Autosomes and Sex Chromosomes 



135 



Measuring '^Goodness of FW in Dihybrids. In Chapter 8, it 

 was pointed out that the test of whether an observed ratio fits a 

 theoretical expectation, such as 3:1 or 1:1, could be made 

 by comparing the deviation of the observed ratio with its stand- 

 ard error and that it could equally well be determined by the 

 X^ method. In a ratio involving more than two terms, however, 

 the usual method of the geneticist is to determine goodness of 



TABLE 3 



Values of x" * 



* Table 3 is abridged from Table III of Fisher: Statistical Methods for 

 Research Workers j Oliver & Boyd Ltd., Edinburgh, by permission of the 

 author and publishers. 



fit by calculating x^. By means of a table like R. A. Fisher's, 

 the probability that a deviation as great or greater than the one 

 in question will occur by chance can be determined from the 

 value calculated for x^. In a ratio of four terms, the method 

 differs from the method used when the ratio contains only two 

 terms, chiefly by the number of degrees of freedom involved. In 

 a 3 : 1 or other ratio involving two terms only, the probability is 

 obtained by using the first line in Table 3. The reason is that 

 if there are a certain number of individuals in the population 

 there may be any number of dominants or of recessives from 

 zero to the number in the population, but when either the num- 

 ber of dominants or the number of recessives is determined, the 

 number in the other term in the ratio is fixed because it must 

 include all the rest. In a ratio of two terms, there is only one 



