CYCLE PROBLEM AND LONG-RANGE FORECASTING 135 



of less importance than, for example, the two-year cycle mentioned in the 

 next paragraph. 



Second: There is a short variation of large amplitude with a length of 

 two or three years. This is well known as the most frequent cycle length 

 found in purely accidental variations and its harnessing into some system 

 must be done with caution. But its consistent dominance in certain form 

 over wide areas gives the impression that it is not accident in the common 

 meaning of that term and that it can be solved. Pending its solution the 

 form for prediction to take, if it is made, will have to be a forecast for the 

 mean of each two successive years instead of for individual years. This 

 amounts to prediction of a smoothed curve. 



Third: In this Arizona region the winter rains have been found to give 

 good correlations with tree growth; what about the summer rains? Summer 

 rains have been found to correlate to a degree with tree-ring growth in 

 Florida (Lodewick, 1930). A study of summer rains can be made also in 

 the Arizona trees, for in the case of "double" rings a definite part of the ring 

 is obviously due to summer rains. There is evidence of the dependence of 

 summer rains in part on the precipitation of the preceding winter but, in a 

 more general way, it is probably related to torrid zone conditions. 



MISSISSIPPI VALLEY 



With regard to great areas like the Mississippi Valley, the problem, it 

 seems to the writer, divides itself into two parts, centering respectively about 

 winter and summer conditions. The activity of winter storms is probably 

 related to the Arizona winter rainfall situation, as both are in the great 

 westerly winds, but the Central Valley has its local records modified by latitude 

 of the mean storm tracks of each season. In this way great complexity is 

 introduced. Any studies that take up, month by month and year by year, 

 the relation between the Mississippi Valley with its complex climatic changes 

 and Arizona with its long and relatively simple records in trees have now an 

 increased importance. Such, for example, are the departure studies now 

 carried on by Dr. F. E. Clements. Measures of rings of trees in favorable 

 localities should be used if possible to establish any climatic connection be- 

 tween these areas in order to use the cycle studies of Arizona for the benefit 

 of the Mississippi Valley. 



Summer conditions in the Valley partake of the torrid zone circulation 

 in the summer thunder storm, modified here also by the latitudes of the large 

 cyclonic storm tracks which still persist but lie farther north than in winter. 



A summer rainfall maximum of thunderstorm type is so characteristic 

 of Arizona that these summer rains are sometimes referred to as "Arizona 

 rains," but they are more widely distributed than that. They characterize 

 the Great Plains lying on the east side of the Rockies. They are quite evi- 

 dent as far north as Montana. Toward Texas they increase in intensity, 

 taking up more and more of the annual total, thus growing more and more 

 like torrid zone rains. In Texas and northern Mexico they even develop a 



