10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Ill 



upon the individual investigator. Here, it is important to know if 

 cross-dating is a prerequisite to the merging of sequences for correla- 

 tion purposes. 



Figure 2 shows the so-called skeleton plots for HPC i to 9 and 

 a master plot made by a synthesis of the nine. The heights of the 

 inked lines bear an inverse ratio to the widths of the growth layers 

 on the wood as judged by the eye. No actual measurements enter 

 the skeleton plots. In order to judge the quality of the relationships, 

 the above figure should be compared with figure 3, which shows 

 excellent cross-dating from a forest-border area. The conclusion is 

 obvious: cross-dating as exhibited by the Holman Pass specimens 

 is of remarkably poor quality. One is tempted to say it does not 

 exist at all, for, if the dates were entirely unknown and within a 

 range of several centuries, one would have difficulty in convincing 

 others that the sequences match growth layer for growth layer as 

 they stand. In the present case the validity of the cross-dating, or the 

 only assurance that the growth layers grew on the dates assigned to 

 them on the skeleton plots, rests on two circumstances: (i) the 

 narrowness of the growth layers designated 1880 and 1893 and (2) 

 the probability that the soil moisture in the zone where the trees 

 grew was sufficiently adequate to prevent a temporary halt in growth 

 during any one growing season. 



A detailed comparison of the growth layers on all specimens for 

 each date in succession (fig. 2) brings out a lack of correspondence 

 that appears to emphasize a certain degree of individuality in the 

 site factors at each tree. 



An analysis of figure 2 was made, and the results were arranged 

 in table i. The record covers 116 years. Out of this length of record 

 only two cases exist, the growth layers designated 1880 and 1893, 

 wherein the growth layers are notably narrow on the eight specimens 

 bearing them. Two cases, 1836 and 1934, could perhaps be classed 

 with the previous two because the one growth layer that does not 

 conform on each exists on a so-called uniform sequence, one on HPC 

 7 and the other on HPC 6. Entries in table i wherein no narrow 

 growth layer exists on any one of the specimens number 55 cases. 

 With the number of specimens in the count disregarded, there are 

 22 cases where one growth layer is atypical, 23 cases where two 

 growth layers are atypical, 12 where three are atypical, and 2 where 

 four are atypical. For more than half the years the sequences are 

 from II to 50 percent out of agreement with one another. 



