skip to primary navigation skip to content


Short-term and spatial variations in infectious disease mortality in England 1600-1837


1. The changing volatility of mortality

The most dramatic declines in year-on-year burial volatility were observed in seventeenth century London, associated with the disappearance after 1665 of the very great epidemic peaks associated with plague. Although London’s plague outbreaks were local rather than nationwide, they did have simultaneous impact in parts of the counties neighbouring the metropolis. We found that during the plague years themselves, parishes in the hinterland of London on the River Thames and its tributaries were most likely to experience synchronous peaks and troughs in mortality with London.

Burial volatility in London suburbs, 1560 to 1743 (annual burials detrended to remove the effect of population growth)

Image as described adjacent

At the national level, we found that the degree of variation in burials from one year to the next began to decline in all sizes of settlement during the eighteenth century. This happened more swiftly in some regions and types of settlement than others. The decline was most apparent in the largest urban centres.

Variance in burials (detrended) by size of settlement 1675-1799

Image as described adjacent

2. Population thresholds and the divergence in adult and child mortality

Improvements in adult life chances were not always accompanied by improvements in child mortality. Wrigley and Schofield drew attention to England’s late seventeenth and early eighteenth century divergence in adult and child mortality at the national level, strikingly illustrated in the ‘back-track’ between 1640 and 1730.

Movements in adult and child life expectancy 1640-1809

Image as described adjacent

We found that the volatility of mortality declined at different rates among adults and children, and that the rate of change was affected by settlement population size. In major cities the annual incidence of child burials varied more than that of adult burials and was most pronounced for children aged 1 to 4 years rather than among infants or older children

In the major cities of Leeds and Liverpool, sudden ‘cliff edges’, or precipitous downward shifts in the variance in annual burials totals for both adults and children were apparent when burials reached 400 per annum, equivalent to a population of some 10,000 persons if we assume a crude death rate of about 40 per thousand. These thresholds were crossed at different times that reflect the timing of population growth in each case (Leeds grew rapidly from the late sixteenth century whereas Liverpool grew explosively from 1660 onwards). In Leeds the reduction in adult burial volatility was more gradual and preceded that for child burials, but in more recently settled Liverpool the reduction in both adult and child burial volatility was simultaneous.

Variance in Leeds burials (detrended to remove population growth) 1603-1778

Image as described adjacent

Variance in Liverpool burials (detrended to remove population growth)

Image as described adjacent

3. Settlement size and the endemicisation of childhood diseases

Causes of death in parish registers

Burial registers recording causes of death and age information start to become more widespread by the second half of the eighteenth century. In London it is presumably the requirements of the Bills of Mortality that produce sporadic parish register recording of causes of deaths with ages. While the Bills required report both the age structure of deaths and deaths by specific diseases, they do not cross-tabulate both, so the parish register information is particularly valuable. In other towns where Bills of Mortality were not produced, or only produced in sporadic years, the parish register causes of death are all the more important for having virtually no corollary.

We found that those burial registers where ages and causes of death are recorded yield consistent patterns of age-specific mortality in different parts of the country. Diseases such as whooping cough, measles and scarlet fever that affect children almost exclusively can easily be distinguished from diseases of adulthood such as deaths in childbed, apoplexy and, of course, ‘old age’.

Cumulative distribution by age of selected causes of death in Leeds 1778-99

Image as described adjacent

Epidemics of childhood infectious diseases

We found that settlement size and growth rate determine whether a location experienced the acute childhood infectious diseases periodically or in every year. In small settlements such as Braithwell in the West Riding of Yorkshire, with a population of just 331 persons in 1801, smallpox, measles and pertussis deaths did not occur in every year. However, when epidemics did break out they were responsible for a large proportion of all deaths that year.

Frequency of selected infectious diseases of early childhood in Braithwell village

Image as described adjacent

In medium-sized settlements of a few thousand inhabitants, such as Hindley in Lancashire, smallpox was present in most years but pertussis and measles were not. When smallpox began to subside in Hindley after 1809 (probably as a result of localised vaccination initiatives) there was a prolonged three year measles epidemic, suggesting that some of the lives saved by vaccination were sacrificed to measles.

Frequency of selected infectious diseases of early childhood in Hindley

Image as described adjacent

In moderately large but relatively slow-growing cities such as York, smallpox was endemic, although measles and pertussis were not. While endemicisation meant that the death toll from smallpox was in most years a less prominent component of overall mortality, fluctuating around 10% of all burials as opposed to the 30% or 40% found in smaller settlements when smallpox epidemics arose, deaths occurred with greater regularity.

Frequency of selected infectious diseases of early childhood in York

Image as described adjacent

In the largest and fastest-growing settlements and their periphery, where tens of thousands of persons lived in close proximity, continual infection of new susceptible individuals ensured that smallpox and pertussis deaths were present in every year, and measles every few years. In Leeds smallpox seems to have increased in severity after 1778, although by the early nineteenth century after vaccination had been introduced it could be entirely eclipsed by measles and pertussis, as during the particularly severe measles epidemic of 1807. In some of the preceding years an outbreak of flux carried away a proportion of the young children who might otherwise have died of smallpox.

Frequency of selected infectious diseases of early childhood in Leeds

Image as described adjacent

4. Drivers of early childhood mortality


Smallpox was a key driver of short-run mortality fluctuations among children by the mid to late eighteenth century. It became increasingly concentrated in early childhood, with deaths attributed to the disease affecting primarily children aged between 1 and 4 years, but also a growing proportion of infants under 1 year. By the first decade of the nineteenth century fewer older children aged 5 and above were dying of the disease.

The changing age incidence of smallpox deaths 1750-1812 in sample urban areas

Image as described adjacent

Diarrhoeal diseases

Causes of death indicative of diarrhoeal diseases included dysentery, griping in the guts, flux and bowel complaint (cholera being a new epidemic disease of later decades of the nineteenth century). These affected adults as well as children, but were most devastating among the young. We found considerable geographical and temporal variation in the mortality burden of diarrhoeal diseases. In low-lying Liverpool and in areas of east London adjacent to the river Thames, water-borne diarrhoeal diseases contributed a large but declining share of all deaths in early childhood in the eighteenth and early nineteenth century. However, in colder, more elevated Leeds their contribution was much more minor.

Early childhood (4q1) causes of death in urban centres 1745-1812

Image as described adjacent

In the London suburbs diarrhoeal diseases had perhaps not always been such a heavy burden as they became by the mid eighteenth century, although it is difficult to find much age-specific cause of death data from earlier periods (the London Bills of Mortality reported causes of death from 1629 but did not cross-tabulate by age) and the descriptors are more problematic. However, in the late sixteenth century when London’s east end was still adjacent to countryside, the available evidence suggests that although strongly affected by plague, even omitting the known plague year of 1593, mortality in the 1 to 5 years age group was little-affected by diarrhoeal disease.

Early childhood (4q1) causes of death in east London 1583-1802

Image as described adjacent

In July to October high ambient and sea temperatures provide optimal growth conditions for pathogens such as the Shigella bacterium responsible for dysentery, and diarrhoeal diseases peak in the later summer and early autumn months. In Liverpool, when diarrhoeal disease mortality peaked, other major infectious diseases such as smallpox temporarily subsided, providing a clear illustration of the often-reciprocal nature of the mortality burden of different diseases.

Seasonality of diarrhoeal diseases and smallpox in Toxteth Park, Liverpool 1801-1812

Image as described adjacent

5. Spatial convergence in mortality fluctuations

We measured the degree of convergence in the mortality experience of different locations by cross-correlating annual burials totals between all pairs of sample parishes in observation, generating. We then represented moderate-strong correlations as lines connecting the parishes on a map of England (see figure). Three regions of spatial convergence in annual mortality experience associated with transport and trade links emerged early on in the area west of the Pennines, in Yorkshire and in the south east.

Spatial convergence between annual burials totals for pairs of parishes 1600-49

Image as described adjacent

London, Norwich and Ipswich were the settlements whose mortality experience was shared with the largest number of other settlements. However, examining the maps at regional level, it is apparent that the central nexus of shared mortality experience is not always a major settlement in its own right, but may also be a centre of trans-shipping. Notable examples were Stratford Bow on the River Lea near London and Howden in Yorkshire, which is the lynch-pin of the Yorkshire region of spatial convergence in 1600-49 shown above.

River connections for Howden, Yorkshire

Image as described adjacent

Over time, correlations between parish pairs actually decreased after 1700 despite improvements in transport infrastructure, contrary to expectation. We attribute this fragmentation was due to the development of a larger number of self-sustaining local mortality regimes dominated by cycles of infectious diseases among the young, as more settlements reached population thresholds where they could sustain pathogens endemically.

Decrease in correlation in annual burials fluctuations between pairs of parishes over time

Image as described adjacent