It has long been known that total complement activity is absent in normal cerebrospinal fluid (CSF). Lewandowsky (1900) was the first to call attention to the fact and this finding has been confirmed by other investigators. There is only the discrepant article of Heidelberger & Muller (1949) whose peculiar results may be questioned, because what they observed in their experiments was the activating effect of CSF over serum complement, never an increase of the amount of complement itself. In 1922, Banchieri, Göckel and Kafka demonstrated that the first component of complement (mid-piece) is always present in normal CSF and the studies in this line of investigation have progressed and other complement components were shown to be present in it. Kuwert et al. (1964) have demonstrated the presence of the first component in all CSF specimens, the fourth and the second in the majority of the specimens, and the third component in approximately 30 per cent of the cases. However, in spite of the presence of the four components in many CSF specimens, these authors were not able to demonstrate total complement activity in any case. This puzzling abnormality led Kuwert et al. to assume that normal CSF complement system is defective. Considering this hard problem carefully, the idea occurred to us that the low protein content would account for absence of total complement activity in normal CSF. In order to investigate this hypothesis, tests for complement activity were performed in a series of 108 normal CSF specimens. These specimens were routinely examined for cell count, proteins, sugar and chloride contents, globulins tests, complement fixation tests for syphilis and cysticercosis and all specimens were within normal values. Specimens containing blood were discarded. All specimens were concentrated in order to get a 20 times enrichment of the CSF proteins, according to Mies method of dialysis in collodion bag under negative pressure of 400 mm Hg. The tests for complement activity were performed as soon as possible (usually within four hours) after the withdrawal of CSF from the patients. For the demonstration of total complement activity the quantitative method of Wadsworth et al. (1931), based on photometric determination of the 50 per cent hemolytic endpoint was used. The titer is indicated by number of complement units in 1 ml of concentrated CSF. The results of this investigation have demonstrated total complement activity in 106 normal CSF specimens. Up to now, to the best of our knowledge this technique of investigation and the demonstration of total complement activity in normal CSF have not been previously described in the literature. The following conclusions can be drawn from this study: 1 — when concentrated 20 times, normal cisternal cerebrospinal fluid showed the presence of total complement activity in 98 per cent of the cases; 2 — the complement titer ranged from 0 to 23 units in 1 ml of the concentrated fluid; 3 — the statement that normal cerebrospinal fluid complement system is defective must be reconsidered; 4 — the results of this investigation open a new field of research in cerebrospinal fluid physiology and pathophysiology.