20 BULLETIN 193 



stations where light intensities were .0374, .0069 and .0154; but none 

 grew at adjacent points where light intensities were .0278, .0246 and 



A study of the foregoing tables soon convinces one that the pres- 

 ent methods of determining light values in the forest are unsatisfac- 

 tory. Too little attention has been paid to the fact that the so-called 

 shade is a discontinuous shade and a constantly changing factor. The 

 variations in light intensity due to clouds, the impossibility of making 

 equal exposures in repeated readings, the variability of the forest 

 cover thus requiring the operator to choose a "typical station for the 

 forest under consideration," and the habit of reading only on bright 

 days about noon; all these make for inaccuracy and emphasize the 

 difficulty, if not impossibility, of determining the relation of forest 

 trees to light by a "study in the forests." 



The results of the work in the forests showed the necessity of 

 devising a habitat wherein continuous shade may occur. Accordingly 

 a series of frames was made of 1x1 inch lumber, each frame being 

 two inches larger than the one preceding it in the stack. The smallest 

 frame was 18 inches square. The larger frames were somewhat bulky, 

 but by using iron braces at the angles a sufficient rigidity was pre- 

 served. Seven of these frames were covered with white cheesecloth, 

 36 threads to the inch, and several with similar cloth which had been 

 dyed black. 



A ditch large enough to admit the arm of the operator was dug 

 under the north side of the frame or frames. When his arm was 

 in place with the photometer under the covers, a black cloth was 

 packed around it completely filling the space between the arm and the 

 ditch to keep out all light from entering around the arm. While a 

 reading was being made under the shades, a second operator made a 

 simultaneous reading in the open. When several covers were used a 

 long time exposure was found necessary, and two or more readings 

 were made in the open to compare with the reading under the covers, 

 thus taking cognizance of possible changes in total light which might 

 occur during the interior exposure. Readings were taken on both 

 clear and cloudy days. Table XIII may be taken as a sample. It dis- 

 plays a series of readings taken between 12 and 12:35 on August 16, 

 1912, when the sky was cloudless and the light in the open was 1. 



