OBJECTIVE To compare and evaluate the clinical failure rates of the chemically-cured composite bonding resins Sondhi Rapid Set (SD) and Maximum Cure (MC) when used in an indirect bonding technique. METHODS In vivo study in the private orthodontic practice of a solo practitioner. METHODS Forty consecutive patients meeting the inclusion criteria were assigned to alternating groups in a split-mouth study design. Group 1 had the maxillary right and mandibular left quadrants indirectly bonded using SD adhesive, while the contralateral quadrants were bonded using MC adhesive. Group 2 had the opposite sides bonded to Group 1. One patient was lost from each group. Over a 6-month observation period, all loose brackets were recorded and the data compared with a Wilcoxon sign-rank test. RESULTS Of the 363 brackets placed in each group, 36 with the SD adhesive came loose (9.9 per cent failure rate) compared with five from the MC group (1.4 per cent failure rate, P = 0.0001). In the maxillary arch, seven brackets from the SD quadrants came loose versus one for the MC (P = 0.109). In the mandibular arch 29 brackets from the SD quadrants came loose during the 6-month observation period compared with four from the MC quadrants (P = 0.001). CONCLUSIONS Both chemically-cured adhesives (SD and MC) examined in this study were suitable for the indirect bonding of brackets. The SD adhesive had seven times the number of breakages than the MC adhesive in both arches (P = 0.0001).