ton-miles were not available until the latter part of the second month, that is to say, the net tonmiles for January would not be complete until March 15 to March 25. The information was taken from the waybills, the source of freight revenue statistics, and the delay in the settlement of interline waybills prevented an earlier closing of the revenue accounts. As a consequence the net ton-miles were received so late that they did not provide a satisfactory basis for the computation of train loading and car loading statistics. Another objection applied to waybill ton-miles as a measure of train- and car-performance. Tonmiles computed from the waybills rarely corresponded with the tons actually moved during the period for which the train-miles were reported, because of the delay in taking the interline waybills into account. There was always a "lap-over" of interline waybill ton-miles omitted from the preceding period, and a shortage of interline waybill ton-miles produced in the current period but not taken into account until the next period. In theory the "lap-over" should have balanced the shortage, but in actual practise the discrepancy was often so great as to invalidate waybill tonnage as a measure of train performance for any particular month. It was decided, therefore, to require that the net ton-miles, like the gross ton-miles, the trainmiles, the locomotive-miles, the car-miles and the train-hours, should be computed from the train reports. All of the basic data, then, would come from the same source. This insures the comparability of all these related data, and definitely allocates the transportation product to the particular period under review. Steps were taken later to utilize the train report ton-miles for revenue accounting and statistical purposes, and to discontinue the computation of waybill ton-miles except in certain states which require a separation of ton-miles between interstate and intrastate. It may occur to the minds of those who are interested in ton-mile statistics purely from the viewpoint of revenue and public service that the substitution of the statistics from the train reports may be less accurate than those from the waybills, and that the use of train report ton-miles as a divisor into freight revenue may affect the integrity of the important unit "Revenue per tonmile." There is, however, no cause for apprehension on that score, as experience has shown that differences in the two sets of statistics are so small as to be negligible. A comparison of the net ton-miles from the waybills (before that basis was discontinued) with those taken from the train reports, shows that for all railroads for five months the variation was but 0.8%. In this test the effect of the "lap-over" items is nullified because they are spread over a period of five months. The variation would be greater in the comparison of a single month. THE TIME ELEMENT IN OPERATING STATISTICS The importance of the time element in operating statistics had not generally been recognized. The majority of railroad men and financiers, are accustomed to think in terms of train-load, i.e., ton-miles per train-mile. Relatively few have been accustomed to think in terms of ton-miles per train-hour. The latter, however, is the better index to efficiency. The train-load, by itself, takes no account of speed. Ton-miles per train-hour are the resultant of load and speed. It is analogous to the horsepower unit. It combines in itself the net effect of the operating policy between the two extremes of loading the locomotive to every ton it can drag at low speed over the ruling grade, and of sacrificing tonnage in order to make the trip quickly. There is always a critical point between the two extremes which under normal conditions will produce the maximum of ton-miles per train-hour at the minimum cost per ton-mile. To illustrate: assume that on a given run there are sections of 1% grade over which a given type of locomotive can haul 1,500 gross tons (tons of car and lading combined) at a speed of 6 miles per hour on these maximum grades. The speed on other sections, of course, will be greater, but we will assume that, with a normal allowance for road delays, the run of 100 miles may be made in 10 hours. At that speed the production would be 15,000 gross ton-miles per train-hour (trainload of 1,500 gross tons times train-speed of 10 miles per hour). As the train- and engine-crews are on an 8-hour day basis, they would be paid overtime for 2 hours. (It is unnecessary here to go into the technicalities of the wage schedules which provide that mileage rates apply unless the miles per hour in freight service are less than 12.5, in which case hourly rates, based on 12.5 miles per hour, apply.) In this case it may be found that a reduction in the tonnage rating to 1,350 tons would permit an increase in the speed and reduce the trip hours to 8, or an average speed of 12.5 miles per hour. This combination of train-load and train-speed will produce 16,875 gross ton-miles per train-hour. The ton-mile production per train-hour is thereby increased from 15,000 to 16,875, and the cost per ton-mile is decreased because of the elimination of overtime. In this assumed case it is plain that the 1,500 ton rating is uneconomical. In the great majority of cases it may not be clear whether there would be any real economy in decreasing the load to increase the speed. These principles are ordinarily considered when the tonnage ratings are established, and it is the intention that they shall be set at a maximum which will not prevent the trains from moving at economical speed. In order to provide for the time element in operating statistics, the compilation of freight trainhours was required. Form 1 shows these basic data as well as gross ton-miles and net ton-miles per train-hour. It is possible, therefore, to trace the relationship between increases or decreases in the train-load and increases or decreases in the train-speed, and to note the combined effect in ton-miles per train-hour. The fluctuations in tonmiles per train-hour may, in turn, be compared with fluctuations in the cost per gross ton-mile, reported on Form 6. The time element has recognition also in Form 5, which shows as the final and inclusive unit of freight car efficiency, "Net ton-miles per carday." This unit is the resultant of three factors: 1. Average ton-miles per loaded car-mile 2. Per cent of loaded car-miles to total car-miles If, for example, the car-load is 30 net tons, the per cent of loaded to total car-miles is 70%, and the car-miles per car-day are 30, the net tonmiles per car-day are 630 (car-load-30 tonsmultiplied by per cent of loaded cars-.70-multiplied by car-miles per car-day-30). An improvement in any one factor favorably influences the inclusive unit; a loss in any one factor adversely affects it. If a campaign of intensive car loading brings about an increase of 10% in the car-load, to 33 net tons, but also causes a slowing up in car movement of 10%, to 27 miles per day, the road is no better off. In fact there is a slight loss, as the ton-miles per car-day will be 624 instead of 630. If, further, the heavier car loading increases the empty car movement, and thereby decreases the per cent of loaded car-miles to total car-miles, say to 67% instead of 70%, the inclusive unit-ton-miles per car-day-will suffer a further loss-597 instead of 630. The interrelation of these factors is often overlooked. From the single viewpoint of car performance there is no advantage in improving one factor if it is done at the expense of either or both of the other two factors. It should be noted, however, that one factor in this composite unit is practically constant from the viewpoint of the roads as a whole. The total number of freight cars varies but slightly from month to month. It is affected only by additions |