Why Prehistoric Herders Didn’t Spit Out Their Watermelon Seeds

Thousands of years ago, Saharans ate the kernels before the fruit became sweet

Watermelon Seeds
The 6,000-year-old watermelon seeds from Uan Muhuggiag (left) border a child eating a modern watermelon. Seeds photo by Susanne Renner / Illustration by Emily Lankiewicz

About 6,000 years ago, a band of herders corralled their sheep into a cave in northern Africa. As the group settled in for the night, they probably munched on seeds, gossiped and gazed at the cave walls, which were adorned with paintings of spear-wielding hunters and dashing prey. The weary nomads wouldn’t have noticed when some of the seeds fell to the ground.

In most cases, the snacks would have decomposed, their story ended. But the cave’s dry, salty air preserved the remains like cured meat. Millenniums later, the seeds resurfaced with other botanicals, bones and artifacts when 20th-century archaeologists excavated the site, known as Uan Muhuggiag, in the Sahara of present-day Libya.

And now, a team of scientists has sequenced one of the seed’s DNA—the oldest-yet genetic code recovered from a plant. The genome reveals the seeds belonged to a 6,000-year-old wild watermelon, which probably had sickeningly bitter pulp. It seems Saharans first consumed watermelon seeds, long before the fruit evolved into the sweet, domesticated crop grown today on farms worldwide. The findings, recently published in Molecular Biology and Evolution, provide clues for archaeologists piecing together the history of watermelon domestication. And understanding the fruit’s past diversity could help genetic engineers design future melons.

“It is an excellent paper,” says Anna Maria Mercuri, an archaeologist who has analyzed pollen from Uan Muhuggiag but was not involved with the new study. Through this kind of work, which blends genetics and archaeology, “we can really improve our knowledge, giving more and more details on past biodiversity, past types of plants and also past behavior of humans,” says Mercuri, a professor at Italy’s University of Modena and Reggio Emilia.

Both archaeologists and geneticists have been investigating watermelon evolution, but often with different methods and goals. By comparing reads of DNA among living species, geneticists have constructed evolutionary trees that show how domesticated watermelon relates to wild watermelons and other species in its plant family, cucurbits, which also includes squash, cucumber and pumpkin. Using this information, they’ve cross-bred domesticated watermelon with evolutionary cousins to develop varieties more resistant to diseases and pests.

For archaeologists, plant remains like watermelon seeds provide information about past societies’ diets and lifestyles. Around the globe, across the ages, foragers have helped useful wild plants proliferate by spreading their seeds and clearing competitive species, among other means. Wild species became domesticated crops, which couldn’t survive or reproduce without human care, and foragers became farmers. While archaeologists can generalize broad strokes of the domestication process, the details differ for each plant and culture.

“Every domestication’s an interesting question,” says Dorian Fuller, an archaeologist and botanist at University College London, who was not involved in the new study. “But watermelon is in a category of crops … which are among the earliest cultivated plants in various parts of the world.” And unlike wheat, rice or maize—domesticated as staple foods to feed the masses—watermelon and other cucurbits weren’t necessarily eaten daily or used the way they are today, Fuller explains. For example, early farmers may have squeezed cucurbit seeds into nutritious oils or hollowed the rinds into containers.

A crop’s evolutionary cousins can offer hints about why its wild ancestor attracted past humans. If, say, all living wild watermelons harbored succulent pulp, then the progenitor of domesticated watermelon probably did, too. In that case, scholars might assume foragers cared for the wild ancestor for the same reason we now farm the domesticate, known scientifically as Citrullus lanatus: to eat something sweet.

But watermelon’s wild cousins aren’t sweet. In addition to domesticated Citrullus lanatus, scientists have identified six wild species of watermelon, or Citrullus members, that grow in Africa and the Middle East. The most distant cousin hardly seems like kin: Native to southern Africa, the vine produces fruit that resemble a dog’s spiky chew toy and roots that can make poison for arrows. The other wild watermelons outwardly look more like Citrullus lanatus but inside have light-colored pulp, usually packed with bitter compounds.

Citrullus naudinianus
Groups in southern Africa reportedly have made poison for arrows from the roots of Citrullus naudinianus, a wild cousin of domesticated watermelon. Bernard Dupont from France via Wikimedia Commons under CC By-SA 2.0

Scientists have identified genetic mutations in Citrullus lanatus that explain its distinctiveness from wild forms. One mutation shuts off the plant’s ability to make bitter compounds; another renders the flesh red. These mutations created a fruit more appealing to human senses, but it remains unclear whether they arose before or after early farmers began cultivating watermelons. Wild watermelons are rarely noshed raw. Some humans still eat wild Citrullus, despite the off-putting pulp. “There’s ways to overcome the bitterness through processing to make them perfectly edible,” Fuller says. Some people create jam by boiling sugar with one Citrullus, the citron melon. West African communities make soups and dry-roasted snacks from the seeds of egusi melons.

Decades ago, Fuller put forth the idea that savory seeds, rather than sweet pulp, initially attracted foragers to certain wild melons and gourds. “The seed is high in edible fats, and it’s storable and transportable,” he says. “We often think of watermelon seeds or pumpkin seeds as snacks, but there’s no reason why in some cases they might be cultivated in part primarily for the seed.”

Yet, archaeologists have struggled to discern how early cultivators used the fruit. Some of the oldest evidence for watermelon eating comes from Egypt during the time of the pharaohs. In a 4,300-year-old tomb, a mural depicts a green-striped, oblong fruit that looks like Citrullus lanatus. Because the melon rests on a table laden with grapes and other sweet fruits, the scene suggests Egyptians ate watermelon for its pulp by this time. However, a millennium later, the 3,300-year-old tomb of the pharaoh Tutankhamun included 11 baskets stocked with a mix of jujube and watermelon seeds.

Watermelon Drawing
A mural within a roughly 4,300-year-old Egyptian tomb depicts a fruit that resembles domesticated watermelon. Susanne Renner and Lise Manniche

“I don’t think it was expected that King Tut would have planted these watermelons,” says Susanne Renner, a plant biologist at the University of Munich in Germany. More likely, Renner explains, the king was supposed to eat the seeds as he journeyed to the afterlife.

Renner started probing watermelon’s past about ten years ago, when she asked Guillaume Chomicki, a PhD student at the time, to build a family tree for the fruit. The task was meant to be a small project, to help him learn genetic techniques used in Renner’s lab. “It turned out to be much, much more interesting,” recalls Chomicki, now a biologist at the University of Sheffield in England.

At the time, botanists thought only four species of watermelon existed and that the sweet one, Citrullus lanatus, was domesticated in southern Africa. The speculation largely arose because an 18th-century naturalist collected a vine near Cape Town, thought to come from a Citrullus lanatus plant. Pressed and mounted on cardstock, the scraggly leaves became sweet watermelon’s type specimen—a particular preserved organism treated as the official representative of its species.

Chomicki extracted DNA from a snip of the type specimen as well as leaves from 80 other fruits, belonging to three dozen species. With these specimens, he hoped to determine how domesticated watermelon relates to the wild ones and how the entire Citrullus group relates to other cucurbits. The analysis raised the number of known watermelon species from four to seven. Chomicki also discovered the sweet watermelon type specimen was not a sweet watermelon. The 18th-century naturalist actually collected leaves from a citron melon, a wild species with a tough, bland interior, like rind all the way through. Native to deserts of southern Africa, citron melon is a relative of Citrullus lanatus, but it’s not the closest cousin. The sweet crop shares more DNA with wild watermelons growing in other parts of Africa, Chomicki and Renner reported in a 2015 study.

“Until our paper … everybody, all the texts and web pages and so on, said the [domesticated] watermelon was from South Africa,” says Renner.

Through similar studies, analyzing modern and centuries-old specimens, the researchers discovered that sweet watermelon’s closest relative sprouts in Sudan today. So northeast Africa is a good guess for where its domestication occurred. But scientists can’t definitively pinpoint the area because the habitats suitable for watermelon may have shifted as Africa’s environment changed across the past 10,000 years. Also, finding the crop’s closest kin doesn’t answer why past humans started cultivating the plant.

To uncover the culture and motivations behind watermelon domestication, the scientists knew they needed much older genomes. Yet they doubted the necessary specimens existed. At archaeological sites, plant bits only survive in special circumstances, like when they’re charred in a campfire, and that heat usually destroys the DNA. As of 2016, the oldest plant DNA scientists had nabbed came from barley grains and corn cobs between 5,000 and 6,000 years old. Based on the Egyptian mural, watermelon domestication likely occurred more than 4,000 years ago.

Renner obtained a few seeds she knew of from archaeological reports, including those recovered from Uan Muhuggiag, the cave in what’s now Libya. Before trying DNA extraction, the researchers scanned these seeds with high-resolution X-rays for a 2021 paper. The detailed images revealed distinctive cracks, suggesting that human teeth bit the 6,000-year-old seeds.

The scientists also obtained seeds from a site along the Nile River in present-day Sudan. Some 3,000 years ago, the spot held a desert encampment near a larger town ruled by Egyptian pharaohs. “What’s special about these seeds is that they’re desiccated,” says Philippa Ryan, an archaeologist with Kew Royal Botanic Gardens in England.

The team managed to extract DNA from both sites’ seeds. “It’s just amazing,” says Renner. The Uan Muhuggiag genome lacked key mutations that determine sweetness and red color. “It was not a watermelon as we know it now,” she explains. The fruit likely had a bitter, white interior. Considering the bite-like cracks, it’s probable Saharans munched these seeds. The pulp’s fate remains a mystery. Perhaps the cave visitors discarded the unpalatable stuff, fed it to livestock or cooked it in a stew. In any case, the archaeologists who excavated the site didn’t find any traces of pulp, which in theory could have been preserved, caked in a pot or trapped in tooth tartar.

The seed from Sudan didn’t yield the stretch of DNA necessary to check for sweetness and color mutations. But both seeds offered enough genetic code to compare them to more recent melons. In addition to the seeds from these two places, the researchers clipped leaves from 47 watermelon specimens, which were originally collected between 1892 and 1927 from five continents, and are now stored at the Kew herbarium. They also had genetic code from a few dozen modern Citrullus, analyzed in previous studies.

The archaeological seeds from Libya and Sudan contained stretches of DNA that matched different modern watermelons. The Uan Muhuggiag seed was genetically closest to egusi watermelons eaten for their seeds today in West Africa. The findings support the idea that past Saharans enjoyed bitter fruits for their tasty seeds.

The results also document the lost diversity of watermelon plants used for food. Early cultivators seem to have been nurturing sweet and bitter types, as well as crosses between species. Ryan, a co-author of the study alongside Renner and other colleagues, wonders: “When did that diversity start to disappear?”

Answering that question will require more watermelon seeds to fill in the geographic and temporal gaps. These lessons from the past could help modern farmers breed more resilient and nourishing fruits.

Get the latest Science stories in your inbox.