He hopes a comparative DNA analysis may get him closer to unraveling a mystery that has haunted him for a decade.

The findings may help establish a genetic link between some villagers in Yongchang County, Northwest China's Gansu Province, and the ancient Romans in the Mediterranean.

When Xie was attending his post-graduate courses in Lanzhou University in 1995, he heard about stories of some ancient Roman soldiers who later ended up in Yongchang County, about 500 kilometers to the northwest of Lanzhou, the provincial capital.

Xie was intrigued, hoping to explore it with his studies in genetic research.

Research forerunners

Xie, however, is a newcomer in the search for the ancestry of the small group of farmers in Zhelai Village of Yongchang County. In June, he went to the village to collect samples from the villagers who have blue eyes, blond hair, big noses and prominent cheekbones. They look more Caucasian than Asian.

According to Song Guorong, a local villager with a good knowledge of Liqian (ancient name of Zhelai Village), Chinese researchers suggested that Liqian might have some links with ancient Rome in the 1930s and 1940s.

In 1955, Homer Hasenflug Dubs, professor of Chinese history at Oxford University, surmised that some of the 10,000 Roman prisoners taken by the Parthians after the battle of Carrhae in southeastern Turkey in 53 BC made their way east to today's Uzbekistan and later enlisted with the Hun chieftain Jzh Jzh against the Chinese Han Dynasty (206 BC-AD 220).

Dubs derived his speculation from ancient Chinese Han Dynasty history annals, which described a battle between the Han empire and Jzh Jzh in western China.

The annals noted that about 150 men from Jzh Jzh's army took up a "fish-scale formation," which Dubs surmised to have been the Roman testudo formation.

Dubs then asserted that these men, captured by the Chinese, then settled and built their own town called Liqian (Li-chien) the Chinese transliteration of "Alexandria."

In 1957, Dubs published his book entitled A Roman City in Ancient China.

Thirty years later, David Harris, an Australian writer and adventurer, read Dubs' book and came to Gansu to search for Liqian, which he called "a city built by Romans in China 1,300 years before Marco Polo entered Cathay."

During his trip, he met Guan Yiquan, a scholar in the history of Central Asia at Northwest University of Nationalities in Lanzhou, who had already probed into Liqian for about 10 years.

Guan, who was a young interpreter for the American Air Force in Chongqing during World War II, discussed in detail the questions Harris raised during his journey to Yongchang.

In 1991, Harris published his book, Black Horse Odyssey, mainly sharing his experiences of the journey.

Meanwhile, Guan was still writing his own work on his research into this possible "Roman city." However, Guan died in 1998, leaving behind a draft of 450,000-Chinese characters.

Guan Heng, Guan Yiquan's son, said he is trying to continue his father's studies and hoping to publish the work one day.

In his letter to Guan Heng, Harris wrote: "Without (Old) Guan's work, we in the West would know so little about the story of the Roman troops in China."

Indeed, today, in an e-mail to China Daily, Harris admitted that there was no new development in the study of "Roman city in China" in the West.

Over the years, a few more scholars have joined in the search.

Chen Zhengyi, a historian at Lanzhou University who had introduced Guan Yiquan to Harris, said he could cite proof from Han Dynasty annals to support these scholars' speculations.

Challenges

So far, their research has remained inconclusive.

Dubs' theory was considered "interesting and provocative" but was criticized as jumping to too many conclusions in his assertions, according to an article on the Pennsylvania State University website.

Yang Gongle, professor with Beijing Normal University, said there has not been sufficient proof to link the villagers with the ancient Romans.

According to Yang's research, Liqian County was established in 104 BC, half a century earlier than the proposed arrival of the Roman soldiers.

Meanwhile, he noted that the fish-scale formation had nothing to do with Roman legion's testudo strategy.

The double wooden palisade, which might have looked like fish scales, was widely used in constructions in Central Asia and India at that time, Yang said.

There is no link between the name Liqian and the Roman legions, Yang argued.

The debate took a new turn after a group of ancient tombs dating back more than 2,000 years were uncovered in Yongchang in 2003 during the laying of the country's giant west-to-east natural-gas pipeline project.

From one tomb, archaeologists found the owner of one tomb to be 1.8 meters tall in life. Some researchers believe this offered more proof that soldiers from ancient Roman legion once lived here.

However, Zhang Defang, director of Gansu Provincial Archaeology Team, pointed out that the tombs were dated to the Eastern Han Dynasty (AD 25-220). The tomb owners should have no relations with the ancient Romans.

The development and wide application of DNA technologies have opened a new approach for researchers like Xie, who are bent on unraveling the mystery.

DNA lends a hand

However, Xie and his colleagues are encountering tremendous complexities.

The area where Yongchang is located was a trade hub along the ancient Silk Road, where people of various ethnicities from as far as the Mediterranean came and went, Xie said.

Moreover, soldiers in the Roman legions were supposed to consist of peoples of different ethnic and national backgrounds.

Because the Roman Empire was at that time at the height of its power and splendor, it had conquered many countries and regions across Europe, Africa and West Asia, he added.

According to Zhou Ruixia, Xie's assistant, they will build up the genetic data from the local villagers with Caucasian features and compare the data with those of European as well as Western, Central and East Asians.

They will report their research results in academic journals in the United States or Britain.

Two years ago, Ma Runlin, a bio-chemist based in Beijing, also collected blood samples from Yongchang people for DNA analysis.

However, he has not finished his research yet.

In an e-mail to China Daily, Ma said he is collaborating with British researchers in the genetic study of the villagers' ancestry.

He does not know when he will finish the research.

"I have backache. I needed to input 1,000 lines of data with 16 numbers in each line yesterday ... We're doing the experiments at the fastest speed we can," the 26-year-old said. "Please don't push me any more."

(China Daily August 24, 2005)