228 



BIOLOGICAL EFFECTS OF RADIATION 



percentages. This question is treated under what is called the sampHng 

 error of a percentage, by which we mean the degree of variability in 

 percentages, as determined by repeated samples of a fixed size, when as 

 far as we can determine these samples have been drawn from the same 

 universe or the same experience. 



A discussion of sampling error can be approached most simply in 

 terms of an actual example and we shall consider cases taken from a paper 

 by Sievert and Forssberg (8), in which they discuss the lethal effect on 

 Drosophila eggs of varying dosages of X-rays at different intensities. We 

 shall make use of their results for the dose 165 r, which are presented in 

 Table 1. 



Table 1. — Mortality Rates of Drosophila Eggs Irradiated with an X-ray 



Dose of 165 r at Varying Intensities 



The question involved is the comparison of the percentages dying at 

 different intensities. These percentages could be compared just as they 

 stand, but since we know from experience that a repetition of the experi- 

 ment would not give an exact duplication of these results, the com- 

 parison becomes fruitful only after we have some estimate of the stability 

 of these results under repetition. If we consider the first two exposures, 

 the evidence can be presented completely in what is known as a fourfold 



table (see Table 2). 



Table 2. — Fourfold Table 



