top of page
  • Writer's pictureTastes Of History

Dispelling Some Myths: Romans in China?

Updated: Feb 15

Every so often, over the last decade at least, our modern-day bards, the media, regurgitate a story about 'European-looking' people in north-west China claiming ancient Roman descent. This tale of a 'lost legion', spawning generations of Chinese villagers with fair hair and green eyes, has an appeal in itself that perpetuates its stranglehold on popular imagination. Most recently the author encountered the story posted on social media and, unsurprisingly, it once again grabbed the attention of various commentators whose belief in its 'obvious truth' seem unquestioning.

At first the stories were interesting and in their telling seemed, well, plausible. Yet on closer inspection can we be so certain that the evidence presented holds up to scrutiny? Ever the sceptic, could there be a much simpler explanation?

The Daily Mail (29 Nov 10), in the footsteps of a Daily Telegraph article from February 2007, reported that DNA tests have shown the residents in a windswept village in Gansu are the blond-haired, blue-eyed descendants of Roman mercenaries who allegedly fought the Han Chinese 2,000 years ago. While no one in the modern town of Lou Zhuangzi is fair, and there is no proof that the Romans ever set foot in Gansu before the Christian era, there has been frequent debate on whether a group of Romans offered their services to the Huns against the Chinese in the battle of Zhizhi in 36 BC before settling in the Gansu village of Liqian. Before Marco Polo's travels to China in the 13th-century, the only known contact between the two Empires was a visit by Roman diplomats in AD 166. Yet the story persists that 145 Romans, taken captive, had wandered the region for years before becoming mercenaries. Indeed, the idea has spawned various theories, TV documentaries, novels and, perhaps significantly, commercial opportunities for the impoverished area.

It was back in 1955 that Homer Hasenpflug Dubs, an Oxford University professor of Chinese history, first proposed the theory in a paper entitled 'A Roman City in Ancient China'. In it Dubs stated that some of the 10,000 Roman prisoners reported by Plutarch as taken by the Parthians after the defeat in 53 BC of Marcus Licinius Crassus’ army at Carrhae in south-eastern Turkey made their way to Zhizhi in Uzbekistan to enlist with Huns against the army of the Chinese Han Dynasty commanded by General Chen Tang. Prof Dubs theorised that this group of legionaries made their way eastwards and thus how a mercenary troop 'with a fish-scale formation' came to be captured by the Chinese 17 years later. Interestingly, Chinese accounts of their victory at the battle of Zhizhi note facing a unit of soldiers numbering more than 100 who used an unusual 'fish scale formation' - now popularly associated with a testudo ('tortoise') - and the use of wooden fortifications foreign to the nomadic Huns.

Dubs postulated that after the battle the Chinese employed the Roman mercenaries as border guards, settling them in Liqian, a short form of Alexandria used by the Chinese to denote Rome. The professor's hypothesis took almost 40 years to reach China. During Chairman Mao's rule, ideas of foreign ancestry were not ideologically welcome and the story was suppressed, and the issue has divided experts ever since. While some Chinese scholars have been critical of Dubs' theory, others went so far as to identify Lou Zhuangzi, a settlement in north-western China on the fringes of the Gobi desert, more than 200 miles from the nearest city, as the probable location of Liqian in the late 1980s.

In the 1990s Chinese archaeologists were surprised to find the remains of an ancient fortification in Liqian that appeared strikingly similar to Roman defensive structures. Yet by 1999, no academic papers had been published on the subject, and no archaeological investigation had been conducted. Despite this the media, residents and local government have remained unfazed. County officials, sensing potential tourist revenue, erected a Doric pavilion in Lou Zhuangzi (pictured opposite), while the county capital of Yongchang decorated its main thoroughfare with statues of a Roman legionary standing next to a Confucian scholar and a Muslim woman, as a symbol of racial harmony. Even entrepreneurs have caught on: in 'Imperial City Entertainment Street' there is a 'Caesar Karaoke' bar.

In 2005, after local authorities loosened control over genetic research, scientists and historians took blood samples from 93 people living in and around Liqian in the hope of that DNA tests would explain the unusual number of local people with western characteristics - green eyes, big noses, and even blonde hair - mixed with traditional Chinese features. Typical of these residents is Song Guorong who, with his wavy hair, six-foot frame and strikingly long, hooked nose, stands out from his short, round-faced office colleagues. He epitomises local sentiment: 'I really think we are descended from the Romans. There are the residents with these special features, and then there are also historical records about the existence of these people long ago.'

Another resident, Cai Junnian, pictured opposite, said his ruddy skin and green eyes meant he was now nicknamed Cai Luoma, or Cai the Roman, by friends. He become a local celebrity, and was flown to the Italian consulate in Shanghai to meet his supposed relatives. Mr Cai said his great-grandfather told him that there were Roman tombs in the Qilian mountains a day and a half's walk away, but he had never connected them to the unusual appearance he inherited from his father.

The 2005 DNA test results confirmed some of the villagers were indeed of Caucasian origin, leading many to conclude their descent from an ancient Roman army, although others remained less certain. As Yang Gongle, of the Beijing Normal University said: '[Yongchang] county is on the Silk Road, so there were many chances for trans-national marriages…The "foreign" origin of the Yongchang villagers, as proven by the DNA tests, does not necessarily mean they are of ancient Roman origin…Even if they are descendants of the Roman Empire, it does not mean they are necessarily from the Roman army. The empire covered a large area. Many soldiers were recruited locally, so anything is possible.' Prof Wang Shaokuan has gone further pouring scorn on Prof Dubs's thesis, saying the Huns themselves included Caucasians, Asians and Mongols thus explaining the unusual characteristics of Liqian/Lou Zhuangzi’s residents.

Even so, on November 21st, 2010, Newstrack India reported that experts at the newly established Italian Studies Centre at Lanzhou University in China’s Gansu province were looking into the possibility that some European-looking Chinese in Northwest China are the descendants of a lost army from the Roman Empire. Excavations were to be conducted on a section of the Silk Road, the 7,000-kilometre trade route that linked Asia and Europe, to see if a legion of Roman soldiers settled in China. Yuan Honggeng, head of the Centre, hoped to prove the legend by digging and discovering more evidence of China's early contact with the Roman Empire.

The Daily Mail story of November 29th, 2010 is interesting for the misleading connections it makes to provide substance to the original Dubs theory and to perpetuate the myth. According to the story’s author, Niall Firth: 'DNA testing of the villagers has shown that almost two thirds of them are of Caucasian origin. The results lend weight to the theory that the founding of Liqian may be linked to the legend of the missing army of Roman general Marcus Crassus.' Mr. Firth fails to mention, however, whether the DNA testing he refers to is more recent than the 2005 study. One can only assume that this is the earlier story regurgitated and thus not new evidence of proof.

Of interest, the article concludes with the bold statement that: 'As part of their strategy Romans also hired troops wherever they had conquered and so many Roman legions were made up not of native Romans, but of conquered men from the local area who were then given training.' The implication is for evidence of some form of Chinese legion, but actually provides readers with a confused view of Roman history and more specifically the nature of military recruitment in the late Republic. In 110 BC the Roman army was composed of propertied men mostly of moderate means, but means nonetheless. They fought for the Republic because it was their duty as citizens. They were the Republic. By 107 BC a series of reforms, often attributed to seven time Consul Gaius Marius, had led to men with no means, who had to be equipped by the state, joining the Army. While the reforms were probably gradual to meet the circumstances faced by an expanding fledgling empire, they were the first steps towards a professional Roman army. In turn this resulted in a greater loyalty of these men to their leaders and their unit than the Republic. Leap forward in time to AD 400 and the legions were predominantly Germanic in origin, and supplemented with foederati, barbarian allies from beyond the Empire’s borders. But in 53 BC, the start point for the myth, this had yet to happen. The legionaries who marched with Crassus would have been Roman citizens, albeit with newly enfranchised Italian allies after the Social War. A century after Crassus and the legions of the Julio-Claudians were probably still mostly Italian. Clearly not an admixture of ethnic Chinese and Romano-Italians.

The recruiting of 'foreigners' was not the exclusive purview of Rome. Pliny, for example, implies that the Parthians actively recruited prisoners for military service and the Roman historian Horace claims that the survivors of Carrhae were integrated into the Parthian army and married to local women. If this is accepted, then it is entirely plausible that the legionaries’ fate was to be moved to the far eastern fringes of the Parthian empire in Turkmenistan for use as border guards against the Huns - a practice familiar to many Auxiliary soldiers serving Rome on its imperial frontiers. There is, however, a problem with timing.

In 20 BC, during negotiations for the recovery of the standards lost at Carrhae between Augustus and the Parthians, it was stated that there were no prisoners to be returned. Various theorists have used this to substantiate the idea that the Romans must have migrated eastward to China as they were obviously no longer in Parthian hands. Unfortunately the negotiations were 33 years after Carrhae at a time when the life expectancy of a male soldier in the late Republic was between 45 and 50 years (excluding the risk of death in battle or other such mortal hazards). If we assume that most soldiers were aged between 17 and 30 years old during Crassus’ ill-fated expedition, then this would place them in the age bracket of between 50 and 63 years old which, when subject to scrutiny, seems highly unlikely and somewhat undermines the theory.

So, how to explain the Caucasian appearance of the Gansu villagers without the Roman hypothesis. Actually there is relatively straightforward explanation for the so-called 'Chinese Romans' as explained in more much detail by Razib Khan in his blog entry 'No Romans needed to explain Chinese blondes'. Based on gene studies, Mr Khan contests that they 'are from the same population mix, roughly, of the Uyghurs': a hybrid population with about a 50/50 West/East Eurasian mix that is the ultimate genetic result of the migrations and assimilations of peoples throughout the Eurasian region between AD 500 and AD 1000. Like many European populations, the Uyghurs have a West Asian and Northern European aspect, but critically they lack the South European ancestry. This is important, because it is dominant in both the Tuscans and North Italians. If the 'Chinese Romans' are genuinely Roman, they will have this specific southwest European ancestry, but even the earlier 2005 DNA tests failed to definitively prove the genetic link. For now, the Gansu villagers 'are simply part of the normal range of variation on the Western borders of China, which was inhabited by European looking peoples like the Tocharians until relatively recently.' The people of Liqian are essentially a Chinese speaking group whose ancestry, with its equal portions of West and East Eurasian influence, creates the illusion necessary for this story’s persistent survival.

Bon appétit!


Recent Posts

See All


bottom of page