The Dog-head feature is at least as old as Ctesias.  The story originated, I imagine, in the disgust with which “allophylian” types of countenance are regarded, kindred to the feeling which makes the Hindus and other eastern nations represent the aborigines whom they superseded as demons.  The Cubans described the Caribs to Columbus as man-eaters with dogs’ muzzles; and the old Danes had tales of Cynocephali in Finland.  A curious passage from the Arab geographer Ibn Said pays an ambiguous compliment to the forefathers of Moltke and Von Roon:  “The Borus (Prussians) are a miserable people, and still more savage than the Russians.....  One reads in some books that the Borus have dogs’ faces; it is a way of saying that they are very brave" Ibn Batuta describes an Indo-Chinese tribe on the coast of Arakan or Pegu as having dogs’ mouths, but says the women were beautiful.  Friar Jordanus had heard the same of the dog-headed islanders.  And one odd form of the story, found, strange to say, both in China and diffused over Ethiopia, represents the males as actual dogs whilst the females are women.  Oddly, too, Pere Barbe tells us that a tradition of the Nicobar people themselves represent them as of canine descent, but on the female side!  The like tale in early Portuguese days was told of the Peguans, viz. that they sprang from a dog and a Chinese woman.  It is mentioned by Camoens (X. 122).  Note, however, that in Colonel Man’s notice of the wilder part of the Nicobar people the projecting canine teeth are spoken of.

Abraham Roger tells us that the Coromandel Brahmans used to say that the Rakshasas or Demons had their abode “on the Island of Andaman lying on the route from Pulicat to Pegu,” and also that they were man-eaters.  This would be very curious if it were a genuine old Brahmanical Saga; but I fear it may have been gathered from the Arab seamen.  Still it is remarkable that a strange weird-looking island, a steep and regular volcanic cone, which rises covered with forest to a height of 2150 feet, straight out of the deep sea to the eastward of the Andaman group, bears the name Narkandam, in which one cannot but recognise [Script], Narak, “Hell”; perhaps Naraka-kundam, “a pit of hell.”  Can it be that in old times, but still contemporary with Hindu navigation, this volcano was active, and that some Brahman St. Brandon recognised in it the mouth of Hell, congenial to the Rakshasas of the adjacent group?

  “Si est de saint Brandon le matere furnie;
  Qui fu si pres d’enfer, a nef et a galie,
  Que deable d’enfer issirent, par maistrie,
  Getans brandons de feu, pour lui faire hasquie.”
      —­Bauduin de Seboure, I. 123.

(Ramusio, III. 391; Ham. II. 65; Navarrete (Fr. Ed.), II. 101; Cathay, 467; Bullet. de la Soc. de Geog. ser.  IV. tom iii. 36-37; J.A.S.B. u.s.; Reinaud’s Abulfeda, I. 315; J.  Ind.  Arch., N.S., III.  I. 105; La Porte Ouverte, p. 188.) [I shall refer to my edition of Odoric, 206-217, for a long notice on dog-headed barbarians; I reproduce here two of the cuts.—­H.C.]