to perpetuate the memory and history of our dead

20.2 the people

We will never know their names. We will never know how many nor exactly where they came from. We can, however, try to tell their story so that they are not forgotten.

The story of the romusha is almost an impossible one to tell in full. Much like a large jig-saw puzzle, we can attempt to add small pieces of the story as they become available. But, unlike such a puzzle, we do not have a full picture to guide us as to where to place any given fragment of the saga. The two basic questions are: Who and How many were there? About the only thing we know for sure is that we will never know the full story of these doomed workers.

20.2.1 How Many?

Gavan McCormack attempted to account for the numbers of romusha by nationality. But any actual statistics are fragmentary at best. Japanese researcher Yoshikawa Toshiharu found that the Thais had completed a census of romusha that were in the large POW camp at Kanchanaburi (KAN). Of 70K romusha, they listed as 23K Chinese (32%); 40K Malay (57%); 7K Burmese and Mon (10%). But there is much lacking in that set of figures. It is well accepted (per Japanese records) that the island of Java alone provided 2.5M workers both male and female. As many as half of those may have been shipped off the island to any number of destinations. It is likely that a large number went to work on the Sumatra Railway, but a significant — if unknown number – worked the TBR. In his PhD thesis, Kurasawa Aiko studied the Javanese who were sent to the TBR. He places that figure at 15,000. He also offers an explanation as to why none were tallied in KAN as that they were kept in the jungle to cut wood for the trains. So their labors did not cease when the TBR was completed. How many others such a task applies to is an unknown figure? Other sources place the Javanese TBR worker population closer to 100,000. Like their Vietnamese counterparts no records have yet to surface regarding their involvement.

It is also accepted as fact that many Thais were also pressed into service. How many is only a guess at 12,000 [1]. The simple fact that the Kanchanaburi census tallied no Thais suggests that they simply melted back into the countryside once the TBR was completed. Another Japanese researcher, Hiroake Toshio, found that of 200K romusha who worked the TBR on the Thai side, 53% were Tamil/Indian; 37% Javanese; 8% Thai/Burmese and 2% Vietnamese. Ota Tsunezo places the number of Burmese conscripted as romusha at 70K with perhaps another 20K Mon hilltribesmen. In Burma, any Burmese or Mon would also likely have fled back to their homeland so the 7,000 Burmese / Mon that the Thais counted likely were sent to KAN from the Thai end of the TBR only.

As to deaths in this diverse group, Hiroake Toshio’s estimate is the lowest at 16% on the Thai side of the TBR. Hugh Clarks sets the percentage at 26% overall on the TBR. Other estimates run as high as 40%! A newspaper report states that in 1990, 700 sets of remains were unearthed in a sugar cane field in Kanchanaburi. Such a place that remained unknown and undocumented for 45 years after the event suggests a massive number of deaths among the 200-270,000 romusha that are thought to have worked somewhere in Thailand. 

Almost as an aside to this side story, the 1943 post-TBR census by the Thais also accounted for 25,000 Japanese and 33,000 POWs at the KAN camp area. That first number seems high and the latter quite low. We know from the US POW accounts that it wasn’t until early in 1944 (FEB-MAR) that many of them finally were consolidated at KAN. So, many of the Allied POWs may simply have not yet reached KAN when the count was made. There is no explanation as to the source of the Japanese nationals, although they must have been almost exclusively military. So one has to wonder about the romusha data from that head count. We also know that many of the POWs were trans-shipped to Saigon and Japan thereby lowering the number of TBR workers encamped at KAN.

[1] One Japanese source cites the number employed on the first 50 kilometer section of the TBR as 5000 with more ‘recruited’ to work in and around Kanchanaburi City and possibly to improve the road to the Tadan Bridge.

20.2.2 Who were they?

Another somewhat surprising aspect is that among the thousands of romusha there was a fair number of women. They are hardly ever mentioned in any of the accounts of the TBR. The archives of the Australian War Memorial in Melbourne contain photographs of some of these women unloading large metal 44-gal drums from rail cars. This indicates that the women were also pressed into service as laborers and were not just families of the romusha men.

The following website also addresses the plight of the romusha:

While McCormack concentrates his efforts on the numbers, it behooves us to try to grasp the human side of what transpired with these people from 1942-47. In more ways than one their lives were turned up-side down. Initially, many of them thought of the Japanese as liberators, fellow Asians who were going to cast off the oppression of European colonization. What we know today as Indonesian was then the Dutch East Indies. India, Malay, Singapore and Burma were British Colonies and although this technically afforded the inhabitants the protection of being British subjects many felt that they were being subjugated by the British.

The first act by the Japanese after conquering a country militarily was to loot it of any and all usable commodities. The natives soon realized that they, too, were little more than a commodity; to be used and disposed of as the war machine demanded. The Bushido code would not allow full assimilation of these ‘other Asians’. With minor exceptions, they were never incorporated into the IJA or Japanese industries; they were simply expendable ‘coolies’. Collectively the Japanese called them romusha.

As noted elsewhere, the initial TBR plan called for the entire project to be worked by these ‘hired’ workers. Large locally-organized recruiting centers were set up particularly across Malaya to hire workers using elaborate lies of wages and working conditions ‘to the north’. Many boarded trains with their meager belongings and families in tow, bound for an ill-defined destination. By and large, they were pushed off the trains at place called BanPong. They crossed the tracks and entered a transit camp on the grounds of an ancient temple turned house of horrors. From there, they were generally marched 100+ Kms along the path of the rail line to the place where they would now ‘work’.

It is known that the first 50 Kms of the TBR were built mainly by paid Thai labor [1]. Although the IJA marker at NongPlaDuk sets 16 SEP 42 as the official start of the TBR, the first of the British POWs had arrived from Singapore in mid-June. Precisely how they were employed for those first three months is ill-defined although it is thought that few were actually involved in construction. The NongPlaDuk camp was mainly dedicated to maintenance and supply activities. LtCol Toosey’s force of 1500 Artillerymen did not arrive at the future bridge site until November. Supposedly the TBR was completed largely by Thai laborers as far as ThaMaKam and perhaps even almost to ChungKai. That is the point where British POWs began their efforts while Toosey was building the bridges.

It is a story for another installment (see Section 20.3), but the use of Thai workers largely ceased in DEC 42 when what has become known as the BanPong Incident occurred. This created a major rift between the Railway HQ personnel, the local Thais and even the gov’t officials in Bangkok.

The larger group of US POWs didn’t arrive at their camps in Burma until JAN 43 and there were already romusha in the area. Precisely which nationalities were involved is not clear. It is thought, however, that the majority of the Asians in the area nearest Thanbyuzayat were Burmese and Mon. In addition to physically laying track, one of the primary duties of the various romusha groups was to clear the jungle and build the make-shift camps for those to follow.

One of the aspects of accounting for the numbers and the plight of these groups is that by and large upon completion of the TBR, the local Burmese and Thais simply melted back into the country-side. They rarely were consolidated to the post-construction camp at ThaMaKam.

Kyoto-Seika’s Prof Boggett relates that in the weeks between 15 AUG and official surrender on 2 SEP, the Japanese Railway HQ oversaw the destruction of most of the records at their various HQs. He goes on to say that the IJA kept their extensive POW records out of mutual military respect and some fulfillment of the Geneva Convention rules. But there were no such rules that applied to civilians so most everything was destroyed. Hence whatever account there was of their numbers or outcomes, was sent up in smoke.

The estimates of the numbers of romusha vary widely. The chart published by the TBRC as shown in section 17i (177K) is among the lowest and includes no Thais. The upper limit is 500,000 and most seem to come in around 300,000. The fate of these people is also a guess with death tolls ranging from 35-50%. But with no firm documents, all of these are ‘best guesses’.

Despite the ‘hired worker’ status that these people supposedly had, once they arrived at BanPong they were herded and treated more like cattle than humans. They had no infrastructure, no support, no supplies and no hope! Allied survivor accounts do not shed much light on their encounters with the romusha groups that were often nearby but rarely intermingled. It is reported that the small contingent of US POWs worked alongside Tamils in the area west of Hellfire Pass.

There were a few post-war studies of ‘returning’ workers that attempted to document their fate. But most of these were from a limited perspective and few offer any true documentation; mostly speculation. For example, one study attempts to document the Javanese who were ‘exported’. It mentions, but does not enumerate, that some went to Thailand. Did they indeed work the TBR; we cannot know for sure.


Clifford Kinvig’s book, RIVER KWAI RAILWAY, provides some insight into the Burmese workers. The puppet government of Dr. Ba Maw responded the call for laborourers and delivered the first 50,000 by June 1942. This number would triple over time. Like their Thai counterparts, they were to construct the first 50 or so kilometers of the TBR starting at Thanbyuzayat. But the work went much slower. Because of a lack of equipment all they could do was move dirt. These workers who became known as the “Sweat Army” (Let Yon Tat) were able to prepare the railway trace, but there were no rails to lay. IJA teams in Malaya were busy dismantling parts of the railway there with the intention of sending that iron north, but transportation was slow in coming. The laying of rails would have to await the arrival of the POWs in OCT-NOV. Apparently, once they experienced the horror conditions in the TBR camps, thousands of them simply melted back into the countryside and found their way home.

Kinvig also introduces us to a new POW group. A group of British (about 500) and Dutch (1200 KNIL soldiers) and 2 Australians who had been captured on Sumatra arrived in Burma. They are referred to as the British Sumatra Battalion. Once there they was divided into two groups. One joined the A Force and the other went south to Mergui. (see Section 8.22)

20.2.3 Their Plight

Which brings us to the status and plight of these workers at the time of liberation of the POWs. There is no reason to suspect that what happened at the many POW camps did not also play out the same way for the remaining romusha. By and large, the IJA cadre and Korean guards simply melted away, leaving a vacuum of responsibility. It is reasonably well documented that while the POWs were first consolidated at Kanchanaburi, then dispersed to many different destinations, many if not most of the romusha stayed in jungle camps where they cut wood for fuel and performed routine maintenance on the rail line. In AUG 1945, they were simply left ‘high and dry’ by their ‘bosses’. Assuming that they made it out the jungle to some semblance of civilization in due course – of which many undoubtedly did not – they were still mostly strangers in a strange land and still far from home, wherever that was.

Once again. Prof Boggett seems to have taken an interest in their plight. He notes, with much dismay, that neither the British nor the Dutch governments seemed the slightest bit interested in their ‘subjects’. While massive efforts were underway to repatriate the hundreds of thousands of POWs of all nationalities; hardly a finger was lifted to assist any of the stranded romusha. He calls it a national disgrace.

From the above website:

“Some Rǒmusha were employed to keep the Burma-Thailand railway operating while captured Japanese personnel, prisoners of war and other Rǒmusha were being transported from locations further up the line.

In other Japanese-occupied territories Rǒmusha were given supplies of food and medical attention by American troops arriving from August–September 1945 on. However, Allied authorities in Thailand and Burma prioritised their own military personnel leaving Rǒmusha perhaps last in line for supplies.

The repatriation of Rǒmusha was managed by a number of different authorities. The British Military Administration in Malaya sent missions to Thailand in November 1945 to aid the repatriation of Malayan Rǒmusha. Dutch authorities in Java did the same for their colonial subjects, although this process was made more difficult by the Indonesian independence uprising. The Allied Bureau of Refugees and Displaced Persons and the Red Cross were also involved in the repatriation of Rǒmusha.

The return of Rǒmusha to their homes was a slow process given the worldwide scarcity of shipping and the priority accorded the repatriation of POWs by Allied authorities. While most Australian prisoners had returned home by the end of 1945, the last ship carrying Javanese from Thailand to Java did not leave until July 1947, nearly two years after the war’s end. Given these difficulties many Rǒmusha remained in Thailand or started new lives elsewhere in the Pacific.

Once back in their home town or region, Rǒmusha were often left to fend for themselves. In Malaya, some received some clothing and a small amount of money, but many received nothing.”

Seemingly, in the immediate post-war period, the throng of romusha were swelled by the former KNIL native Javanese soldiers. Political unrest and rebellion against the Dutch colonial forces was on the rise across the Dutch East Indies. It was proving difficult to arrange for the repatriation of these men. Reportedly, they and many of the romusha were still in the Kanchanaburi area well into 1947. Little is documented about their return journey.

As part of his seminal work on this group, Professor David Boggett sat for some interviews to discuss their plight:

In the above videos, Prof. Boggett touches on the history of how and why there were Indians of Tamil ancestry in Malaya in 1942 as the IJA overran the country.

Cambridge Univ article about the methods of recruitment.

20.2.4 Conclusions

Japanese historians such as Kurasawa Aiko, Gotō Kenichi, and Satō Shigeru are pioneers of research on rōmusha. Nevertheless, they have focused largely if not exclusively on Indonesia and in particular on Java, because Java was Japan’s main source of forced labourers in South East Asia. Javanese rōmusha were shipped from Java, mostly via Sumatra, to neighbouring regions, or used on the spot, including, notoriously, as forced labourers on the Burma-Siam Railway. The majority of the labourers on the “Death Railway” were not, in fact, the Allied POWs, but people from Burma or Malaya. It should also not be forgotten that the rōmusha of the Burma-Siam Railway project were themselves not the only forced labourers in the “Greater East Asia Co-Prosperity Sphere”. Forced labour was used in various sectors, but not all of it was manual labour. Besides, the conditions of forced labourers could differ considerably, even on the Burma-Siam Railway.

Indeed, the prospect of recruitment for the construction work on the Burma-Siam Railway was without doubt greatly feared by all of Malaya’s local populations, their fear prompted chiefly by the reports of returnees about the appalling working conditions on the “Death Railway”. Ironically, many of those returnees had volunteered to work on the railway, but fled when they realized the truth of the situation in Siam.

At the same time as the Japanese increasingly used force to recruit labour, the local population found ways to develop avoidance strategies. For example, to escape rōmusha, some locals took to specializing in the production of certain goods necessary for the Japanese troops in the region. Others collaborated with the occupying force in other ways, such as the young men who were confronted with the choice of being frogmarched into a batch of rōmusha or volunteering to join paramilitary outfits like the heiho. Some locals did construction work other than that on the Burma-Siam Railway, perhaps working on other railway projects, or building airfields and other kinds of construction works. They were, in turn, guarded and overseen by other locals who had joined a collaborating paramilitary unit, an aspect of rōmusha that is only one “irony of the history” of the Japanese occupation of South East Asia. Further, it illustrates an unintended consequence of the actions of the occupying power, whose aim had been to recruit labour for the Burma-Siam Railway, by coercion, if necessary. Yet, the ultimate effect was that members of the labour force chose less coercive labour relations, and even if their agency led them into collaboration rather than open resistance, the imposition of particular labour relations could sometimes be eluded.

It appears as though the Tamils were shifted from Southern Indian to Malaya over many decades and were used mainly as manual laborers. One might speculate that even in Malaya they were on the lowest rung of whatever caste system existed; working as ‘coolies’ and hardly more than slaves. It is of little wonder that when the Japanese made offers of ’employment in the north’ and pro-offered the exorbitant sum of $US1 equivalent per day, thousands of these illiterate families literally jumped on the northbound trains!

Of course, the IJA had no intention what so ever of honoring any type of ‘contract’. Having been offered 90-day labor contracts, they were enslaved for 2-3 years. As they were pushed off the trains at Wat Don Toom in BanPong, they had only a glimpse of the horrors that would befall them and their accompanying families. In addition to the Thai-Burma (Death) Railway, it is estimated that an additional 60-100,000 romusha from Malaya were sent to work the much lesser known rail link termed the Kra Isthmus Railway.

Because, as Boggett explains in his writing and the above videos, the IJA destroyed every shred of paper work concerning the romusha, we will never know the true numbers nor the distribution of where the romusha worked the TBR. For the most part, when Allied POW survivor accounts even mention having contact with romusha, they fail to identify the ethnicity of those workers. It is generally believed that the the tens of thousands of Burmese (to include Mon and perhaps even some Karen) worked exclusively on the TBR within the confines of the own country. The same is true of the ‘hired’ Thai workers who laid most of the track from NongPlaDuk to Kanchanaburi and possibly on to ChungKai before British POWs began work towards the Khao Poon cuttings.

It is reported that dark-skinned workers (most likely Tamils) were employed in and around the Railway HQ at KAN, loading railway components (iron rails, wooden ties and ballast) on to barges and later trains to ferry those necessary items farther along the TBR.

The above video depicts many who appear to be romusha. One has to wonder exactly who made this film and for what purpose. No narration to reveal maker.

It is estimated that over 100,000 Javanese were also exported to the the TBR. But precisely where they worked is unclear (undocumented). An unknown number of Vietnamese (aka French Indo-Chinese) were imported into Thailand during the war, but if they worked the TBR directly is also not documented. They may have been employed in other tasks farther to the east of Thailand.

In the third video above, Prof Boggett mentions that the ONLY two memorials that he is aware of are the Thai-anusorn obelisk built by the IJA Camp Commandant in 1944 and the War Memorial in Ipoh that, when it was renovated in the 1950s, added “the Death Railway” to the list of memorialized events.

There is also the little known memorial (with an inscription date of 1957) at the Wat Yuan cemetery that contains the remains of as many as 10,000 romusha discovered in the decade or so post-war as road improvements and building took place in mid-town Kanchanaburi. This is mentioned in passing by Boggett. Finally, in 1990, as many as 400 sets of remains (thought to be largely Tamils) were unearthed from a sugar cane field in downtown KAN.

[see the other sections of this website where these memorials are discussed in more detail.]

20.2.5 Asian perspectives and background history of the Tamils:

During the WWII Japanese army used more than 250,000 Tamils in the construction of 415 KM railway between Siam and Burma to transport army supplies. During this project, it was initially believed that half of them (around 60,000) perished. However, recent research revealed that about 150,000+ Tamil Indians were killed during the duration of the Siam railway project. They fell victim to snake bites and insect bites, diseases like cholera, malaria & beriberi, massacre, torture, rape, committed suicide, etc. as they were unable to bear the burden.

In order to wipe out cholera, Japanese forces launched huge massacres against the Indians, killing massive numbers of the Tamil Indians daily. Handfuls of Tamil Indians also died weekly from overwork as they were worked to death like slaves.

In Hindi language(?) but a powerful depiction.

Sikhs on the TBR!

Many of the comfort women came from the same ethnic groups as the romusha:

Leave a Reply

Your email address will not be published. Required fields are marked *

6 + 1 =