The Inter-Generation Period. 
501 
generation was by no means complete when the records were 
compiled, a presumption supported by the observed fact that its 
numbers, instead of being more than those of the seventh gen¬ 
eration, are considerably less. But a compensation for this 
deficiency is made by the unusual length of the first period, due 
to the fact that there is no place for a mean when but one per¬ 
son is considered, and that the condition of this one, a pioneer 
in the unsettled wilderness, seems to have been such as made it 
by no means easy for him to find a wife, and he remained un¬ 
married until the age of thirty-four. It is also worthy of at¬ 
tention that the sums of these periods, when taken in succes¬ 
sive groups of three, are each near an exact century, the largest 
excess, of only four years, being in the sum containing the effect 
of the pioneer’s isolation, and the equal deficit where it is caused 
by the generation still incomplete. It is further evident that 
the greatest irregularities in the generation lines are in those 
places where the data are imperfect; as, for example, at the 
extreme left of the diagram, where the only line of descent is 
that of a member of the third generation who was born twenty- 
one years after the marriage of his parents. If the data for the 
older branches of that family had been available, the generation 
line must have been elevated to a marked extent. The records 
of the branch presenting such unusual irregularities at the ex¬ 
treme right of the diagram are likewise very imperfect. 
The families denoted B, C, and D are not presented graphi¬ 
cally; but the tabular exhibits of the dates of birth show less 
nearly uniform inter-generation periods than those of family A; 
yet in no case does the mean of the seven periods differ from 
the third of a century by more than 1.6 years. The sums of 
three successive periods, however, present much greater irregu¬ 
larities, rising in one case to 115 years. It may be noticed, 
also, in B, C, and D, as in A, that the intervals between the 
seventh and eighth generations are very short, and the same 
evidence of incompleteness in the eighth generation is shown 
in each case by the small number of births recorded. Therefore 
it may be well to neglect the eighth generation and find the 
mean of the six preceding periods; which modification increases 
the length of the mean period in A from 33.3 to 34.5, in B 
