10 E. Loomis— Results from an examination of the 



shows the line of the least mean pressure, being about 29-6 

 inches on the meridian of Greenwich and increasing somewhat 

 as we proceed either east or west from that meridian. These 

 lines are drawn chieflv from data collected by Alexander Bu- 

 chan. (See Edinburgh Phil. Trans , vol. xxv.) 



We perceive then that at all places near the southern 

 margin of the chart the mean pressure of the atmosphere is 

 greater than it is further northward, and this is generally suf- 

 ficient to cause an average surface wind from south to north 

 although the wind advances from a warmer to a colder region. 

 On the other hand, at places near the northern margin of the 

 chart the mean pressure of the atmosphere is somewhat greater 

 than it is further south, and this force combined with a lower 

 mean temperature causes a surface wind from north to south. 

 Here then are permanent causes producing winds from opposite 

 directions near the upper and lower portions of the chart, and 

 these must be a permanent source of storms independent of 

 those inequalities of pressure which arise from causes of a 

 more local nature. 



The average path of storms in their progress from America 

 to Europe is apparently modified by the line of greatest mean 

 pressure. This line has a more northerly position in Europe 

 than it has in America, and this may be the reason why storm 

 tracks generally bend northward in advancing from America to 

 Europe. There are some minor particulars in which storm 

 paths are apparently modified by the line of greatest mean 

 pressure ; but instead of attaching importance to coincidences 

 which may prove to be accidental, it is more prudent to wait 

 and see if these peculiarities are confirmed by further obser- 



Oscillations of the barometer in different latitudes. 



For the purpose of determining in what region of the globe 

 the oscillations of the barometer are the greatest, I have pre- 

 pared a table showing the mean monthly oscillation of the 

 barometer at as many stations as possible in high northern lati- 

 tudes. A few of the numbers in the following table are de- 

 rived from Kaemtz Meteorology, edited by C. V. Walker, p. 297. 

 The other numbers have been collected by myself from various 

 sources which are indicated in the last column, and some of the 

 results have required a careful discussion of many years' obser- 

 vations. Column fourth shows the average monthly range of 

 the barometer for the three winter months, and column fifth 

 shows the same for the three summer months expressed in 

 English inches 



As some of these numbers depend upon observations of only 

 one 3^ear, and therefore do not represent mean values very ac- 



