142 
AGRICULTURE: REED AND HOLLAND 
lem like the present, we are more concerned with the temperature summation 
than with the mean temperature unless the range is very large. 
It is clear that we shall not arrive at a correct value if we take an arithmeti- 
cal average of the maximum and minimum daily temperatures, because we 
do not in that way take any account of the time during which either prevailed, 
or of the range of temperatures. For example, the minimum temperature on 
a given day may be 50° and the maximum temperature 90°, an average of 70° 
but if the maximum temperature prevails for only two hours out of the twenty- 
four, while the temperature varies between 50° and 65° for most of the daily 
period, it is obvious that the mere arithmetical mean, 70°, is a false expression 
of the temperature. The values must be weighted in order to give an aver- 
age which correctly represents the temperature condition. 
A method of measuring temperature summations has been employed which 
is believed to be fairly satisfactory. It consisted in finding the product of 
hours multiplied by degree of temperature above 40°F. and is expressed in de- 
gree-hours. A degree-hour may be regarded as one degree of effective tem- 
perature acting for one hour. The point 40°F. was arbitrarily chosen as a 
basal point, at or near which plant growth will proceed. The method of ob- 
taining the summation of effective temperature consisted in measuring with 
a planimeter the area between the pen tracing and the 40°F. line on thermo- 
graph records obtained from a self registering thermograph situated about 100 
yards from the plantation of sunflowers. This method gives a direct index 
of temperatures above the 40° point, but does not take into account the 
efficiency of temperatures as assumed by the van't Hoff-Arrhenius principle. 
The coefficient of correlation between the degree-hours and the increase in 
height of the sunflower plants for each seven-day interval was calculated. Its 
value turned out tober = 0.199 =t 0.187. There are some indications here 
of a positive correlation, but, since the probable error nearly equals the co- 
efficient in magnitude, no reliance can be placed upon the existence of a 
correlation. 
Reference to the graph showing temperature summations in figure 2, shows 
little correspondence with the curve representing growth increases, except in 
the first twenty-one days of the period. 
In a somewhat similar way we have investigated the possibility of a cor- 
relation between growth rate and the coefficient of the evaporating power of 
the air, the latter value being obtained from the readings of a spherical porous- 
clay atmometer-bulb located about one hundred yards from the plants. The 
coeffixient of correlation for these values was even less than that in the fore- 
going case, being 0.041 ^ 0.202. The coefficient in itself is so small as to 
lack significance, and when compared with its probable error it fails entirely 
to indicate any correlation between these two factors. 
Thtese statements are not to be construed as arguments against the effect 
of temperature and transpiration upon the rate of growth of plants. Our ar- 
gument is merely intended to emphasize the greater importance of the inter- 
