The Craftsman’s Purse

It was in the early 800s, and vikings raiders were sweeping across Northern Europe as western kingdoms collapsed into civil war. But by the shores of southern Sweden near the mouth of the Helge Å, one family prospered. Their home was simple—a shallow pit measuring 3×3 meters (10×10 feet) with a house pitched over it like a tent. A loom stood in one corner and a hearth in another. They feasted on fish, made pots for their leftovers, spun wool, wove cloth, carved amber, casted bronze, and traded for beads of glass.

This small house in early Åhus measured 3.4 m (11 ft) by 2.8 m (9 ft). The two circles along the bottom left corner represent postholes from an earlier structure demolished before construction. The darker circles stand where the posts went to support the roof. The two dashed circles on the inside of the bottom right likely supported a standing loom. And the cluster of smaller circles at the top right stands for the jumble of rocks and pottery that remained from the family hearth. With an entire family confined to such a tiny space, it’s amazing that anything could fall to the ground unnoticed. (Tyra Ericson-Borggren, Report on Archaeological Excavations at Åhus, 1989-91.)

Admittedly, their lives haven’t attracted much attention. They left more trash than some of their neighbors, but less than others—if that’s any measure of wealth or importance. But when I saw the beads they let fall to their floor, I was stunned. For here amid the humble dwellings of Åhus, a handful of drawn and segmented glass beads had been abandoned among a family’s fishbones and broken pots.

These beads are admittedly small, but they could be precious. No one in Northern Europe could make such things, although some were beginning to try. A few families of glassworkers, it seems, shuttled back and forth between this small community of Åhus and a similar settlement on Denmark’s North Sea coast. They had little to work with, perhaps just a handful of glass pieces pilfered from old Roman mosaics, but they were learning to melt these bits of glass, wrap them around iron mandrils, and work them into beads both simple and advanced.

This bead was likely made by one of the glassworkers who trafficked between Ribe and Åhus in the early 700s. Before making the bead, the glassworker prepared one twisted rod of red, white, and blue, and one rod of white and red. He or she then wound a plain blue glass bead around a metal mandril, wrapping the twisted cane of red-white-blue around its center and then tracing wave patterns with the red and white rod on either end. Altogether, this is an impressive work of craftsmanship, and it ended up being buried with an elite woman on Bornholm, where it was found earlier this year. (Bornholms Museum, Rønne, Denmark.)

The beads from this house at Åhus were made differently. If glass can be wound at a temperature of about 700° C (1300° F), the glass for these beads must have been heated to somewhere around 1000° C (1800° F). A small air bubble was blown into this liquified glass, which was then stretched into a long hollow tube. For the green bead in the back (now turned white from reactions with the soil), this was the final step. Once it had cooled, the tube was cut into small cylinders like the one we see here. The blue beads went one step further. They were rolled on a surface with regular ridges, giving the beads the appearance that they were made in several segments, as if they were several small beads stuck together. The narrow waists were convenient breaking points, and similar beads typically show up with anywhere between one and five segments. Sometimes these fractures got messy, like the blue double bead on the back left.

Beads abandoned in the dirt by a family living in Åhus, Sweden, ca. 800. (Lund University Historical Museum, Lund, Sweden.)

The other beads, though much deteriorated, show evidence of yet another step. At some point in the production process, a fine metallic dust was added to make them shine like silver or gold. Over the long term, this has caused weaknesses in the glass, and many of these silver- or gold-foiled beads have started to peel like onions. The gold bead on the bottom left has lost some of its surface glass, and just above it is a piece of surface glass that came from a silver-foil bead, although the core of the bead itself is missing.

But here’s a close up of two beads that staggered me. The two-segment bead in the middle is what right looks like. The beads on either side, however, are missing their holes. There’s no way to string these beads into a necklace, or even to use them as beads. The closest furnaces capable of making these these beads were in the glass workshops of Syria and Iran. These beads then went through Constantinople and up the Danube to get to the Baltic. Or they went the longer way round via Marseilles, up the Rhône, and down the Rhine, before hitting the North Sea and rounding the tip of Jutland. Either journey is impressive, but I’m more impressed that nowhere along the way did anyone stop to check their cargo and discard these beads as worthless.

Three segmented beads found in a house at Åhus, Sweden. They were all made in one of the major glass centers of the Near East sometime around the year 800. The beads on the right and left are missing their perforations, meaning that they had no functional use as beads. (Lund University Historical Museum, Lund, Sweden.)

There’s many possible explanations for this. For now I’ll withhold my conjectures. But whether early medieval merchants never paused to take stock of their wares, or whether they didn’t see these hole-less beads as worthless, either case provides much food for thought about how medieval people moved and valued their goods. And this simple family from Åhus, whose stories have been compressed into a mere 9.5 m2 (100 ft2), somehow had not only the economic reach to participate in exchange networks stretching all the way to the Mediterranean and the Middle East. They also shared in an interregional set of economic values that made beads like these worth collecting—in such great numbers that some even fell unheeded to the floor and were ground underfoot.


Colors of the Viking Age

We have little evidence for the colors of the Viking Age. Most people would have lived in a world of colors that drew on and deviated from the natural world—split wood, smelted metal, smoked meat, dyed wool. Few of these things preserve their colors through long years in the archaeological record. Wood decays to dirt, metal corrodes to rust, meats reduce to bone, and what wool survives has often changed into the reddish hues of the earth where it was buried. Only glass beads retain their colors and are in fact protected by the dirt which shelters them from the sun.

But bead researchers have often neglected to appreciate just how precious the colors of their artifacts truly are. Most researchers seem to have been daunted by the beads’ varieties as well as by the difficulties of describing colors which are, after all, a very subjective thing. Each person understands and describes colors differently, filtered through eyeballs that are as unique as an individual’s DNA, and conceived of in a language unique to each person’s experience of the world.

Viking Age Scandinavians had more colors than your average crayon box. But while glass preserves these colors remarkably well, the ways in which we perceive and understand color have changed. How would a viking sort these beads? Which beads belong together? (Bornholms Museum, Rønne, Denmark.)

Nevertheless, I’ve been using the Munsell Color System to help me describe the colors that I see. Basically, this means referencing a book full of color chips representing the array of colors that most people can see and differentiate. Munsell in collaboration with the Society of Bead Researchers has helpfully published a slim volume of the 176 colors most commonly found in premodern glass. It’s kind of like selecting a paint chip at the hardware store. And although a color chip might not exactly match the object I’m looking at, and although it might be affected by the light of the sun or the warmth of a light bulb, these chips can still serve as a useful reference point that most people will recognize and agree upon.

The Munsell Color System, with hues wrapped around the equator, black on the bottom, white on the top, and moving from gray at the center toward purer colors on the surface. (Image courtesy of Munsell Color. Used with permission.)

Perhaps most importantly, by using the Munsell system to describe colors rather than resorting to vague and overlapping terms like blue, teal, and turquoise, I can quantify what I see. Munsell colors indicate not only hue (think: red, green, blue) but also value (light or dark) and chroma (pure and bright versus dun and gray). These three values can be plotted onto a 3D map, which lets me look at how the colors of the Viking Age cluster into groups. Right now, I’m using a model which places gray at the center and radiates into purer colors at the surface, with richer colors fading to black in the south and brighter colors shading to white in the north. This is the basic Munsell model that’s been used and tested by thousands of researchers for well over a hundred years.

So while observers today might be tempted to describe the colors of the Viking Age in terms of red-green-blue (a color model meant to stimulate the yellow-green-violet receptors in the human eye) or cyan-magenta-yellow (absorbent colors that can mask unwanted wavelengths from background whites in our computer screens and on printer paper), people in the Viking Age seem to have sorted their colors differently. Here’s a quick map of all the colors that commonly appear in premodern glass. If a dot appears darker, it’s because it actually represents a range of values (shades of light and dark) which are nevertheless the same hue and intensity. For example, tomato red and brick red lie on top of each other, because both can be made from the same base color with the addition of either white or black. The dark dot at the center actually represents both white and black, but without any additional color shading added.

The chart on the right shows how the Munsell system organizes all the hues of the rainbow. The dots on the left indicate all the hues likely to appear in premodern glass. Darker dots mean that these hues occur in varying brightnesses. (Munsell Hues image courtesy of Munsell Color. Used with permission.)

And here’s the colors of 1,429 beads that I’ve looked at. This represents only a subset of my data, selected from beads that were most likely made locally in Scandinavia. The glass itself would have been made and mixed in the Middle East—Egypt, Syria, and Iran were the early medieval centers for glass production—but the colors here represent the bits of glass selected by traders who traveled routes north between about 700 and 900 AD. The beads plotted here thus represent color choices made by several generations of people living between the Near East and Northern Europe.

The chart on the left shows the colors observed for 1,429 Viking Age glass beads. Note how this represents only a small selection of the possibilities for premodern glass shown in the chart above. There are very few reds or yellows, few pure greens, and almost no purples or oranges. But dark dots show the popularity of colors near the center spanning from yellow to blue, with an especially wide variety of blue colors.

I think this chart is interesting in itself, since it shows some of the selections being made by glass makers, traders, and consumers during the Viking Age. This is important because pretty much anyone who could get glass did, and they showed it off to those who couldn’t both in their daily dress and in their public ceremonies. Glass was among the most modest and most valued of goods, and in many of the graves for individuals who merited few gifts, a glass bead or two is all that survives. But where this data gets really exciting is when we start grouping the colors. This is a pretty straightforward statistical analysis that looks for concentrations of data points and then tries to draw lines around the looser sets of data that surround them. Here’s what the color chart looks like when I broke it into five clusters.

This is one of several cluster analyses I ran on my dataset of ancient glass beads. Although the results varied and my methods need to be refined, this particular clustering shows just how different Viking Age perceptions of color might have been from our own.

I must preface my comments by saying that I’m still learning how to choose and run statistical analyses, so my conclusions are by no means final. But what we can see here is really exciting. The blue dots represent the smallest variety of colors but the largest group of beads. These are dark blues that range from a purplish navy blue to a deep aqua green. But the red dots represent a broader range of colors that are blues pure and bright. The green dots represent lighter blues, overlapping with greenish colors like teal and turquoise and shading into even more yellowish colors like mustard and olive.

At the other end of the spectrum, the black dots represent a range of deep reds, browns, and oranges. I’m personally impressed by how closely these colors cluster around the colors of natural amber—there were lots of amber beads in my study but I excluded them from this analysis of purely glass beads. The light blue dots, however, represent a really loose cluster of bright reds and yellows. These were rarer colors during the Viking Age, so neither of them showed up in enough of a concentration to be modeled as a group of its own.

All this suggests that the Norse traders who plied Russian rivers might have perceived both dandelion yellow and fire engine red similarly as bright variants of a more fundamental amber hue. The vikings raiders who traversed the North Sea might meanwhile have seen more difference in light and dark shades of blue than in diverging hues of yellow, green, and purple. The Scandinavians who made, traded, or wore these beads all seem to have been people preoccupied with the intensity and depth of colors, rather than on our more recent interests of shade and hue.

This necklace has fewer beads of dark blue than I’ve found in most necklaces I’ve looked at, but it captures better than most the rainbow of colors evident in the larger data set, with few reds and yellows but a greater sensitivity towards blues and blue-greens. From a burial on Bornholm ca. 775. (National Museum of Denmark, Copenhagen, Denmark.)

To offer credit where credit is due, Munsell Color has taken an interest in my research and even helped me access some of the materials that make it possible. They’ve recently published a brief article I wrote about classifying ancient colored glass. Plotting Munsell Colors is, however, a remarkably new method of analysis. I’m especially indebted to foregoing research by Roy G. D’Andrade and A. Kimball Romney on the relationship between Munsell Colors and human color perception, and by Lana Ruck and Clifford T. Brown on the application of Munsell Color mapping to archaeological contexts.


Counting Beads from Viking Age Bornholm

It can be bewilderingly difficult to describe a simple thing. At least, that’s my experience now that I’m back in Copenhagen examining Viking Age beads. At present, I’m studying beads from an important cemetery at Bækkegård on the Danish island of Bornholm. A government administrator by the name of Emil Vedel excavated most of the cemetery in 1876, blitzing through 156 graves in a mere 13 days. This means that records are minimal and the handling of some artifacts was haphazard.

Beads attributed to Burial 66 at the Bækkegård cemetery on Bornholm. Other items in the grave allow us to date the burial to 750-775, or right at the cusp of the Viking Age. But to the original excavators, all the beads looked much the same and got mixed together by color. So although we know that Burial 66 received 20 blue beads in her grave, we can’t be sure that these are the same ones. (National Museum of Denmark, Copenhagen, Denmark.)

Beads in particular suffered from Vedel’s treatment. He recorded colors and numbers for each grave but then threw all the beads that he thought were worth keeping into containers for shipment to the National Museum of Denmark. The curators in Copenhagen resorted the monochrome beads into strings for each grave—although it’s unlikely that many of these beads actually came from the graves that they’re now associated with—and then they strung all the mixed polychrome and mosaic beads onto strings of their own.

Vedel’s catalog of beads at Bækkegård from his invaluable Nyere undersøgelser angaaende Jernalderen paa Bornholm (1878). His categories of beads include mosaic, painted, varied, blue glass, green glass, other glass, yellow clay, and red clay. It’s not certain what Vedel thought the difference was between “painted” and “varied” glass, nor what exactly belonged to his “other” category. His “clay” beads are now known to be opacified glass. (Emil Vedel, Nyere undersøgelser angaaende Jernalderen paa Bornholm, 1878.)

Nevertheless, Vedel’s initial 156 graves plus an additional 12 that were found and excavated thereafter make Bækkegård the site of nearly half of all Iron and Viking Age graves found on Bornholm. And although other large cemeteries have been found in southern Scandinavia, including the massive grave field at Lindholm Høje on the Danish mainland, Bækkegård remains a site of great interest. The people who buried there preferred inhumation to cremation burial, which means that the artifacts these people were buried with remained largely intact, whereas those in a cremation cemetery like Lindholm Høje have mostly been burnt or broken beyond recognition. So while Bækkegård’s beads might not be attributable to any particular graves, they remain valuable as the one of the largest groups of necklace beads that can be attributed to a single site in Viking Age Scandinavia.

Beads attributed to Burial 103 at the Bækkegård cemetery on Bornholm. Vedel recorded that yellow “clay” (actually glass) was the predominant material used for this burial’s necklace beads. Although these might not be the exact same beads that Vedel actually excavated, we know from the larger picture of cemeteries on Bornholm that yellow beads went out of style around 600, so I can safely exclude this necklace from my study while minimizing the number of Viking Age beads that I potentially miss. (National Museum of Denmark, Copenhagen, Denmark.)

And so, whereas earlier studies of Bækkegård and Bornholm have relied on Vedel’s simple listing of colors to deduce how necklace styles changed over time, I am one of the first researchers to examine the individual beads. They seem to be strung still on the original threads used back in the 1870s, and each bears a tiny tag with the catalog number written in a black ink that faded long ago.

Beads attributed to Burial 109 at the Bækkegård cemetery on Bornholm. This mix of colors is appropriate for the period around 630-650, but some of the beads mixed into this reconstruction almost certainly date to the Viking Age. Although the blue beads from Bækkegård appear in many diverse styles, the light that shines through them usually appears in one of only three shades. With further research, I may be able to associate these colors with particular glass production centers in the Eastern Mediterranean, whereas the diverse styles might be attributed to more local bead makers working within Scandinavia itself. The red bead at the right doesn’t have bubbles like the blue beads next to it, revealing that this is a bead was made of amber rather than glass. (National Museum of Denmark, Copenhagen, Denmark.)

At this point, I don’t know what’s important, so I’m trying to take as many discrete measurements as possible for future analysis. I’ve got a data set of 261 monochrome beads thus far, representing styles that were used between 600 and 900 AD. For each bead, I currently have 44 columns of data. Most of this is as simple as I can make it, including date and time of analysis, museum and catalog number, provenance location, material, method of production, notes on condition or damage, shape (profile and cross-section), basic dimensions of the bead and its perforation, and the colors reflected and transmitted by the glass (i.e. sitting on a white surface versus having light shine through it). I also have a few columns where I can write notes about observations that my current data scheme doesn’t quite capture, such as irregularities in condition, shape, or color. All told, it takes me about 2-3 minutes to catalog each bead, plus time for photos.

Mosaic and polychrome beads from the Iron Age and Viking Age cemetery at Bækkegård on Bornholm. The beads from Bækkegård comprise a beautiful but dizzying array of styles and techniques—the hard part for a researcher is figuring out how to record all these details. What details are important? And how do I record them discretely so that they can be correlated across a large data set taken from beads across Scandinavia? (National Museum of Denmark, Copenhagen, Denmark.)

Next week, I start tackling the polychrome and mosaic beads. To be honest, I’m daunted by the prospect. There’s so much variety and innovation during the Viking Age that I’m not quite sure how to capture it all in a meaningful way. I’ve consulted as many catalogues for other bead classification systems as I can find, and they’re all broadly divergent. This has been inspirational, since it allows me to pick the best pieces from each one. But it’s also discouraging, since it’s a clear reminder that even basic data collection is never unbiased, and the choices that I make about data collection now will determine the quality of my dataset for future analysis and interpretation.