president's address. SECTIOX Gl. 52S 



statistics are kept on suitable lines to disclose fully the real nature 

 of the railway business. It is always a great advantage when^ a 

 common scheme is adopted in regard to railway statistics of a series 

 of States between whom there is railway traffic. 



18. Analysis of Trade and Commerce. — Possibly on few economic 

 subjects has there been more confusion of thought than concerning 

 the nature of value, of v.hich there are at least four kinds — viz., use, 

 esteem, cost, and exchange value. In the majority of economic 

 questions, it is the last which is most frequently under consideration. 

 The mathematical formulation and general theory of the subject has 

 somewhat recently been woi-ked out with great thoroughness by 

 Walsh. Time will not permit of more than the briefest reference to 

 the matter, but it may be pointed out that in analysing exchange 

 value relations, the possible errors of arithmetic, harmonic, and 

 geometric averaging, all need to be considered. The difficulty about 

 price relations may be seen by a very simple illustration. We propose, 

 let us say, to analyse the significance of imports of two years, ex- 

 pressed in pounds sterling. If we take the ruling prices of the former 

 year and apply them to the quantities of the latter year, the ratio of 

 the deduced and actual value is in general different from what is 

 obtained by taking the ruling prices of the latter year and applying 

 them to the quantities of the former year. The method followed by 

 the Commonwealth Bureau is that adopted by the Board of Trade, 

 and has been employed not because of its validity from the standpoint 

 of theoiy, but simply because it places the Australian comparison on 

 the same basis, however imperfect, as the British, and it certainly 

 gives a result of some value. 



This is one of the matters in which international agreement as 

 to method is a desideratum. Dutot's, Carli's, Scrope's, Young's, 

 Probisch's, Lehr's, Nicholson's, and other methods have all been 

 thoroughly analysed, and the adoption of a universal method has 

 been ablv discussed bv Walsh. 



IV.— THE PRESENT ASPECT OF STATISTICAL PROBLEMS. 



1. Return to Original Conception of Statistics. — What has been 

 veiy roughly indicated in the preceding remarks discloses the fact that 

 by virtue of its inherent jiiature the science of statistics is not merely 

 resuming its original range and dignity, but transcending that range. 

 The wide reach of statistical method, and the demands it makes in 

 regard to the necessary mathematical equipment, mark it out as on 

 the broadest planes of human knowledge. It concerns itself with the 

 totality of human affairs. No longer can mere tabulations of informa- 

 tion be regarded as statistic, nor compilers of and commentators on 

 statistical tables as statisticians. Much of the meaningless tabulation 

 which in the past has passed current for statistics will in the near 

 future be dispensed with, while penetrating investigations, such, for 

 example, as those with which the founders of " Biometrika " and 

 similar publications are associated, will take their place. 



Not only will statistics become more incisive, not only will mere 

 voluminous and increasing detail be abandoned, not only will inquiries 

 be intelligently directed to-ward intelligently conceived ends, but even 

 the field embraced will be enlarged. In a word, it may already be 

 said that the original field has been fully resumed. 



