Abstract
The axial-vector-current divergences do not transform under SU(3) as they naively appear to transform. The consequences of this observation are drawn, correcting and extending an earlier paper of the author.