No. 603] MENDELIAN CLASS FREQUENCY 149 



an operation, and we may resort to an approximate 

 method. The simplest one is to take .67449 o- on either 

 side of the median, which is approximately determined 

 by remembering that the median lies between the mean 

 and the mode and approximately twice as far from the 

 mode as from the mean. The criterion of whether this 

 method of fixing the quartile limits may be safely applied 

 will be found in the value of the skewness, 8k. In prac- 

 tical work this approximate method will give sufficiently 

 accurate results unless the skewness is very large, say 

 >0.6. 



We have by definition 



Hence having calculated the values of mean, mode and <r 

 by (i), (ii) and (iii) we can readily obtain (viii), since 

 d = mean — mode. 



We may now pass to the consideration of some numer- 

 ical examples, by means of which certain facts can be 

 better brought out than by further theoretical discussion. 



5. As a first and simple example we may take some 

 data, recently published by F. L. Platt,^^ on the results of 

 mating Blue Andalusian fowls. On account of the fre- 

 quency with which the Blue Andalusian case is cited as a 

 paradigm in Mendelism, coupled with the great dearth in 

 the literature of exact statistics of actual matings of this 

 breed of poultry, it seems especially worth while to dis- 

 cuss these statistics furnished by Mr. Piatt, on the au- 

 thority of Mr. W. J, Coates, a breeder of Andalusians. 



Table I gives the data, and in the last line, the Men- 

 delian expectation expressed in the form suggested in this 

 paper. 



The occurrence of the ''dark reds," which Mr. Coates 

 informs us had a pattern like a Red Game, is a phenom- 

 enon not mentioned in textbook accounts of Mendelian in- 

 heritance in the Blue Andalusian. In the present con- 



12 Piatt, F. L., "Western Notes and Comment," Beliable Poultry Journal, 

 Vol. XXlir, p. 665, 1916. 



