metal, slag, ceramics | archaeometallurgy
Nov 4 2018

Reconstruction of earliest bronze church bell is ringing for the first time

Bastian Asmus

The Carolingian bell is ringing

This reconstruction is unique,

because this bell is not only cast in the shape of the Canino bell, as it has already been done for some decades on different occasions and for scientific investigations. See for this the various excellent works of the colleague Hans Drescher . No, this reconstruction is to be made as a modern original, following the same instructions of Theophilus Presbyter concerning bell making; using the same bell metal and  the same moulding material. In short to do it, as it was supposedly done in the olden days.

This reconstruction in accordance with the production process tries to reconstruct and implement the original production process in such a way that the resulting object does not differ from the original not only in material, shape and form, but also in the way of its production.

Bell tower for the beehive bell on the Galli campus. Tower, yoke and bell can now be tested for their suitability for the coming years.

Bell tower for the beehive bell on the Galli campus. Tower, yoke and bell can now be tested for their suitability for the coming years.

Applied Archaeometallurgy

As the name suggests, applied archaeometallurgy deals with the applicability of archaeologically, historically and scientifically informed interpretations. The aim is to reconstruct a process that can be applied in practice and that could have been implemented technologically for the epoch in question. It is thus closely related to experimental archaeology. In contrast to this, it additionally formulates the wish for applicability and practicability in the sense of pragmatically acting craftsmen. In addition to the pure functioning of a method, they also have to take care of other aspects such as the production costs, raw material availability, raw material procurement, livelihood and sales of their products.

The Campus Galli Bell, the Tower and the Yoke

The reconstruction of the bell production took place over the last three years and has already been described in numerous articles, e.g. here or here, or here. Beside a specialist publication  the experiments also led to a reconstruction for the Romanesque Bartholomew Chapel in Paderborn, which will be hung there in August 2019.

For the reconstruction, a bell tower and a yoke were designed by the craftsmen of Campus Galli, which can now be tested for their practical suitability over the next few years. Sources  on yokes for early  bells is very thin: Only one yoke of the Haithabu bell is preserved . The description of Theophilus Presbyter needs more interpretation  and  is still waiting for a practical implementation.

In the afternoon of 26.10.2018 the time has come: The bell rings for the first time, and compensates with its sound from the Carolingian period for variuos failed attempts in the past three years.


Jul 6 2018

A Romanesque Bell for a Romanesque Chapel

Bastian Asmus

Visiting the amazing bell foundry Grassmayr in Innsbruck, Austria, which houses the Romanesque bell from Hachen.

A Romanesque bell for the Romanesque Bartholomew chapel! What a fantastic development based on experimental archaeology! After working since 2015 on the reconstruction of the bell casting technique described by Theophilus Presbyter , which finally led to the long-awaited bell for the Campus Galli  earlier this year, there is another step to be made: For not only does the metropolitan chapter of Paderborn want the public to experience how a Romanesque bell was cast around the year 1000, no, this bell is to be installed in the Romanesque Bartholomew Chapel, which celebrated its 1000th anniversary in 2017.

The Bartholomew Chapel is not just any Romanesque chapel, it is an important architectural achievement of the high Middle Ages, which with its Byzantine style elements has no direct precursors or subsequent buildings. This oldest hall church in Germany was built in the Ottonian period as a palatine chapel during the construction of the new Paderborn palatine.

Bartholomäuskapelle, der Dachreiter für die romanische Kapelle ist noch leer!

The Bartholomew chapel in Paderborn, dated to the early 11th century. The ridge turret is still empty! There the new Romanesque bell will be installed.

A Romanesque bell in the making

The St. Bartholomew’s Chapel is given a bell that matches its date of origin. As a model for the new bell, a Romanesque bell from North Rhine-Westphalia was chosen. The bell of Hachen can be admired today in the bell museum of the bell foundry Grassmayr in Innsbruck, which acquired the bell in 1938 from the bell foundry Heinrich Hupert from Brilon. The Grassmayr brothers allowed me to measure the bell exactly, so that a new creation of this bell type can succeed. This is not about making an exact copy of the Hachen bell. Rather, it can be seen as a model for casting a “new Romanesque” bell. The technology of bell casting is similar to that used to make the Canino bell, i.e. the lost wax process as described in Theophilus Presbyter.


Romanesque bell from Hachen, Germany

The Romanesque bell from Hachen is housed in the bell museum Grassmayr in Innsbruck. In comparison to the much older Canino bell, this bell exhibits a much more pronounced waste and shoulder section.

By 1017, the church had already had at least two centuries of experience in the casting of bells. In addition to increased wall thickness of the rib the shape has also developed from an egg-shape to a more pronounced waist and shoulders. Also the sound holes are often only present as a typological rudiment. Next to the technical casting questions, here are a few questions concerning the conceptual side of bell development:

  • can the different dimensions of the bell be put in a certain ratio?
  • how was a bell designed in the Middle Ages?
  • what were the measurement options?
  • how exactly can the sound be predicted?

All ribs of contemporary bells were studied and compared. Drescher has already done very valuable work here , and many of the ideas presented here have their basis in his work.

Digression: The fingerprints of the Hachen bell founder

During the examination of the Hachen bell I discovered the fingerprints of my colleague who left them in the bell model a 1000 years ago. Even for an archaeologist this was an extremely impressive moment. It also goes to show the surface quality the casting method was capable to produce.

1000 year old fingerprints on Romanesque bell detected.

A 1000 years old fingerprints of my “colleague” in the handle of the Romanesque bell from Hachen! Width of the handle shown: 33mm.

Still no strickle board in the Romanesque period

How can we imagine the manufacturing process? Has a cleric seen another bell and developed a desire to also possess a bell? Was this wish expressed like: “I would like a bell like the one from Mainz” or was it completely different? I don’t think I can answer those questions that quickly. But since Theophilus still does not speak of the use of a strickle board a hundred years later, we must assume that the ribs of the bells were made in a different way. A direct comparison of different ribs was therefore not possible in the medieval period. This would explain also explain why the sound finding of the early bells had to be based on chance at first. It is assumed here that only the documentation of the bell rib may have led to controlled experiments on the influence of the rib on the sound.

The production of the Romanesque bell for Paderborn poses a similar problem for me. I have a client who knows a bell he’d like to have. I can look at the bell and measure it, but I cannot use the strickle board to to transfer the rib, as this was unknown in the 11th century. I can transfer the dimensions recorded with the gripping compass with the gripping compass. This inevitably results in inaccuracies which, however I suspect were not a problem then. Nevertheless, during working at the bell you wonder whether there were any underlying rules to lay out the measures. Drescher proposed that the diameter of the sound  bow provided a base measure, from which the rest of the bell may have been constructed.

The Hachen bell has a height to sound bow ratio of 5:4, as well as a ratio of the diameter at the shoulder to the total height of the bell 1:2, and a ratio of about 1:1 strike ring diameter to the height of the sounding body. Further clear ratios could not be determined so far, in particular to the curvature of the bell waist.


The project is documented as a video log and recorded in several short films. So far, the first four working days have been completed. Note that working days are not the same as the process days, as the relatively massive casting core takes several days to dry. So far, the process has been ongoing since June 23, 2018, so that eleven process days have already passed by the time this article is written (July 3, 2018)


Day 1 and 2 Romanesque bell – setting up shop and making of the core

Day 3 and 4 Romanesque bell . finishing the core

Day 5 and 6 Romanesque bell – making of wax model

Day 7 to 10 Romanesque bell – finishing the wax model

Tag 11 bis 15 Romanesque bell – mould, pit and furnace

Tag 16 bis 18 Romanesque bell – Casting the bell!!



May 6 2018

Successful Carolingian bell-founding

Bastian Asmus

 Carolingian bell-founding on Campus Galli

A Carolingian bell for Campus Galli  has been cast successfully. Three experiments were necessary to cast a complete bell. On 28.4.2018, for the first time in hundreds of years, a Carolinigina bell was successfully cast according to the treatise of Theophilus Presbyter. The sound is already breathtaking – although the bell has not yet been worked on – and shows the harsh sound so typical of for these early bells. I am thrilled! The bell of Canino, the earliest known bronze cast church bell is the model for the bell that we cast in this experiment. It was not only reconstructed to the original dimensions, but more importantly, its original production method was also reconstructed, following the excellent medieval manuscriput of Theophilus Presbyter]. Thus the  bell-founding experiments on the Campus Galli have finally come to a happy  end. The bell can still be regarded as a raw casting on the Campus Galli  until Whitsun. Then, after I have finished it, it will be installed on campus.

Henkel der Bienenkorbglocke

Applied archaeometallurgy: The crown of the Carolingian bell of experiment no. 3 form 2018.


The bell

The shape of the beehive bell is based on the oldest known cast bronze bell: the Canino bell. It has a diameter of 39 cm at the sound ring and weighs 44 kg. Note the triangular sound-holes of this bell from the 8th/9th century. Theophilus also describes them in the 12th century in his bell-founding chapter. According to the unanimous opinion of the bell entrepreneurs, however, these do not contribute to the sound.

Canino bell drawing

Drawing of the bell of Canino in the original publication [zotpressInText item="{4ICDYHGU}"].

Carolingian bell casting in the Middle Ages can only be achieved in a team effort

Große Drehlade zum Formen von Glocken

Large lathe for bell-founding. Making the core . Unlike later, early bells were made horizontally, and in the lost-wax process!

Besides many technical details which will be discussed later on, the most important conclusion is that a large and well-coordinated team is necessary to successfully master the bell-founding process. This year we have succeeded in doing this in a particularly good way. This also allows some conclusions for the past, because good communication is necessary to successfully execute the entire process.

Previous experiments and some background on medieval bell-founding

Some time ago I started a larger project: the casting of a medieval bell. To that end  I spend some time at the  Campus Galli near Meßkirch, south western Germany. At the weekend of the 27./28. June 2015 a core of the bell was made. The casting of the medieval bell will take place from September 16 to 20. The Campus Galli is project that aims at building a Carolingian monastery with authentic methods only. The Campus Galli thus provided an excellent infrastructure to conduct this archaeometallurgical experiment. In the first part of this report we will talk a bit about the history of bells, where they are concerned with Christianity, in the second part I will discuss the manufacturing process and in the third part I will report on the experimental reconstruction  of the casting of a medieval bell.

A very incomplete introduction to early church bells

Oldest known church bell from Canino, Italy

The oldest surviving church bell is the bell from Canino, Upper Italy. It stems from the 8th/9th century and is now at the Vatican museums.

Cast bells are much older than Christianity and the earliest cast bronze bells are small clapper bells found around  pre-Shang China, roughly 1900-1600 BC , but this shall not be of further interest here. The oldest surviving cast bronze church bell is from the ninth century AD. It is the bell of Canino in Upper Italy , nowadays housed in the Vatican museum . From the fifth century bells were used by Christians.

The bells from the 8th to the 12th century did differ in several respects from bells we know nowadays. They possessed a different shape, reminiscent of old beehives. Because of the simpler shape and the sound. The sound of the bells was less defined than today. The following sound snippet is from the oldest resonant bell still in existence in Germany:

via wikimedia commons, User: 2micha

It is the sound of the Lullus bell from Bad Hersfeld, Hassia, cast in 1038 AD. The image to the left shows the differences in the bell shape. The early church bells were more or less of a uniform thickness throughout. In the beginning there was no sound ring of any note. This only evolved over time and the differences in thickness, which are responsible for the sound, became more pronounced in the following centuries. Another difference is the mode of production, whereas modern bells foundries use a two part mould, which is made using a clay model of the bell, until the 12th or 13th century AD bell founding relied on the lost wax method.

Bell shape of early Christian cast bronze bells and modern bells.

Shape of bells in the 9th century compared to modern times.

Experiences and observations made

My experience of the broken core can also be read indirectly in the schedula diversarum artium of the Benedictine Theophilus Presbyter for bell casting .


Theophilus Presbyter De fundendis camapanis

Detail from Theophilus Presbyter’s de diversis artibus, Harley MS 3915. We see Theophilus warning to dry each layer of core clay properly, before applying the next one.

Quo facto, sume ipsum lignum et circumpone ei argillam fortiter maceratam, inprimis duobus digitis spissam, qua diligenter siccata, suppone ei alteram, sicque facies donec forma compleatur quantam eam habere volueris, et cave ne unquam superponas argillam alteri nisi inferior omnino sicca fuerit.

Once that is done, take the wooden spindle and cover it with vigorously worked clay, first of two fingers thickness. Is this thoroughly dried, put the next about it, and repeat this until the mould is ready, as big as you want it. And make sure that you never apply a layer on the other before the underlying clay is completely dry.

His insistent warning is not followed by a consequence, but it can be assumed that similar mishaps happened to the craftsmen of that time: If you don’t follow the instructions – for example to keep to an ambitious schedule – what happened to me happens: The (too) damp clay core breaks out of each other. Why is that so?

The clay core is made in  horizontal fashion i.e. as long as the moulded clay is damp, it tends to tear off as it hangs freely. If the clay is too damp, it tears off due to its own weight. Perhaps this is also one of the reasons why from the 13th/14th century on one started to form bells standing. Here the moulded clay can also move when moist, but the clay core does not hang freely, and nothing can tear off. The worst that can happen is that the core collapses a little.

Bild eines gebrochenen Kernes für eine karolingische Bienenkorbglocke


Der Lehmkern der Bienenkorbglocke wird auf einer Holzspindel aufgebaut. Im Unterschied zu späteren Jahrhunderten, werden karolingische Glocken horizontal liegend geformt.

The clay core of the bell is mounted on a wooden spindle. Unlike later centuries, Carolingian bells are shaped horizontally.   Image:

See the following images for an illustration how the core is made


wooden spindle

Inserting the spindle in the large lathe.

Loam is applied in layers to an oaken wood spindle. Each layer should be well dried before applying another layer. Theophilus suggests layers about two fingers thickness. These layers can be dried relatively easily with a small fire under the revolving spindle, however wind works even better. Since the core is mounted horizontally, the weight of the core acts on the freely suspended part of the core. The problem is the total moisture content of the core, because a moist core tends to tear apart. If the core becomes too heavy, it tears apart in the middle.

Bell-founding: Schematic of the bell's core

Schematic of the bell core: 1. wooden spindle     2. first loam layer        3. second loam layer      4. competed core      5. wax model of the bells body.

The addition of straw and horse manure – sometimes referred to as organic tempering counteracts this by producing a fibre composite. The organic fibres in the moulded clay increase its tensile strength. De facto this means that the thus treated, very lean moulding loam has improved plastic properties in the moist state. This results in a sufficient stability not to tear apart. In addition, the organic admixture increase moisture transport and reduces drying times. Moreover, during firing of the mould these fibres burn and ensure improved gas permeability of the moulding material.

Carolingian bell-founding: Almost finished with the core for the bell.

Second last layer of core loam.  Photo: © Simone Napierala 2015.

The last layer of the mould core consists of fine moulding loam, so that the surface of the casting becomes smoother. It has no addition of straw, only the fibres of the horse manure.


Nevertheless, some challenges remain

Although the bell is complete, the properties of the moulded clay to be examined have also led to undesirable effects. On the one hand, only a poorer casting surface quality was achieved this time, and on the other hand, the mould broke for the first time. By damming and prudent action during casting, the bell could be saved, even though the mould  broke below the handle. Nevertheless, this has had a negative influence on the handle, which is clearly visible at the top of the picture.

Future work

The next bell I will cast this way is the Hachen bell. Today it can be seen in the Glockenmayr Museum in Innsbruck, Austria. I am allowed to measure due to the courtesy of Johannes Grassmayr in the museum there. In this experiment the focus is on the refinement of proven  moulding loams.