300 
Proceedings of the Royal Society 
vegetable world into herbs, shrubs, and trees. Under these circum- 
stances the interference of the naturalist may be less impertinent 
than might at first sight seem probable. 
§ 5. Before attempting classification, it is necessary to come to some 
definite agreement as to the nature of statistics ; and after laying 
aside the popular belief that it is an inexpressibly dreary accumula- 
tion of numbers by which anything whatever may be proved, we 
find that at least two hundred non-coincident definitions have been 
given by statisticians. Many of these assert statistics to be a 
science , many again regard it as a method ; while some, including 
the most recent foreign authorities, claim that it is at once both. 
But the sciences (using even the widest classification, that of Herbert 
Spencer) are logic, mathematics, physics, chemistry, astronomy, 
geology, biology, psychology, sociology, and ethics ; the methods 
of science (according to Bain) are simply observation and definition 
(classification), induction and deduction. We do not find statistics 
in either category. Some statisticians, however, hold the sound 
view that statistics is simply a quantitative record of the observed 
facts or relations in any branch of science,* and I have ventured to 
condense and define this view into a diagram, as follows : — 
Record of Facts (at given time). 
Qualitative. 
Quantitative. 
Y erbal. 
Numerical. 
Linear. 
Plane. 
Solid. 
Statements. 
j 
Graphic 
Statistics. 
§ 6. If this definition be correct, we obtain history by superposing 
or combining successive records, and this view is identical with that 
* For a valuable discussion of recent opinion as to the nature of statistics, in 
which this latter view is substantially maintained, see Hooper, “On the 
Method of Statistical Analysis” ; Journ. Statist. Soc. Lond., vol. xliv., 
March 1881. 
