70 • BURTON EDWARD LIVINGSTON 



A group of cups, with one of previous standardization, are 

 first operated without reading for a period of from 12 to 24 

 hours, to allow the water films in the walls to come into equilib- 

 rium, after which they are operated on the rotating table until 

 the water loss from each is sufficiently great (say from 20 to 50 

 cc.) to make errors of reading negligible. The corrected read- 

 ing of the cup of previous standardization is divided by that 

 of each cup to be standardized, thus giving what is termed the 

 coefficient of correction in each case. The coefficients are at 

 once seen to be numbers by which the readings of the respective 

 cups are to be multiplied, in order to give the loss which would 

 have been shown from a cup with coefficient of unity, operated 

 under the same conditions and for the same time. By "corrected 

 reading" of the previously standardized cup is here meant the 

 product obtained by multiplying the actual reading by the co- 

 efficient derived at the earlier test. 



The group is operated for a second period and the calculations 

 are repeated, giving a second coefficient in each case. Where 

 the two coefficients thus obtained agree, or show differences not 

 greater than 0.03, the cups are regarded as satisfactorily stand- 

 ardized. They are then removed from the table, dried and 

 wrapped in paper, the cup number and the mean coefficient from 

 the two tests being written upon the wrapper. Cups' which show 

 larger fluctuations in the value of the coefficient are left on the 

 table and operated, with the same or' other standard cups, till 

 a coefficient is fixed upon, or until they prove themselves impos- 

 sible of satisfactory standardization, in which case they are dis- 

 carded. 



The basis of standardization is still the old 6 cm. form of cup, 

 but only for historical reasons. The first standardizations, car- 

 ried out at the Missouri Botanical Garden and at the Desert 

 Laboratory, in 1907, amounted simply to comparing the readings, 

 under similar conditions, from a large number of 6-cm. cups. It 

 was found that about half of these agreed very closely among 

 themselves and that their water loss for any test was about the 

 average loss from the entire group, some seventy-five cups alto- 

 gether. The largest number of cups with agreeing rates were 



