Career Patterns in the Ch’ing Dynasty

Front matter

THE OFFICE OF THE GOVERNOR GENERAL

KENNETH G. LIEBERTHAL AND RICHARD H.

ROGEL CENTER FOR CHINESE STUDIES

Chapter II: Chapter IIT: Chapter IV: Conclusions

Appendix 1:

Appendix 2: Appendix 3: Appendix 4:

Notes Glossary Bibliography

Historical Survey and Powers of Office Ethnie Composition and Dynastie Control Career Patterns

Professional Mobility: Determinants of Success and Failure

Percentage of Complete Data by Subject and Ethnie Group

Memorials by Lin Tse-hsti and T'ao Chu Official income of Governors-general Examples of Regulations and Penalties for the Conduct of Governors and Governors-general

111 131

2.1

2.2

3.1 3.2

3.3 4.1 4.2

4.3

4.4

4.5

4.6

List of Tables

Number of Governor-generalships by Period and Politieal-ethnie Group, 1644-1911

Number and Duration of Governor-generalships Held by

Political-ethnic Group, 1644-1911 Governors-general with the Chin-shih Degree Ch'ing Local Officials Who Began Their Careers with Inherited Positions (by ethnie group)

Local Officials by Ethnie Group, 1796-1908 Average Age at Attainment of Chin-shih

Average Age at First Appointment as Governor-general

Length of Tenure in Provincial Posts

Percentage of Governor-general Tenures Ending in Demotion or Dismissal (general averages, and by political-ethnic group)

Percentage of Governor-general Tenures Ending in Demotion or Dismissal (by region) Number of Governor-general Tenures Ending in Demotion or Dismissal (by period)

45 53

57 66 74

Acknowledgments

We are most grateful to Professor Evelyn Sakakida Rawski, who read the manuscript at an early stage and offered a number of extremely useful suggestions. Ms. Nancy Evans offered editorial assistance throughout the preparation of the original manuscript. Ms. Patricia Saywell helped in the preparation of the final editorial changes. Mrs. Ruth Maloney of the University of Toronto and Mrs. Joan Pearson at Simon Fraser University offered graciously of their time and talent in much of the typing of the manuscript.

We gratefully acknowledge the financial assistance of the University of Toronto in the research for the book and in its preparation for publication.

Finally we should like to claim that our editors, Ms. Janis Michael and Ms. Janet M. Opdyke, must be paragons of patience and good spirit.

ix

Introduction

The office of governor-general (tsung-tu) was the highest provincial post throughout the Ch'ing dynasty. As such, it was a vital link in the control of a vast empire by a very small and alien ruling elite. This is primarily a biographical and statistical analysis of the incumbents of that office. By analyzing the biographical data of those who held the position of governor-general, much may be learned about the nature of the office itself. However, the main objective of the study is to provide information on career patterns, that is, the variety of different posts held from the first official appointment to that of governor-general, of an important cross section of successful Ch'ing bureauerats. By plotting and analyzing the different patterns their official careers took, we should be able to determine what kind of men reached the top of China's provincial and national administration during the final centuries of China's imperial history; the qualifications that were required; the factors which prompted rapid promotion or sudden disgrace. We should also be able to determine the extent to which these and other factors varied markedly among Manchu, Mongol, Chinese Bannerman, and Han incumbents and whether changes throughout the dynasty can be detected in policies concerning the office or in the career patterns of its personnel. If such detection is possible, this study may lend support to the view that late imperial China was not static, but a society undergoing significant change.

A number of important studies in Japanese, Chinese, and Western languages have appeared that are either devoted to a number of the problems with which we are concerned, or that contain important insights on these issues. While some of these are more exhaustive in their analysis of certain issues, none, to the best of our knowledge, concentrate specifically on the office of governor-general or study it with as broad a sweep of interests as is attempted in this analysis. A good deal of general information on the office itself is available in such studies as Ch'ü T'ung-tsu's Local Government in China under the Ching, and the earlier work of Hsieh Pao-chao.! No study of Ch'ing government is possible without frequent use of Thomas Metzger's indispensable book The Internal Organization of Ching Bureaucracy.

Important early works such as John King Fairbank's study on the Manchu- Chinese dyarehy have analyzed, at least for short time periods, the problems faced by the alien Manchu rulers in controlling the large, ethnically distinct, indigenous population of Han Chinese and in implementing a dyarchy, problems xi xii of particular importance to the office of governor-general.? Valuable insights into the office and some of its better-known incumbents, especially those of the nineteenth century, are available in such studies as Franz Michael's work on the Taiping Rebellion and Ssu-yü Teng's study of the Nien Rebellion.? Also of great value are: J. Y. Wong, Yeh Ming-ch'en: Viceroy of Liang Kuang, 1852-58; Liu Kwang-ching, "Li Hung-chang in Chihli: The Emergence of a Policy, 1870-1875," in Albert Feuerwerker, et al., eds., Approaches to Modern Chinese History; Daniel Bays' article in Modern Asian Studies, "The Nature of Provineial Authority in Late Ch'ing Times: Chang Chih-tung in Canton 1884- 1889," as well as his book-length study of Chang, China Enters the Twentieth Century: Chang Chih-tung and the Issues of a New Age, 1895-1909; and S. A. M. Adshead's work in Papers on Far Eastern History (The Australian National University), "Viee-regal Government in Szechwan in the Kuang-hsii Period (1875-1909).". On many questions related to career patterns or factors encouraging or impeding social mobility, there is an enormous amount of literature already in existance, including a study by R. M. Marsh and the seminal works of Ho Ping-ti.4

More specific studies on the office of the governor-general and its incumbents throughout the Ch'ing dynasty normally have included the office of the governor (hsiin-fu) as well and have tended to limit their range of interest to matters related to the ethnie identity of those who held the posts (collectively known as tu-fu). Although, as we shall deseribe briefly below, the distinctions between the offices of governor-general and governor were few and generally unclear, we have not included governors in this study for several reasons. Our work is primarily a case study of career patterns. If its main purpose was a thorough analysis of the office of governor-general, more attention to the role of governors would have been required to elucidate as fully as possible those elements that distinguished the governor-generalship from the governorship. However, for a study of career patterns, we believe the governors-general alone offer a sufficiently large sample to make the data statistically significant and our interpretation of them reliable. They offer a sample, moreover, for which biographical data is a good deal more complete than for the much larger group of governors. Therefore, in most areas of our statistical analysis we have greater confidence in the reliability of our data on the governors-general, and the extrapolations we make from them than we do in the comparable data for governors. Finally, while the responsibilities of the two offices were deliberately overlapped to provide a system of checks and balances, the governors-general were in rank and status at the top of the provincial hierarchy. (Normally the governor-general held the rank of 2A, while the governor's rank was 2B, though either might hold more senior ranks by virtue of posts held concurrently in the central government.°) Unlike governors, who frequently had the final step to the governor-generalship yet to make, the governor-general was more likely to retire or be promoted to one of the highest offices xiii in the capital after reaching the pinnacle of provincial service. Thus, his eareer pattern was more likely to be near completion than that of a governor.

The first book-length study to attempt a reasonably complete statistical breakdown of both sets of officials throughout the dynasty, emphasizing echanges in ethnic composition by period, was published in Taipei in 1963 by Fu Tsung-mao.® Unfortunately, Fu's study is based on a sample of 571 governors and governors-general rather than all who held either post and therefore tends to distort the data on many of the issues of concern to us. His sample was drawn from the group of office-holders who warranted biographical entries in the Ch'ing-shih kao and Ch'ing-shih lieh-chuan, and these men, by definition, were more successful than their excluded colleagues.

An important early article by Lawrence D. Kessler, "Ethnic Composition of Provincial Leadership During the Ch'ing Dynasty,” appeared in the Journal of Asian Studies in 1969.” While our statisties do not always agree, our cónclusions on aspects of common concern generally do. Kessler's study is unlike ours in that it includes governors, although for purposes of many of his analyses he isolates the governors-general. It also differs from our analysis in that it is limited to the question of ethnic composition.

Our master list of governors-general with initial ethnic identification was compiled from the Ch'ing-shih and checked against other sources such as Yen Mao-kung's 1931 compilation, Ch'ing-tai cheng-hsien lei-pien; the official banner history, Pa-chii t'ung-chih; Ch'ien Shih-fu's Ch'ing-chi chung-yao chih-kuan nien-pao; and Wei Hsiu-mei's Ch'ing-chi chih-kuan piao. Biographical data was obtained from one or more of the Ch'ing biographical collections such as Ch'ing-shih lieh-chuan, Kuo-ch'ao ch'i-hsien lei-cheng ch'u-pien, Pei-chuan chi, Hsii pei-chuan chi, Pei-chuan chi pu, and Kuo-shih lieh-chuan. In addition, data was also culled from the biographical sections of the Ch'ing-shih, Ch'ing-shih kao, Pa-ch'i t'ung-chih, Liao-hai ts'ung-shu (which contains some biographies of Manchus who are not included in the Pa-ch'i t'ung-chih), and local gazetteers. Obviously many of the most important governors-general appear in Hummel's Eminent Chinese of the Ching Period, and some of those who held the office very late in the dynasty appear in Boorman's Biographical Dictionary of Republican China.

The collected writings (wen-chi) of leading governors-general were also used for checking biographical information, particularly the chronological biography (nien-p'u) and posthumous biographical writings included in the collected writings of some governors-general. The nien-p'u provided dates of birth and death, and recorded when examinations were passed, appointments received, and official duties taken up, as well as the reasons for administrative punishments, important events in the official's career, and key information on family background.®

Most collected writings have a section which consists of memorials written by the subject. These often will include two particularly valuable kinds of xiv memorials. One type, the hsieh-en che, was written by an official on the occasion of his appointment to express his gratitude to the emperor. These provide the date on which the appointment was to take effect. The second, the tao-jen che, reported the official's arrival at the new office. As officials were frequently appointed to posts that were never actually assumed, these memorials are particularly valuable in verifying that a post was indeed taken up and in providing the date upon which the new appointee arrived at his office.

The memorials of governors-general were extremely important for other reasons. Official works such as statutes, the regulation of boards, and imperial ediets provide a picture of what was supposed to happen, but do not always provide a reliable description of what actually took place. The memorials offer a more complete picture of the way an official functioned, how he got along with colleagues and subordinates, the limits of his power, and his relationship with the emperor.” In some collected writings, information in the memorials is supplemented by copies of official correspondence, dispatches, and comments on the reports of subordinates; these add yet another dimension to the portrait of an official in action.

A large number of rather arbitrary and discretionary decisions had to be made in the process of compiling and analyzing the data on which this study is based. For most categories of analysis, the complete sample was broken into three political-ethniec groups: Han Chinese, Manchus, and Chinese and Mongol Bannermen. We have followed standard practice by considering the bannermen as a single group for purposes of comparison with the Han Chinese, particularly where questions of ethnic control, loyalty, and dynastic confidence or lack of it were involved. In areas for which sufficient information was available, a separate analysis of the Mongol data was performed. However, as only thirteen Mongols served as governors-general throughout the dynasty and complete biographical information is available for only some of them, the sample was in most instances too small to analyze separately. In some areas information on the Mongols was excluded from consideration because it seemed to distort the analysis rather than clarify it. We also separated the complete sample into groups according to the native provinces of the governors-general, which provided a basis of comparison within the Han Chinese group.

In compiling our master list of 504 governors-general, we adhered to customary procedure by including the governor-generalships of water conservaney and grain transport. We excluded those who served only as acting officials and those who were given the full title but appear for one reason or another never to have actually assumed the post. While these exclusions may have distorted certain aspects of the analysis, it seems to us that their inclusion would have presented more significant aberrations. As a result, our complete sample does not coincide with that used in similar studies such as Kessler's. 10

KV

Data was gathered for the twenty categories listed in appendix 1, but availability varied markedly according to the ethnic group, period of tenure in office, and native province of a governor-general. For the entire sample, information was more complete in some areas than in others. For example, in over 90% of the cases, information was available on degrees held or other initial qualifications, the post held immediately prior to a man's first appointment as governor-general, and reasons for eventually leaving the office. This contrasts sharply with the category in which information was least complete, namely, ages of appointment and retirement, for which the availability of complete data on our entire sample was only 37.5%. Nevertheless, it was encouraging that for the full sample, 37.5% was the lowest figure for completeness of information. This percentage would be recognized by most analysts engaged in this type of research as more than adequate for such a large sample. Indeed, age of appointment and retirement was the only one of twenty categories of bigraphical information in which, for the entire sample of governors-general, the amount of relatively complete data fell below 60%. In the majority of categories the figure was well above s096. 12

On the other hand, complete biographical data in all categories of analysis over the entire sample was available for only 22.6% of the governors-general. The fact that this figure varied from 37.8% for the Han Chinese group to 15.4% for Mongols, 14.6% for Chinese Bannermen, and 10.8% for Manchus, increases the possibility of distortion in our comparative statistical analysis. However, these differences lose some significance in light of the relatively complete information obtained for each group in our twenty categories of data (see appendix 1). There, the lowest percentage of available information occurs in the age of appointment and retirement category for all ethnic groups, where availability varies from 58.5% for the Han Chinese to roughly 32% for Chinese Bannermen, 31% for Mongols, and 14% for Manchus. However, in the vast majority of categories, information return for all groups is well over 50%. Thus, both in absolute and comparative terms, our analysis should prove generally reliable. In the case of the Han Chinese group, all twenty information categories have over 50% complete data. For Manchus, only two categories (age of appointment and retirement, and years between first official appointment and first appointment as governor-general) report less than 50% complete data. Chinese Bannermen have three categories with complete information below 50%, the above two, plus number of years between the last degree and the first appointment as governor-general (32%, 42.7%, and 43.7% respectively). The small Mongol group has a total of five data categories with less than 50% complete information, but in four of these the percentage is over 45%.

In terms of differences among Han Chinese representatives, the least amount of data was available for those who came from the Northeast, Kansu, xvi

Chihli, Kwangtung, Anhwei, Kiangsi, and Hupei. It is difficult to read any significance into this, and low availability probably occurs because a relatively high number of men from these provinees had less than distingushed careers, and most of them served in the early years of the dynasty, for which biographical information is more scarce. A much more serious problem was the relative weakness of biographical data for the earliest periods of the dynasty. This implies problems of possible distortion against which no foolproof check was available, since the early years saw a predominance of the Chinese Bannermen and the later years a predominence of the Han Chinese. Generally speaking, our analysis increases in reliability as the dynasty wears on.

Despite the problems inherent in this kind of study, we have chosen not to adopt any hard and fast rule for evaluating the reliability of our information on the basis of sample size and the ratio of complete information. Rather, we have sacrificed the attraction of rigorous and consistently applied absolute standards, and our criteria for maximum discretionary judgments are based essentially on common sense. Normally we do not assign a high degree of reliability to figures based on less than 30% information availability, and percentages higher than 30% are considered suspect if the sample involved is very smali. Usually, when we have noted trends or even presented hard statistics for which we feel the sample is too small or the information only marginally adequate, we have stated our reservations in the text. In some eases we have attempted to check statistical patterns against the patterns which emerge from larger samples based on estimated information. For instance, in the case of age of first appointment as governor-general, the figures for the sample with actual ages given was checked against another sample based on estimated age. In these cases, age was estimated by using the average age of attainment of the chin-shih degree for each ethnic group, providing, of course, that the year the degree was awarded is known. In most eases, the pattern that emerged from such estimates tended to confirm the analysis done on hard data. In any event, interpretations based on or influenced by this kind of estimate are always clearly noted in the text.

The data in this study were gathered and evaluated entirely without the assistance of a computer. Errors that may be discovered, mathematical or interpretive, are thus entirely man-made, and we make no attempt to "pass the buck" to technology or technician. Despite these many problems and pitfalls, we are confident that the general patterns and trends we trace are accurate, even if not all the markings along the way are as clear to view or free of error as we would wish.

CWHTK CSWP CWHTK HCCSW HP HT

HTSL LPCFTL LPTL

SL

THL

Abbreviation of Sources

Chiing-ch'ao hsit wen-hsien t'ung-k'ao Huang-ch'ao ching-shih wen pien Chiing-ch'ao wen-hsien t'ung-k'ao Huang-ch'ao ching-shih wen hsii-pien Ta-Chling hui-tien (1733 and 1899 editions) Ta-Chiing hui-tien shih-li (1899 edition) Ch'in-ting ch'ung-hsiu liu-pu ch'u-fen tse-li Ch'in-ting Li-pu tse-li

Ta-Chling li-ch'ao shih-lu

(Shih-i ch'ao) Tung-hua lu xvii

Chapter 1 Historical Survey and Powers of the Office

In the Chinese provincial bureaucracy, the offices of governor-general (tsung-tu) and governor (hsiin-fu) were collectively known as tu-fu. As we shall see the scope and purpose of the tu-fu system changed over time. Generally speaking, however, each province (except Chihli and Szechwan) had at the top of its adminstration a governor, and, with some exceptions, a governor-general was put in charge of two provinces. These two officials had direct control over all aspects of provincial or local administration (terms often used interchangeably). The provincial hierarchy included district magistrates, sub-prefects, prefects, intendants of circuit (taotai), and judicial and financial commissioners, as well as other posts related to the examination system and salt monopoly. The governors-general and governors had military, judicial, and financial powers and were generally responsible to Peking for the peace, order, and good government of the provinces over which they had jurisdiction. The relationship between the two officials as well as changes in their duties over the course of time will be described below.

By the time of the Ch'ing dynasty, the governor-generalship, sometimes. referred to by Europeans as the "vice-regal," was the highest office outside the capital. It originated, however, centuries earlier during the Ming dynasty. In 1441 a rebellion broke out at Lu-ch'uan in Yunnan Province and the minister of the Board of War, Wang Chi, was dispatched to direct the campaigns against the rebels with the additional title of tsung-tu chiin-wu (in complete charge of the direction of military affairs).! Thereafter, a number of high-ranking, central-government officials were given this title when on military missions in the provinces. From the reign of the Ching emperor (1450-57) on, all high-ranking, central-government officials sent to the provinces on military assignments were given the title tsung-tu.

During a southern military campaign led by the emperor Wu-tsung (1506- 22) the emperor called himself tsung-tu chiin-wu. It then became inappropriate for officials to take this title and, in place of it, those sent to the provinces on military missions were called tsung-chih (to control all). In the middle of the Chia-ching reign (1522-67) it was decided that the term "chih" (to control) would not be used by officials and the title tsung-chih was changed back to tsung-tu. <

The office of governor appears to have originated in 1391 when the Hungwu emperor dispatched the heir apparent to "hsiin-fu" (tour and soothe) the province of Shensi. In 1421 the Ch'eng-tsu emperor dispatched twenty-six high officials to hsun-fu different parts of the empire. When an official returned to the capital and submitted his report, his hsiin-fu commission was complete. The governors sometimes held additional titles such as fu-chih liu-min (to soothe and settle the refugees), tsung-li ho-tao (to superintend the river works), cheng-ch'ih pien-kuan (to manage the border defense), or t'i-tu chiin-wu (to superintend the military affairs).” Later, officials were sent as imperial delegates to designated areas and by 1430 this custom "fell into a stable pattern." In the following years, governors in various provincial, special frontier, or strategic locations began to serve as resident coordinators for longer and more indefinite periods of time.

As military crises dictated regional coordination on a broad scale, the governor-generalship system became more institutionalized. The need for this kind of coordination was described by a Ming author noted for his writings on governmental institutions.

When a disturbance erupted in one province, the governor-general could mobilize the military supplies and soldiers of the neighboring provinees to make a concerted effort to attack the rebels from the front, block them from the rear, and press them from the left and right. .. . After this was over, troops were sent back to their [original] provinces. ... There is a saying that it is more effective to hit with the fists than to fillip the ten fingers.4

While the position was essentially a military one, governors-general were also appointed to supervise river conservancy and grain transport.® The first resident governor-general was appointed in 1452 when a rebellion of the Miao tribe broke out in Kwangsi, and the government wanted to maximize coordination between that province and neighboring Kwangtung. In 1465 the governor-general of Liangkuang (Kwangsi and Kwangtung provinces) was concurrently governor of Kwangtung.® Most other governor-general appointees during the Ming were responsible for more than two provinces, for example, one official had jurisdiction over Szechwan, Shensi, Honan, and Hunan, and another was responsible for Chekiang, Fukien, Kiangsu, and Kiangsi.!

The practice of appointing governors-general became increasingly common in the second half of the fifteenth century. Many remained in the hinterland for long periods of time and served concurrently as governor of one of the provinces under their jurisdiction. After 1453, any minister or vice-minister sent to the provinces, either as governor-general or governor, was given the concurrent title vice-censor-in-chief to enhance his status vis-a-vis the provincial censors. This was the court's reaction to its belief that in earlier years the governors-general had feared the censors and gone too far out of their way to please them. These Ming practices began to change the offices of governor and governor-general from temporary commissions in the provinces held by officials who retained their substantive appointments at the capital, to fully institutionalized posts at the top of the provincial hierarchy.

The final step toward institutionalization was taken by the Manchu rulers. In the early decades of the Ch'ing dynasty, a number of important changes in the tu-fu system occurred. Between 1661 and 1665 each province was provided with its own governor-general. Governors at subprovincial levels were eliminated, and by 1667 each province had its own governor as well. The regime continued to experiment with the regional jurisdictions of these offices well into the eighteenth century.

Strategic considerations frequently dictated changes in the jurisdictional boundaries of governor-generalships. For example, the area administered by the governor-general of Ch'uan-shen (Shensi, Szechwan, and Kansu provinces) was changed in 1748 to cover only Shensi and Kansu. A new post was created to handle Szechwan because the area was considered too large and strategically sensitive to be governed effectively from Sian in Shensi province.’ Other posts were split temporarily for similar reasons, as when the Min-che region (Fukien and Chekiang), was divided in 1648.10

On at least two oceasions the competence of individual appointees determined the redefinition of jurisdictional boundaries. In 1727, for instance, Kao Chii-chuo, governor-general of Min-che, was perceived by the emperor to be "short of ability to look after the affairs of two provinces." A new governor-general, Li Wei was appointed to administer Chekiang province, although the emperor made it clear that this was not to be considered a precedent.

Just as governor-general posts could be divided for less capable incumbents, new ones could be created for exceptionally competent officials or court favorites. During the Yung-cheng reign, T'ien Wen-ching, governor of Honan, was appointed governor-general of that province and later of Hotung (Honan and Shantung provinces), new jurisdictions created because no other governor-generalship was available. The emperor made it clear that "this appointment reflects special imperial grace and should not be taken as a precedent."!2 In other cases, changes were implemented in an attempt to effect greater coordination of areas sharing common characteristics or problems. In 1734, the province of Kwangsi was removed from the jurisdiction of Kwangtung and temporarily placed with Yunnan and Kweichow because "Kwangsi and Kweichow are areas inhabited by the Miao people. If they are not placed under the control of one governor-general, coordination will often present a serious problem."!3

The thirty-one different governor-generalships were, for the most part, amalgamated into eight jurisdictions by 1760. The sole exception was the governor-generalship of the three northeastern provinces, which was established in 1907.14 The provinees of Shantung, Shansi, and Honan were ruled directly by a governor who did not fall within the jurisdictional purview of a governor-general. On the other hand, two provinces, Chihli and Szechwan, were without governors and were ruled directly by governors-general who resided at Pao-ting-fu and Chengtu, respectively. Kiangsu, Anhwei, and Kiangsi, commonly referred to as Liangkiang, were grouped together and administered by a single governor-general with headquarters at Nanking. The other provinces were grouped in pairs: Shensi and Kansu (known as Shen-kan) with headquarters at Lanchow; Fukien and Chekiang (Min-che) at Foochow; Hupei and Hunan (Hukuang) at Wuch'ang; Kwangtung and Kwangsi (Liangkuang) at Canton; and Yunnan and Kweichow (Yun-kwei) at Yunnan-fu or Kunming.

In addition to the provincial posts, other governors-general were appointed to administer river conservancy and grain transport. The number of river conservancy posts varied from one to three, but in their most settled form they earried such titles as director-general of the conservation of the Yellow river and Grand Canal, southern section (chiang-nan ho-tao tsung-tu) and director-general of the conservation of the Yellow river and Grand Canal, eastern section (ho-tung ho-tao tsung-tu) The grain official was called the director-general of grain transport (ts'ao-yun tsung-tu) . All three officials held the same rank as the provincial governors-general (2A) and often served in provincial posts as well. In addition to these, the court from time to time appointed special governors-general.

Concurrent Titles

The early Manchu rulers appear to have experimented liberally with concurrent titles for governors and governors-general. Both offices began with concurrent titles on the Board of War, but, as we shall see later, the military authority and titles of governors changed in different locations over time. The policy on granting concurrent titles to governors-general in the early years is not entirely clear. It appears that prior to 1692 the Ming tradition of providing governors-general with appointments on the Board of War and the Censorate was generally followed. It is not certain that all governors-general in these early years held concurrent posts, but it is elear that those who did always held them with the Board of War and the Censorate.

In 1692 the practice of giving governors-general concurrent titles became more regularized. At that time governors-general promoted from governorships were appointed concurrently junior vice-president of the Board of War and the Censorate. An official promoted to governor-general from a board vicepresideney received a concurrent post as either senior or junior vice-president of the Board of War and the Censorate, depending upon his seniority on a board prior to his new appointments.!° In 1723 the regulations were changed to inelude the stipulation that those who had held a previous appointment as president of a board would henceforth be given the concurrent titles president of the Board of War and president of the Censorate. The only variation on this general rule was that the governors-general of Ch'uan-shen and Liangkiang, because of the great strategic importance of their regions, would be appointed concurrently president of the Board of War and junior president of the Censorate, regardless of their previous appointments. 16 The last major change appears to have taken place in 1749. Thereafter, whatever the region of their appointment, governors-general received the concurrent titles junior president of the Censorate and, at the discretion of the emperor, either president or vice-president of the Board of War. 17