Free-ranging feral swine (Sus scrofa) are known to be present in at least 32 states of the USA and are continuously expanding their range. Infection with pseudorabies virus (PRV) occurs in feral swine and the primary route of transmission in free-living conditions seems to be venereal. Between 1995 and 1999, naturally infected feral swine and experimentally infected hybrid progeny of feral and domestic swine, were kept in isolation and evaluated for occurrence of latent PRV indigenous to feral swine in sacral and trigeminal ganglia and tonsil. Sacral ganglia were shown, by polymerase chain reaction (PCR) amplification of the thymidine kinase (TK) gene of PRV, to be the most frequent sites of latency of PRV. Nine (56%) of 16 sacral ganglia, seven (44%) of 16 trigeminal ganglia, and five (39%) of 13 tonsils from naturally infected feral swine were positive for PCR amplification of TK sequences of PRV. These tissues were negative for PRV when viral isolation was attempted in Vero cells. DNA sequencing of cloned TK fragments from the sacral ganglia of two feral swine, showed only one nucleotide difference between the two fragments and extensive sequence homology to fragment sequences from various domestic swine PRV strains from China, Northern Ireland, and the USA. The hybrid feral domestic swine, experimentally inoculated with an indigenous feral swine PRV isolate by either the genital or respiratory route, acquired the infection but showed no clinical signs of pseudorabies. Virus inoculated into either the genital or respiratory tract could, at times, be isolated from both these sites. The most common latency sites were the sacral ganglia, regardless of the route and dose of infection in these experimentally infected hybrids. Nine of 10 sacral ganglia, six of 10 trigeminal ganglia, and three of 10 tonsils were positive for PCR amplification of TK sequences. No virus was isolated from these tissues in Vero cells. The demonstration of the sacral ganglia as the most common sites of latency of pseudorabies viruses indigenous to feral swine, supports the hypothesis that these viruses are primarily transmitted venereally, and not by the respiratory route as is common in domestic swine, in which the trigeminal ganglia are the predominant sites of virus latency.