102 MAINE AGRICULTURAL EXPERIMENT STATION. I91I. 



investigations was to compare the "error in weighing the parts 

 of a boiled egg with the error in weighing the parts of an egg 

 separated fresh. The weight of the fresh albumen may be 

 determined by using a scale pan to receive it when the egg is 

 broken, and weighing with it the filter with which the yolk and 

 shell are dried. This filter should either be previously balanced 

 by another filter which is placed on the weight pan, or it may 

 be weighed before use and its weight subtracted from the 

 weight of the albumen and filter. 



Balancing or weighing the filters was found by experience 

 to require about one minute each. An experiment was under- 

 taken to determine the error which would be involved in con- 

 sidering filters taken from the same package to be of equal 

 weight. This experiment involves two assumptions. The first 

 is that the weights of 50 filter papers taken in succession from 

 one package of 100 filters is a random sample of the weights of 

 the filter papers of the same make, number and size. The sec- 

 ond is that the distribution of the filters (as to weight) in the 

 package depends entirely on chance, i. e., is random. The filters 

 used were Carl Schleicher and Schull's No. 597, diameter 150 

 mm. Fifty filters were weighed. The weights ranged from 

 1.33 to 1.83 gms. The mean weight was 1.5636. The weight 

 of each of these fifty papers was written upon a card, and these 

 cards were then shaken in a cylindrical box. Four wires were 

 stretched through the box in different planes and at different 

 levels. One hundred drawings of two cards at a time were 

 then made, the two cards drawn being returned to the box and 

 the box shaken between successive drawings. The difference 

 between the weights recorded on the two cards drawn together 

 ranged from 0.00 gr. to 0.40 gr. The mean difference was 

 0.1346 gr. In 88 of the 100 drawings the difference was below 

 0.25. In other words this indicates that if one takes a pair of 

 filter papers at random from a package and assumes that they 

 are the same weight he will be in error in his weights because 

 of this assumption 0.1346 gr. on the average. 



In five (26 to 30 Table IV) of the eggs separated fresh the 

 filters used to balance the ones on which the yolk and shell were 

 dried were taken from the pile without weighing or balancing. 

 The error in weights is not perceptibly more in these than in 

 the other cases where weighed or balanced filters were used. 



