502 
Chandler—The Inter-Generation Period. 
from 34.3 to 35.5, in C from 32.7 to 34.3, and in D, in which 
the eighth generation is evidently exceptionally incomplete, 
from 31.6 to 33.7. 
These corrections seem to be in the direction of greater ac¬ 
curacy, and at first they indicate a strong probability that the 
correct period is somewhat greater than one-third of a century. 
But, on the other hand, it is necessary to recognize an uncer. 
tainty exceedingly difficult to estimate, but presenting a strong 
probability that the error in mean periods, derived like these 
from family records and printed histories, is in the opposite di¬ 
rection, and that the true period is shorter than the one thus 
obtained. For in all such histories the lines of descendants 
from sons are likely to be recorded far more fully than those 
from daughters, which are generally widely scattered in the 
records of the families into which they had married; and, since 
daughters are likely to marry at an earlier age than sons, the 
probable resulting error is as stated. 
If investigations in the opposite direction could be made with 
equal ease, that is, if a person’s ancestry could be traced as 
easily as the descendants of a common ancestor, not only would 
this difficulty be avoided, but there would also be eliminated 
from the computation the effect of the dates of birth of children 
dying unmarried or without offspring, and who therefore should 
not be considered in determining the period of successive gen¬ 
erations. But such an investigation requires far greater labor 
and patience than the search in the opposite direction, since the 
facts are so widely scattered in different books of record. 
Table E, however, shows a tolerably complete record, also 
represented graphically by Diagram 2, of the ancestry of two 
children, the mean date of whose births was 1876, extending 
back to include 11 of the 512 ancestors of the tenth generation. 
The mean of these nine periods is 32.2. But it is evident that 
the earlier of these generations are represented by too small a 
part of their members to furnish a reliable basis for a conclu¬ 
sion. There are, however, no vacancies in the last five genera¬ 
tions, only three of the thirty-two members of the sixth are 
missing, and more than half of the seventh generation have been 
