288 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. JO 



In figure 107 we have plotted the twelve-month means of the 

 departures of the air-pressure at various stations. The unbroken 

 curves are drawn directly, while the broken curves are inverted. 

 The values are departures from normals ; those for the stations of 

 curves III-IX are computed for the thirty years from 1877 to 

 1906. 



The curves of figure 107 show especially two very distinct types : 

 the North-Atlantic type (curves I to III), and what we may call the 

 Indo-Maylayan South-American type (curves IV-VIII). 



The North Atlantic type of pressure curves is, in our figure, 

 represented by curves I-III, from the low-pressure region (curve 



II for Stykkisholm, Iceland) and the high-pressure region (curve 



III for Ponta Delgada, the Azores) of the North Atlantic. In curve 

 I is plotted the difference between the Azores pressure maximum 

 and the Icelandic pressure minimum (cf. fig. 94, I). The vertical 

 scale (in mm.) for curves I and II has been reduced to the half of 

 that of curves III-IX. To the Atlantic curves I-III ought to be 

 added the curve for the pressure difference between 30° N, 30° W 

 and Sao Thiago, in the region of the NE. trade winds (see fig. 91, 

 VIII). 



All these curves have a striking resemblahce to each other, the 

 curve of the pressure maximum (III) agreeing very closely with 

 the inverted curve of the pressure minimum (II). An increase 

 of pressure in the region of the Azores pressure maximum conse- 

 quently coincides as a rule with a decrease of pressure in the region 

 of the Icelandic pressure minimum, and vice versa, as was already 

 pointed out by Hildebrandsson [1897, etc.] and Hann [1904] . Hence 

 the pressure gradients are simultaneously increased or decreased 

 over the whole region of the North Atlantic (cf. fig. 91, VI, VIII; 

 fig. 109, II, III). 



Unfortunately we have had no opportunity of examining any 

 sufficiently long series of barometric observations from stations in- 

 side the tropical low-pressure region of the Atlantic. Thus we 

 do not know the nature of the barometric fluctuations in that region. 

 But if we may judge from the observations at Port au Prince and 

 Fort de France (in the West Indies, see fig. 71, VB and VIB) the 

 pressure in the tropical Atlantic regions fluctuates chiefly directly 

 as that of the Azores high-pressure region and inversely as that of 

 the Icelandic minimum. [It has to be considered that pressure 

 curves for stations lying outside the maximum or minimum regions 

 of the North Atlantic, may show certain resemblances to the one 



