Previous reports have established that vaccinia virus (VV) recombinants expressing G, F, or N protein of respiratory syncytial (RS) virus protect small animals against intranasal challenge with live RS virus. This work demonstrates that a variety of parameters affect the protection induced by recombinant viruses. The route of vaccination, the subtype of challenge virus, and the species used influenced the antibody titers and extent of protection. During these studies, observations were also made on the subclass of antibody generated, and pulmonary histopathological changes induced by challenge after vaccination were noted. The effect of route of inoculation on host response was examined by vaccinating mice intranasally, intraperitoneally, or by scarification with a recombinant VV expressing the RS virus G glycoprotein. Intranasal vaccination induced 25-fold-higher titers of antibody to RS virus in the lung than the intraperitoneal route did, but both routes resulted in complete suppression of virus replication after intranasal challenge 21 days after vaccination. Scarification was a less effective method of vaccination. The antibody induced by recombinant VV in mice was mostly immunoglobulin G2a (IgG2a) with some IgG2b. No antibody to RS virus was detected in the IgA, IgM, IgG1, or IgG3 subclass irrespective of the vaccination route. The G and F glycoproteins were shown to elicit similar subclasses of antibody. However, animals vaccinated with the G and F vectors differed strikingly in their response to challenge by heterologous virus. Mice or cotton rats vaccinated with recombinant VV carrying the G gene of RS virus were protected against challenge only with homologous subtype A virus. Vaccination with a recombinant VV expressing the F glycoprotein induced protection against both homologous and heterologous subtype B virus challenge. The protection induced in mice was greater than that detected in cotton rats, indicating that the host may also affect immunity. Finally, this report describes histological examination of mouse lungs after vaccination and challenge. Vaccinated mice that were subsequently challenged had significantly greater lung lesion scores than unvaccinated challenged mice. The lesions were primarily peribronchiolar and perivascular infiltrations of polymorphonuclear cells and lymphocytes. Further work will establish whether these pulmonary changes are a desirable immune response to virus invasion or a potential immunopathogenic hazard. The results have important implications for planning a strategy of vaccination against RS virus and emphasize potential dangers that may attend the use of recombinant VV as vaccines.