It is seriously debated whether the presence of immunoglobulins in CSF is due to local production, diffusion of proteins through the blood-brain barrier, or both. Tourtellotte et al [1] strongly suggest that in both healthy and diseased individuals, immunoglobulins are synthesized extravascularly and subsequently diffuse into the CSF. Cohen and Bannister [2] demonstrated that lymphocytes from the CSF of patients with multiple sclerosis could produce IgA and IgG in vitro. Although such evidence suggests the likelihood of local production of immunoglobulins, others believe that elevated protein levels in CSF are due to migration of immunoglobulins from serum to CSF because of damage to the blood-brain barrier [3]. Thus, the elevated IgG level in CSF could be an expression of such an impaired blood-brain barrier. A further implication of this hypothesis is that the more the function of this barrier is impaired, the more extensive is the transudation. This includes the migration of large molecules such as IgM during the inflammatory state, a condition which increases not only the number of theoretical filters but also the size of the pores. IgG levels in CSF and serum were elevated in all five infections studied except viral encephalitis, in which this value remained normal. The serum IgA level was elevated in all five lesions, but the IgA level in CSF was elevated only in viral encephalitis, Guillain-Barré syndrome, and meningeal carcinomatosis.(ABSTRACT TRUNCATED AT 250 WORDS)