Some probability distributions have infinite sets of outcomes and infinite variance, and in these cases, even if independence and identical distribution hold, the sum may not converge to a normal distribution.