186 DOMESTICATED ANIMALS AND PLANTS 
The first class is, on the average, decidedly inferior because 
only partly improved, and though high-class individuals occasion- 
ally occur, even they are next to worthless for breeding purposes, 
because, under the law of ancestral heredity, the influences are 
so diverse that regressions and reversions will be common, even 
inevitable. The second class, on the other hand, may be virtu- 
ally and even actually high-class, pure-bred animals, whose 
records may have been lost by fire or other accident, or are 
otherwise untraceable. Such animals may be every whit as use- 
ful for everyday purposes as are registered stock, but the 
impossibility of knowing or stating their blood lines of course 
destroys their sale value as breeders. 
Now and always the great mass of our farm stock will be 
unrecorded animals. The business of the improver is to raise 
the quality of this stock to the nearest possible approach to the 
best recorded blood. This is the best we can hope to do, for 
there will always be a few best animals, and these are really 
the only ones worth recording. It adds nothing to the value 
of an inferior pure-blooded animal to record it, — indeed, it is 
better that such animals be not recorded, — and one of the 
first steps in practical improvement is to get rid of the pedi- 
gree scrub, meaning by that, those animals of good breeding 
which are themselves worthless. 
We have, then, two great classes of animals : first, those whose 
ancestry is known and recorded; and second, those whose an- 
cestry is not known. Manifestly, most of the best animals and 
all of those valuable for securing additional improvement are 
in the first class. 
Systems of breeding. With these facts before us we are ready 
to discuss the relative merits of different systems of breeding, 
which may be briefly outlined as follows : 
1. Mixed breeding, in which no attention whatever is paid 
to ancestry. It has the merit of cheapness and the disadvantage 
that no further improvement need be expected. If any syste- 
matic attempts should be made toward selecting to a constant 
