In regard to what tonnage should be included, I must differ a little from Mr. Quereau. The work that the locomotive does is partly in pulling itself, you might say, but it is a fact that the best designed locomotive will pull more weight behind the tender, that is, the best designed locomotive is that which has the greatest power with the least weight. Now, let us consider two engines of the same steam pressure and the same size cylinders and wheels; they would both have the same tractive force, and if we were to figure on the weight, including the engine and tender, they would each pull the same amount and would each be esteemed equally good engines. But if one engine, by the benefits of more modern design, we will say, is 20,000 pounds lighter than the other, of course it would be in favor of its tonnage, and it seems to me it would be only just to let the engine have the benefit of it and take the weight behind the tender.

I think Mr. Quereau will see what I mean—if we have an engine that is of more modern design, a better design, it will haul more weight behind the tender than another engine which may be just as powerful, taken from a cylinder standpoint. I think that is worth considering, as it is of more importance to know, in regard to any engine, what we are getting out of it in pulling than as a heat machine.

Now, the point which Mr. Quereau makes in regard to the statistics, in having total and net ton mileage separated, I think that is very beneficial. Of course, the general officers are interested in what the net mileage is, while the motive power department is more interested in the total ton mileage, and I think it would be a very good plan to have the total ton mileage and net, or revenue ton mileage, kept entirely separate.

The plan suggested in regard to weighing each train, of course, is admirable, and it would be most desirable, but I am afraid there would be so many practical difficulties in the way that it would be almost impossible. In other words, if we were to submit a plan that would require a lot of track scales, I think the question would likely be turned down on the first instance. It would be a good plan to do it, but it is almost impracticable.

In regard to the tonnage of empty cars, I would like to say that the committee (of which I was chairman a year or two ago) of the Master Mechanics' Association, appointed for the purpose of investigating that subject, gave the following formula for the percentage

which was to be allowed on empty cars. It depended on grades, and I think that is a pretty fair way to establish it. The formula was:

where "e" was the proportionate increase in resistance and “y” was the resistance due to grade. That is, on a level. This formula gave

about 30 per cent increase in resistance of empty cars over loaded cars, in tons, and on grades the resistance decreased as the grade increased, as this formula, dividing 1.8 by resistance of grade plus 6 would make allowance for the steeper grade, and the less the effect it would have on empty cars, which made a convenient method, because it was stated merely in percentages, and it could be worked so that when there was a grade of a certain per cent the despatcher would add 5 or 3 per cent to the weight of the train, and as that was constant for any one grade, they could have a list of the grades on the division and could decrease the loads 3 or 5 per cent, whatever the factor was, corresponding to the grade.

There is another point in regard to the ton mileage which is suggested here, but not entered into very thoroughly. Of course a good many of us are now keeping fuel records on the ton mileage, but I think it is just as interesting to keep the lubrication and the repair account on this mileage basis. We had a case the other day where a master mechanic seemed to be quite delighted with the fact that he had the lowest consumption of oil per engine mile of any in that division. I simply got out a little memorandum book that I had, and I showed him that while it was a fact that he had the lowest consumption per engine mile, that the tonnage hauled was so much less that the amount of oil per ton mile was precisely the same in each case; and I think we have every reason to expect that, in a greater or less degree, the increase in the size of our engines and naturally, also, in the size of the journals and the weights becoming greater, we must expect more lubrication necessary. At the same time those engines are expected to haul higher tonnage. If they do not, we have made a bad investment. But though large engines may decrease engine miles per pint of oil, yet if they show an increased ton mileage, I think we have every reason to be satisfied with our showing, and I think that this may explain some apparently unexplainable increases in lubrication.

MR. J. F. DEEMS (C., B. & Q. R. R.): I do not know, Mr.

President, that I have anything to add to what has already been said. While the ton mile basis for motive power statistics, or, in fact, any other statistics, is manifestly quite defective, I presume it is the best practicable unit that we can get; but in discussing this question. it reminds me of what Prof. Breckinridge said one day about some subject that we had under discussion,-he said it had as many sides to it as a globe, and this is almost the same.

I certainly agree with what Mr. Quereau says about the impossibility of making anything like an intelligent comparison between different roads, and that is doubtless equally true as between different divisions of the same road, and I do not know just how much better off we are when we come to make comparison between any one division of the same road from year to year. Perhaps somewhat better, but there is hardly a year passes by that there is not sufficient change in the capacity of the cars handled, and in the engines, and perhaps in the grades over a given division to make it necessary to figure that very carefully, or the statistics, as collected, will be misleading.

In the matter of loads and empties, it seems to me that is a matter that needs to be considered very carefully; and another point that has occurred to me is, you take a division over which the gradients are mostly in one direction, perhaps giving a rise of some seven or eight hundred feet in 150 miles. We will say one month in 1899 the traffic may be ascending the grade, while in 1900 the bulk of it would be descending; that is an element that will affect the results very materially, and unless it is watched closely the statistics collected would be decidedly misleading.

I recall a matter that was brought to my attention some time ago. One of the large trunk lines had been giving a great deal of attention to increasing the train rating, with a very marked success so far as the increase of the load was concerned, but the first thing they knew one of their competitors was getting most of the stock and time freight. Then the traffic department woke up and got after the operating department, and the result was a very marked reduction in train rating and increase in speed. That would vitiate any figures which might be collected.

There are so many things that belong in statistics of that kind that it seems to me the very best we can do in ton mile statistics is for every one to be watched very carefully. Yet I do not mean to say that they should not be collected, neither do I mean to say

that they should not be consulted, nor do I mean to say that we should not be in a measure guided by them, but for all these the factors or items by which the statistics are influenced have to be watched very closely.

I know of a place where they recently have been showing diagrams to secure this showing graphically for oil. They give each engineer a space on a big card, perhaps eight or ten feet square, and show his oil record in that way, which occurred to me, when I saw it, as being a very good thing. I think it appeals to the enginemen more quickly, perhaps, than it does to put down figures.

I notice something here that I am free to confess I cannot see how you are going to do it. Mr. Quereau says it is customary to credit switch engines with an arbitrary mileage, etc., and then he refers to work train service. I am really unable to see how it will be profitable to get anything like a proper ton mile basis for work train engines. with the character of work they are engaged in. It would be possible with the switch engines by simply taking the arbitrary mileage that is adopted, but I cannot see how it can be done for the work train engine. You can get a figure, I admit, but the question is, how reliable is it?

MR. G. W. RHODES (B. & M. R. R.): Mr. Chairman, this subject of ton mile statistics is one that I have watched with some little interest and with a good deal of disappointment, at first, because it necessarily changes our well established basis of comparison, viz., cars hauled one mile, or as more commonly expressed, car mile unit. It seems to me some of us consider the metric system in the same way; we do not want to learn a lot of things over again, and therefore we would rather have our present system continued. When the ton mile statistics for motive power came up, I at once saw that the question of what it cost per engine mile, per loaded car mile and for data that we are tolerably familiar with, would be lost; that our past history would be lost in so far as comparisons were concerned, and that we would have to start off with new statistics, and for that reason I have kept rather quiet on the subject and not said very much either in favor of or against it. At the same time, I felt every day, with the increasing differences in car capacities the car unit was becoming more and more incorrect, and that the inevitable result was the ton unit. We have cars now varying in capacity all the way from fifteen tons up to fifty tons. It must be apparent that this car unit.