N-acetyl-β-D-glucosaminidase (NAGase) levels and somatic cell counts (SCC) were determined monthly for 6 months in the bulk milk of 181 suppliers (1063 samples). A highly significant correlation (r = 0·74; P < 0·001) was found between supplier's bulk herd milk geometric mean NAGase activity and SCC. Monthly trends which grouped suppliers into various categories defined by different NAGase and SCC thresholds showed that a similar overall pattern was obtained with both SCC and NAGase. However, further analysis indicated that 18% of the bulk milk which had SCC < 500 × 103/ml had NAGase levels > 25 units. Also, 34% of samples with SCC > 500 × 103/ml had NAGase levels < 25 units. Overall, 24% of all samples did not have corresponding NAGase and SCC results. When the bulk milk of 2 commercial dairy herds was tested monthly over 12–18 months whilst the infection status of all quarters in the herds was regularly monitored, those herds with low incidence of mastitis (5% quarters infected) had significantly lower bulk milk SCC and NAGase levels than those with a high incidence of mastitis (22% of quarters infected). This suggests that NAGase measurement on bulk herd milk would be a simple means of monitoring infection status of dairy herds and of rapidly classifying a supplier's milk as being of low, medium or high SCC status. The possible combined use of SCC and NAGase levels in bulk milk monitoring schemes is discussed.