Going to the single-phase substations, unquestionably these should be without attendants, which means self-cooled units. Self-cooled units within the last few years have been developed so that there would be no trouble in providing transformers of 1500 to 3000 kv-a., and I think we could take about 2000 kv-a. as the size that would be ordered under present conditions. With 64,000 kv-a. to be transformed, we would have 32 units. The cost of these was taken at $4 per kv-a. I at first considered this too high. However, on second thought, considering that the transformers are self-cooled and that it has only been by very recent developments that we have gone up to such sizes, it is probable that $4 as compared with the $2.50 is an approximate figure.

Mr. Hobart pointed out the fact that three-phase units are cheaper than single phase. That may be taken wrongly, and I wish to say that a three-phase unit of the same capacity will cost some 15 per cent more than a single-phase, but three single-phase units having the same aggregate capacity as one three-phase unit will probably cost 10 to 15 per cent more than one three-phase unit. Then there is the question of voltage. On the three-phase and d-c. system, we have 30,000 for the hightension line. The secondary would be in the neighborhood of 800 to 1000, depending on whether the direct current was 1200 or 1500 volts. On the other hand the single-phase would be 11,000 in all probability. At first thought the 11,000-volt secondaries would cost more than the 1000-volt, and I believe Mr. Hobart so stated, but I doubt if the cost would be more in the large units that we are considering, like 2000 kv-a., and I believe that the 11,000-volt secondary would be perhaps 3 to 5 per cent cheaper than the lower voltage, due to the very heavy current on such large capacities. I will close with one other comment. The last point brought out in Mr. Hobart's paper refers to the development of the static converter or rectifier. I wish to point out from a transformer standpoint the fact that the employment of static rectifiers would permit the use of 60 cycles, as he has stated, rather than 25 cycles. If this is accomplished, and the rectifier is developed to a stage where it can be used for such work, and 60 cycles is adopted, I wish to point out that the transformers would be some 15 to 20 per cent cheaper than those quoted in the paper, which are 25-cycle transformers. In other words, that change in frequency will allow of 15 to 20 per cent saving in the transformers. This is a considerable item. C. M. Green: What is the comparative efficiency of the two? W. C. Smith: Before answering that I would like to get an idea from Mr. Hobart as to my assumption that the three-phase synchronous converter stations would be about the same capacity as the single-phase.

H. M. Hobart: It is my opinion that the greater drop in the line with alternating current, owing largely to skin effect in the rails, and the desirability from the operating standpoint, of cutting the line up into sections would make it well to have

numerous substations. I do not see why you should forego that advantage in the number of single-phase substations, even though you have 11,000 volts. The system will comprise a great many route miles and I hardly think it would be expedient for a railway with the very dense traffic considered in this paper, not to be able to cut up its line into as small sections as, at most, ten miles, whatever system is used.

C. M. Green: Do I understand you to mean ten miles from the station or place the station in the middle?

H. M. Hobart: I mean along the line of the railway. I think you would want to divide that up into ten mile stations anyway for a line with such heavy traffic as that with which my paper deals.

W. C. Smith: I mentioned the fact that three-phase transformers would cost less than single-phase transformers of the same aggregate capacity. It is also true that at the present time they would probably be more efficient; so that, choosing arbitrarily the three-phase at 8000 and single phase at 4000 kv-a., the probabilities are that the efficiencies would be about the same in the two systems. I think that would amount to a small consideration.

B. A. Behrend: I have consistently refrained in the past from expressing my personal opinion on the merits or demerits of the single-phase system as such, or of any particular part of the single-phase system, as the generating station, the transformer station, the substations, locomotives, etc. Perhaps few men have been so privileged as myself in obtaining an insight into the operation and also the construction and design, as well as cost and the waste of thought, labor and money expended in singlephase electrification, as I have, but I have resisted the great temptation for the last seven years in particular and for the last twenty years in general, to air my views in public. Mr. Hobart's judicial paper, treating this subject in a manner which takes it out of the range of polemics, allows me to express my opinion without compromising myself or the large manufacturing institutions with which I used to be connected for a great many years.

First allow me to express my entire agreement with Mr. Hobart's results. I believe he has stated these very fairly as to the cost of the generating stations. I believe his statement is not quite correct that units for three-phase current generation of 15,000 or 16,000 kw. at 1500 rev. per min. represent the limit at that particular speed. I myself worked out a year and a half ago a 25,000 kw. unit which, if it should ever be built, I have not the slightest hesitation in saying would be successful, and within the conservative guarantees as to temperature rise which it is now customary to make. I believe Mr. Hobart has given us an excellent summary, and after saying this, allow me to make a few critical remarks.

My criticism first of all is this, that the paper deals with but

one-half of the problem, and unfortunately the least important half. The generating station, electric generators, transformers and the line, to my mind, are a mere bagatelle in comparison with the problem at the other end, the locomotive, etc. The problem of operating your locomotives successfully after you have built them is more important. Mr. Hobart's paper is distinctly analytic and analyzes the problem up to the critical point --the locomotive. The method of analysis is distinctly orthodox. The single-phase system has come into large use. I designed some single-phase generators in 1892, and my master designed some 22 years ago, which are still in operation. Mr. Hobart knows

of one large successful single-phase power plant at Frankfort on the Main, built in 1892. We must take a heterodox view if we are to look at the single-phase system from the right angle. We must go back on the rules of standardization of the American Institute of Electrical Engineers. We must forget temperature rises as measured by thermometers. We must forget a great many things about electric generators and motors. Unless we do so, unless we turn heterodox, viewing this whole problem from a practical angle, we cannot understand the single-phase system, why it came and why it has done such wonderful good to the electrical engineering industry. I say we must become heterodox. Why? First take the generators-they are very large. Mr. Hobart is right in regard to his diagram only to some extent. Mr. Hobart shows two machines at 4000 kw. I would substitute one of 8000, which would reduce our cost a little. We should obtain a somewhat greater simplicity, which is quite essential. That is a minor matter, however, and does not enter into our argument, because the cost of the power station would be only 10 per cent greater. We must not figure in hundreds when our investment is in millions, and the money a railroad has to invest in a problem is not measured by the cost of the power house. Let your power house be twice as expensive and it will be still all right, if the operation of the locomotives and everything else were ideal; so I say we must be heterodox. We must size up the whole situation and view it not only as the power station-not important alone or the locomotives-not important alone-but we must take the whole problem and view it as a unit. If the power house breaks down because the electric generators are not designed properly or because some important things are overlooked, very well, change it, and you may be able to obtain a successful generation of power. Let me assure you that after a great many trials and a great many mistakes one of the large corporations in this country engaged in the manufacturing of machinery, succeeded in making the New Haven road a success. I hold no brief for this company or for the New Haven road, but I do want to reiterate that the success of that single-phase power transmission from Woodlawn to Stamford has been a landmark in the engineering business.

Now the next step in the solution of the problem is how can you simplify it-how can you eliminate sources of trouble. Reliability is the whole thing. This must be capitalized-it should be expressed in figures. You cannot do it, but let me assure you that any experienced banker, in trying to form an opinion as to whether his client's funds should be placed in any enterprise, first of all looks not at the financial standing but at the personnel connected with the enterprise. It is the personnel that counts. In regard to this railroad problem, it is the reliability of your railroad, of your electrification problem, which counts in the end, and that cannot be expressed in figures.

The last pages of the paper touch a question of polemics again, the problem of single-phase railways. As I have said, I hold no brief for single-phase railways. I am expressing my opinion for what it is worth. My opinion is that the single-phase system has a field and that the d-c. system has a field; and I have held that opinion for 20 years. The essential point is that both these systems have a field, and let us acknowledge those fields. Let us have the decency to say that a system is all right under certain conditions, but is all wrong under other conditions. Let us remember that 1500 volts for railway purposes has not been tried out thoroughly. I built a number of those high voltage d-c. systems that operate today, with commutators worn fantastically-but still going and delivering current-and therefore 1 believe that the development of the two systems will be side by side to some extent. The single-phase system, if all the kinks are taken out, which I hope will happen in the end, will have its place. Whether it was properly applied in the case of the New Haven road I do not know. I would like to ask Mr. Hobart whether he would not be so good as to give us his own personal views of the single-phase system, taking the whole system together, for trunk line electrification? I personally do not believe I shall live to see many trunk lines operated by electricity, because it will take such an enormous amount of money that I fear it will be a long time before the railroads can make the purchasers willing to defray the expense of these electric problems. H. M. Hobart: I do not see why we should use electricity where steam is better. I think the continuous-current system has its field, the single-phase has its field, and steam its field. Electrical transmission is appropriate where you can get a reasonable load factor. Where you cannot, there are other and more simple methods. These other methods may comprise the use of electric motors to drive the axles, and still not be electrical in the sense of transmitting electricity from a stationary generating plant to a moving train. I cannot recommend the continuous current system for all cases. It was in 1910 that I mentioned this in discussing a paper entitled "The Economics of Railway Electrification read before the Institution of Mechanical Engineers, and which advocated the 1500-volt, continuous-electricity system for a road a hundred miles long,

with one train per hour in each direction, a stop every three miles and a scheduled speed of 33 miles per hour. In my contribution to the discussion I stated that for this road the single phase system would be cheaper than the continuous, and that steam would be cheaper than either. I gave quantitative calculations in support of my contention. These will be found on pages 1239 to 1245 of the Proceedings of the Institution of Mechanical Engineers (London), for 1910.

C. M. Green: It is my purpose to speak of one very small section of this important subject, which we have before us for discussion, and that is the replacement of the synchronous converter by the mercury arc rectifier. I have had the good fortune for the past six years to be associated with the development of the rectifier for series arc lighting up to 9000 volts, 4 amperes, and on multiple work up to 350 volts, 40 to 50 amperes. The growth of the rectifier has been phenomenal, and I consider the future brighter than the past. The reliability of service is of the first and utmost importance, and the life of rectifier tubes and their service varies all the way from absolutely nothing up to over 14,000 hours, and I should be very much disappointed if in five years from now there will not be some tubes which have run over 25,000 hours, and inside of ten years I predict that we shall have tubes in service over 30,000 hours. Seven years ago tomorrow I had the privilege of running 60 kw. (10,000 volts and 6 amperes load) of electrical energy on a single tube, 25,000 volts alternating current across the anodes of the tube. The load for the tube was very excessive, however it did not give out during the short run. Since that time I have seen very much larger currents rectified with lower voltages. I have seen tests run up as high as 1000 amperes and over 1000 volts, but not at that amperage. Frankly speaking, I expect to see the mercury arc rectifier replacing the synchronous converter for 600 and 1200-volt service for railway and other work. It of course means a considerable amount of development and patience on the part of some of the operating companies in order that this result may be accomplished. You can readily appreciate that experimentally the tests must necessarily be of comparatively short duration for the simple reason that the amount of energy required is large, and when consumed on water-barrel or resistance load it runs into dollars with great rapidity so that after comparatively short tests it will be necessary to put the apparatus out into commercial service where some use may be made of the rectified energy. This furthermore has the advantage that under these conditions the rectifier is performing useful work and the expense of operating it is almost forgotten.

A Member: About what power factor would you expect to get with a mercury arc rectifier of large capacity on the alternating-current supply?

C. M. Green: I should be very much disappointed if we could not readily obtain on a three-phase supply a power factor of at