330 
Proceedings of Royal Society of Edinburgh. [july 4 , 
cannot be accepted as evidence that he is not at that time married ; 
and the fact that a married man is not mentioned as having children, 
cannot be accepted as evidence that he has none. Another circum- 
stance, which diminishes the value of Lodge’s Peerage for statistical 
investigation, is the practice adopted in some families of giving, 
among the collateral branches, the names, &c., only of the “ last 
surviving” uncles, aunts, &c., omitting all mention of those who 
are dead. It is clear that we may fall into very serious errors if we 
draw conclusions from incomplete information of this kind, and 
that very careful consideration is necessary to determine what use 
may safely be made of it. 
Such were the reasons which led me, on the former occasion, 
to reject all the facts relating to the collateral branches ; but 
further consideration showed me that, although the information 
as to the collateral branches is, on the whole, much less trust- 
worthy than that as to the immediate family, yet we cannot 
safely lay down the rule, to take the latter and reject the 
former; for, as w r e have seen, no strict line can be drawn between 
the two; the uncles and aunts and cousins being sometimes in- 
cluded in the immediate family, and sometimes among the collateral 
branches. In fact, we find, on comparing the editions for different 
years, that uncles and aunts and cousins, who are given in one year’s 
Peerage as members of the immediate family, will, after the lapse of 
some years, when the peer has died, be transferred to the collateral 
branches in the new volume. The distinction, therefore, between 
the collateral branches and the immediate family, is one that cannot 
be acted upon in practice; and we must seek for some other distinc- 
tion. 
Even among the immediate family, the information given is not 
always full and trustworthy ; and the facts given in Lodge’s 
Peerage for one year, sometimes differ from those given in the 
edition for another year, or in Foster’s work to be presently men- 
tioned. Careful examination soon showed me that the cases where 
the information is most defective and least trustworthy, are those 
of recent titles. When a man is created a peer who has been 
married many years, the information as to his children is often less 
full, and apparently less accurate, than in the case of peers who 
inherited their titles ; and I therefore came to the conclusion that? 
