A critical remark on a system approximation in queueing theory

It is shown that some inequalities for the waiting times in GI/G/m are not true, which are stated by some authors as "obviously true". In the proof of a continuity theorem for the waiting time distributions in GI/G/m an error is corrected.