1851.] 



on the Vishnu Purana, 



3 



In Figure 2 T is the true 

 place of the sun when it ap- 

 pears to rise at V, T V being 

 33' according to received opi- 

 nion. Then T P S = dh, the 

 error in hour- angle due to 

 T V. Hence from the sphe- 

 rical triangles S P T, S P N, 

 and the plane triangle S V T S 

 we obtain 



T 



dh = T V cosec P S. cosec S P N sec P N. 

 = 33'. cosec 66§.° cosec 72.° sec 35° 24'. 

 Therefore dh = 46-4' 



m. sec. 



= 3 " 5*6, or nearly one-sixteenth of a muhurtta. 



Having the preceding approximate effect of refraction, we re- 

 calculate the latitude for a new hour-angle of 107° 14' for sunrise ; 

 and, on so doing, the corrected latitude comes out 34° 16'. 



The round numbers given by the writer imply that his estima- 

 tions were rough ; and this is corroborated by his making the length 

 of the shortest night equal to that of the shortest day, when the 

 two must differ by the sum of the effects due to four refractions. 

 There seems however no reason to doubt that the author of the 

 passage was an inhabitant of some place lying between 34° and 

 36° north latitude ; indeed the very confusion apparent in his astro- 

 nomical views renders his testimony more trustworthy, by proving 

 that his remarks were drawn from observation and not from theory. 



When we turn to the map, we find Cabul, Cashmeer and a few 

 other towns within the limits above-mentioned ; but Hindustan, 

 according to its modern boundaries, is seen to lie generally without 

 and to the south. This circumstance, agreeing as it does with 

 Brahminical traditions of a migration from the north-west, appears 

 to merit further investigation. 



E. B. P. 



