Epidemiologic and clinicopathologic data from 11 macaques with naturally acquired SIV infection--10 of which have died--were compared with those from 34 rhesus monkeys that have died of experimental SIVmac infection. Several differences, including gender affected, age at time of death, and the occurrence of certain opportunistic infections, could be explained by the experimental design; others remained unexplained. The most striking difference was the 41% incidence of meningoencephalomyelitis in the experimental group and its absence in naturally SIV-infected animals.