features of the natives of this province; and it is highly probable that the Canton and Fo-kien people were originally the same race as the tribes which still remain unreclaimed on the east side of Formosa."[1] (Supply.  Vol. p. 260.) Indeed Martini tells us that even in the 17th century this very range of mountains, farther to the south, in the Ting-chau department of Fo-kien, contained a race of uncivilised people, who were enabled by the inaccessible character of the country to maintain their independence of the Chinese Government (p. 114; see also Semedo, p. 19).

["Colonel Yule’s ‘pariah caste’ of Shao-ling, who, he says, rebelled against either the Sung or the Yuean, are evidently the tomin of Ningpo and zikas of Wenchow.  Colonel Yule’s ’some aboriginal tribe between Fo-kien and Che-kiang’ are probably the zikas of Wenchow and the siapo of Fu-kien described by recent travellers.  The zikas are locally called dogs’ heads, which illustrates Colonel Yule’s allophylian theories.” (Parker, China Review, XIV. p. 359.) Cf. A Visit to the “Dog-Headed Barbarians” or Hill People, near Fu-chow, by Rev. F. Ohlinger, Chinese Recorder, July, 1886, pp. 265-268.—­H.C.]

NOTE 4.—­Padre Martini long ago pointed out that this Quelinfu is KIEN-NING FU, on the upper part of the Min River, an important city of Fo-kien.  In the Fo-kien dialect he notices that l is often substituted for n, a well-known instance of which is Liampoo, the name applied by F.M.  Pinto and the old Portuguese to Ningpo.

[Mr. Phillips writes (T.  Pao, I. p. 224):  “From Pucheng to Kien-Ning-Foo the distance is 290 li, all down stream.  I consider this to have been the route followed by Polo.  His calling Kien-Ning-Foo, Que-lin-fu, is quite correct, as far as the Ling is concerned, the people of the city and of the whole southern province pronounce Ning, Ling.  The Ramusian version gives very full particulars regarding the manufactures of Kien-Ning-Foo, which are not found in the other texts; for example, silk is said in this version to be woven into various stuffs, and further:  ’They also make much cotton cloth of dyed thread which is sent all over Manzi.’  All this is quite true.  Much silk was formerly and is still woven in Kien-Ning, and the manufacture of cotton cloth with dyed threads is very common.  Such stuff is called Hung Lu Kin ‘red and green cloth.’  Cotton cloth, made with dyed thread, is also very common in our day in many other cities in Fuh-Kien.”—­H.C.]

In Ramusio the bridges are only “each more than 100 paces long and 8 paces wide.”  In Pauthier’s text each is a mile long, and 20 feet wide.  I translate from the G.T.

Martini describes one beautiful bridge at Kien-ning fu:  the piers of cut stone, the superstructure of timber, roofed in and lined with houses on each side (pp. 112-113).  If this was over the Min it would seem not to survive.  A recent journal says:  “The river is crossed by a bridge of boats, the remains of a stone bridge being visible just above water.” (Chinese Recorder (Foochow), August, 1870, p. 65.)