Recreating Ancient Beauty

By Eboni John, published as part of the Undergraduate Series

The society of ancient Rome was just as obsessed with cosmetics and beauty as we are today. Indulging in the use of items such as white lead foundation, ash-based eye-shadow and poppy petalled blush, it is clear that the Romans’ idea of perfection centred around pale skin, large dark eyes and blushed cheeks (see e.g. Pliny the Elder, Natural History 21.73.123, 35.56.194). However, what these cosmetics may have achieved for beauty, achieved very little when it came to health. This is because the Romans used all sorts of deleterious ingredients in their cosmetics such as crocodile dung, nightshade, urine and the infamous lead – to name a few (Olson, 2008, 61-72). What exactly then, are the possibilities of recreating some of these ancient Roman cosmetics today?

To answer this question, I will be covering the recreation of two Roman cosmetics:
1) A 2000-year-old Roman face cream used for coverage
2) A Roman beauty mask used to soften and in some cases, whiten the skin.

Roman Face Cream

The Roman face cream dates from the 2nd century AD and was discovered in London in 2003. Due to the container’s expensive nature (tin was a relatively precious metal at the time) it is thought that the cream was used by a Roman aristocrat, with a function similar to that of modern foundation.

The ingredients for the cream were revealed by Katherine Mansell: it contained 40% animal fat (likely derived from a goat then boiled); 40% starch (likely obtained from boiling wheat or roots); 20% synthetic tin oxide (cassiterite). The starch would have been added to the fat ‘to reduce the greasy feeling of fat on the skin’. The tin would have then made the cream white (Mansel, 2004). 

It is surprising that this cream does not contain the ingredient lead, which was frequently used in cosmetics at the time. This suggests that the Romans were possibly becoming more aware of the lead poisoning that plagued their cosmetic products, as this particular chemist appears to have identified tin as being a non-toxic ingredient.

When visually comparing the two products is it clear that the colour is one of the largest differences, with the recreation displaying a brilliant white in comparison to the originals faded grey. The texture also looks a lot less crusted and granular than the original. While the cream did not smell entirely pleasant, it did provide a very adequate form of coverage.

Roman Face Mask

A Roman face mask to soften the skin was a must-have when it came to skincare. The ingredients for this mask are provided by Susan Stewart:


Almond oil; rosewater; water parsnip (boiled); lily root (ground into a fine powder using mortar and pestle); eggs (Stewart 2007, 32-60).

I found that the smell of the face mask was rather pleasant, but those who sampled my mask said otherwise. The final products colour of light beige was exactly how I visually imagined it, but the texture turned out to be runnier than I expected. Upon sampling the mask, my volunteer reported that it did make her skin feel slightly softer but left an oily residue on her face that failed to wash off immediately with water.

Sourcing Ingredients and Substitutions

In preparation for recreating these cosmetics, I had to acquire the right ingredients. Unfortunately, particular ingredients proved rather hard to come by and so have been substituted.

1. Animal fat: Originally the fat would have probably been derived from a cow or goat but the only animal fat I could come across was goose fat.
2. Synthetic tin oxide: This ingredient was not a cost-efficient one to come by as it was often sold in bulk. I have, therefore, substituted it with zinc oxide which I was able to find in small quantities for a reasonable price. I also felt safe in my knowledge that it was a safe ingredient as it is used in many modern cold creams.
3. Lily root: I found it was not easy to come across lily root as it is not something typically sold in supermarkets or online. So, I substituted it with powdered orris which is typically found in perfumes.

Because I have chosen to use substitutes, these replicas cannot be considered exact recreation of the originals. However, I do believe that the cosmetics I have prepared hold some resemblance to the originals I have attempted to recreate. After all the ancients were no strangers to substituting ingredients as preserved substitution lists have shown.

From this experience of recreating ancient Roman cosmetics, I have found that it is no simple or easy task. The difficulty is mostly derived from acquiring the right ingredients. I often found myself stopping to consider the authenticity of the ingredients I thought to use. For example, I had to stay clear of starches found in pasta, as it would not have been an available form of starch at the time.

We must also keep in mind that these two recipes were not particularly hard or dangerous to follow, yet I still found myself substituting the original ingredients for those that were more available. Therefore, if you were to attempt recreating some of the more complex cosmetic remedies, the difficulty of acquiring the authentic ingredients and the risk of encountering hazardous ingredients is sure to increase along with the complexity.


My name is Eboni Alis John. I am 22, and a recent graduate of Cardiff University where I studied English Literature & Ancient History. I am a book fanatic that has always been keen to travel and write about my experiences. After writing my ancient cosmetics blog post for my third-year Greek and Roman medicine module, I was totally inspired by the level of creativity it allowed for. Since this assignment, I have begun planning to create my own blog that will focus on travel advice and my experiences exploring the countries of eastern Asia (Thailand, Vietnam, Malaysia). I am now looking to start a career in teaching abroad as this is what I believe truly fuels my passion.


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.