128 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, I40 



section. But partial growth layers are not the only variables — the na- 

 ture of the outer surface of a growth layer varies from sharp to in- 

 definite, to diffuse, and even to a faintness that makes the surface 

 scarcely identifiable. 



The incidence of partial growth layers is directly proportional to the 

 number of intervals of localized canibial activity and maturation, and 

 these intervals depend upon the fluctuations of growth factors or of 

 their stimulating qualities. If soil moisture is of great importance to 

 tree growth, then partial growth layers should no doubt characterize 

 the lower forest border. Further, the more extreme the forest-border 

 conditions are, i.e., the more rapid and more intense the fluctuations 

 of soil moisture, the greater the incidence of partial growth layers. A 

 greater incidence of partial growth layers carries with it more rapid 

 variations among different radii of the same cross section. 



Therefore, as rainfall decreases, as amplitude of rainfall varia- 

 tions increases, and as lower forest-border conditions become more 

 extreme on the side of aridity (i.e., plant growth is more and more 

 dependent upon the individual local rain), different radii vary with 

 each other within shorter distances and to a greater degree. 



Nearly every example used in this work illustrates some form or 

 characteristic of multiplicity. Even so, a few pertinent cases will be 

 cited to illustrate variation among radii on the single cross section. A 

 comparison of figure 2, plate 5, and figure i, plate 6, and figure 2, plate 

 6, and figure i, plate 7, illustrates the differences which can exist on 

 opposite radii. 



Specimen TTC 33-io-a contains dated growth for 1938 to 1942, 

 five years. An over-all analysis of the section gives: 



Table 50. — TTC 33-io-a 



1938 — I see 



1939 — 2 sce-l-i psce+i sL 

 1940 — I sce-f-i psce 

 1941 — I sce-|-i psce 

 1942 — 2 sce+i sL 



Three points should be mentioned. First, with seven sharp, complete, 

 entire growth layers and two sharp lenses and three partly sharp, 

 complete, entire growth layers, different radii must show variable 

 numbers of sharp contacts. Second, four radii, 90 degrees apart, con- 

 tain 9, 10, 7, and 9 sharply bounded growth layers. How could these 

 be correlated with each other if just the four radii, and not the entire 

 section, were available? Third, if we momentarily adopt the assump- 



