The data used in the study are available in four files:

1. Regional IMR.ods is an Open Document Spreadsheet containing the data behind table 1 of the paper. It shows which rural Registration Districts (RDs) used in the study (those meeting our definition of rural and having data) were in which counties and which counties were in which 1851 Census Regions. Each RD has births and infant deaths for the 1850s and the 1900s, from which region-level rural infant mortality rates (IMRs) are calculated.

2. RD IMR by decade.ods is an Open Document Spreadsheet containing IMR by RD by decade. This data was used to produce table 2 of the paper and figures 2-4. All the RDs used in the study are listed in rows, organised in the clusters of similar IMR trajectory into which we sorted them (see paper). The meanings of the columns are mainly self-explanatory: they are:

	ID		The RD's ID number
	regdist		Name of the RD
	county		The county in which the RD was situated
	avgelev		Average elevation above sea level of the RD
	cluster		Our label for the different clusters used in the paper (see figures 2 and 3 for their IMR trajectories and geographical distributions).
	distance	Distance of the RD's centroid from central London (km)
	IMR50 etc	IMR for the 1850s, etc IMR00 is the value for the 1900s.
	changeIMR	IMR00 minus IMR50

All the data comes from the publicly available GB Historical GIS, obtainable from the UK Data Archive, with the exception of the elevation data. We calculated this in GIS from raw elevation data obtained from Consultative Group on International Agricultural Research Consortium for Spatial Information, ‘SRTM 90m digital elevation data’, http://srtm.csi.cgiar.org/ (accessed on 20 Oct. 2015). 
 

3. RD1911.shp is a GIS shapefile containing the 1911 boundaries of all RDs in England and Wales.