An early development vehicle experienced an unusually high rate of windshield breakage. Most breaks were identified as due to impact, but the severity of impact was low. It was reasoned that the windshield should possess a greater level of robustness to impact.Many theories were put forth to explain the breakage data. It was universally agreed that the unusual breakage rate could be due to only one condition, but its source was indefinite. The condition present must be tensile stress. One of three situations were considered regarding its source: 1) the tensile stress was present in the glass after manufacture due to improper annealing; 2) the installation of the windshield into the vehicle body put the glass into stress; 3) some combination of the other two sources.A gray-field polariscope was used to measure the stresses of the windshield from both the manufacturing process as well as the installation in the vehicle. Readings of the manufacturing stresses were low, and compared to those of the supplier. They did not correlate with the observed breakage. Readings of the installation stresses were high enough to explain the breakage and correlated with the failure mode as observed.Based upon these results, the polariscope was further used to evaluate any proposed design changes and validate them as effective, robust measures for reducing the probability of breakage in service.