THE CURVE OF THE BIG TREES. 



155 



Still another factor produces an error in this same direction. As the trees get older 

 they tend to throw out buttresses at the base. In the center of the buttresses the rings 

 of growth are well developed, wide, easily measured and counted, and are rarely missing. 

 In the portions between the buttresses they are thinner, more crinkled and hence hard 

 to measure, and are more apt to be missing than in the neighboring buttresses. When 

 the tree is younger and there are no buttresses, the rings preserve nearly the same width 

 throughout the entire circumference, and a measurement in one place is as good as in 

 another. Where the buttresses exist, however, a measurement along them gives an 

 impression of growth greater than is warranted by the facts. If the buttresses and the 

 depressions between them are averaged together, the average growth of an old tree appears 

 to be distinctly less than where the buttresses alone are measured. It would be possible 

 to avoid a part of the error due to buttresses by choosing only young or unconmionly 

 symmetrical trees which flare but little. We were confronted, however, by the paramount 

 necessity of finding as many old trees as possible; therefore we felt obliged to take every 

 tree of great size which gave promise of giving fairly accurate measm-ements. The errors 

 involved in measuring along the buttresses are less serious than those which would have 

 arisen had we measured along the hollows, where there was danger of losing a considerable 

 number of rings. After the danger of the buttresses was reahzed, we tried, so far as 

 possible, to take our measurements from points half-way between buttress and hollow, 

 but this in many cases proved impracticable, and we were thrown back upon the good, clean 

 sections of the buttresses. 



From what has been said above, it is evident that the corrected curve of growth of 

 the sequoias, as given in figure 38, is wrong in two respects: (1) the early portions fall 

 unduly low, partly because of scarcity of data and consequent imperfections in the cor- 

 rection for age and longevity, and partly because those corrections are deliberately designed 

 to ehminate any effects due to a change of climate occurring in a cycle of 3,000 years or 

 more; (2) the later part of the curve rises too high on account of the flaring buttresses 

 at the base of the trunks of the trees. With our present incomplete data these two errors 

 can not be ehminated by any strictly mathematical correction. The best that we can do is 

 to adopt some standard unconnected with the trees, and swing the curve of growth so 

 that at a few critical points the relative height of the two is the same, taking the greatest 

 care, however, that we do not alter the sinuosities. For this purpose I have adopted the 

 fluctuations of the Caspian Sea as given in Chapter XVII of the "The Pulse of Asia." 



