466 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1955 



problems, such as those presented by false annual rings, and, second, 

 an identification which is not merely probable but absolute with a 

 selected segment of the master chronology. 



Absolute identification is possible by the forecast-and-verification 

 method. Given a tentative matching of test specimen and master 

 chronology by some ring characters, corresponding additional charac- 

 ters (locally absent rings, check segments of the ring sequences out- 

 side the test-dated interval, etc.) are sought; with a sufficient number 

 of verifications the probability of chance correlation becomes vanish- 

 ingly small. Since such verification depends on fairly close congruence 

 of both ring sequences, such parallelism should appear in the com- 

 pared measured growth curves and in the correlation coefficients, es- 

 pecially of the check or forecast intervals. 



The unqualified archeological dating to the year which tree-ring 

 analysis makes possible under favoring circumstances is, from a world- 

 wide point of view, of highly limited application. Even in the South- 

 west, ruins yielding only juniper or hardwood beams cannot be di- 

 rectly dated; in many other regions the absence of favorable species 

 makes the construction of sufficiently long master chronologies ex- 

 tremely difficult or quite impossible. This deficiency has been met in 

 a most unexpected way. 



In the past few years a new method has been devised by W. F. 

 Libby, then of the Institute for Nuclear Studies of the University 

 of Chicago, for the dating of wood and other materials by the meas- 

 urement of the amount of decay in the radioactive isotope of carbon, 

 C", which these materials contain. This very elegant and powerful 

 technique, now in process of active development, is applicable to a 

 wide range of organic and carbonate matter and appears able to pro- 

 vide dates from about 300 to some 30,000 years in the past with a prob- 

 able error that is satisfy ingly small. By supplying an absolute time 

 scale for regions where the construction of a master tree-ring chro- 

 nology anchored in the present is impossible, C^* dating greatly in- 

 creases the value of possible relative tree-ring dates in some of these 

 regions. 



DENDROCLIMATOLOGY 



Principles. — Evaluation of all the influences responsible for the 

 observed absolute growth of trees is obviously a far more complex 

 matter than that of setting up a centuries-long tree-ring index of 

 climate. Nevertheless, even with the latter limited objective, the 

 numerous pitfalls in analysis and interpretation of ring growth permit 

 no simple generalizations of the results of work in dendroclimatology. 

 It will perhaps be sufficient to sketch certain broad outlines as follows : 



1. A most fundamental property of all phases of ring growth is 



