Objective To determine the durability and effectiveness of a black copper cement (BCC) and a conventional glass ionomer cement (GIC) when used to restore primary molars following partial caries removal (PCR) and to compare these results with conventional cavity preparation and restoration. Design Split-mouth randomised controlled clinical trial. Setting Department of Paediatric Dentistry, Dundee Dental Hospital, Dundee, 1998–1999. Subjects Patients with previously unrestored, matched carious cavities in non-pulpally involved primary molars. Interventions Three treatment groups: (1) Partial caries removal followed by lining with BCC and restoration with GIC (PCR:BCC); (2) Partial caries removal and restoration with GIC alone (PCR:GIC), and (3) Complete caries removal and conventional restoration (CR). Restoration durability and effectiveness was assessed both clinically and radiographically over 24 months. Main outcome measures Median survival time (MST) of restorations. Results Forty-four patients (F: 31; M: 13), mean age 6.8 years (range: 3.7–9.5), had 120 restorations placed (PCR:GIC: 43; CR: 41; PCR:BCC: 36). Eighty-six molars (29 patients) (PCR:GIC: 30; CR: 29; PCR:BCC: 27) were reviewed at 24 months. The median survival times (MST) with 25% and 75% quartiles in parenthesis were as follows: PCR:BCC, MST = 24 months (6, 24); PCR:GIC, MST = 24 months (24, 24) and CR, MST = 24 months (24, 24). The MST for PCR:BCC restorations was significantly less than for PCR:GIC and CR restorations (W = 1163.5, P = 0.028 and W = 1081.0, P = 0.004 respectively). Conclusion There were no differences in the proportions of restorations lost between restoration types, although PCR:BCC restorations did have significantly more abscess/sinus formation over the 24-month study period.