Boiling hot oil: on the assessment of temperature in late medieval processing of linseed oil

By Indra Kneepkens

Indra Kneepkens is a technical art historian, specializing in the materials and techniques of late medieval panel painting. She is currently finalizing her dissertation, which is focused on the use of processed linseed oils and paint additives in the painting practice of the fifteenth and early sixteenth century.

Raw linseed – the basis for linseed oil

From art technological sources, such as recipes and manuals for the preparation of paints, and from the analysis of paint samples, we know that late medieval and early modern craftsmen heated oils for use as a binding medium in paints, as well as for the preparation of varnishes. As a technical art historian, I research and reconstruct these oils and varnishes, to be able to establish the effects they have on paints. My aim is to understand how the development and use of these materials influenced the painting process and the final appearance of art works. Unfortunately, 14th-  to 17th-century art technological sources that include heat-treatment of oil typically do not give absolute temperature indications. This is no surprise, because thermometers were developed in different stages between the late 16th and early 18th century. Therefore, craftsmen had to rely on their senses for the assessment of temperature.

Cennini’s handbook has been transcribed an translated repeatedly, most recently by Lara Broecke in 2015

That controlling temperature was important to late medieval craftsmen is illustrated by a remark in the Libro dell’arte (ca. 1390), that was written by the Italian painter Cennino Cennini. In his recipe for a heat-treated oil the writer almost personifies fire, warning us that it would willingly burn down the house if it got a chance to reach into the pan.[1] And indeed, it seems likely that the temperature of oil was sometimes raised pretty high, to the point where potentially explosive gasses form. So how did pre-thermometer craftsmen determine the temperature of their oils?

Even if not in absolute terms, early modern sources do contain clues about the kind of temperatures that were considered useful in oil processing. In most cases, oils were heated over fire, in ceramic, bronze or copper vessels. As wood or peat fires can easily reach temperatures over 500°C, and the vessels were able to withstand these, it was clearly possible to reach the auto-ignition point of linseed oil (ca. 343°C). Most likely, temperature would have varied from case to case, depending on factors like fuel type and the distance between the vessel and the heating source.[2]

Sometimes sources mention a strong reduction of the oil, or they describe the texture of the end result, comparing it to fluid honey for example. In varnish recipes a string-test reoccurs, in which a finger is dipped into, and then lifted from a cooled drop of the mixture of boiled oil and resins, to see if a thread is formed.[3]My own experiments have shown that to reach a sufficient thickening of an oil or varnish, and to dissolve most resins, the materials must be heated for quite some time and at a fairly high temperature. Finally, empirical tests that remind us of everyday cooking may have been used to keep an eye on the temperature of oils. In Arte de la Pintura (1649), Pacheco describes how bread, garlic and feathers were stuck in boiling oil; if they appeared browned or scorched, the cooking process was completed.[4]

During my experiments over the past few years, it has become evident that one can make a reasonably accurate temperature assessment based on how the oil behaves during heat-treatment. A swirling movement for example, was noticed in oils from a temperature of ca. 90°C, while repugnant fumes typically started to develop around 200°C.

Figure 1. Foam developing on oil heated in a ceramic vessel

I also realized that the development of foam, which is commonly mentioned in recipes, may have been a clear indication of temperature. But it is also closely related to the type of cooking vessel that was used. I did not see any foam until I exchanged my glass laboratory beakers for a glazed ceramic pot. Foam appeared when the oil reached a temperature of 100°C. It must have been caused by water evaporating from the clay body, as I had rinsed the vessel before use. Bubbles also formed when garlic and bread were boiled in oil, from ca. 61 and 80°C respectively, with a high point again around 100°C. Although the garlic started to darken earlier, both bread and garlic clearly browned around 180°C, and completely blackened between 230 and 240°C.

In another experiment thirteen feathers, of different size and from various birds, were dipped in hot linseed oil until they started to curl. Surprisingly, all feathers curled within a range of 27°C, between 237 and 264°C. These experiments not only support the relative accuracy of these empirical methods, but also indicates that temperatures of ca. 100 and 200-250°C were meaningful to early practitioners.

Figure 2. Raw linseed oil (far left) heated at 150 (1-4 hours, jars 2-5 from the left), 200 (8 hours, larger jar) and 300 degrees C (1-4 hours, jars 7-10 from the left).

In more systematic tests, linseed oil was heated to 150 and 300°C, for one, two, three, and four hours. Interestingly, the color and thickness of the oils that were heated at 150°C appeared more or less unaltered after the experiment. In tests where these oils were mixed into paints, they behaved very similar to raw linseed oil. At 300°C the oils thickened and darkened considerably. They affected paints in a very significant way, causing them to level, and making it possible to create smooth glossy films without any visible imprint of the brush. Paints with these oils were also less prone to yellowing. An oil that was heated at 200°C for eight hours however, still made paints level perfectly, but it also caused extreme yellowing.

So knowing how to assess temperature and balancing it over time must have been crucial skills for those who prepared heat-treated oils and varnishes. Experiments have shown that knowledgeable individuals would have been able to make a fairly accurate assessment of temperature, using their senses. They could note changes in the appearance and behavior of oil and indicator materials, and manually test its viscosity. Although there are several indications that temperatures of 200°C and higher were preferred, it makes sense that craftspeople would adapt the temperature to the materials at hand and the desired end result. The lack of unambiguous temperature indications in oil processing recipes reflects this adaptive use of temperature and a reliance on the senses that was expected of craftspeople before the invention of thermometers.

[1]Broecke, Lara. Cennino Cennini’s Il Libro dell’ Arte: A new English translation and commentary with Italian transcription.London: Archetype, 2015, 127, chapter 91.

[2]Aldeias, Vera, Harold L. Dibble, Dennis Sandgathe , Paul Goldberg, and Shannon J.P. McPherron . “ How heat alters underlying deposits and implications for archaeological fire features: A controlled experiment” Journal of Archaeological Science67 (2016): 66.

[3]Broecke 2015, 127. Neven, Sylvie. The Strasbourg Manuscript. A Medieval Tradition of Artists’ Recipe Collections (1400-1570). London: Archetype, 2016: 132-135, no. 92-94.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.