Bordetella bronchi.septica vaccines, w]?ich were killecl by formalin, merthiolateor heat and ennulsified with Freunds incomplete adjuvant, were examined on their protec-tive effects in guinea pigs. Restarts indicated that merthiolate-killed vaccine was mosteffective anuong the 3 vaccines and alntost all guinea pigs receivccd 10 -20 mg of the vaccinewere protected from the establishment of carrier statc:, as well ;ts formation of pneumoniclesions, when cltallengcd intranasally with 10 cells of a virulent strain. Formalin-killedvaccine was also effective but not so potent as merthiolatc-killed vaccine. No protectiveeffect was demonstrated itt a vaccine heated at l0OC for 1 hour.Protective action afforded by mcrrtltiolate-1<illecl vaccine appe;tred 20 days after vaccitta-lion, became greatest 50 to 80 days ;tnd then decreased gradua113, but persisted for at least150 days. Serum agglutinins appeav-cd earlier than the developmcrnt of protective action andremained at high Liters for at least 180 days. Although there was a general correlationbetween protective potencies and serum aggltatinin Liters, discrepancies were observed insome guinea pigs. Therefore, agglutinin production was not always accepted as evidenceof protective potency, Remarkable protectixe effects were demonstrated in all virulentt strains of B. broncltisep-tica originated from gtuinea pigs. I-Iowever, no protective effects was shown in an avirulentvariant derived from a virulent strain by subculturing 20 times on MacConkey agar.Offspring delivered from mothe-cs in good immunity showed evident protection againstchallenge inoculation and higla agglutinin liters for 4 weeks after birth.