*POPULAR SCIENCE MONTHLY.*

tegrity of results. We must know the number of all the people, the total value of products and of capital, the aggregate wages paid in manufactures, and various other data where there would be little or no value unless all are included in the results. This class of statistics demands higher statistical qualifications, both in preparing for the enumeration and in the tabulation, classification and analysis of the results.

The third kind of statistics are those secured through a special investigation of certain representative facts. For instance if it is desired to learn the cost of producing iron and steel, it is not necessary, as in the previous case, to secure data for all the establishments engaged in such production. A few representative works offer ample information for determining cost of production. So, in endeavoring to ascertain the course of wages and prices, it is not essential to secure aggregates or data relating to all prices or to the wages paid to all the people employed.

The practical work of official statistical offices is divided into three parts also—first, the collection of data, which involves the preparation of schedules and instructions; second, the tabulation and the presentation of the results obtained, and, third, the analysis. No statistical table should ever be used without consulting carefully all textual treatment thereof, the accompanying notes and the analysis.

Facts may be presented in two ways—in tables, comparative and otherwise, and by the graphical method. The latter is popular and very effective in displaying proportions. The difficulty with it is that one cannot in a speech or in an article quote the diagrams, but it has a very important place in scientific statistics. The graphical method is carried to an absurd degree at times, but it nevertheless offers to a certain class of minds the very best method of determining results. In the final reflections upon statistics, however, one is drawn to the figures themselves.

With these preliminary statements, the general subject for discussion to-night brings us to the question: Is statistics a science or a method? It is not a matter of much consequence whether statistics constitutes a science or is simply a method. English writers on statistics generally consider it a method; continental writers, a science. American students often lean to the continental view. It is true that statistical research can be called a scientific method of determining facts and for studying various phenomena from which laws relating to life, production, distribution, consumption, etc., etc., can be drawn; and the method must be considered scientific, because by it the facts can be clearly stated, classified and analyzed, elements which make science in every department.

We speak of the science of botany, for one reason because all the