DOI: 10.1148/rg.266065079
RadioGraphics 2006;26:1837-1844
© RSNA, 2006
CT: A New Nondestructive Method for Visualizing and Characterizing Ancient Roman Glass Fragments in Situ in Blocks of Soil1
Roel J. Jansen, RT,
Martin Poulus, RT,
Jaap Kottman, MSc,
Tessa de Groot, MSc,
Dirk J. Huisman, MSc, PhD and
Jaap Stoker, MD, PhD
1 From the Department of Radiology, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands (R.J.J., M.P., J.S.); and the National Service for Archaeological Heritage (Rijksdienst voor het Oudheidkundig Bodemonderzoek), Amersfoort, the Netherlands (J.K., T.d.G., D.J.H.). Presented as an education exhibit at the 2005 RSNA Annual Meeting. Received April 25, 2006; revision requested May 10 and received June 26; accepted June 28. All authors have no financial relationships to disclose.
Address correspondence to R.J.J. (e-mail: r.j.jansen{at}amc.uva.nl).
 |
Abstract
|
|---|
A rare, complete ancient Roman burial site was discovered near the Dutch village of Bocholtz. In addition to many preserved grave offerings, there were countless fragments of deteriorated glass objects still buried in the ground. This glass was in very poor condition, however, and there was no possibility of excavating it directly. Instead, archeologists working at the site decided to dig up blocks of soil containing the glass fragments. High-resolution spiral computed tomography (CT) with multiplanar reformation, shaded-surface-display rendering, and volume rendering was used to obtain detailed information about the position, number, and form of the deteriorated glass fragments. CT-guided removal of the soil made it possible to restore some of the objects excavated from the blocks. In five of the 14 excavated objects, a correct Isings classification could be made based on the CT findings. In addition, CT was very important for the reconstruction of the layout of the burial chamber, the compilation of a list of grave contents, and the positioning of these contents within the chamber.
© RSNA, 2006
|
Introduction
|
|---|
In November 2003, a farmer in the Dutch town of Bocholtz was plowing his field when he noticed a spot where the plow jarred. He investigated the spot and discovered a Roman sarcophagus made of sandstone. The plow had broken the lid of the sarcophagus, which was otherwise still intact. The National Service for Archaeological Heritage (Rijksdienst voor het Oudheidkundig Bodemonderzoek) performed an excavation, which soon led to a major discovery: a rare, complete Roman burial site. The grave consisted of an underground burial chamber with timber floor, walls, and ceiling (Fig 1). The burial chamber contained "grave goods" made of silver, bronze, glass, and iron, indicating that the deceased was a person of wealth and importance. This discovery provided an opportunity to study ancient Roman burial customs in that area of the Netherlands, about which customs little or nothing was known (1–3).
View larger version (117K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Drawing illustrates the reconstructed burial chamber. Beside the sarcophagus near its longest side are two cylindric bottles and a dish. Two blocks of soil that were dug up in this vicinity contained one of the bottles and the dish; the bottle beneath the folding chair was removed "en bloc" (ie, in a block of soil) but was not examined with computed tomography (CT). The rest of the glass objects are seen in the corner on the left. The glass in this location was removed by digging up six blocks of soil. The bronze bowl surrounded by the three standing cylindric bottles was also present in one of the blocks (cf Figs 3–5).
|
|
Only two glass bottles were found undamaged, apart from some mild deterioration in the form of iridescence. Next to intact glass objects, countless fragments of glass were discovered in situ in the silt beside the sarcophagus. This glass was in very poor condition. It was broken and in some cases pulverized and therefore impossible to excavate directly from the site without the loss of information and context. The only way to retrieve and preserve the objects was to lift them out en bloc. The size of the blocks was determined by the concentrations of glass fragments on the surface. The block with the highest concentration of glass fragments was lifted by sliding a steel plate underneath it (Fig 2), a technique that produced the largest block of soil. However, this block did not fit into the CT scanner and had to be cut into three smaller blocks, with subsequent damage to glass fragments at the cutting surface.
View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2. Photograph illustrates how the block with the largest concentration of glass was lifted by sliding a steel plate under it. This block was then wrapped in plaster to prevent damage during transport. The smaller concentrations of glass were removed from the grave with a shovel and put in crates.
|
|
The blocks with smaller concentrations of glass were removed en bloc with a shovel and put in crates. The remaining space in the crates was then filled with soil to protect these blocks during transport (Figs 3–5). The largest block was not put in a crate but was wrapped in plaster to prevent damage.
View larger version (110K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. CT scan shows a bulb-shaped bottle (black arrow) that is partly filled with soil and air owing to a connection between the interior of the bottle and the open air. The image was obtained perpendicular to the longitudinal axis of the bottle. The Isings type of the bottle is unknown. Double-headed arrows indicate the loose soil that protects the block. A part of the floor of the burial chamber is visible as a dark discoloration (white arrows).
|
|
View larger version (159K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. (a) Maximum-intensity-projection (MIP) image obtained perpendicular to the axial section demonstrates parts of the handle and neck (arrows) of the bulb-shaped bottle shown in Figure 3. The bottle is decorated with a glass thread. Fragments of the handle and neck of a second bulb-shaped bottle are also visible (arrowheads). The glass from both bottles is fragmented and of poor quality, and the Isings types are unknown. (b) Three-dimensional shaded-surface-display image obtained perpendicular to the axial section shows the same parts of the two bulb-shaped bottles. Arrows indicate the decorative glass thread on the first bottle. Arrowheads indicate fragments of the handle and neck of the second bottle. The white dots surrounding the fragments of glass are pebbles in the loose soil that protects the blocks.
|
|
View larger version (68K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. (a) Maximum-intensity-projection (MIP) image obtained perpendicular to the axial section demonstrates parts of the handle and neck (arrows) of the bulb-shaped bottle shown in Figure 3. The bottle is decorated with a glass thread. Fragments of the handle and neck of a second bulb-shaped bottle are also visible (arrowheads). The glass from both bottles is fragmented and of poor quality, and the Isings types are unknown. (b) Three-dimensional shaded-surface-display image obtained perpendicular to the axial section shows the same parts of the two bulb-shaped bottles. Arrows indicate the decorative glass thread on the first bottle. Arrowheads indicate fragments of the handle and neck of the second bottle. The white dots surrounding the fragments of glass are pebbles in the loose soil that protects the blocks.
|
|
View larger version (73K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5. Photograph shows the partially reconstructed second bulb-shaped bottle (left) and a large fragment of the first bottle (right), which is decorated with rings of glass (cf Fig 4). After excavation, the bottles appeared to be Isings type 88b and a variant of Isings type 88b, respectively.
|
|
Each block contained several glass fragments that were part of a then unknown number of glass objects. To obtain a complete inventory of the burial site, it was essential that the glass fragments inside the blocks yield the necessary information. The two undamaged bottles were not hidden inside the blocks but were excavated separately. One damaged bottle was found beside the sarcophagus under an iron chair and was not included in this study.
In this article, we evaluate the potential of CT in identifying and characterizing glass fragments in blocks of soil from a Roman burial site and the possibility of obtaining sufficient information about the glass fragments with CT to facilitate excavation or identify badly deteriorated objects without excavation.
|
Materials and Methods
|
|---|
The CT examinations were performed in the context of a research project at the Department of Radiology, Academic Medical Center, University of Amsterdam, exploring the use of radiologic techniques in archaeology. No grant support was available for this study, and the CT examinations (
200. per examination) were paid for by the Department of Radiology.
To standardize the CT parameters, a customized protocol with good spatial and contrast resolution was used for each block.
Eight irregularly formed blocks of soil approximately 40 x 25 x 20 cm in size were fully scanned with a multisection CT scanner (MX 8000; Philips Medical Systems, Best, the Netherlands) using the following parameters: 120 kVp, 200 mAs, 1.3-mm section thickness, and 0.6-mm increments. The blocks were oriented in the scanner with the largest dimension in the Z direction. The reconstructed spiral CT scans were transferred to a workstation (MxView, Philips Medical Systems) for postprocessing. Multiplanar reformatted, MIP, three-dimensional (3D) volume-rendered, and 3D shaded-surface-display images were obtained.
The axial CT scans, coronal and sagittal multiplanar reformatted images, 3D volume-rendered images, MIP images, and 3D shaded-surface-display images were studied by an expert on ancient Roman glass prior to the excavation of the blocks, assisted by radiology technicians experienced in the application of radiologic techniques in archeology. The number and types of objects, their position in the block, and the quality of the glass were determined. The objects were classified according to a system developed by C. Isings, PhD, the author of Roman Glass from Dated Finds (4), a pioneering work on the typology of glass vessels from the Roman Empire that was published in 1957. The book continues to enjoy an international reputation as a valuable reference tool for the study of Roman glass. A so-called Isings number is assigned to an object based on the descriptions given in the book. These descriptions are themselves based on dated finds of more or less complete pieces of glass. Dr Isings described and made drawings of 134 various types of glass vessels. All known dated findings of each type, along with their origins, are listed.
The soil in six of the eight scanned blocks was at least partially removed to verify the CT findings with the excavated objects: The soil in five blocks was completely removed, guided by the CT findings, whereas the soil in one block was only partially removed for museum display purposes (Figs 6, 7). Of the two remaining scanned blocks, one was used as a sample for studying the effects of soil on glass degradation, and in the other no attempt was made to remove the objects because CT showed the glass to be of very poor quality (Fig 8). Nevertheless, CT revealed 11 objects in these two blocks.
View larger version (103K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6a. (a) CT scan obtained perpendicular to the longest side of the block shows a metal (bronze) bowl (curved arrow) enclosing a cylindric bottle. Double-headed arrow indicates the sides of the neck of the bottle, which is Isings type 51b. An Isings type 82b2 perfume bottle, or unguentarium, is visible on the left (straight arrow). The unguentarium is situated on the edge of the block (cf Fig 7) and is filled with air because of a connection between its interior and the open air. (b) MIP image obtained perpendicular to the axial section shows the unguentarium (open arrow). Solid arrow indicates another cylindric bottle (Isings type 51b), whose glass is pulverized and in poor condition. Arrowhead indicates glass fragments of an incomplete bottle appearing as a curved gray layer. (c) MIP image obtained at a 45° angle from the axial section shows the bronze bowl (black arrow) enclosing the Isings type 51b cylindric bottle, whose handle (white arrow) and neck are clearly depicted. However, the glass is fragmented and in poor condition.
|
|
View larger version (80K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6b. (a) CT scan obtained perpendicular to the longest side of the block shows a metal (bronze) bowl (curved arrow) enclosing a cylindric bottle. Double-headed arrow indicates the sides of the neck of the bottle, which is Isings type 51b. An Isings type 82b2 perfume bottle, or unguentarium, is visible on the left (straight arrow). The unguentarium is situated on the edge of the block (cf Fig 7) and is filled with air because of a connection between its interior and the open air. (b) MIP image obtained perpendicular to the axial section shows the unguentarium (open arrow). Solid arrow indicates another cylindric bottle (Isings type 51b), whose glass is pulverized and in poor condition. Arrowhead indicates glass fragments of an incomplete bottle appearing as a curved gray layer. (c) MIP image obtained at a 45° angle from the axial section shows the bronze bowl (black arrow) enclosing the Isings type 51b cylindric bottle, whose handle (white arrow) and neck are clearly depicted. However, the glass is fragmented and in poor condition.
|
|
View larger version (81K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6c. (a) CT scan obtained perpendicular to the longest side of the block shows a metal (bronze) bowl (curved arrow) enclosing a cylindric bottle. Double-headed arrow indicates the sides of the neck of the bottle, which is Isings type 51b. An Isings type 82b2 perfume bottle, or unguentarium, is visible on the left (straight arrow). The unguentarium is situated on the edge of the block (cf Fig 7) and is filled with air because of a connection between its interior and the open air. (b) MIP image obtained perpendicular to the axial section shows the unguentarium (open arrow). Solid arrow indicates another cylindric bottle (Isings type 51b), whose glass is pulverized and in poor condition. Arrowhead indicates glass fragments of an incomplete bottle appearing as a curved gray layer. (c) MIP image obtained at a 45° angle from the axial section shows the bronze bowl (black arrow) enclosing the Isings type 51b cylindric bottle, whose handle (white arrow) and neck are clearly depicted. However, the glass is fragmented and in poor condition.
|
|
View larger version (50K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8a. (a) CT scan shows the remains of a high bowl (solid arrow) and two large plates. Double-headed arrow indicates layers of heavily corroded glass. A part of the floor of the burial chamber is visible as a dark discoloration (open arrow). (b, c) MIP (b) and 3D volume-rendered (c) images obtained perpendicular to the axial section clearly depict the base rings of one large, two medium-sized, and two to three small plates; two stacked bowls; and two to three small bowls (solid arrows) (cf d). Open arrow indicates the high bowl (cf a), whose glass is less corroded than that of the other objects. However, all of these objects are severely corroded and heavily fragmented, and no attempt was made to excavate them because of their very poor condition. (d) Drawing made after the CT examination illustrates the five to six plates, two stacked bowls, and two to three small bowls whose base rings are depicted in b and c.
|
|
View larger version (126K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8b. (a) CT scan shows the remains of a high bowl (solid arrow) and two large plates. Double-headed arrow indicates layers of heavily corroded glass. A part of the floor of the burial chamber is visible as a dark discoloration (open arrow). (b, c) MIP (b) and 3D volume-rendered (c) images obtained perpendicular to the axial section clearly depict the base rings of one large, two medium-sized, and two to three small plates; two stacked bowls; and two to three small bowls (solid arrows) (cf d). Open arrow indicates the high bowl (cf a), whose glass is less corroded than that of the other objects. However, all of these objects are severely corroded and heavily fragmented, and no attempt was made to excavate them because of their very poor condition. (d) Drawing made after the CT examination illustrates the five to six plates, two stacked bowls, and two to three small bowls whose base rings are depicted in b and c.
|
|
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8c. (a) CT scan shows the remains of a high bowl (solid arrow) and two large plates. Double-headed arrow indicates layers of heavily corroded glass. A part of the floor of the burial chamber is visible as a dark discoloration (open arrow). (b, c) MIP (b) and 3D volume-rendered (c) images obtained perpendicular to the axial section clearly depict the base rings of one large, two medium-sized, and two to three small plates; two stacked bowls; and two to three small bowls (solid arrows) (cf d). Open arrow indicates the high bowl (cf a), whose glass is less corroded than that of the other objects. However, all of these objects are severely corroded and heavily fragmented, and no attempt was made to excavate them because of their very poor condition. (d) Drawing made after the CT examination illustrates the five to six plates, two stacked bowls, and two to three small bowls whose base rings are depicted in b and c.
|
|
View larger version (95K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8d. (a) CT scan shows the remains of a high bowl (solid arrow) and two large plates. Double-headed arrow indicates layers of heavily corroded glass. A part of the floor of the burial chamber is visible as a dark discoloration (open arrow). (b, c) MIP (b) and 3D volume-rendered (c) images obtained perpendicular to the axial section clearly depict the base rings of one large, two medium-sized, and two to three small plates; two stacked bowls; and two to three small bowls (solid arrows) (cf d). Open arrow indicates the high bowl (cf a), whose glass is less corroded than that of the other objects. However, all of these objects are severely corroded and heavily fragmented, and no attempt was made to excavate them because of their very poor condition. (d) Drawing made after the CT examination illustrates the five to six plates, two stacked bowls, and two to three small bowls whose base rings are depicted in b and c.
|
|
|
Results
|
|---|
In the blocks in which the soil was at least partially removed, 14 glass objects and two metal objects (knife and bowl) were unearthed (Figs 3–9). All 16 objects were correctly identified using the results of CT. A CT-based Isings type could be assigned to five of the 14 glass objects. Six objects were disrupted during the cutting of the blocks at the excavation site; consequently, portions of these objects were present in more than one block. However, this situation did not lead to incorrect interpretations of the CT findings. It should be noted that CT did not help detect a thin horizontal layer of glass splinters in one of the blocks, which proved to be part of a cylindric bottle.
View larger version (102K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9a. (a–c) MIP images obtained in perpendicular planes show the wide portion (black arrows) and neck (white arrow) of an Isings type 51b oval bottle. The bottle appears to be deformed and fragmented. The glass fragments are of poor quality. (d) Photograph shows three large excavated fragments of the bottle.
|
|
View larger version (76K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9b. (a–c) MIP images obtained in perpendicular planes show the wide portion (black arrows) and neck (white arrow) of an Isings type 51b oval bottle. The bottle appears to be deformed and fragmented. The glass fragments are of poor quality. (d) Photograph shows three large excavated fragments of the bottle.
|
|
View larger version (105K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9c. (a–c) MIP images obtained in perpendicular planes show the wide portion (black arrows) and neck (white arrow) of an Isings type 51b oval bottle. The bottle appears to be deformed and fragmented. The glass fragments are of poor quality. (d) Photograph shows three large excavated fragments of the bottle.
|
|
View larger version (73K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9d. (a–c) MIP images obtained in perpendicular planes show the wide portion (black arrows) and neck (white arrow) of an Isings type 51b oval bottle. The bottle appears to be deformed and fragmented. The glass fragments are of poor quality. (d) Photograph shows three large excavated fragments of the bottle.
|
|
On the basis of the CT findings combined with the findings at the excavation site, detailed information about the position, number, and form of the glass fragments was obtained. Most glass objects could be classified as plates or bowls, which together formed a dinner set. Fragments of six cylindric bottles and three small jugs that were probably used to hold fluids or beverages for use during the meal were unearthed. Together with four bronze objects, a silver spoon, and an iron knife, the glass objects served as tableware for a luxurious "meal" for the deceased. The CT scans were of paramount importance for the reconstruction of the burial chamber and its contents. Subsequent sampling and excavation of the scanned blocks helped confirm the CT findings. Determining the number and types of glass objects—especially the completely disintegrated plates and bowls—would have been nearly impossible with conventional methods.
The position of the objects in the burial chamber was reconstructed by combining field documentation (drawings and photographs) showing the position of the soil blocks with the interpretations of the CT scans of these blocks. This information would have been impossible to obtain without the CT scans, which provided the level of detail needed for such reconstruction.
The general scientific conclusions about the burial chamber are as follows (5): The burial chamber contained a sandstone sarcophagus and 52 objects that could be considered as grave goods. The objects were made of glass (n = 28), bronze (n = 17), iron (n = 6), and silver (n = 1). With regard to function, the objects could be divided into six categories—personal hygiene, hunting equipment, tableware, writing utensils, adornment, and miscellaneous—with objects in each category having a specific location within the burial chamber. The nature of the objects indicated that the deceased had knowledge of and, most likely, access to activities that were typical of the Roman world.
Inside the sarcophagus, the cremated remains of the deceased were found. Because of the presence of hunting equipment, it is thought that the grave belonged to a man. Specialized research of the cremated remains showed that the man was between 20 and 34 years old when he died. The objects indicate that his death occurred at the end of the 2nd century or the beginning of the 3rd century AD.
The man to whom the grave belonged was a resident of a nearby farmstead, which was excavated at the beginning of the 20th century. The main building of the farmstead was Romanesque, with a portico, underfloor heating, mosaic floors, frescoes, and several bathrooms. Together with the grave goods, these findings indicate that the man must have been wealthy and probably served an important public function.
Research into this Roman burial site confirmed hypotheses about burial habits from the era in question. The burial site itself was the first of its kind that could be carefully examined. This was also the first time that the residence of the deceased was known.
|
Discussion
|
|---|
Roman glass consists of 70% quartz sand; 20% sodium oxide, which was gathered in the ancient Middle East (eg, Egypt); and 5%–7% calcium oxide. Most ancient Roman glass was green or greenish-blue owing to an iron oxide mixture of Fe2O3 and FeO. Iron oxides in minor quantities were nearly always present in the raw materials used at that time. Adding specific metal oxides and varying the furnace conditions produced glass of different colors. For example, copper oxides were used to create turquoise, pale blue, dark green, and ruby red glass; and manganese oxides were used to create yellow and purple glass and colorless glass. In addition, gold, silver, and the oxides of cobalt, tin, and antimony were frequently used as coloring agents. The glass objects found at Bocholtz consisted mainly of clear and transparent green glass. For the colorless glass, antimony (0.2%–0.7%) was used as a decolorizer. In those days, colorless glass was expensive.
The oxides of the heavy metals used for coloring the glass absorb x rays very well; therefore, the glass had an attenuation very different from that of the surrounding soil.
CT was especially important in finding and identifying glass objects that were too disintegrated to excavate. These objects were left in the soil to prevent their further damage. However, CT also made possible the excavation and restoration of other objects with greater accuracy and less damage.
To our knowledge, this is the first study in which CT has been used to visualize glass in situ in the soil in which it was buried. Although Roman glass was examined in this case, because the chemical composition of glass remains very similar to that of glass in Roman times, we assume that the same examination method could be used for any type of glass.
|
Footnotes
|
|---|
Abbreviations: MIP = maximum intensity projection, 3D = three-dimensional
|
References
|
|---|
- Isings C. Roman glass in Limburg (Archaeologica Traiectina, IX). Groningen, the Netherlands: Wolters, 1971.
- Isings C. Glass from Roman barrows at Esch. Annales du 2d congrès des ‘journées internationals du verre’Leyde 30 juin-4 juillet 1962 (Edition du Secrétariat general permanent à Liège). Leiden, the Netherlands: University Press, 1962; 69–74.
- van den Hurk LJAM. The tumuli from Roman period of Esch, province of North Brabant. Nijmegen, the Netherlands: Catholic University Nijmegen, 1986.
- Isings C. Roman glass from dated finds. Groningen, the Netherlands: Wolters, 1957.
- de Groot T. Resultaten van de opgraving van een Romeins tumulusgraf in Bocholtz (gem. Simpelveld), Amersfoort (Rapportage Archeologische Monumentenzorg 127). Amersfoort, the Netherlands: ROB, 2006.
Top
Abstract
Introduction
