Yip et al (Reference Yip, Chao and Chiu2000) demonstrated that, in England, the seasonal variation in suicide rates in the 1980s and 1990s decreased considerably when compared with that in the 1960s and 1970s. From monthly suicide frequencies, they concluded that current data hardly show any seasonal effects on suicide rates, and they predicted that seasonal variation in suicide rates would disappear completely in the years to come.

Although we fully agree with Yip et al (and several other authors) that there is a global decline in the amplitude of seasonal variation in suicide rate, we do not agree with the conclusion that seasonal influences are beginning to fade away. We came to this conclusion by a recent study of train suicides (i.e. suicide by jumping before a moving train) in The Netherlands (Reference van Houwelingen and Beersmavan Houwelingen & Beersma, 2001). In this study (n=30) we confirmed the absence of a seasonal pattern in suicide rates as observed in 28-day intervals. We did, however, observe a strong seasonal influence on 24-hour patterns. Whereas the winter season showed two daily peaks in suicide rates, at around 9-11 am and 7-10 pm, the summer season revealed one major peak around 12-4 pm and a smaller peak shortly before midnight. The timing of the major summer peak is in the trough between the two winter peaks.

This more subtle influence of time of year on suicide rates adds a different dimension to what has been considered seasonality in suicidal behaviour and may generate new ideas concerning relevant factors involved. In train suicide data, seasonal influences are clearly present. This may also be true of other methods of suicide. In order to see this, time of day and time of year have to be taken into account simultaneously.