CAUSES OF THE GLACIAL PERIOD 569 



southerly regions. It can be seen in Colorado, where a square crossed 

 by 22.8 storm tracks per year lies south of two which are crossed by only 

 17.8 and 18.8. Just west of this Colorado square, and in the western 

 part of the same State, a square crossed by 13.2 tracks lies south of one 

 crossed by 12.8 and of another crossed by 12. Still more to the north- 

 west we have 9.0 south of 6.9 and 7.8. In times of numerous sun-spots 

 there seems to be an unmistakable tendency for this incipient southern 

 storm belt to become more highly developed. A study of figure 7 and of 

 the others of the same kind proves that this must be the case. In times 

 of many spots storminess decreases in the central United States and in- 

 creases on either side. If this process goes far enough, we are bo-und to 

 have an area of actual deficiency of storms in the center and two storm 

 belts on either side. The case is like that of a heap of sand from which 

 sand is shoveled to either side. No matter how high the central heap 

 may be, its level will ultimately fall below that of the heaps on either 

 side, provided the shoveling be continued long enough. 



In order to make the matter as concrete as possible, I have prepared 

 the diagrams shown in figure 19. They represent the number of storm 

 tracks in the given longitudes under four different conditions. The 

 figures at the bottom represent latitudes; the height of each curve shows 

 the number of storms. Each curve begins at zero in a latitude sufficiently 

 far north so that no storm centers actually pass over it. Going south — 

 that is, toward the right in the diagrams — the number of storms rises to 

 a maximum and then decreases with more or less regularity. The first 

 set of curves shows the conditions in longitude 60° to 65° west, the next 

 in longitude 70° to 75° west, and so on to 120° to 125° west. In each 

 set there are four curves. The upper, marked "Min," shows the distri- 

 bution of storminess at times of minimum sun-spots according to figure 

 7. The next line, marked "Av," shows the average distribution for 30 

 years according to figure 3. The third line shows the conditions at times 

 of maximum spots according to figure 7. Finally, the lower line repre- 

 sents the conditions which would prevail if the conditions which now 

 exist during times of maximum sun-spots Avere magnified sixfold. This 

 does not mean that the actual number of storms has been multiplied by 

 six, but that the departures have been so multiplied. In other words, if 

 at one particular point the curve marked "Max" show^s that the number 

 of storms at times of maximum spots now averages two in excess of the 

 average for. all the years for which data are available, this excess has been 

 multiplied by six and becomes 12 in the "Max. X 6'^ curve. Thus the 

 average number of storms for all years may be 20. During the nine 



