GEOGRAPHY. 



37 



Here I should be inclined to place the New 

 Eosse^ of the Periplus, and to identify, as do 

 Vincent and Stuch, the Pyralaon Islands, with 

 Pemba, Zanzibar, and Mafiyah. M. Guillain, 

 by a careful calculation of distances, would 

 transfer the site further north to the natural 

 channel between Patta, Mandra, Lamu, and the 

 mainland. But whilst errors of numerals are 

 easily made, and readily copied in manuscripts ; 

 and whilst mistakes of distance can easily be 

 accoimted for, it is hard to believe that the 

 Phoenician, Egyptian, and Greek merchants 

 would have neglected the finest harbour and the 

 best site for trade upon the whole Azanian 

 shore. Moreover, there is nothing in the text 

 to prevent the Pyralaon Islands being those off 

 the Benadir, and the Posse being about the 



1 In the edition of Charles Muller (Pains, 1845) the word 

 Kaivrjc disappears, and the sentence becomes /cat ttjc; Xeyo^f j t/c 

 litjpvxoe, ' and the so-called fosse.' Certainly the term ^iw()v^ 

 would suit the Mombasah Canal better than the Channel of 

 Patta, and the former is the only ' digging ' where liuman 

 labour can possibly have been applied. Thus Pliny (v. 3) ex- 

 plains the name of the city Hippo Diarrhytus, ' from the chan- 

 nels made for irrigation.' That the scanty population of Arabs 

 at ancient Mombasah was incapable of excavating a canal 600 

 metres long is no proof that the work was not done. The 

 Sultan of Mombasah could bring into the field 5000 wild 

 archers, and, similarly, in the Brazil the most astonishing 

 works were effected by a handful of Portuguese, assisted by 

 hordes of Tupy savages. 



