464 PROCEEDINGS OF SECTION P. 



which the inexpert, and some of us perhaps who think we 

 are expert, are exposed, through the too ready comparison with 

 each other of figures which apparently are applied to facts 

 of a like kind, but which really cover dissimilar facts. Such a 

 discussion becomes more and more indispensable, I think, on 

 account of one of the most important causes of the increased 

 diffusion of statistical knowledge in recent years — the extensive 

 development of statistical abstracts, hand-books, year-books, 

 manuals, dictionaries, statistical atlases, and such like works of 

 reference. Accustomed to see quantities, which are really dis- 

 similar in kind, placed together under the same heading, which 

 is done primarily for the mere purpose of reference, we come to 

 neglect the dissimilarity in our speech, and, by and by, in 

 thought. The numbers of different communities are com- 

 pared as if numbers alone were something in themselves, with- 

 out any thought of the diflferent qualities of the units : 

 production, imports and exports, and money wages in different 

 communities are spoken of as if they in all cases meant the 

 same things, and without any preliminary discussion of what 

 the figures I'eally do mean. All this is essentially mischievous, 

 and is contrary to the most elementary lessons in the study of 

 statistics. It is the part of the student to re-act against the 

 temptation to which he is exposed to use works which are only 

 good for reference in this haphazard fashion. 



POPULATION STATISTICS. 



At the risk of being common-place through enforcing con • 

 siderations which no one will dispute, I propose to begin with 

 the foundation statistics of all— those of population. It is 

 obvious at the first sight, when the statement is made, that for 

 very few purposes can the populations of different countries be 

 placed together as if the units wei'e the same. The peoples of 

 Europe and the United States are as a rule units of a very different 

 value from the units of population in Hindoo, Chinese, negro, and 

 aboriginal communities. Even among European peoples them- 

 selves there are enormous differences. 



It follows, then, that many questions of first importance for 

 which statistics of population are used, cannot be discussed at 

 all without reference to the quality of the units. The fact has 

 only to'be stated to be admitted. Among such questions, for 

 instance, is the question of the area that a given population 

 will support. The plain of Bengal, say, supports some seventy 

 million Hindoos — the population, in numbers, of the United 

 States. But if the consuming power of the Hindoo were at all 

 like that of the average man of the United States, how many 

 could Bengal support 1 The same, mutatis mutandis, comparing 

 even a French or German with a United States population. 

 The units in the difterent cases are entirely different. The area 

 of the United States might suffice with the same total 

 value of production that it now has for the support of 

 perhaps twice as many French or German as it could 

 support of people of the actual type of those now planted 

 on the soil of the United States. The question may be turned 

 about another way. Along with the increased capacity of 



