tag="{http://www.w3.org/1999/xhtml}a">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.

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.

It clearly appears from this table that the effect of correlation between Ability and Environment is to increase, and not to diminish, the closeness of association between Success and Ability. Indeed, if the correlation were perfect, Success would become an equal measure both of Ability and of Favourableness of Environment.

These arguments are true for each and every branch of Success, and are therefore true for all: Ability being construed as Appropriate Ability, and Environment as Appropriate Environment.

The general conclusion is that Success is, statistically speaking, a magnified, but otherwise trustworthy, sign of Ability, high Success being associated with high, but not an equally high, grade of Ability, and low with low, but not an equally low. A few instances to the contrary no more contradict this important general conclusion than a few cases of death at very early or at very late ages contradict the tables of expectation of life of a newly-born infant.

Chapter VI.—Nomenclature of Kinship.

Specific kinships are such as “paternal uncle” or “maternal uncle,” as distinguished from the general term “uncle.” The phrase “first cousin” covers no less than eight specific kinships (four male and four female), not taking the issue of mixed marriages into account. Specific kinships are briefly expressed by a nomenclature in which fa, me, bro, si, son, da, Hu, Wi, stand respectively for father, mother, brother, sister, son, daughter, Husband, Wife. Each of these syllables is supposed to have the possessive 's added to it whenever it is followed by another syllable of the set, or by the word is when it is not. Example: Let the person from whom the kinships are reckoned be called P, and let Q and R be two of P's kinsfolk, described respectively as fa bro and me si son. That means that P's father's brother is Q, and that P's mother's sister's son is R. It is a simple and easily intelligible nomenclature, and replaces intolerable verbiage in the description of distant kinships. My correspondents used it freely, and none of them spoke of any difficulty in understanding it. Its somewhat babyish sound is soon disregarded.

Those relationships that are expressed by different combinations of these letters differ specifically; therefore, in saying, in the next chapter, that each person has “roughly, on the average, one fertile relative in each and every form of specific kinship,” it means in each and every combination of