280 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.55. 



The principal interest in the present paper must be the decisive 

 evidence which it furnishes of the very different climatic conditions 

 formerly existing in this desert region and the rather far-reaching 

 correlations which it is possible to make — results which in both cases 

 are not at all dependent on the correct botanical determination of 

 the species. Before discussing these results it would be profitable 

 to consider briefly the present climatic and physiographic features 

 of this section of Peru. Physiographically Peru may be divided into 

 three well-defined longitudinal regions: (1) The Coast region of 

 sandy arid desert or semidesert, crossed by rivers in narrow, more 

 or less fertile valleys according to the water supply; (2) the Sierra 

 or Andean region, which I need not stop to characterize; and (3) tlie 

 Montaiia, or region of tropical forests of the Amazon basin. 



In northern Peru the Piura desert extends along the coast from 

 the Gulf of Guayaquil southward to the Morrope Valley, a distance 

 of about 200 miles. It is traversed by three rivers — the Tumbez, 

 Chira, and Piura — the first two having their sources in the inner 

 Cordilleras and hence not dry part of the time. The climate is less 

 arid than farther south, fogs and garuas are more frequent, and at 

 intervals of several years there are occasional heavy showers. The 

 country is practically treeless. The valleys contain chaparral of 

 Prosopis, and along the river banks willows and groups of palms ; and 

 the irrigated cultivation of cotton, cane, corn, alfalfa, and vineyards 

 is possible where the streams are of sufficient size and not ijiter- 

 miltent. 



I might add that the entire sea border of Peru is a semidesert, 

 Arica dividing the true rainless desert country of Chile from the 

 sterile country extending northward from Arica (latitude 18° oO') 

 to north of Tumbez (latitude 3° 20'), which, though normally rain- 

 less, as just stated, receives the benefit of the drizzly garuas and sea 

 fogs which are reflected in the less pronounced desert character of 

 this northern country. 



Beginning with John Ball ^ the flora of the Coastal zone of west- 

 ern South America has been discussed by von Eggers^, Wolffs, and 

 Guppy*. The Strand is characterized as the Sesuvium zone, re- 

 placed some distance north of Tumbez by the Mangrove zone of the 

 Gulf of Guayaquil. 



The region immediately behind the beach from Tumbez southward 

 Ls a typical desert or semidesert, consisting of sand-covered barren 

 hill slopes with shifting dunes (mendanos) and varied only in the 

 transverse valleys by vegetation more fitting to the latitude. About 

 25 or 30 miles northeast of Tumbez there is a remarkably sudden 



1 Ball, J., Notes of a Naturalist in South America, London, 1887. 



2 Eggors, H. von, Das Ktistengebiet von Ecuador. Deutsche Geogr. Blatter, vol. 17, 

 Heft 4, 1894. 



» Wolff, T., Geografia y Geologia del Ecuador, 1892. 



* Guppy, H. B., A Naturalist in the Pacific, vol. 2, pp. 474-501, 1906. 



