Miscellanea 
TABLE II. 
Analytical Constants for Variation of Fecundity in Brood Sows. 
435 
Constant 
Poland China 
Duioc Jersey 
4-1538 
1 -4259 
58-0100 
-0284 
•1684 
3-3622 
-3622 
-6392 
-0336 
-0701 
-1429 
2-0381 
7-4353 
7-8754* 
7-2924 
5-8922 
1 -8494 
116-1018 
-0167 
-1293 
3-3442 
"1 
K2 
SkewnesH ... 
Modal divergence .. 
Standard deviation 
Mean 
Median 
Mode 
Modal frequenc}' .. 
Origin 
-3442 
-6381 
-0198 
-0539 
-1308 
2-4274 
9-3372 
9-7631* 
9-2064 
1344-04 
5-7270 
950-2781 
1039-0000 
7-7770 
850-5350 
The modal divergence has values of 0-131 and 0-143, and the probable errors of these 
quantities on the assumption of norniahty are +0-026 and +0-021 respectively. Since the 
lower value is more than five times its probable error, there can be no doubt that we are dealing 
with skew variation. 
Turning to the kiirtosis, it is seen that ft -3 is positive and equal to about 0-35. The 
probable error of ^2 on the assumption of a normal distributit)n is +0 04 for the two cases. 
This indicates that these curves are significantly leptokurtic. The deviation from the meso- 
kurtic condition of the normal curve is more than eight times its probable error. 
Since /ci ( = 2/32 - 3/3i - 6) is positive and K2 greater than 0 and less than 1 we have the 
conditions for a Type IV curve in each case. The equations to the two curves are as follows : 
For the Poland China 
The frequency polygons and fitted curves for these two distributions are shown iu Figures 1 
and 2 respectively (p. 436). The graduation is very close in both cases. The distributions for 
several other years were calculated far enough to determine that they would lead to the same 
type of curve. This indicates that the present curves give us the manner of the distribution of 
fecundity as measured by size of litter in these breeds of swine. So far as the writer knows, 
distributions of fecundity in other organisms have led to either Type I or Type II (or 
occasionally to a normal curve). That is to say, from the meagre data at present available 
it would appear that fecundity distributions do not usually lead to skew curves with unlimited 
range in both directions, but on the contrary to some type of limited range curve. 
950-2781 
For the Duroc Jersey 
850-5350 
2/ 
* These are the observed values of the median. 
55—2 
