Bovine herpesviruses (BoHVs) types 1 (BoHV-1) and 5 (BoHV-5) are alphaherpesviruses of major importance to the bovine production chain. Such viruses are capable of establishing latent infections in neuronal tissues. Infected animals tend to develop a serological response to infection; however, such response-usually investigated by antibody assays in serum-may eventually not be detected in laboratory assays. Nevertheless, serological tests such as virus neutralization (VN) and various enzyme-linked immunosorbent assays (ELISAs) are widely employed to check individual or herd status of BoHV infections. The correlation between detection of antibodies and the presence of viral nucleic acids as indicatives of infection in infected cattle has not been deeply examined. In order to investigate such correlation, 248 bovine serum samples were tested by VN to BoHV-1 and BoHV-5, as well as in a widely employed (though not type-differential) gB ELISA (IDEXX IBR gB X2 Ab Test) in search for antibodies to BoHVs. Immediately after blood withdrawal, cattle were slaughtered and trigeminal ganglia (TG) excised for DNA extraction and viral nucleic acid detection (NAD) by nested PCR. Neutralizing antibodies to BoHV-1 and/or BoHV-5 were detected in 44.8% (111/248) of sera, whereas the gB ELISA detected antibodies in 51.2% (127/248) of the samples. However, genomes of either BoHV-1, BoHV-5, or both, were detected in TGs of 85.9% (213/248) of the animals. These findings reveal that the assays designed to detect antibodies to BoHV-1 and/or BoHV-5 employed here may fail to detect a significant number of latently infected animals (in this study, 35.7%). From such data, it is clear that antibody assays are poorly correlated with detection of viral genomes in BoHV-1 and BoHV-5-infected animals.