According to this showing, maintenance of equipment, which we held in our analysis to be about one-half a constant expense and independent of traffic, especially after 1900, appears to have actually outrun the expansion of ton-mileage and passenger business. How largely this is due to actual purchases for the sake of future growth is not determinable. And maintenance of way outlay--one of our largely constant expenses--has increased, in fact, more rapidly than conducting transportation, which we held to be mainly variable. But these figures are confused by the failure to differentiate in the accounts, mere maintenance from actual improvements and additions to plant. Expenditures for these latter purposes, charged to operating expenses rather than to capital account, have been so enormous during these years of prosperity that they confuse the true facts utterly. It is to be hoped that now with the revised statistics since 1906, which will permit a clearer definition of these expenditures in detail, an analysis covering a series of years will bring out the real relationships. Equally important is the fact that these years have been characterized by rapid and extensive rises, both of prices and wages. Had our table covered a longer series of years the results would have been more clear. Until such an analysis be made, it will suffice for our purpose, viz., the analysis of the principles of railroad rate making, that we adhere to our first general conclusion, namely--that of the total expenditures of a railroad _at any given time_ about two-thirds of them are constant, while only one-third vary with the ups and downs of the volume of traffic. Comprehending in survey a long period of years, it might happen, as Acworth concludes, that nearly one-half of the total expenditures were entirely fixed in character, leaving the other half as dependent upon the amount of transportation effected.