Ancient Proteins From Unwashed Dishes Reveal the Diets of a Lost Civilization

Material pulled from ceramic sherds reveals the favored foodstuffs in the 8,000-year-old city of Çatalhöyük in Turkey

The beginning of excavations at Çatalhöyük. Wikimedia Commons/Omar Hoftun/CC 3.0

Archaeologist Eva Rosenstock was never the potsherd type. She preferred the work of stratigraphy: sorting through the layers of remnant artifacts where humans once lived, learning about changes over time. With ceramics, she says, “you either love it or you don’t.”

That divide over sherds is especially notable at Çatalhöyük, a famous Neolithic town in Anatolia, Turkey. Rosenstock worked as one of the principal investigators at the West Mound area of Çatalhöyük, where the population appears to have declined and then disappeared around 5700 B.C. Compared to the East Mound—the more famous part of the prehistoric city—the West Mound is flush with pottery.

“You have like one sherd per bucket of earth that you dig [on the East Mound],” Rosenstock says. The population then shifted to the West Mound "around 6000 B.C." and the amount of pottery "explodes to kilos and dozens of kilos of ceramics that you dig out of the soil.”

Broken bits of bowls and jars remain, some with decorations such as vibrant red stripes. But for the most part, Rosenstock remained uninterested in the shattered vessels—until another researcher noticed something odd. Calcified deposits were found in the ceramic vessels but nowhere else. If those deposits showed up on other objects, like bones or human-made tools, they would likely have been a product of the environment where they were buried. But deposits found exclusively on the inside of the ceramics pointed to another explanation.

“It was really clear that this must have to do with the stuff that was inside this bowl,” Rosenstock says. She wasn’t sure what to do about the strange finding until she learned about the work of Jessica Hendy. An archaeologist from the University of York, Hendy’s research involves extracting proteins from dental calculus on fossilized teeth and analyzing the molecules to learn about the diets of ancient humans. When Rosenstock approached Hendy to discuss applying the same method to the flaky material on the inside of the Çatalhöyük ceramics, Hendy was eager to dive in.

Potsherd Food
Examples of calcified deposits from modern and ancient vessels at Çatalhöyük. a Examples of CaCO3 accretions from a modern tea water pot with extensive calcified deposits used near the research project compound Çatalhöyük, b a close-up of calcified deposits, c a relatively intact vessel (not analyzed in this study) demonstrating bowl shape and extent of calcified deposits and d a selection of four sherds analyzed in this study showing deposits adhering to the inside surface of the ceramic sherds. Jessica Hendy et al.

The results of that years-long collaboration are described in a new paper in Nature Communications, revealing just how effective dirty dishes can be in helping archaeologists decode the past.

“This is the oldest successful use of protein analysis to study foods in pottery that I’m aware of,” Hendy says in an email. “What’s particularly significant is the level of detail we were able to see from the culinary practices of this early farming community.”

The potsherds yielded proteins from numerous plants—barley, wheat, peas and bitter vetch—as well as the blood and milk of several species of animal, including cows, sheep and goats. Of even greater interest to the researchers was the precision with which they could identify the proteins. They didn’t just see barley, but could identify the specific signature of endosperms, the edible part of the plant. The material was stored in ceramic containers in a way that suggests it was probably used to make some kind of porridge.

The milk offered even more insight, as the researchers could distinguish whey from other parts of the liquid—and in one jar they found only whey, indicating the ancient Anatolians were actively transforming the milk into something like cheese or yogurt. “Here we have the earliest insight into people doing this kind of milk processing,” Hendy says. “Researchers have found milk in pottery in earlier times, but what’s exciting about this find and this technique is that we can see actually how people are processing their dairy foods, rather than simply detecting its presence or absence.”

Caroline Solazzo, who works on protein analysis in textiles at Smithsonian’s Museum Conservation Institute, was impressed by the study. “The work was done by a very good team of experts in ancient proteomics studies,” Solazzo says. “It seems that proteins can be better extracted from the accumulation of visible residues in the crust than from the ceramic wall, which is an interesting result for future studies of this type.”

To identify the proteins, Hendy and her team took samples from the potsherds and put them through a mass spectrometry machine. This “shotgun” approach is different from past protein analyses, which involved looking for specific proteins rather than doing a catch-all examination. Proteins are made of specific chains of amino acids. Some proteins, like osteocalcin (which is found in bone), are made of only a couple dozen amino acids, while others form chains of thousands of the building blocks. To decipher the protein puzzle left behind in the jars from Çatalhöyük, Hendy and her team compared their results to a database of known proteins.

Reliance on a reference catalogue is one of the hurdles of this type of research, because the analysis is only as good as the database. Such archives tend to contain lots of data on commercially significant species like wheat, Hendy says, whereas less common plants remain underrepresented. Due to gaps in the data, the researchers couldn’t identify everything in the batch—but they still managed to unlock a wealth of information.

Çatalhöyük Map
Map of Çatalhöyük. a Site location and b site plan of Çatalhöyük, adapted from Hodder33. Çatalhöyük consists of two distinct mounds; the East Mound, dating to circa 7100–5900 BC, and the West Mound, dating to circa 6000–5600 BC. Jessica Hendy et al.

Hendy and Rosenstock aren’t the first ones to use proteins as windows into ancient life. In 2008, researchers looked at proteins trapped in clay pots that belonged to the Inupiat of Alaska around 1200 A.D. They found signs of seal muscle in the vessel, providing evidence of the Alaskan native’s diet. And bioarchaeologist Peggy Ostrom managed to extract proteins from the 42,000-year-old leg bone of a horse discovered in Juniper Cave, Wyoming.

The question of how long such proteins survive is hard to answer at this point, because the technique is so new. Rosenstock and Hendy speculate that the proteins survived in their potsherds thanks to limescale buildup on the vessels (think of the white buildup around your faucets or tea kettles). But scientists won’t know just how long proteins can survive until they pull samples from many more sites of different ages and different environments.

“We would love to use this technique to identify the diverse cuisines of past societies and how culinary traditions have spread around the world,” Hendy says.

As for Rosenstock, she’d like to learn more about whether certain foods at Çatalhöyük were always eaten together for reasons of nutrition—the way rice and beans create a more nutritious meal together due to the combination of amino acids. She also says that after this exciting discovery, her mind is finally changed about potsherds. “It got me really interested in ceramics, in the end.”

Get the latest Science stories in your inbox.