As for Laplace, the main purpose of the CLT for Poisson was to be a tool in
classical probability calculations, not so much to be a mathematical theorem in
itself. Poisson did therefore not explicitly formulate any conditions for the central
limit theorem to hold. It seems clear from his proofs and examples that he assumed
the variances of the components of the sum to be bounded so that the variance of
the sum would be of the order n. He does not say this explicitly though. He did,
however, discuss a few counterexamples where the CLT does not hold.
One example of where the CLT does not hold are so called Cauchy-distributed
variables where the probability density takes the following form: