Coinfection of a cell culture with a human and avian influenza A virus had yielded a recombinant virus with high neurovirulence for mice. This study reports on the comparative pathogenesis of central nervous system infection in mice between the parental human and the recombinant virus using the immunohistologic peroxidase-antiperoxidase method and virus assay of tissue suspensions. The human virus replicated poorly in mice and did not replicate in the brain even after intracerebral inoculation. In contrast, the recombinant virus replicated to high titer in the lung and brain with resulting viremia after inoculation of young mice by the intracerebral, intraperitoneal, or intranasal routes. Different populations of cells in the brain became infected after inoculation by each of the three routes: choroid plexus, and ependymal and subependymal cells after intracerebral inoculation; cells in perivenous areas, neurons in the olfactory bulbs and trigeminal ganglia and nuclear groups in the brainstem and midbrain after intranasal inoculation. Intraperitoneal inoculation resulted almost exclusively in the perivenous spread of the virus. The intranasal inoculation suggested that virus entry into the brain both by spread along nerve cell processes from the nasal mucosa to the brain and trigeminal ganglia and subsequent perivenous spread after viremia developed following virus replication in the lung. To dissect these two mechanisms we inoculated neonatal mice that had acquired high levels of serum antibody by nursing from actively immunized mothers. Intraperitoneal inoculation of these mice failed to cause infection, whereas intranasal inoculation resulted in the same pattern of cellular spread through the olfactory and trigeminal pathways as noted previously. This proved that this recombinant influenza virus could invade the central nervous system after infection via a natural route of infection. This highly neuroinvasive agent provides one example of the extent of virulence which can be acquired by recombination of apathogenic influenza viruses and raises a note of caution for adequate control of those agents generated in the laboratory.