GIS Cartographer and Honorary Research Fellow
School of Geographical & Earth Sciences
University of Glasgow
The Hearth Tax was a property tax levied in England and Wales from 1662 until 1689 in the early modern period and levied on each hearth in a household and was by proxy a tax based on wealth or poverty. Poverty mapping is not a new concept and was undertaken by Charles Booth between 1886 and 1903 when he researched and published several poverty maps in a seventeen-volume study of wealth and poverty in London entitled “Life and Labour of the People of London”. Full details of this can be found in the excellent online article by former Hearth Tax Research Project Manager Dr John Price (now Senior Lecturer, Goldsmiths, University of London) “London Hearth Tax Mapping: A ‘Charles Booth’ for the Seventeenth Century?”
An example of the Booth Poverty Map 1891
Since the time of Booth’s mapping, several projects have been active in the thematic mapping aspects of the Hearth Tax for selected counties or cities in England and Wales. The https://www.thebritishacademy.ac.uk Hearth Tax Research Project at the University of Roehampton has been involved in mapping the Hearth Tax and its aspect of poverty mapping, using modern Geographical Information Systems (GIS) since 1999. It was then that I took over the general mapping from Ian Agnew cartographer at Cambridge University and began using GIS software for the Hearth Tax analysis and mapping. The first set of Colour Distribution and General Maps I created in collaboration with Professor Margaret Spufford used a combination of GIS software (ArcGIS) and graphics software (Adobe Illustrator) and were for the Kent volume at the Hundred administrative unit level published in 2000. All hundred (and lathe) boundaries were manually digitised (digitally traced) at 1:126,720 scale (Ordnance Survey, half inch to one mile, 1881) with GIS software on a large format digitising tablet to create enclosed area polygons (shapefiles) which were digitally linked to attribute tables (dbase) containing aggregated hearth tax data percentages for each hundred. This allowed a set of colour distribution maps to be automatically generated in the GIS for the percentage of hearths per hundred (and similar administrative units) according to the number of hearths e.g., 1 hearth households, 2 hearth households, 3-4 hearth households, 5-9 hearth households, over 10 hearth households, exempt hearth households, hearth and household density per 1000 acres. This methodology of using 3 and over hearth households, together with the use of sub-regions, was put forward by Tom Arkell to analyse distributions of wealth and poverty (Arkell 2003).
To allow access to flexible cartographic design capabilities and high-resolution eps digital output, the maps were imported into Adobe Illustrator graphics software for final cartographic design and layout (see full explanation of the mapping procedures in the section below ‘Creating the Hearth Tax Maps’). An additional larger format monochrome map was also produced to show the hundred / lathe boundaries and their associated names. The Kent volume provided the initial template for mapping the Hearth Tax for 14 subsequent volumes for the Hearth Tax Research Project at the University of Roehampton, published by the British Record Society, up to and including the most recent volume for the county of Norfolk due for publication in 2022 and also for the Huntingdonshire and Northamptonshire volumes which are both at an advanced stage.
Hearth Tax Research Project – Kent hundreds, percentage of households with one hearth
GIS Data and Basemap Sources
The publication of a very significant GIS dataset occurred in 2000 and proved to be of great benefit to the ongoing mapping of the Hearth Tax. As a result of a six-year project, the Great Britain Historical GIS (GBH GIS) published a comprehensive, consistent and accurate cartographic record of the changing administrative and reporting units of the UK since the creation of the modern system of civil administration in the mid-nineteenth century in a GIS shapefile spatial format. The project was partly funded by the Hearth Tax Research Project. The base mapping was digitizing using the 1:126,720 (half- inch to one mile) County Administrative Diagrams first published by Ordnance Survey between 1906 and 1910 and critically for the Hearth Tax, included pre-1881 Parish level boundaries. With digitized parish boundaries now available, the task of recreating the parish boundaries to match those at the time of the Hearth Tax would be significantly reduced, resulting in faster map production and reduced mapping costs.
Extract from the Great Britain Historical GIS database for part of Kent
Two other secondary map resources used for Mapping the Hearth Tax was the Phillimore Atlas (Humphery-Smith, 2003) and Historical Parishes of England & Wales (Kain & Oliver, 2001). The Phillimore Atlas is a paper based volume with small scale county maps showing pre-1832 parish, tithe and parochial boundaries and is a useful reference to the general location and registration dates of pre-1832 parishes that don’t exist on the Historical GIS database.
Extract from the Phillimore Atlas for part of Kent
Kain & Oliver published their book and maps in 2001, the maps were in GIS and Adobe Illustrator formats with Parish boundaries based on 1840 Tithe and 1760-1860 Parliamentary Enclosure ‘Cadastral’ maps overlaid on 1:63,360 (one inch to one mile) Ordnance Survey maps dated 1945-48. The Kain & Oliver parish boundaries are more detailed than the Historical GIS boundaries and also contain parishes from an earlier period. However, these were deemed too detailed for the Colour Distribution Maps required for the Hearth Tax county volumes which were to be produced at the relatively small 11x 15cm size, the maps were very useful for identifying and locating pre-1881 parish boundaries.
Extract from the Kain & Oliver GIS for part of Kent
Creating the Hearth Tax Maps
From 2000 onwards, a similar mapping procedure was adopted for the creation of the series of colour distribution maps to be produced for each published county volume. In addition, a key map showing parish names, terrain and topography maps were created, plus enlarged parish maps for major town or cities where required.
A standardised specification for the classification bands to be used in the colour distribution maps was produced by Professor Margaret Spufford and Professor Peter Spufford and a colour specification for each band was created in Adobe Illustrator. Note: a 25%-35% band means that 24.999% is outside and 25.001% is inside the band.
Hearth Tax Classification and Colour Bands
The following procedure was adopted to move from a GIS database to a final Encapsulated PostScript file (Eps) for each colour distribution map to be supplied to the printer for publication.
Step 1 – use ArcGIS (ArcMap) to extract the appropriate county parish boundary area polygon shapefile from the Historical GIS database as the primary source and update them in ArcGIS using the Phillimore and Kain & Oliver mapping as secondary sources, to finally match each GIS parish to the parish list created from the published Hearth Tax assessments / returns.
Step 2 – merge the GIS parishes shapefile in ArcGIS with the Hearth Tax percentages tables created in Excel format, using the parish names as the common field for linking the data.
Step 3 – create a layout template in Adobe Illustrator for the maps to match the size of the County Volume to be published (normally 11x 15cm size).
Step 4 – import the ArcGIS shapefile into Adobe Illustrator graphics software using the Avenza MaPublisher plugin.
Step 5 – use Adobe Illustrator in combination with the MaPublisher legend tool to create each distribution map, including the legend from the GIS attribute table containing the Hearth Tax statistical percentages.
Step 6 – create an Adobe Acrobat Pdf file for checking purpose by the county volume editors.
Step 7 – export each final checked map to Eps format for publication.
Mapping the Hearth Tax with Adobe Illustrator and MaPublisher
Additional maps, such as a key map showing parish names or index numbers, terrain and topography maps, plus enlarged parish maps for major town or cities where required, were also created in Adobe Illustrator. With the advancement of the cartographic design and output capabilities of GIS software (e.g., ArcGIS; QGIS), it is now possible to complete the full process from GIS database to a final Encapsulated PostScript file (Eps) and may be considered for future published volumes.
Margaret Spufford’s interest in the social structure of early modern England led to the use of hearth numbers as proxies for social ranks, ranging from poor households with 1 hearth to the homes of the gentry (5-9 hearths), which were linked to the architectural evidence (Spufford, 2000). The maps below from the Essex and Westmorland volumes provide some examples of this. Each county volume contains the standard set of maps to enable comparisons between counties, as well as some maps to draw out the particular contexts.
A recent and interesting ‘spin-off’ from the mapped volumes has been a series of consolidated colour distribution maps for England, combining all published counties in one map for each of the following – 1 hearth households, 3 & over hearth households, exempt hearth household’s hearth density and household density. These maps which were created for research purposes for Dr Andrew Wareham (Director of the Hearth Tax Research Project and Reader, School of Humanities and Social Sciences, University of Roehampton), clearly show a North-South poverty divide and is especially pronounced on the 1 hearth and 3 & over hearth household maps.
Consolidated One Hearth Household maps
Consolidated Three & Over Hearth Household maps
The latest stage of the Hearth Tax Research Project is Hearth Tax Digital (beta version), which utilises the GIS shapefile parish boundaries created for the published Hearth Tax county volumes to map the Hearth Tax graphically and interactively on a web browser. This work is being jointly undertaken by the Centre for Information Modelling (University of Graz, Austria) and the British Academy Hearth Tax Research Project. The mapping function uses the same key as above, but also allows users to apply their own parameters for searches (as well as providing static options to view maps), and the mapping data is presented on Google maps.
Mapping the Hearth Tax – Challenges
Parish, township and borough boundaries are the main administrative units used in recording the location of households for the Hearth Tax and proved to be the most challenging to establish for the period of the Hearth Tax 1662-1674. Ancient parishes are often defined as those which existed prior to the 1597 and 1601 Elizabethan Poor Law Acts. Civil parishes in their modern sense were established in 1894 by a Local Government Act. The earliest maps to show these boundaries were Tithe maps produced by private surveyors around 1840 and Parliamentary Enclosure ‘Cadastral’ maps 1760-1860. These were the main sources used by Kain & Oliver in their mapping of the 1832 parish boundaries for England & Wales. The first Ordnance Survey map was published in 1801 and covered the county of Kent, but it was many years before it produced detailed maps nationally. For many places, the oldest large-scale (1:10,560, six to one mile) Ordnance Survey map that included parish boundaries date from the 1860s, 1870s or 1880s. As stated previously, the Great Britain Historical GIS created a national GIS database for pre-1881 Parish level boundaries using the Ordnance Survey 1:126,720 (half-inch to one mile) maps dated 1906 and 1910.
There were however noticeable differences in the alignment of parish boundaries that had to be considered when the Historical GIS and Kain & Oliver shapefiles were overlaid in ArcGIS software, due to the differing Ordnance Survey base maps used to digitize them and possible projection / scaling issues and differences. The differing dates of the map sources also highlighted many instances where a small parish or parishes had been merged, to create a single larger parish and occasionally where large parishes had been split to create smaller parishes. A further complication was that the Historical GIS extended coastal parish boundaries to the Low Water Mark, whereas on the Kain & Oliver maps the parish boundaries followed the High Water Mark. All these issues could be resolved successfully with careful use of ArcGIS editing and spatial adjustment tools, however this was often a very exacting and time-consuming task.
Comparing GIS data for part of Kent
Other issues relating to boundaries was the situation where an existing parish was named on the Hearth Tax assessments / returns, but no corresponding parish boundary existed on any of the GIS files. However, in most of these situations the parish appeared on the small scale and less detailed Phillimore Atlas, together with its date of establishment. This made it possible to scan and georeference a section of the map on the Phillimore Atlas in ArcGIS, digitise the boundary and make any adjustments to match adjacent parish boundaries an underlying Ordnance Survey map, or using OpenStreetMap online. Another common issue was the merging of parishes between the earlier pre-1832 Kain & Oliver data and the later 1881 Historical GIS data, which was the basemap used for the Hearth Tax mapping. This was resolved relatively easily by overlaying the two GIS files and edited the boundaries using the split polygon and trace tools.
The issue of adjusting the parish boundaries from Low Water Mark to the High Water Mark was most noticeable in areas of tidal flats and large river estuaries e.g. The Wash, Severn and Morecambe Bay. This was very time consuming as it involved digitising the coastline using online Ordnance Survey and OpenStreetMap maps and clipping out the parish areas below high water.
While boundaries created the biggest challenge in creating the required GIS data for the mapping, there were also issues relating to the Hearth Tax statistical tables, and their merging with the GIS data. The Hearth Tax percentages tables were created in Excel format and the parish names were used as the common field for linking the data. This required some considerable editing of the parish names, as there had to be an exact match with regard to number of parishes, spelling and the use of capital and lower-case lettering. The main issue was often minor, but critical differences in the nomenclature adopted by the GIS and the Hearth Tax return for parish names e.g. Wells-next-the-Sea or Wells next the Sea; Ackington & Park or Ackington and Park; Fulmodeston cum Croxton or Fulmodeston with Croxton etc. in Norfolk.
Over the past twenty or so years, I have enjoyed my involvement with the British Academy Hearth Tax Research Project, as it has continued to challenge me to adopt and learn new practices and procedures using GIS and graphic software. Hearth Tax Digital is an important new step in keeping the project alive. Margaret Spufford and her associates have long had a vision of completing the Hearth Tax and its mapping for all counties. We are now a long way along the road to achieving this goal, but have many miles yet to travel.
List of references:
T. Arkell, ‘Identifying Regional Variations from the Hearth Tax’, The Local Historian 33.3 (2003), pp. 148-78
C. Humphrey-Smith, The Phillimore Atlas and Index of Parish Registers (Chichester, 2003)
R. Kain and R. Oliver, The Historical Parishes of England and Wales: An Electronic Map of Boundaries before 1850 with a Gazetteer and Metadata (History Data Service, 2001)
M. Spufford, ‘The Scope of Local History and the Potential of the Hearth Tax Returns’, The Local Historian 30.4 (2000), pp. 202-21