440 PHYSICAL GEOGRAPHY OF THE SEA, AND ITS METEOROLOGY. 



857. The instruments used for these observations were 

 Discussion of observa- niost part the old-fasliionecl 

 ^'^^^^- marine barometer, to which 

 no corrections have been apphed. The dis- 

 crepancies of this table evidently arise from 

 the lack of number sufficient to mask these 

 sources of error, or from the influence of the 

 land, and not from any difference as to the 

 mean height of the barometer along the 

 same parallels at sea in any one of the three 

 divisions. In this discussion, the observa- 

 tions of each group and every band were 

 arranged according to the month. These 

 monthly tables are not repeated here, but 

 they do not indicate any decided change 

 in the barometric pressure in high southern 

 latitudes according to the season. The 

 barometer there stands low the year round. 



858. Eesorting to the graphic method. 

 Barometric curve at and usmg the table (p. 439) for 

 ^^^- the pm^pose, the barometric 

 curve of the diagram (Plate XYI.) has been 

 projected fi^om pole to pole. 



859. Professor Schouw has given us the 

 Ditto over the land, mcau height of the barometer 



for 32 places on the land between the paral- 

 lels of 33'^ S. and 75° 30' N. They afford 

 materials for the annexed diagram, and show 

 the exceptional character of the meteorolo- 

 gical influences which rule on shore v/hen 

 compared with those which rule at sea. 

 There is barely a resemblance between this 

 profile of the atmosphere over the land and 

 the profile of it (Plate XYI.) over the sea, so 

 different are these influences. The UTegu- 

 larities over the land are chiefly owing to 

 the difference in the amount of precipita- 

 tion at one station as compared with the 

 amount at another. Those islands, as the 

 Sand^\dch and Society, which are so situated 

 as to biing down a heavy precipitation. 



for the 



