ANTIPODAL SOUTHERN CONTINENT. 443 



extending round the globe through the poles. The idea of 

 Crates thus approaches the reality, Avith the exception of the 

 equatorial ocean. It was perhaps based on the same sup- 

 posed necessity of an equal distrilnition of land and water in 

 order to maintain the equilil)rium of the globe which 

 influenced Mercator to postulate the existence of an antipodal 

 continent. The device of Crates survives in the orb which 

 forms one of the symbols of royalty at the coronation of our 

 sovereigns. 



Another element in the conception of a Terra Anstralis 

 was added by Marinus of Tyre and Ptolemy of Alexandria. 

 Eratosthenes (226 B.C.) had made the eastern coast of the 

 African continent terminate in about 12° N. at the Land of 

 Cinnamon, whence it was supposed oriental spices were 

 brought. Hipparchus (160 B.C.) prolonged the coast in- 

 definitely towards the south. But Marinus (about 1 00 A.D.), 

 perhaps basing his }"epresentation on misunderstood reports 

 of Greek travellers, diverted the African coast towards the 

 east at about 15° 30' S., and produced it to about the 

 longitude of the Golden Chersonese, at which point he made 

 it trend northwards to meet an extension of Asia, trending 

 southwards immediately beyond the Magnus Sinus, or Gulf 

 of Siam. The Indian Ocean thus became an inland sea. 

 This scheme was not upset until the true form of Africa had 

 been discovered, and even after that time it survived in 

 many sixteenth-century maps in the eastern and isolated 

 position which they gave to Zanzibar. 



The theory of Marinus was adopted by Ptolemy (about 

 151 A.D.), whose great weight as an astronomer gave 

 authority to all he promulgated. His astronomy, ri avvra^ig 

 fxtyiartj, was translated into Arabic in the cali])hate of 

 Mamun, 813-833, and was familiar to the Arabs under the 

 name Almagest (al fxtjiarr]). But their geogra])hical science 

 owes more to an anonymous work supposed to tiate from the 

 eighth century and to be of Greek origin, supplemented jjy 

 information supplied by the Arabs themselves. This is the 

 liasm al arsi, or description of the earth. We have not suffi- 

 cient data to enable us to judge whetiier the author of this 

 work accepted the Greek or the l^tolema^an scheme of the 

 southern hemisphere. Later Aralj geographers combined 

 the two. The Tabula Rotunda in the Oxford MS. of Edrisi 

 (1154) shows a round earth encircled by the Homeric Ocean 

 and with the Ptoleniffian extension of Africa ; but the Indian 

 Ocean is not surrounded by land, being connected with the 



