Main Content

Student initiatives

Experimental archaeology: Creating a cylinder seal

Head: Prof. Dr. Elisabeth von der Osten-Sacken

Run by: Turna Somel, Florian Klatt and Jürgen Reitz

During the Summer Semester 2017, during a seminar held by Prof. Dr. Elisabeth von der Osten-Sacken at the Center for Near and Middle Eastern Studies at Philipps-Universität Marburg, an archaeological experiment was carried out on creating a cylinder seal. The aim of the experiment was to test various hypothetical techniques for producing a cylinder.

In this report, the process of the experiment is presented in four stages: Hewing a blank, drilling, grinding, and producing a cylinder diagram. Due to technical and time constraints, we decided not to make any flint drills ourselves.

1.    Hewing the blank 

To begin with, we were given access to square sandstone blocks. The first step was to make these blanks into a roughly round shape. We used different types of stone to experiment with trying to shape them by hitting them. One problem was trying to fix the stone in place. We tried holding the blank by hand. However, it was not possible to hit the stone hard enough and there was an increased risk of injury. Another method was to place the cylinder blank on top of another stone. This caused the blank to move around uncontrollably because of the uneven surface of the supporting stone. In the end, the best solution was to hold the stone in place by hand on a stable and level surface. This made it possible to make stronger, more controlled hits and reduced the risk of injury.
We discovered that an effective hitting method was striking the edges of the blank at an angle from above. This broke off larger pieces of the blank, but was totally controlled. At the top and the bottom, in particular, there was a tendency for larger pieces to break off than intended. This happened with two out of the three blanks. Another method we tried to round off the surface of the cylinder was to scrape away excess material using shards of flint. Due to how soft sandstone is in comparison to flint, this was successful. However, this method became less effective as the flint shards wore down. We hypothesize that this method would not be effective with harder rock types.

2. Drilling

During the first meeting, we attempted to drill an area of a pre-hewn blank. We tried the drill already described, but without a bow drill. We drove the drill by rubbing our hands together quickly. The blank was held in place vertically by a second person and we tried to drill through it vertically downwards. However, we soon discovered this was ineffective because of the lack of pressure from above and the lack of speed. Another drawback was the friction on our hands, which made it impossible to drill continuously. During this first attempt, we also realized that we would have to flush out the drill hole with water to remove the particles that broke off. Without using water, drilling did not progress very far.
During our second attempt, we implemented what we had learned. We used a bow drill to drive the drill. We also ensured that water was available to flush out the drill hole. Without a mechanism to hold the blank in place, two people had to drill while one person held the blank and flushed the hole with water as needed. Another person operated the bow drill by wrapping it around the drill and moving the bow horizontally forwards and backward. We discovered that it was better to first drill a shallow hole into the blank by hand. This made it easier to position the drill and ensure it fit safely.

Initially, we tried to use the bow drill without additional pressure from above; however, we soon realized that we needed a capstone to stabilize the bow drill and increase its speed. In particular, the samples that already had a hole proved to be most suitable for the capstone. The drill could be stabilized by inserting it into the hole and drilling a small amount. We also discovered that it didn’t matter whether the string was clamped to the arch with high or low tension.

During our third attempt, we experimented with another method of stabilizing the drill. To avoid the string slipping around the shaft of the drill too fast, which meant we had to frequently readjust the bow drill, we tried etching an indentation into the drill shaft. We discovered that this made drilling more difficult as it prevented the bow drill from moving. We also realized that the bow of the bow drill had to be dry so that it could fit tightly around the shaft.

3. Grinding

After drilling through the cylinder seal, we began the grinding process. We used several types of grindstone in the experiment. Both the seal stones and the grindstones were regularly rinsed with water while grinding to ensure a good result. First of all, we used a raw sandstone block for grinding. The material of the sandstone block was very soft, which made grinding very time consuming as only a small amount of material was ground away from the seal. After this, we used a sandstone slab with indentations carved in. Although the harder sandstone of the plates smoothed the seals better, we realized that the seal blanks were damaged and had bumps that could not be ground away with the sandstone within an acceptable amount of time. Because of this, we decided to use modern grindstones to remove rough bumps and damage. The modern grindstones used in this stage were coarse, so we then used the finer grained sandstone plates to achieve a smoother surface.

After the seals were rounded off and most of the damage had been removed from the surface, we then used very fine-grained abrasive materials used in modern stone processing. This allowed us to produce a very smooth and glossy seal surface in a short amount of time; however, there were still small, deep bumps that became apparent when carving the seal patterns. In ancient times, very fine-grained stone or quartz sand was probably used to achieve a result like this, but this would probably take a lot of time.

4. Production of the seal images

We planned to carve the seal images into the seal surface using a replica of the device depicted. However, this attempt was unsuccessful. Because the gear was mounted outside of the frame and the axle wasn’t straight, we were unable to control it. The drill attached to the other end of the device was also unusable for the same reasons.

Although several techniques were used in ancient times to produce seal images, we only used the scratching technique. Since the wheel-cut and ball drill designs didn’t work, the scratching technique seemed to be the most logical given the time and resources available to us. As well as the scratching technique, we decided to try out other techniques like filing and ‘micro-chipping’.

We made tools from flint. Microliths were particularly useful for carving patterns into the stone. However, using flint tools caused hand injuries, which was a drawback. Ideally, we would have been able to create handles out of wood or leather. Carving the first lines was difficult. However, once the first line was carved, it became easier to continue working. An abrasive material could have perhaps made this process easier. We also found that the flint tools quickly became blunt. It was, therefore, necessary to sharpen them or swap them for new ones.

The remaining rough patches were incorporated into the seal image in several motifs. We realized that the lines carved did not have to be particularly deep to leave a visible impression, even when trying to roll unfinished motifs into modeling clay.