272 Pearl and Surface. 
higher producers than other pullets of less intensely 
selected ancestry. 
2. The coefficients of variation clearly show that the ‘‘unregistered”’ 
birds are relatively less variable than the “registered.” There is 
only one exception to this rule and the difference in that case is 
very small. This means that the daughters of ‘‘200-egg”’ hens instead 
of conforming more closely to a particular type of egg production, 
as would on general grounds be expected, actually conform less 
closely to type than do birds of less closely selected ancestry. 
Discussion of Results. 
The answer given by the investigations here summarized to the 
questions stated as the title of this paper is definite and clear. So 
far as the character fecundity (egg production) in the domestic fowl 
is concerned long continued and carefully executed experiments give 
no evidence whatever that there is a cumulative effect of the selection 
of fluctuating variations. 
The facts brought out by this work indicate clearly enough that 
all birds which have equal records of performance in respect to egg 
production are not alike in their ability to transmit fecundity. Judged 
by the performance of their offspring a group of “200-egg” hens, 
while very homogeneous so far as performance records are concerned, 
must be very heterogeneous in regard to the constitution of the 
germ cells relative to the character fecundity. Some ‘‘200-egg”’ hens 
are apparently capable of producing offspring with very high 
laying capacity (cf., data in Table II). Other ‘‘200-egg” hens 
exactly similar in all observable respects lack this capability. These 
facts are of exactly the same order as those which have been brought 
out by Mendelian work, showing that the constitution of the soma 
furnishes no certain criterion of the condition or constitution of the 
germ cells. 
But the assumption, tacit or expressed, which lies at the 
foundation of mass selection methods in practical breeding is that the 
soma does as a matter of fact give a working criterion of the 
constitution and potentialities of the germ cells. One breeds from 
the superior individuals in regard to somatic characters because he 
expects that the offspring will be superior. Is this expectation well 
founded? Altogether much evidence is accumulating from widely 
different sources to show that simple selection of superior individuals 
as breeders can not alone be depended upon to insure definite or 
