From observation of lightning activity in severe storms with VHF radar the echoes from lightning discharges following one another within a time interval typical for a multistroke cloud-to-ground flash and spaced in range by several km are designated ''associated discharges.'' The hypothesis of association between lightning flashes is tested against an alternative hypothesis of independence between them by using methods of statistical analysis for series of events. The hypothesis of independence is rejected at a level of significance of 5%. The probability for lightning flashes in multicell storms to be associated increases when time intervals between flashes are very short (<100 ms for the investigated storm).