Accelerated life tests on avionic PWAs (printed wiring assemblies) showed no relationship between measles and premature electrical failures. For the study, test samples with measles and crazes bridging up to 100% of the distance between adjacent conductors were exposed to different combinations of thermal cycling, humidity and DC voltage bias. The exposures were designed to accelerate electrochemical failure processes such as CAF (conductive anodicfilament) growth between conductors. The study was conducted in two phases. In Phase I, measled and non‐measled sites on six modules (including both epoxy/fibreglass and polyimide/fibreglass PWBs were tested. Bias voltages of 10,100 and 200 volts DC were applied during 10‐day tests and currents were monitored to detect the appearance of electrical shorts between individual conductor pads. Microscopic examination revealed no detectable change in the size or appearance of the measles during the exposure. Currents measured during the exposure were generally in the nanoamp range. They were similar for both measled and control sites and all sites passed the minimum criterion for insulation resistance of 500 megohms. Two of the modules were then tested at 320 and 500 volts. Short circuit failures were produced, but metallographic examination revealed them to be artifacts associated with reworked solder pads. There was no correlation between the incidence of failure and the presence of measles. In Phase II, test sites on two new polyimide/glass modules were exposed at voltages up to 500 V for a series of 5‐ or 10‐day periods, interspersed with precise measurements of insulation resistance under ambient conditions. Although no failures were produced in these tests, the ambient measurements showed some loss in insulation resistance with exposure time. Further testing revealed that the loss of insulation resistance was reversible and was unrelated to the presence or absence of measles.