Class II cavities with the cervical margin just below the C-E junction and with two cervical retention grooves were prepared in intact human premolars, 31 cavities in teeth stored frozen and 29 cavities in teeth stored in 1%
chloramine solution. The teeth from both storage conditions were assigned to two groups; in one group, the cavities were cleaned with a water spray, in the other group, with a cleaner containing 0.2%
EDTA and 1% NaF. All cavities were lined with
Vitrebond and restored with P50. Cervical contraction gaps were disclosed by a fluorescent resin penetration technique. The length and width of each gap were analysed. A contraction gap was revealed in 17 of the 31 cavities in teeth stored frozen (54.8%) and in 22 of the 29 restorations in the teeth stored in
chloramine (75.8%). The width of the gap was significantly greater in teeth stored in
chloramine (p = 0.0004). No statistical difference in gap formation was found between cavities cleaned with water or with the
detergent containing 0.2%
EDTA and
sodium fluoride. Since the cavity cleansing procedure was not a significant factor in gap formation, apparently removing the superficial
smear layer with the cavity cleaner does not impair the bonding of
Vitrebond to the dentin. However, gap formation appeared to be significantly affected by the storage conditions of the extracted teeth. It is suggested that teeth to be used in bond strength and leakage experiments should be stored frozen rather than stored in 1%
chloramine solution.