Chiang Kai-shek’s “secret deal” at Xian and the start of the Sino-Japanese War

Using newly available archives, particularly the diary and the presidential papers of Chiang Kai-shek, this article challenges the conventional interpretations of the Xian Incident (1936), in particular the widely held belief that the kidnapping of China’s leader Chiang by two rebellious generals forced him to form a united front with the Communist Party to confront Japanese aggression, and of the outbreak of the Sino-Japanese War 7 months later. It puts forth the interpretation that full-scale war between China and Japan was started not by Japan but by Chiang after a Japanese provocation, and the united front was only formed after Chiang ordered his best army units to attack Japanese forces in Shanghai in August 1937 turning it into the largest land battle after the First World War. It must be noted, however, that Japan acted provocatively and aggressively in a local incident outside Beijing a month earlier. Chiang decided on war not because he reached an agreement with the Chinese Communists to form a united front whilst a captive in Xian but because in Xian he received a signal from Josef Stalin that the Soviet Union would support him in a war with Japan. Chiang read Stalin right and the Soviet Union became the largest supplier of weapons to China in the first 4 years of China’s 8-year war with Japan. The hitherto unknown or “secret deal” Chiang made in Xian was an implicit one with Stalin, not with the Chinese Communist Party or its man on the spot Zhou Enlai.


Introduction
T he Xian Incident of December 1936 was an event of monumental importance. It caused the Chinese leader Chiang Kai-shek (Jiang Jieshi) to rethink and, indeed, reverse his priorities that had major long-term consequences. Until then he was committed to exterminate the Chinese Communist Party (CCP) as a prerequisite to building up the necessary national capacity for China to defend itself against Japan, and used compromises to slow down unrelenting Japanese aggressive moves in Northern China. At Xian Chiang found out that a workable alternative existed for confronting Japanese aggression, which gave him the option to stand firm against the next Japanese aggressive move. As a result, when Japanese forces in North China provoked the Marco Polo Bridge Incident in July 1937, Chiang did not appease the Japanese and took actions that transformed another local incident into the starting point of China's War of Resistance against Japan (1937)(1938)(1939)(1940)(1941)(1942)(1943)(1944)(1945).
This chain of events dramatically changed the fortune of the exhausted and heavily depleted CCP struggling to survive in the poor northwest of China. It enabled the CCP to revive and expand dramatically and exponentially during the war, ultimately seizing control of the mainland of China in 1949. Without the Xian Incident it is doubtful if full-scale war between China and Japan would have started in 1937, though it most probably could not have been avoided for long. Whether the CCP could otherwise have survived Chiang's "final push" in his extermination campaign cannot be known. What was certain was that the Incident enabled the CCP to reverse its fortune and get out of the blind alleyway to which it had been driven deeper and deeper by Chiang after 1934 (Chen, 2001). Even though Chiang simultaneously conducted parallel covert negotiations with the CCP as well as the Soviet Union prior to the Incident, his programme to exterminate the CCP only ceased as a result of the dramatic events in Xian. It was a turning point for the CCP. Even top Communist leaders like Zhou (1980) saw it at the time (a paper dated 29 December 1936).
Most existing scholarly works and personal accounts of the Incident focus on the intrigues and dramas in Xian, the CCP headquarters in Baoan, the Chinese capital Nanjing and the Soviet capital Moscow. 1 Even though there is not one agreed narrative of the Xian Incident, the prevailing view is that while the CCP did not stage the incident it played a pivotal role through Zhou Enlai in ending it. It is widely believed that Zhou secured the release of Chiang Kai-shek by getting Chiang to agree to form a united front with the CCP against Japan. This formed the basis for China to stand firm against Japan the following summer when the Japanese capitalized on an incident in the Marco Polo Bridge area outside Beijing to stage a more general aggression in Shanghai, in respond to which Chiang led the Chinese nation to resist in a war that lasted 8 years. This paper challenges this conventional wisdom.
Fascinating as the drama in Xian is, the real significance of the incident is how it affected the Chinese government's decision to respond when Japan made its next major aggressive move. Most works on the Sino-Japanese War gloss over how China finally made a stand against unrelenting Japanese aggression (Ch'i, 1982;Mitter, 2013;Lary, 2007;van de Ven, 2003;Ma, 1993). Even those that relate the Xian Incident to the start of the war works on the basis that the Incident "radically changed the position of the Nationalist Government, as it indicated that Chiang could no longer force his anticommunist campaign on his military subordinates" and seized on the "new mood of national unity" and "resolved to resist any further Japanese provocation" (Shyu, 2001). This overlooks the reality that Chiang could not decide to act on a matter with existential implications for the country just because the "national mood" changed. The reality was that while the Incident did change Chiang's sense of priority and that of his government there was a more powerful and decisive factor at work.
The long-standing official account of the Incident, based on what Chiang would like the world to think, asserts that the young Marshall Zhang Xueliang, who staged a mutiny and held Chiang captive in Xian, reversed course and freed Chiang as he was moved by how determined Chiang was to resist Japanese imperialism after reading Chiang's Diary. 2 This has rightly been discounted by Young-Tsu Wong. Wong also dismisses the alternative interpretation put forward by John Garver and Hans van de Ven that the Soviet Union played a key role in ending the kidnapping. Wang argues that Chiang was not "released by order of Moscow" but "because he met Zhang's demands" (Wong). In Wong's revisionist assessment although Chiang "could have disregarded what he had agreed to do in Xi'an … his sudden rise in popularity upon his release, due to his identification now with the rising tide of anti-Japanese nationalism, made it more than ever difficult for him to oppose resistance in favor of continuing the civil war" (Ibid.: 24). But others, from Yang Kuisong to Jay Taylor, prefer to stress the important role played by the Chinese Communist Party in ending the Incident (Yang, 1995;Taylor, 2009). Apart from Taylor none of the others examined in any detail how Chiang reacted in captivity and why he made a deal. Nor did they probe into the link between the Incident and the start of the war. Taylor's emphasis on the personal rapport between the Communist leader Zhou Enlai and Chiang as the key factor is, however, unconvincing. Chiang was too much of a realist and not enough of a sentimentalist for that. To settle the controversies over how the Incident was ended, one must examine how Chiang assessed his own predicament in captivity and the conditions for his release, and see how they influenced his decision to go to war with Japan the following summer.
Newly available sources, the most important of which are Chiang Kai-shek's diary at the Hoover Institution at Stanford University and his presidential papers at the Academia Historica in Taipei, have now make it possible to re-evaluate the deal that Chiang made in Xian for his own release. Also valuable is the Xian Diary of TV Soong (also at the Hoover Institution), as he was the key intermediary who did most of the actual negotiation between Chiang and his captors. 3 For this purpose Zhang's oral history account at Columbia University and various publications based on interviews with him after he regained freedom are not particularly illuminating, as Zhang refused to speak about the Incident in detail.
The deal was more complex than the mainstream interpretation that "Chiang's hand was forced in December 1936, when he was kidnapped by … Zhang … [and as a result] came to an accord with the CCP, and the two sides agreed to hold back from conflict with each other, while waiting for any further moves from Japan" (Mitter, 2004). On the basis of these recently available archival sources this paper sheds light on how Chiang assessed his predicament in captivity and secured his own release. It also examines how the settlement in Xian underpinned Chiang's confidence to risk going to war with Japan before a "united front" with the CCP was agreed.
As it will be explained, there were in fact two parallel elements to "the deal" that secured Chiang's release. One was negotiated between Chiang's agents and his real captors, which was what Young-Tsu Wong focused on. The other, and from Chiang's perspective the really important one, was an implicit understanding he thought he had with Stalin. There was no written agreement between Chiang and the CCP or the Soviet Union to form a united front against Japan though Chiang did make verbal undertakings to the Communists and his real captors. What Chiang thought was most important was, however, something else. It was that he had ascertained in Xian the intentions of the Soviet leader Josef Stalin. As he tried to work out a basis for his own release Chiang came to the conclusion that Stalin would support him in a war against Japan, a matter about which he did not have confidence hitherto. This was crucial to Chiang's decision to hold firm, half a year later, over an incident that involved the missing of a Japanese soldier on exercise in the Marco Polo Bridge area in the outskirt of Beijing on 7 July 1937. Since the missing soldier was found and tension eased, the scope for a settlement similar to what put an end to similar incidents in the previous few years existed. But Chiang decided to make a stand on this occasion, instead of appeasing the Japanese militarists as he had previously done. The big difference this time was that he now believed he could count on Soviet support in the event of escalation into a full-scale war.
The raison d'être of this paper is not to review the plotters' intentions, which cannot be established beyond doubt, 4 though it does ascertain carefully the sequence of events leading to the mutiny, as well as the role the Chinese Communist Party and the Soviet Union played in ending it. In so doing, it challenges the mainstream view in recent Chinese scholarship, represented in particular by the work of Yang Kuisong, that the Chinese Communists acted on their own and not on Stalin's instruction (Yang, 1995). By using the presidential papers and the personal diary of Chiang Kai-shek, and the accounts by Zhang Xueliang, it re-constructs the events that led to Zhang staging the kidnapping. While Zhang had contemplated such a drastic option for a few months the fateful decision was only made after a violent verbal confrontation between him and Chiang 2 days before the mutiny. By relying on the same original sources as well as TV Soong's diary for the Xian Incident it also re-examines the role Zhou Enlai played. It argues that Jay Taylor's conclusion that Zhou's special personal relationship with Chiang was pivotal needs to be set in the context of realpolitik (Taylor, 2009: 129-136). Zhou's diplomatic skills certainly helped to broker a deal but Chiang moved forward as he thought he had got an implicit understanding with Stalin, not because he had a special rapport with Zhou or trusted Zhou. This paper further examines the nature of this implicit understanding Chiang thought he had with Stalin and why something apparently so tenuous was subsequently allowed to leave marks of great historical importance. It shows that while a captive in Xian, Chiang believed that he was albeit indirectly and only implicitly dealing with Stalin. He thought Stalin gave him to understand that if he would lead China in a united front against Japan, the Soviet Union would provide support. It was Chiang's belief that Stalin would deliver on this understanding that gave him confidence to start a second front in Shanghai after the Marco Polo Bridge Incident and commit the overwhelming majority of his best army against the Japanese in the largest land battle since the Great War. When he did so he also hoped this would persuade the West to seek to protect its own interests in Shanghai by brokering a peace favourable to China. But he was willing to risk escalation to full-scale war only because he was confident that while the West would probably fail him Stalin would support him. As it turned out once the Japanese found it necessary to reinforce massively the Shanghai front against unexpectedly strong and effective Chinese military operations, the crisis escalated to a point that full-scale war could no longer be avoided. This paper further shows that the "second united front" between the Kuomintang and the CCP was not agreed at Xian but was hastily agreed after the Battle of Shanghai had already started.

The incident
Chiang Kai-shek was personally aware that Zhang Xueling, leader of the Xian mutiny, had had dealings with the CCP almost immediately after Zhang and the CCP started their secret dialogues in January 1936, more than 10 months before the Xian Incident. 5 Indeed, it was the failure of forces under Zhang's command to attack the remnants of the CCP in Northwest China that took Chiang to Xian on 4 December 1936. Chiang believed that by personally superintending Zhang's command, he could require the Northeast Army (under Zhang's command) and the 17th Route Army (a corps size unit under former Northwest Army general Yang Hucheng) to co-operate with Central Army units in the region to deal a final crushing blow to the remnants of the CCP forces.
The mutiny and kidnapping took Chiang completely by surprise, though the risk that the Northeast Army might rebel did cross his mind in late November (Shilue gaoben, 25 November 1936). When he was captured the first issue about which he sought clarification was whether his captors were from the Red Army (Zhu, 1994, 79; account of Sun Mingjiu). It was not until his captors explained they were from Zhang's elite bodyguard regiment and were escorting him back to their commander that Chiang realized Zhang had really staged the mutiny. Chiang did not know Yang was a co-conspirator until he saw soldiers of the 17th Route Army on duty in the streets of Xian as he was being taken to Yang's headquarters, where Zhang and Yang were (Shilue gaoben, 12 December 1936).
Even though the attitude of the Soviet Union under Stalin and the CCP leadership were critical to finding a solution to the Xian Incident the mutiny was staged without their prior knowledge (Zhang). 6 This was notwithstanding the reality that the two key conspirators (generals Zhang and Yang) had significant parallel dealings with the CCP, and had hoped to secure Soviet military supplies if they could form some kind of a grand alliance including the CCP as a partner in Northwest China. The idea of holding Chiang captive and forcing him to stop the civil war against the Communists and lead them in a War of Resistance against Japan was originally put forward by Yang (Shilue gaoben, 15 December 1936;Dou, 2005). Zhang was at first hesitant.
Zhang, who lost his father to Japanese military assassins in 1928, had for some time been highly dissatisfied with Chiang's determination to exterminate the Communist Party before resisting Japan. He had repeatedly tried to remonstrate with Chiang. His last attempt before Chiang arrived at Xian took place at the end of October 1936, when he did so with the backing of the strongly entrenched leader of Shanxi Yan Xishan (Zhu, 2006). 7 Chiang rebuffed them. This was the end of the matter for Yan, but Zhang gained the mistaken impression that Yan would support him if he should act against Chiang's central government (Zhu, 2006: 228-229).
What finally prompted Zhang and Yang to seriously consider this extreme measure was Chiang's decision to take overall personal charge in Xian. This put them on the spot (Yang, 1995: 281). They were presented with a choice of either implementing Chiang's order to attack the CCP forces, with which they had a secret truce agreement, or be re-deployed to Fujian and Anhui, respectively, and face the consequences later. Zhang and Yang thus explored the option of kidnapping Chiang to force him to abandon his "pacify the country first" policy and adopt instead their "resist the external threat first" approach. They did not reach an agreement to act in the first 6-7 days of Chiang's stay in Xian.
The final decision to seize Chiang (and those senior generals who accompanied him to Xian) was made after Zhang and Chiang ended their meeting on 10 December in an emotionally charged verbal altercation that revealed how irreconcilable their basic difference was. 8 This took place the day after Zhang went to meet with students who went on a demonstration demanding the government resist Japanese aggression. Anti-Japanese sentiments among the educated population had been running high in the major cities of China in the autumn of 1936 (Sun, 1993). When Zhang met student demonstrators on 9 December he was moved by their patriotism (Wang et al., 2005: 130; memoirs of Ju Yiqiao). Knowing that security forces had been ordered to suppress them if necessary Zhang pleaded with them to disperse and promised them that within a week he would do something to show he had not let them down (Wang, 2004: 307-308).
At their meeting the following day, on 10 December, Zhang tried to persuade Chiang as he had promised the student demonstrators. Chiang abusively dismissed Zhang's recommendations as naive and touched a raw nerve that had been exercising Zhang. Chiang reportedly said the students who disturbed the peace and through their demonstrations sabotaged the government's policy should be suppressed-by the use of firearm if necessary (Ibid.: 319; Zhang speaking to Tang Degang after regaining freedom). Chiang himself remorsefully noted that evening that he had been too harsh and rough on Zhang, but he had no inkling how great an impact this emotionally charged meeting had (Shilue gaoben, 10 December 1936). It led to Zhang deciding that the only way to persuade Chiang was to kidnap him and force him to agree to reverse the policy priorities. There is no evidence that Zhang had made up his mind before this dramatic verbal confrontation.
Zhang wanted Chiang to lead the whole country to stand firm against an aggressive Japan before consolidating control in the country. 9 The hastiness of the decision was reflected in the deployment of troops in Xian and in executing the coup. Zhang did not have time to move units of his substantially bigger and more powerful Northeast Army to Xian, which would have made sense if Zhang had decided to stage the mutiny sooner. 10 The security of Xian at the time was in the hands of Yang's 17th Route Army, not a long-standing ally of Zhang. Within the city Zhang had only one bodyguard regiment of the Northeast Army (Zhu, 1994: 52; Wang Yuting's account). 11 When he finally decided to act Zhang took on the role and responsibility as the leader and invited Yang to serve as the deputy leader of the mutiny. Even though Yang had a substantially larger force in Xian that could be deployed to seize Chiang, Zhang sent his own relatively small unit of bodyguards instead.
Yang did not order his men to join Zhang's bodyguards in this critical operation. One cannot absolutely and totally rule out the possibility that the two leading mutineers trusted each other though it was unlikely. A much more likely explanation is that Zhang did not trust Yang's forces with Chiang's life, as he merely wanted to capture Chiang, not to kill or harm him. 12 Zhang's plan suited Yang as it gave him plausible deniability in case the mutiny should fail. In such an eventuality Yang had the option to turn on Zhang with superior force, putting all blame on Zhang, and stay on the right side of Chiang.
Zhang's troops tried to seize Chiang at his Xian residence in the early hours of 12 December, less than 48 hours after their oral altercation. As his small detachment of bodyguards put up resistance, Chiang escaped by scaling a wall, badly hurting his back as he fell into a ditch in the dark and sought refuge on a hillside. He was found and taken prisoner before 9 o'clock in the morning.
Chiang's initial reactions to his kidnapping took Zhang and Yang by surprise. Instead of opening negotiation, Chiang adamantly refused to discuss anything with them and declined food. In the moment of shock Chiang appeared to think at least as much about his place in history as his immediate safety. Whether it was in the hill where he was captured or in Yang's headquarters Chiang insisted first and foremost on his dignity not being offended. 13 No doubt he must have privately calculated that he was too great a prize for Zhang and Yang to execute on the spot. In any event his intransigence left Zhang and Yang in a quandary. Zhang's plan to force Chiang's hand got nowhere. They promptly sought advice from the CCP and specifically requested Zhou Enlai, whom Zhang greatly admired, to help move the matter forward in Xian (Dou, 2005: 169). 14 When the CCP leaders in Baoan learned of the incident by telegram around noon on 12 December, they were both shocked and jubilant. Mao Zedong reportedly laughed wildly while some, including General Zhu De, thought it best to kill Chiang immediately (Zhang, 1974(Zhang, : 1237Chen, 2001: 320). In the end it was decided that the Party must keep Zhang and Yang on side while it sought direction from the Comintern-and Mao drafted the telegram to Moscow.
In his memoirs Zhang Guotao, who was present in Baoan as one of the top leaders though not in a dominant position, recalled that Zhou Enlai was dispatched to and arrived in Xian on 13 December, and the Comintern reply (drafted by Stalin himself) was received that same evening (Ibid.: 1240-1242). 15 According to Zhang, Stalin saw the Incident as staged by the Japanese to cause a civil war in China and reduce its capacity to resist Japanese aggression, and the Soviet Union must not fall into this trap. Stalin considered Chiang the only person who could lead China to resist Japan and he instructed the CCP to ensure the Incident would be resolved by forming a united front with Chiang as the leader (Zhang, 1974(Zhang, : 1240. 16 Zhang's account is credible as we now know Stalin did see that it was in the Soviet Union's interest that Chiang should lead China to resist Japanese aggression so that the Japanese militarists could not turn their attention to the Soviet Union. Zhang's account has, however, been challenged by the Chinese historian Yang Kuisong who insists that the key decision to resolve the Xian Incident peacefully and by working with Chiang was reached by the CCP leadership before Stalin's telegraphic instructions were received (Yang, 1995: 322-332). 17 He has provided evidence that the telegram from Stalin was not in fact received until 3 days later, on 16 December which was also the day when Zhou actually arrived in Xian though Yang got the latter's date wrong himself. 18 He also suggests that the telegram could not be deciphered and one that could be read was received on 20 December, the day after the Politburo meeting in which the key decisions were made. An independent source confirms that telecommunication between Baoan and Moscow, which was only established in June that year, was unreliable for some months and thus the possibility of genuine difficulties in telecommunication with Moscow happened cannot be eliminated (Kampen, 1989, Chiang Diary, 14, 24 and25 December 1936). 19 This contradicts the interpretation of Hans van de Ven who suggests unequivocally that the telegram caused the CCP to make a "u-turn". According to him Mao Zedong and Zhou Enlai decided on 17 December to reverse the CCP's 15 December position of demanding Chiang be handed "over to a people's tribunal" into one of calling "for a peaceful settlement and the protection of the safety of Chiang" ( Van de Ven, 2003: 187). In this matter the reliability of van de Ven's interpretation is marred by the reality that it would have been very difficult, if not impossible for Zhou and Mao to have reached a joint decision and acted together on 17 December as Zhou was by then advising Zhang and Yang in Xian while Mao was physically in Baoan. To act together they would have to co-ordinate by telegrams, which were controlled by Zhang and Yang at the Xian end. Van de Ven has not provided any evidence to suggest telegraphic coordination took place or that Zhang and Yang made telegraphic facilities in Xian available to Zhou for private and confidential communication with Baoan. While Baoan could and must have used this facility, which was monitored by Zhang and Yang,  an anti-Japan united front and the CCP should seek a peaceful resolution of this incident. 20 As to Yang's version no explanation has been provided, by him or anyone else, as to why it took the CCP 2 days before it asked Moscow on 18 December to resend this long awaited for and apparently extremely important telegram, after asserting that this telegram could not be deciphered. 21 This could not have been a procedural failure or oversight by telegraphic clerks since several reminders were sent to Moscow before Stalin's reply was first received. Likewise, Yang has not explained why the CCP decided on this matter on 19 December, after having decided to seek Comintern direction first and requested the re-sending of Stalin's instructions the previous day. No one has explained the compelling reason why the CCP must decide how to react on 19 December, rather than wait for Stalin's telegram to be transmitted again if the original could not be read at all. 22 Whether Mao and the CCP leadership decided to reverse within less than a week their initial consensus over Chiang-to put him on a show trial-under a direct telegraphic order from Moscow or not, they acted as Stalin had suggested or desired. The reality was that even if the key message in Stalin's original telegraphic reply could not be read, the views of the Communist Party of the Soviet Union had already been publicly articulated, authoritatively in an editorial of the Pravda newspaper, the official organ of its Central Committee, on 14 December (Yang, 1995: 322). 23 Unless the CCP allows unrestricted access to its archives of this period in full it is not possible to establish beyond doubt if the most important part of Stalin's telegraphic instructions were successfully deciphered before the CCP reversed its policy or changed its position. However, there is no compelling reason to dismiss the thrust of Zhang Guotao's recollection though he misremembered the exact date-for both Zhou's arrival in Xian and receipt of Stalin's telegram, two events that did, as he correctly recalled, happen on the same day. 24 Apart from saying both events took place 3 days earlier than they did, which made no difference to the sequence of the major events he recounted, there is no significant inaccuracy in the rest of his recollection (Chen, 2001: 320, Shilue gaoben, 10 December 1936). 25 Indeed, according to Zhang Xueliang, Zhou knew Stalin's position when they met in Xian as Zhou told him that Stalin did not approve of the incident and wanted Chiang to remain the leader of China (Zhang,Vol. 39: 25).
On the issue of whether the CCP acted on Stalin's instruction John Garver takes the view that Mao probably came to realize the great value of fomenting a war between the Kuomingtang regime and the Japanese. Garver thinks this might have contributed to the change in the CCP's position. This could be the case. An alternative or additional, as the two are not mutually exclusive, explanation would be that the key message in Stalin's telegram came through sufficiently clearly, particularly if read in juxtaposition with the Pravda editorial. The request, transmitted 2 days later, for resending Stalin's telegram was to ascertain if there were specific details Stalin outlined as the basis for a peaceful settlement, not the basic policy. 26 Hence, the CCP leadership felt confident to reverse its earlier decision even before Stalin's telegram giving the specifics was received again. Whether Mao changed his mind before he received Stalin's telegraphic instructions in full or not, "the fact remains that Moscow did intervene and that intervention helped bring about the reversal of the CCP's initial radical position" (Garver, 1991: 173).
Changed priority All through his captivity and for two decades afterwards Chiang thought the Communists were behind Zhang and Yang in staging the mutiny and kidnapping (Li, 2007: 12; record of Zhang's interview with Kuo Guanying and Tang Degang, June 1989). Indeed he could not believe that Zhang staged and led the mutiny when his bodyguards first reported to him that they were under attack from soldiers of the Northeast Army in his Xian residence. 27 Chiang's belief that the Communists were behind this had a significant impact on how he assessed his predicament and worked out a solution.
The Xian Incident did have a profound impact on Chiang. He had to confront the prospect of death as soon as he was found by rebel soldiers. In the first meeting Chiang had with Zhang as the latter's prisoner, Chiang felt he could not dismiss the prospect of being executed after some sort of a show trial (Ibid.). Although unknown to him at the time, this was essentially what the CCP decided as their preferred outcome in an enlarged Politburo meeting held the following day, on 13 December (Yang, 1995: 299-301). It would be unusual if Chiang did not consider this possibility. In contrast, his repeated demand that his captors should either shoot him or treat him with dignity and respect was almost certainly meant to assert his moral authority and to strengthen his hand if he should negotiate subsequently as he lost control of the situation. They were not indications that he wished, as it is implied in his published account of the incident, to provoke his captors to kill him. 28 But the threat to his life was real and was confirmed in the assessment of TV Soong, who went to Xian on 20 December to seek a peaceful resolution. Soong concluded that "the Generalissimo's life was in the greatest danger" (Soong's Xian Diary, 1936). Chiang could not but reflected hard on what he had done and on alternatives that he could have chosen.
Chiang undoubtedly also pondered what drove Zhang, the most patriotic of all former warlords whom he treated as a protégé, 29 and an "old comrade" Yang Hucheng to such an extreme measure. 30 Although Chiang asserted that Zhang became repentant as Zhang was moved by his determination to resist Japan after reading his diary, the reverse was almost certainly closer to the truth (Chiang Diary, 14 December 1936). 31 Zhang did read Chiang's diary and noted entries showing Chiang's determination to resist Japanese aggression but it had little impact on him (Zhang,Vol. 84,51). Zhang could not have been moved as Chiang claimed, as Zhang's problem with Chiang was not over intention or determination but over priority-resist Japanese aggression first, or wait until the CCP had been exterminated (Li, 2007: 4; record of Zhang's interview with Kuo Guanying and Tang Degang, June 1989). Even after Zhang had released Chiang and voluntarily escorted Chiang back to Nanjing to face the consequences of his act of mutiny, 32 Zhang still insisted to Chiang that the most important of the eight conditions Zhang initially raised-reorganize the government-must be implemented (Shilue gaoben, 27 December 1936; see below for the list of the eight conditions).
In Xian it was Zhang's willingness to take responsibility and his obvious sincerity in requesting Chiang to take leadership of a national united front to resist Japan that must have forced Chiang to reflect. 33 It did not take long for Zhang and Yang to disagree intensely over how to move forward. Chiang noticed this as he noted, in his diary, Zhang's repeated efforts to ensure soldiers from his own bodyguard regiment, rather than Yang's men who were in control of Xian, would take charge of his personal safety (Chiang Diary, 14, 24 and 25 December 1936). The patriotism of Zhang and the articulated wish of the nation to stand firm against further Japanese aggression could not have escaped him as Chiang examined his predicament. 34 While this did not amount to compelling pressure on Chiang to reverse his policy priority and accept the terms of his captors, he had to ponder whether adhering strictly to his policy priority was worthy of dying in the hands of the mutineers or the Communists. Should he be killed it would almost certainly result in a civil war that would facilitate Japanese aggression against China. 35 In the early days of captivity Chiang had to confront seriously the prospect that he might not emerge alive.
It was not until the fourth day of his incarceration, the day when the government in Nanjing started bombing the rebels from the air, on 16 December, that Chiang felt his personal safety was less at risk. This was because, in his assessment, Zhang's remark that he needed 4-7 days of peace to work out a solution meant Zhang must be seeking direction from Moscow. In his diary he noted: "my heart was at ease as the Soviet Union could not endorse their rebellious conduct and the Soviet Union never took Zhang seriously" (Chiang Diary, 16 December 1936;my translation). On the following day as the rebels reduced their specific demands from eight to four, Chiang thought this was the result of instructions from the Comintern (Chiang Diary, 17 December 1936). The fact that Chiang's analysis was wrong since neither was due to instructions from Moscow is irrelevant. What was critically important was that Chiang believed the rebels would ultimately follow Stalin's instructions. Whilst still a captive, Chiang confided in his wife, who had gone to Xian on 22 December to secure his release, that the key was held not by Zhang whom she knew well but by the Communists (Shilue gaoben, 23 December 1936).
We now know, from his diary and papers of the 1930s, Chiang was determined to resist Japanese aggression, and he devoted a huge amount of resources to prepare for the eventual fight with Japan. Indeed, before the Xian Incident, Chiang had been negotiating with the Soviet Union for about a year for Soviet support for China's resistance against Japanese aggression in return for a political solution to the CCP. 36 While the Chinese government had hoped to get a mutual defence treaty with the Soviet Union, the latter preferred a non-aggression pact (Garver, 1991: 153). The Soviet refusal to sign a mutual defence treaty raised doubt as to its commitment to help China in the event of full-scale war between China and Japan.
Prior to the Xian Incident, Chiang had to focus on securing Soviet weapons and support in the event of war with Japan as he could not get a mutual defence treaty from Stalin (Chen, 1994). 37 The fact that such negotiations were taking place was important. 38 In the year-long covert negotiations between Stalin and Chiang both took a utilitarian and calculating approach and "each was wary of the other's intentions" (Garver, 1988: 45). Before he became a captive in Xian, Chiang was not certain that the Soviet Union would support his government should it go to war with Japan. With Chiang believing in this juncture that Stalin's intervention would be vital to his release he would take any message or powerful gesture from Stalin seriously.
Such a gesture indeed materialized. During Chiang's captivity Zhou Enlai let him know that his son, Chiang Ching-kuo (Jiang Qingguo), who went to the Soviet Union in 1925 to study but had been kept as a hostage in the Soviet Union, would soon be allowed to return to China. 39 Once this was mentioned to Chiang, he would have seen it as confirmation of the correctness of his analysis that he really was dealing, albeit in a shadowy way, with Stalin. However Zhou might have articulated it Chiang could not have failed to see that only Stalin, not the CCP, could promise the release of Ching-kuo. The issue of Ching-kuo's release had in fact been used as a bargaining chip earlier in the decade. In June 1931, an agent of the Comintern's Far Eastern Bureau, Hilaire Naulen, was captured by Chiang's government, which mistook him to be the head of the Far Eastern Bureau. The Soviet Union made an offer to trade the release of Naulen and his wife for the return of Ching-kuo, but this was rejected by Chiang (Yang, 2002). The CCP was not involved in this negotiation. Thus, when Chiang was informed of Ching-kuo's impending release, he could not but have seen it as an important message from Stalin. Thinking he knew Stalin for what he was, it would be reasonable for Chiang to think Stalin was essentially telling him that they should look at the big picture in finding a way to end the Xian Incident. 40 As to what Stalin wanted specifically Chiang could only surmise, as he chose not to negotiate directly with Zhou Enlai whether Zhou could actually represent Stalin or not. From Chiang's perspective what Stalin wanted would have to be something in the strategic interest of the Soviet Union. He could reasonably have concluded that Stalin wanted him to live and lead China to resist Japanese aggression so that Japan could not turn its attention to the Soviet Union or, at least, get pinned down in China by the Chinese Army supported by the USSR. What he was willing to concede was to open or, rather, re-open negotiations with the CCP to form a united front, something about which he had had parallel negotiations with the CCP for a year (Mitter, 2013: 64). In return he wanted an implicit Soviet promise to provide support in their resistance against Japan in the event of war. This was not an unreasonable assumption for him in light of the covert negotiations his representatives had with the Soviet Union and the CCP in parallel. Chiang understandably preferred to negotiate an agreement after his release rather than under duress in Xian, though he had made a moral commitment to find a way to reach an agreement subsequently.

The deal
The deal that Chiang made to secure his own release consisted of two parts. They were a commitment along the line of what Chiang speculated Stalin wanted, and the specific verbal agreement that Chiang's brother-in-law TV Soong (who held no government office at the time) and Madam Chiang (nee May-ling Soong) negotiated with Zhang, Yang and Zhou.
Chiang adamantly refused to get personally drawn into the latter negotiations and authorized the Soong siblings to reach an oral agreement with them. Chiang was willing to reverse the priority between fighting Japanese aggression and suppressing the Communists to avoid a civil war and his own death on two conditions. The first was that he was not made to appear to yield to his captors to save his own skin. For this reason Chiang would not negotiate directly with Zhang, Yang and Zhou though he endorsed the principles underlying the verbal agreement with them. He would not overtly do a deal under duress or compromise on maintaining his sense of dignity. 41 The second condition was that the Communist Party must declare their support for his government. An understanding could be reached as Zhou knew Chiang's sensitivities and showed great diplomatic skill.
Although Zhou had repeatedly asked to see Chiang, Chiang declined to meet him until the evening of 24 December. By then the Soong siblings had already reached a verbal agreement with Zhang, Yang and Zhou. This meeting happened because Zhou insisted on it. Chiang finally conceded but on the condition that it was only a meet and greet affair (Shilue gaoben, 23 December 1936). Chiang's record shows that this meeting did happen and no negotiation took place (Shilue gaoben, 24 December 1936). But Chiang reversed his previous decision not to have any exchange of substance with Zhou the following morning. Since the only event that happened between the evening of 24 December and the morning of 25 December was the "meet and greet" affair, this would have to be the occasion when Zhou passed on, most probably by way of polite pleasantries, the all important message from Stalin that Chiang Ching-kuo was well and would be returning to China soon (Shilue gaoben, 24 December 1936). What else could have caused Chiang to change his mind over the issue of speaking to Zhou personally? Indeed, Chiang met Zhou again in the presence of TV Soong in the morning of 25 December, the day of his actual release. This was the only meeting in which Chiang agreed beforehand that he would discuss with Zhou the basis for future cooperation and, in reality, the terms of his release. According to Soong, Zhou explained: [F]or one year the communists had tried to avoid fighting in order to preserve the national strength. They have not made any capital out of the Sian incident and the measure suggested were the same that they put forward months ago. Now they want his personal assurance (1) to stop剿共[jiaogong or exterminate the Communists]; (2) to enrol the communists to fight Japan; (3) to be allowed to send some representative to explain to him in Nanking. 42 In Chiang's own personal account, which was not published and was undoubtedly written to cast himself in a positive light but its substantive points are corroborated in Soong's diary entry, he told Zhou: You must know that your wish of I saying "No extermination of the Communists from now on" is something that I cannot say now. You have to know that I have devoted my whole life to achieve national unification and the centralized command of the armed forces of the nation … If you and your colleagues will not undermine national unity, will obey the Central Government, and fully accept my command as part of a united army, I will not only not seek to destroy you but will also treat you in the same way as other army units. 43 To this Zhou simply replied: "The Red Army will accept Mr Chiang's command, will support unification under the Central Government, and will not undermine it" (Ibid.; my translation). 44 Notwithstanding the positive gloss put on this exchange it probably captured the spirit of the exchange. 45 This understanding could not have been reached so easily without Zhou's diplomatic flair and his insistence on passing Stalin's message in the "meet and greet" meeting. Zhou gave Chiang time to ponder and work out the meaning of Stalin's message. He ensured Chiang did not feel his sense of dignity and authority had been affronted. As Soong noted in his diary, after this brief exchange Zhou left and said to Soong "the Generalissimo was exhausted so he would not talk too much to him but that there are certain things" to be confirmed between him and Soong, adding that "as an old subordinate he knows the Generalissimo is a man of his word" (Soong's Xian Diary, 25 December 1936). The basic deal was, in plain language, to open negotiations after Chiang's release for a united front against Japan under Chiang's leadership, with details of arrangements to be worked out later and on the understanding that Soviet assistance would be available in the event of war with Japan. In other words, Chiang and Zhou did not reach an agreement to form a second united front between the Kuomintang and the CCP in Xian.
As to the more specific, and from Chiang's perspective secondary, part of this deal, the starting point was based on the original eight demands that Zhang and Yang announced immediately after the kidnapping and without the input of the Communists. They were: 1. Reorganize the government in Nanjing … to enable all parties to join in and take part to save the nation. 2. Stop all civil wars. 3. Immediately release all patriots arrested in Shanghai.
4. Release political prisoners in the whole country. 5. Allow popular movements that promote the national salvation movement. 6. Protect the political rights and freedom for all people to form political parties and hold demonstrations. 7. Truly implement Sun Yat-sen's will. 8. Immediately convene a national salvation meeting (Zhu, 1993;Telegram from Zhang and Yang to Liu Xiang, 12 December 1936).
With the injection of the CCP's own demands and Yang's concern over whether Chiang would go back on whatever agreement he might reach once he returned to Nanjing, a somewhat different set of ten points were agreed in the final unwritten understanding. Reached between Zhang, Yang and Zhou on the one side and the Soong siblings on the other they were, according to Zhou: 1. Kong Xiangxi (HH Kung) and Soong will form a new government that excludes pro-Japanese individuals. 2. Withdrawal of Central Army units from the northwest. 3. Release of "patriotic leaders". 4. End the campaign to exterminate the CCP. Rename the Red Army and put it under a united command when a War of Resistance against Japan is started, 3 months later. 5. Reorganize the Kuomintang within 3 months and involve others in the government. 6. Staged release of all political prisoners. 7. CCP to be allowed to operate openly when a War of Resistance is started. 8. In terms of foreign policy: ally with the Soviet Union and liaise with Britain, the United States and France. 9. Chiang to take responsibility for the Incident and resign from the office of Premier. 10. The CCP agrees to support TV Soong against the pro-Japan faction within the central government and keep a secret agent in Shanghai to liaise with Soong (Zhou, 1980: 72-73; Report to the CCP dated 25 December 1936). 46 After he returned to Nanjing, Chiang did not keep all the specific promises his agents made to Zhang, Yang and Zhou. But he honoured the brief verbal undertaking he himself made to Zhou. Chiang appeared to have made a distinction between the two though the latter was in effect included in the former. The latter had to be implemented for two reasons. It was a matter of personal honour for him. More importantly, this implicit undertaking to reverse the national priority was predicated on Stalin agreeing to come to China's assistance in a future war with Japan, something he had wanted for some time. He treated the rest of the unwritten understanding agreed to by the Soong siblings as a list of additional policy changes to be implemented if possible. This suggests the apparently changed atmosphere in the country was not enough to require Chiang to fulfil all the terms for his own release.
In dealing with Zhang and Yang, Chiang tried to appear magnanimous. Although the kidnapping was for all intents and purposes a mutiny none of its leaders was given a death sentence. Zhang, the leader of the mutiny who voluntarily surrendered himself, was promptly tried and sentenced to 10 years imprisonment in a court-martial (Sima, 1989). Chiang then requested clemency and put him under the supervision of the Military Commission, which in effect meant putting Zhang under house arrest in a comfortable setting (Shilue gaoben, 4 January 1937). 47 Yang, who remained in Xian, tried to negotiate specific arrangements that would protect his power base. After the Northeast Army imploded as an organization in the absence of Zhang and units of his own 17th Route Army switched allegiance to Nanjing in early 1937, Yang was required to relinquish his command, and travel overseas at government expense for 6 months (Fu, 1989). He was removed from the Army List, and was put under de facto house arrest after he returned to China. 48 Although working out the future of Zhang and Yang was politically delicate, it was largely a matter of striking a balance between pre-empting future mutinies and projecting an image that Chiang was fair, honourable, strict with military discipline and yet apparently magnanimous. Chiang could not afford to let the rebels keep the northwest outside of the central government's authority after staging a mutiny. It would otherwise have rekindled residue warlordism. He achieved his objective by a careful deployment of central government forces to put pressure on the rebels-a clear violation of one of the terms for his release-and allowed internal tension among the rebels to unravel their coalition. In this Chiang largely succeed, which implied the non-fulfilment of some of the key terms Zhang and Yang imposed for his release.
The truly monumental decisions Chiang had to make after the Xian kidnapping concerned what to do with the CCP, China's relations with the Soviet Union and the national expectation that his government would stand up to further Japanese aggression. All three issues were closely intertwined. Indeed, if Chiang thought the formation of a united front with the CCP was not critical to securing Soviet aid, there would have been no greater need for him to honour his pledge to Zhou than for the understandings reached with Zhang and Yang. While Chiang did feel his sense of honour should require him to uphold the gentlemen's agreement he did break other promises. In the end, as will be explained below, it was his belief that national interest took precedent that caused him to treat the CCP differently for considerations over the big picture.
Decisions leading to war Chiang had been determined to stand up against Japan and remove repeated Japanese humiliation at least since the Mukden Incident of 1931 when the Japanese Army forcefully seized control of Manchuria, though he also saw the need to accommodate repeated Japanese aggressions until China had built up the capacity to resist (Shilue gaoben, 21 and 22 September 1931). 49 While the outburst of popular patriotism during the Xian Incident must have affected him, it does not mean he had as a result made up his mind to go to war when the Japanese make their next aggressive move. To act on the articulated patriotism would have been emotionally easy for him. But as a hardnosed realist whose decisions had implications for national survival he could not afford to act on emotion or popular expectation. In terms of his determination, it should be recognized that even though he kept this secret he had already ordered in October 1936, before the Xian Incident, the stockpiling of aviation fuel, pre-positioning of ammunition in strategic locations and planning by government departments to relocate inland (Shilue gaoben, 6 October 1936). The efforts to build defensive structure in the event of a conflict with Japan focused on the Shanghai-Nanjing area, with additional defensive installations also built further up the Yangzi River all the way to Wuhan (He, 2011). The critical issues for him after the Xian Incident remained, however, how to secure the resources to resist Japan, and when and where to stand and fight.
After 1931 Chiang had worked on the basis that the Chinese government needed several years to build up its military and political capacity and to eliminate the CCP and residual warlordism to stand united against the might of Japan. This was the fundamental consideration behind the policy of "pacifying the country before resisting the external threat" (攘外必先安内) (Yang, 2002: 393-399). The Xian Incident gave him an opportunity to rethink this strategy. Starting in early 1937 Chiang gave higher priority to the avoidance of a civil war (Chiang Diary, 5 February 1937).
But the reality continued to be that China did not have the military and logistical capability to defeat Japan. Even its capacity to resist a major Japanese military incursion over any period of time was limited. A top-secret document among Chiang's papers reveals that in 1937 the Chinese Army had only 190 million bullets in stock, whereas by the calculation of its planners the Army would need 300 million bullets to sustain 50 divisions in a month of full-scale war. 50 The only way that China could resist a major Japanese incursion was for it to secure external support and supplies. Chiang thought that the democratic countries like Britain, France and the United States would offer no help in the event of China going to war with Japan (Mitter, 2013: 76-77). Hence, to reverse China's defence priority he must first get the Soviet Union committed to support it in war-knowing full well that the price was to find a political, rather than a military, solution to the CCP.
To proceed, Chiang needed to ascertain in parallel the intentions of both Stalin and the CCP leadership. The attitude of Stalin was pivotal, as the real issue was the reliability of Soviet military assistance in the event of war. The reaffirmation of Stalin's goodwill and willingness to help materialized in April, as the Soviet Ambassador Dimitri Bogomoloff informed Chiang that "the Soviet Union would provide China with military hardware should a war with Japan break out" (Yu, 2006). This was quickly followed by the promised return of Chiang Ching-kuo to China. 51 Chiang did not immediately respond to these Soviet moves as the Soviet Union still insisted on signing a mutual non-aggressive treaty, which would have required the Chinese Government to accept the Soviet Union's privileged position in Outer Mongolia and over the Chinese Eastern Railway (Sun, 2006). Chiang dropped reservation over this after the start of the Marco Polo Bridge Incident. On 8 July 1937, the day after the incident started, Chiang pondered if the incident would turn into "a total War of Resistance", for which China was not yet ready. As he did so he decided the "most critical" issues were to reach "an agreement with the USSR for the supply of military equipment and the conclusion of a Soviet-Chinese treaty" (Garver, 1987). 52 The attitude of the CCP was also significant as Chiang conceded the CCP "should be accepted back into the fold under certain conditions, but it must be kept within defined perimeters" (Chiang Diary, 5 January 1937). He was not prepared to allow the CCP use a war against Japan to rebuild its political and military might so that it could seize power after the end of the war. It was with this in mind that Chiang insisted on several conditions for forming a united front with the CCP, viz.: no military command above divisional level (which was a roughly 10,000 men unit); no maintenance of a territorial base; and the transformation of the CCP forces into regular units of the National Revolutionary Army of specified strength (Chiang Diary, 6 and 10 March, 17 May and 7 June 1937). This marked a great concession on Chiang's part as the maximum number of Communist troops he had previously been willing to allow in the secret negotiations that took place prior to the Xian Incident was only 3,000. 53 The other matters involved were relatively easy to fudge but the CCP refused to accommodate Chiang's core demands. As late as June 1937 Chiang still tried to persuade Zhou that the CCP should accept the central government's bottom line so that an agreement for a united front could be reached (Shilue gaoben, 9 June 1937). Prior to the Marco Polo Bridge Incident Chiang thought the CCP was insincere in its offer to cooperate but he saw no alternative to ARTICLE PALGRAVE COMMUNICATIONS | DOI: 10.1057DOI: 10. /palcomms.2014 negotiation (Chiang Diary, 15 March 1937)-as long as he was still seeking a Soviet commitment to help China in the event of war with Japan.
In the end the Marco Polo Bridge incident happened before Chiang could complete his negotiations with Stalin and the CCP. It forced his hand. Chiang had to make choices and monumental decisions with little time to consider their long-term consequences.
Although the initial incident was contained after the missing Japanese soldier was found, Mao surmised that Chiang would have to back down in his conditions for forming a united front if the Japanese forces resumed hostilities (Sheng, 1992). On 27 July, the date the Japanese Kwantung Army launched a new assault against Beijing after a lull Mao thought war could no longer be avoided and instructed Zhou not to negotiate further with Chiang until the latter started making concessions (Jin, 1998: 443). Mao's assessment was accurate-Chiang noted in his diary that evening that he did not see how war could be avoided (Huang, 2012).
On 22 August, the day after Chiang's government signed a non-aggressive pact with the Soviet Union and 9 days after fighting started in Shanghai, Chiang quietly conceded to Mao's terms for a united front (Sheng, 1992: 169). But no formal agreement was signed. The Communist forces became the 8th Route Army simply by an instruction issued by Chiang. Chiang accepted "the united front" as a price to secure Soviet military aid.
In fact when the Marco Polo Bridge Incident happened Chiang adopted a dual track approach. He did not simply decide to stand and fight. On the one hand he tried to settle this new Japanese provocation by negotiation but only if diplomacy could restore the status quo ante (Chiang Diary, 14 July 1937). On the other hand, mindful of Stalin's promise of support, he prepared for war and was not willing to accept humiliating terms similar to those that had settled previous incidents. Thus, he quickly took steps that put China's armed forces on a war footing and prepared to open a second front if the Japanese militarists should push further. The latter materialized when the Japanese assaulted Beijing on 27 July and seized it the following day (Chiang Diary, 27 July 1937).
In addition to rushing reinforcements to North China Chiang deployed the most elite divisions of the Central Army, the German trained and equipped 87th, 88th and 36th divisions, to Shanghai. As he did so he violated the 1932 agreement that settled the Shanghai Incident of 1931. 54 On 11 August he instructed them to advance to designated assault areas. Chiang ordered them in the evening of 13 August to launch a full scale attack on Japanese Marines in Shanghai and open a second front in an area where he had made the best preparation for war (Benton and Tsang, 2006;He, 2014;: 235). In so doing he took a risk. Counting on Soviet support-albeit at a price-Chiang was prepared to fight a "long" War of Resistance but he also hoped that a robust defence of Shanghai, where the Western powers had vast interests, could get them to intervene and force Japan to agree to a settlement less prejudicial to China's interests. In the end the battle captured the imagination of both combatant nations and marked the point of no return in transforming the Marco Polo Bridge Incident into the full-scale war of Resistance.
Indeed, the day after the Marco Polo Bridge Incident Chiang sought to strengthen China's capacity to fight by securing Soviet support. On 8 July Chiang approached the Soviet Union for a mutual defence agreement, which he had long sought (Sun, 2006: 147). After the battle of Shanghai started he asked the Soviet Union if she would consider joining the war (Yang, 2008). 55 Stalin rejected both requests as his motive was to tie the Japanese down in China so that they could not attack the Soviet Union. He had no wish to get the Soviet Union directly involved in the war. However, Stalin did fulfil his side of the bargain that Chiang thought was implicitly raised in Xian, though not without conditions. The Soviet Union responded to Chiang's request for support by insisting on signing a non-aggressive pact as a condition for providing weapons to China (Chiang Diary, 2 August 1937). Once large scale fighting with the Japanese started Chiang could no longer afford to worry about the longterm implications for Chinese sovereignty in Outer Mongolia and the Chinese Eastern Railway. He authorized the signing of such an agreement on 21 August, less than 10 days after fighting commenced in Shanghai (Qin, 1981; full text of nonaggressive treaty dated 21 August 1937). In return the Soviet Union delivered 400 aircraft and other matériel valued at over 100 million Chinese yuan by the middle of October (Sun, 1993: 112).
Indeed, Soviet aircraft and pilots challenged the Japanese dominance of the sky after the Chinese Air Force suffered horrendous losses by the end of the battle of Shanghai (Chiang Diary, 2 October and 28 November 1937;: 259). To illustrate the scale of Soviet support in the early years of the war, Chiang requested matériel to equip 20 divisions of the Chinese Army from the Soviet Union in 1938 alone. 56 In the course of the Wuhan battle in the third quarter of 1938, Stalin indicated that he would consider providing arms for up to 60 divisions and an additional 500 aircraft (Shilue gaoben, 6 October 1938). Chiang did not get everything he requested but Soviet matériel was critical to sustain China's war efforts up to March 1941 (Qin, 1981: 295). It was provided to serve Soviet interest and formally paid for by Chinese strategic minerals (Yu, 2006: 16-17). The additional informal price Chiang had to pay was to allow the CCP to continue to exist. The most important assumption Chiang made over the intentions of Stalin when he was a captive in Xian was vindicated.

Conclusions
Although the attitude of Stalin was critical to the resolution of the Xian Incident, Hans van de Ven is not fully justified to pronounce that the CCP simply reversed itself on receiving a direct telegraphic order from Stalin to do so. Stalin's attitude undoubtedly influenced how the CCP handled the incident. It was also seen by Chiang as the most important factor in securing his own release. But van de Ven's contention is not supported by available evidence. Young-Tsu Wong's claim, from the opposite spectrum, that Chiang was released not by order of Moscow but because he met his real captors' terms contradicts evidence available as well. Equally, Yang Kuisong's assertion that the CCP made the key fateful decision on its own is unconvincing. The reality was that the CCP and Mao reversed their initial position to one that dovetailed that of Stalin within a week. The weight of evidence suggests the top CCP leaders did not make this change under a direct telegraphic order from Stalin on16 December but did so after they came to know Stalin's views.
One must not lose sight of the reality that Stalin's attitude represented at most one of the key factors for a solution. It would not have been decisive if Chiang did not react as Stalin hoped he would. Respond Chiang did. Once he saw that Stalin had signalled him to look at the big picture he found it possible to endorse in front of Zhou an agreement his agents reached with his captors and Zhou. If TV Soong's assessment in his diary is correct even as late as 24 December, the day Zhou passed on Stalin's message that Chiang Ching-kuo would soon be allowed to return to China and 1 day before Chiang's release, the deal could still have collapsed and Chiang might have refused to endorse it personally. Such an assessment was corroborated by Chiang's own private record (Shilue gaoben, 24 December 1936). Whether Chiang would have done a deal with his captors if he had not received Stalin's signal cannot be known with certainty. But once Chiang thought he got confirmation of Stalin's intention in the evening of 24 December, he stopped being reticent in dealing with his captors who were prepared to release him. He cooperated with Zhang and had a substantive conversation with Zhou that secured his own release within 24 hours. Yang, who was concerned about his own safety after Chiang returned to Nanjing, was left in the cold by both Zhang and the Communists.
Zhou's diplomatic flair greatly facilitated this process but the evidence available does not confirm the popular belief in China, reiterated by Jay Taylor, that Zhou played an indispensible and pivotal role. Neither Zhou, nor the CCP nor Stalin could have required Zhang and Yang to release Chiang. What really made a settlement possible were Stalin's astuteness in signalling Chiang his commitment to support China in a war against Japan, and Chiang's success in reading and responding positively to this signal. Zhou's greatest contributions were two. First he insisted on having the "meet and greet" meeting on 24 December and thus created an opportunity to pass on Stalin's message. The second was to assess correctly what Chiang would tolerate in how to endorse the deal without appearing to have done so under duress, and thus maintained his sense of dignity and vanity. An alternative Communist representative might not have Zhou's touch but as long as he did not offend Chiang directly, it is reasonable to conclude that Chiang, who was always looking at "the big picture", would still have tolerated the deal, once Stalin's message and "sincerity" was conveyed.
The Xian Incident did have great consequences. To be sure it did not change Chiang and make him suddenly determined to fight Japan. It also did not produce an oral agreement for the forming of a second united front between the Kuomintang and the CCP, as has been taken for granted. Reaching an understanding with Zhou Enlai that he would, after his release, negotiate with the CCP did not amount to reaching an agreement. Previous negotiations between them did not produce anything like a united front and the risk of failure remained real. The reality was that Chiang had been determined to resist Japanese aggression for several years, and had had covert negotiations with the Soviet Union and the CCP as he prepared to confront forcefully Japanese aggression. What the Incident did do was to induce Chiang to alter his priority as he changed the assessment of the likelihood of Soviet support in the event of a war with Japan.
Before Xian Chiang preferred to unify the country, which included exterminating the CCP, to get China ready to go to war against Japan. After the Incident he decided that he and his government should take a robust stance against the next aggressive move of the Japanese militarists. For this he needed external support. The only realistic source was the Soviet Union.
Notwithstanding the year-long covert negotiations that preceded the Xian Incident it was as a captive in Xian that Chiang felt sufficiently confident that Stalin was looking at the big picturehow to confront and survive the rising ambition of the Japanese militarists against their respective countries-as he was. He thought he got confirmation of this when Stalin made the grand gesture of promising to release his son Chiang Ching-kuo. It showed that Stalin wanted him alive and would support him to lead China in a war resisting Japan. Chiang did not reach a specific and clear agreement over this with his Communist interlocutor Zhou Enlai but he could see the Soviet Union's self interest would make such an implicit deal a real commitment. But since the price for Soviet support and his own release was for Chiang to form a united front with the CCP, he agreed in principle to work for one after his release.
While the Xian deal was critical to Chiang's decision to make a stand over the Marco Polo Bridge Incident, the former did not lead to the latter. Up to 7 July 1937 Chiang had reached no agreement with the CCP (and did not implement roughly half the terms for his own release in Xian), let alone make joint preparations, in anticipation of an incident like that which took place in the Marco Polo Bridge area. The Incident took him by surprise. His reactions were not based on a pre-conceived plan. The changed political situation and atmosphere, unfolding events and above all his new assessment of Stalin's intentions guided Chiang in his responses. Under his leadership China drifted into the War of Resistance as his attempt to open a second front in Shanghai escalated out of control. The War of Resistance was not the result of an agreement reached with the CCP at or shortly after the Xian Incident to form a united front to wage a full-scale war against Japan.
After Xian Chiang focused on finding a way to incorporate the CCP into the mainstream politics of China on the conditions that the CCP would adhere to the Three Principles of the People, give up holding on to a territorial base and integrate its military forces into the National Revolutionary Army. Chiang was willing to let CCP members and soldiers to join in the war that he was going to wage against Japan in due course. But there was no set timetable. Chiang agreed to let the CCP survive for a while but not to revive. This was the price he knew he had to pay to secure Soviet military aid in the event of war with Japan. Chiang still tried to eliminate the CCP as a threat by political means (Yang, 2009).
But the CCP had a different agenda. It was not willing to be assimilated by the KMT or be integrated into a KMT led and dominated political framework. Chiang was still negotiating with the CCP when the Marco Polo Bridge Incident happened. In spite of the public rhetoric Mao did not respond to Chiang's appeal to patriotism. He saw an opportunity and seized on the Japanese aggression to force Chiang's hand.
Chiang had to make a choice. He decided to stand and fight unless diplomacy could restore the status quo ante, and urged the CCP to be patriotic and support the war efforts in the spirit of the proposals he had already put to it. Mao capitalized on Chiang's predicament and insisted on his own terms. As events unfolded Chiang made significant concessions to the CCP to form the united front to secure Soviet matériel. Chiang could afford to make a stand against the Japanese as he felt confident that Stalin would supply him with weapons and ammunition, provided he would pay the price Stalin wanted vis-à-vis the CCP.
By allowing the CCP to keep its territorial base, independent existence and for its forces to be organized into a corps of three divisions called "the Eighth Route Army" with its own general headquarters Chiang basically accepted Mao's terms. In return he obtained vital Soviet weapons and military supplies that sustained China in the early years of the war. Chiang brought his regime scope to survive the Japanese onslaught at the price of, with the benefit of hindsight, allowing the CCP to revive and prepare to contest the future of China after the end of the War of Resistance against Japan.
The release of Chiang's presidential papers in Taipei and his diary in the Hoover Institution has been indispensible in helping to explain why Chiang did what he did. This in turns provides a good impetus to review and rethink some of the assumptions of the history of Republican China, in this case the causal relationship between the Xian Incident and the start of the War of Resistance against Japan, as well as the Nanking government's assessment of the relevance and significance of the CCP and the Soviet Union. 53 Shilue gaoben, 21 February 1937. CCP accounts suggest Chiang was willing to tolerate a much larger force. 54 For the Shanghai Incident and its settlement, see Jordan (2001). 55 Chiang did not give up hope until at least December 1937, and possibly, according to John Garver, the middle of 1938 (Chiang Diary, 5 December 1937;Garver, 1987: 305-313). 56 Shilue gaoben, 5 January 1938; and Tejiao dangan, 080106-063, 08A-01742, Table  showing Main weapons and equipments needed for 20 divisions, 1938.