Salving, Smitting and Smorring

Salve and smit are preparations applied to live sheep, while smorring was a dressing applied to woollen sails to reduce air permeability. All three products were traditionally mixed with varieties of animal fat, together with either, or both, Stockholm tar and ochre. This has led to a re-appraisal of the highly comminuted cattle long bone fragments, popularly referred to as “soup kitchen” waste, as a possible source for the necessary fat and emollient for such mixtures, prior to the ready availability of cheap imported butter, fish and whale oils in the 19th century. A small-scale trial empirical study has been undertaken to examine the plausibility of this suggestion.


Before the invention of sheep dip in the late 19th century, sheep were salved with a mixture of tar and fat as a preventative against parasites, such as lice, ticks and keds, and to increase the water shedding property of the fleece before winter. This practice was particularly associated with upland common grazing in northern England. The ratio of the mix was roughly two parts fat to one part tar (Hartley & Ingilby 1981, 89).

The use of tar by shepherds has a much longer history. The stock equipment of the shepherd in sixteenth century literary sources always includes the tar box. For example, Sir Philip Sidney’s “Dialogue Between Two Shepherds” has the lines:

What! Is thy bagpipe broke, or are thy lambs miswent;

Thy wallet or thy tar-box lost, or thy new raiment rent?

The tar used is known generically as Stockholm tar, produced in Scandinavia from pine wood or roots by a method akin to charcoal burning. Shepherds have also traditionally used a tar-based salve for the nicks on the skin of sheep made during shearing and in the prevention and treatment of blowfly strike. Tar is generally considered too astringent to use neat and was mixed with either or both oils and fats, to dilute the sting of the raw tar and ease application. The shepherd’s tar box for this usage appears to have been such a universal item of equipment that it seems to have passed unremarked. Few specimens survive in museum collections. Ingram (1977, 21; 1979, 25) illustrates a wooden example, the museum of English Rural Life has a further wooden one and the Lincolnshire museum service has a more recent tin box, all of which would have been strapped to the shepherd’s forearm. The lack of a lid on these examples suggests a solid salve. The consistency of the raw tar appears to have changed. Modern Stockholm tar is semi-liquid, whereas in the early 20th century it was stiff (Hartley & Ingilby 1981, 89). This is an immediate example where even known historic recipes cannot be followed precisely because of the difference in the raw materials. Empirical trials are necessary to establish the properties of the modern ingredients in order to produce a comparable result.

Smitting and Raddling

Smit marks are the temporary ownership identification marks on the fleece of sheep. These were used in conjunction with horn brands or lug (ear) marks on polled breeds, such as the Herdwick, to allow identification of sheep from a distance, particularly on common upland grazing (Winchester 2000, 105-9). Modern smit marks are made with aerosol sprays, while both lug marks and horn brands have been replaced by ear tags. Raddle is colouring applied to the chest of the ram before he is loosed with the ewes in the autumn, so the ewes are marked when they are mated.

Both smit and raddle were mixtures of red or yellow ochre, naturally occurring iron rich earths, and whatever fat or oil was readily available to the shepherd. Loose raddle powder is still used but modern preparations are generally mixed with mineral or vegetable oils. The consistency varies with the requirement of the individual shepherd, those who paint raddle on with a brush make a more liquid mix than those who apply it with a trowel.


Smorring was the way of proofing the woollen sails, which were the norm for Norse seafarers, to catch the wind (Jorgensen 2005, 65). Modern experiments in conjunction with the Roskilde ship museum in Denmark and Tommervik Textile Trust in Norway (Jorgensen 2005; Cooke and Christiansen 2005) have demonstrated the efficacy of both tar and ochre based smorring mixtures for handmade woollen sails on reproduction Viking ships. Viking sailors were farmers too, so would have been familiar with the uses and properties of tar, ochre and fat on the fleece of live sheep, so the extension of these treatments to woollen textiles for outdoor use would have been entirely logical. The very word tarpaulin is a reminder that waterproof textiles other than sail cloth were a necessity.

The Roskilde sail experiments have shown that both hard and soft fats are needed for the process, that the mixture may need diluting with water for application to a large area of textile and that more than one coating, of differing consistency, may be required.

The Animal Bones

Enormous deposits of highly fragmented cattle long bones have been found at sites such as Carlisle (Rackham 1991) and Zwammerdam (Mensch 1976) and a smaller collection associated with one vicus building at Piercebridge (Rackham and Gidney 2009, 30). The two latter sites were Roman deposits, while the Carlisle context was 8th-9th century AD. The original Zwammerdam report suggested this might have been a “Roman soup kitchen”. Stokes (2000) suggested a range of further possibilities.

In contrast to these rich deposits, a known post-medieval naval dockyard in London, with documentary evidence for large numbers of beef cattle being prepared on site to provision ships, produced a paucity of cattle limb bones (West 1995). West (1995, 35) further observed that the cattle hides were being exported to Holland by the navy victuallers, so postulated that the cattle bones may have accompanied them. In the 17th century, Holland was the centre for the manufacture of the best sailcloth (Nyberg 1998, 294). Trade links between Holland and the Royal Navy appear to have been well established, with the procurement of sailcloth from Holland in the 18th century for the navy (Nyberg 1998, 312).

The possibility thus suggested itself that the product represented by the deposits of highly fragmented long bones could also be used in the preparation of smorring and that the English naval dockyard bones could have been exported to the sail yards of Holland for this purpose. The Zwammerdam finds could suggest some antiquity for such practice. In the 18th century, sails of British fishing-vessels were tanned with mixtures of grease, tar and ochre, though the royal navy did not approve of the practice (Steel 1794, 85).

The author has prepared numerous reference skeletons of cattle and has found that gently simmering complete bones produces a thick jelly stock with a bone marrow based surface layer of fat, without the need to smash the bones into tiny pieces. However, industrial scale production with the relatively small riveted iron plate or sheet metal copper alloy cauldrons available in antiquity would have dictated the necessity to comminute the bones to increase the quantity of bone that would fit in one vessel. Smashing the bones into tiny pieces would also reduce the amount of water required to simmer the bones, resulting in a smaller quantity of more gelatinous stock without the need for further reduction. Rather than using water to dilute the tar-based mix, the possibility of the jelly stock being used for this purpose warrants consideration, particularly for applications other than sails. When the preparation was warmed, this would liquidise for ease of application, while when cold it would be a semi-solid jelly rather than a runny, sticky mixture and so easier to handle, store and transport. Buttermilk was similarly used in some preparations of sheep salve, to help it spread better (Hartley & Ingilby 1981, 89).

The experiments

It is freely acknowledged that the current experiments are empirical rather than truly experimental. The purpose is to examine whether the hypotheses about the bone assemblages and salving or smorring mixes are sufficiently robust to warrant further controlled experiment. Epistemology of craftsmanship (Jorgensen 2005, 67) acknowledges that it is not always possible to define such processes in words or quantify results that may depend on a range of background variables. As medieval cookery books put it, one continues a process “until it is enough”.

Breaking the bones

Previous work by Paul Stokes (2000) in conjunction with the author has shown that it is possible to replicate archaeological finds of fragmented bone. This project has acted as a reminder of the processes involved. Modern marrow bones are discarded intact by the butcher. Breaking fresh beef bones requires considerable effort on the part of the author and a suitable implement, in this instance a 20th century bullock splitting axe. The first two bones processed were both humeri that had been sawn in half by the butcher. These were first attacked with a small axe, weighing about 1kg. This made little impact on a distal trochlea despite the expenditure of much effort. Recourse was then made to a felling axe, weighing about 2.5kg. This has a broad, flat back to the head, reminiscent of instructions in Anglo-Saxon Wortcunning and Leechdoms to break bones with the back of an axe (Stokes 2000, 69). Neither the sharp nor blunt sides of the axe head were successful in fragmenting the bones. Finally, the bullock splitting axe, weighing about 3.3kg, was resorted to, which succeeded in shattering all four half bones, not just the half bones weakened by blows from the other two axes. The weight of the axes appears to be directly related to their efficacy. The second couple of bones processed were a radius with ulna and carpals attached and a tibia with tarsals attached. Only the bullock splitting axe was used. Considerable effort was expended in dismembering the distal articulating groups of bones contained within the soft tissue and so not visible as individual bones, other than the proximal end of the calcaneum. Repeated blows were needed to break the shafts and fragment the ends, particularly of the distal radius and tibia which had the carpals and tarsals, respectively, still in articulation. The bones bounced off the chopping board on the floor for the first few blows. Once the shaft was weakened, the shaft broke with much splintering. The splinters that were not attached to the periosteum flew off in all directions. The larger fragments were collected but many very small fragments were scattered all round the floor. The house dog cleaned up those pieces with soft tissue attached. This method of fragmenting marrow bones can therefore produce the tiny unidentifiable fragments that are a feature of soil sample residues. Both the scattering of tiny fragments on impact and the acid etched fragments in canine faecal remains may thus result from such robust breakage of intact beef marrow bones.

The bone fragments of the two humeri were slowly simmered for about 6 hours to produce a thick jelly with a layer of fat on the surface. A rapid scan of the fragments, prior to full cleaning, revealed that only the two lateral halves of the distal trochlea were immediately identifiable as having diagnostic “zones” which would be recorded in an archaeological assemblage. This is welcome corroboration that the “zone” methodology is robust in identifying the representation of individual bones and is not swamped by recording many fragments deriving from single bones.

The radius and tibia were simmered at a higher setting for a shorter time. The warm stock initially looked similar to the first batch but remained semi-liquid on cooling, rather than setting to jelly, and the surface layer of fat was softer and thinner.

The remaining four bones were two tibiae with attached tarsals, one humerus and one femur. Only the femur was broken into small fragments. The remaining bones were broken in half. It was immediately apparent that far fewer bones could be accommodated within a larger vessel, a greater volume of water was required to cover the bones within the vessel and the simmering time was greatly extended, to approximately 24 hours. Despite the extended cooking time, once the bones had cooled it was apparent that much of the marrow remained within the bone cavities and had not floated to the surface. The remaining fat and jelly from the previous two batches were added to this final batch. The initial investment of effort incurred in smashing the bones into small fragments appears to be amply rewarded by a shorter cooking time, with attendant reduction in fuel use, and a superior product in terms of concentration of the liquid stock to jelly and quantity of surface fat.

Although heath and safety issues were very far from paramount in the design of this project, the highly fragmented bones were far safer on the stove as the contents were well below the rim of the vessel. Emptying the vessel was also safer as the weight of the hot container was easily manageable. Previous experience of preparing intact cattle bones for reference specimens by this method has shown that a large vessel full to the brim of intact bones has several inherent dangers. Temperature control is far more crucial to stop the surface fat layer boiling over and causing a fire. The weight of the full vessel is too great to lift when hot. Lifting the bones out of the hot stock is messy and a potential burn and fire hazard with drips of hot fat. The smaller vessel of broken bones could be lifted off the stove and poured into a colander on top of an empty vessel standing on the floor, to separate the bones from the stock, with virtually no mess.

At the end of cooking the third pan of bones, before turning off the heat source, the temperature of the surface fat layer was 85 degrees C and the underlying stock was 100 degrees C. The temperature gradient is important as fat can, obviously, reach a far higher temperature than the water-based stock. The third batch of jelly and stock was not used as, due to thundery weather conditions, it spoiled. This is a reminder that there were practical considerations for the autumn slaughter of livestock. Coincidentally, the peak slaughter time, when cattle bones would be readily available, was also the time when sheep used to be salved.

Initial Results: the cloth samples

Three samples of smorring preparations were made with Stockholm tar but without ochre, using the fat and jelly from the first batch of boiled bones, the highly fragmented humeri. A lightweight twill wool cloth was used to test the application of the mixtures, which were painted onto one surface leaving a margin round the edges. Following the descriptions of traditional salves and smorring, the mixtures were heated gently to aid amalgamation.

Surface fat and tar – Mixed together well and painted on easily. The cloth remains flexible but very sticky, with good penetration to the underside and spreading to the edges.

Jelly and tar – Despite heating, the tar remained as globules floating within the jelly. Painted on easily. The cloth was touch dry by the following day but stiff, with penetration to the underside but incomplete spreading to the edges.

Fat and jelly in the proportion present in the pan. – Much better integration of the mix than the previous one. The cloth is less sticky than the first and more flexible than the second, with good penetration to the underside and the edges.

A fourth mixture was made with hard beef tallow and tar. Despite the difference in the consistency of the fat used compared to the first mixture, the results are practically identical: very sticky but flexible cloth with the mix well soaked through to the underside and edges. All four of these trial pieces have been coloured very dark brown by the tar.

A small range of mixes incorporating raddle powder were prepared in conjunction with members of the architectural and archaeological society of Durham and Northumberland, as part of an outdoor activity event. These were very empirical field mixes using combinations of either the first or second batches of jelly and fat from boiling the bones, together with Stockholm tar and raddle powder. Despite the intense blackness of the Stockholm tar, the addition of raddle powder masks this thoroughly, with good colour obtained with both red and yellow raddle. The addition of raddle also inhibited the penetration of the mix to the underside and gave a touch dry surface later on the day of application. The flexibility of the cloth samples varies, with some being stiff and some supple. These preliminary samples confirm the experience of Cooke, Christiansen and Hammarlund (2002, 209), who observed that ochre was a good filler of the voids in the fabric and furthermore has anti-bacterial properties which help to prevent the wet sail from rotting. Combinations of tar and ochre would therefore also act as preservatives of woollen textiles.


This very restricted trial study has succeeded in demonstrating that further controlled experiments with bone grease and jelly as ingredients of smorring mixtures would be of value. Such grease and ochre mixes as waterproofing might have been more readily available and affordable to agricultural and urban communities than, for example, the lanolin used in the construction of the leather boat Brendan (Severin 2000, 41) or the beeswax used on medieval style canvas tentage by D. Greenhalgh (pers. comm.). The routine preparation of fat and stock based salving, smitting and smorring mixtures may be a further interpretation for the presence of the highly fragmented cattle limb bones encountered on archaeological sites.


Cooke, B., Christiansen, C. & Hammarlund, L. 2002. Viking woollen square-sails and fabric cover factor. International Journal of Nautical Archaeology 31(2): 202-210.

Cooke, B. & Christiansen, C. 2005. What Makes a Viking Sail? In F. Pritchard & J. P. Wild (Eds) Northern Archaeological Textiles NESAT VII. Oxbow Books: Oxford 70-74.

Hartley, M. & Ingilby, J. 1981. Life and Traditions in the Yorkshire Dales. Dalesman Books: Clapham.

Ingram, A. 1977. Shepherding Tools and Customs. Shire Publications Ltd: Princes Risborough.

Ingram, A. 1979. The Country Animal Doctor. Shire Publications Ltd: Princes Risborough.

Jorgensen, L. B. 2005. Textiles of Seafaring: an Introduction to an Interdisciplinary Research Project. In F. Pritchard & J. P. Wild (Eds) Northern Archaeological Textiles NESAT VII. Oxbow Books: Oxford 65-69.

Mensch, P. J. van 1979. A Roman soup-kitchen at Zwammerdam? Berichten van de Rijksdienst Oudheidkundig voor het Oudheidkundig Bodemonderzoek 24: 159-65

Nyberg, G. G. 1998. Dutch Looms for Sailcloth from the 16th and 17th Centuries. In L. Bender Jorgensen & C. Rinaldo (Eds) Textiles in European Archaeology: Report from the 6th NESAT Symposium. Goteborg University.

Rackham, D. J. 1991. The Animal Bone from Post-Roman Contexts. In M. R. McCarthy The Structural Sequence and Environmental Remains from Castle Street, Carlisle. Excavations 1981-2. Fascicule 1. Cumberland and Westmorland Archaeological and Architectural Society Research Series No. 5. Kendal

Rackham, D. J. & Gidney, L. J. 2009. The Animal Bone Report: Part 1. In H. E. M. Cool & D. J. P. Mason (Eds) Roman Piercebridge. Architectural & Archaeological Society of Durham and Northumberland Research Report 7: Digital Chapter 12.

Severin, T. 2000. The Brendan Voyage. The Modern Library: New York.

Steel, D. 1794. The Elements and Practice of Rigging and Seamanship. London.

Stokes, P. R. G. 2000. The butcher, the cook and the archaeologist. In J. P. Huntley & S. Stallibrass (Eds) Taphonomy and Interpretation. Oxbow Books: Oxford 65-70.

West, B. 1995. The Case of the Missing Victuals. Historical Archaeology 29 (2): 20-42.

Winchester, A. J. L. 2000. The Harvest of the Hills: Rural Life in Northern England and the Scottish Borders 1400-1700. Edinburgh University Press.

Comments are closed.