Juba[142] again makes five Fortunate Isles: Ombrios, Nivaria, Capraria, Junonia, and Canaria, all fertile but uninhabited. Large dogs were found, however, in the last-named, and two of these had been brought to Juba himself, who called the island after them. Date-palms also abounded. Juba also, according to Pliny, discovered the Purple islands (Purpurariae) off the coast of Mauretania, which have been carelessly identified by some with the Madeira group, though wanting the two essential conditions of Juba's description: (1) producing Orchil; (2) lying very close to the shore of Mauretania. Lançarote and Fuerteventura agree with Juba's conditions on these points,[143] but then why are they made a separate group from Nivaria, etc., which are undoubtedly the main body of the Canaries? Juba's account is the most clear and valuable we have from ancient geography, dealing with the Canaries, and is far better than that[144] of the Alexandrian geographer. Ptolemy lays down the Fortunate Islands--assuming the Canaries to be meant--incorrectly both in latitude and longitude, in a position really corresponding better to that of the Cape Verdes. Hence it has been supposed that he confounded the two groups in one; whereas the Cape Verdes, lying out to sea 300 miles from the Continent, are not likely to have been known, even in his day. An error in position is so common with Ptolemy that it is quite unnecessary to be disturbed by it. But he clearly had some definite knowledge that islands existed in the ocean to the west of Africa, and in his map he probably reproduces the statements of others, without first-hand information of his own, assigning such a position as suited best with his theories. For he not merely brings the southernmost of the Fortunate Isles down to 11° N. lat., but scatters the group through 5° of latitude, placing the northernmost in latitude, 16° N. His names vary much from Juba's, for he gives us six: Canaria, the Isle of Juno, Pluïtala,[145] Aprositus (the Inaccessible), Caspiria, and Pinturia or Centuria; at the western extremity of these, after the example of Marinus, he drew the first meridian of longitude.[146] [Footnote 142: Copied by Solinus and many mediæval writers (see Pliny, _Hist. Nat._, vi, 31). Juba's work was dedicated to Caius Cæsar, B.C. 1, when just about to start on an expedition to the East. Ombrios, from its mountain lake, has been identified with Palma; Nivaria more easily with Teneriffe and Canaria with Grand Canary; Junonia is difficult to fix, as we have the statement that a second and smaller island of the same name is in its neighbourhood; Capraria is supposed to be Ferro. The remaining two of our modern archipelago, Lancarote and Fuerteventura, are supposed by some to be the "Purpurariae" of Juba.] [Footnote 143: And are therefore accepted as the Purpurariae by D'Anville Gossellin, Major, and, with some hesitation, by Bunbury.]