Love in a cold climate?

There’s going to be a baby boom next month, according to many newspapers – the Daily Mail, the Daily Express, the Daily Telegraph. the Mirror, the Sun, and others.
Why? We had a cold winter and couple unable to get out of doors found alternative amusements.
The stories are based on pathetically flimsy foundations: evidence from a single town, Portsmouth, that 600 babies are expected to be born there in September, a 20 per cent increase on the normal rate. These expectations arose from the number of 12-week scans carried out in March and April.
The seasonality of births has often been studied, but never fully explained. In Europe, births peak in spring and early summer, with a second peak in September. In the US, there is a trough in March and April, with a peak in September. Why this different pattern exists and why it is so persistent is puzzling.
The peaks are significant, ranging in England & Wales from about 4 per cent above the average month in June, and in September/October, to as much as 7 per cent below average in February (which is a short month, of course). In 2004, for example, there were 639,721 live births in England & Wales, an average of 53,310 a month. But in June there were 55,6129, 4.3 per cent above the average, and both September and October were nearly as high (55,556 and 55,368). February’s babies totalled only 49,350, 7 per cent below average.
So a rise in births in September and October is normal. The September peak is most easily explained by the Christmas holiday, which provides a relaxed atmosphere and nothing much else to do. If the gestation period is 38 weeks from conception (the measure used for IVF births where the moment of conception is precisely known) and the Christmas holiday covers week 52 and some or all of week 1 of the following year, the peak would be expected between 20 September and 3 October. There may be an element of planning involved, too, as would-be parents also have a preference for summer births.
Other data confirm a rise in sexual activity over Christmas. Condom sales peak then, while terminations are highest in the first quarter of the year. There is some evidence that diagnoses of new sexually transmitted infections also rise in the first quarter. But the issue is not whether Christmas leads to an increase in births, but whether the effect is increased by a cold or snowy Christmas. Here there is little supporting evidence.
The most powerful counter-argument is that a September peak in births occurs in almost every Christian country in either hemisphere, according to David Lam and Jeffery Miron, two economists from the University of Michigan who published extensively on the subject in the 1990s. So the September peak is linked to the holiday, but not to the temperature prevailing during the holiday.
They did conclude, though, that high summer temperatures in the southern US depressed conceptions in July and August. My, it’s just too hot ... But this doesn’t apply in Canada, where summers are cooler.
As for September 2010, what can we expect? A small rise in births is normal, but a 20 per cent rise across the country, as confidently expected by Portsmouth midwives, would be staggering. The cause in Portsmouth must be simply a case of clustering, which will most likely be followed by a dip in successive months.
It is possible to compare December temperatures, as recorded by the Central England Temperature series, with births the following September. The tables below compare eight Decembers with unusually cold average temperatures with eight that were unseasonably warm. There does not seem to be any effect of temperature in this snapshot comparison: but perhaps focussing only on Christmas might produce a different answer.    
One thing the tables do hint at is that the September peak may be becoming more clearly established in recent years. A more detailed study would be needed to prove this, but it is not unreasonable to suggest that the growing length of the Christmas break has more to do with this than the temperatures prevailing during it.