WiosTROPP — Brasil and the Legendary Islands of the N. Atlantic. 241 



charts, founded on experience and, as their name implied, showing the route 

 from port to port, avoided the conventional, and strove to draw information 

 from every source. Nothing can mark the line between the scholar's map of 

 the world and the merchant's chart better than the contrast between the great 

 map of Eichard of Haldingham, about 1280, and Angelino Dulcert's work. 

 The latter gives in the 1325 map " Insula de montoniis sive de brazill," and in 

 the Catalan portolano, 1339 (Periplus viii),' "Insvda de Berzil," to the west of 

 the Shannon's mouth, and the " Insil diculy,"^ to the north-west of Ireland 

 (Per. viii). The " Charta navigationis " called " Portolano Laurenziano 

 Gaddiano sen Atlante Medicaeo " follows the last in giving " Insul de Berzi " 

 and "Insil dach" in similar positions (Per. x). The island corresponding to 

 Brazil is also found in the Venetian map of 1367 and the far superior Catalan 

 map' of 1373 (Per. xi), which marks a large circular island "Insula Berzil," in 

 the same place, west from the Shannon,* as also St. Brendan's Isle and Mam 

 (the later Maida or Asmanda) to the south of Brazil. It will be remembered 

 that the compass was apparently known in the thirteenth century to 

 Eoger Bacon, and later to Dante's tutor, Brunette Latini; but the latter 

 says that sailors would not have it on board, regarding it as " an infernal 

 spirit."* It only became popular about 1360, so that its use, the making of 

 the early maps, and the " appearance " of Brasil, all seem to be contempo- 

 raneous. The Solerio map of 1385 is even more interesting, for it marks two 

 Brasils, the " Insula de berzil " to the west of Ireland and " Brazir " in a 

 group of islands off the coast of Spain, the latter being evidently the mountain 

 of Brazil in Terceiro, though the Azores are quite misplaced here 

 (Periplus xvii). 



Fifteenth Centuky. — The I. de Brazi is shown among the Azores, and a 

 nameless isle, the mythic " Brazil," to the west of Ireland in the Venetian 

 map of Andrea Bianco in 1436 (Per. xx)." Another chart as early, or earlier, 



' For simplicity of reference, I only refer to Nordenskiold's Facsimile Atlas (Fac.) and Periplus 

 (Per.), even when the maps occurring in them have heen reproduced in other accessible works. 



-An island, "Artetum," occupies its place on Haldingham's map, 1280. Captain Thomas 

 (Proe. Soo. Antiqq., Scotland, 1875, vol. x, p. 7013) regards Daciili as Sanchule or St. Kilda. I 

 venture to suggest (from it being always shown so close to Ireland, and from St. Kilda being out of 

 the old trading tracks) that it is " Trahuli," or Inistrahull, off Donegal, if not Rockall (Eokol or 

 Rockell in maps from 1697 to 1715). 



^ Not the earlier one given on Plate X5, which also shows tlie Island. 



* Brasil is usually shown to the west of North Kerry (from Tralee to the Shannon) ; much has 

 been made of its variable position ; but this mobility it shares with real islands, like Japan, Ceylon, 

 and Madagascar, on the maps of the same period : indeed, Ireland itself is often misplaced. 



= Much is collected by Nordenskiold ; Periplus, p. 47 ; Weise, " Discovery of America," pp. 61, 62; 

 and others, on this important question. 



^ He also gives an unknown island, " Scorafixa " (perhaps Stockfixa or Stockfish), supposed by 

 some to he Newfoundland. The word " Baccalaos," used in later days, is said to hvae the same 

 meaning. 



[33*] 



