Objective: To investigate the effect of the size of the space between the restoration and the dentinal wall of the tooth (i.e. the dentinal portion of the gap) on the development of secondary caries. Methods: Tooth-resin-matrix composite specimens were mounted on custom-made gap-model stages. Specimens were divided into four groups (n=10). Group 1 had a uniform gap size of 30μm throughout both enamel and dentin. Group 2 had a 30μm enamel gap size with a 530μm dentinal gap. Group 3 had 525μm gaps in both enamel and dentin. Group 4 had 525μm and 1025μm gaps in enamel and dentin, respectively. Specimens were attached to plastic Petri plates, gas-sterilized and then incubated in a microbial caries model with S. mutans TH16 in (1% sucrose tryptic soy broth for 1 h, 4 times/day, and with a buffer solution for the rest of the day). After 8 days of incubation, tooth specimens were sectioned and stained with a rhodamine B solution. Digital images were taken under a confocal microscope and analyzed for lesion size at the enamel outer lesion (EOL), enamel wall lesion (EWL), dentin wall lesion next to the DEJ (DWL-A) and dentin wall lesion at 750µm from the DEJ (DWL-B). Results: No difference in EOL size was found between the groups. DWL-A and -B were larger in group 3 than groups 1and 2. Larger DWL-B was found in group 3 than group 4. Group 4 had marginally significant larger EWL than groups 1 and 2 (p=0.0652 and p=0.0648, respectively). Also, group 4 had marginally significant (p=0.0607) larger DWL-B than group 1. Conclusions: Based on the results of this study, it can be concluded that the presence of additional space at the dentinal wall area did not affect secondary caries development as long as the enamel gap was small. However, with enamel gaps of ≈500 µm, the presence of the additional gap space at the dentinal wall led to the development of smaller dentinal wall lesions at the deeper parts of the simulated cavity. Also, in uniform gaps, the size of the interface was positively correlated with size of the dentinal wall lesions.