Cows in a 250-cow Holstein-Friesian herd were allocated at random to be treated with either a homoeopathic nosode or a negative control, both treatments being applied by means of an aerosol spray to the vulval mucous membranes. A total of six treatments were given over a period of three days and milk samples were taken for the determination of somatic cell counts (SCC) on days -3, 3, 7, 9, 14, 21 and 28. Individuals applying the treatments or carrying out the SCC determination were unaware of which animals were receiving which treatment. Owing to the wide natural variations in SCC, the trial had only a 71 per cent possibility of detecting a 30 per cent difference in SCC between the two groups. There were no significant differences between the SCC of the two groups on any sample day, but there were significant variations between the SCC on different days (P=0.003) in both groups.