CYCLOGRAM ANALYSIS 



27 



These gross rings are very rare in a measured sequence because they are easily 

 seen on the stump and the radial cut from the stump is selected to avoid 

 erratic growth. Obvious fire or pest injuries or intervals where the rings could 

 not be correctly identified are excluded from the record. 



Merging occurs in parts of almost every group average. The several 

 trees whose growth curves are to be combined usually end at the same time, 

 perhaps in the very same year, but the beginnings extend over a long interval. 

 One or two curves may begin a hundred years before the others. The 

 shorter curves must be entered without producing fictitious maxima or 

 minima. The common method of treatment is to make a plot of all the 

 individuals on a large scale for several years before and after the merging 



A.D 350 



360 

 Fig. 13 — Merging diagram. 



point. Such a plot is shown in figure 13. An incoming or an outgoing 

 curve then easily shows whether its terminal values are in accord with the 

 other tree records. If only one or two terms are discordant they may be 

 omitted. If ten or fifteen terms are involved, the curve considered has its 

 final 9 values (beginning or end, as the case may be) weighted in a diminishing 

 scale from 0.9 to 0.1 ending at the final term with the weight of 0.1. Then 

 an average by weight merges the new curve into the group without any jar. 

 In rare cases a series of 20 has been used for merging. 



Technique of Long Records — The writer has been fortunate in developing 

 two accurately dated long ring records having climatic significance, one of 

 3200 years in the big sequoias of California, and the other of 1900 years in 

 the pines and firs of the Colorado Plateau, chiefly in Arizona and New Mex- 



