The Curious History of the Hakka and the Tulou

How the rural Chinese of Fujian convinced the US government they had built missile silos.

Much of this article can be regarded as a hypothesis rather than an entirely factual piece. There is far too much contradictory information of variable quality to form many concrete conclusions. It has been written referencing papers that would probably be best read in Chinese, by someone with a better grasp of Chinese history than me and a few days of fieldwork. It would be interesting to have a series of unbiased people research sections of this work to draw their conclusions and to present their agreement or disagreement. I had hoped that the National Geographic/IBM Genographic Project would shed more conclusive light on Hakka migration and origins, but as far as I can see, it does not. At the end of the article are my personal conclusions, made quite subjectively; I hope no one finds offence in them.

The Hakka

The Hakka are a large Chinese language group and probably a sub-group of the dominant Han. Within China, they are dispersed mainly across the boundary area between Jiangxi, Fujian and Guangdong, also in Sichuan and Taiwan. There are probably about 75 million people worldwide who would claim Hakka ancestry, and although the 2007 Atlas of China does not distinguish the Hakka as a separate ethnic group from the Han, there are probably 40 million in China

The term Hakka is probably relatively recent, arising only during the Qing Dynasty in the last 400 years. In the Hakka language, it is pronounced Haagga, and in Putonghua as Kejia, but it derives from the southern Chinese language meaning "families who are guests".

The Hakka are also dispersed widely overseas and were part of the driven migration of Chinese from the mid-C19th. In the United Kingdom, most Chinese restaurants are probably run by Hakka, but they are most widely settled in South-East Asia.

The Origins of the Hakka

There is confusion about the origins of the Hakka people; this is made more difficult by pseudo-scientists hypothesising along lines which may either justify twentieth-century political events or satisfy individual aspirations as regards their genealogy.

In the simplest terms, the Hakka probably migrated from the Huang He (Yellow River) valley during a disputed period between 220 B.C. and 420 A.D. The fact that this has not been more accurately dated is probably unambiguous evidence the Hakka are genealogically confused by interbreeding with other minorities and the majority Han. Due to a sequence of migrations, the linguistic and cultural lineage may be of prime importance, whereas the bloodline may be very confused.

The Hakka language was possibly the spoken language of the Chinese Imperial Court up to and including the Tang dynasty. From this, it can be concluded that the Hakka language was not unique to their culture but simply a carryover from when the Hakka were possibly closer to the centres of power. When the centralised empire collapsed, or once the power base shifted north again under the Yuan and Qing Dynasties, the Hakka appeared to have been at their most mobile and defensive.

The collapse of the Ming Dynasty was possibly the final straw when Fujian formed the last bastion of Ming resistance. Their potential support for the more southern-minded Ming may have resulted in the Manchurian Qing forcing many Hakka into the tight base area of the Wuxi Mountains, which have been notoriously difficult to access.

The later diaspora from the mid Nineteenth century onwards is not of particular relevance to this piece. Still, it should be noted that the Hakka were heavily involved in a series of rebellions at this time, particularly in the leadership of the Taiping uprising. This period in China had been preceded by floods, droughts and incredibly cold periods. The result had been famine. Those in the most marginal areas would have suffered more than those in more favoured lands, and by this time, the Hakka were in marginal farming zones.

Hakka "colonisations" in Taiwan and Sichuan came about through political choices or economic incentivisation.

Over the millennia, both the Hakka and the Han attempted to distance themselves from the gene pool groups to the North of China, often called the Xiongnu and associated with the Huns. Research into Hakka roots amongst the northerly groups has proved inconclusive. The only conclusions that can be drawn are that the Han and the Hakka are mixed in origin and probably contain elements of the northern tribes' ancestries. The migratory waves out of the central Asian area took place over tens of thousands of years. Initially, they were hunter/gatherers, then pastoralists, but continuous southern migrations of agriculturalists followed them. In turn, they were replaced on their farms by the pastoralists, now turned farmers, who pushed them out. The extent of bloodline mixing is uncertain but is sure to have been very high. As the southbound waves collided with the groups settling from the south, this process would have been further confused.

By the time of the arrival of the Hakka in the Jiangxi, Fujian and Guangdong area, they were a group that were self-identifying rather than genealogically different. The Hakka further diluted their genealogy by intermarrying with the local people and would often buy unwanted daughters from these people to serve as wives for their sons. Today it is almost certainly only language and culture that provide any sense of separateness between the Hakka and the Han Chinese.

However, blood typing suggests that the Hakka have more in common with the Japanese and Koreans than the Han. This connection could be explained if it were accepted that the Japanese and Korean people and cultures shared a common ancestry with the Hakka. Indeed, many people in Fujian look significantly more like the Japanese than many other Chinese. However, from the C13th onwards, Japanese incursions into southeast China could easily have led to the introduction of a Japanese gene pool base, first from regular piracy and later during the Sino-Japanese Wars.

The Migrations of the Hakka

It has been proposed that there was a single outward migration from Shanxi, in the Yellow River basin, moving southwards and eastwards into Korea and Japan. Such a migration would go a long way to explaining the appearance of "ancient" Chinese cultures in these two countries at similar times. The Yayoi culture of Japan, which arose during the Qin/Han Dynasties, bears striking relationships to the Chinese civilisation of the time; but this could be a Qin or Han derivative rather than specifically Hakka.

What could be the propulsion mechanism? A few causes suggest themselves:

· The fall of the Qin Dynasty in 206 BC

· The fall of the Western Han Dynasty in 9 AD

· The fall of the Xin Dynasty in 23 AD

· The fall of the Eastern Han Dynasty in 220 AD

· Any of the confusing changes C3rd – C6th AD

Whatever the force behind this first migration, the Hakka were probably pushed into the Chang Jiang (Yangtze) floodplain. The subsequent migrations were perhaps caused by:

· The fall of the Tang Dynasty in 907 AD

· The arrival of the Yuan Dynasty in 1271 AD

· The fall of the Ming Dynasty in 1644 AD

However, it is also unclear who moved. Were these mass movements of an entire people, or simply leadership groups shifting on, moving to dominate a different set of peasants?

The Hakka keep better records of their family lineage than most Chinese groups, and they have been very significant in politics and warfare. Sun Yat-sen, Deng Xiaoping, Lee Teng-hui (President Taiwan 1988-2000) and Hing Xiuquan (Leader of the Taiping Rebellion) are just a few of the Hakka to have risen to prominence. The Hakka held twenty-five per cent of the first PRC politburo. These examples perhaps indicate an experience of involvement in power structures.

There is also no surprise that so many overseas Chinese are of Hakka origin. From the start of the Qing Dynasty, the Hakka were primarily confined to the coastal provinces of Fujian and Guangdong, with access to the ports of Xiamen and Guangzhou. Access to a port means access to a ship, be it to Taiwan, Singapore or the United States. They were also confined to less productive land, which may have meant emigration was more desirable for the Hakka than the other southern Chinese based in more fertile areas.

It is generally accepted that the Hakka moved first from the Huang He floodplain to that of the Chang Jiang. It has to be supposed they were agriculturalists, as most have been until recently. Arriving in southern China would have meant either disputing land ownership with the locals or accepting less productive and unwanted lands. The move into the now recognisably Hakka "home" area coincided with the introduction into China of some of the New World staple crops, which could thrive in the lower latitude but higher altitude conditions and were pre-adapted for growth on steep slopes.

Time spent on the flood plain of the Chang Jiang would also have encouraged the Hakka to take up fishing and to become very competent in micro irrigation techniques. Using clay in irrigation systems may provide a glimpse into the Hakka's later history when they began to build the tulou (earth buildings).

The Hakka group did not appear to have started to move into its present "home" area until the commencement of the Yuan dynasty when settlement in this area would already have been comprehensive. If they were, as is suggested, already proud of their own distinct culture, possibly pre-dating the first unification of China 2200 years ago, the arrival of the Mongol over-lordship may have been the final thrust pushing them into the mountains. It would further explain their migrations under both previous dynasties and the later Qing when these were predominantly northern in characteristic.

Chinese history sees a regular pattern of pastoralist/agriculturalist conflicts in the North, which found its ultimate expression in the Great Wall. It regularly forced a southerly migration for many peoples from the Huang He River basin. However, it makes the Hakka's migratory tendency even more difficult to track; they were simply one group of many that moved. Today they are of the greatest significance because they have become the largest language group.

In each move, people from North of the Great Wall would have interbred with people to the South before the migrations. Thus each migration was a mixed bag. Even though subsequent migrations may be forced by the intrusion of the Huns, Turks, Xiongnu, Mongols or whoever, in the wake of each, the new settlers again would interbreed. Thus the following migration would once again be a mixture of blood types and genes.

The dynastic changes that impacted the migrations of the Hakka almost all coincide with a climatic change in central Asia. The fall of the Eastern Han in 220 AD and the partitioning of the Chinese Empire had been preceded by climatic variations and crop failure. The Jurchen's takeover of northern China as the Jin came in another period of climatic instability at the start of the C12th, as did the Mongol's establishment of the Yuan Dynasty in the late C13th. The Ming Dynasty, driven out by the Manchurian Yuan, also coincides with climatic variation in the mid-C17th.

After the Hakka had fully established themselves in the axial Jiangxi/Fujian/Guangdong region, they were still attempting to shrug off the tag of strangers when the climatic deterioration of the nineteenth century occurred. The period of inclement weather from 1800 to 1840 preceded some of the worst instabilities of the Qing, including, of course, the Hakka-led Taiping Rebellion, which provided the first big shove in destabilising China's last Imperial dynasty.

The Development of Tulou Housing

The Hakka will now forever be associated with the tulou, particularly the massive circular structures that NASA and the CIA identified from space as missile silos during the Cold War. But it is not clear whether the tulou was a Hakka invention or a housing type already used in this region that the Hakka cottoned onto and adopted as their own.

Two developments appeared to have led to the classic circular tulou we recognise today. One branch began with classically structured housing for the wealthy, which can still be found around China today; some perfect examples exist in Suzhou. The other branch began with perimeter fortifications of a village site. As time passes, the housing and fortified village branches develop higher walls as a reaction to the greater need for protection. Eventually, the housing branch becomes the large fortified square tulou, and the fortified village becomes the circular tulou. Both strands of development would have had cross-fertilisation of concepts and reactions to the same problems.

The different branches of development indicate wholly different backgrounds of the people involved in initiating their construction; the housing strand from wealthy families and the fortification strand by terrified peasant villagers. One train of thought is that the poor villagers were of the She minority group or other local people and that the wealthy house owners were Hakka. As their defensive needs converged and the Hakka's possible wealth diminished, their housing requirements became closer. Many of the circular tulou were occupied by non-Hakka groups quite recently, and this strongly suggests that the circular strand of the tulou may pre-date the Hakka arrival in the area.

The word “tulou” can be traced back to the county records of 1573. However, the oldest extant tulou, Yuchang Lu, dates from 1308. Whether this was definitively prior to the arrival of the Hakka in the area is arguable, it was initially home to five clans, and these did not all have family names that are common amongst the Hakka. It was likely constructed in reaction to the new Mongol leadership in Beijing, who, like most central governments until the birth of the People’s Republic, had limited control over the Wuyi Mountains of Fujian. The tulou has five stories with 54 rooms on each floor, but there was a design flaw, and this has led to the housing lining up as much as 15° from the perpendicular, pointing to a somewhat hasty construction. However, with its individual wells for each cooking area, the courtyard suggests a planned structure, as does the building’s longevity.

The local She group are also from northern China, although, like the Hakka, their origins are not entirely clear. Their original migration from the North may or may not have coincided with that of the Hakka. However, they appear to pre-date the Hakka regarding their arrival in Fujian.

Interestingly, the Japanese were the first non-Chinese to study the tulou in detail. The Japanese culture bears some resemblances to that of the Hakka; one supposition is that the Hakka and the Japanese had a common cultural or ancestral line dating from the Qin Dynasty or before.

During the Warring States period (475-221 BC), one of the prime activities was wall building, culminating in linking several walls to form the first great wall under the first all-China Emperor Qin (220-206BC). These walls were made of rammed earth on the lower lands; only in the mountains were they made of stone, and similar construction techniques may be used in building a tulou.

The square strand commences with a villa-like structure resembling the Chinese elite's palatial houses. Sometimes referred to as the Phoenix House, it symbolises wealth and status. This strand gradually becomes larger and more fortified as the external buildings are placed inside the defensive wall.

The early circular strand has similarities to simple defensive structures across Eurasia and Africa and develops through a motte and bailey stage before becoming a building unique to this area. Only the ultimate developments of the tulou are outlined below. As previously stated, there are many variations, but the gigantic square and circular fortress-like constructs hold the most interest and perhaps define the pinnacle of the architectural technique.

Although some of the large tulous are still partially inhabited, more recent developments have seen many former residents abandon tulou life; however, the structures are still used in several ways. They are utilised as:

• storage areas and the central courtyard as work yards

• walls to which lean-to housing attached

• a source of raw material for the construction of more modern buildings

• tourist attractions

The most recently constructed tulou was built as late as 1967. However, this is a rare modern addition and may have been built in response to uncertainties in the Cultural Revolution. Certainly, the need for protection from bandits was still an issue in the early Republic. The occupation of the coast of Fujian by the Japanese would also have seen the tulou prepared for defensive purposes.

It is likely that the earliest large tulous were built from the C13th, although this is not a certainty. It is expected that the more significant fortified buildings were a reaction to times of instability and that in times of relative peace, lesser fortified dwellings would have been built; thus, there is no clear starting point for their construction. Indeed, the gross instability of the C19th, and the period during the overthrow of the Ming, would have been two such times that fortified homes were regarded as a necessity. Japanese piracy between the C13th and C16th and internal banditry up into the C19th were also a threat. It has been suggested that the tulou were sufficiently daunting for the pirate bands to be dissuaded from attacking them. In 1934, a tulou was attacked with army cannon fire after a peasant uprising; it proved remarkably resistant.

During the early Yuan Dynasty, it is clear that the She people were in active rebellion against the centralised leadership. Around this time, we have evidence of the first defensive tulous; this may have coincided with the arrival of the Hakka in the area. The She had been previously troubled by land-grabbing Han lords in Guangdong, causing them to move into the Wuyi Mountains, where they, as late arrivals, had to occupy less attractive farmland. The Hakka were ethnically less distinctive; they were more a disaffected conservative Imperialist Han splinter group. This reasoning explains their presence in the region and perhaps their relationship with the local agriculturalists.

The Tulou Building

Since the recognition of the tulou by the World Heritage Organisation, many tulous have been extensively repaired and brought back to something nearing their original appearance with centralised assistance. In an attempt to gain tourist revenue, private funding has seen many more tulous partially or fully repaired. However, the interiors are often wholly turned over to small commercial enterprises in these cases.

It seems odd that many commentators have failed to mention that tulou’s central area is an open courtyard, the roof and extended eaves only cover the walls and the wooden housing pinned to them. From the outside, at ground level, it may appear as if the roof extends across the whole structure; this is not the case, although some smaller rectangular earth buildings have an entire roof covering.

In these structures, as soon as roofing is removed, the walls erode quickly, and the housing structure becomes immediately unstable. In addition, as soon as occupation ceases, the construction begins to degrade. Continuous upkeep is required; otherwise, the wood rots and the walls become exposed. An excellent example of this can be seen in the modern Hukeng Township, where across the river from a line of modern guest houses can be seen a semi-derelict tulou, which is both partially inhabited and undergoing restoration.

The great outer walls of the majority, but not all, of these larger structures, are made from a mix of earth, sand, some layers of wood or bamboo and lime. Underpinning the wall is often a section made of stones. The foundations were designed in such a way as to make digging through or under the outer wall extremely difficult but also had the effect of raising the earth sections away from the water table. The earth walls taper towards the top and lean inwards, adding stability to the structure.

The earth walls could have been constructed in one of two manners. They could use earth blocks, fitting them one on the next, similar to bricklaying. Alternatively, they could fix planking to either side and tamp down the earth layer after layer. Much of the Great Wall was constructed in this latter manner. It is equally likely that a combination of the two techniques was utilised.

Although the walls vary in thickness, they are generally between 1-2 metres thick, depending upon the overall size of the structure. In the case of the circular tulou, this can range from less than 20 metres to a massive 90 metres in diameter. In height, the tulous range from two stories to five, the tallest being around 26 metres high. The largest rectangular structure, the Yijing Tulou, measures 136m by 76m and is five storeys high.

In the walls, there are four types of opening. There is a single doorway on the ground floor, framed with granite. This entrance usually is about 2 metres high and 1.5 across and has a double door made of heavy wood with various vertical and horizontal bolting devices. The portal is protected from above with arrow slits and a gap presumably through which unpleasant liquids could be poured or even water to douse a fire attack. Some larger tulous have more doors; it can be imagined that if hundreds of people were in occupation, a single door simply might not work. On the ground floor, the only other openings are some small flues to lead smoke out of the kitchen areas, but it is unclear whether these are original features. Generally, there are no windows on the first and second floor, except in the smaller tulous when there are windows on the second floor upwards. From the third floor upwards, there are always windows, generally around 70 cm square, with white paint around them on the outer side.

The roof is pitched with a central ridge and extensive internal and external eaves, which protect the residents’ homes and keep heavy rain off walls. It is a wood construction topped with baked clay tiles. One interesting feature is that heavier tiles are occasionally used to weigh down the standard tiles, which can be found in linear patterns.

The entrance passageway is about 6 metres long. In many cases, two pounding devices are situated on a stone bench running along either side. These consist of a large stone cup with a 2m wooded lever on a fulcrum fitted with a stone head; they were presumably for pounding grain or dried root crops into flour. Their location is less easy to explain, but it can be assumed the whole community shared them, and they needed to be located in a dry area.

The courtyards varied from tulou to tulou. There were no buildings inside at one extreme, only a cobbled floor; at the other, there were so many structures the area resembled an enclosed hutong with narrow pathways between them. The central buildings appear almost invariably made of stone or large clay bricks, surfaced with a clay skim, with baked clay tiles on the roofs. In some of the circular tulous, these buildings were curved following the outline of the basic shape.

These courtyard buildings appeared to be used for storage, cooking and washing, although some were quite obviously designed for animals, probably pigs or chickens. In some of the larger tulous, they may also have provided communal or educational spaces. In two examples, the structure at the centre was the ancestral temple, although in most tulous, this is set into the far wall opposite the entranceway.

In Gaobei Village sits the “King of the Tulou”. It was constructed in 1709 and may be considered the most spectacular tulou. Although the outer wall is a modest four floors, it has a diameter of 73m, two inner rings, the outer of which is two storeys high, and a central ancestral hall. Unusually it has four entrances and, at one time, housed as many as eighty families, perhaps amounting to a total population of more than 400 people.

Nearby, Wuyun Lu is rectangular and pre-dates the “King of the Tulous” by 140 years. In other clusters, notably Tianluokeng, the square tulou pre-dates the circular. It is perhaps the best indication that the one is not a direct development of the other.

The situation of the temples gives one the feeling these structures have a religious purpose, but this is not the case. Unlike Western religious temples, the Buddhist-Daoist-Confucianist prayer centre is devoted to ancestral respect rather than worshipping a Godhead. Today these altars are a mixed bag with a leaning towards Buddhism. Still, in many cases, their origin was probably Daoist, and the critical aspect of the local belief was that the dead could influence the living. Today, most altars are still maintained, although the extent of their upkeep is variable.

One of the more interesting tulous, partly because of its differences rather than its conformity, is Hegui Lou in Pushan Village in the township of Meilin. Built in 1732, it is five storeys high, with 28 rooms on each floor. It has been constructed on the valley floor, thus taking up potentially valuable farming land and is overlooked by a hill at the rear, making defence more difficult. Strangely, it has been constructed on marshland, and the springy cobbles in the courtyard provide evidence of the height of the water table.

In Hegui Lou, the rooms on the top floor have a collection of coffins, wooden and narrow with sliding tops, which almost look as if they had already seen service. Hakka burials are initially made in coffins, but once the bodies have decomposed sufficiently, they are exhumed and the remains placed in an urn to be re-buried. Presumably, the coffins here had been used once and would be used again. There were also some examples of old bathtubs and a trunk made in England for Ezra Samuel & Co, Rangoon.

Generally, the housing, which is pinned to the inside of the wall, is of timber construct reaching up to the tiled roof. In most cases, the first floor is used for the kitchens, equipment storage, washing facilities, etc. The second floor provided a dry area, presumably for longer-term storage of basic foods. The third floor and above are the living quarters, the only rooms with windows in the external wall.

The rear of each dwelling is the massive circumference wall. The front wall, the dividing walls and the flooring are made of wooden planks, supported by round-section posts and rectangular-section beams. They would have provided visual privacy but would prove less effective as regards noise and vibration. Four wooden stairways provided access to the corridors, which looped around the front of the living chambers. Theoretically, each family occupied a vertical slice of the tulou, thus having their own living, cooking and storage spaces.

Much is made of the fact that the tulou clans operated communally and that they each lived in the same sized space. As a guiding principle, this is fine, but observation indicates that this was not always the case; some rooms, particularly at the cardinal points, appear more extensive and different in design, although these could have been community rooms. In the case of Huaiyuan Lou, which was built 1905-09, the mansion-style home (Phoenix Building) in the centre could easily have been the residence of the wealthiest man in the village, who possibly funded the construction of the whole tulou in fear of the forthcoming downfall of the Qing Dynasty. It is also one of the most effectively fortified structures having four observation posts, three water channels over the gate and a three-metre stone wall at the base.

One of the greatest fears for a tulou resident must have been a fire. Either an attack by burning arrows or a simple accident in the kitchens would have led to a conflagration that would have proved difficult to escape; the presence of only a single door in these circumstances could be seen as foolish. However, the earthen walls are virtually fireproof and getting sufficient arrows over the tops of the roof and into the wooden structures would have proved difficult, although not impossible; the reason for the height of the tulou likely was to prevent just such an attack. The possibility of accidental damage would have been assisted by the nearby presence of the wells and the use of fire-resistant wood. As previously stated, water supplies and chutes above the main gate would have facilitated dowsing in times of trouble.

Earthquakes are relatively common in the area. The earthen walls are far less likely to suffer damage from this source than ones made entirely from stone or brick. The self-supporting mechanism of the inward sloping walls, particularly in the circular tulous, would have further strengthened the structure.

A further advantage of the circular tulou is that each circle section was constructed to a standard plan. In the rectangular tulou, different designs are required not only for the corner wall and dwellings but also for the residences abutting them on either side.

It isn't easy to understand the advantages of a rectangular tulou over a circular one. One consideration might be that it is only possible to extend a circular tulou by either going upwards or by destroying the structure and starting again. Factors such as land availability, terrain and the clan's architectural knowledge would have also influenced decisions. The Tianluokeng cluster in Shuyang Township has a central square tulou, Bunyoun Lou dating from 1796. It is surrounded by four much later tulous, dating from this century, the latest being constructed in 1966. At this site, they chose the circular over the square, even though they had a square example to use as a template. All five tulous are of similar size, each having three floors and between 26 and 32 rooms. It is highly likely that in this particular case, the choice was dictated by terrain, as they perch on a ledge halfway up a steep hillside. This cluster is one of the most photogenic as a group as not only are the tulous huddled together they are overlooked by an excellent vantage point.

By the C21st, UNESCO recognised 46 Fujian tulou sites as exceptional examples of building tradition and function. The term "Hakka tulou" is in common usage; a Google search will turn up four times as many entries as "She tulou" and only marginally less than "Fujian tulou". In the future, the term Hakka will become even further imbedded into the architecture of fortified earth buildings; yet it is impossible to ascertain whether the tulou is of Hakka origin and, in all likelihood, it is not.

Personal Conclusions

I have headed this piece "personal conclusions" because there is a degree of subjectivity in their construction. In researching the Hakka and the tulou, finding articles written from an entirely objective viewpoint has been challenging. In many articles, theories are stated and upheld as either fact or accepted truth. I would be delighted to have my conclusions proved incorrect, which would lead to greater enlightenment regarding the histories involved.

Key to my conclusions is the question, "What is a Hakka?" There is a distinct difference between what a Hakka is today, and what a Hakka was in the first and early second millennia.

Whereas a Taiwanese student of mine baulked at the idea of being descended from a Hakka group and quite clearly wished to be considered Han, many people claiming Hakka descent do so because it makes them feel different. The growth of this concept of wanting to belong to a minority group has distorted what that group is and was, particularly when the main thrust has come from outside mainland China.

Here are my personal conclusions:

• The She people are part of the ethnic Miao group that moved out of northern China two thousand years ago. In the late C13th and early C14th, they moved out of Guangdong to escape Han (or Hakka) Chinese land-grabbing.

• The Hakka are a group of the Han who had maintained some traditions dating back to the C3rd BC. They had once been part of the Imperial set-up in which the Hakka language was utilised as the court language. Whilst they moved from their original home areas close to Xian, this was for economic and political reasons; they were not a tribe pushed out by hostilities; they were a group of wealthy citizens looking for an area to dominate.

• The dates for these movements are unclear as the Hakka still identified themselves with the Imperial court and would still have had relationships with it throughout their migrations.

• If the Hakka settled in the Chang Jiang or later the Zhu Jiang (Pearl River) floodplain, it was still as "lords" of each area. They enjoyed the support of single centralised Chinese rule but received no such support whilst the nation was fragmented.

• In their migratory moves, the Hakka interbred with the local populations, whether Han or ethnic minority groups.

• The Tang Dynasty was the last to use Hakka as its prime language; from then on, the Hakka were on their own as a group and had to start using their personal wealth to protect themselves.

• The Hakka fought against the Mongols of the Yuan Dynasty, beginning their retreat into the more mountainous areas of Jiangxi, Fujian and Guangdong.

• The She people had already established farming communities in the mountains and started developing fortified villages.

• As overlords in the areas they commanded, the Hakka used and continued to promulgate their language and build elaborate housing.

• Hakka funds and She human resources led to the design and construction of the tulou, as can be seen today, drawing from both Hakka and She knowledge, to protect themselves from bandits, pirates and Mongol control.

• The Hakka supported the Ming, even though the court's language had changed, but problems with piracy and bandits still beset them.

• As the Ming Dynasty collapsed, the Hakka helped finance the attempt to bribe the Manchurians to leave China.

• By this stage, the wealth of the ruling Hakka had been considerably eroded, and as a bloodline, they were indistinguishable from the local Han

• The Manchurian's Haircutting Order was particularly unpopular amongst the Hakka and led to a further withdrawal into the mountains.

• With both She and local Han assistance, the Hakka joined the forces of Geng Zhongming when he supported Wu Sangui in rebellion against Qing rule. They also endorsed Zheng Chenggong (Coxinga) in his continued defiance of the Manchurians.

• The Hakka objection to the Qing continued and represented a withdrawal from mainstream politics. Establishing a Han sub-culture in the Wuyi Mountains included using their distinct early Chinese language, Hakka.

• In the tough times of the early nineteenth century, the Hakka-speaking mountain folk migrated from their limited and marginal lands to other areas of China and overseas.

• From time to time, encapsulated by the Taiping rebellion, groups of Hakka speakers continued to rebel against the "foreign" Qing.

Bibliography

General. Much information was derived from the signage at each of the tulou sites, which is in excellent English and from our guide and driver, Jiang Hua Feng.

“Ethnic Groups”, Atlas of China, SinoMaps Press, 2007, ISBN988-98630-2-2, Kunyu Publishing Co. Ltd.

“Farm Economy in the Pumao Area, 1823-34 - A Case Study of Agricultural Labour Productivity of Late Imperial China”, by Bozhong Li, published 2002

Huang Hanmin in: "Chuugoku minkyou no kuukan o saguru", Keiichirou Mogi, Kenchiku Shiryo Kenkyusha Co. Ltd., 1991 (Attempts have been made to request permission to use the diagram from this source but it has not been possible to make contact. If there is any objection to its use the diagram will be withdrawn from the article).

“Fujian”, Lonely Planet – China, Lonely Planet Publications Pty Ltd, 2007, ISBN978-1740599153

“Architecture in Fujian”. Sikong Xiaoyue, China National Geography, Volume 1, Issue 3, Sept 2009, ISSN 1009-6337

“The Genographic Project”, National Geographic/IBM, 1996-2010, https://genographic.nationalgeographic.com/genographic/lan/en/atlas.html, accessed 7-12/1/2010

“Hakka- An Important Element of Chinese Culture”, Siu-Leung Lee, Jan 2001, Asia Wind, www.asiawind.com, accessed 6-12/1/2010

“Hakka People” Wikipedia, 2010, http://en.wikipedia.org/wiki/Hakka_people, accessed 5-12/1/2010

"Climate change." Encyclopædia Britannica. 2010. Encyclopædia Britannica Online. 07 Jan. 2010 <http://www.britannica.com/EBchecked/topic/121632/climate-change>. N.J. Shackelton and N.G. Pisias (1985), D.G. Martinson et al. (1987), J. Imbrie et al. (1984), and D.F. WIlliams et al. (1988) in S.C. Porter, Quaternary Research, 32; © 1989 University of Washington

“Recent climate impacts on the limnology and ecology of Lake Baikal”, Anson Mackay, 4 December 2007, Baikal Science.org, http://www.baikalscience.org , accessed 5-12/1/2010

“Minority Groups In Southern China: Achang To Hakka”, © 2008 Jeffrey Hays, Facts and Details. http://factsanddetails.com/china, accessed 5-12/1/2010

“Clan homes in Fujian”, Jens Aaberg-Jørgensen, 2004-06, China Dwelling, http://www.chinadwelling.dk, accessed 5-12/1/2010

“The She and Yao Ethnic Groups”, MSD China, 2005, http://www.msdchina.org, accessed 5-12/1/2010

“The Hakka Earthen Buildings”, Sinoway Travel, 2004-09, http://www.sinowaytravel.com , accessed 5-12/1/2010

“Chengqilou”, China M Tours, 2002-09 http://www.chinamtours.com, accessed 5-12/1/2010

“Kejia Tulou Folk Custom Village - Fujian Tulou”, comeonlq, Dec 2009, Global Times, http://forum.globaltimes.cn, accessed 5-12/1/2010

“Hakka Religion and Expressive Culture”, Countries and Their Cultures, 2010, http://www.everyculture.com, accessed 12/1/2010