Authors' reply

We are pleased that Cornéllisen & Halberg find that our recent association of month of birth and suicide complements their earlier findings in Minnesota. However, they regarded the use of stacked data, at the outset of analysis, as monthly counts as a limitation which they claim can be remedied by focus upon broader chronomes.

In their analysis of suicides in Minnesota from 1968 to 2002 they claimed to have confirmed the concept of transyears, both a near-transyear and a far-transyear. They also claimed to have found the 1-yearly component to be bigger in the longer dataset (Halberg et al, 2005). They also reported a 20-year cycle in Minnesota suicides, which is not dissimilar to what they believe exists in many other phenomena of psychiatric interest such as religiosity, wars and crime. Halberg et al (2005) stated that the ‘photic and thermic calendar year which have been the main focus in suicide research, should now be extended to include not just the effect of seasons but magnetoperiodisms, including the newly discovered near-transyear.’

Cornéllisen & Halberg are therefore interested in looking at the data from England and Wales to investigate not only whether transyears can be aligned with calendar-yearly components, but also whether during the span examined the calendar year or the transyear is larger in amplitude. Interestingly they have made almost identical comments regarding another study on autism (Bolton et al, 1992). We are not sure whether they had access to the unstacked data for autism and month of birth, and if so what was the outcome of their analysis?

Providing that we have definitive evidence to substantiate the above claims, we agree that it would be most interesting to compare, before stacking, the chronomes of suicides at birth and death on the same population.