Eighteen normal horses were assigned to 1 of 3 treatment groups to investigate the effects of IM or intrathecal (IT) administration of ovalbumin on serum and cerebrospinal fluid (CSF) antibody production. Horses of group 1 were injected intramuscularly with ovalbumin and adjuvant, while horses in treatment groups 2 and 3 received ovalbumin intrathecally or intravenously, followed by IM injection as in group 1. Serum and CSF antibody titers were tested in group I every 30 days for 4 months, while serum and CSF were collected in group 2 and 3 horses at postvaccination day 60. Horses of group 1 (IM) developed a serum antibody titer that peaked at postadministration day 60 (1:24,320 +/- 7,680) (mean +/- I SEM). Anti-ovalbumin antibodies were detected in CSF, and titers paralleled that of the serum, although at a much lower concentration (peak, 1:166 +/- 87). Horses of groups 2 and 3 developed significantly (P = .02) lower serum titers (1:720 and 1:2,067, respectively), but the difference in CSF titers did not achieve statistical significance (P = .06). The results confirm that antigen-specific antibody can be found in the CSF of horses in which antigen is not administered intrathecally. This may affect the interpretation of CSF analysis in diseases such as equine protozoal myeloencephalitis. Further, the findings suggest that IT injection of the soluble antigen ovalbumin induces a state of antigenic tolerance in the horse. The clinical significance of this finding remains unknown at this time.