Introduction: The aim of this prospective longitudinal clinical study was to evaluate the short-term and long-term changes in dental-arch dimensions in patients treated with either an acrylic splint rapid maxillary expander alone (RME-only) or a rapid maxillary expander combined with a mandibular removable Schwarz plate (RME-Sz) in the early mixed dentition, followed later by fixed appliances in the permanent dentition. Methods: The dental casts of 27 RME-only patients were compared with those of 23 RME-Sz patients and 16 untreated controls (CTRL) with constricted maxillary arches at 4 times: pretreatment (T1), after expansion but before fixed appliance therapy (T2), after fixed appliance therapy (T3), and at long-term observation (T4). The mean ages for the treated groups were approximately 9 years at T1, 12 years at T2, 14 years at T3, and 20 years at T4. Arch width, arch depth, arch perimeter, and molar angulation were assessed in all subjects at all observation times. T1-T2, T2-T3, T3-T4, and T1-T4 changes were compared statistically in the treated groups with respect to the CTRL. Results: Treatment with an RME-only or an RME-Sz followed by fixed appliances produced significant short-term and long-term increases in maxillary arch widths compared with the CTRL. The RME-Sz led to significantly more favorable results than the RME-only protocol: (1) significantly greater increases in the transverse width of the mandibular arch and mandibular arch perimeter in the long term, and (2) uprighting of the mandibular posterior teeth buccally, thus allowing for an amount of maxillary expansion that was clinically effective for the correction of moderate tooth size-arch size discrepancies. In the overall observation interval, the significant increases in maxillary and mandibular arch perimeters in the RME-Sz group were 3.8 and 3.7 mm, respectively, when compared with the CTRL. The RME-only protocol produced modest long-term increases in maxillary arch perimeter (2.6 mm); the average long-term increase in mandibular arch perimeter (2.0 mm) in the RME-only group was not statistically significant. Conclusions: The RME-Sz led to significantly more favorable results than the RME-only protocol.