22 



north side of the greenhouse. Some types of houses, especially 

 those known as the two-thirds span roof, are boarded up on the 

 north side, and the loss of light in this style house could perhaps 

 be counteracted by using some good reflecting surface. 



The reflecting properties of snow are of course generally known. 

 Some tests made on a sunshiny day, when the ground was covered 

 with snow, gave some rather interesting but perhaps not conclusive 

 results. In one instance, where one tube was exposed to direct 

 sunlight and snow reflection, and the other mxrely to the direct 

 sunlight and light from the sky, we found a difference of about 17 

 percent, in the light intensity in favor of the first case. 



LOSS OF LIGHT FROM LAPPED GLASS AND SHADOWS. 



Three series of experiments were made in a three year old 

 greenhouse to determine the loss of light from, lapping glass, in 

 which we compared the light transmitting properties of lapped 

 glass, where the laps had becom.e more or less opaque from an 

 accumulation of dust, etc., with those of glass not lapped. 



In these tests the light recording tubes were covered with 

 opaque black paper, and exposed alternately with the lapped and 

 unlapi^ed glass. Slits of exactly the same size were made in the 

 paper, of approximately the same width as the lap, and in this 

 way the tubes received in each case only the light which passed 

 through these slits. The glass was typical second quality green- 

 house glass, and the lapped areas were about 3-8 of an inch. The 

 results are shown in the following table. 



Table VIII showing the loss of light from lapping glass. Av- 

 erage of three experiments. 



From the results of these experiments it will be seen that 

 there was a loss of 1 1 percent, of light throvigh the lapped areas. 

 It should be borne in mind, however, that the tests were made in a 

 comparatively new house, where the lapping was not wide and the 

 accumulation of dust comparatively small. 



We obtained similar results in another series of three experi- 

 ments, made with the same idea. The exposures were of two 

 hours duration, and an average of the three tests gave 309 for the 



