Introduction

Historians agree the increase in British agricultural productivity between 1550 and 1880 was the result of major structural and technological innovations, notably: enclosure of open fields and commons, the adoption of new field rotation systems, the greater use of soil conditioners and fertilisers, and the improvement of livestock through selective breeding1,2,3,4. But this is where the consensus ends. Enduring disagreement as to when and how rapidly these developments occurred persists due to the lack of representative farming data, particularly for livestock5,6,7, prior to the start of annual agricultural returns in the 1860s. Consequently at least five periods of agricultural ‘revolution’ have been proposed5.

While the analysis of animal bones has provided valuable insights into the pastoral economy8,9,10, early-modern zooarchaeological assemblages are rarely of a chronological resolution that permits examination on a scale below that of a century. In contrast, historic legal documents written on sheepskin parchment (Fig. 1) record the day, month and year the agreement was signed; a date likely only a few months after the death of the animal from which the parchment was produced11. These documents provide an exceptional resource for high-resolution investigations of animal and land management strategies through stable isotope analysis, a tool used for reconstructing diet12, discriminating the use of organic and inorganic fertilisers13, exploring stocking densities14 and identifying transhumance15. Recent analysis demonstrates that parchment is a viable analyte for isotope analysis, recording dietary and physiological signals from the weeks and months before death16,17. Therefore to provide new insight on the timing, extent and drivers of agricultural change, we undertook the isotopic analysis (δ13C and δ15N) of 658 historic legal documents written on sheepskin parchment (Table 1).

Figure 1
figure 1

Legal deeds. (a) Title deed concerning the ownership of land in Enfield, Middlesex, signed and sealed 15th January 1499 (sample DL035); (b) title deed concerning property in Hanbury Woodend, Staffordshire, signed and sealed 11th August 1728 (sample DL110).

Full size image
Table 1 Collection information of parchment documents analysed in the study (n = number of samples). Collagen quality criteria outlined in the methods section. The artificial collections contain documents of different provenance, while the others have grown organically around a single property.
Full size table

Results and discussion

Full δ13C, δ15N and elemental composition results are presented in Supplementary Dataset 1, with summary statistics provided in Table 2, and plotted chronologically in Fig. 2. Of the 658 samples analysed, 23 failed to meet collagen quality criteria and were excluded from the analysis.

Table 2 Summary of δ13C and δ15N values by 25-year periods (n = number of samples). Due to the small sample sizes, parchment dated between AD 1499–1599 and 1925–1969 are presented as single groups.
Full size table
Figure 2
figure 2

Isotope values from historic sheepskin parchment. (a) δ13C (top) and δ15N (bottom) of individual skins plotted against the year the document was signed; (b) 5 and 10-year averages of δ13C (top) and δ15N (bottom) from AD 1650 to AD 1900.

Full size image

Procurement of skins

Parchment produced δ13C values from − 24.3‰ to − 15.9‰ (mean: − 22.4‰). Except for seven skins from the late 19th (n = 1) and twentieth century (n = 6) with values higher than − 20‰, all are consistent with sheep raised in the British Isles grazing on C3 grasses in agreement with the documents’ provenance18. Those with values above − 20‰ indicate the consumption of C4 plants19, although with the global trade in both livestock and fodder in these later decades, it is difficult to say if it is the animal or their feed which are of a non-domestic origin.

Parchment produced δ15N values from 5.6‰ to 14.2‰ (mean: 9.1‰), with many higher than those typically observed in terrestrial herbivores. Sheep bone collagen from the same period has yielded δ15N values from 3.6 to 10.2‰ (mean: 6‰)20,21, therefore the elevated values in parchment can be attributed in part to inter-tissue isotopic discrimination (mean δ15N(skin-bone) + 1.1‰17) and the impact of parchment production (mean δ15N(skin-parchment) + 0.3‰16). Additionally, these values likely reflect the finishing of sheep on well manured crops and pasture in an effort to increase their weight before slaughter. Across the English downlands where the parchment making industry was principally located11, crop cultivation was supported by large flocks that were folded on fallow arable fields to enrich the thin soil with their dung22. To provide a concentrated covering, sheep were kept in moveable pens in high densities23,24; a practice with the potential to substantially elevate δ15N values in the crops and the consumer tissues25,26. This supplementary feeding over the last 6–8 weeks27 would not be detectable isotopically in bone collagen but would be in skin collagen due to its faster turnover rate28,29. More detailed assessment of regional patterns is not possible as the location the legal agreement concerns, or the stationer through whom the parchment was sold, typically bear no relation to the location the animal was raised11.

It is unlikely these high values reflect the use of young lambskins exhibiting a trophic enrichment from suckling, which elevates δ15N in the offspring’s tissues around 2‰ above that of the mother30. Lambs were typically weaned around 16–18 weeks old during this period31; too young to yield a skin of the ~ 70 × 50 cm typical of legal deeds. Legislation prohibiting the importation of leather gloves ensured lambskins were reserved for these and other expensive goods, while cheaper adult skins were used for parchment—as specified in contemporary manufacturing guides32,33,34.

Episodes of famine and disease

The unusually high δ15N of some skins may indicate they come from sheep which had been in poor health before death. Acute physiological stress due to starvation or illness typically results in an elevation of δ15N due to catabolism of the body’s own protein resulting in a fractionation characteristic of trophic level increases35,36. While most nutritional stress studies have focussed on inert tissues, recent analysis demonstrates that the rapid turnover of skin collagen enables the recording of short-term stresses isotopically17.

A series of elevated nitrogen values coincide with an outbreak of ‘sheep-rot’ during the nineteenth century. Sheep-rot (bacterial pododermatitis) is a highly contagious and painful condition characterised by necrosis of the interdigital skin causing severe lameness, a reluctance to graze, emaciation, and ultimately death if untreated37. Epidemics appeared throughout the early-modern period, but a particularly virulent form spread across Britain between 1828 and 1831, killing an estimated 8 million sheep; a quarter of the national flock38. The twenty-one deeds covering these years have a mean δ15N of 9.6‰, comparable with that of all samples, but five of these (five separate documents across three collections) have δ15N values > 10.5‰, including one with an extraordinary high δ15N value of 14.2‰, the highest observed in all samples.

It is tempting to interpret these five as victims of the rot which experienced acute inappetence and physiological stress prior to death. The disease can take months to reach an advanced stage and persist in chronic form for a duration more than sufficient to be recorded isotopically in skin39. Contemporary accounts indicate sheep suffering from rot were sent to market as farmers sought to recoup their losses40, providing the opportunity for their skins to be sent to a parchment maker. Yet the skins showed no visible sign of malnutrition, such as an impression of the ribs due to excessive thinness34. Dearth and disease were perennial challenges for farmers prior to the advent of pesticides and antibiotics and one must be careful not to simply find historical accounts that support these anomalies. However this and other correlations invite future investigations of parchment for paleopathology studies using additional biomarkers of stress and nutritional deficiency.

Agricultural change after the Napoleonic Wars

δ13C values between the early sixteenth century and early nineteenth century display remarkable uniformity, with almost all falling between − 24 and − 22‰. When grouped into 25-year periods, no statistically significant difference was observed between quarter centuries (Table 3), with just 0.1‰ variation in mean values between 1600–1624 and 1800–1824 (Table 2). In contrast, δ15N values exhibit a high degree of variability, although again with no statistically significant variation in mean values between 25-year periods.

Table 3 Significance of differences between 25-year periods in isotope values (Mann–Whitney U test). Number of samples in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.
Full size table

The absence of statistically significant variation across the sixteenth to eighteenth century is surprising considering contemporary agricultural innovations, such as new field rotation systems which replaced fallows with grass leys and nitrogen-fixing root vegetables and legumes5. Biometrical data indicates an increase in the size of sheep across these centuries8, and as the dimensions of the appendicular skeleton are heavily influenced by diet41 it has been suggested that this size increase may reflect nutritional improvement42. Our results, as well as those of Fisher and Thomas43 for cattle, suggest that early size increases were not accompanied by isotopically detectable changes in environment or diet, and were more likely to have been the result of genetic modifications through the introduction of new stock and selective breeding.

A significant change in isotopic values occurred from the early-nineteenth century, with mean values for both δ13C and δ15N increasing across the nineteenth century (Table 3). There is considerable ambiguity in interpreting complex data of this nature, with multiple possible readings of this increase. We suggest this reflects the shift in animal and land management in the wake of the Napoleonic Wars (1803–1815). The war created boom conditions for arable farmers, stimulating an extension of the land under cultivation beyond even the limits of the Second World War’s reclamation campaign44. As hostilities ceased and the continental blockade lifted, the era of high prices was replaced by decades of deflation. With the fresh supply of cultivable land almost exhausted and innovations introduced during the seventeenth and eighteenth centuries now running into diminishing returns7, farmers sought new ways to enhance their profits and productivity. Many met the fall in prices with intensive mixed farming, known as ‘high farming’ or ‘high feeding’, which achieved high outputs by maintaining large numbers of livestock on imported feeds, producing more manure, which in turn increased soil fertility and ultimately grain yields.

Previous improvements had essentially involved the recycling of materials produced on the farm itself. The essence of what F.M.L Thompson termed the ‘Second Agricultural Revolution’ (1815–1880) was that “it broke the closed-circuit system and made the operations of the farmer much more like those of the factory owner”4, relying on inputs imported from outside the farm and indeed the country, particularly oil-cake fodder and bonemeal fertiliser44,45. Oil-cakes were a by-product of rape, linseed and cottonseed oil extraction. Virtually all were derived from imported materials, at first Prussia, then Russia from the 1820s, India from the 1850s, and Egypt from the 1860s4. American maize was added to oil cakes from the 1840s45. Their consumption increased substantially during the nineteenth century, growing from 35,000 tons in the 1820s to 740,000 in the 1880s. Crops grown in continental Europe are characterised by higher δ13C values than those grown in the British Isles, a difference which is recorded in the isotopic composition of the animal tissues18,46,47. The consumption of oil-cakes made from imported crops, including C4 maize, is likely the key driver in the increasingly higher δ13C values seen from 1815 onwards.

The growing use of organic fertilisers such as bonemeal in the nineteenth century is reflected in the significant elevation in mean δ15N between 1800–1824 and 1825–1849 (Table 3). Bones were either crushed into half-inch pieces, ground into bonemeal or dissolved in sulphuric acid to be converted into superphosphate7. The quantity of domestic and imported bones is estimated to have increased from 30,000 tonnes in the 1810s to 115,000 by the 1880s4. Even the bones of those who fell at Waterloo were ground for fertiliser, performing, as one farmer regarded, “the less glorious, but more useful purpose of producing wheat for their brothers at home”48. By the mid-eighteenth century, bones were supplemented with the nitrogenous Peruvian seabird guano and Chilean nitrates7 as increasingly distant sources of nutrients were used to replenish exhausted British fields. Despite the elevated δ15N values of marine bird guano (> + 26‰49) there is no clear evidence of the 1840–1880s ‘guano craze’ in sheepskins from this period, though contemporary surveys suggest it was infrequently used on the chalk downlands where the parchment making industry was principally located50.

Conclusions

The isotopic compositions of sheepskin parchment indicate a substantial change in animal and land management from the second quarter of the nineteenth century. This transformation was driven by renewed and unfettered access to continental and American markets for manures and animal feed, which developed into ‘High Farming’ as agricultural production adapted to more intensive patterns of feeding and manuring. This quantitative data is a new and significant addition to the theory that there were staged episodes within a long ‘agricultural revolution’ much of which was contemporaneously with full industrialisation51. More broadly, the findings demonstrate that parchment is an extraordinary high-resolution biomolecular archive through which centuries of environmental history, agricultural history and animal health can be explored.

Materials and methods

Sampling

Samples were obtained from 645 individual skins from a total of 477 deeds. Of the deeds with multiple pages, each was of a size (> 70 × 50 cm) to indicate they came from a single animal. Samples (approx. 1 × 5 mm) were removed from the edge of each skin from an area devoid of any ink, stamp, glue, seal or surface marking to avoid contamination.

Species identification via peptide mass fingerprinting, presented in Doherty et al.52, identified 622 (96.4%) as sheep (Ovis aries), whilst the remaining 23 (3.6%) could be classified as sheep or goat (Capra aegagrus hircus) as separation between the species was not possible due to a lack of diagnostic peptides. Sheepskin was preferentially used over goat of calfskin (Bos taurus) for the production of legal documents from the twelfth century onwards in England, Wales and Ireland due to its susceptibility to delaminate when scrapped serving to deter fraudulent textual erasure and modification52. No British legal deed has been identified as goatskin through biomolecular analysis52,53. It is therefore highly likely that these 23 skins are from sheep and as such were included in the analysis.

Isotopic data from thirteen eighteenth and nineteenth century British legal deeds reported by Campana et al.54 have been included in all statistical analyses, bringing the total to 658. While the original DNA analysis could not determine the species, subsequent identification via peptide mass fingerprinting has identified these as sheep (pers. comm. M.J Collins). Samples were prepared following the same methodology as this study.

Stable isotope analysis

Samples were prepared for stable isotope analysis at BioArCh facilities, Department of Archaeology, University of York, following the methodology outlined in Doherty et al.16. Lipids were removed via solvent extraction, dichloromethane/methanol (2:1 v/v), by ultrasonication for 1 h, with the supernatant removed and solvent replaced every 15 min. The samples were subsequently demineralised in 0.6 M HCI at 4 °C for 6 h to remove residual calcium carbonate/hydroxide, rinsed with distilled water, and gelatinised in 0.001 M pH 3 HCI at 80 °C for 48 h. The supernatant containing the collagen was filtered (60 μm Ezee-Filter™, Elkay Laboratories, UK), frozen and freeze-dried.

Prepared collagen (0.9–1.1 mg) was weighed out in duplicate in 5 × 3.5 mm tin capsules (Elemental Microanalysis, Okehampton, UK) and analysed at the Natural Environment Research Council Life Sciences Mass Spectrometry Facility (NERC LSMSF) in East Kilbride, UK, where isotope ratio determinations were carried out on a ThermoElectron DeltaPlusXP (Thermo Fisher Scientific, Bremen, DE) with an Elementar Pyrocube elemental analyser (Elementar UK Ltd). Sample data were reported in standard delta per mil notation (δ ‰) relative to V-PDB (δ13C) and AIR (δ15N) international standards. Three laboratory reference materials were interspersed within the measurement run to correct for linearity and instrument drift. Each of the laboratory reference materials is checked regularly against international standards USGS40 and USGS41. Following the calculations outlined in Szpak et al.55, the total analytical uncertainty was estimated to be ± 0.18‰ for δ13C and ± 0.20‰ for δ15N.

To account for the changing ratio of atmospheric 13CO2 to 12CO2 due to increased anthropogenic fossil fuel emissions, parchment δ13C values were corrected following Dombrosky’s56 Suess model to the average atmospheric δ13C of AD 1760 (− 6.4‰), prior to the Industrial Revolution (Supplementary Dataset 1). This atmospheric correction increased the amount of 13C in samples after AD 1807 by 0.1–0.9‰.

Collagen quality indicators

Parchment produced an average collagen yield of 69.4% (range 32.4–98.3%), and %C ranged from 37.1 to 47.3% and %N from 12.3 to 16.9%; all consistent with modern parchment and skin16. Samples produced C:N ratios ranging from 3.1 to 3.5. Skin collagen has a theoretical C:N ratio of 3.1116 and the elevated ratios in parchment are likely due to collagen hydrolysis during the liming process. Twenty three samples produced C:N ratios > 3.5 and were excluded from the analysis.

Statistical analysis

Statistical testing was carried out using the IBM SPSS Statistics 27 software package. Shapiro-Wilks test for normality indicated δ13C data did not conform to a normal distribution (P < 0.05), thus the resultant statistical tests were non-parametric in nature. Significance of difference in isotope values between period 25-year period groupings were evaluated with a Mann–Whitney U test.

Ethical compliance

No experiments were performed on live animals, or animal tissue deriving from previous experiments. The authors were not involved in the life or death of animals from which the parchment was made.