Purpose: To determine whether the frequency of microleakage at the margin of the tooth-restoration interface in primary molars restored with high-viscosity glass-ionomer cement (HVGIC) differs from that of primary molars restored with amalgam. Materials and Methods: The HVGIC restorations were performed according to the ART method. A total of 19 naturally exfoliated primary molars (10 amalgam and 9 ART/HVGIC) with clinically assessed intact restorations fulfilled the inclusion criteria. SEM analyses using replicas were performed to assess the microgap presence and size of the tooth-restoration interface The depth of a microgap was measured from histological sections after infiltration of 50% (w/v) AgNO3 solution. Independent variables were restorative material, age of restoration, and number of tooth surfaces. Data analyses included Fisher's Exact version of the chi-squared test, independent t-test and one-way ANOVA. Results: More ART/HVGIC than amalgam-restored teeth were gap free (p = 0.14). The mean gap size for teeth with a gap between amalgam (322.0 µm) and ART/HVGIC (201.0 µm) restorations did not differ statistically significantly. AgNO3 infiltration was not influenced by restorative material or age of restoration. Infiltration in multiple-surface was higher than in single-surface restorations (p = 0.02). Conclusion: The occurrence of marginal microleakage in primary molars restored by ART/HVGIC is no different from that observed in primary molars restored with amalgam. Microleakage was deeper in multiple-surface than in single-surface restorations. Despite the presence of microleakage, restoration functionality up to exfoliation is secured, reinforcing the importance of biofilm removal and behavioral education of patients. Keywords: dental caries, primary teeth, amalgam restorations, atraumatic restorative treatment, marginal microleakage, microgap, silver nitrate infiltration, scanning electron microscopy