appears possible in about two thirds of cases of mumps meningoencephalitis without preceding or concurrent salivary gland involvement. The clinical findings and routine laboratory data on spinal fluids and blood are presented; these findings and data are in accord with those previously reported by others.