The evidence for an increased incidence of rates of depressive conditions, defined as depressed mood, depressive syndromes or depressive disorders, is reviewed. Findings from prospective studies, family genetic studies, community surveys, repeated cross-sectional surveys, admission data and suicide statistics are summarized. Methodological problems and difficulties in the analysis and interpretation of these results are then summarized. Because converging results from studies conducted with various designs, sampling and measurement strategies are convergent, it is concluded that the increase over time of depressive phenomena is a plausible finding. However, its magnitude remains unknown and is probably smaller than sometimes claimed. Furthermore, studies have failed to demonstrate that the increase in rates was specific to depressive conditions. Finally, it is noted that these epidemiological enquiries have failed to provide cues on the possible mechanisms underlying these secular changes.