height of the adult males, to whom my figures refer, was nearly? 5 feet 8 inches, and the value of their " normal-talent " (which is a measure of the spread of distribution) was very nearly zl inches. From these data it is easily reckoned that Class U would contain men whose heights exceed 6 feet i i inches. Even they are tall enough to overlook a hatless mob, while the higher classes, such as V, W and X, tower above it in an increasingly marked degree. So the civic worth (however that term may be defined) of U-class men, and still more of V-class, are notably superior to the crowd, though they are far below the heroic order. The rarity of a V-class man in each specified quality or group of qualities is as 35 in io,ooo, or say, for the convenience of using round numbers, as I to 300. A man of the W class is ten times rarer, and of the X class rarer still ; but I shall avoid giving any more exact definition of X than as a value considerably rarer than V. This gives a general but just idea of the distribution throughout a population of each and every quality taken separately so far as it is normally distributed. As already mentioned, it does the same for any group of normal qualities ; thus, if marks for classics and for mathematics were severally normal in their distribution, the combined marks gained by each candidate in both those subjects would be distributed normally also, this being one of the many interesting properties of the law of frequency.
