Yip et al (Reference Yip, Chao and Chiu2000) reported that the seasonal variation of suicides in England and Wales had declined in recent decades. This does not appear to be the case in the USA. Monthly data for the number of suicides were available for the period 1960-1998 (National Center for Vital Statistics, 1960-1999; additional data J. McIntosh, personal communication, 2001). Chi-squared tests were conducted for each year for the monthly distribution (correcting for the number of days in each month), and contingency coefficients calculated in order to control for the number of suicides each year. There was no linear trend in the contingency coefficients over the 39-year period (Pearson r=0.033). The month with the peak number of suicides varied over the 39-year period, but not in any linear fashion (see Table 1 for a summary by decade).

Table 1 Summary by decade of suicides in USA

	Mean number of suicides per day
	1960s	1970s	1980s	1990s
January	54.59	69.05	78.08	84.12
February	55.61	70.22	79.92	84.24
March	58.75	72.93	83.471	86.55
April	60.621	74.681	83.18	86.80
May	59.69	74.52	81.47	86.27
June	57.26	72.54	81.02	87.561
July	56.36	70.56	81.54	86.33
August	56.76	71.75	81.52	86.34
September	56.47	72.22	79.51	85.35
October	56.47	71.31	77.56	82.80
November	55.33	69.40	78.11	80.86
December	53.722	66.652	74.142	76.812