The type A neurotoxin produced by Clostridium botulinum is a potent neuromuscular blocking agent which causes paralysis by preventing the release of neurotransmitter from motor neurones. This property has led to the use of the toxin in the treatment of a number of neuromuscular diseases involving muscle spasms. At present, the only recognised assay with the specificity and sensitivity to estimate accurately the potency of botulinum toxin in clinical preparations is bioassay, in which lethality is used as the end point. Refinement of this assay, with respect to the end point, was explored on the basis of the development of flaccid paralysis of muscles following subcutaneous injection of the toxin at the inguinocrural region. Potency estimates, relative to in house reference preparations, for different therapeutic preparations obtained using flaccid paralysis as a scored response gave excellent agreement with estimates obtained in independent assay using the currently required control method. This study demonstrates that an alternative, more humane bioassay for potency testing of clostridia neurotoxins gives valid estimates equivalent to those currently in use.