Latitude and the timing of childbirth

Background

Last month, I was introduced to an intriguing graph, which is reproduced below:

The source of the graph is a Business Insider article from 2014, which is reporting on a study published in the Proceedings of the Royal Society. The graph shows an interesting relationship between the peak birth month of a nation and the latitude of the nation.

My immediate reaction on seeing the graph was to consider the timing of conception, rather than birth. Based on the graph above and some simple mental calculation, it appears that conception in the far north generally peaks around September, and in the tropics it generally peaks around January. This makes the pattern seem a little less mysterious to me, given the way average temperature, and the length of daylight hours, depend on latitude.

Peak conception time by daylight hours

Based on the supporting information provided with the study mentioned above, the new graph below shows the relationship between the daily number of hours of sunlight at the time of peak conception for a variety of nations and the latitudes of the nations.

The dashed curves in the graph represent the minimum and maximum daily hours of sunlight possible for each latitude. We can see that in many nations, conceptions peak at times of the year when daylight hours are at their shortest — or in other words, when nights are longest. At more northern latitudes, there is more variation in this pattern; however, the vast majority of peaks still occur when daylight hours are on the shorter side of their range. This is somewhat more pronounced if we limit our attention to the peaks that are marked as “significant” in the supporting document, as shown here:

Here is a histogram of daylight hours for all of the significant peaks:

The vast majority of peak conception times occur when there is just a bit less than 12 hours of light in a day. Half of them (61 out of 132) occur when the number of hours of daylight is between 11 and 12.

Technical details

Unlike my usual practice of using Python or R, everything here was produced in Microsoft Excel, using the following main steps:

I exported the table of data available at the end of this PDF file to Excel.
I calculated a new column of peak conception months by subtracting 9 from the “mean peak [birth] month” column, modulo 12.
I calculated a column of estimated daylight hours at conception with the help of the “sunrise equation” described here.
I also calculated the minimum and maximum daily hours of sunlight possible for each latitude by the “sunrise equation,” using the minimum and maximum possible declination of the sun.
I graphed the results, clearly.

If you want to take a look at my Excel workbook, feel free to get in touch.

January 14 note: I have made some large corrections since the first version of this post.