292 
On the Systematic Fitting of Curves 
Suppose we attempt to fit a curve of the form 
2/ = 2/o(l + -, 
aj 
to the above data. The values of the constants in terms of the moments are given 
in Phil. Trans. Vol. 186, A., pp. 367 et seq., and we find 
2/ = 342-187 1 + 
47-13579 
1 - 
12-11059 
The mean being at _y + -4945, and the mode, which is the origin, at j + "7970. 
Here j denotes a fecundity of 9/15, and 1/15 is the unit of fecundity. Fig. 2 
shows that we have a very reasonable fit, — a curve which effectively represents the 
phenomenon. 
/ 
— 1 
>0 
1 1 
11 
V 
/ 
10 
) — 
— oty 
— 40 
— 30 
— 20- 
—10- 
o V,a V,5 % 'r,5 «/„ V,5 %5 '>'.3 'Via "/a "/a 'V.a 
l<''ecunditij. 
Fig. 2. Fecundity of 2000 Brood Mares, 8 or more coverings. 
(6) Illustration II. Half the battle of curve-fitting is to select a suitable 
type of curve for representing the observations. Mere increase of the number of 
