1908.] PUBLIC DOCUMENT — No. 33. 73 



day, their mean found, and that figure employed as the tem- 

 perature of the day. This method has given results which were 

 obviously very inaccurate as to the sum of heat for the time, 

 and much more variable on some days than on others. How- 

 ever, in comparing difterent sets of observations taken in this 

 same manner, the variations have averaged up with each other 

 fairly well and relatively correct comparisons could be made. 



For several years the division of horticulture has been carry- 

 ing forward a series of investigations in this field by methods 

 not hitherto applied to this interesting sulyect. The novelty 

 and value of our methods consist in their being very much more 

 accurate than any previously employed. Instead of depending 

 on public meteorological reports for the computation of accu- 

 mulated temperatures, we have employed the recording thermo- 

 graph. This instrument makes a complete and continuous 

 record, showing exactly the quantities of heat to which it has 

 been exposed. 



Greater accuracy was secured, secondly, by placing the ther- 

 mograph in close proximity to the plants under observation. 

 The temperatures recorded are therefore the exact temperatures 

 to which the plants were subjected. When it is understood 

 that previous investigators have been forced in many cases to 

 accept meteorological records taken many miles from the plants 

 under observation, it will be seen that this feature of our work 

 constitutes a considerable improvement. 



In the third place, much greater accuracy was secured in 

 methods of computing sum temperatures. Having a perfect 

 record from the thermograph, there remained only the problem 

 of securing an exact measurement of the heat quantities thereon 

 represented. This problem was solved by the use of the pla- 

 nimeter. The thermograph record appears in the form of an 

 irregular line having a generally horizontal direction. If the 

 height of this line, representing degrees of temperature, be 

 measured from some l)ase line (as, e.g., the zero of the ther- 

 mometer) , we may readily construct a figure which offers an 

 exact geometrical representation of the c|uantity of heat which 

 we seek to measure. Such figures are shown in Fig. 1. Hori- 

 zontal distances represent degrees of heat ; so that the product 

 of length l)y height, giving the area of the figure, gives also the 

 quantity of accumulated heat. 



