¹ Saint-Denys, Hervey de, Ethnographie des peuples étrangers à la Chine, Genève, 1883, II, 461Google Scholar; Tuan-lin, Ma, Wên-hsien t'ung-k'ao (Wên-k'u, Wan-yu edition), 2602Google Scholar. Saint-Denys translated ‘after the yung-hui period’ as ‘vers l'époque des années yong-h-oei’. In this essay the expressions ‘Funan’ and ‘Chên-la’ which are Chinese ones, are used only when quoting from Chinese sources or from studies by scholars who have accepted the expressions as representing political realities in Cambodia. Reasons will be given later in the essay why the author believes that Chinese nomenclature confuses rather than clarifies pre-Angkorian history.

² Aymonier, E., Le Cambodge, III, Paris, 1904, 429–30, 439Google Scholar; ‘Le Founan’, JA, x^e Ser., i, janvier-février 1903, 132–3, 149–50; ‘Le Siam ancien’, ibid., mars-avril 1903, 190.

³ Pelliot, P., ‘Deux itinéraires’, BEFEO, IV, 1–2, 1904, p. 290, n. 4Google Scholar; 403–4. Pelliot pointed out that the conquests, according to the text, took place after the 650–6 reign period. Ch'ih-t'u's position on the east coast of the Malay Peninsula is definitively established by ProfessorWheatley, Paul, The Golden Khersonese, Kuala Lumpur, 1961, ch. iiiGoogle Scholar.

⁴ Pelliot, P., ‘Le Fou-nan’, BEFEO, III, 2, 1903, 248–303CrossRefGoogle Scholar; ‘Deux itineraires’, 131–413.

⁵ Gerini, G. E., Researches on Ptolemy's geography of eastern Asia, London, 1909, 202–6Google Scholar.

⁶ Briggs, L. P., The ancient Khmer empire, Philadelphia, 1951, 50, 54Google Scholar. Briggs assigned the conquests to the 650–6 period.

⁷ Wheatley, , op. cit., 46, 278Google Scholar; Wolters, O. W., Early Indonesian commerce, Ithaca, 1967, 163Google Scholar. Ferrand also identified Chia-cha with Kedah; Ferrand, G., ‘Le K'ouen-louen’, JA, xi^e Ser., xiii, mars-avril 1919, p. 249, n. 1Google Scholar. Dupont, who was particularly interested in pre-Angkorian Cambodia, ignored these toponyms.

⁸ Hsin T'ang-shu (Po-na edition), 222C, 2a.

⁹ For the suggested reconstruction of ‘Śrī Kumāra’ and ‘Śrī Devavarman’, see Pelliot, , ‘Deux itinéraires’, 404Google Scholar. A Ch'ih-t'u Brahman was called Chiu-mo-lo = Kumāra; Sui-shu (Po-na edition), 82, 4b.

¹⁰ i.e. T'ang hui-yao, T'ung-tien, T'ai-p'ing yü-lan, T'ai-p'ing huan-yii chi, and T'ai-p'ing huang-chi. Ssŭ-ma Kuang ignores the missions of 638. The Wên-hsien t'ung-k'ao and the Yü-hai (facsimile of an edition of 1337 in the National Central Library, Taipeh, 1964, 2898)Google Scholar quote the Hsin T'ang-shu.

¹¹ See Rotours, R. des, Le traité des examens, Paris, 1932, 56–64Google Scholar, for the intentions and performance of the compilers.

¹² On the T F Y K see Rotours, des, op. cit., 91–2Google Scholar, and Teng, Ssū-yu and Biggerstaff, Knight, An annotated bibliography of selected Chinese reference works, third ed., Cambridge, Mass., 1971, 89Google Scholar. Professor Hiraoka has noted that the 1642 edition of the T F Y K is frequently undependable; Hiraoka, T. and others, ‘Tō-dai shiryō kō’, Kyōto Daigaku Jinbun Kagaku Kenkyūjo sōritsu nijūgo shūnen kinen ronbunshū, Kyoto, 1954, 676ffGoogle Scholar. The present writer is grateful to Professor C. A. Peterson for calling his attention to the textual status of the 1642 edition of the T F Y K. This edition (Peking facsimile reproduction of 1960) is quoted below, but passages, for which parallel passages do not appear in texts other than the T F Y K, have been checked against the Southern Sung wood-block print at the Seikado Bunko, Tokyo, by means of xerox copies kindly supplied by Professor Akira Nagazumi. The wood-block print is the oldest extant version of the T F Y K and the earliest we can get to the original contents of the text. The passages in the 1642 edition which have been checked concern: the Funanese attack on Champa reported in 643 (999, 11721b), the missions of Chiu-mi and Fu-na in 671 (970, 11402b), Jayavarman I's mission of 682 (970, 11403a), the Wên-tan mission of 717 (974, 11445a), the Chên-la mission of 750 (971, 11413b), the Wên-tan mission of 753 (971, 11414a), the Wên-tan mission of 771–2 (976, 11461b), the Wên-tan mission of 798 (976, 11462b), and the Chên-la mission of 814 (972, 11417b). There are no significant differences between the Sung and 1642 texts.

¹³ T F Y K, 970, 11398b.

¹⁴ T F Y K, 970, 11402b.

¹⁵ T F Y K, 970, 11403a; 970, 11403b.

¹⁶ Coedès, G., The indianised states of Southeast Asia, Honolulu, 1968, 65–70Google Scholar; 72–3.

¹⁷ Coedès, , op. cit., 85Google Scholar.

¹⁸ T F Y K, 995, 11688a.

¹⁹ B. Karlgren, Grammata Serica recensa (reprinted from BMFEA, 29, 1957, and henceforth cited as GSR), 104a, 823a.

²⁰ The equivalence of -liäng and -lyāng, the stressed syllable, is not in doubt. In connexion with the first double vowel, Mrs. Judith Jacob informs the author that ‘a variety of short vowels, a, i, and u, and sometimes both i and u, occur in both Old Mon and Old Khmer in first short syllables. The writers were trying to indicate a neutral vowel for which they had no symbol’. Professor E. G. Pulleyblank informs the author that m u is in the rising tone, which was used at this period to represent short vowels in foreign words, and miu would have been suitable to represent a neutral shwa, necessarily rounded after a labial in Chinese. The in the transcription is not to be regarded as a distinct and separate ‘vowel’.

²¹ Cœdès, G., Inscriptions du Cambodge (henceforth cited as 82IC), 1, 30Google Scholar.

²² I follow the most recent definition of the terms pramān and viṣaya, provided by ProfessorJacques, Claude in ‘Études d'épigraphie cambodgienne: VII’, BEFEO, LIX, 1972, 198–9Google Scholar. Pramān is a geographical term, referring to definable territory, while viṣaya is an administrative term, for which ‘province’ may be cautiously used. Viṣaya is an expression which begins to be used in the epigraphy of the second half of the tenth century. In the twelfth century Malyāng was known as a visaya; IC, VI, 314Google Scholar.

²³ Ta-kuan, Chou, Chén-la fêng-tu chi (Li-tai Hsiao-shih series, pén 31, no. 103, Shanghai, 1940), 15aGoogle Scholar; Pelliot, P., ‘Mémoires sur les coutumes du Cambodge’, BEFEO, II, 2, 1902, p. 173, n. 1Google Scholar. Chou Ta-kuan's transcription is Muo/Mu-liang ; GSR, 802a, 735a. Pelliot, identifying Malyāng, was guided by Aymonier, and the identification has been upheld by Cœdés; Cœdés, G., The indianised states, p. 355, n. 147Google Scholar. Coedes remarks that, phonetically, Malyang would have been pronounced as ‘Molieng’ or ‘Moling’; Cœdés, G., BEFEO, XXXII, 1, 1932, p. 80, n. 1Google Scholar. Professor John McCoy has informed the author that the Hva-i i-yu, a Sino-Mongolian glossary of 1389, supports the value mo for the character at that time. The author is grateful to Mrs. Jacob and Professors Pulleyblank and McCoy for their advice on the sounds represented by the Chinese transcriptions of place-names discussed in this study. Chou Ta-kuan describes Malyāng as a chün , for which ‘prefecture’ is a more appropriate term than ‘province’, which Pelliot used. Chün and chou had been used interchangeably for ‘prefecture’ since at least Sui times; Wén-hsien t'ung-k'ao, 315, 2470.

²⁴ IC, V, 209Google Scholar. For a discussion of the geographical significance of this reference to Malyāng, see Cœdés, , BEFEO, XXXII, 1, 1932, p. 80, n. 1Google Scholar.

²⁵ IC, IV, 155Google Scholar. This inscription does not mention the status of Bhimapura, but IC, III, 11Google Scholar, also of the eleventh century, describes Bhīmapura as a province (viṣsaya). An inscription of 1145 mentions the ‘district/village’ (sruk) of Pūrvāśrama in the province (viṣsaya) of Malyang. The inscription is from near the present town of Battambang; IC, VI, 314Google Scholar. Pūrvāsŕama is also mentioned in Vat Baset inscriptions of 1036 and 1042. Vat Baset is near Battambang; IC, III, 9, 23Google Scholar.

²⁶ Cœdés, , The indianised states, 170Google Scholar. The suppression of the revolt was sufficiently important to be commemorated on the bas-relief of the Bayon.

²⁷ GSS, 933r, 350a.

²⁸ Soothill, W. E. and Hodous, L., A dictionary of Chinese Buddhist terms, London, 1937, 370bGoogle Scholar.

²⁹ G. Cœdés, ‘Index des noms propres de l'épigraphie du Cambodge’, 1C, VIII.

³⁰ Hsüan-tsang transcribes Vrji as P uet-l et-zi; GSR, 500a, 403a, 961y. He also transcribes Vārāṇasī as Pvâ-lâ-(?)ni-siḙ; GSR, 2511, 6a, 563a, 869a. Pelliot suggests that the seventh-century transcription represents ‘a la rigueur’ Na-va-na-(gara), which Cœdés subsequently identified from an inscription of 664 as ‘Naravaranagara’; Pelliot, , BEFEO, III, 2, 1903, 295Google Scholar; GSR, 350a, 500a, 350a; Cœdés, G., BEFEO, XLIII, 1943–6, 4Google Scholar.

³¹ (P u- …) and (^B'uet …); Soothill, , op. cit., 370a, 228aGoogle Scholar.

³² A frequent example is (b'uát-mua = varman); GSR, 276b; Karlgren, B., Analytic dictionary of Chinese and Sino-Japanese, Paris, 1923 (cited as AD), 593Google Scholar.

³³ Soothill, , op. cit., 347aGoogle Scholar; GSR, 25q, for = b'uâ. Also see Soothill, , op. cit., 449bGoogle Scholar, for = vana = b' wak… = GSR, 771m.

³⁴ Soothill, , op. cit., 248bGoogle Scholar; GSR, 740z, in respect of . Mrs. Jacob has remarked that, since pura does not appear in the transcription, vana-, pronounced alone, may well have undergone a change of form from to. In this case the Khmers may have pronounced vana in such a way that the Chinese heard a p. The Chinese transcribed Vrji as B'uât-dź' a, where both syllables begin with a voiced consonant, but they also transcribed this word as P uet-l et-źi. In the latter form, the unvoiced p precedes two words beginning with voiced consonants, presumably to give greater effect to the stress in -rji.

³⁵ P. Dupont suggests that sruk was the equivalent of -pura; BEFEO, XLIII, 1943–1946, p. 116, n. 5Google Scholar.

³⁶ Harsavarman I's year of accession is according to Jacques, C., BEFEO, LVIII, 1971, 175Google Scholar.

³⁷ IC, VI, 266Google Scholar; IC, I, 196Google Scholar (of 1041). The origin of the connexion between this family and Vanapura is unknown. Śrī Vīralakṣmī is said to be of the Vraḥ Sruk; IC, VI, 266Google Scholar. Vat Ek, near Battambang, was in the Vrah Sruk; IC, IV, 153, 155Google Scholar.

³⁸ IC, VII, 139Google Scholar.

³⁹ IC, VII, 49Google Scholar; BEFEO, XLIII, 1943–1946, p. 121, n. 7Google Scholar.

⁴⁰ IC, V, 89–90Google Scholar, describing Jeng Vnaṃ as a pramān. The Sdok Kak Thom inscription of the eleventh century refers to the viṣsaya of Adripāda in the context of Jayavarman II's reign and to the viṣsaya of Jeng Vnaṃ in Jayavarman III's reign; BEFEO, XLIII, 1943–1946, 96, 111Google Scholar.

⁴¹ The Bo Ika inscription of 868, from Korat province, is in a region described as ‘outside Kambudeśa’; IC, VI, 85Google Scholar.

⁴² IC, I, 53Google Scholar; IC, VII, 154Google Scholar. The former inscription, of 928, mentions a pramān and also a sruk of Vīrendra. IC, V, 154Google Scholar, provides the site of the city of Vīrendra; it is north of the Dangreks and was the provincial capital. The inscription of 928 states that the province of Jeng Vnam contained a sruk of Vīrendrapattana; IC, I, 30 and 54Google Scholar. It would have been a different place from the Virendra north of the Dangreks.

⁴³ IC, VII, 49Google Scholar.

⁴⁴ IC, V, 90Google Scholar, referring in 896 to the pramān of Jeng Vnaṃ.

⁴⁵ BEFEO, XLIII, 1943–1946, 114Google Scholar. Aymonier and Groslier suggest that Bhadragiri should be identified with the Tangko peak in the north-western part of the Dangreks; Aymonier, , Le Cambodge, II, Paris, 1901, 265Google Scholar; Groslier, G., BEFEO, XXIV, 3–4, 1924, 366Google Scholar. The peak is north to north-west of the Sdok Kak Thom site.

⁴⁶ BEFEO, XLIII, 1943–1946, p. 129 and n. 6Google Scholar.

⁴⁷ ibid., 74.

⁴⁸ de Lajonquière, E. Lunet, Inventaire descriptif des monuments du Cambodge. Cartes, Paris, 1911Google Scholar.

⁴⁹ IC, V, 23–4Google Scholar . This inscription, of Khan Noi, dated 637, is in Khmer and mentions a Mratañ Khloñ Jyeṣṭhapura.

⁵⁰ In Aymonier's day the Chup Smach was the main route for travellers from Laos to the Tonlé Sap and Bangkok; Aymonier, E., ‘Notes sur Ie Laos’, Cochinchine Française. Excursions et Reconnaissances, IX, 21, 1885, 11Google Scholar.

⁵¹ ibid., 10–11: ‘For an entire day, the traveller will walk in a dismal shadow which oppresses him like a nightmare, seeing nothing but enormous tree trunks.… When he leaves this gloom, he will greet the rays of the sun with joy, no matter how warm it is’.

⁵² On this forest see Delvert, J., Le paysan cambodgien, Paris, 1961, map 5Google Scholar. Lunet de Lajonquiére's map shows a number of small temples and reservoirs in the fertile depression at the foot of the range, through which streams flow into the Tuk Chun river; Lajonquiére, Lunet de, Inventaire descriptif des monuments du Cambodge, in, 405–13Google Scholar, and ibid., Cartes.

⁵³ GSR, 992n, 405p.

⁵⁴ See Pulleyblank, E. G., ‘The Chinese name of the Turks“, JAOS, LXXXV, 2, 1965, 121–5Google Scholar, on the use of entering tone words to represent a single consonant in a non-final position. Professor Pulleyblank transcribes as m t; ibid., 123. Could K eu-m ět be an attempt to represent Kam, noted by Cœdés as a probable abbreviation of a title of someone connected with Maleng = Malyāng in 987? See IC, VI, 186, note (1)Google Scholar.

⁵⁵ For the temples in th e Svay Chek area see de Lajonquiére, Lunet, op. cit., 371–91Google Scholar. The area was sufficiently important in Yaśovarman I's reign at the end of the ninth century to justify the establishment of a royal monastery; IC, III, 65Google Scholar. Jayavarman V, in the second half of the tenth century, sought to recruit scholars from this area; IC, II, p. 62, n. 5Google Scholar.

⁵⁶ AD, 342; GSR, 806a. For Professor Pulleyblank, pointing out that the character was reserved for foreign words, writes k â; Asia Major, NS, XI, 2, 1965, 202Google Scholar.

⁵⁷ AD, 1047; GSR, 1129a. For Professor Pulleyblank writes kau; Asia Major, NS, X, 2, 1963, 220Google Scholar.

⁵⁸ Wolters, , Early Indonesian commerce, 163Google Scholar.

⁵⁹ As in gagana and garuda, both of which begin with ; Soothill, , op. cit., 317a and 315bGoogle Scholar.

⁶⁰ Sung-shih (Po-na edition), 489, 21b.

⁶¹ Sastri, K. A. Nilakanta, The Cōḷas, second ed., Madras, 1955, 219Google Scholar. According to the Sung-shih, the voyage took 1,150 days to reach Canton; Sung-shih, 489, 23a.

⁶² See Soothill, , op. cit., 224bGoogle Scholar. The Chinese transcriptions are: Ka- a (), AD, 342, GSR, 47b; Ka- a (), AD, 342, GSR, 47a; Ka-dź' a (), AD, 342, GSR, 45h.

⁶³ T F Y K, 970, 11398b.

⁶⁴ IC, III, 36Google Scholar.

⁶⁵ An inscription from this site shows that th e site was called ‘Giripura’ in the eleventh century; IC, III, 39. Giripura was the name of the āśrama and not of the areaGoogle Scholar.

⁶⁶ IC, III, p. 37, n. 1Google Scholar, and IC, IIIḷ, 45Google Scholar.

⁶⁷ IC, III, 61Google Scholar. Also see IC, III, 63Google Scholar, for a similar reference to Gajapura.

⁶⁸ IC, VII, 14Google Scholar. The Khmer rendering of the name is ‘Gahjaḥpure’.

⁶⁹ IC, III, 37, dated 949Google Scholar.

⁷⁰ IC, V, 207Google Scholar, dated 1003. The administrative status of Jaroy Sangke is not supplied. A Koh Ker inscription of 921 gives a list of sniks, including… y Sangke, and the indistinct part of this name may have been Jaroy; IC, I, 48Google Scholar. Sruk Sangke may also be mentioned in the Kralanh Thorn inscription; IC, VII, p. 25, n. 3Google Scholar. Another inscription from near Koh Ker, of the tenth century, mentions a sruk Sangke; IC, I, 186Google Scholar.

⁷¹ IC, III, 5 and 9Google Scholar.

⁷² Soothill, op. cit., 420aGoogle Scholar.

⁷³ Professor Pulleybiank informs the author that -áu- was sometimes used for a foreign -o-. Tau , for example, is used as manyōgana for the Japanese to.

⁷⁴ IC, IV, 314Google Scholar.

⁷⁵ Lewitz, S., ‘La toponymie khmére’, BEFEO, LIII, 2, 1967, 431Google Scholar. A representative of the sruk of Manggalapura is mentioned in another Vat Baset inscription, together with a representative of Purvasrama; IC, III, 16Google Scholar.

⁷⁶ The representatives of Sangko and Manggalapura were present at the time of the purchase of the land. The Purvasrama representatives were among those who planted the boundaries of the land. The land was at Vak Tongting (ibid., 8), but the location of this place is unknown.

⁷⁷ An excellent description of the geography of north-western Cambodia is given by Lunet de Lajonquiére, Inventaire descriptif, III, 1–7Google Scholar. On p. 3 he stresses the dividing effect of this river system.

⁷⁸ Part of Sangkhah province's location is indicated in IC, III, 37Google Scholar, which connects the viṣsaya of Sangkhah with the āśrama of Phnom Prah Net Prah. IC, III, 45Google Scholar, comes from the site of Prasat Sangkhah, not far from the previous inscription. Cœdés thinks that Sangkhah may be the ancient name of the temple.

⁷⁹ The Sdok Kak Thom inscription refers to Stuk Bansi in the province (viṣ) of Amoghapura; BEFEO, XLIII, 1943–1946, 117Google Scholar. The rice-fields of Ganesvara, also in Amoghapura (ibid., 115), provided support for foundations in Bhadrapaṭṭana in the Dangreks province; ibid., 114–15. The same rice-fields also helped to support foundations in Stuk Bansi in Amoghapura (ibid., 115), and Dupont believed that Bhadrapaṭṭana was th e eastern neighbour of Stuk Ransi (ibid., p. 121, n. 2) and that some of the Stuk Ransi temples were west of the Makkak river; ibid., 74. The context of the Sdok Kak Thom inscription, from which this information comes, is the progressive development of lands in the extreme north-western part of Cambodia by Śivakaivalya's descendants, and Dupont was justified in assuming tha t the Amoghapura lands in question were in this part of Cambodia. Amoghapura's connexion with the lands to the west of the Makkak river may also be reflected in IC, III. 55, an eleventh-century inscription which states that an Amoghapura provincial official witnessed the demarcation of th e boundaries of land given to the Bantay Prav temple near Svay ChekGoogle Scholar.

Groslier believed that the Amoghapura province extended to th e east a t least as far as the Prasat Sangkhah region. He noted that revenue from Amoghapura was given to this temple; BEFEO, XXIV, 3–4, 1924, 361Google Scholar;IC, III, 53Google Scholar. The gift is not, however, evidence of the location of Amoghapura. Groslier also noted that the rice-field of Stuk Veng, which he believed to be in Amoghapura, was given to a temple at Cung Vis near Prasat Sangkhah; art. cit., 364–5. For Cung Vis’ position, See IC, III, p. 83, n.7Google Scholar. Yet the Koh Ker inscription, which Groslier quoted to connect Stuk Veng with Amoghapura, does not contain this information; IC, I, 47–71Google Scholar.

⁸⁰ Aymonier, , Le Cambodge, II, maps facing p. 304Google Scholar. Also see map 3 in J. Delvert, Le paysan cambodgien. In the seventh century the Tonle Sap was larger than it is today. Professor Delvert believes tha t it once reached to Mongkolborei and Sisophon; Delvert, J., op. cit., 57Google Scholar. A vivid description of the situation in the summer months is contained in Brien, , ‘Aperçu sur la province de Battambang’, Cochinchine Française. Excursions et Reconnaissances, X, 24, 1885, 343–1Google Scholar, in respect of the Battambang area. The writer refers to “…un spectacle des plus singuliers et des plus étonnants: la navigation á toute vapeur en pleine forêt!’.

⁸¹ The somewhat rare temple sites south of the river system extend from Mong towards Battambang and thence towards Mongkolborei; Lunet de Lajonquière, Inventaire descriptif des monuments du Cambodge. Cartes.

⁸² The only seventh-century inscriptions so far discovered from the north are three from the Aran region, two of which are dated 637 and 639. The names of kings are not mentioned; IC, V, 23–4Google Scholar.

⁸³ The author is grateful to Professor R. Billard, who, in a letter dated 23 June 1969, informs him that the inscription's horoscope shows that its date corresponds to Wednesday, 14 June 657, and about 8.30 a.m. local time. For this inscription see IC, II, 193–5Google Scholar. Its exact provenance is unknown. Professor Boisselier would not be surprised if a connexion exists between it and a lintel from the Baset hill, about 10 miles north-east of Battambang; Boisselier, J., ‘Arts du Champa’, Artibus Asian, XIX, 1, 1956, 204Google Scholar.

⁸⁴ Barth, A. and Bergaigne, A., Inscriptions sanscrites du Cambodge et Champa (henceforth cited as ISCC), Paris, 1885, 28Google Scholar.

⁸⁵ See Cœdés, G., The indianised states, 68–9Google Scholar, for the chronology of this period.

⁸⁶ The first was reached by sea from Champa. The second has been identified as Kompong Chnang near the southern entrance to the Tonle Sap; Cœdés, G., The indianised stales, p. 355, n. 147Google Scholar. The third was on the coast. For these identifications, see Pelliot, , BEFEO, II, 2, 1902, p. 138, n. 4, p. 138, n. 9, p. 170, n. 4Google Scholar; Pelliot, , Memoires sur les coutumes du Cambodge de Tcheou Ta-kouan, Paris, 1951, 70, 95–6Google Scholar.

⁸⁷ Chên-la-fêng-tu, 15a; Pelliot, , BEFEO, II, 2, 1902, 173Google Scholar; GSR, 281a, 289d.

⁸⁸ Today the Angkorian p has become b; IC, II, 3Google Scholar. An example of the equivalence of the two consonants is provided by Pa- , the first syllable in the Chinese transcription of ‘Basan’, the Khmer capital in 1371; Ming-shih (Po-na edition), 324, 12b. The following explanation is proposed for the consonants at the end of the Chinese words. Chou Ta-kuan came from Chekiang province, and most of his informants in Cambodia were probably southern Chinese. In the transcription of foreign words in modern Cantonese and lower Yangts e dialects the final consonant in the first syllable is lost or assimilated, but the final consonant in the second syllable is retained. These patterns are likely to be as old as Chou Ta-kuan's period. Thus = pwát-si-wi, mentioned by Chou Ta-kuan, has been restored by Cœdès as(ta) pasvi; Pelliot, , Mimoires, p. 65, n. 1Google Scholar. The consonant in the first syllable is also muted in numbers 24, 25, and 36 in Pelliot's list of Chinese transcriptions; Mémoires, 62–70. The retention of the final consonant in words of one or two syllables is illustrated by numbers 17, 30, and 34 in Pelliot's list. The author is grateful to Professor McCoy for advice on this subject.

⁸⁹ Pavie, A., Cochinchine Française. Excursions et Reconnaissances, IV, 12, 1882, 526–9Google Scholar.

⁹⁰ Mouhot, H., Voyage dans les royaumes de Siam…, Paris, 1868, 183Google Scholar.

⁹¹ Delaporte, L., Voyage au Cambodge, Paris, 1880, p. 132, n. 2Google Scholar.

⁹² ibid., 137. ‘Baset’ is not mentioned in the Portuguese and Spanish texts quoted in Groslier, B.-P., Angkor et le Cambodge au xvi^e siècle, Paris, 1958Google Scholar. The letters of the French missionary, Langenois, who was in this area from 1790 to 1795, may contain information about Baset. On Langenois, see Groslier, , op. cit., 134Google Scholar.

⁹³ Pavie, , art. cit., 526–9Google Scholar. Aymonier suggests that masonry from the temple site was used to build the new town of Battambang; LeCambodge, II, 292–3Google Scholar. The author is grateful to Mrs. C. A. Trocki, who informs him that, according to the Battambang Chronicle, in 1838 the Thai general Bodin was ordered to establish the foundations and wall of the city of Battambang. This account is somewhat different from Pavie's.

⁹⁴ Professor Delvert suggests that the Dambang river had been the more important arm of the Sangke river; Delvert, J., op. cit., 58Google Scholar. Also see Delaporte, L., op. cit., p. 132, n. 2Google Scholar. Pavie quotes a legend which connects the name of the town with that of the river; Pavie, , art. cit., 528Google Scholar. On the derivation of ‘Battambang’ from dámbay, a ‘club’, see Lewitz, , art. cit., 394Google Scholar.

⁹⁵ Because it lists 10 provinces, including ‘Battambang’ and not ‘Baset’, the ‘Histoire d'un centenaire roi du Cambodge au XVII^e siècle’ most clearly reflects the prevalence of the usage of Battambang; de Villemereuil, A. B., Explorations et missions de Doudart de Lagrie, Paris, 1883, 325Google Scholar.

⁹⁶ Sūryavarman I established a linga here some time before 1018; IC, VI, 269Google Scholar. Jayavarman VII erected a hospital here; IC, III, 25Google Scholar. Jayavarman VII's wife established a Mahesvara image here; IC, II, 179Google Scholar.

⁹⁷ Boisselier, J., art. cit., 204Google Scholar.

⁹⁸ IC, III, 3–33Google Scholar. A lintel in the Vat Baset complex is thought to be as old as the end of the tenth century, indicating early buildings on this site; Dagens, B., ‘Étude iconographique de quelques fondations de l'époque de Sūryavarman I ^er’, Arts Asiatiques, XVII, 1968, 188Google Scholar.

⁹⁹ Martini, F., ‘De la signification de “ba” et “me”’, BEFEO, XLIV, 1, 1947–50, (pub.) 1951, 202–4, 207Google Scholar.

¹⁰⁰ IC, IV, 155Google Scholar, of the first half of the eleventh century. Narendragrama is the name by which Vat Ek was known.

¹⁰¹ The foundation of Vat Ek was probably not long after 1018; Dagens, B., art. cit., 176Google Scholar.

¹⁰² In China, too, chün is used ambiguously to refer to the area under the jurisdiction of the prefecture or to the town which served as the seat of the prefecture. Thus, according to Chou Ta-kuan, Malyāng would have meant a town and its immediate environs, though Malyāng was also the name of an Angkorian province. At the end of the twelfth century Jayavarman VII distributed 23 Jayabuddhamahanatha images throughout the empire. This figure rather than Chou Ta-kuan's ‘more than 90 prefectures’ reflects the number of substantial territorial units in the kingdom at that time, which included some in the Menam basin. Chou Ta-kuan's figure corresponds more closely to the 50 and more towers on the Bayon of Jayavarman VII, which Mus suggested represented ‘at least religious or administrative centres’ of a province; Cœdés, G., Angkor, Hong Kong, etc., 1963, 65Google Scholar.

¹⁰³ IC, V, 209Google Scholar. The Palhal inscription states that Jayavarman III, in the second half of the ninth century, chased elephants from Pursat towards Malyāng; Cœdés, G., BEFEO, XIII, 6, 1913, 27Google Scholar. Pursat is to the west of the southern part of the Tonle Sap and some distance south of Tano.

¹⁰⁴ IC, IV, 155Google Scholar, reveals that Vat Ek was in the ‘holy sruk’ and also in Bhimapura. IC, III, 11Google Scholar, describes Bhīmapura as a viṣsaya in the eleventh century. IC, IV, 71–3Google Scholar, shows that Bhimapura was regarded as a ‘territory’ (praman) in the tenth century.

¹⁰⁵ GSR, 102n, 1240c; 560e, 417a. B'uo-mai may have been intended to represent Bhima-pura, but it is more likely to stand for Bhūmi. Bhūmi, which earlier meant a large extent of land, became equivalent with a sruk; Lewitz, , art. cit., 407–8Google Scholar. In this case, part of the name is missing. The sruk of Bhumyakara is mentioned in an unpublished tenth-century inscription from the Battambang region: Jacques, C., ‘Supplement a l'index des noms propres’, BEFEO, LVIII, 1971, 185Google Scholar.

¹⁰⁶ The toponym ‘Sangke’, the name of a woody plant associated with lac production as well as of the river flowing through modern Battambang and, according to Pavie, of the ancient village on the site of modern Battambang, is also very old, appearing at least as early as the eleventh century; IC, V, 207Google Scholar. The reference is to ‘Jaro y Sangke’ and is on the Tano inscription. Mme. Lewitz defines jaroy as ‘avancée de terre dans l'eau’; Lewitz, , art. cit., 420Google Scholar.

¹⁰⁷ The significance of the Bhimapura toponym has not been determined. Bhimapura and Amoghapura are seventh-century as well as Angkorian toponyms; ISCC, 42. In Angkorian times both had associations with north-western Cambodia. The Chinese evidence of the seventh century mentions neither of them. They could have been alternative names for two of the principalities mentioned in th e Chinese records, imposed on conquered lands by southern over-lords. In the seventh century they were under the control of the lord of Tamrapura, another unknown toponym.

¹⁰⁸ Aymonier, E., Geographie du Cambodge, Paris, 1876, 54Google Scholar.

¹⁰⁹ Porée-Maspero, E., Étude sur les rites agraires des Cambodgiens, I, Paris, 1962, 100Google Scholar.

¹¹⁰ IC, V, 206Google Scholar.

¹¹¹ Barus, Dvāravatī, and Malayu are familiar examples of ancient toponyms which lingered on for centuries. The antiquity of the ‘Lavapura’ name has come to light; Boeles, J. J., ‘A note on the ancient city called Lavapura’, JSS, LV, 1, 1967, 113–14Google Scholar.

¹¹² The inscriptions mention Tamrapura, Ugrapura, Pasengga, Indrapura, Vyādhapnra, Dhruvapura, Bhavapura, Amoghapura, Bhimapura, Cakrangkapura, and Jyeṣṭhapura.

¹¹³ Sui-shu (Po-na edition), 82, 5b.

¹¹⁴ Chiu T'ang-shu (Po-na edition), 197, 2b.

¹¹⁵ The T F Y K, in the passage quoted on p. 357 above, states that the ‘dress and speech (of the four tributary countries of 638) are similar to that of Lin-yi (Champa)’. This description of people in north-western Cambodia must be ignored. The Chinese court officials would have been perplexed by the identity of envoys from a hitherto unknown part of South East Asia, especially when their envoys insisted that their countries were not subordinate to any Khmer kingdom known to the Chinese. A cultural comparison with Lin-yi, the best known country in this part of mainland South East Asia, may have been regarded as an acceptable description.

¹¹⁶ Personal communication by Mr. Dam Chhoeurn, who discovered the inscription. The inscription, which mentions no king's name, describes gifts to the god Śri Prabhāsasomeśvara.

¹¹⁷ IC, v, 23–4Google Scholar.

¹¹⁸ Pavie, A., Cochinchine Française. Excursions el Reconnaissances, V, 14, 1882, 295Google Scholar.

¹¹⁹ E. Aymonier, ibid., IX, 21, 1885, 13.

¹²⁰ Aymonier, , Géographic du Cambodge, 54Google Scholar.

¹²¹ Aymonier describes how, when one descends th e Dangreks, one sees the scattered peaks an d hills of Sisophon, Battambang, Phno m Srok, Chongkal, and Siem Reap; Cochinchine Française. Excursions et Reconnaissances, IX, 21, 1885, 12Google Scholar.

¹²² On the subject of the veal see Delvert, , op. cit., 127–30Google Scholar.

¹²³ See Lunet de Lajonquiere, op. cit., passim. The inscriptions of the region often contain toponyms which incorporate what Lewitz calls ‘elements hydrographiques’; Lewitz, , art. cit., 418–24Google Scholar. A reservoir was built in the Gajapura area in the tenth century; IC, VII, 14Google Scholar. Sūryavarman I built a dyke an d reservoirs in th e same area; Jacques, C., ‘La stèle du Phnom Sres’, BEFEO, LIV, 1968, 616–17Google Scholar.

¹²⁴ The lands nort h of the river system are today regarded as being among the poorest in Cambodia; Delvert, , op. cit., 632–4Google Scholar.

¹²⁵ On the canals of southern Cambodia, see Paris, P., ‘Anciens canaux khmers’, BEFEO, XLI, 2, 1941, map facing p. 372Google Scholar.

¹²⁶ Delvert, , op. cit., 88, 97–9, 634–40Google Scholar.

¹²⁷ Lunet de Lajonquière, op. cit., Cartes.

¹²⁸ , C. and Mourer, R., ‘The prehistoric industry of Leang Spean, province of Battambang, Cambodia’, Archaeology and Physical Anthropology in Oceania, V, 2, 1970, 127–46Google Scholar.

¹²⁹ Boisselier, J., ‘Arts du Champa’, Artibus Asiae, XIX, 1, 1956, 204Google Scholar.

¹³⁰ Boisselier, J., ‘Une statue feminine inédite du style de Sambor’, Arts Asiatiques, II, 1, 1955, 25–31Google Scholar.

¹³¹ ibid., 25.

¹³² Bhattacharya, K., Les religions brahmaniques, Paris, 1961, 98Google Scholar.

¹³³ Boisselier, J., ‘Travaux de la mission archéologique francaise en Thailande’, Arts Asia-tiques, XXV, 1972, 53Google Scholar.

¹³⁴ ISCC, 28. Barth refers to it as being ‘short and proud, worthy of a conqueror’; ibid., 27.

¹³⁵ IC, VII, 152Google Scholar.

¹³⁶ Dupont remarks that the liṅga foundations, attested by these inscriptions, commemorate the taking possession of the soil; Dupont, P., La statuaire prêangkorienne, Ascona, 1955, 76Google Scholar.

¹³⁷ Cœdés, , The indianised states, 66–9Google Scholar.

¹³⁸ Jacques, , BEFEO, LIX, 1972, 217Google Scholar.

¹³⁹ ISCC, 38–44.

¹⁴⁰ Pelliot noted a statement in the T'ang hui-yao that Chên-la (the northern base of these conquerors) began to bring Funan (the southern overlordship) into submission in the 535–45 period; ‘Deux itinéraires’, 368. He also noted Funanese missions in 559, 572, and 588; ibid., 389.

¹⁴¹ Cœdés, , The indianised states, 69Google Scholar.

¹⁴² IC, II, 193–5Google Scholar.

¹⁴³ IC, II, 149–52Google Scholar, found not far south of Ba Phnom. The author is grateful to Professor Roger Billard, who, in his letter of 23 June 1969, supplied this date.

¹⁴⁴ Cœdés, G., BEFEO, IV, 4, 1904, 691–7CrossRefGoogle Scholar; IC, II, 70Google Scholar. The exact provenance of the inscription is unknown.

¹⁴⁵ For his death after 690, see Cœdés, , The indianised states, p. 291, n. 70Google Scholar, which corrects the date in IC, II, 40Google Scholar.

¹⁴⁶ 657 (IC, II, 149–52Google Scholar); 664 (IC, VI, 8Google Scholar); 667 (ISCC, 70); 673 (IC, I, 14–15Google Scholar); 674 (IC, II, 12Google Scholar). Two undated inscriptions refer to the king in a similar manner; IC, I, 10Google Scholar; IC, III, 163Google Scholar.

¹⁴⁷ IC, VII, 56: ‘His power was well known’Google Scholar.

¹⁴⁸ IC, IV, 60Google Scholar. Jayadevi's relationship with Jayavarman I is established in IC, VII, 51Google Scholar.

¹⁴⁹ Barth, A., BEFEO, II, 3, 1902, 239Google Scholar. A cintārdtna is attributed to him. The inscription extols his military skills.

¹⁵⁰ IC, I, 11Google Scholar, verse xii.

¹⁵¹ IC, II, 195Google Scholar, with th e date supplied by Professor Billard. It is remarkable that two inscriptions, some distance apart, have identical dates. The reason may be that the astrologer of this already powerful king had prepared a calendar of auspicious days for that year.

¹⁵² The Han Chey inscription refers to a king's departure to war in autumn. His enemies' moats were dry; ISCC, 17. Fighting in the flooded zone of north-western Cambodia would have been impossible during the wet season in the middle of the year.

¹⁵³ See Dupont, P., ‘La dislocation du Tchen-la’, BEFEO, XLIII, 1943–1946, 17–55CrossRefGoogle Scholar, for a study of the eighth-century evidence.

¹⁵⁴ T F Y K, 970, 11403a and 11403b.

¹⁵⁵ See p. 356, above, for the intentions of the Hsin T'ang-shv's compilers.

¹⁵⁶ IC, V, 49–52Google Scholar.

¹⁵⁷ IC, II, 12Google Scholar. For the meaning of ājñā see de Mestier du Bourg, Hubert, ‘Ājñā praçasta, çāsana’, JA, CCLV, 3–4, 1967, 375–82Google Scholar. The name ‘Purandarapura’ is at least as old as 14 June 657, when it is mentioned in an inscription which also mentions one of Jayavarman I's counsellors; IC, II, 151Google Scholar.

¹⁵⁸ Jacques, C., ‘Deux inscriptions du Phnom Bakheṅ’, BEFEO, LVII, 1970, 62Google Scholar.

¹⁵⁹ IC, VII, 14Google Scholar, verse xxxii.

¹⁶⁰ IC, VII, 56Google Scholar.

¹⁶¹ Cœdés, , The indianised states, 36Google Scholar, discusses the transcription of ‘Funan’. Professor Jacques, however, reminds us that the equivalence of the local term bnaṃ ‘mountain’ and fu-nan is only a hypothesis. He notes that no inscription mentions a ‘king of the mountain’, the ruler of an important kingdom; Jacques, C., Annuaire, 1971–1972 (École Pratique des Hautes Études, IV^e section), 610Google Scholar.

¹⁶² Chin-shu (Po-na edition), 97, 10b.

¹⁶³ Llang-shu (Po-na edition), 8b, 10a.

¹⁶⁴ Professor Jacques shares the present writer's doubts concerning the suitability of the terms ‘Funan’ and ‘Chên-la’ as categories for Khmer history; Jacques, , Annuaire, 1971–1972, 609–10Google Scholar.

¹⁶⁵ CTS, 197, 2b; HTS, 222C, 3a. The division is said to have occurred after the shên-lung reign-period, which ended on 31 August 707. A mission had arrived from ‘Chên-la’ between 5 June and 3 July 707; T F Y K, 970, 11404a.

¹⁶⁶ T F Y K, 970, 11404a.

¹⁶⁷ T F Y K, 970, 11405b.

¹⁶⁸ T F Y K, 971, 11414a; 975, 11458b (753); 976, 11461b; 999, 11719a-b (771–2); 976, 11462b (798). The last mission came to the capital in the first month of 798 (22 January-20 February), and Pelliot's statement that 799 was the year of the mission should be amended; Pelliot, , ‘Deux itineraires’, 212Google Scholar.

¹⁶⁹ T F Y K, 972, 11417b.

¹⁷⁰ T F Y K, 972, 11417b (814). Only once does the T F Y K refer to ‘Water Chên-la’, which is in a chapter on ‘raids’ and not on ‘missions’. ‘Water Chên-la’ attacked Champa in 838; T F Y K, 995, 11688a.

¹⁷¹ T F Y K, 957, 11259a, is identical with this passage. Wên-tan can be restored as M uen-tân, and the name may be derived from ‘Mun’, the river which drains the Korat Plateau and enters the Mekong north of the Dangreks in the neighbourhood of Vat Phu; Cœdés, G., BEFEO, XXXVI, 1, 1936, 2Google Scholar.

¹⁷² is better rendered as ‘dry land’, contrasting it with the ‘wet land’ in the south.

¹⁷³ Sui-shu (Po-na edition), 82, 7a.

¹⁷⁴ A tenth-century inscription may reflect the same perception of mountains and waters; IC, IV, 96Google Scholar, and notes (5) and (6). This inscription includes references to an ancestor of the Vat Phu principality and also to Rudravarman, who ruled in southern Cambodia in the first half of the sixth century.

¹⁷⁵ Cœdés, G., BEFEO, IV, 4, 1904, 691–7CrossRefGoogle Scholar.

¹⁷⁶ IC, I, 3–5Google Scholar; ISCC, 16–21, and Cœdés, , The indianised states, 72Google Scholar.

¹⁷⁷ Jacques, , Annuaire, 1971–1972, 608Google Scholar.

¹⁷⁸ T F Y K, 999, 11721b.

¹⁷⁹ Peliiot, , BEFEO, IV, 1–2, 1904, 390–1Google Scholar

¹⁸⁰ T F Y K, 970, 11399a. The next mission came in 653; ibid., 11401b.

¹⁸¹ T F Y K, 970, 11399a.

¹⁸² T F Y K, 999, 11719b.

¹⁸³ T F Y K, 999, 11721b. Vassals sent ‘request’ missions in 320 (11720a); 450 (11720a); 488 (11720a); 450 (11721a); 472 (11721a); 493 (11721a); 508 (11721a). The T'ang hui-yao also contains an example of a request mission which was not regarded as a tributary one; THY, 36, 667 (Hsin-lo in 686), which is not recorded in the T H Y's section on this country or in the T F Y K. In 663 the Khmer king successfully requested that a Buddhist teacher should be allowed to return from China to Cambodia; Pachow, W., JGIS, XVII, 1–2, 1958, 16Google Scholar. No Khmer tributary mission is registered under this date. The Khmer king was Jayavarman I, who protected some Buddhist foundations in 664; IC, IV, 6–9Google Scholar.

¹⁸⁴ Pelliot, , BEFEO, IV, 1–2, 1904, 195, quoting the two T'ang historiesGoogle Scholar.

¹⁸⁵ In ‘Jayavarman II's military power: the territorial foundation of the Angkor empire’, JBAS, 1973, 1, 21–30, the author invokes the earlier history of north-western Cambodia as a basis for interpreting Jayavarman II's military achievements. ProfessorJacques's, ‘La carriere de Jayavarman II’, BEFEO, LIX, 1972, 205–20Google Scholar, with its revised chronology for this reign, was not then available to the author. Both Professor Jacques and he agree, however, that Dupont had greatly underestimated the scale of Jayavarman IPs territorial power.