A new, investigational, parenteral form of sparfloxacin was compared with ceftriaxone in the treatment of experimental endocarditis caused by either of three penicillin-susceptible streptococci or one penicillin-resistant streptococcus. Both drugs have prolonged half-lives in serum, allowing single daily administration to humans. Sparfloxacin had relatively low MICs (0.25 to 0.5 mg/liter) for all four organisms and was also greater than or equal to eight times more effective than the other quinolones against 21 additional streptococcal isolates recovered from patients with bacteremia. Ceftriaxone MICs were 0.032 to 0.064 mg/liter for the penicillin-susceptible strains and 2 mg/liter for the resistant isolate. Both antibiotics resulted in moderate bacterial killing in vitro. Rats with catheter-induced aortic vegetations were inoculated with 10(7) CFU of the test organisms. Antibiotic treatment was started 48 h later and lasted either 3 or 5 days. The drugs were injected at doses which mimicked the kinetics in human serum produced by one intravenous injection of 400 mg of sparfloxacin (i.e., the daily dose expected to be given to human adults) and 2 g of ceftriaxone. Both antibiotics significantly decreased the bacterial densities in the vegetations. However, sparfloxacin was slower than ceftriaxone in its ability to eradicate valvular infection caused by penicillin-susceptible bacteria. While this difference was quite marked after 3 days of therapy, it tended to vanish when treatment was prolonged to 5 days. In contrast, sparfloxacin was very effective against the penicillin-resistant isolate, an organism against which ceftriaxone therapy failed in vivo. No sparfloxacin-resistant mutant was selected during therapy. Thus, in the present experimental setting, this new, investigational, parenteral form of sparfloxacin was effective against severe infections caused by both penicillin-susceptible and penicillin-resistant streptococci.