Alice Reid
For the last 11 years there have been fewer births on Boxing Day than on any other day of the year, with Christmas Day and New Year’s Day also having very low numbers. In contrast, there were more babies born on 28 September than any other day, and late September to early October has been the most popular time to be born over the last 30 years or so. The lack of births on festive season holidays is due to fewer inductions and planned caesareans over the Christmas bank holidays, while the late September peak has been attributed to Christmas and New Year conceptions.

Sir Edward Burne-Jones, The Star of Bethlehem (1887-1891). Birmingham Museums Trust.
Over the past 30 years, Christmas Day, Boxing Day, and New Years Day have also been very unpopular days to get married, although the spring and summer bank holidays have proved very popular (and the summer months in general), along with Valentines Day. The majority of marriages have taken place on Saturdays (although this has become less popular over the time period), with Friday being the next most popular day of the week.
But what of further in the past? Has Christmas always had fewer than its fair share of births and marriages, and were summer weddings and autumn births always more likely?
Changing seasonality of marriage
The graph below illustrates seasonal patterns of marriage. Red and yellow indicate more marriages than would be expected on an even spread over the year (taking account of the number of days in each month), and blue indicates fewer.
It’s clear that there have been significant changes in the popularity of different seasons for getting married since the 1540s. October and November were very popular times for marrying in the 16th century, and March the least popular, with January and July being a bit more popular than the surrounding months.
Over the following 250 years the autumn peak and avoidance of March remained, although with less prominence, and a late spring peak emerged for a while, with December also increasing in popularity. There is no monthly data available for 1837-1947, but the patterns changed dramatically during that period, with the emergence of a strong preference for summer weddings, punctuated by a period in the 1950s-60s when March was very popular.
High-days: religion and marriage practices
Under the Roman Catholic church, marriage was prohibited at certain times of the year: Lent (the 40 days leading up to Easter), Advent (roughly the four weeks leading up to Christmas), and Rogationtide (a shorter period leading up to Ascension Day). Although the Church of England, when it was established in 1534, did away with these prohibitions, old habits died hard and this can explain both the unpopularity of marriage during March and December, and the gradual phasing out of these traditions.

A Marriage Ceremony. Illustration from ‘A Book of Roxburghe Ballads’.
The unusual pattern of the 1650s (and particularly the relative rise in Lent and Advent weddings) is also linked to religion (or the lack of it), in the form of the imposition of civil marriage between 1653 and 1657. This was deeply unpopular, and although the act was only announced in August 1653, with a requirement to post banns before any wedding, many couples rushed to marry before it came into force on 30 September. Over three times as many couples married in September 1653 than in the previous September, with a corresponding drop in October and November weddings.
Surprisingly, there were more weddings during this period than the surrounding years. This may partly be due to the inclusion of non-conformist marriages during these years (which were not usually included in Church of England registers). But it has also been argued that many people married twice: once in front of a JP for the law, and once in front of a priest for their conscience.
Hiring days and seasonal labour demands
The summer trough and autumn peak in marriages can be linked to the seasonal demand in labour, and seasonality in the way that labour was hired. Over the summer months, harvesting was very demanding labour which allowed little time off. Arable areas were particularly affected by high workloads at this time of year, so summer weddings were unpopular in arable areas of south-east England. The pastoral areas in the north and west also started the period with autumn peaks, but moved towards early summer weddings.
Ann Kussmaul argues that this is a sign of early market integration in England: before this, pastoral areas could not rely just on livestock for food: they needed to grow some grain crops. With market integration, grain could be imported from other areas of England, and pastoral farms were freed from summer harvests, enabling summer weddings.

John Frederick Herring I, Harvest (1857). Yale Center for British Art.
The nature of marriage in pre-industrial England, which involved newly-weds setting up a new household, contributed to the popularity of post-harvest weddings in the autumn. The plentiful work during the harvest season will have added to the young couple’s savings (indeed some were given the whole year’s wages at that point), giving them the funds to rent and furnish a place of their own. William Farr (often credited with setting up the modern system of registration of births, marriages and deaths) wrote in 1885 that ‘The accumulations of autumn supply a store of food, and the harvest wages of young swains in agricultural districts are often wisely invested in the furniture of a cottage.’
Annual hiring fairs added to the popularity of autumn weddings. During the pre-industrial period farm workers, labourers, and servants were usually hired for a year at such events, which took place during the autumn. The end of the year’s service was a good time for a couple to marry and move into their first joint home.
Over time there was a move towards hiring farm workers by the day rather than for a full year, and this decreased motivations for an autumn wedding and dampened seasonal patterns. Although some marriage seasonality continued in agricultural areas, it became dwarfed by the increasing proportions of the population who were living in cities and therefore were unaffected by the agricultural calendar.

A farmer and his wife contemplating their full-grown crop in a field ready for harvest. Wellcome Collection.
Holidays
For urban populations, the ability to take time off has generally been the biggest influence on the seasonality of marriage. Farr noted that, in the mid-19th century, both Christmas and Whitsun (the seventh Sunday after Easter and traditionally a holiday week when fairs would often take place) were popular for weddings.
Sir William Beveridge (who later wrote the 1942 report which was influential in the founding of the British welfare state) credited the increasing popularity of summer weddings to the establishment of additional bank holidays in 1871. ‘Sir John Lubbock’, he wrote of the promotor of the Bank Holidays Act of 1871, ‘has had much to do with altering the matrimonial customs of a nation’. A preference for a summer wedding, with a higher chance of good weather and a larger number of non-workdays to choose from, remains today.
Taxation and disease
Three areas of disruption stand out from these long changes in customs. The most obvious is an 18-year period in the 1950s and 1960s, when March weddings became surprisingly popular. This is entirely due to the introduction of tax breaks for those marrying late in the financial year (which ends on 5th April). While there were also more weddings in the first few days of April than would be expected, this is dwarfed in the monthly figures by far fewer in the rest of the month.
More recently, the disruption caused by the Covid lockdowns and restrictions on public gatherings can be detected, with virtually no weddings during periods of lock-down in 2020 and 2021, and what few weddings there were concentrated into other times of the year.
Changing seasonality of births
The graph above shows that the seasonality of births has also changed dramatically over the long time period, although it was far less pronounced than for marriage (note that although the colours are similar, the births graph covers a smaller range).
In the preindustrial period there was a peak in births in February and March. This drifted later towards April by the early 19th century, becoming less pronounced as it did so. A smaller peak in the autumn died away almost completely. By the mid–20th century, a spring peak had emerged, and there was a small disturbance caused by the end of the second world war. Since the 1970s, today’s September peak is in evidence.
Births are much less amenable to planning than marriage, so it is unsurprising that seasonality is lower. But given that it is commonly accepted that people did not plan births or use contraception in the past, it is surprising that seasonality in the 16th and 17th centuries was considerably higher than that in 20th and 21st centuries. Before discussing the difference, we will first look at the pattern in the pre-industrial period.
Seasonal labour (pun intended)
The seasonality of births largely reflects the seasonality of conceptions (assuming that seasonality in miscarriages and stillbirths is relatively small). This means that in the preindustrial period there was likely to be a peak in conceptions in April-July, and a big drop between August and November. The demands of the harvest season appear to have reduced the opportunity or energy for sexual intercourse, while better weather in late spring and early summer as well as Easter and Whitsun holidays had the opposite effect.

Alfred James Munnings, Whitsuntide – A Gala Day. Harris Museum & Art Gallery.
Beveridge noted that not only are holidays a good opportunity for marriage, but they are ‘also the natural opportunity for mating without marriage’. Indeed the seasonality of births conceived outside of marriage (whether or not they were born within marriage) is very similar to the seasonality of all births shown in the graph.
From marriage to birth
The seasonality of first births conceived during marriage, however, was quite different to that for all births, and implies conception in the first few months of marriage, followed by birth in roughly the same month as marriage but a year later.
By the mid-20th century, women were having far fewer children than in the preindustrial period, meaning that nearly half of all children were first children. This means that the seasonality of first births has a larger effect on the seasonality of all births, and we can see evidence of this in an echo of the tax-induced March weddings.
The more recent September peak in births also roughly matches the August peak in marriages, suggesting a continued link with marriage. Despite the fact that more births now take place outside marriage than in, many marriages still take place when a couple feels ready to start a family. The gap between marriage and birth is slightly larger, and this could be due to a combination of longer honeymoons and a desire to delay trying to conceive until after the honeymoon.

Telemaco Signorini, The honeymoon (1862-3).
Alternative explanations for the popularity of September births include couples deliberately trying to have children who will be older in their school year, or the influence of the festive season, perhaps in a combination of Beveridge’s idea of holidays as ‘a natural opportunity for mating’ and couples who decide to start trying ‘after the new year’ (as I did – resulting in my first child having a late September birthday).
How do we know and why are there gaps?
Data for the mid-19th century onward represent actual births, but data for the early period are derived from parish registers, and therefore reflect baptisms rather than births. This slightly complicates the picture as baptisms necessarily take place after birth, so any gap between birth and baptism can distort seasonality. In the 16th century, however, baptism generally happened very soon after birth, so baptism month is a good indicator of birth month. By 1800 there was an average gap of about a month between the two events, so the shift of the February-March peak in births is likely to be exaggerated.
There are gaps in the series because although birth and marriage records exist, they are not available for systematic analysis, and the Registrar General did not publish summary monthly totals until 1938 for births and 1943 for marriages. We know something of seasonality in this period because the Registrar General published events by quarter of the year, but it is not as finely grained as the monthly data and cannot easily be shown on the same graphs.
References and further reading
- Birth heat map 1995-2014: https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/livebirths/articles/howpopularisyourbirthday/2015-12-18
- Birth heat map 2001-2021: https://www.ons.gov.uk/visualisations/dvc2381/heatmap/index.html
- Marriage seasonality since 1947: https://www.ons.gov.uk/visualisations/marriages/marriages/index.html
- Beveridge, W. H., ‘Marriage and Birth Seasons’, Economica 3:10 (1936), 133–61.
- Durston, C., ‘‘Unhallowed Wedlocks’: the Regulation of Marriage during the English Revolution’, The Historical Journal 31:1 (1988), 45-59.
- Kussmaul, A., ‘Time and Space, Hoofs and Grain: The Seasonality of Marriage in England’, The Journal of Interdisciplinary History 15:4 (1985), 755–779.
- Wrigley, E.A and Schofield, R.S., The Population History of England 1541-1871: A reconstruction (CUP, 1981), pp. 286-305.
Tags: baptism, birth, marriage, seasonality