APPENDIX 2 



recognized in addition to those which are standard throughout mathematics, such 

 as the use of n for the ratio of the circumference of circle to its diameter, and e 

 for the natural base of logarithms. A most useful general rule has been proposed 

 by Fisher, viz. that Greek letters should be used to denote parameters and the 

 corresponding Latin letters to denote the corresponding statistics. 



In addition the following special usages are widely, though by no means 

 universally, adopted: — 



Z" or S to indicate summation, with superscripts and subscripts to show the 

 range of summation where necessary, 

 n for the number of individuals or observations, 

 p for a probability or proportion, with q = i-p. 

 X and y for variates, with x as the independent and y as the dependent variate 

 in regression analysis. 

 X for the mean of x. 

 CT or s for the standard deviation or standard error. N.B. s is preferable 

 for estimated and a for theoretical standard deviations on Fisher's 

 scheme. A subscript is used to denote the nature of the standard 

 deviation, thus s^ is the standard deviation of x, s- that of x, etc. 

 V for the variance, with the appropriate subscript, as with s and a. 



I = I - 1 for the amount of information yielded by a body of data. 



(-;) 



for the average information yielded by one observation or individual 



p for the theoretical correlation coefficient, 

 r for the estimate of p. 

 b for a regression coefficient. 



t for the ratio of an observed deviation to the corresponding estimated 

 standard deviation. 

 X^ for the ratio of an observed sum of squares to the corresponding 

 variance fixed by hypothesis, 

 z for half the natural logarithm of the ratio of two estimated variances. 

 V.R. or F for the ratio of two estimated variances. 



Variations of these conventions are used in agronomical and other statistical 

 work as well as, or sometimes more especially than, in genetico-statistical analyses. 



432 



