An exploratory analysis was carried out to verify the use of symetrical and asyimetrical distributions for describing data sets of carrot, rape and Stevia rebaudiana (Bert.) Bertoni seeds stressed during different times under the temperature of 40°C. The symetrical distributions were the Normal and the Lognormal, and the asymetrical ones were the Gumbel and the Weibull distribution. The highest numerical value of the log likelihood function, used as goodness-of-fit test, was obtained with the fit of the Gumbel distribution. The results from this work suggest that others distributions different from the Normal can be used as an alternative to describe data sets from the accelerated ageing test.