M. Barth says that this identification would agree well with the testimony of his inscription XVIII.  B., which comes from Angkor and for which Campa is a part of the Dakshinapatha, of the southern country.  But the capital of this rival State of Kamboja would thus be very near the Treang province where inscriptions have been found with the names of Bhavavarman and of Icanavarman.  It is true that in 627, the King of Kamboja, according to the Chinese Annals (Nouv.  Mel.  As. I. p. 84), had subjugated the kingdom of Fu-nan identified by Yule and Garnier with Campa.  Abel Remusat (Nouv.  Mel.  As. I. pp. 75 and 77) identifies it with Tong-king and Stan.  Julien (J.  As. 4 deg.  Ser.  X. p. 97) with Siam. (Inscrip.  Sanscrites du Cambodge, 1885, pp. 69-70, note.)

Sir Henry Yule writes (l.c. p. 657):  “We have said that the Arab Sanf, as well as the Greek Zabai, lay west of Cape Cambodia.  This is proved by the statement that the Arabs on their voyage to China made a ten days’ run from Sanf to Pulo Condor.”  But Abulfeda (transl. by Guyard, II. ii. p. 127) distinctly says that the Komar Peninsula (Khmer) is situated west of the Sanf Peninsula; between Sanf and Komar there is not a day’s journey by sea.

We have, however, another difficulty to overcome.

I agree with Sir Henry Yule and Marsden that in ch. vii. infra, p. 276, the text must be read, “When you leave Chamba,” instead of “When you leave Java.”  Coming from Zayton and sailing 1500 miles, Polo arrives at Chamba; from Chamba, sailing 700 miles he arrives at the islands of Sondur and Condur, identified by Yule with Sundar Fulat (Pulo Condore); from Sundar Fulat, after 500 miles more, he finds the country called Locac; then he goes to Pentam (Bintang, 500 miles), Malaiur, and Java the Less (Sumatra).  Ibn Khordadhbeh’s itinerary agrees pretty well with Marco Polo’s, as Professor De Goeje remarks to me:  “Starting from Mait (Bintang), and leaving on the left Tiyuma (Timoan), in five days’ journey, one goes to Kimer (Kmer, Cambodia), and after three days more, following the coast, arrives to Sanf; then to Lukyn, the first point of call in China, 100 parasangs by land or by sea; from Lukyn it takes four days by sea and twenty by land to go to Kanfu.” [Canton, see note, supra p. 199.] (See De Goeje’s Ibn Khordadhbeh, p. 48 et seq.) But we come now to the difficulty.  Professor De Goeje writes to me:  “It is strange that in the Relation des Voyages of Reinaud, p. 20 of the text, reproduced by Ibn al Fakih, p. 12 seq., Sundar Fulat (Pulo Condore) is placed between Sanf and the China Sea (Sandjy); it takes ten days to go from Sanf to Sundar Fulat, and then a month (seven days of which between mountains called the Gates of China.) In the Livre des Merveilles de l’Inde (pp. 85, 86) we read:  ’When arrived between Sanf and the China coast, in the neighbourhood of Sundar Fulat, an island situated at the entrance of the Sea of Sandjy, which is the Sea of China....’  It would appear from these two passages that Sanf is to be looked for in the Malay Peninsula.  This Sanf is different from the Sanf of Ibn Khordadhbeh and of Abulfeda.” (Guyard’s transl. II. ii. 127.)