through the purchase of new cars, by the retirement of old cars, and by the fluctuations in the daily number of cars belonging to private car lines and Canadian roads. These changes are relatively slight in their effect on the total, so that it may be said that the net ton-miles per car-day for the roads as a whole will fluctuate almost directly with increases or decreases in the volume of freight traffic. This is not so true of individual roads which have some control over the cars on their lines. In periods of thin traffic, each road endeavors to reduce its number of cars belonging to other roads. This has a tendency to shift the balance of surplus cars as between railroads and regions, but, of course, has no effect on the grand total.

In one other notable particular, the new plan recognizes the prime importance of the time element, that is in locomotive utilization. Heretofore, there were no complete data to show the distribution of the hours in the locomotive day. Form 4 contains the most radical elaboration of orthodox statistical practise, as it provides for the division of serviceable locomotive time between that spent in productive road service, that spent at terminals "standing by" both before and after the road run, and that spent in the enginehouse between trips. The latter item is subdivided further to show how much of the time the locomotive is undergoing repairs or receiving other attention at the hands of the mechanical department forces, and how much of the time it remains idle in the enginehouse awaiting call from the transportation department.

An examination of the details of the hours of serviceable locomotives is facilitated by the requirement that the hours under each subdivision on the report shall be expressed also in percentages of total serviceable locomotive hours. Thus it is easy to compare the percentage of time on the road, at terminals and in enginehouses. When traffic is heavy it is desirable, of course, to show a high percentage of time in productive road service, and to take steps to control the unproductive hours at terminals and in enginehouses. When traffic is subnormal, it is inevitable that the time in the enginehouse (or as stored locomotives) will increase, but there is the same necessity for watching terminal time, as the crews are paid for the hours "standing by" at the same rate as on the road. The percentages, of course, show wide variations as between roads, reflecting differences in traffic conditions, in physical facilities, and in the policy of locomotive assignment-whether to single crews, double crews, multiple crews or to pooled crews. It is not safe to draw general conclusions from the figures alone without first hand information as to local conditions.

No attempt has been made by the Railroad Administration thus far to use the statistics for road-by-road comparisons. The figures, as reported, are summarized and published, but as comparisons with the preceding year will not be possible until the October, 1919, reports are received, the full value of the report for comparative purposes will not be apparent until the full year has elapsed. But even without the last year comparison the figures for the first year have

been of value, as they give a clear picture which localizes the extent of the non-use of power. There is force to the assertion that under existing conditions, with a surplus of locomotives, the value of the data is not as great as when there is a shortage of power. The continuation of the record, however, provides a bench-mark for comparisons of future performance, and will have an educative value as all concerned learn to appreciate the full significance of the figures. The low percentage of time on the road will surprise many who have little conception of what it really is. The high percentage of time at terminals (in certain instances) will throw light on overtime payments. The data should be of importance to supervisory or executive officers in passing upon recommendations for the purchase or transfer of power, or as to the necessity for improvements at terminals and enginehouses.

The requirements of Form 4 brought some protests from roads which had no statistics of distribution of locomotive-hours, and which consequently were put to some additional expense in compiling the figures. The answer of the Railroad Administration was that while it recognized the difficulties which lie in increasing the percentage of hours on the road in productive service, yet it maintained that effective remedial measures may not be applied without a complete knowledge of the facts, not from casual observation or off-hand statements, but from a current and comprehensive record.

It is pertinent at this point to refer to one feature which is subject to misunderstanding, and concerning which the United States Railroad Administration has been criticized. Prior to federal control, the Railroads' War Board inaugurated a monthly Summary of Freight Operation which among other statistics, showed what was termed, "Per cent of freight locomotives in shop or awaiting shop." No clear definition was given, but it was generally understood by the reporting carriers to apply only to the locomotives held out of service for general or classified repairs, which are made in the general shops, and was not meant to embrace locomotives held out of service for running repairs or other light repairs which in most cases are made in the enginehouses, although often made in the general shops.

Under federal operation the monthly summary above referred to was continued without change in basis until October, 1918, when the standardized statistics became fully effective. Under the new plan, the policy is to hold the operating department to a high standard in locomotive utilization, and the dividing line between serviceable and unserviceable locomotives was set at those which are held 24 hours or more for repairs of any kind, whether running repairs or classified repairs. The record is kept on an hourly basis, and the average number of unserviceable locomotives per month is obtained by dividing the monthly aggregate hours of locomotives held 24 hours or more for repairs, by the total hours in the month. This change in method naturally brought about an apparently large increase in

the percentage of unserviceable locomotives. The percentage (in freight service) on the last report on the old basis (September, 1918) was 14.8%. On the first report on the new basis (October, 1918) it is shown as 25.1%. Actually there was practically no difference in the condition of the locomotives in the two consecutive months. The difference is due entirely to the change in basis which was made under war conditions with a view to showing conditions in their worst light so that all concerned might be impressed with their responsibility for keeping locomotives employed to their maximum productive capacity. The use of the word "unserviceable" is somewhat strained, as it is not fairly accurate (although technically correct) to say that a locomotive which is held 24 hours for an hour's repairs, is "unserviceable.' Yet the line had to be drawn definitely, and it was set at 24 hours' delay for repairs of any kind.

As already stated the Administration has been criticized because its reports for July, 1919, show 27.2% of freight locomotives unserviceable, while the July, 1918, report on the Railroads' War Board basis shows the percentage as 14.1%. Practically all of the difference is due to the change in basis. The current summaries now bear the footnote:

The factor of "unserviceable locomotives" here used is a factor designed to be correlated with performance in transportation and is not designed to reflect and does not reflect the physical condition of the equipment. The factor reflects not merely the need for repairs but also the extent of delay in obtaining the repairs; and does this not merely with respect to classified repairs, the