42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Ill 



For group lo contrasted in the same manner : average variation, 

 average departure, average departure from mean variation, and corre- 

 lation with rainfall increased. 



For group 7 contrasted in the same manner : average variation and 

 average departure increased ; average departure from mean variation 

 decreased very slightly ; correlation with rainfall increased ; and 

 internal agreement of the trends among the several trees increased. 



For group 7 (restricted) contrasted in the same manner: average 

 variation, average departure, and average departure from mean vari- 

 ation decreased. 



The substance of the above summary is that the characteristics of 

 group II, made up of wet-site trees, and group 7 (restricted) agree 

 with those of rainfall whereas most of the characteristics of group 10 

 and group 7 disagree except for correlation with rainfall. Also, intra- 

 correlation on the wood rises in quality zvith greater rainfall. 



Obviously, these findings must be translated into a method whereby 

 study of the wood alone can be made to reveal changes in rainfall. 

 Two of the trees from the drier sites (in group 10) for some reason 

 reacted oppositely in comparison with the remainder of the collection 

 and when combined into group 7 (all trees) overbalanced the influence 

 of the remainder save for the one characteristic, average departure 

 from mean variation. It is clear in respect to the Holman Pass trees 

 that agreement of variation among the trees rises with increased 

 rainfall. Such increase in rainfall would be expected to lessen the 

 variation of rainfall within short distances on the ground as it affects 

 growth and .thus permit greater agreement among the trees. This is 

 well shown by table 2 especially among groups. Reasoning from a 

 knowledge of rainfall characteristics one can expect average variation 

 to decrease with increase of rainfall. 



Therefore, in regard to a study of the wood alone for evidence of 

 rainfall changes, the Holman Pass collection suggests the use of the 

 following methods: (i) the amount of agreement in directional 

 variation, including correlation and trend parallelism, among the trees 

 themselves ; (2) the change in average variation, average departure, 

 and average departure from mean variation among growth layers of 

 trees grown under conditions, or in a zone, at least as moist as those 

 for group II or for group 7 (restricted). 



RESUME 



In bringing together a resume of results it is well to recall the 

 several aspects of the study: The site chosen extended from well 



