By Eirik Gralókkr (Michael S. Matthews)

Lamps that burned animal fat for illumination can be traced back to the Paleolithic, some 40,000 years ago (de Beaune & White, 1993). They represent an important advance in humanity’s technology, producing a controlled light source that could extend human activities into the evening and overnight hours of the day and even into permanently dark areas such as caves. In use, one or more wicks, typically made of plant fiber, rest in the fat and function like the wick of a candle does: they hold a flame and bring the melted liquid fuel near it, where in liquid or gaseous form the fuel burns to produce light and heat. This technology, with some variation, has remained in continuous use across much of the world across over the last fifty thousand years of human history.

While stone and clay lamps have been known from both archaeological and ethnographic contexts for over a century, it is only in recent years that it has become possible to analyze the organic residues present in archaeological finds to determine what fuels these lamps burned. Other scientific measurements on reconstructions also have examined their light production, fuel consumption rate, and other aspects of interest. In this paper I’ve distilled what I have learned about the history and construction of animal-fat oil lamps, with a focus on finds from northern Europe. To make this topic more manageable, I have purposefully avoided considering both the plant oil-burning lamps that were used widely in more temperate and tropical climates (i.e., Mediterranean or African contexts) and the few metal-bodied lamp finds (e.g., from the Oseberg ship burial) whose use likely was restricted to only the most high-status individuals. This excludes, for example, the many Roman-era lamps and related finds from Middle Eastern and Egyptian cultures that were fueled by olive oil or other plant-derived oils.

Paleolithic Lamps

The first Paleolithic lamp identified as such was found in 1902 in association with the cave art at La Mouthe, France; it consisted of a piece of burned sandstone with a hollowed out top and an image of an ibex engraved on its underside. In an important review, De Beaune and White (1993) identified some 285 lamps in museum collections, from 105 different Upper Paleolithic sites in France, dating between 11,000 and 40,000 BCE. These lamps are made of limestone or sandstone and all are fist-sized or smaller. The simplest style is essentially a rock that is naturally flat to concave on top, sometimes with grooves added so that the excess hot fat drains off rather than swamping the wick. These flat lamps likely were temporary expedient choices, a found rock that was minimally worked if at all, because this type wastes more fuel in use than more elaborately crafted examples do. Measurements using reconstructions of these Paleolithic lamps show that this style typically produced somewhat less light than a modern candle does.

The most common style of more deliberately constructed lamp from the Paleolithic features a simple shallow bowl hollowed out to hold the melted fat, sometimes with a groove in which the wick rested; lamps also may be more elaborately worked to include a handle and sometimes additional decoration (de Beaune, 2002). The bowl has gently sloped sides and is typically wider than it is deep. De Beaune and White (1993) noted based on their experience with reconstructions that sandstone lamps become too hot to hold when burning, so these tend to need handles. Preserved burn marks show that the wick on lamps with handles rested on the end of the bowl away from the handle, while on examples with no handle the wick rested in a different place around the rim each time it was used. Chemical analysis verifies that these lamps burned animal fat, most likely from horses, cows, or pigs, rather than oil from plant sources. The authors’ reconstructions showed that wicks of lichen, moss, or juniper (bark) were the most effective materials. These choices are consistent with both the very limited archaeological evidence (which includes these wick materials and also grasses—likely the fluffy tops of cottongrass, genus Eriophorum) and with more modern ethnographic evidence; for example, similar lamps with lichen wicks were still in use by Inuit peoples into the modern era (Hough, 1898).

Mesolithic to Neolithic Lamps

Moving to a somewhat more recent era of prehistory, Robson and colleagues (2022) documented the widespread use of ceramic, oval low bowls for lighting across a 2,000-year period, from the 6^th to 4^th millennium BCE, in the circum-Baltic region. A wide variety of names are used in different site reports for what on inspection are very similar-looking ceramics, characterized by oval/oblong/elliptical outlines and a shallow, slightly rounded to flat bottom. Some include handles or knobs on their sides. The dimensions of these lamps ranged from 6-30 cm in length and 2-15 cm in width, but less than 5.5 cm in height. 

Robson and colleagues (2022) reported that finds from the eastern Baltic tend to have incised decorations around the rim in the form of oblique hatching or incised parallel lines; they are made with clay that was tempered with crushed shells, plants, rock debris, and/or sand. Finds from the western Baltic traditions were less frequently decorated and their clay was tempered with feldspar/quartzite, lime, plants, and/or sand. Both eastern and western Baltic finds were made using pinching and/or coiling techniques, and both show carbonized surface deposits, often on the ends or rims, that reflect their use as oil lamps. Analysis of organic residues from 115 such bowls from 25 archaeological sites (Robson et al., 2022) confirmed that fats from marine, terrestrial, and freshwater animal sources were burned for illumination. The authors’ comparisons with cooking pot residues from the same sites also support the flat bowls’ use as oil lamps, as the bowl residues are higher in lipids (fats, oils, or waxes) and lower in nitrogen (indicating proteins) than the cooking pots’ residues are. Some chemical signatures also indicate that these residues had undergone extended heating. Lipid residues in both types of vessels suggest their origins “could conceivably include fish, birds feeding on aquatic organisms, aquatic mammals, including beaver and seal, and molluscs” (Robson et al., 2022, p. 37). These analyses suggest that people used a variety of local animals as both food and as sources of lamp oil.

Robson et al. (2022) note that the production and use of these clay lamps appears to have died out at around the same time that farming came into widespread practice in this region, though the same aquatic resources continued to be utilized as food (i.e., the sources of lamp oil remained available). Similar lamps continued in use farther north among Arctic peoples, who also made them of other materials including unfired clay, soapstone, and even perishable materials such as birch bark. The authors speculate that perhaps sedentary farm life had different lighting needs that may have been met in other ways, such as through the introduction of tallow candles.

Viking-age Lamps

Oil lamps are present in finds dating to the Viking age, approximately 800-1100 CE, both in England and in Scandinavia. Below I consider some finds from each area.

England

In eastern England (where there was extensive settlement by and interactions with Scandinavian peoples) pottery lamps were a common product prior to around the 12^th century, after which time they appear to have been replaced by lamps made of stone, metal, or glass, or by tallow candles. These lamps are less frequently found in western and southern England (Blinkhorn et al., 2017). Eastern English pottery lamps appeared in two different forms; both had a simple cup shape on top to hold the fuel and a wick, with bases formed into either a spike (called a cresset lamp, designed to be hung in a wire cage or pushed into soft earth) or a flat foot (a pedestal lamp, which could stand on its own on any available level surface). Pedestal lamps somewhat resemble a parfait glass in shape, but without a neck between the bowl and the foot. Both of these styles looked interesting, but I decided to try making the pedestal style (see Figures 1 and 2 in Blinkhorn et al., 2017 for cross-sectional drawings) due to not having a suitable way to mount a cresset lamp.

Blinkhorn and colleagues (2017) analyzed three pottery lamps excavated in 2012 at the garden of the Edward Jenner Museum, Berkeley, Gloucestershire, the site of a late-Saxon metalworking workshop during the 10th/11th centuries. Their analysis of the organic residue in the three lamp finds shows that two likely were fueled with ruminant fat (tallow) and the third with a mix of ruminant and non-ruminant fats (e.g., pork lard). Furthermore, the authors state that “from all three lamps we can confidently exclude beeswax, plant oils, aquatic fats (ie fish oil) and butter” as possible fuels (Blinkhorn et al., 2017, p. 111).

Blinkhorn and colleagues observed that this style of lamp often is found in the context of craft production workshops (with the exception of those working with textiles), rather than in homes or high-status places like churches or castles. They hypothesized that these lamps may have produced the best available artificial lighting for conducting craft work. In an example of experimental archaeology, they compared the light output from replicas of these lamps with that of another common form of lighting at the time, the rush light. Rush lights are prepared by stripping off the outer layer of dried stems of the common rush (Juncus effusus), dipping the inner pith into melted tallow, and lighting the end of the resulting tiny torch; rush lights are considered to have been the least expensive type of portable artificial lighting available during this period. Blinkhorn and colleagues compared rush lights to a replica oil lamp with three wicks. The rush lights burned for 17 to 30 minutes each, with a “flickering light of generally low intensity, although with short-lived bursts of brighter light” (p. 113) and required regular adjustments, while the oil lamp produced a more consistent and brighter light and had enough fuel to continue burning for multiple hours on a single fill. The authors noted that three candles would be required to match the output of a single three-wick lamp, and that candles would cast more shadows. The oil lamp would have been more stable in use than the rush lights were, though the authors suggested it would be less stable than a candle whose holder was affixed to the wall. They further suggest that this might explain why these oil lamps are found less frequently in households than in production centers. The authors concluded that the oil lamp would have produced the best illumination under which to conduct industrial or craft activities.

Baltic Region

In the eastern Viking regions around the Baltic, extensive interactions occurred between the region’s varied populations over a period of many centuries. Trade in northern animal furs existed already during Roman times and it experienced a resurgence from the Migration era through the Viking age and into the early medieval era. For example, excavations at the Viking trading town of Birka (occupied from the 8^th century through the 970s) have found hundreds of paw bones from the processing of fur-bearing animals (Callmer et al., 2024). The authors also note that decorative Finnish pottery is found consistently at Birka over its history and across Swedish sites in the Mälaren region until the mid-tenth century, though I have not found reports specific enough to identify any oil lamps that may be present among these assembled finds. I learned from various sources that locally produced pottery is uncommon but not completely absent from lower Scandinavia, apparently mostly due to a lack of usable clay deposits in much of the region.

Birka, near present-day Stockholm, was an important trading center from circa 750-960. Numerous excavations have been conducted in this area since the late 17^th century CE, first to locate the settlement and then later to examine some of the thousands of grave sites and other structures identified there (these include, most recently, a shipyard—see Isaksson et al., 2022). However, many earlier Birka finds were not as well documented by their excavators as more recent ones are, making it difficult in many cases to establish their specific context. One such find is a ceramic oil-fueled lamp from Birka in the shape of a small bowl with a central post (Historiska museet/SHM, object #270759_HST). This interesting find is the other piece I chose to recreate.

My Lamp Project

After reading about the history of these oil lamps, I decided to try making ceramic lamps following each of these two Viking-age finds. This first required learning a bit more about clay and ceramic technology, before I could start getting my hands dirty.

Pottery and Ceramic Technology

Clay

Clay consists of very small particles of less than 0.002mm in diameter (< 2 µm; particle size decreases going from gravel, to sand, to silt, to clay). These particles consist of highly weathered minerals, mostly hydrous aluminium phyllosilicates such as kaolinite, Al₂Si₂O₅(OH)₄ (https://en.wikipedia.org/wiki/Clay) and usually with very little organic matter present. When heated to a high temperature clay transforms into a ceramic, removing chemically bound water molecules and locking the particles together into new crystalline forms (ceramics) such that their aggregated mass can no longer be dissolved in water. Humans have used clay to make pottery for thousands of years, doing so in nearly every part of the world where suitable clay deposits are found.

Firing Temperature

The temperature range at which the transformation from clay to ceramic structure occurs is in the range of 1,700°F to 2,100°F (900°C to 1,150°C), a range described as Cone 04- 6 (Kiln Frog, 2023). However, the transformation is also time dependent such that it can occur at temperatures as low as 1,112 °F (600 °C) (https://en.wikipedia.org/wiki/Pottery#Firing). This initial transformation produces a porous to semi-porous ceramic known as earthenware or terra cotta. These lower temperatures are fairly easy to achieve in a well-built campfire. Midrange firing temperatures produce less-porous stoneware at 2,100°F to 2,300°F (1,150°C to 1,260°C) or Cone 6 -8, while high firing temperatures of 2,300°F to 2,700°F (1,260°C to 1,480°C) or Cone 10 produce non-porous vitreous stoneware or porcelain (Kiln Frog, 2023). Midrange and higher firing temperatures are more difficult to achieve, such that they typically require firing in a specially designed kiln structure; they also may only be successful with some clay bodies, e.g., those higher in kaolin, as other clays will melt at these higher temperatures. While there are a few clay deposits in the southeastern U.S. that were used successfully in recent historic times for high fired ceramics, these are both uncommon and are inappropriate for the vessels I aimed to recreate here.

Tallow and Wicks

Since I was making the lamps from foraged materials, I decided to also make the wicks from material I gathered. While I have read about and even helped process flax into linen, I had a tiny scrap piece of hand-woven linen left from another project that I was able to reuse for this purpose. I gathered juniper bark from a downed Eastern red cedar tree (a species of juniper) that grew in my area.

Tallow. Tallow is animal fat that has been rendered, converting it from a solid to a lower-melting liquid form and making it more stable for extended storage at room temperature. Tallow generally refers to fat from cows or sheep but may also be made from other animals (e.g., horses); a similar process produces lard from pig fat. Tallow has been used for lighting for thousands of years in lamps, candles, and rush lights, and for other uses such as skin care and (in more recent times) in the production of soap and for deep frying food. The more desirable tallow for human consumption comes from around the animal’s internal organs, though its origin within the animal is not a concern with tallow used for lighting or other purposes. Suet or raw fat is converted to tallow via extended heating at a low temperature, and/or by boiling with water to extract undesirable components. The rendered fat floats on the water, making it relatively easy to separate from the other components.

Wicks. Used as a fuel for lamps, tallow requires a wick just as would be used in a candle. Several plant-based materials are documented historically as wick materials. For these lamp reproductions I have used linen (made from flax) and juniper bark.

Sourcing Materials and Lamp Construction

This was my first-ever attempt at making pottery. Last summer I dug out a few shovels of yellow-green clay from my front yard while planting a tree. I mixed this clay thoroughly with water to separate any foreign matter, of which there was very little. Then I left it in a container for several months to settle and dry back out to a workable consistency. 

My readings suggested that in the archaeological finds either sand or organic matter was used to temper the clay; temper reduces the stresses in the material during firing (and hence, reduces the likelihood a piece will crack or break) by introducing internal voids. I collected some sand from a dry drainage ditch down the street from where I live. I also collected and coarsely chopped a handful of seed hulls from underneath my bird feeder. Since I was not sure which temper would work better with my clay, I made a batch with each kind. 

Lamp Construction and Firing

Once the clay was dry enough to work, I mixed a few handfuls of clay with approximately 20% by volume of seed husks (first batch) or sand (second batch). I formed coils of the tempered clay by rolling a blob between my hands and attached the resulting coils around the edge of a flattened piece, pushing the wetted sticky edges together and working them with my fingertips until the joint was no longer visible. Working in this way I added coils until I reached the desired height. I then added a short fat coil, on end, inside the Birka lamp in the same manner. The Saxon-style lamp does not have a central post, but it required some additional working to form the foot pedestal and to flare the top edges to match the archaeological finds.

I dried the lamps for a few weeks indoors, where they benefitted from the dry winter air. Firing was accomplished in my backyard fire pit using wood that I had cut and stacked to dry last year. I built a large fire and let it burn down to a deep bed of coals, putting the dry clay pieces nearby but outside of the flames and rotating them every minute or two until they were almost too hot to touch. I then set a metal grill on the bed of coals (to hold the pieces upright), set the pieces on the grill, and covered them with a large metal mixing bowl. In period they probably would have been covered with scrap shards from broken vessels to protect them from being damaged by falling wood as the fire above burned down, but as my first attempt, I didn’t have any shards so using the bowl seemed like a reasonable substitution. I then built up a large fire around and over the top of the bowl and let it burn down to ashes over a period of half a day before removing the fired work. The lamps after firing had transformed from the grayish color of the raw clay to a pleasant light reddish color, with a consistency resembling a common clay flower pot (which also are terra cotta).

Fuel and Wicks

I originally intended to purchase some tallow at a local farmer’s market, but it was sold out everywhere I checked. So instead, I bought some beef at the grocery store and trimmed off all the marbling, then baked all this fat overnight at a very low temperature (250-275F) to render it into tallow. The next morning, I strained out the solid pieces and filled the test lamps with the still-warm liquid tallow. 

I sewed a linen wick from a tiny scrap of hand-woven linen left over from a different project for the Birka-style lamp. I used rolled juniper bark I collected to make three wicks for the Saxon-style lamp. In a process like that used for making a tinder bundle, one rolls the dry peeled bark back and forth between the palms until it is all finely shredded. Then, the shredded bark is twisted into a loose 2-ply cord similar in appearance to a fat piece of yarn.

Testing

As the pictures show, both lamp styles and both kinds of wick material produced a strong, steady flame consistent with the findings from experimental archaeology. As such, I consider this project to be a success. 

Though some event sites no longer allow candles or other open flames, it would be instructive to try working on some other A&S projects indoors at night by the light these lamps produce, to see how well they serve this purpose. I find that this project has provided me with some unique ways to feel more closely in touch with the experiences of our historical ancestors, for whom this kind of relatively steady and consistently bright illumination must have represented a welcome technological advance.

Process Photos (all by the author):

Raw and very gooey clay

Chopped seed husk temper

First attempts

Second batch, sand tempered clay with a pot and a coil in progress

Second attempt, a more accurately shaped pedestal-style lamp

Beginning, middle, and end of the firing process

Fired pieces after the fire has died out. The small flatter lamp is based on a soapstone lamp from Jarlshof in the Shetlands that dates to just after the Viking era (National Museums Scotland, X.HSA 754)

Fitting a tubular linen wick sewn from scraps

Tallow before and after rendering, and strained but still warm and liquid

Juniper bark on the log

Three juniper wicks in use in the English pedestal-style lamp

Large flame from a central linen wick in the Birka-style lamp

References

Blinkhorn, P., Cramp, L., Prior, S., Glass, E., & Horton, M. (2017). Fiat lux: Functional analysis of three Saxo-Norman pottery lamps from Berkeley, Gloucestershire. Medieval Archaeology, 61(1), 104–116. https://doi.org/10.1080/00766097.2017.1295911 

Callmer, J., Gustin, I., & Roslund, M. (2024). Baltic Finnic and Scandinavian social interaction in the fur-hunting North, c AD 700–1200. Medieval Archaeology, 68(2), 331–357. https://doi.org/10.1080/00766097.2024.2427530 

de Beaune, S. A. (2002). Palaeolithic lamps and their specialization. Bulletin of Primitive Technology, 23, 60-67. https://shs.hal.science/halshs-00803851v1 

de Beaune, S. A., & White, R. (1993). Ice Age lamps. Scientific American, 268(3), 108–113. http://www.jstor.org/stable/24941409

Hough, W. (1898). Origin and range of the Eskimo lamp. American Anthropologist, 11(4): 116-122. Available: https://doi.org/10.1525/aa.1898.11.4.02a00010 

Isaksson, S., Fjellström, M., Kalmring, S., & Holmquist, L. (2022). En vikingatida varvsplats vid Kugghamn, Birka : Arkeologiska undersökningar av L2022:2719, Birka, Björkö,Adelsö socken, Uppland, augusti 2020 och 2021 [A Viking Age shipyard at Kugghamn, Birka]. Retrieved from Arkeologiska forskningslaboratoriet website: https://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-206390

Kiln Frog (2023, July 24). Pottery vs. Ceramics | What’s the Difference? [Blog post]. https://kilnfrog.com/blogs/frogblog/pottery-vs-ceramics

Robson, H. K., et al. (2022). Light production by ceramic using hunter-gatherer-fishers of the circum-Baltic. Proceedings of the Prehistoric Society, 88, 25-52.  https://doi.org/10.1017/ppr.2022.12 

A selection of oval bowls dating to the Ertebølle culture of the western Baltic, circa 6500-6000 BCE (Fig. 3 from Robson et al., 2022)

A hanging soapstone oil lamp from Jarlshof in the Shetlands:

National Museums Scotland, X.HSA 754 

This steatite lamp, now cracked, was found at Jarlshof in Shetland and the museum says it was probably carved from local stone. The lamp was used between 1100 and 1300. It is perforated on both sides of the bowl, allowing it to be hung. The bowl of the lamp would have been filled with oil on which a wick floated.

Dimensions: 5.10″ L x 2.80″ W x 1.30″ deep; cup 1.80″ D x 0.70″ deep

http://nms.scran.ac.uk/database/record.php?usi=000-100-102-595-C

One of the three similar footed-style pottery lamps excavated from Jenner garden, Berkeley (SF23) and dating to the 10^th to 11^th century CE. Light brown ceramic with a dark grey core. This drawing is Figure 1 in Blinkhorn et al. (2017).

Ceramic oil lamp from Birka with a central wick post:

Burnt earth oil lamp. Settlement find, Svarta Jorden, Birka, Adelsö Parish, Uppland.

Historiska museet/SHM 

Object number: 270759_HST

• Weight 0.229 g
• Diameter 80 mm
• Height 45 mm
• Dated 800-1100