749 



the tempestuous winds from the south *. The 

 meteorologic journal of M. Rob redo proves 

 that these differences are alike remarked far 

 from the coast, in the interior of the island. It 

 is not the mean of the months that differs more 

 near the tropic of cancer than near the tropic 

 of Capricorn, it is rather the extreme heights 

 owing to accidental causes. At the boundary 

 of the southern torrid zone, the extreme oscil- 

 lations -f of the barometer attain only 21 mil- 

 limeters (9.3 U ) ; at the extremity of the northern 

 torrid zone, they are often 25 millimeters, 

 sometimes 30.5 mm (13.3 li ). The southern hemi- 

 sphere, south of the parallel of 23°, contains a 

 very small portion of land ; and the atmosphere 



* The hurricanes are not in general accompanied by such 

 an extraordinary lowering of the barometer as is imagined 

 in Europe. I possess 56 barometric observations made by 

 the captain of a ship, Don Tomas de Ugarte, nearly from 

 hour to hour, at the Havannah, during the terrible hurricane 

 of the 27th and 28th of August, 1794. When the tempest 

 was most violent, the column of mercury sunk only 5 lines 

 {11.3mm), Kirwan asserts however, that at the island of 

 Saint Bartholomew, the barometer has been seen to lower in a 

 hurricane (1792), 42 millimeters. Irish Trans., vol. viii, 

 p. 387. Is this fact as well certified as a lowering of 25 

 millimeters at the Isle of France ? (Moreau de Jonnes, Hist, 

 phys. des Ant. } Tom. i. p. 420). See on the barometric 

 heights observed on the coast of Chili, Espinosa, Memorias 

 de los Naveg. Esp., Tom i, p. 129, 134, 179. 



t In December and March. See Mem. de Lisboa, Tom. ii, 

 p, 397. 



