1 6 INTRODUCTION. 



Now, since the operations of good administration are in their 

 nature gradual, and for their successful issue depend rather upon 

 uniform attention to their progress than upon occasional violent 

 efforts to adjust them to the current of affairs, it will be seen that 

 the most useful teachings are those gained from a continuous 

 record of events ; for these may be expected to recur with time, 

 while great catastrophes can seldom be provided for, or, in fact, 

 prevented, better than by the daily discharge of the duties 

 pertaining to direction. 



If there be a science correlative to the art of administration, 

 it must, like every other physical science, be founded on the 

 comparison of accumulated observations. 



Since the accuracy of the knowledge sought can be no greater 

 than the exactness of the data from which it is derived, in order 

 to make a proper comparison it is important that the observations 

 be as free from error as possible, and that they be measured by 

 a common standard. 



Errors of observation may be divided into two general classes ; 

 the instrumental, and those due to the personal bias of the 

 observer; the former referring to the standard itself, and the 

 latter to the application of the standard and the record of the 

 measurement. 



Whatever be the standard of measurement, it suffices for 

 comparison if it be generally accepted, if it be impartially 

 applied and if the results be fairly recorded. 



In regard to personal errors of observation it is generally 

 admitted that truth is most nearly approached when, having 

 observers of equal goodness they are most numerous, and when 

 they individually know least of the immediate consequences of 

 what they report. 



Hence the most truthful records will be had when each observ- 

 er's share of the work is reduced to a minimum ; for the number 

 of observers being the greatest possible, each one will have so 

 much the less to do, and being, therefore, better able to do 

 his share, each will feel more responsible for the accuracy 

 of the aggregate result and will besides be most likelj" to have 



