THE SHIFTING OF CLIMATIC ZONES. 205 



"The effect of these complicated changes * * * is to move the isogonic lines which lie east of 

 the line of no change westward and those west of the no-annual-change-line eastward, or, in other 

 words, the lines are being crowded together from both sides of the line of no change. This effect 

 considered in company with the known changes in dip and intensity * * * implies that the 

 magnetic pole has moved during the past twenty years chiefly southward, the west component of 

 motion being greatly subordinate to the southerly one." 



The hypothesis Hes near that the storm track may center at the magnetic pole and may 

 move with the magnetic field. If such an hypothesis can be entertained, the agreement 

 here shown may be considered to be in harmony with it. 



In conclusion, I should like to point to the desirabilitj' of having similar maps made for 

 Japan and Europe, the only other areas of high storm frequency in the northern hemisphere. 

 Lack of opportunity and the inaccessibility of the material have made it impossible to 

 construct these maps, but they are necessary in order to give us a true insight into the 

 shift of the storm track. In constructing such maps or in making others of the United 

 States, it is particularly desirable that areas of less than 5° square should be employed, 

 in order to bring out greater detail. Moreover, a single year should be the unit, and the 

 maps for single years should be compared or combined by the use of overlapping means in 

 such a way as to determine whether the location of storms varies from times of maximum 

 to those of minimum sun-spots.* 



This brings us to the end of Professor Kullmer's contribution. Its main significance 

 hes in the fact that the average storm followed a course slightly farther south and west 

 during the period from 1899 to 1908 than during the period from 1878 to 1887. To 

 put the matter in another way, between the times of the two maps the zone of prevailing 

 westerly storms moved very shghtly equatorward from its earlier position. I would not 

 be understood as magnifying the shght difference between the two maps. Doubtless 

 conditions may soon once more become the same as they were during the period covered 

 by Dunwoody's map, and there is no reason to suppose that any great change has taken 

 place. The important point is that here we have direct evidence that the climatic zones of 

 the world are at the present day subject to minor shiftings back and forth, and that these 

 shiftings on a small scale produce the results which our assumed larger and more prolonged 

 shiftings appear to have produced upon a really important scale. Meteorologists almost 

 universally recognize the fact of such minor shiftings, but it is valuable to have the matter 

 definitely recorded in maps which can readily be compared. 



We have now seen that whether single years or longer periods such as decades be taken 

 as our unit of time, the location of the storm tracks shows a tendency to vary. Let us now 

 consider the same question, using a time unit of much greater length. 



In discussing the changes of climate which appear to have taken place in Asia I have 

 elsewhere shown that they appear to be of the same nature as those of the glacial period, 

 the difference being one of degree, not kind. In the western hemisphere the evidence points 

 to the same conclusion. Hence it will be of value to consider some of the latest results of 

 researches upon the conditions of glaciation in polar regions at the present time. These 

 results have been gathered into convenient form by Professor W. H. Hobbs, in his book 

 upon "Glaciers," a volume which contains a large number of valuable facts as well as some 

 suggestive theories. To meteorologists and glaciologists one of the surprising results of 

 recent explorations of Greenland and the Antarctic continent has been the discovery that 



* Since this chapter was in print, Professor Kullmer has constructed year maps of the storm frequency of the 

 United States from 1874 to 1912, and of Europe from 1876 to 1891, the only period for which data are available. 

 These maps are on the basis of areas 2.5° in hititude by 5° in longitude. Tliey confirm tlie conclusions here set 

 forth as to the shifting of the storm track, and also show that there is a distinct and unmistakable relation between 

 the location and number of storms on the one hand and the number of sun-spots on the other. (See page 2.53.) 



