Subtitle: How a siege that ended without a final storming of the city produced mass civilian death — the formation of the siege, the food blockade, civilian escape routes, the questions each side must answer, and why the death toll remains unsettled.
Free China Movement (FCM) historical research
Overview
The 1948 siege of Changchun was one of the gravest civilian catastrophes of the second phase of the Chinese Civil War.
From late May through October 1948, the Northeast People’s Liberation Army maintained a months-long encirclement of Changchun, which was held by Republic of China forces. As airfields, overland supply routes, and routes for movement were progressively cut, food shortages inside the city became acute. Large numbers of civilians tried to leave, and many were stranded in areas between the outer blockade lines and the defenders’ positions.
There is still no consensus on the total number of civilian deaths. Contemporaneous U.S. diplomatic cables, archival research from Academia Historica, published military telegrams, and later historical studies nevertheless support a basic conclusion: large numbers of civilians died abnormally under conditions of starvation and siege.
1. The siege was not an ordinary urban battle
From late May through October 1948, the Northeast People’s Liberation Army subjected Changchun to prolonged encirclement. Food, transportation, and military supplies deteriorated, and large numbers of civilians died under abnormal conditions. Key chronological anchors include the loss of Dafangshen airfield on May 24 and the cutting of the air route; on May 30, Lin Biao convened leaders of the Northeast Bureau and Northeast Military Region and decided on an encirclement strategy toward Changchun. Sources differ somewhat on when the “siege” began, so “late May” is defensible but should not be treated as the only possible starting point. On October 17 the 60th Army switched sides; on October 19 the New 7th Army laid down its arms; and in the early hours of October 21 Zheng Dongguo surrendered at the Central Bank building, ending the siege.
U.S. diplomatic records from 1948 provide contemporaneous outside observations. On June 15, U.S. Consul General in Mukden (Shenyang) Angus Ward sent Telegram No. 258 to Washington, describing Changchun as under an intense siege: artillery fire rose from about 100 rounds a day to a peak of 800 (falling to around 500 by June 13), the city’s northern airfield could not be used because of shelling, Nationalist airdrops required fighter escort, and prices for rice and kaoliang had risen roughly fifty-fold. (FRUS 1948, vol. VII, doc 893.00/6–1548.) The value of such records lies in their timing: they were written while the siege was under way, rather than decades later in political memoirs. A follow-up cable of June 21 said the siege had “diminished in intensity but continued.”
From a military-history perspective, Changchun was a classic war-of-attrition siege: encirclement, supply denial, and sustained pressure were used to compel surrender rather than a direct assault on the city.
On September 9, 1948, a report drafted by Luo Ronghuan and sent to Mao Zedong in the names of Lin Biao, Luo Ronghuan, and Tan Zheng, titled “Report on the Encirclement of Changchun,” stated that hungry civilians were pouring out of the city by night and day, were driven back, gathered in the zone between the two sides’ security lines, and that “many died of starvation as a result.” The report also stated: “Do not allow hungry civilians to leave the city; those who have already come out must be blocked and sent back” (Selected Military Writings of Luo Ronghuan). Contemporaneous U.S. cables provide outside corroboration: on April 13 Ward reported that refugees who had withdrawn from Jilin into Changchun and wanted to return to Jilin were not allowed to leave Changchun, and refugees approaching Communist outposts were told to return to Changchun to “consume food” (FRUS 1948, vol. VII, doc 893.00/4–1348). On August 11 Ward reported that Communist forces continued a tight blockade of Changchun and that “the civilian food shortage has reached a critical stage” (doc 893.00/8–1148). The full chain of command and implementation of the blockade still require item-by-item verification against original military archives.
2. The central question: could civilians leave?
Historical responsibility for the Changchun siege cannot be understood only through the outcome of “food shortages inside the city.” Three questions must also be asked: who controlled the roads into and out of Changchun; whether unarmed civilians were permitted to pass through the encirclement lines; and when and through what channels military commanders learned that the prolonged blockade was causing civilian starvation and death.
If military authorities knew that large numbers of civilians remained in the city and nevertheless used restrictions on supplies and population movement together as a means of compelling the defenders to submit, historical research must examine how those decisions were made, how the chain of command worked, and how much decision-makers knew about the consequences for civilians.
The humanitarian cost of Changchun therefore cannot be judged only by whether a final large-scale assault occurred.
3. Why food became decisive
As the siege continued, food ceased to be merely a necessity of life and became central to both the battle and survival. Existing stocks, military requirements, airdrop capacity, market-price movements, requisitioning and rationing policies, whether besieging forces permitted food or civilians to pass, and the fate of refugees all directly affected the scale of death.
The research must keep several levels distinct:
- existing food stocks and market supply inside the city;
- food requisitioning, rationing, and controls imposed by Nationalist military and local authorities;
- the scale of airdrops and outside supply and who received them;
- the scale of civilian flight;
- the besieging forces’ specific policies toward incoming food and outgoing civilians;
- the situation of refugees outside the city or between the two armies’ positions.
Reducing all of this to the sentence “there was a famine in the city” does not explain how the catastrophe developed or why it lasted for months.
4. The Republic of China side also requires historical scrutiny
Investigating the besieging force’s blockade policy does not absolve Nationalist forces or the Republic of China government of responsibility.
Before the full encirclement, Changchun already faced shortages of food and ammunition and growing defensive pressure. The Nationalist government subsequently used airdrops, food controls, and population evacuation, but never fundamentally solved supply or evacuation. Archival research cited by Academia Historica quotes Zheng Dongguo’s August 15, 1948 telegram to Chiang Kai-shek as saying the defenders needed 150,000 jin of food per day, while only about 1.45 million jin had actually been received by airdrop between May 31 and August 9. Air-force records separately give a figure of 1,457 tons dropped in June and July. The periods and accounting bases differ, so the two figures should not be directly converted or compared; their coexistence is itself worth examining. (Xu Hao, “The 1948 Battle of Changchun,” Academia Historica Bulletin, no. 71, 2022.)
A complete investigation must continue to ask:
- Why was the city held even as supply and relief prospects deteriorated?
- Why was civilian evacuation not undertaken earlier and on a larger scale?
- How did military requisitioning and food controls affect ordinary residents?
- As the chance of relief declined, how was the military value of continued defense weighed against civilian costs?
- Did authorities at different local and central levels have the ability to change policy but fail to do so?
These questions do not replace investigation of the besieging force’s blockade policy, and they do not cancel it out. Different responsibilities must be assessed separately on the evidence.
5. What questions must be asked about the Communist blockade?
For the Northeast People’s Liberation Army, research should focus on questions that can be tested against records: how operational orders described the siege; whether they explicitly ordered the denial of food or other civilian necessities; what policies applied to unarmed civilians leaving the city and to refugees already stranded near the blockade lines; which command levels received reports about hunger, death, and refugee conditions; whether there were discussions of relaxing the policy or opening passages; and how those policies changed over time.
These questions should be answered as far as possible through Northeast Field Army orders, telegrams, meeting records, operational logs, and contemporaneous materials from participants, rather than inferred solely from later memoirs or political narratives. One published document that provides a starting point is a September 11, 1948 telegram from Lin Biao, Luo Ronghuan, and others to Xiao Jinguang and others concerning the handling of refugees leaving Changchun, reproduced in Historical Materials of the Chinese People’s Liberation Army: Liaoshen Campaign (PLA Press, 1993).
No single civilian death toll for the 1948 siege of Changchun has been jointly validated by the major sources. This article therefore does not present either the highest or lowest estimate as established fact.
Published works and later studies contain estimates ranging from more than 100,000 to more than 300,000. These figures reflect different source bases, population denominators, periods, and definitions of death; none amounts to a verified name-by-name list. The estimates and their provenance are listed below, but none is adopted here as definitive.
Major reasons for the differences include:
— different definitions of the siege’s start and end dates;
— whether city residents, refugees outside the city, and people who died between the two armies’ positions are all included;
— whether disease, malnutrition, exposure, and combat deaths are counted together;
— differing prewar and wartime population baselines;
— incomplete household-registration, burial, relief, hospital, and refugee records;
— the fact that some estimates come from memoirs or later synthetic calculations rather than contemporaneous statistics.
This article uses the following formulation:
“The 1948 siege of Changchun caused mass civilian death, much of it directly related to prolonged food shortages, starvation, disease, and civilians being trapped in a wartime blockade environment. Estimates differ significantly, commonly ranging from above 100,000 to above 300,000. Until archival and demographic evidence can be cross-checked more systematically, FCM does not adopt any single figure as the final civilian death toll.”
FCM will not average competing figures or privilege a higher number merely because it has greater political impact.
Published estimates include the following. Each remains an attributed estimate, not an adopted conclusion:
— Central Daily News, October 24, 1948: “no fewer than 150,000 bodies” outside the city (a battlefield body count; methodology unspecified);
— Shang Chuandao, mayor of Changchun in 1946–48: 120,000 “died of starvation and illness,” said to be based on a post-entry count (published in 1985; underlying basis not independently verifiable);
— Zhang Zhenglong, White Snow, Red Blood (PLA Press, 1989): uses a 120,000 figure based on more than one hundred interviews and telegrams;
— Xu Yan of the National Defense University: about 120,000, based on a total of bodies buried by both Communist and Nationalist sides (2010 article; method described);
— Lung Ying-tai, Big River, Big Sea: Untold Stories of 1949 (2009): “from 100,000 to 650,000; take the middle and it is 300,000” — an averaging of an estimated range and therefore not an acceptable data point;
— Du Bin, Changchun Starvation War (2017): 370,000–460,000 (basis remains to be checked);
— Li Fasuo (2017): 58,063 deaths (a precise figure whose derivation remains to be checked).
The Changchun City Gazetteer records a decline from 611,000 people at the beginning of 1948 to 179,000 on October 19. That is net population decline, not a death count. Treating the difference as casualties is methodologically incorrect.
To verify any major estimate, researchers need to establish at least: who proposed it; where it first appeared; what population baseline it used; what period it covered; whether it included refugees, soldiers, disease deaths, and combat deaths; and whether independent sources corroborate it.
A more reliable death toll will require cross-checking 1947–48 Changchun population and household-registration records, pre- and post-siege population change, Nationalist rationing, requisitioning and evacuation records, Northeast Field Army blockade and refugee-handling orders, hospital and relief records, burial and body-disposal records, contemporaneous foreign diplomatic reporting, postwar local gazetteers, survivor testimony, and military archives from both sides.
Three questions must remain distinct:
First, how many people died? — a demographic-history question.
Second, how did they die? — a war and social-history question.
Third, which military and political decisions caused or aggravated those deaths? — a question of historical responsibility.
No single unverified number can answer all three.
6. How should the narrative of “peaceful liberation” or “victory without bloodshed” be reread?
Some later narratives emphasize that Changchun did not experience a final large-scale urban assault, using phrases such as “victory without battle,” “without shedding blood,” or “peaceful liberation.”
But “there was no final large-scale assault on the city” and “there was no mass civilian death during the siege” are entirely different propositions.
A siege is itself part of warfare. Food supply, population movement, refugee handling, and prolonged blockade can all produce severe humanitarian consequences. The cost of the 1948 siege therefore cannot be judged solely by how control of the city finally changed hands.
Any victory narrative should be tested against contemporaneous evidence, with civilian experience restored to the center of the historical record.
7. Legal assessment should be separated from historical fact-finding
This article does not at this stage classify the siege of Changchun as a “war crime,” “crime against humanity,” or “genocide.”
Any future legal analysis would need to distinguish: A. the domestic and international legal rules applicable in 1948; B. international humanitarian-law rules developed and codified later; and C. modern historical and legal assessments of siege warfare, civilian starvation, and restrictions on population movement.
Until the applicable law, elements of conduct, mental elements, and evidentiary chain have been systematically analyzed, FCM will not substitute modern legal labels for historical fact-finding.
8. Changchun needs a public, reviewable record of historical responsibility
FCM believes research into the 1948 siege should not stop at finding a death figure that serves a present-day political position. It should build an evidentiary record that outside researchers can review.
Priority materials include:
- Northeast Field Army siege orders and daily operational records;
- telegrams and meeting records involving Lin Biao, Luo Ronghuan, and the Northeast Bureau;
- records of the Nationalist First Army Corps, 60th Army, and New 7th Army;
- Republic of China Ministry of National Defense, Executive Yuan, and local-government records;
- 1948 diplomatic cables from the United States and other countries;
- food prices, rationing, and population records inside the city;
- refugee registration, relief, hospital, and death records;
- oral histories of survivors, defenders, and besieging forces;
- postwar party histories, military histories, and local gazetteers;
- private diaries, family letters, and burial records capable of verifying individual identities and deaths.
These sources should be cross-checked rather than selectively quoted.
9. What can the present evidence establish, and what remains uncertain?
At present, the evidence relatively firmly supports the following conclusions:
- Changchun experienced a siege lasting several months;
- supplies inside the city deteriorated severely;
- civilians suffered mass starvation, disease, and abnormal deaths;
- large numbers of civilians attempted to leave the city;
- military and administrative decisions by both besiegers and defenders directly affected civilian survival conditions.
Core questions still requiring further verification include:
- the precise total number of civilian deaths;
- the specific policies and actual implementation governing civilian passage through the blockade at different stages;
- what command authorities knew in real time about the scale of civilian deaths;
- how strong a causal link can be established between particular orders and particular deaths;
- complete data on the distribution of food between soldiers and civilians;
- the number of victims who can be individually identified.
Keeping these questions open avoids undermining genuine accountability with certainty that the evidence does not yet support.
Sources and research note
Source hierarchy:
- contemporaneous diplomatic records from 1948;
- military orders, telegrams, government archives, and population records;
- archivally grounded historical scholarship and university-press works;
- memoirs and oral histories, clearly identified as such;
- secondary syntheses only where stronger sources are unavailable.
Main source leads:
- U.S. Department of State, Foreign Relations of the United States, 1948, China volume: Ward’s June 15 Telegram No. 258 (doc 893.00/6–1548, history.state.gov/historicaldocuments/frus1948v07/d236) and the June 21 follow-up are directly cited here; other relevant cables include April 13 (doc 893.00/4–1348), August 11 (doc 893.00/8–1148), August 24 (doc 893.00/8–2448), and October 14 (doc 893.00/10–1448);
- Academia Historica archival scholarship: Xu Hao, “The 1948 Battle of Changchun,” Academia Historica Bulletin, no. 71 (2022), citing the Academia Historica “National Government” files and Chiang Kai-shek files;
- U.S. Army Military Review / Army University Press historical analysis of the siege;
- demographic and archival studies addressing casualty estimates;
- survivor accounts and documentary compilations proposing higher estimates, cited only with explicit attribution;
- Historical Materials of the Chinese People’s Liberation Army: Liaoshen Campaign (PLA Press, 1993), reproducing the September 11, 1948 refugee-handling telegram by Lin Biao, Luo Ronghuan, and others;
- Changchun City Gazetteer (1999), useful for population data but not for treating population decline as a death count;
- Changchun Municipal Archives, The Liberation of Changchun, 1948.10.19 (China Archives Press, 2009), for the October 19 chronology.
FCM’s method is not to treat any figure as a “final number” until its original source, statistical method, and population base have been independently checked.
Conclusion
Ordinary residents of Changchun did not decide to launch the Chinese Civil War.
They did not decide military deployments, the siege, or whether the city should continue to be held, yet they bore the gravest consequences of those military decisions.
More than seven decades later, the purpose of studying Changchun is not to replay a propaganda battle between the Kuomintang and the Communist Party, but to restore civilians to the center of the history.
Who issued which orders, who knew civilians were dying and when, who had the power to change policy but did not act — and how many people died, and who were they? Those questions should not cease simply because one side ultimately won the war.
A free society confronting its own history should not require victims to yield to a victory narrative.
Languages / 語言： 简体中文 · 繁體中文 · English · Français · Deutsch · Español