NO. I GROWTH LAYERS IN TREE BRANCHES CLOCK ET AL. 289 



the amplitude and the wavelength of growth activity provided its in- 

 tensity is plotted against time. The physiological activity of a growth 

 flush, no matter where and when intensified, produces a growth layer, 

 entire or partial, complete or incomplete. 



Our work on trees grown under extreme lower forest-border condi- 

 tions and in the forest border itself shows that multiple flushes within 

 one year can, and do, yield multiple growth layers within an annual 

 increment. From the forest interior, where it is thought that an 

 annual increment seldom if ever contains multiple growth layers, 

 out to the forest edge and into the scattered trees beyond, the 

 tendency toward and the presence of multiplicity increase until it 

 becomes of high order as illustrated by the trees studied in the 

 present report. The error in the simple counting of sharply bounded 

 growth layers, interpreting them to be annual, is therefore rather 

 large among trees grown under extreme forest-border conditions. 

 Within the true forest border as commonly understood, where con- 

 ditions are more favorable to tree growth than at Lubbock, Tex., the 

 range of this error is estimated to be from zero to 15 percent, the 

 average error being about 5 percent. 



It is abundantly clear, therefore, that growth rings cannot be used 

 for precise dating of historical events if the trees grew in or near the 

 forest border. A tree-ring calendar centuries in length, from the 

 lower forest border, would, it is believed, have its accuracy enhanced 

 for purposes of archeological dating if it contained numerous "bench 

 marks" or check points, if it could be duplicated by other materials 

 or records, or if it could be duplicated by other students. Even so, er- 

 rors due to indistinguishable intra-annual growth layers could not be 

 wholly eliminated. 



It is also abundantly clear that growth-layer sequences from forest- 

 border areas, where multiplicity is characteristic, cannot be used for 

 precise rainfall or cycle studies. The reasons are quite obvious. 



Although growth layers from forest-border trees do present difficul- 

 ties in chronological reckoning, it must be remembered that all trees, 

 whether from the forest border or forest interior, require soil moisture 

 and that many of them must meet the problem of soil-moisture deficit. 



In spite of possible chronological difficulties, growth layers may be 

 used in an analysis of environmental factors in at least two ways : ( i ) 

 by selecting trees that have grown under conditions where the annual 

 cycle of growth is seldom broken into secondary cycles, that is, trees 

 grown toward the interior and away from the lower forest border, 

 and (2) by analyzing growth patterns. The chief gross elements in a 

 pattern are multiplicity (or its absence), partial growth layers, vari- 



