The Scientific Goal of Replicating Greco-Egyptian Perfume Recipes
Ground cinnamon (left) and myrrh (right). Preparation of Mendesian perfume. Photo by Sean Coughlin.
Our goal is not to become perfumers, but to figure out the chemistry behind the recipes. We believe we can use replications to help us do this. And we think this can add to our understanding of the material, cultural and scientific context in which these recipes were composed.
We can’t know ahead of time what a correct interpretation of a recipe will look like. In fact, there may not even be a correct interpretation of a recipe: like today, there were almost certainly countless variations for each perfume type, and the same recipe likely resulted in very different perfumes depending on the quality of the ingredients, the skill of the perfumer and the time and place it was produced.
That’s why even though studying the text of a recipe is an important first step, we still need to try them out, testing and documenting different interpretations to see what more general processes might be going on chemically. We therefore treat the replication of each recipe as we would if we were replicating a scientific experiment. The only difference is that the methods in recipes are often obscure to us and we need a method to fill in what is missing.
Recovering Missing Knowledge
Much like recipes today, ancient recipes take a lot technical knowledge for granted. Much of that knowledge and skill was probably passed down from master to student in the workshop; some of it was present in the apparatus and tools designed by craftspeople for the process; and some of it was present in the mind and memory of those who collected the aromatic resins, woods, fruits and flowers. This makes our job more difficult (and more fun), but we can still develop hypotheses about how to interpret or fill in what is missing.
Our hypotheses come from the different possible meanings for ancient Egyptian and Greek terms that we’ve compiled in our lexicon, and we supplement these with what we know from archaeological evidence and from perfumery techniques still in use today.
We also don’t stop at just one hypothesis. One of the unique aspects of our approach is that we test multiple such hypotheses to see what happens. We believe experimental replication of different interpretations stand a better chance at helping us understand the text, either by confirming or disconfirming a philological hypothesis, or by being consistent or inconsistent with archaeological or other historical evidence.
Chemistry as Reference Point
Replication does this by helping us to understand the chemical reality behind the texts. Organic compounds, because they act the same way now as they did in ancient Alexandria, are stable reference points for us when testing the effects of different ingredients and techniques. Using tools like GC-MS, spectroscopy, NMR, and our senses of smell, touch and sight, we observe, at least in part, what it would mean if one of our interpretations of the text were true.
The observation of different interpretations can help us to understand why our sources recorded the recipes they did. It therefore opens up a new way of exploring the material, cultural and scientific context in which they were composed.
Key sources for Techniques of Perfumery
We are currently replicating the procedures for five different perfumes. The recipes were chosen because they are key-sources for certain techniques:
They are preserved in Egyptian, Greek and Latin sources and are attested in the pre-Alexandrian period, either in Greek or Egyptian sources.
There are extant visual representations or other forms of archaeological evidence suggesting how they were prepared.
Their production is characterized by or known for employing a unique procedure.
The order of the experiments was chosen so that each successive replication addresses more complex and articulated research questions (more complex procedures with more ingredients) and incorporates the lessons from earlier experiments.
The Recipes
Myrrh resin (malmal), Somalia. Likely Commiphora myrrha (Nees) Engl. Photo by Sean Coughlin.
Year 1. Stakte. Heating Gum-REsin in Water.
In our first year, we began tests to replicate the production of stakte (στακτή). Stakte is commonly thought to be an oil derived from myrrh and some scholars have argued that it is similar or identical to an oil called mḏt in Egyptian sources, which was extracted from ‘ntyw (possibly myrrh). Our aim was to replicate the processes recorded in our sources for extracting the stakte oil by combining resin with water and pressing the sediment that results.
Heating oil in a bain-marie. Photo by Sean Coughlin.
Year 2. Mendesian. Refining vehicle Oils.
We are currently replicating the production of the vehicle oil of the Mendesian (Μενδήσιον) perfume. The Mendesian comes from the deme of Mendes (likely near present day Tell-Timai) and it came to be emblematic of Egyptian perfumes during the Ptolemaic period. One of the common features of the recipes preserved for this perfume is its use of an oil called balaninon (βαλάνινον ἔλαιον). The identification of this oil is disputed (see Littman, Silverstein, Goldsmith, Coughlin and Mashaly, 2021), but whatever it was, it had a distinctive way of being prepared for use in perfumery. Theophrastus tells us it was heated for ten days and nights so that it would become more receptive of whatever scent is added to it. We have replicated this procedure with several variations and are testing its effects on the composition and sensory characteristics of different oils used in Egyptian and Greek perfume production.
Maceration of myrrh and calamus in wine, a possible initial step in the styptic process of Metopion (in this case derived from susinum process). Photo by Dora Goldsmith.
Year 3. Metopion. stypsis (increasing the receptivity of oils).
Today, perfume is often described using a metaphor from music: individual scents are ‘notes’ of different registers (base, middle, top) and together they form a chord. Ancient Egyptian and Ancient Greek perfumers used different models to describe how the ingredients of a perfume are related. This suggested they understood something different about the process. One such model comes from mordants used in dyeing natural fibers. This is called stypsis, and in perfumery it was a way of making an oil receptive to a particular scent. Certain resins, roots and herbs were thought to change an oil or fat like a mordant acts on a fabric: it causes the oil to become both more receptive to the quality the craftsperson was trying to apply and also increase its fastness, it’s ability to retain the quality for a long period of time. Different final scents would require different styptic agents, just as different mordants can produce different colours. Our Metopion experiments explore this process. Metopion was one the most costly Egyptian perfumes to enter the Greek record after Alexander’s occupation of Egypt and well-known for its complex styptic process.
Relief showing lilies being collected, likely for susinum perfume. 6th century BCE. © 2009 Musée du Louvre / Erich Lessing. (permalink).
Year 4. Susinum. Maceration / enfleurage.
The susinum (σούσινον) or lily perfume is depicted in several ancient Egyptian reliefs and recorded in early Greek medicine—which includes the works attributed to Hippocrates—and philosophy. Recipes in the Greek tradition that still exist likely originate in Alexandrian medical tradition. The process was labour-intensive: it required over 3,000 lilies to be harvested and processed over a period of three days. We will be seeking partners to grow a large number of lilies in order to understand how the process would have worked and to capture the scent of the lilies in its stypic-treated oil.
Egyptian alabastron for holding oils etc. Ptolemaic Period. Wellcome Collection. Attribution 4.0 International (CC BY 4.0)
Year 5. Smoked Oil. to replicate An early distillation procedure.
This will be the most complex procedure. We will replicate what is likely an Alexandrian recipe preserved in a much later source (Aetius of Amida) for a perfume that uses a rudimentary form of distillation per descensum, a process common in later alchemical sources where the sample reservoir is heated from above and the distillate is below. Performing the replication of this recipe will involve testing designs for a complex apparatus based on archaeological and textual evidence . We will share more about this exciting procedure in the future.
