delivered the majority opinion of the Court. Hammond, C. J., and Marbury, and Finan, JJ., dissent. Dissenting opinion by Finan, J., in which Hammond, C. J., and Mar-bury, J., concur, at page 605, infra.

This appeal principally involves the question of whether the decree of the Circuit Court of Baltimore City (Harris, J.) dated November 21, 1966, reversing order 56137, dated February 18, 1966, of the Public Service Commission of Maryland, because the Commission’s order was not supported by substantial evidence on the record considered as a whole, was correct. We have concluded that the lower court’s order was correct and will affirm it.

The facts, which we must consider in some detail, are vo*582luminous. There were seven volumes of testimony taken before the Commission containing over 1000 pages. There were some 27 witnesses and many documentary exhibits. The Joint Record Extract in this Court consists of 746 printed pages.

There were two cases instituted before the Commission by the Brotherhood of Railroad Trainmen (the Brotherhood). One case, No. 5939, was instituted on November 30, 1964, against the Pennsylvania Railroad Company (the Pennsylvania) ; the other case, No. 5946, was instituted later in 1964 against The Baltimore and Ohio Railroad Company (the B & O). Both cases were consolidated for trial before the Commission and were consolidated for trial before the Circuit Court of Baltimore City on appeal to that court.

In No. 5939', the Brotherhood challenged Rule 99 (in regard to flagging) of the Pennsylvania as changed and promulgated by that railroad’s order of October 18, 1964, making Rule 99 non-applicable (1) to trains operating under the automatic block signal system for following movements on the same track, and (2) to tains operating under manual block signal systems for following movements on the same track where Rule 316 is in effect, except when required by train order or time table special instructions, because the elimination of flag protection of stopped trains by flagmen allegedly “created an unreasonable risk of harm to train crews and, in case of passenger trains, to the travelling public.” The relief prayed was that the Commission require the Pennsylvania “to rescind its order dated October 18, 1964, in the State of Maryland so that full protection will be given to standing trains from other trains on the same track.”

In No. 5946, the Brotherhood challenged a General Order of the B & O, effective January 1, 1965, which would make Rule 99 (in regard to flagging) non-applicable to trains operating under automatic block signal systems for following movements on the same tracks, because the elimination of flag protection allegedly “creates an unreasonable risk of harm to train crews and, in case of passenger trains, to the travelling public.” It was alleged by the Brotherhood that the General Order was in violation of the Commission’s orders in case No. 5768, passed on January 8 and July 14, 1962, which required *583the B & O to provide flag protection on tracks controlled by an automatic block signal system “particularly when trains are stopped at places other than station stops.” The relief prayed for was that the B & O comply with the prior orders in case No. 5768 and that the B & O be temporarily restrained from placing in effect its General Order modifying Rule 99 until a hearing was had and the matter determined.

In order to understand fully the contentions of the respective parties in the present case, it is necessary to review the development of the methods of operating the railroads involved. This background testimony is largely undisputed. We will first consider the testimony in regard to the operation of the Pennsylvania.

Prior to the early 1900’s and the beginning of the use by the Pennsylvania of the automatic block signal system, trains were operated on a time table and train order system. This meant that when a train stopped there was no signal protection behind it except the flagman who would walk back of the stopped train a sufficient distance to protect against an approaching train proceeding at a speed up to 50 miles per hour. Rule 99 provided as follows:

“99. When a train stops under circumstances in which it may be overtaken by another train, the flagman must go back immediately with flagman’s signals a sufficient distance to insure full protection, placing two torpedoes, and when necessary, in addition, displaying lighted fusees.

“When recalled and safety to tlie train will permit, he may return.

“When conditions require, he will leave the torpedoes and a lighted fusee.

“The front of the train must be protected in the same way when necessary by the fireman.

“When a train is moving under circumstances in which it may be overtaken by another train, the flagman must take such action as may be necessary to insure full protection. By night, or by day when the view is obscured, lighted fusees must be dropped off at proper intervals.

*584“When day signals cannot be plainly seen, owing to weather or other conditions, night signals must also be used.

“Conductors and enginemen are responsible for the protection of their trains.

“When a pusher engine is assisting a train, coupled behind the cabin car, and the flagman that protects the rear end of the train is riding in the cabin car, the requirements as to the use of fusees should be met by dropping them off between the cabin car and pusher engine on the track the train is using, and not between that track and an adjacent track.”

(The B & O rule was substantially the same).

The torpedoes mentioned are placed on the train tracks so that an oncoming train’s wheels will explode them. Their noise is like that resulting from the discharge of a shot gun. A fusee is a flare which during the night may, on a straight track under normal weather conditions, be seen for approximately one to two miles. In addition to torpedoes and fusees, a flagman is equipped with a flag with which he may signal an oncoming train during daylight and with a lantern with which he may signal such a train at night. The locomotive engineer on a straight track in clear weather could see the flagman for a distance of one mile. It will be noted that the flagman is required to go back immediately from the stopped train “a sufficient distance” to insure full protection, and what this “sufficient distance” might be, was left to his judgment.

Beginning at the turn of the century, the block signal system began to be used as a method of informing and alerting the engineers of the train to conditions ahead. There are two types of block signal systems: one the automatic block signal system which operates automatically through electrical circuits; the other, the manual block system, in which the signals are operated manually. This latter type of block system is confined to certain limited tracks at the present time. The automatic block system controls 88% of the train mileage and all of the important mileage of the Pennsylvania in Maryland.

As the present case is principally concerned with the auto*585matic block signal system, the present operation of that system should be briefly described. The line of track is divided into segments — or blocks, ranging in length from one to two miles. At the entrance to each block there is located a signal visible to an approaching train. These signals are actuated automatically through electric currents which normally flow through both rails of a block. The circuits are closed by the wheels of any train which is occupying the block and this activates a change in the signal displayed. A similar signal change occurs when the circuit is interrupted by a broken rail or by a possible failure within the signal mechanism itself. If the block is occupied by a train, the signal at the entrance to the block will be “Stop and Proceed” (Rule 291), which requires a following train approaching the signal to stop and then to proceed at a speed, not in excess of 15 miles per hour, which will permit the train to stop short of any obstruction. When the block ahead of the block in question is occupied, the signal at the entrance to the block in question will display “Approach” (Rule 285), which requires a following train to operate at curtailed speed prepared to stop at the next signal. When the block in question as well as the block immediately ahead are unoccupied, the signal will generally display “Clear” (Rule 281). On some tracks, the signals are designed to display a signal requiring a following train to curtail speed when the block governed by the signal as well as the block immediately ahead are unoccupied, but the block twice ahead of the signal is occupied.

Many of the tracks of the Pennsylvania are signaled in both directions, or have “reverse signalling,” which means that they are equipped with automatic block signals governing the operation of trains in both directions. In Maryland 28.0% of automatically signaled track falls within this category and throughout the Pennsylvania System approximately 21.9% of its track has reverse signalling. This is generally present on the high speed tracks (the authorized speed limit for passenger trains in both directions is 80 miles per hour on 78% of the track mileage), and where the tracks have reverse signalling, there is exclusive reliance on the automatic signals for protection against head-on collisions.

The testimony indicated that at the present time the auto*586matic signal system is almost infallible. On the tracks of the Pennsylvania there are located 7,566 automatic traffic signals capable of displaying two or more aspects. On every day each one of these signals will change its aspect a number of times, so that during a year several million signal changes or “operations,” as the changes are termed by the Signal Department, occur. These signals protect approximately 6,918 miles of track and 48,700,000 train miles of operation annually. The official failure reports show that there have been an average of only 5.2 failures a year, which were only potential sources of hazard to train operations, and less than a fifth of that number presented a substantial risk. The tables from the annual reports of the Interstate Commerce Commission Bureau of Safety and Service for the ten year period from 1955 to 1964 indicated that during the five year period 1960 to 1964 “false proceed” failures on the Pennsylvania System averaged 5.2 a year and during the five year period 1955 to 1959, such failures numbered 9.0 a year. During the fiscal year ending July 1964, signal figures throughout the United States on all railroads was 99.76% safe and the Pennsylvania’s average was even higher. Of the “false proceed” failures occurring during the ten year period only one occurred in Maryland and none of the “false proceeds” aspects resulted in an accident.

The design of the automatic signal system requires that if there is a failure it “fails safe,” that is, when there is any physical or electrical failure in the system, it will result in the displaying of a signal aspect which is safer than would occur in the absence of the failure.

In addition to the operation of the automatic signal system itself, all of the Pennsylvania locomotives operating in Maryland on automatically signaled tracks have cab signals as a supplementary safeguard. These cab signals consist of a miniature panel in the engineer’s cab to display in miniature the signal aspects which are displayed by the wayside signals. Their function is to display at all times the same signal aspect as was presented to the engineer upon entering the block in which he is traveling, except that any changes in track occupancy ahead occurring after he is within the block are reflected by the cab signal.

*587When the automatic block system began to be used at or about 1900, it was necessary to have additional methods of protection by human means. As early as 1905 the “fail-safe” principle was incorporated but the necessity for supplementary protection by human means continued. By 1910, however, as a result of many innovations and developments, the automatic block signal system had reached such a high degree of efficiency and reliability that many railroad men thought that supplemental protection was unnecessary. The Pennsylvania thereafter, on an expanding scale, relied exclusively on automatic signals to protect trains traveling in opposite directions at high speed on track equipped with reverse signaling.

In 1945, the Pennsylvania made an important change in Rule 99. It added the following note:

“Note — When trains are operating under automatic block signal system rules, the requirements of Rule 99, insofar as protecting against following trains is concerned, will have been complied with when full protection is afforded against trains moving at Restricted speed.”

“Restricted speed” is defined in the “Definitions” section of the Rules as follows:

“Restricted Speed — Not exceeding 15 miles per hour prepared to stop short of train, obstruction or switch not properly lined and to look out for broken rail.”

The testimony offered on behalf of the Pennsylvania indicated that prior to 1945 flagmen had in {act, for the most part, already assumed that the engineer operating trains on tracks subject to the automatic block signal system did obey the rules and was operating the oncoming train at restricted speed. This meant that the engineer was prepared and able to stop the train upon the engineer’s sight of the standing train. This logical assumption led the flagmen generally to abandon the former practice of walking various distances to the rear of the standing train to place torpedoes or fusees on the track. Rule 99 provided that the flagmen should go back with his signals “a sufficient distance to insure full protection,” the “sufficient dis*588tance” being left to the judgment of the flagman. It might be observed that the term “flagman” is not defined in the Rules and the duties imposed on the flagman were delegated to a trainman who was available to perform other services, as for example, a conductor or a fireman. In the present Pennsylvania Rule 99, the word “flagman” is not used, but the words “member of the crew” are used instead. If the oncoming train was heavy, long and fast, a flagman could possibly leave to go back a mile or more to insure “full protection” if he assumed that the engineer was not obeying the Rules and the automatic signal and was not proceeding at restricted speed. On the other hand, if the flagman assumed that the oncoming train was operating under restricted speed, it was obvious that the necessity of going back any substantial distance, with its attendant delay, discomfort and even danger, was no longer present. As the flagmen, by and large, made the logical assumption, they merely got off the stopped train and stood at or near the end of the train. The assumption by flagmen of compliance with the Rules by the engineer operating a train in automatic block territory was incorporated in the 1945 amendment to Rule 99 and the general practice regularized by that amendment. Rule 99 with the 1945 amendment continued in effect until October 18, 1964, when the Pennsylvania added the amendment which precipitated the present litigation. It may be observed that the Brotherhood did not challenge Rule 99 as amended by the 1945 amendment during the 19 year period the amendment was in effect.

The 1964 amendment added two “Notes” which superseded the Note added by the 1945 amendment. The 1964 amendment is as follows:

“Note — When trains are operating under automatic block system rules, the requirements of Rule 99 do not apply for following movements on the same track.

“Note — When trains are operating under manual block signal system rules, the requirements of Rule 99 will not apply for following movements on the same track when Rule 316 is in effect, except when required by train order or time table special instructions.”

Rule 316, referred to in the second note of the 1964 amend*589ment in regard to operation under manual block signal system rules, as also amended on October 18, 1964, applies only to designated portions of track of the Pennsylvania. It prohibits a block operator from admitting a second train to a block already occupied by one train, except that a freight train may be allowed to enter a block already occupied solely by a train within yard limits and upon display for the benefit of a following train of the modified “clear block” signal described in Rule 280.

The Brotherhood and the Pennsylvania both agree that the 1964 amendment to Rule 99 was not motivated by considerations of operating economy by reducing the number of employees required to man a train. Flagging is still required under the Rules in certain circumstances, such as against trains, on adjacent tracks in the event of an emergency stop or ontrackage governed by “Rule 317” manual block rules, which,, as an alternative to Rule 316, permits a second train to be admitted to an occupied block in manual block territory. Except within yard limits when Rule S-93 applies, flagging is still required when Rule 317 controls.

The testimony on behalf of the Pennsylvania indicated that the 1964 amendment was added because of the conviction of the company’s operating officials that, on tracks protected by automatic block signals, flagging was both unnecessary and undesirable. It was unnecessary because if the permitted assumption that the following train was moving at restricted speed was correct, then no accident could occur, in that the engineer would be operating so that he could and, indeed, would, stop, within the range of his vision. It was undesirable because, in their opinion, the existence of a flagman with purely ritualistic or perfunctory duties increased the danger of collision by providing for a “divided responsibility” whereby the engineer might believe that he has a second or last chance, as it were, and as a result would possibly be misled or lulled into a false reliance upon a warning, subject to all the uncertainties and failures inherent in the nature of the flagging operation.

The evidence in the present case is overwhelming in its support of the correctness of the position that flagging on tracks. *590governed by the automatic block signal system is both unnecessary and undesirable.

There was a long and careful study of the proposed 1964 change in Rule 99 made by the Advisory Sub-Committee to the .'System Train Rules Committee of the Pennsylvania, and also •by the System Train Rules Committee itself. This latter committee consists of the Superintendent of Passenger Transportation, the Assistant Chief Mechanical Officer, the System Engineer for Communication and Signals and the Manager of Safety. John W. Rathvon, Manager of Operating Rules, is Chairman •of both committees. He is also a member of the Operating Rules Committee of the Association of American Railroads. He has had approximately 28 years of service with the Pennsylvania, beginning as a laborer, advancing as gang foreman, engineer house foreman, train master, transportation engineer, and finally as Manager of Operating Rules. He explained the operation of the Rules in detail and testified that both committees •of the Pennsylvania unanimously approved the 1964 amendment •as giving a safer operation. He explained the type of examinations which are given to engineers and what their practice is in the operation of trains in automatic and manual block system territory.

George C. Vaughn, General Manager of the Eastern Region •of the Pennsylvania, a graduate of Lehigh University in 1930 with a degree in Civil Engineering, had been with the Pennsylvania for 35 years. He was charged in a general way with the responsibility of the safe operation of the Pennsylvania and seeing to it that the employees lived up to the Rules. He ex■plained in detail the reasons why the 1964 amendment resulted, in his opinion, in a safer rule.

In addition, T. J. Klauenberg, Chairman of the B & O Rules 'Committee, Joseph A. Bonelli, Manager of Safety of the Pennsylvania, and Hubert E. Middlekauff, Chairman of the Operating Rules Committee C & O — B & O, all men of great experience with the safe operation of railroads, gave their opinion supporting the principles underlying the 1964 amendment and their reasons for that support.

A number of other employees of the Pennsylvania and the !B & O testified in regard to the actual practice in the operating *591of trains on tracks governed by automatic and manual block signal systems, the lack of necessity of flagging and its impairment of safe operation under present conditions.

The testimony established that the principles underlying the 1964 amendment have been accepted by the vast majority of railroad officials in the United States. The 1964 amendment has a parallel in the Association of American Railroads Model Code of operating regulations which most railroads in the United States follow closely, but not exactly.

The Pennsylvania operates in 13 states and the District of Columbia. The 1964 amendment has been unchallenged in seven states, including New York and the District of Columbia, where the amended rule was in effect for almost three years when the present case was heard before the Commission. The 1964 amendment has been challenged by the Brotherhood in six states. When the case was heard, of the six Commission proceedings, two were pending, two cases had been decided in favor of the Pennsylvania (Delaware, and Maryland at the lower court level with two modifications), two were decided adversely to the Pennsylvania (New Jersey, and Pennsylvania where an appeal was pending). In the six challenged states the 1964 amendment Rule was in effect for almost three years, except in New Jersey and Pennsylvania, where modified versions of the 1945-64 Rule were then in effect as a result of orders of the Commissions in those states. It is thus seen that there was actual operating experience under the 1964 amended rule for almost three years over the entire Pennsylvania System.

The statistical data submitted by the Pennsylvania is impressive and uncontradicted. This data established that for the 12j4 year period beginning in January 1954, the Pennsylvania experienced an average of 5.3 rear-end collisions per year, based on Interstate Commerce Commission (I.C.C.) classifications, only 12% of which were of such severity as to justify investigation by the I.C.C. During this same period, the number of rear-end collisions occurring on main track protected by automatic signals averaged 2.16 a year. During the eighteen months period following the October, 1964 Rule Change, however, the Pennsylvania had only five rear-end collisions — an average of 3.33 rear-end collisions per year — but only two of the rear-end colli*592sions occurred on main track protected by automatic signals, or an average of only 1.33 a year. This is to be contrasted with an average of 2.16 a year rear-end collisions during the prior 12j4 year period, or a 38% decrease in rear-end collisions since 1964 when the October, 1964 Rule Change was made. It is interesting to note that the two accidents on main tracks which occurred subsequent to October, 1964, took place under special circumstances in which the October, 1964 Rule Change was not applied.

The operating experience of the Pennsylvania, by the uncontradicted evidence, is similar to that of other railroads in the United States. The Chesapeake & Ohio Railroad Company (C & O) put its change of Rule 99, similar to the Pennsylvania 1964 change, in effect in October, 1963. Since that time and up to the time of the hearing before the Commission in the present case, the C & O had experienced no rear-end collisions in automatic block signal territory and had received no complaints in regard to its change in Rule 99 in any of the states through which it operated. Of the approximately 5000 miles of track, 2290 miles of track are governed by the automatic block signal system, and of the 2290 miles, 950 miles are in multiple track territory consisting of two, three or four tracks.

The New York Central Railroad Company, the Pittsburgh and Lake Erie Railroad Company, The Seaboard Air Line Railroad Company and the Norfolk and Western Railroad Company have had operating experiences similar to that of the Pennsylvania since their comparable changes in Rule 99.

The testimony also establishes without contradiction that the use of the diesel locomotive has resulted in the increase of safety factors, in that the windshield of the modern diesel locomotive gives a vastly improved vision ahead as compared with the old steam engines and the diesel locomotives may be operated at very low speeds which enable the engineers to continue the movement of the train, but to stop within the range of the engineer’s vision.

Engineers as part of their training not only must have the required mechanical knowledge, but also must become intimately acquainted with the physical characteristics of the route over which they operate the trains. This includes accurate knowl*593edge of the location of the signals, curves in the tracks, location of tunnels, type of grade and the like. As one witness aptly described it, the engineer knows his route “like the back of his hand.”

In operating the engine, an engineer must keep his foot on a pedal to continue the power; if his foot is removed, the power is discontinued. In most instances there is a second employee in the cab with the engineer who repeats the signals to the engineer. In many of the Pennsylvania trains there is an “inter-com” system of radio communication between the cab of the locomotive and the caboose or other portions of the train.

The Brotherhood offered the testimony of eleven witnesses, offered copies of eleven reports of the I.C.C. involving cases of rear-end collisions and several photographs showing curves in tracks at various locations, including tunnels and a location just east of the Main Street crossing at Oakland, Maryland.

The eleven I.C.C. reports were offered into evidence through William E. Fravel, assistant to the Brotherhood’s National Legislative Representative. Mr. Fravel testified that he had made a study of a number of recent rear-end collisions which had been investigated by the I.C.C. and had selected from them the eleven reports offered in evidence. It may be inferred that these eleven reports were the ones most favorable to the Brotherhood’s position in the present case. None of the rear-end collisions occurred in Maryland. None involved a train of the B & O. One collision involved a train of the Reading Company, two involved trains of the Pennsylvania and the remaining eight reports involved trains of other railroad companies throughout the United States. The date of the earliest report (No. 3900) was October 7, 1960; the date of the latest report (No. 4005) was March 22, 1964. The I.C.C. reports indicated that in nine cases one of the assigned causes was the failure to control the speed of the following train in accordance with the signal indication; in five cases, inadequate flag protection; and, in two cases, the failure of the block signal system. It should be observed, however, that in no case did the I.C.C. assign inadequate *594flagging as a contributing cause of the accident, except where the carrier’s own regulations required flagging under the circumstances of the accident. In the case involving a train of the Reading Company (No. 3984), it is clear that flagging was not required. It was in fact performed and proved to be ineffectual. In one of the cases involving the Pennsylvania at Corliss, Pennsylvania (No. 4018), it was clear that although flagging was not required, it was in fact performed and proved to be ineffectual. In the other case involving a train of the Pennsylvania at Massilon, Ohio1 (No. 4005), flagging was performed but the engine crew of the following train did not see the flagman so that, here again, flagging was ineffectual. In several of the cases, the reports indicate that the violation of a signal indication by the following engineer was the direct cause of the accident. In none of the reports is there any basis, other than mere speculation, for believing that full compliance with the carrier’s flagging rules would have served any purpose in avoiding the accident and, indeed, in four reports there was affirmative evidence that flagging would have been ineffectual. In short, the eleven reports tend to support the reasons for the Pennsylvania’s 1964 change in the rules.

Mr. Fravel testified that the Brotherhood was “opposed to the elimination of the duties of a flagman” for the reason that “signals, even though infrequently, do fail” and that “until the time comes that all these things are perfect and we can work out all the details of human failures and so forth, the flagman is necessary as an extra measure of safety to protect the lives of the people whom we represent.”

The Brotherhood, in addition to Mr. Fravel’s evidence, introduced the testimony of seven conductors, one from the B & O and six from the Pennsylvania, who had performed flagging operations, and also the testimony of R. K. Bartlett, a locomotive engineer employed by the B & O, and of A. W. Sims, a locomotive engineer employed by the Pennsylvania.

Mr. Bartlett principally testified in regard to what he considered to be the impossibility of operating a train of the B & O at a sufficiently low speed on what is known as Seventeen-Mile Grade in the Western Maryland portion of the B *595& O System in order to stop the train within the range of the locomotive engineer’s vision. He was of the opinion that a flagman was necessary to give an added safety feature. Mr. Klauenberg and two former locomotive engineers of the B & O, Messrs. Hotchkiss and Hollen, who prior to promotion operated B & O trains over Seventeen-Mile Grade (and who even now on occasion operate such trains on Seventeen-Mile Grade) indicated that they had no difficulty in operating the trains at speeds at which they could stop within the length of their vision and the witness Hollen testified that he had personally observed Mr. Bartlett’s operation on Seventeen-Mile Grade and that Mr. Bartlett had his locomotive under control prepared to stop within his range of vision. Moreover, Mr. Bartlett testified that he had never run through an automatic signal and was then stopped by a flagman. He also testified that when he was promoted to engineer he passed an examination in regard to the physical characteristics of the track telling “exactly where all the signals were, the proper names of them and the positions of them and what kind of aspects I could get on them and telephones and switches and so forth and so on.” It may be noted that Seventeen-Mile Grade is only a small fraction of the 272 miles of track of the B & O controlled by the automatic block signal system.

Mr. Sims testified in regard to certain curves on the tracks of the Pennsylvania which restrict the locomotive engineer’s vision and that in these situations a flagman would give the engineer “a little bit more time in cases of maybe pulling a freight train.” On cross-examination, however, Mr. Sims testified that he had not put reliance “as such” in the past on flagmen in the operation of freight trains, that he liked to feel the flagman was out there but he never could be sure that he would be out and stated “I have to guard against this also. I can’t exceed the speed just for my own sake. I have to live up to those rules to the limit.” The other witnesses for the Brotherhood indicated that they had continued to go back from the standing train a sufficient distance, placed torpedoes or fusees and that this practice was an added safety factor. Witnesses for both the Pennsylvania and the B & O indicated that the flagmen would, as a general matter, merely stand at *596the end of the train and that this act was a useless one which did not contribute at all to safety and, on the contrary, was now a hazard as it caused “divided responsibility.”

The Brotherhood pointed out that in a case in 1958, involving a claim for cancellation of a 60 day suspension of George J. Lapidus for the violation of Rule 99, counsel for the Pennsylvania in its brief took the position, in opposing the cancellation, that Rule 99 required the flagman to go back immediately with flagman’s signals a sufficient distance to insure full protection, placing two torpedoes and when necessary, in addition, displaying lighted fusees, and that the failure of Mr. Lapidus to follow this procedure in order to insure full protection resulted in the collision which caused the death of the engineer in the approaching train. The claimant’s claim for cancellation was denied. R. E. Glosser, assistant to Pennsylvania’s General Manager of Transportation, explained that the position of counsel in that brief did not represent the position of the operating officials who had jurisdiction over the observance of the operating rules. It was also clear in the Lapidus case that he failed to obey Rule 99 as it was generally interpreted in 1957 in that he failed promptly to walk to the rear of his train. The position in the brief that the flagman should have done more than this was unnecessary to support the suspension and represented an argument by counsel at variance with the policy of the operating officials.

The B & O had operated under Rule 99 which provided that when a train stopped under circumstances in which it might be overtaken by another train, the flagman “must go back immediately with flagman’s signals, a sufficient distance to insure full protection, placing two torpedoes on the rail, and when necessary, displaying lighted fusees in addition.” It proposed to add to a somewhat simplified revision of Rule 99 on January 1, 1965, the following exception:

“EXCEPTION ■ — ■ When operating under Automatic Block System Rules 505 to 519, inclusive, rear end flag protection for TRAINS or ENGINES is not required against following movements on the same track, except as provided by *597Rules 98 (B), 511, 512, 514 and 515. (Does not apply to other On Track Equipment or where Special Instructions required otherwise.)”

It will be noted that this exception eliminates the necessity oí rear-end flag protection for trains or engines against following movements on the same track in automatic block signal territory, except that there will remain the obligation on a member of the crew on the stopped train to flag in emergencies, to protect adjacent tracks, etc., as provided in the B & O’s Rules 98 (B), 511, 512, 514 and 515.

The Commission, on February 18, 1966, rendered its opinion and order in the consolidated cases. After reviewing the testimony, the changes in Rule 99 involved and the contentions of the respective parties, the Commission concluded that although the electronic equipment used by the railroads was excellent and the incidence of failure minimal, failure could and did occur so that continued flagging of following vehicles in automatic block territory had a definite safety value. The Commission rejected the position of the Pennsylvania and the B & O that continued flagging for following vehicles in automatic block territory created a condition of divided responsibility and was less safe than no flagging at all. The Commission indicated that with respect to the 1945 amendment by the Pennsylvania of Rule 99 complete confusion existed and that the witnesses placed different interpretations on the effect of the footnote. It concluded that the proper provision for Rule 99 so far as the Pennsylvania was concerned was the provision of that Rule as it was prior to the 1945 amendment and that the proper provision for Rule 99 so far as the B & O was concerned was as it presently existed. The Commission’s order provided that the Pennsylvania and the B & O promulgate appropriate rules in accordance with its opinion, setting out the terms of the rule to be promulgated.

The appeal to the Circuit Court of Baltimore City followed, and, as already indicated, that court, after filing an opinion reviewing the testimony1 and staling its reasons, signed an order *598on November 21, 1966, reversing the Commission’s order on the ground that it was not supported by substantial evidence on the record considered as a whole. The lower court did in its final order provide for two modifications of Rule 99 which have been accepted by both the Pennsylvania and the B & O. In any event, since neither of the railroad companies has appealed, these modifications would not be before us, Schneider v. Menaquale, 187 Md. 202, 49 A. 2d 330 (1946), and we will not pass upon their propriety or validity. The Pennsylvania raised in its brief and in argument the point that the Commission’s order was invalid because it was a burden on interstate commerce and invaded a field pre-empted by the federal government. In view of our opinion that the Commission’s order was unsupported' by the substantial evidence on the record considered as a whole,, we do not find it necessary to consider that interesting question.

The only question in this case is whether the order of the Commission “is unsupported by substantial evidence on the record considered as a whole,” which is one of the statutory grounds-for reversal set out in Code (1957), Article 78, section 97. This-ground for the reversal of a determination by an administrative body on an issue essentially legislative in character presents-many logical and even constitutional difficulties. We are fortunate in this case, however, to have had a consideration of these difficulties and their resolution in the recent decision of the Court in State Insurance Commissioner v. National Bureau of Casualty Underwriters (No. 600, September Term, 1966) and the companion and consolidated case of State Farm Mutual Automobile Insurance Company v. State Insurance Commissioner (No. 639, September Term, 1966), 248 Md. 292, 236 A. 2d 282 (1967). In the State Insurance Commissioner cases the majority of the Court sustained the validity of the ground for reversal of the State Insurance Commissioner’s order in regard to rates set forth in Code (1957), Article 48A, section 245(2) ; i.e., if the Court found that the order or decision was “not sup*599ported by the preponderance of the evidence on consideration of the record as a whole.” The Court held that this ground for reversal by the Court did not violate the provisions of Article 8 of the Declaration of Rights of the Maryland Constitution providing for the separation of powers, in that the legislative intent was that the test was equivalent to the “weight of the evidence” or “substantial evidence” test which “does no more than permissibly require the Court to decide * * * whether a reasoning mind reasonably could have determined that the factual conclusion reached was proven by the weight of the evidence on the record as a whole.” 2

In the opinion for the Court in the State Insurance Commissioner cases Chief Judge Hammond cited with approval the meaning which the Supreme Court of the United States had given the term “substantial evidence on the whole record” in Universal Camera Corp. v. National Labor Relations Board, 340 U. S. 474, 488, 71 S. Ct. 456, 464, 95 L. Ed. 456, 467 (1951), in which Mr. Justice Frankfurter, for the Supreme Court, stated: “The substantiality of evidence must take into account -whatever in the record fairly detracts from its weight. This is clearly the significance of the requirement * * * that courts consider the whole record.” Also cited with approval was the language of the Court in Board v. Oak Hill Farms, 232 Md. 274, 282-83, 192 A. 2d 761, 766 (1963) in commenting on Universal Camera as follows :

“The holding of the Supreme Court that in determining substantiality of evidence there must be taken into account whatever in the record fairly detracts from its weight emphasizes the observation that there *600is a thin line between a test based on substantial evidence and one based on weight of evidence. Some statutes or rules have directed the courts to pass on the weight of evidence in administrative appeals, and courts have followed the direction in a number of decisions, federal and state, without feeling that they were exercising an impermissible non judicial function in asserting and exercising the power to ascertain whether agency findings are in error because clearly erroneous or against the weight of the evidence or the clear or the manifest weight of the evidence. 2 Am. Jur. 2d, Administrative Law, Secs. 661, 681, 682. See also 4 Davis, Administrative Lam, Secs. 29.08, 29.09, and Jaffe, Judicial Review: Question of Pact, 69 Harvard L, Rev. 1020, 1052, et seqj” (Emphasis in original).

In footnote 6 in Oak Hill Parms, five of the cases cited were appeals from legislative action by the administrative body and these cases are listed in footnote 3 in the majority opinion in the State Insurance Commissioner cases.

As the Court has cited with approval Professor Jaffe’s article “Judicial Review: Question of Fact," in 69 Harvard L. Rev. 1020 (1956), some of Professor Jaffe’s observations are helpful. He stated:

“Judicial review is designed, as has been said, to make sure that there is record evidence which provides a rational or logical basis for the finding and for the consequent presumption that the finding was in fact the product of reasoning from evidence. This must mean evidence in the case and in the context of the case. To abstract out of a case that part of the evidence which can be made to support a conclusion is to imagine an abstract case, a case that was never tried. A conclusion based on such abstracted evidence may be ‘rational’, but it is not a rational decision of the case which was in fact tried. Evidence which may be logically substantial in isolation may lose its logical relevance, even its claim to credibility, in contest with other evidence.” (Emphasis in original). Id. 1027.

*601In Gooding v. Willard, 209 F. 2d 913, 916 (1954), the United States Court of Appeals for the Second Circuit in considering the permissible extent of judicial review under the federal Administrative Procedure Act, 5 U.S.C.A., section 1001, et seq., Act of June 11, 1946, stated, in an opinion by Chief Judge Chase, for that Court, in commenting on the effect of the decision of the Supreme Court of the United States in Universal Camera Corporation:

“As is pointed out in the opinion in the last mentioned case, ‘substantial evidence’ means more than evidence which, considered by itself alone, would be sufficiently persuasive to induce the trier of fact to give it the credence and weight essential to support findings. It must have those characteristics to such an extent that in the setting made by the entire record the trier may reasonably find in accordance with it after giving due consideration to whatever else is shown both in opposition or in accord.”

In view of the exhaustive review of the authorities in the opinion for the Court in the State Insurance Commissioner cases, it w-ould serve no useful purpose to prolong this opinion with any additional review of the authorities.

In considering, therefore, whether the lower court was correct in finding that the order of the Commission was not supported by substantial evidence considering the record as a whole, we must take into account whatever in the record fairly detracts from the weight of the evidence offered by the Brotherhood to support its position that the Pennsylvania amendment of 1964 to Rule 99 and the B & O’s proposed amendment to Rule 99 were unsafe and that a safe operation in automatic block signal territory as covered by the 1964 amendment and the proposed amendment did require that flagging be continued. We may not consider the evidence offered by the Brotherhood in isolation to determine whether or not it was “substantial”, but must consider whether that evidence “in the case” and “in the context of the case,” as Professor Jaffe indicates, was substantial. Assuming, arguendo, that the evidence offered by the Brotherhood might be “substantial” in isolation, it is not, in *602our opinion, “substantial” in context with the other evidence in the case.

As we have seen, the uncontradicted evidence established that the perfection of the automatic block system has proceeded to such a point that the possibility of error is most remote. Indeed, it has been so perfected that as far as the movement of fast trains on tracks equipped with reverse signaling is concerned, there is exclusive reliance upon the automatic block signal system. The automatic block system had really reached sufficient perfection to justify an elimination of flagging in 1945, but this unproductive practice — a vestigial survival from a less sophisticated period of train operation — was continued by the Pennsylvania, but with the assumption by the flagman that the following train was operated by the locomotive engineer at restricted speed, that is, at a speed “not exceeding 15 MPH prepared to stop short of trains, obstructions or switch not properly lined and to look out for broken rails.” The evidence is clear that regardless of what certain flagmen did after the 1945 amendment— and some flagmen did continue the old pre-1945 practice, as witnesses for the Brotherhood testified — the old practice of going back from the end of the stopped train, placing torpedoes and fusees, etc., was unnecessary and did not, in fact, contribute to a safe operation. On the contrary, the evidence is really conclusive that the old practice had become a detriment to safety as it did indeed result in a “divided responsibility” giving the locomotive engineers in whom the safe operation of the train in automatic block system territory was placed “a second chance,” as it were, to be advised of a standing train and thus weaken to some extent the degree of care and alertness which the locomotive engineer knew he must have. The instances given by witnesses for the Brotherhood in which a flagman might help in the avoidance or mitigation of the degree of injury from a rear-end collision would be obviated if the locomotive engineer obeyed the rules and stopped the train until he obtained further information.

Most importantly, the eleven cases before the I.C.C. offered by the Brotherhood do not support its position when carefully analyzed and, indeed, as we have indicated, rather support the policy for which the Pennsylvania and the B & O contended.

*603It would be difficult to find more impressive evidence indicating that the position of the Pennsylvania and B & O is correct than the uncontradicted figures showing that since the 1964 amendment the Pennsylvania has had fewer rear-end collisions than before and that the C & O has had wo rear-end collisions since its new policy went into effect in 1963. Moreover, this has been the experience of railroads generally throughout the United States and represents the policy set forth in the Model Rule Book promulgated by the Association of American Railroads.

The officials of the railroads having a vast experience with safety considerations, without exception, supported by their testimony the position of the Pennsylvania and B & O in this case. The Brotherhood did not produce a single “safety expert” to support its position.

The testimony of the two locomotive engineers and of the conductors -who had acted as flagmen from time to time (whose testimony was substantially challenged by the testimony of other witnesses), the photographs of various curves, the records of the eleven I.C.C. cases which did not support the position of the Brotherhood, cannot, in our opinion, mount up to substantial evidence on the record as a whole in the context of the overwhelming evidence contrary to the position of the Brotherhood offered by the Pennsylvania and the B & O in this case.

The Pennsylvania and B & O emphasized in their briefs and in the argument before us that the Brotherhood had the burden of proof in establishing its position that the 1964 amendment of the Pennsylvania and the proposed amendment of the B & O were unsafe and that it clearly failed to meet this burden of proof. In our opinion the question of burden of proof was not really involved, in that if there had been substantial evidence -on the record as a whole offered in support of the position of the Brotherhood, and the Commission had accepted such substantial evidence as establishing the Brotherhood’s position, the Brotherhood would, ipso facto, have met the burden of proof. The only question in the case did not involve “burden of proof,” but only whether there was “substantial evidence on the record -as a whole” to support the order of the Commission.

We are aware that the Supreme Court of New Jersey, in the *604recent case of In re Complaint of Brotherhood of Railroad Trainmen, 49 N. J. 174, 229 A. 2d 505 (decided April 24, 1967), has ruled in favor of the Brotherhood in regard to the amendment of October 18, 1964, to Rule 99 by the Pennsylvania in the State of New Jersey. This decision may readily be distinguished from the case at bar on two grounds: (1) the statutory test for judicial review in New Jersey is different from the test in Maryland, and (2) the factual data in support of the position of the Pennsylvania as recited in the opinion in the New Jersey case appears to be far less than the factual data introduced by it in the case at bar. In New Jersey, the statutory test on review by the provisions of R.S. 48: 2-46 N.J.S.A. is that the court on review of a decision of the Board of Public Utility Commissioners may only reverse “when it clearly appears that there was no evidence before the board to support the same reasonably or that the same was without the jurisdiction of the board.” (Emphasis supplied). In Maryland, as we have indicated, the test is whether the Commission’s order is supported by “substantial evidence on the record as a whole,” which is different from the New Jersey test as set forth in the New Jersey statute and construed by the Supreme Court of New Jersey. Cf. Application of Greenville Bus Co., 17 N. J. 131, 137-38, 110 A. 2d 122, 125 (1954), in which the opinion was written for the Supreme Court of New Jersey by Mr. Justice Jacobs who also wrote the opinion for that Court in In Re Complaint of Brotherhood of Railroad Trainmen, supra. Then too, in the opinion in In Re Complaint of Brotherhood of Railroad Trainmen, it was stated in regard to the Pennsylvania’s position that the elimination of flagging requirements would increase rather than decrease the safety of operations; “And while several railroad employees expressed their opinions to that effect, they did not produce any significant factual documentations or any independent expert studies in support.” (49 N. J. at 180, 229 A. 2d at 508). As has been already indicated, the situation in the case at bar is far different, as the Pennsylvania produced definitive data based on actual operating experience not only by it, but also by the C & O and railroads generally throughout the United States. There are also other factual differences in the case at bar from the factual situation recited in *605the opinion in In Re Complaint of Brotherhood of Railroad Trainmen.

Since the lower court properly found, in our opinion, that the Commission’s order was not supported by substantial evidence on the record as a whole, we will affirm its order reversing the Commission.

Order affirmed, the appellant to pay the costs.

filed the following dissenting opinion in which Hammond, C. J., and Masbuky, J., concurred.

I am unable to agree with the majority of this Court because, in my opinion, the lower court erroneously substituted its judgment for that of the Public Service Commission.

The evidence presented before the Commission was both exhaustive and detailed, as the opinion of this Court clearly indicates, and the members of the Commission took advantage of these hearings to closely question the witnesses, observe their demeanor and weigh both the relevancy, credibility and objectivity of their testimony. The evidence propounded by the railroads attested to the high degree of reliability and increased safety record produced by the Automatic Block Signal System, with wffiich the Brotherhood in large measure agreed. Witnesses for the railroad were of the belief that the continuance of flagging and other manual devices did not add to the protection of trains, and in fact contributed to unsafe practices insofar as train engineers would not rely as much on automatic signals if they expected the added precaution of a flagman would provide sufficient warning.

Opposed to this evidence were the testimony and exhibits of the Brotherhood, which recognized the efficiency of the Automatic Block Signal System, but indicated that it was not infallible, and on numerous occasions had failed to properly warn approaching trains of obstructions on the track ahead. In addition, it produced evidence of many curved stretches of track in Maryland which posed a threat to engineers in that, even if they were traveling a block under restricted conditions, they would be unable to see far enough ahead to determine the location of trains and to stop in time. In such cases, a flagman could walk *606far enough behind the lead train to provide a warning and drop torpedoes and fusees along the track to warn approaching equipment. There was also testimony that the use of a flagman alleviated hazardous conditions encountered at night in mountainous terrain when trains must proceed at an exceptionally slow speed while descending steep grades. In general, the Brotherhood argued that the flagman was the added safety feature providing protection in the event of failure of the Automatic Block Signal System.

After seven days of hearings, the Commission issued its Order, which stated in part:

“It is our opinion that the flagging of the following vehicles in automatic block territory has a definite safety value. While we agree that the electronic equipment used by the railroad is excellent and the evidence of failure minimal, the fact remains that failure can and does occur. Human error is even more likely to occur. We believe from all the evidence that there are times governed by weather and the terrain when a failure to provide adequate flagging of following vehicles could be characterized as negligence per se. We are not impressed with the strongly urged argument on the part of the Companies that the requirement for adequate flagging creates a condition of divided responsibility and therefore is less safe than no flagging at all.”

This Order was set aside by Judge Harris in the Circuit Court of Baltimore City on the basis that the Brotherhood did not meet the required standard of proof.

An order of the Public Service Commission will not be set aside on appeal unless, inter alia, “such order is unsupported by substantial evidence on the record considered as a whole.” Md. Code Ann., Art. 78 § 97 (1965 Repl. Vol.). Although the terms “substantial evidence” and “record as a whole” appear unambiguous, their application to precise sets of facts has frequently resulted in confusing and indefinite guidelines for courts to follow. In an early attempt at breathing substance into the “substantial evidence” test, Justice Frankfurter in Universal Camera Corp. v. NLRB, 340 U. S. 474, 71 S. Ct. 456 (1951), recognized the problem while interpreting the scope of review section of the Taft-Hartley Act, now 29 U.S.C. § 160(f) :

*607“Congress has merely made it clear that a reviewing court is not barred from setting aside a Board decision when it cannot conscientiously find that the evidence supporting that decision is substantial, when viewed in the light that the record in its entirety furnishes, including the body of evidence opposed to the Board’s view.

Since the precise way in which courts interfere with agency findings cannot be imprisoned within any form of words, new formulas attempting to rephrase the old are not likely to be more helpful than the old. There are no talismanic words that can avoid the process of judgment. The difficulty is that we cannot escape, in relation to this problem, the use of undefined defining terms.” 340 U. S. at 488, 489.

Justice Frankfurter’s statement underlines the difficulty of clearly articulating the extent of review, without creating further semantic confusion over “substantial evidence.” However, when he endeavored to clear the water by paraphrasing the standard of judicial review he perhaps unintentionally created a few waves which have yet to settle.

“The substantiality of evidence must take into account whatever in the record fairly detracts from its weight.” 340 U. S. at 488.

It should be noticed that the majority opinion in the instant case virtually adopts the language employed by Justice Frankfurter in Universal Camera, supra, 248 Md. at 601, 238 A. 2d at 528. Further, the majority opinion quotes Judge Hammond (now Chief Judge) in Board v. Oak Hill Farms, 232 Md. 274, 282, 192 A. 2d 761, 766 (1963), that the Supreme Court’s use of the term weight in Universal Camera “emphasizes the observation that there is a thin line between a test based on substantial evidence and one based on weight of evidence.” It is my belief that the court below crossed that thin line, and, rather than determine whether the total record permitted the findings reached by the Commission, indulged in the non-judicial activity of weighing the evidence.

*608It is understandable that the language of Universal Camera would lead some courts into the trap of confusing “substantial evidence” with “weight of the evidence.” Professor Jaffe, to whom reference was made in the majority opinion, recognized that some jurists would get caught when he said:

“But no doubt the ‘whole record’ test may with certain judges become a ‘weight of evidence’ test. There are a number of reasons for this. First it may be misunderstood. The test teaches us that the probability of an inference, the logical evidentiary support of an inference, is determined not by any one bit of evidence but by the whole record. And in the opinion in Universal Camera the Court says: ‘The substantiality of evidence must take into account whatever in the record fairly detracts from its weight.’ Thus if the value or ‘weight’ of the evidence in support of an inference must be assayed in order to determine its substantiality, it is understandable that some judges may feel that they are called upon to ‘weigh the evidence.’ The judge must, therefore, be vigilantly aware of what he is doing and stop short of such an exercise of power. Once he had determined that there is a reasonable probability of the fact found by the agency he is functus officio. It matters not that the whole record would support a contrary inference or that in the opinion of the court the contrary inference is more probable or even much more probable. The court may not weigh the worth of competing inferences. In the words of Mr. Justice Frankfurter a court may not ‘displace the Board’s choice between two fairly conflicting views, even though the court would justifiably have made a different choice had the matter been before it de novo! ” Jaffe, Judicial Review. Question of Fact, 69 Harv. L. Rev. 1020, 1028 (1956).

Judge Harris’s memorandum opinion reveals that, although he understood that the court’s power of review was limited to finding “substantial evidence on the record considered as a *609whole,” he nevertheless indulged in the weighing of the evidence. For example, in discussing the interpretation placed upon the I.C.C. railroad accident reports submitted by the Brotherhood, the lower court stated:

“* * *, nor do they show convincingly that the absence of such rear-end flagging would increase the number of rear-end collisions. Certainly there is no preponderance of evidence to that effect. In fact, there is some undisputed evidence of some improved safety by Pennsylvania since the change of the rule.” (Emphasis supplied.)

Further on, the opinion states :

“After a careful consideration of all the evidence, however, it is the opinion of this Court that the evidence produced by the Brotherhood, as the Complainant before the Commission, did not meet the required burden of proof. A close scrutiny of the record indicates that the preponderance of evidence, on the contrary, shows * * (Emphasis supplied.)

The lower court apparently was of the impression that the Brotherhood held the burden of producing sufficient evidence to convince the Commission (and the Court) that the elimination of flagmen’s duties would have a damaging effect on the safety of railroading. The fact is that the accident reports did indicate situations where the presence of a flagman would have presented accidents, and tlie lower court exceeded its power of review when it determined that the reports did not carry sufficient weight to “convince” the Commission. Misapplication of the court’s power of review is also evident from its conclusions that the railroads were able to rebut the Brotherhood’s argument by a “preponderance of the evidence.” The law is settled that the reviewing court may not base its decision on the manner in which the scales tip.

Chief Judge Hammond in the recent case of State Insurance Commissioner v. National Bureau of Casualty Underwriters, 248 Md. 292, 236 A. 2d 282, 291 (1967) stated the corn*610mon denominator that must apply to any valid test of judicial review:

“Whichever of the recognized tests the court uses— substantiality of the evidence on the record as a whole, clearly erroneous, fairly debatable or against the weight or preponderance of the evidence on the entire record —its appraisal or evaluation must be of the agency’s fact-finding results and not an independent original estimate of or decision on the evidence. The required process is difficult to precisely articulate but it is plain that it requires restrained and disciplined judicial judgment so as not to interfere with the agency’s factual conclusions under any of the tests, all of which are similar. There are differences but they are slight and under any of the standards the judicial review essentially should be limited to whether a reasoning mind reasonably could have reached the factual conclusion the agency reached. This need not and must not be either judicial fact-finding or a substitution of judicial judgment for agency judgment.” (Emphasis supplied.)

See also Baltimore Tank Lines v. Public Service Commission, 215 Md. 125, 137 A. 2d 187 (1955); Howard Sports Daily v. Weller, 179 Md. 355, 18 A. 2d 210 (1941) ; Stern, Review of Findings of Administrators, Judges and Juries: A Comparative Analysis, 58 Harv. L. Rev. 70, 89 (1944).

It is therefore immaterial that this Court viewing all of the evidence may have reached a different conclusion than that of the Commission, or hearing the matter de novo may have arrived at a contrary result; that is not the test. The question is, viewing the record as a whole was there substantial evidence of rational probative force to support the Commission’s finding? In my opinion the evidence presented by the Brotherhood of Railroad Trainmen, considered in context with the record as a whole, constitutes such evidence.

I would reverse the order of the lower court. Chief Judge Hammond and Judge Marbury have authorized me to say that they concur in this dissent.