In this study, we have compared two in vivo assay methods to measure the type A botulinum toxin neutralising activity of specific immunoglobulin G (IgG) and its fragments (F(ab')(2), Fab', Fab) purified from pentavalent botulinum antisera raised in goats. Each assay method was repeated on three separate occasions in mice and relative potencies calculated with respect to a type A equine reference antitoxin. The conventional assay, which measures the number of mice surviving typically after 72 or 96 h following the intraperitoneal administration of a mixture of toxin and antitoxin, gave the following order of potency IgG>F(ab')(2)>Fab'>Fab (6.8>4.7>3.5>2.6 IU/mg). Differences in potency are likely to be due to differences in the pharmacokinetics of the antitoxins, which are related to their molecular weight. The alternative local flaccid paralysis assay, where toxin and antitoxin are injected subcutaneously into the left inguinocrural region, gave results with a narrower range of activities: IgG>Fab'>F(ab')(2)>Fab (6.0>5.9>5.5>4.6 IU/mg). Comparison of the two assay methods showed no significant differences for IgG, F(ab')(2) or Fab', although the Fab fragment was significantly more potent in the non-lethal assay probably because of the reduced influence of antitoxin pharmacokinetics in this localised assay. These findings show that a local flaccid paralysis assay provides a less time consuming and more humane alternative to the lethal assay for the potency testing of botulinum IgG and F(ab')(2) antitoxins.