This double-blind cross-over study was undertaken to assess the effect of systemically administered metronidazole when used as an adjunct to periodontal surgery for the treatment of moderate and advanced periodontitis. The effect of metronidazole was compared with that of placebo in patients undergoing modified Widman flap procedures in two areas of the same jaw which could be matched for type of tooth and severity of the periodontal disease. Clinical and microbiological parameters were examined prior to surgery and then 7 days, 1 month, and 3 to 6 months, postoperatively. The clinical parameters recorded were pocket depth (PD), Sulcus Bleeding Index (SBI), probing attachment level (PAL), and patients' preference and pain score. Subgingival plaque samples were studied with dark-field microscopy for differential bacterial count. Pocket depths and SBIs were reduced significantly at all stages, in both groups. Probing attachment levels increased at 7 days, to significant levels only in the metronidazole group, subsequently PALs decreased in both groups with no significant differences between the groups. Although the differential bacterial count altered markedly in both groups at all times, only the straight rod count at 1 month was significantly (P less than 0.05) lower in the metronidazole group. Metronidazole with surgery did not exert a significantly greater beneficial effect than placebo with surgery.