Employees of a factory producing antiknock additives for gasoline were examined psychologically after an exposure of an average of 14 years. A neurobehavioral symptom questionnaire, tests of single and choice reaction time, a cancellation task and a digit symbol test were the neurobehavioral variables. Total lead in urine and trimethyllead in urine showed different patterns of correlation with the neurobehavioral measures. Intellectual abilities (logical reasoning), age, and job years were controlled by partial correlation statistics as possible confounders. Referring to the low level of 21 micrograms lead per 100 ml blood and regarding the dose-response relations reported in the literature, the results support the hypothesis of a special neurotoxicity of the alkyllead compounds.