While carrying out the fatigue tests we inevitably get censored specimens, that is still unbroken after running out the base cycles (2.10^6, for example). Those specimens bear important information, but so far we do not use it while building the fatigue curve. What could you propose?