A reanalysis of 4-week data from a recently published study (Feighner et al. 1979) clearly demonstrate that in a significant subgroup of highly depressed patients who also suffer from at least moderate anxiety, significant treatment differences in favor of Limbitrol over AT were present not only after 1 or 2 weeks, but also after 4 weeks of treatment. These differences were not present in the less sick patients, a population in which AT was slightly favored. These findings suggest that not only is Limbitrol more effective than AT alone for a heterogeneous group of depressed patients with some concomitant anxiety during the early phases of treatment, but remains more effective that AT for a significant subgroup of highly depressed patients after 4 weeks of treatment.