70 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I40 



or resumes activity locally for a lens. It must not be forgotten that 

 somewhere on the area of a half-lens the free edge of the densewood 

 no doubt "migrates" so as to join a contiguous band of densewood. 



The sketches in text figure 12 suggest a different mode of origin 

 for half-lenses. Although five stages are given, all manner of transi- 

 tions have been observed. Our cross sections show many examples of 

 stages 2 and 3 (see pis. 3, fig. i ; 6, fig. 2 ; 8; 22, fig. i ; 24, fig. i ; 29, 

 and others). 



Fig. II. — The problem of the half-lens diagrammatically expressed. Dense- 

 wood, b, taken by itself is not the half-lens. Sequence a-b-c-d should be in- 

 terpreted in terms of cambial activity and maturation processes. 



The identity of the half-lens is not easy to visuaHze in three dimen- 

 sions. Structurally, classification need only be descriptive; but 

 physiologically, classification should be as nearly genetic as possible, 

 serving as a guide to the course of cambial activity and variation of 

 growth processes. 



The matter of cambial activity becomes complex indeed. Text fig- 

 ure 13 shows the complex part of the annual increment for 1939. 

 Over more than half the circuit 1939 is a growth layer of fairly uni- 

 form width. A circle of dark parenchyma cells extends entirely 

 around the circuit at a uniform distance from the start of the growth 

 layer, thus indicating that the rate of cambial activity was uniform 

 throughout most of the growth. Later, a portion of the cambium be- 

 came intermittently active, the activity punctuated by intervals of 

 slow and of no growth. As a result, 1939 bulges outward in a series 

 of lenses and half -lenses. The temptation is strong to call B and E 

 (text fig. 14) the half-lenses rather than A and D. To say that B and 

 E are the half -lenses is nearly tantamount to saying that a growth 



