THE WEATHER CYCLE 65 



must not expect periods, due to sun spots or 

 any form of solar change, to be ' always the 

 same in length. 



The statement that the sun spot period is 

 i i . 1 1 1 1 years in length, looks very precise ; but 

 that period is only the average of nearly two 

 centuries of records, and any one solar cycle 

 may differ in length from the mean by as much 

 as two years. * If, therefore, our weather 

 cycles are connected with the sun spot period, 

 they also will vary from 9 to 13 years. 



We cannot expect maximum to follow 

 maximum with the punctuality of clock work. 

 There is no sanctity to the number n, any 

 more than the number 13 is unhallowed. 

 The great variation in the length of the solar 

 period shatters for ever any hope of our getting 

 certain predictions as to the weather of future 

 years from simple averages. Averages are 

 useful, but they are easily misused. In trying 

 to determine weather cycles we must not put 

 our trust in the mere application of simple 

 addition and division to haphazard collections 

 of figures. " Professor Armstrong once had a 

 German chemical student, who, considering his 

 nationality, was abnormally lazy but normally 

 systematic, and who used to determine the 

 weight of his precipitates by asking all the 

 students in the laboratory to guess the amount, 

 and then taking the average. That student's 



* The precise length of the sun-spot cycle as given by Wolf, (1875) 

 is ii-iin etc. + 2-030 years. 



