PREFACE   xxxix

worthies, or 5.50 each ; four sorts containing between them 7 names, or an average of 1.75 each. These figures are self-consistent, being each the sum of two practically equal constituents, and they are sufficiently numerous to be significant. The remarkable differences in their numbers, I 1.5o, 5.50, 1.75, when they ought to have been equal, has therefore to be accounted for, and the explanation given above seems both reasonable and sufficient.

CHAPTER X.-CONCLUSIONS.

The most casual glance at Table VII. leaves no doubt as to the rapid diminution in the frequency of noteworthiness as the distance of kinship to the F.R.S. increases, and it would presumably do the same to any other class of noteworthy persons.

I n drawing more exact conclusions, the returns must be deemed to refer not to a group of 207 F.R.S., because they are not a fair sample of the whole body of 467, and, for reasons already given, they are too rich in noteworthiness for the one and too poor for the other. "F hey will, therefore, be referred to the number that is the mean of these two limits-namely, to 337. 1 am aware of no obvious guidance to any better hypothesis.

The value of the expectation that noteworthiness would be found in any specified kinsman of an