66 STEPPES AND DESERTS. 



of the Haradsh, on the margin of the African Desert, reminds the 

 geologist of the augitie vesicular amygdaloid, phonolite, and green- 

 stone porphyry, which are only found at the northern and western 

 boundaries of the Steppes of Venezuela and of the plains of the Ar- 

 kansas, as it were on the hills of the ancient coast line. (Humboldt, 

 Relation Historique, torn. ii. p. 142; Long's Expedition to the 

 Rocky Mountains, vol. ii. pp. 91 and 405.) 



( 7 ) p. 26. " When suddenly deserted by tJie east wind of the tropics 

 in a sea covered with weed." 



It is a remarkable phenomenon, well known among sailors, that, 

 in the vicinity of the African coast (between the Canaries and the 

 Cape de Verde Islands, and particularly between Cape Bojador and 

 the mouth of the Senegal), a west wind often takes the place of the 

 general east or trade-wind of the tropics. It is the wide expanse of 

 the Desert of Sahara which causes this westerly wind. The air over 

 the heated sandy plain becomes rarefied, and ascends, the air from 

 the sea rushes in to supply the void so formed, and thus there some- 

 times arises a west wind, adverse to ships bound to the American 

 coast, which are made in this manner to feel the vicinity of the heatr 

 radiating desert without even seeing the continent to which it belongs. 

 The changes of land and sea breezes, which blow alternately at cer- 

 tain hours of the day or night on all coasts, are due to the same 

 causes. 



The accumulation of sea-weed in the neighbourhood of the Afri- 

 can coast has been often spoken of by ancient writers. The locality 

 of this accumulation is a problem which is intimately connected with 

 our conjectures respecting the extent of Phoanician navigation. 

 The Periplus, which has been ascribed to Scylax of Caryanda, and 

 which, according to the researches of Niebuhr and Letronne, was 

 very probably compiled in the time of Philip of Macedon, describes 

 beyond Cerne a quantity of fucus forming a weed-covered sea-r-a 

 kind of " Mar de Sargasso;" but the locality indicated appears to 

 me to differ very much from that assigned in the work entitled "De 

 Mirabilibus Auscultationibus," which long bore> unduly, the great 

 name of Aristotle. (Compare Scyl. Caryand. Peripl. in Hudson, 

 vol. ii. p. 53, with Aristot. de Mirab. Auscult. in opp. omnia ex. 



