Is there a Cumulative Effect of Selection? 263 
Chief Results of Selection Experiment I. 
In considering the results of the long selection experiment I the 
the three following questions present themselves as of first importance. 
I. What is the general character of the variation exhibited in 
first-year egg production (fecundity) of the domestic fowl? Do we 
have here typical fluctuating variability ? 
2. Did the mean or average annual egg production per bird 
increase or decrease or remain unchanged during the course of the 
intensive and long continued selection practised in this experiment ? 
3. Was there any definite change in the variability of the flocks 
in respect to egg production during the course of the experiment? 
Did the birds come truer to type in respect to egg production at 
the end of the experiment than they did at the beginning? 
Variation in Annual Egg Production. A careful biometrical 
analysis of variation in annual egg production (fecundity) was made 
with our data. The methods of Pearson were used in this analysis. 
Without going into detail regarding the results of this part of the 
work it may be said that the data warrant the following conclusions: 
(a) Variation in egg production (fecundity) in the domestic fowl 
is, so far as can be determined, continuous and “fluctuating.” 
It is of such character as to give rise to unimodal frequency curves, 
similar to those known for variation in other characters and organisms. 
(b) The observed frequency polygons were all capable of 
satisfactory graduation by one or another type of Pearson’s generalized 
probability curves. The values of the analytical constants obtained 
indicate that, in general, these poultry fecundity data demand limited 
range curves for their graduation. The majority of the curves are 
of Pearson’s Type I (Skew curve of limited range). 
c) The skewness or asymmetry of these variation curves is, 
generally speaking, not great, though in most cases significant in 
comparison with its probable error. 
The most significant point for our present purpose is that in egg 
production we are dealing with a character which varies strictly in 
the usual or normal “fluctuating” continuous manner. 
The general character of these variation curves may be illustrated 
by a single example. Figure 1 gives the frequency polygon and 
fitted curve for variation in egg production exhibited by 275 birds 
kept in flocks of 150 birds each in the laying year 1905—06. 
