INFLUEXCE OF THE ROMAN EMPIRE. l91 



tudc and latitude, and to the length of the day ; but, njtivith- 

 standmg the constant reference to the advantages of astro 

 noraical results over mere itinerary measuremMits by land and 

 sea, it is, unfortunately, impossible to' ascertain, among these 

 uncertain positions (upward of 2500 of which are given), the 

 nature of the data on which they are based, and the relative 

 probability which may be ascribed to them, from the itinera- 

 ries then in existence. 



The entire ignorance of the polarity of the magnetic needle, 

 and, consequently, of the use of the compass (which, twelve 

 centuries and a half before the time of Ptolemy, under the 

 Chinese Emperor Tsing-wang, had been used, together with a 

 luay measui'ci', in the construction of the magnetic cars), 

 caused the most perfect of the itineraries of the Greeks and 

 Romans to be extremely uncertain, owing to the deficiency 

 of means for learning with certainty the direction or the line 

 which formed the an^le with the meridian.* 



In proportion as a better knowledge has been acquired, in 

 modern times, of the Indian and ancient Persian (or Zend) 

 languages, we are more and more astonished to find that a 

 great portion of the geographical nomenclature of Ptolemy 

 may be regarded as an historical monument of the commercial 

 relations existing between the West and the remotest regions 

 of Southern and Central Asia.t We may reckon the knowl- 

 edge of the complete insulation of the Caspian Sea as one of 

 the most important results of these relations, but it was not 



* See a collection of the most striking instances of Greek and Roman 

 errors, regarding the directions of different mountain chains, in the in- 

 troduction to my Asie Centrale, t. i., p. xxxvii.-xl. Most satisfactory 

 investigations respecting the uncertainty of the numeiical bases of Ptol- 

 emy's positions are to be found in a treatise of Ukert, in the Rkeiniscke 

 Mustnm fur Philologie, Jahrg., vi., 1838, s. 314-324. 



t For examples of Zend axid Sanscrit words which have been pre 

 served to us in Ptolemy's Geography, see Lassen, Diss, de Taproban. 

 Insula, p. 6, 9, and 17 ; Bumouf 's Comment, sur le Yacna, I. i., p 

 xciii.-cxx. and clxxxi.-clxxxv. ; and my Examen Crit. de VHist. de la 

 Giogr., t. i., p. 45-49. In a few cases Ptolemy gives both the Sanscrit 

 names and their significations, as, for the island of Java, " barley island," 

 'ladadiov, 6 arj^awEL KpcSfjg v^aog, Ptol., vii., 2 (Wilhelm von Humboldt. 

 Ueber die Kawi-Sprache, bd. i., s. 60-63). According to Buschmana 

 the two-stalked barley, Hordevm distichon, is still termed in the princi- 

 pal Indian languages (as in Hindostanee, Bengalee, and Nepaulese, and 

 in the Mahratta, Guzerat, and Cingalese languages), as well as in Per 

 sian and Malay, yava, dschav, or dschau, and in the language of Orissa 

 yaa. (Compare the Indian translation of the Bible, in tbo passage Job. 

 vi., 9, and 13, and Ainslie, Materia Medica of Hindostan, Madras, 181? 

 p. 217.) 



