Glass was highly valued across the Roman Empire, particularly a colorless, transparent version that resembled rock crystal. But the source of this coveted material — known as Alexandrian glass — has long remained a mystery. Now, by studying trace quantities of the element hafnium within the glass, researchers have shown that this prized commodity really did originate in ancient Egypt.
It was during the time of the Roman Empire that drinks and food were served in glass vessels for the first time on a large scale, said Patrick Degryse, an archaeometrist at KU Leuven in Belgium, who was not involved in the new study. “It was on every table,” he said. Glass was also used in windows and mosaics.
All that glass had to come from somewhere. Between the first and ninth centuries A.D., Roman glassmakers in coastal regions of Egypt and the Levant filled furnaces with sand. The enormous slabs of glass they created tipped the scales at up to nearly 20 tons. That glass was then broken up and distributed to glass workshops, where it was remelted and shaped into final products.
But what many people really wanted was colorless glass, so glassmakers experimented with adding different elements to their batches. Producers in the Levant are known to have added manganese, which reacts with iron impurities in sand. The manganese-treated glass still retained a bit of color, however, said Gry Hoffmann Barfod, a geoscientist at Aarhus University in Denmark who led the study, which was published this month in Scientific Reports. “It wasn’t perfect,” she said.
Glassmakers also tried adding antimony, with much better results. “That made it completely crystal clear,” Dr. Barfod said.
And expensive: A price list issued by the Roman emperor Diocletian in the early fourth century A.D. refers to this colorless glass as “Alexandrian” and values it at nearly double the price of manganese-treated glass. But the provenance of Alexandrian glass, despite its name, had never been conclusively pinned to Egypt.
“We have the factories for the manganese-decolorized glass, but we don’t have them for the Alexandrian glass,” Dr. Barfod said. “It’s been a mystery that historians have dreamed of solving.”
Motivated by that enigma, Dr. Barfod and her colleagues analyzed 37 fragments of glass excavated in northern Jordan. The sherds, each an inch or two long, included Alexandrian glass and manganese-treated glass from the first through the fourth centuries A.D. The sample also included other specimens of glass known to have been produced more recently in either Egypt or the Levant.
The researchers focused on hafnium, a trace element found in the mineral zircon, a component of sand. They measured the concentration of hafnium and the ratio of two hafnium isotopes in the sherds.
Glass forged in different geographic regions had different hafnium signatures, Dr. Barfod and her collaborators showed. Egyptian glass consistently contained more hafnium and had lower isotope ratios than glass produced in the Levant, the team found.
These differences make sense, Dr. Barfod and her colleagues propose, because the zircon crystals within sand are inadvertently sorted by nature.
After being expelled from the mouth of the Nile, sand sweeps east and north up the coast of the Levant, propelled by water currents. The zircon crystals within it are heavy, so they tend to settle out early in the journey on Egyptian beaches. That explains why glass forged in Egyptian furnaces tends to contain more hafnium than Levantine glass, the researchers suggest.
When researchers analyzed the sherds of Alexandrian and manganese-treated glass, they again found distinct differences in hafnium. The manganese-treated glass had hafnium properties consistent with being produced in the Levant, as expected. And Alexandrian glass, the clearest of the clear when it came to transparent glass, chemically resembled Egyptian glass.
It’s rewarding to finally pin down the provenance of Alexandrian glass, Dr. Barfod said, adding, “This has been an open question for decades.”
But it’s still a mystery why glasses from Egypt and the Levant exhibit different ratios of hafnium isotopes. One possibility is that the zircons containing certain isotopic ratios are bigger, denser, or bulkier, which affects their movement, Dr. Barfod said. “We don’t know.”
Analyzing the chemistry of Egyptian and Levantine beach sand would be a logical way of confirming these findings, Dr. Barfod said. “The next step would obviously be to go out and get sand from both places.”