38 BULLETIN 1233, U. S. DEPARTMENT OF AGRICULTURE. 



in April and May, possibly because at that time the higher air is 



cooled by passing over snow fields, while locally the ground may be 

 bare. The daily minima, it is found, are likely to show the greatest 



depression at the ground, in the coolest periods with extremely calm 

 nights. 



As has been indicated, the exposure of the thermometers at the 

 ground was not such as to insure their recording strictly the tempera- 

 ture of the air itself at the time of the highest temperatures each day. 

 Because this was realized, at the end of the period above described, 

 the equipment, with thermograph added, was placed in a standard 

 shelter, tree from direct insolation or reflected insolation, and well 

 ventilated through the floor to insure radiation at night. The site of 

 the shelter was only a few feet from the other site, but was more fully 

 clothed with vegetation. Since this change was made the higher 

 maxima near the ground have been at all times just about counter- 

 balanced by the lower minima, and the result has been, as before, a 

 higher daily range for the ground position, but essentially the same 

 mean temperature. The data for April, 1918, are especially interest- 

 ing, because during this month, in the earlier arrangement, the 

 greatest excess of ground heat was recorded. (See Table 7.) 



Table 7. — Departure of ground temperatures from temperatures 20 feet above ground 

 at the control station (with similar exposure of instruments) . 





April, 1918. 



July, 1918. 





-1.3 

 -1.8 



+2. y 



+0.2 



Mean maxima 



+2.2 





—1.9 





-2.0 



Mean duil v range 









1 See note, p. 29. 



Although these data are hardly extensive enough to invalidate 

 those previously presented, it is probably safe to state that in an open 

 situation of practically neutral aspect there is no important differ; 

 between the mean temperatures near the ground and those 20 feet 

 above the surface; but that the ground location may easily show daily 

 ranges exceeding by 3° or 4° those of the atmosphere above. Hence, 

 of course, the probability of frost is much greater at the ground than 

 on the tower where our main temperature record lias been secured. 

 It is seeli from Table 4 that the probable minimum for the second 

 decade of June, when expressed in terms of around temperatures, may 

 \>c a> low as 32° F., and for the first decade of September as low as 

 34° P. It is thus apparent thai the frost-free season may actually be 

 10 or 1 5 days shorter at the ground than would he indicated by Table 

 4; and of course it is the seedling at the ground level that is most 

 sensil ive to frost . On the morningof September?, L9 18, for example, 

 the ground temperature was 30.2"° F., ana the frost was severe enough 

 to kill tomatoes but not other garden vegetables. The minimum 

 temperature 20 feet above ground was 32.5 . 



