NO. I GROWTH LAYERS IN TREE BRANCHES — CLOCK ET AL. 183 



Longitudinal variations, present in branches, are amply illustrated 

 by the foregoing tables and discussions. On the whole, the variations 

 are neither so numerous nor so striking as they are among different 

 radii of the same section. At least a person is left with such an im- 

 pression, probably because the composite picture of one section is 

 compared with the composite of another. 



A study of sections as individuals and as parts of branches illus- 

 trates clearly the transition of one growth-layer type into another, 

 the need for a classification of growth layers, and the basis of a classi- 

 fication as a descriptive localization in time and in space. 



Both circuit and longitudinal variation are high among the branches 

 of trees grown under lower forest-border conditions. With increase 

 in altitude, that is, with increase in rainfall and with decrease in am- 

 plitude or rainfall fluctuations, variations of growth layers around and 

 along branches decrease. Simplicity of classification increases. 



No general rule of longitudinal variation emerges from a study of 

 many branches from many trees. The amount of variation, or mul- 

 tiplicity, along a branch may remain the same, or it may increase in- 

 ward or outward on the branch. On a single branch one annual incre- 

 ment may increase in multiplicity outward, whereas another annual 

 increment may increase in the opposite direction. A rough census of 

 the behavior of the annual increments in the different branches showed 

 nearly double the number of cases where multiplicity decreased out- 

 ward over the number which remained the same or increased. 



RELATION OF DIAMETER FLUSHES TO TIP FLUSHES 



The relationship between diameter flushes and tip flushes is closely 

 connected to the longitudinal habit of growth layers in a single branch. 

 Ordinarily, of course, one tip flush is thought to represent one grow- 

 ing season. Actually, tip flushes may be dual, or even multiple, and 

 their relations to diameter flushes rather complex. 



No question arises as to the identity of a second tip flush if tip 

 growth has been measured accurately or closely observed day by day. 

 In such case, the presence of terminal bud scale scars between the two 

 flushes of the same season corroborates the evidence of the measure- 

 ments. The second tip flush is commonly much shorter than the first, 

 but this fact cannot be used alone as evidence of two flushes. Indeed, 

 on occasion, the first flush is the shorter, e.g., 1942 of Con T 2-6, 0.8 

 cm. and 2.8 cm. ; of Con T 2-7, 0.9 cm. and 3.3 cm. ; and of Con T 

 2-8, 0.3 cm. and 1.6 cm. In the same set of branches. Con T 2-5 

 showed 2.8 cm. and Con T 2-4 showed 2.0 cm. in a single flush for 



