Article more suitable for mature students
Find out more18th Jun 2012
Mexicolore contributor on the Aztecs Raul Ybarra
This article has generously been specially written for us by Raúl Ybarra, M.Sc., an independent researcher and metalsmith with a special interest in pre-Hispanic jewellery casting.
Introduction
The ‘lost-wax’ technique of jewellery casting began in the northern half of South America, in Colombia, Peru and also Costa Rica, and reached Mesoamerica between 800 and 600 BCE. ‘Lost-wax’ casting is the most complex of all jewellery-making techniques as it involves a series of crucial stages, mistakes during any of which can affect the quality of the end product. Pieces of jewellery made with this technique are larger in volume, design and texture than those obtained using any other method.
In pre-Hispanic societies, any metal used in jewellery had a mystical significance, tied to its transformation, continuity and regeneration, rooted in ancient views of the origin of the universe. Each design would be associated with magical powers that would be transferred to its bearer during rituals and ceremonies. In Mesoamerica small bells were the most widely produced pieces of jewellery, and incorporated a rich variety of shapes, depicting animals, plants, gods and mythical beings.
Whilst many anthropological interpretations are nowadays placed on the shapes, textures, lines, motifs and so on that were used in the past to decorate jewellery, it’s important to consider too the functionality of these elements from the technical point of view of the goldsmith himself, as they both allow and guarantee superior results after casting. This can be seen clearly in the example of string bells from the Tarascan region of central Mexico: here, a wax thread is placed downwards from the base of the ring from which the piece hangs to the edge of the bell’s aperture. This thread serves three purposes: 1) to keep the threads together, to prevent them separating, 2) as a secondary casting channel, ensuring that the liquid metal advances more quickly inside the casting mould, and 3) as a decorative element, lending balance and harmony to the entire piece.
It’s more than likely, therefore, that in the mind of the pre-Hispanic goldsmith, creativity and technical procedures went hand-in-hand during the fashioning of a jewel: he had to juggle simultaneously cultural and religious elements alongside the important physical details that would allow the piece to be cast in metal (its thickness and size, melting temperatures for both metal and mold, etc.).
Today, our most important source of information for pre-Hispanic jewellery manufacture is the descriptions contained in the Florentine Codex, compiled between 1548 and 1561 and provided by local craftsworkers for Fray Bernardino de Sahagún. Seeing in this document the simplicity of materials needed to produce a jewel (clay, charcoal, beeswax, copal incense, tubes...) one can’t help but wonder if Aztec goldsmiths didn’t possess some ‘secret to ancient casting’ in order to create such fine and delicate pieces of jewellery – which today can only be matched with the aid of modern technology. Indeed, what method today could be considered the ‘Holy Grail’ of pre-Hispanic goldsmithery?
The literature on pre-Columbian metallurgy contains few studies by writers who have experimented with ancient techniques of casting jewellery using the lost-wax method, and unfortunately a few of these have introduced modern equipment and/or materials into their research, making it hard to draw firm conclusions from their results as to exactly how things were done in ancient times.
Aware of this, I have carried out tests designed specifically to reproduce as faithfully as possible the conditions of a pre-Hispanic goldsmiths, basing my work on the images and writings contained in the Florentine Codex, and on a close study of the shape, size and texture of original bells, in order then to reconstruct the casting of a number of different shapes – from charms to bells – using copper and bronze metals.
Apart from establishing an experimental workshop (Pic 4), I carried out research on the shape, size and texture of a number of original pre-Hispanic jewellery pieces exhibited in Mexican archaeological museums (in Mexico City, Michoacán, Oaxaca, Colima, Nayarit and Veracruz), with the aim of integrating the information obtained into the design of the test pieces, and of determining some of the key casting factors – such as the size and thickness of the moulds, the proportions of the casting channels, etc.
The experimental stages followed in making the test designs, based on the ancient lost-wax formula of creating a wax model that is subsequently turned into metal through the process known as ‘casting’, were these:-
a) Mixing of beeswax with copal incense
b) Forming the core in clay and charcoal
c) Creating the design in wax
d) Making the clay mold
e) Inserting the casting channel
f) Forming the casting funnel
g) Removing the wax and firing the cast
h) Casting and pouring of the metal.
a) Mixing of beeswax with copal incense. According to the Florentine Codex, the raw material for jewellery-making in pre-Hispanic times was beeswax and white copal (incense). We found in earlier tests in our workshop that the addition of just 1% of white copal to the beeswax was sufficient to give the latter the key properties of malleability (able to be hammered into sheets), stretchability and compression needed for successful jewellery design. (Concentrations of 5% or more render the wax too soft to be worked into models). We used the 1% formula in the current research.
b) Forming the core in clay and charcoal. In order for the figures to be hollow and not solid after casting, a clay and charcoal core has to be created, the same size and shape as the final object. The cores used in our work comprised 40% coal dust and 60% clay.
c) Creating the design in wax. After preparing the beeswax and copal incense mix, the designs were created based on archaeological examples from different pre-Hispanic regions of Mexico. In such cases the wax thread was wound round the clay and charcoal core.
d) Making the clay mold. Step 1: a semi-liquid mixture of coal dust (50%) and clay (50%) was applied to the surface of the finished wax designs. Step 2: a semi-solid mix of the same composition was applied on top to form the mold. Step 3: the molds were left to dry for two to four days. The molds were made following the copa style of original molds found in a tomb at Monte Negro, Quindío, central Colombia and described by Dr. Olsen in her 1972 study Two Prehispanic Cire Perdue Casting Molds from Colombia.
e) Inserting the casting channel. Once the molds had dried, the cylindrical shaped wax casting channels were added to each design. Ventilation channels were not used.
f) Forming the casting funnel. The mold’s funnel was then made of charcoal and clay and placed around each design’s casting channel, to direct the metal inside (Pic 9).
g) Removing the wax and firing the cast. Once dried, the molds were placed in an oven for two hours, for the wax inside to melt and for the clay to be fired.
h) Casting and pouring of the metal. The metal used to form the designs was pure copper and bronze (copper with 5% tin). The metal was melted in a spoon-shaped crucible made of the same materials as the molds. The crucible and metal were placed in a clay brazier, into which air from the lungs was blown through several reed tubes with a nozzle at one end, allowing the air to be introduced in amongst the coals, and so raising the temperature of the brazier. Once the metal had melted, it was poured inside the mold, which could subsequently be opened to determine the quality of the jewellery piece.
Results
Of 45 castings carried out (19 in copper and 26 in bronze), 32 complete designs (12 in copper and 20 in bronze) and 13 incomplete were obtained (Pic 11). One reason for some incomplete results was the temperatures of the clay mold and/or the metal at the point of pouring being too low. We also observed that some finished designs bore spots of excess metal on their surfaces, caused by metal filling a space accidentally created inside the mold when it was made.
Observations and Conclusions
When the designs were being modeled in wax we were able to observe that the technique of wrapping the clay core in wax threads proved the most efficient and effective in producing pieces of consistent thickness and shape. If a design with a smooth surface was required, all that was needed was for these threads to be rubbed down to achieve the desired effect. This matches our observations of jewellery designs and pre-Hispanic bells with smooth surfaces, in each of which faintly visible in the metal were the original marks made by the wax threads.
The absence of ventilation channels (or signs of these) was one of the most important characteristics found in the designs on display in different museums. This helped us gauge the experimental ratios of charcoal and clay to use in the molds in order to achieve the right level of porosity (empty space) which would lead to complete designs in finished metal.
The absence of a bracket or support for the clay core was found not to affect its internal strength, and complete designs could be achieved since the semi-watery covering of charcoal and clay, on coming into contact with the exposed part of the core, formed a temporary support after the drying stage. This underlines, albeit indirectly, the importance of using this type of mixture in the original process; it also encourages the detailed reproduction of textures in the metal design.
Metal-pouring tests were carried out placing the molds vertically (90°) and at an angle (45°), as well as employing a metal reservoir. The results showed no significant differences. These tests were made since much stress is placed in the academic literature on placing the mold at an angle during casting, to allow gases and the air contained inside the mold to escape. Our tests showed that in order to achieve finished pieces, given a mold with sufficient porosity, neither the angle of casting nor the presence of ventilation channels was found to be crucial.
We noticed that the most important factor in obtaining finished designs was the temperature of the mold and of the metal at the moment of casting: any variation of temperature would lead to failure due to the premature solidifying of the metal within the mould.
Studies of the structure and form of pre-Hispanic jewellery and bells, alongside detailed analysis of the Florentine Codex, proved crucial for understanding more of just how these designs could have been made, and for succeeding in reconstructing some similar pieces.
Our results confirm that it is possible to cast jewellery successfully using the lost wax technique, without the need for modern equipment, material or technologies, and that the Florentine Codex contains all the technical details needed.
Finally, it’s clear that further studies are needed to fully explain the manufacturing processes involved in the different styles of jewellery (earrings, charms, pendants, rings, etc.) made by our pre-Hispanic gold and silversmiths. At the present time we still cannot claim that the secret of ‘ancient casting’ has finally been revealed.
Picture sources:-
• All images from the Florentine Codex scanned from our own copy of the Club Internacional del Libro facsimile edition, Madrid, 1994
• Pic 1 and pic 2 (image from the Codex Tepetlaoztoc): photos by Ian Mursell/Mexicolore
• All other photos courtesy of Raúl Ybarra.
NOTE: For a full bibliography, click on the Spanish language version of this article, below...
Piers Watson
5th Nov 2019
Hello, I would still very much like to speak with Mr Raúl Ybarra if possible. Is there any way you can send me his contact details or let him know that I’d very much like to speak with him? Thank you, Piers
Mexicolore
We seem to have lost touch with him, which is a shame. We’re sending you separately the latest email address we had for him.
Emiliano
4th Mar 2019
Hola: Felicidades por su sitio. En torno a este artículo tengo una duda: la abeja que se utilizó en los experimentos era de Apis mellífera o de alguna otra especie de abeja (mellipona)?
Mexicolore
Gracias por tu mensaje. No tenemos suficiente conocimiento de este tema, así que le hemos escrito al autor Raúl Ybarra, pidiendo su ayuda...
Piers Watson
22nd Oct 2018
I would like to contact Mr. Raúl Ybarra but his website appears not to be working. I am also experimenting with casting copper bells using clay moulds and charcoal fires. Do you know how I can contact him? Thank you
Mexicolore
Sorry a) for the long delay in responding and b) that Raúl’s website is down. We’ve written to him to seek clarification... On getting a reply from him we’ll put you directly in contact with him.
Tania Cortes
18th May 2017
Hi. On Florida beach found several hundred pieces round metal, various sizes/ thickness resembling coins. Have lots of sediments.Some silver traces; most 99.8% iron.core very hard though.Could be molds;seem to contain native signs. I’m a lay person having problems identifying. can send you pics &/or mail phys samples if you share findings/results
Mexicolore
Please do send us images and we’ll try and give some guidance...
Mexicolore contributor on the Aztecs Raul Ybarra