28 CLIMATIC CYCLES AND TKEE GROWTH 



ico. Consideration of the best form for the expression of such data will 

 appear in a later chapter. Some features of that consideration are very- 

 general and may be stated here. 



The sequoia curves are continuous records. Weak and defective records 

 had to be eliminated and occasionally some merging had to be done. But 

 the plateau pines of Arizona and New Mexico have relatively short records. 

 They show maximum length of a little over 500 years and a common length 

 of less than 200. Hence typical records that were exceptionally long became 

 very valuable and much dependence has been placed on them. In spite of 

 these the number of mergings is large in the Arizona records. This is not 

 all a loss; each "joint" has had years of careful consideration; that means 

 in many cases a holding back of the results for years until the evidence was 

 convincing. Such care was considered worthwhile with the plateau data 

 because they approximate rainfall records so closely. 



Compressed Curves — Long records give a unique opportunity of comparing 

 cycle data from two separate geographical locations; they permit the study 

 of short cycles back in antiquity, but more especially they give us a chance 

 at cycles a century and more long. Adaptation had to be made of these very 

 long curves to the capacity of our analyzing instrument which can handle 

 400 or 500 terms at a maximum. Again 30 to 40 years has been the maximum 

 cycle length adequately analyzed in curves plotted on normal scale of 2 mm. 

 to the year. Hence the long curves have been "compressed" into shorter 

 lengths; that is, the time scale has been reduced so that longer cycles would 

 come within the capacity of the instrument. Different tests have been 

 made for effecting this compression, such as smoothing, taking means of 10, 

 means of 5, overlapping means of 5. Since our minimum cycle seen on the 

 instrument is 5.0 units, simple means of five years, not overlapping, have 

 come to be our usage. 



This should not be done on standardized values because of the danger 

 that standardizing, although not affecting short cycles, could more easily 

 influence long cycles; it must be done on the original measures of each indi- 

 vidual tree. Five-year sums so made are plotted and standardized anew for 

 each individual tree and averaged together in a new group table. These 

 new curves then either for individuals or for the group offer a cycle range of 

 25 to 150 years or more. To get yet longer cycles, we again go back to the 

 original measures and gather the 5-year means into new groups of 5, standard- 

 ize them individually, merge and average them to get cycles between 125 and 

 750 years or more. That is perhaps all we are justified in trying to get from 

 our long curves. 



Curve Character Figure: Mean Sensitivity — Mean sensitivity is the mean 

 percentage change from each measured yearly ring value to the next. It 

 uses the reading from a base and can not be obtained merely from a series 

 of departures from a mean. It is an important character in cross-dating and 

 therefore has a climatic significance, telling us essentially which curves are 

 presenting climatic data in the simpler terms. Low sensitivity occurs in the 



