land and Bengal. In density the two countries are as nearly equal as can be (Table CXIII), but in Bengal 4.8 per cent. of the population is urban, and in England 61.7 per cent. (Tables XCI and CXII).
The fact is, that in order to show the proximity of human habitations by density figures, the unit of territory must be too small for use in statistical studies. Only by means of maps and cartograms can the average density be made to portray the conditions of residence, considered with relation, not to the land, but to the people themselves. It is this fact, doubtless, that leads Professor Mayo-Smith in Statistics and Sociology to treat of density of population in the chapter on "Physical Environment;" of concentration of population in the chapter on "Social Environment." Density is far more dependent upon natural conditions than is agglomeration.
But it must be admitted that the study of agglomeration by means of percentages of the total population dwelling in centres of a specified size offers some difficulties that are escaped when the comparison is limited to density. What, for instance, is the real significance of the terms "urban population," "rural population?" Does urban population include the dwellers in villages and small towns as well as those in cities? What is the line of division between urban and rural districts?
In ancient times the distinction between urban and rural populations was easily drawn, because the term "urban" was suited only to the few large cities. There existed, indeed, smaller centres of population like our towns or villages, but these bore few or none of the characteristic marks of the city; they were closely identified with the scattered population, which was devoted to agriculture. The development of the arts and sciences, the prosecution of industry, and political activity — all the social forces going to make up civiliza