both are _low_; they will
produce Successes of Grade III. This is still more clearly seen by
sorting the results into Table II., from which it is clear that a
high grade of Success is statistically associated with a high, but
less, grade of Ability, a medium with a medium, and a low grade of
Success with a low, but less low, grade of Ability.

TABLE II.--ABILITY INDEPENDENT OF ENVIRONMENT.
_____________________________________________________________________
| | | |
| Grades of | | |
| Success. | Contributory Combinations. | Corresponding Abilities. |
|___________|_____________________________|___________________________|
| | | | | | | |
| I. | AE | AF | BE | 2 of A | 1 of B | -- |
| II. | AG | BF | CE | 1 of A | 1 of B | 1 of C |
| III. | CG | BG | CF | -- | 1 of B | 2 of C |
|___________|_________|_________|_________|_________|________|________|

Secondly, suppose A, B, C to be correlated with E, F, G, so that A is
more likely to be associated with E than it is with F, and much more
likely than with G. Similarly, C is most likely to be associated with
G, less likely with F, and least likely with E. The general effect of
these preferences will be well represented by divorcing the couples
which differ by two grades--namely, AG and CE, by re-mating their
constituents as AE and CG, and by re-sorting them, as in Table III.
The couples that differ by no more than one grade are left
undisturbed. The results now fall into five grades of Success, in
four of which each grade contains two-ninths of the whole number, and
one, the medium Grade 3, contains only one-ninth.

As remarked previously, the grades are not supposed to be separated
by equal steps. They are numbered in ordinary numerals to distinguish
them from those in Table II.

TABLE III.--ABILITY CORRELATED WITH ENVIRONMENT.
_____________________________________________________________________
| | | |
| Grades of Success. | Contributory | Corresponding Abilities. |
| | Combinations. | |
|____________________|_______________|________________________________|
|