Objective: The purpose of this study was to assess, via a photoelastic resin, the expansion of gallium restorative alloys under conditions similar to those found in the clinical situation. Method and materials: Two gallium alloys, Galloy and Gallium GF II, were tested, along with a high-copper amalgam, Dispersalloy, and a low-copper alloy, New True Dentalloy. The gallium alloys were tested as (1) uncontaminated, (2) contaminated with water, Ringer's solution, or a cell culture medium, and (3) immersed in these fluids at times ranging from 5 minutes to 3 days. The gallium and amalgam alloys were condensed in a hole drilled in a block of photoelastic resin and observed for 3 months. The amount of stress was recorded on color slides taken through polarized light at regular intervals. The photographs of the color bifringen stress patterns at 1 and 3 months were ranked by two independent evaluators for least to greatest observed stress. Results: Disperalloy had the least expansion, followed by uncontaminated Galloy and Galloy contaminated with water. Next came Galloy contaminated with cell culture medium, Galloy contaminated with Ringer's solution, contaminated New True Dentalloy, and Gallium GF II. The last group was Gallium GF II contaminated with any of the three solutions. Both gallium alloys immersed in the three fluids showed a strong edge effect, and by 6 weeks many of the gallium alloys had extruded from the mold. Conclusion: These results corroborate the findings of some clinical studies that have shown that these gallium alloys can potentially cause catastrophic failures.