The question as to how far immediate heredity is a cause of differences found between individuals, can only be answered by measuring how much more alike members of the same family are in a given trait than people picked at random, and then making allowance for similarity in their training.  The greater the likenesses between members of the same family, and the greater the differences between members of different families, despite similarities in training, the more can individual differences be traced to differences in ancestry as a controlling cause.  The answer to this question has been obtained along four different lines:  First, likenesses in physical traits; second, likenesses in particular abilities; third, likenesses in achievement along intellectual and moral lines; fourth, greater likenesses between twins, than ordinary siblings.  In physical traits, such as eye color, hair color, cephalic index, height, family resemblance is very strong (the coefficient of correlation being about .5), and here training can certainly have had no effect.  In particular abilities, such as ability in spelling, the stage reached by an individual is due primarily to his inheritance, the ability being but little influenced by the differences in home or school training that commonly exist.  In general achievement, Galton’s results show that eminence runs in families, that one has more than three hundred times the chance of being eminent if one has a brother, father, or son eminent, than the individual picked at random.  Wood’s investigation in royal families points to the same influence of ancestry in determining achievement.  The studies of the Edwards family on one hand and the so-called Kallikak family on the other, point to the same conclusion.  Twins are found to be twice as much alike in the traits tested as other brothers and sisters.  Though the difficulty of discounting the effect of training in all these studies has been great, yet in every case the investigators have taken pains to do so.  The fact that the investigations along such different lines all bear out the same conclusion, namely, that intellectual differences are largely due to differences in family inheritance, weighs heavily in favor of its being a correct one.

The fifth factor that might account for individual differences is environment.  By environment we mean any influence brought to bear on the individual.  The same difficulty has been met in attempting to measure the effect of environment that was met in trying to measure the effect of inner nature—­namely, that of testing one without interference from the other.  The attempts to measure accurately the effect of any one element in the environment have not been successful.  No adequate way of avoiding the complications involved by different natures has been found.  One of the greatest errors in the method of working with this problem has been found just here.  It has been customary when the effect of a certain element in the environment is to be ascertained