658 Journal of Agricultural Research vol. xx.No. 8 



The fourth method consisted in freezing the whole tree by surrounding 

 and covering it with a two-walled metal vessel containing ice and salt. 



The apparatus is shown in Plate 80. 



The factors that determine the amount of damage done and that 

 need to be controlled in the experiment are: 



1 . The kind of buds. 



2. Their stage of development. 



3. The minimum temperature. 



4. The humidity. 



5. The duration of the freeze. 



6. The rate of thaw. 



The first three are of most importance. By keeping the other factors 

 fairly constant and varying the fifth and sixth, little difference in the 

 results was noted. In almost every case in nature, as well as in our 

 experiments, the humidity just as freezing occurs is practically 100 per 

 cent. Transpiration into a closed vessel will ultimately give this result, 

 and the best desiccating agents will not keep the humidity down appre- 

 ciably. This holds true also in the orchard simply by the cooling irre- 

 spective of the transpiration, because even in such a dry section as 

 the arid West, with a humidity as low as 50 per cent and a cool spring 

 day of perhaps 45 F. noon temperature, the dew point would be 27. 5 

 F., which is about the temperature at which slight damage is caused. 

 Where the humidity is higher, as it is in most places east of the Rocky 

 Mountains and west of the Sierra Nevada Mountains, the dew would 

 collect and the humidity would be 100 per cent even before the buds had 

 cooled to the danger temperature. In all the work here reported the 

 humidity was practically 100 per cent. 



While the whole tree was being frozen, several minimum thermometers 

 were suspended at different places in its branches, and the air was stirred 

 by an electric fan driven with storage batteries. The humidity was 

 determined by a continuous reading hygrometer, and the rate of thaw 

 and duration of freeze were recorded by means of a thermograph that 

 was placed in the branches. 



The cost of the different methods is about the same for freezing the 

 same number of buds. Adjoining limbs and adjacent trees were thinned 

 to the extent that the branch or tree had been thinned by the frost, 

 and the yields in the fall were noted for comparison. A greater varia- 

 tion in the factors, and thus a greater number of different experiments, 

 can be secured for the same expenditure by freezing the branches on 

 the tree rather than the whole tree. 



The results of the natural and artificial freezing experiments are 

 presented in Tables II to IV. 



