Methods of counting apply especially to the study of human societies, because in them the individual units are widely different, and it is with the differences almost as much as with the resemblances between individuals that science is concerned. In the inorganic sciences and to a less extent in the sciences of plant and animal life, the units taken for purposes of investigation resemble one another far more closely than hu man individuals do. or at least their resem blances are far more obvious and for man's pur poses more important than their differences. In other than the social sciences. therefore. the ob servation of one or a few units may serve as a basis for general statements about the group, but in human societies it is frequently necessary to ascertain the existence, or non-existence, of a particular characteristic in every member of the group. This involves counting, and results in a statement in some numerical form, as a percent age, an average, or a series of figures showing the distribution of the characteristic among the members of the group. This indicates a funda
mental reason for the imperfect development of methods of measurement in the social sciences and a reason that statistics or the science of measuring aggregates of units, as distinct from individual units, is more necessary and has a wider range of applications in the study of social sciences than it has in the study of natural sci ences. There is, therefore, much ground for the opinion of those who define statistics as the nu merical study of social facts, or the numeri cal investigation of man's social life. But it seems more correct to hold that, for example, an enumeration through the century of the num ber of auroras observed each year and an obser vation through the same period of the extent of spots on the sun, and an arrangement of these two measurements in such a way as to show that they have fluctuated in close conformity, is to be included in statistics.