mold the character during these so-called "plastic
years of childhood." They showed (2) that the resemblance between twins
was two or three times as great as between ordinary children of the same
age and sex, brought up under similar environment. There seems to be no
reason, except heredity, why twins should be more alike. The data showed
(3) that the twins were no more alike in traits subject to much training
than in traits subject to little or no training. Their achievement in
these traits was determined by their heredity; training did not
measurably alter these hereditary potentialities.

"The facts," Professor Thorndike wrote, "are easily, simply and
completely explained by one simple hypothesis; namely, that the nature
of the germ-cells--the conditions of conception--cause whatever
similarities and differences exist in the original natures of men, that
these conditions influence mind and body equally, and that in life the
differences in modification of mind and body produced by such
differences as obtain between the environments of present-day New York
City public school children are slight."

"The inferences," he says, "with respect to the enormous importance of
original nature in determining the behavior and achievements of any man
in comparison with his fellows of the same period of civilization and
conditions of life are obvious. All theories of human life must accept
as a first principle the fact that human beings at birth differ
enormously in mental capacities and that these differences are largely
due to similar differences in their ancestry. All attempts to change
human nature must accept as their most important condition the limits
set by original nature to each individual."

Meantime other investigators, principally followers of Karl Pearson in
England, were working out correlation coefficients in other lines of
research for hundreds of different traits. As we show in more detail in
Chapter IV, it was found, no matter what physical or mental trait was
measured, that the coefficient of correlation between parent and child
was a little less than .5 and that the coefficient between brother and
brother, or sister and sister, or brother and sister, was a little more
than .5. On the average of many cases the mean "nature" value, the
coefficient of direct heredity, was placed at .51. This gave another
means of measuring nurture, for it was also possible to measure the
relation between any trait in the child a