of Edinburgh, Session 1880-81. 
311 
human, civilised or savage : — for savage and animal societies, some 
columns here and there of course simply remaining blank. It is 
extremely simple, too, of understanding, and may, therefore, on all 
these grounds, satisfying as it does all the desiderata of a classifica- 
tion, legitimately claim a trial of convenience in use. In so far as the 
author’s own studies have extended, it has proved eminently service- 
able and suggestive ; and, moreover, if it be admitted to be a better 
classification than its predecessors, it is entitled provisionally to super- 
sede them for working purposes, according to the universal practice 
of the preliminary sciences, even although itself open to criticism. 
§ 23. Such, then, being the classification in its most general and 
abstract form, its completion — a task even more than that of any of 
the preliminary sciences — needing innumerable lifetimes broad and 
long, would require subclassification into the minutest details of 
social life and the filling up all the major and minor tables for each 
given society, with the facts of which so many are already gathered 
into economic and statistical libraries, and so many are being 
periodically collected, but of which perhaps even more await inves- 
tigators, and the notation of all these by all the resources of 
graphic statistics. Thus, with the comparison, too, of each record 
with those of other communities and of antecedent times — in other 
words the comparison of statistics with statistics, and history with 
history, it is hard to speculate how vast would be the outcome of 
elucidated laws.* 
But while this complete application is not within our reach, it 
must not be supposed that no application is or can be made to 
practice, nor that the present is a mere untested scheme. On the con- 
trary, a very considerable number of volumes of actual statistics, 
journals of societies, census returns, and works on special subjects, 
have been gone through, without the discovery of any facts relating 
to any given society which could not be immediately referred to their 
places on the tables, while the facts relating to relations between 
different societies arranged themselves conveniently as links between 
their respective sets of tables. + 
* It is interesting to compare these with the in many respects similar tables 
employed by Mr Spencer. See his “Descriptive Sociology.” 
t The reader may conveniently verify this statement by running through 
any such book, say a number of the Journal of the Statistical Society, or a 
VOL. XI. 2 R 
