The approach for Zn and Co herd status was not valid, because all herds were finally classified as adequate ‘Zn’ and ‘Co’ herds, and these were consequently dropped from the analysis. In all statistical analyses, two-way interactions were investigated once a main-effects model had been achieved