162 DOMESTICATED ANIMALS AND PLANTS 
selected for high oil produced in the ninth year a crop as high 
as 7.29 per cent oil and one as low as 2.58 per cent. Note, 
also, the rapid rate at which the distributions separate from 
each other, —so rapid, indeed, that in the fourth crop (1900) they 
no longer overlap but entirely part company ; that is, the lowest 
of the high oil is higher than the highest of the low oil.! 
If the offspring of the exceptional parent is in many cases so 
decidedly exceptional, how did the tradition start about the 
mediocre sons of great men? Naturally enough. Some of these 
sons are truly mediocre, even inferior, as we have seen, and in 
this, as in other matters, a few cases make a great impression, 
provided they are sufficiently striking. Every preacher’s son that 
goes wrong attracts special attention, —even more attention than 
does the long line of divines like the Edwardses or the Adamses, 
in which greatness almost invariably descended from father to 
son for many generations. This impression is akin to that 
other popular fallacy that people choose opposites in matrimony ; 
that is, that tall people prefer short mates; dark-haired prefer 
light ; phlegmatic prefer vivacious, ctc. Now the facts are, so 
far as they have been studied, that people prefer and choose 
their like to a surprisingly large degree. For example, the 
correlation or ratio of correspondence between husbands and 
wives amounts to 0,28 in stature and about the same in eye 
and hair color, whereas if they tended to choose opposites, 
it would be negative, and if they were indifferent, it would be 
zero. The fact is, that if we see one tall woman with a little 
husband, or the reverse, the grotesqueness of it all strikes our 
attention and we remark about it, reminding ourselves again of 
the “law of dissimilars”’; whereas we fail to notice the large 
number of properly assorted people that pass and repass, and 
thus overlook the real law that men and women in general mate 
by similarities and get along best when they do so. These few 
illustrations will show the need of accurate and somewhat exten- 
sive observation before hastening to generalization, 
1 For a fuller discussion, see “ Principles of breeding,” pp. 492-499. 
