The presence of Aujeszky's disease virus in cerebrospinal fluid of experimentally infected pigs was studied using the techniques of virus isolation and PCR. Pigs, some of which were previously vaccinated against Aujeszky's disease, were inoculated with different doses of the Aujeszky's disease NIA-3 strain. At the time of death or sacrifice, a sample of cerebrospinal fluid was taken and tested for the presence of virus using the mentioned techniques. Virus was isolated only from one sample, while it was detected by PCR in most of them. The higher sensitivity of the PCR technique and the possible presence of antiviral antibodies in the cerebrospinal fluid are reasons that can be argued to explain this fact. By PCR, the virus was detected more efficiently when digested cerebrospinal fluid cells were used as DNA source than when using whole cerebrospinal fluid, suggesting that the virus could be cell-associated. Aujeszky's disease virus could not be detected by PCR in pigs which survived the acute phase of the infection and were euthanased at 8 weeks post-inoculation, when they were latently infected. This indicated that the cerebrospinal fluid is not an adequate sample for the diagnosis of latency. Since Aujeszky's disease virus was detected from most of the tested samples, we believe that this could be an adequate procedure for the quick diagnosis of Aujeszky's disease.