266 galton.org
statistical averages for reasons that are quite unconnected with theology,
and which I will explain and illustrate. Briefly, it is the very purport and
claim of statistics to isolate the effect of specific influences from all
others, whether known or unknown, that may act concurrently with them.
Suppose a large number of silk-worms to be tended by a caretaker, and
that an observer watched his proceedings as well as he could, but only
during the day-time and through a telescope. We will further suppose his
observations to show that the worms were of various breeds, and that they
were fed in various ways, irrespectively of their breeds, and that the
observer desired to discover the relative effects of breed and feeding on
the growth of the worms. There can be no doubt as to the principle on
which he would work; he would classify his observations so as to
compare race with race, and he would reclassify them to compare nurture
with nurture. By this well-understood treatment he would isolate the two
classes of influence.
Now suppose the caretaker had a custom wholly unknown to the
observer, of feeding the worms in various ways during the nighttime, how
would that affect the statistical conclusions? I answer, only by increasing
the amount of individual deviations from the average result, so that, other
circumstances remaining the same, the observer would not attain the same
constancy in his averages unless the number of observations in his groups
was larger than before. Let us consider the ways in which the interference
of the caretaker might act.
(1) Suppose he favoured a particular race by giving food to every
individual of it during the night-time, then the effect would be that every
individual of that race, by virtue of his belonging to the race, would be
benefited. The observer who noticed the generally thriving condition of
worms of that race would be justified in accepting it as a racial
characteristic, for it would be the consequence of the race of the worm.
(2) Suppose the caretaker gave additional food in the night to the
particular set whom he had fed the best during the day-time. The observer
would rightly ascribe the more or less thriving condition of that set to the
peculiarities of their nurture.
(3) Suppose the caretaker acted conversely, feeding those in the night-
time whom he had inadequately fed in the day. If the night and day
feeding were of equal importance, the observer would find the effects of
Nurture to be nil, and rightly so. If they did not balance, he would notice
the differential effect.
Thus far we see that the relative total effects of Nature and Nurture
