Transpiration as a Factor in Crop Production. 
23 
color from the effect of moisture. Paper infiltrated with certain 
other chemicals than cobalt chloride may also be used. Since 
the rate of change in color depends upon the rate of water applica- 
tion to the paper, it is possible to make a somewhat relative com- 
parison of the rate of water loss from two leaves, but it is impos- 
sible to make an accurate quantitative measurement. 
Livingston (1912, pp. 121-123) describes a standardized cobalt 
test and suggests that it " should be of value in many problems of 
physiological ecology and of agriculture. The study, for example, 
of the relative drouth-resisting powers of different varieties, an 
important one in agriculture, should be much facilitated by these 
means." 
In a later paper (1913), "The Resistance Offered by Leaves to 
Transpirational Water Loss, " Livingston 1 gives tables and charts 
comparing hourly records from the standardized cobalt test with 
evaporation from white and brown atmometers, and with actual 
"relative" transpiration from plants. It is evident that the hygro- 
metric paper gives only approximate comparisons and cannot be 
considered quantitatively reliable. The method is wrong in 
principle for determining the plant response to normal climatic 
conditions. That portion of a leaf covered by an opaque paper 
held in close contact with it by means of a glass plate fastened with 
clamps for 10 minutes or longer certainly is not exposed normally 
to the surrounding atmosphere. The same may be said of the 
water bottle with which it is compared. It would be impossible 
to prevent the relative temperatures and other conditions of the 
leaf and water bottle from fluctuating during the reading, which 
destroys the possibility of standardization. 
FIELD ESTIMATES. 
Field estimates 2 of the amount of water used by plants are 
based upon initial and final soil-moisture samplings together with 
rainfall records during the period between moisture determi- 
nations. Since no accurate estimate can be made of the run-off or 
evaporation of moisture from the soil, results secured cannot be 
regarded as indicating the amount of transpiration. 
COLLECTION OF TRANSPIRED WATER. 3 
The method of collecting and weighing the transpired water 
has been used rather extensively. Some closed vessel containing 
1 Other investigators who have used a cobalt or similar test are Stahl 
(1894, pp. 117-145) and Merget (1878, p. 293). 
2 This method has been used by the following investigators, some of 
whom realized its limitations: Briggs and Shantz (1913), Leather (1911), 
Powers (1914), Widtsoe (1912), and Willard and Humbert (1913). 
3 This method has been used in recent years by Shreve (1914), Roberts 
and Freeman (1908), and Cannon (1905). 
