and specific causes. These data should then be analyzed to determine whether the problem deserves a high priority and, if so, the kinds of research or action that may be required. Such a national data collection and analysis system can help point to potential problems, assist in establishing programs, and monitor the effectiveness of safety programs.

Certainly the Federal Railroad Safety Act has played a role in helping make railroads a safe mode of freight transportation, a role we would not like to see misdirected because of common misconceptions about railroad safety. Misconceptions, for example, concerning track-related accidents, the carriage of hazardous materials, the importance of grade crossing safety, human factors, and so on. I think it must be emphasized that no new regulation or legislation is needed. What is needed is careful analysis of the total situation and of reports provided to the Federal Railroad Administration in order that funds and activities can be directed in ways that will provide the greatest payout in terms of improved safety. And it must always be remembered that there are trade-offs involved. Some safety requirements-however well-intentioned-might actually reduce safety if they result in greatly increased rail freight costs. For example, it might be theoretically possible to design a railroad tank car that would be absolutely impervious to rupture, but if the cost of using the car resulted in a diversion of hazardous traffic to the highways the result would be more, not fewer, fatalities from hazardous materials transportation.

Railroads represent by far the safest form of transportation of hazardous materials. That is one of the main reasons that shippers choose railroads for 70 percent of the business.

The fact is highway tractor-trailer trucks, which transport only 30 percent of all hazardous materials, are involved in 90 percent of all hazardous materials accidents. According to federal statistics and the National Transportation Safety Board, trucks are involved in 70 percent of the fatalities and 80 percent of the total injuries of all such accidents.

I would also like to caution you that perceptions can be misleading. A superficial reading of train accident statistics might lead one to believe that this would be a fruitful area for new safety regulation. Noting that track accounts for more derailments than any other single cause, it might seem logical to assume that significant expenditures of money for track and more stringent regulation of track quality would produce great safety benefits.

Yet, if all train accidents in 1978-the last year for which we have full statisticshad ben eliminated, it would have reduced the fatalities by only 3.,7 percent and injuries by only 2.3 percent. If track had been eliminated as a cause of accidents, it would have reduced fatalities from train accidents by only 5 percent.

I do not mean to leave the impression that railroads are not concerned about such accidents. They are greatly concerned because their real impact is economic.

Poor track, which requires that trains be operated at low speeds to minimize the possibility of serious accidents is an economic problem. For this reason, railroads are stretching scarce resources to improve the railroad rights-of-way.

In 1979, spending for capital improvements and maintenance reached a new allhigh high of $14 billion. This followed record capital spending of $12 billion in 1978, $9 billion in 1977, and $8 billion in 1976. And I should point out that these records were achieved in the face of industry earnings which remained at inadequate levels. In 1979, the railroad industry had an income of only $922 million and a rate of return on investment of a mere 2.7 percent, one of the lowest of all major industries in this nation.

In 1979, $4.6 billion was invested in track, roadbed, and facilities, compared to the previous record level of $4 billion in 1978. Another $6 billion was spent last year for equipment maintenance, up from the $5.2 billion price tag in 1978.

Actions in the past, such as slowdowns, served to stabilize the incidence of costly derailments, but not, of course, minor derailments. However, preliminary figures for 1979 seems to indicate that railroad spending in recent years is paying off. These figures show a decline in all derailments of 18 percent during the first nine months of the year.

Perceptions can also be deceiving with regard to hazardous materials. There is no denying that an accident in which liquefied petroleum gas explodes is spectacular and may rightly be classed as a disaster. Nor would I argue that any accident involving release of dangerous products or the burning of volatile substances is not a serious matter.

But I must emphasize that it would be grossly misleading to characterize the movement of these materials by rail as a clear and present danger to population centers. In 1979, there was not a single death due to a hazardous materials accident.

Out of more than 1.1 million carloads of hazardous materials shipped by rail last year, there were only 1122 instances of hazardous materials releases from derailments.

This is not to say that the railroads are complacent about hazardous materials safety. Although the cars primarily used for the transport of such products as Liquefied Petroleum Gas (112 and 114) are almost all owned by shippers, the railroad industry is cooperating fully in the retrofit program authorized by FRA. Shelf couplers have been installed on all such cars and the complete three-part retrofit of coupler, head shield, and insulation is scheduled to be completed by the end of this year.

The AAR is currently involved in a program, jointly funded by FRA, to test and study the smaller 105 tank cars to determine if any added protection is necessary to bring them to the level of safe operation we have now in the retrofitted 112 and 114 cars. Because the results should be available in mid-1981, it seems inappropriate at this time to pursue legislative or regulatory solutions.

Where then are the areas upon which attention can be focused fruitfully? I think it must be said that there is no single area where the cause of accidents is so marked as to provide an easy "handle" by which to bring about a noticeable reduction.

Employee safety provides an illustration of this.

Some 99.8 percent of all reported employee casualties are injuries, and only 50 percent of these involve lost work days. Since 1967, there has been a 22 percent increase, however, in injuries per million employee hours, and the AAR has initiated an extensive human factors research program. But just as there are no easy answers to derailment accidents, there are no easy answers to employee safety. The causes? Stumbled, slipped, and tripped caused 25 percent of all employee injuries in 1978; use of tools, 10 percent; handling materials, 8 percent; handling ties, 6 percent; striking or struck, 6 percent; getting on or off, 5 percent; coupling, 4 percent; operating switches, 4 percent; and so on.

The Federal Railroad Administration and the Association of American Railroads have been engaged in research on safety since the expansion of their program in 1970.

While other factors are responsible for a greater portion of fatalities and injuries, derailments are responsible for the most potentially dangerous of railroad accidents the release of hazardous materials. It was, therefore, concluded by both organizations that attention to derailment prevention was crucial and that many of the aspects of derailment prevention required better materials, or better inspection methods, or better operating practices, or appropriate combinations of these. Major research activities undertaken by both organizations in some cases separately and in some cases in cooperative, jointly managed programs include the following.

TANK CAR SAFETY RESEARCH

In 1970 both institutions began the study of the processes in a derailment that lead to the violent rupture of tank cars carrying flammable compressed gas. It was clear, after much study, that a release of product, a subsequent fire, and processes that had to do with the overheating of the tank car shell, lead to the ultimate release of the product and the rocketing of the ends of the ruptured tank car. Studies starting in 1971 were directed at means of controlling this sequence. These culminated in 1977 in the issuance of HM-144 and retrofitting of the 112 tank car fleet with head protection and with high-temperature insulation. This project will be completed by the end of 1980. By that time, there should be a virtual end to the violent rupture of these tank cars. Information learned in these programs is clearly applicable to other kinds of problems, particularly head puncture during derailment, and may be found to be applicable to other cars carrying hazardous materials. The reason for attention to the containment of a product in a tank car that may be involved in a derailment is that derailment prevention is much more difficult. Examination of mainline derailments reveals that they are distributed over more than 150 different causes, most of which account for less than one or two percent of all mainline derailments. Notwithstanding_this problem, serious work has been in progress for several years in such areas of improved rail materials, improved rail inspection, improved track structures, and improved dynamic response. These programs are drawing on advanced technology and the university community as well as on the effective work being performed by the manufacturing industry that supports and supplies the railroads.

2 Preliminary 1979 data.

TRACK-TRAIN DYNAMICS

Among the most important programs in progress in this field of safety is TrackTrain Dynamics. This program began in 1972 and is now in its seventh year. It is a program supported by the FRA, individual railroads, the railroad supply industry, the Transportation Development Agency of Canada, and the Association of American Railroads. Through this program, there have been developed a set of guidelines for train handling that describe precisely to a locomotive engineer the optimum manner of maintaining his train in a stable, dynamic condition under a wide variety of terrain, climatic, and consist conditions. These guidelines have been widely distributed throughout the railroad industry and constitute a new basis for revision of individual railroads' train handling guidelines. They have been very extensively adopted and utilized and have eliminated a large number of heretofore unexplained derailments.

In the Track-Train Dynamics program a large number of dynamic, analytical models have been developed. These models have established the dynamic response of cars and equipment. They have been used to assess probable behavior of new equipment and have resulted in some extensive changes. They have been used to analyze derailments and have taken out of contention some views as to what might have caused the derailment. They have been used to check train handling practice and to refine them. They have been used to establish guidelines for track maintenance by making possible studies in ranges where operations are not feasible because of potential risk of accident. In short, these analytical models have had wide application and continue to be important in defining the basis for improved practices within the industry. Under Track-Train Dynamics a number of efforts are in progress to provide the locomotive engineer with more information about how he can handle a train with more explicit information on forces in the train than is currently available. Under the program, guidelines for the design of new classes of freight cars, more dynamically stable, are being issued in anticipation of their extensive use. In short, this program has provided the FRA and the railroad industry with deep insights into opportunities for improvement and the provided specific needs of translating those findings into action.

FACILITY FOR ACCELERATED SERVICE TESTING

Another major cooperative effort is the Facility for Accelerated Service Testing, also known as FAST, operated at the Transportation Test Center in Pueblo, Colorado. In this program, the FRA provides the operation of a train that circles a 4.8 mile loop many times a night so as to expose the track structure and the equipment to an operating environment between five and ten times the intensity level of conventional railroad operations. By having a single consist operate over the track, variables can be isolated and comparative studies made between equipment or track components with much more precision than is possible in revenue service with a variety of trains and cars that must be moved over the system. Deficiencies in some equipment have shown up before that equipment was introduced in service. Opportunities for choosing advanced technology have emerged and are being adopted by many railroads in their procurement decisions.

LOCOMOTIVE CONTROL COMPARTMENT COMMITTEE

The Locomotive Control Compartment Committee charged with the responsibility for improving safety in locomotive cabs was organized in 1970. This committee of labor, FRA, and management representatives has the responsibility for studying the safety issues in the locomotive cab and recommending measures to alleviate these problems. Early in the work of this committee, an analysis was made of the causes of injuries in cabs. The location of the door handle was found to be unsatisfactory thus leading to serious hand injuries, including the loss of fingers from occupants of the locomotive cab. A redesign was institutes which is now in universal application in new equipment and being retrofitted on many older pieces of equipment that has totally eliminated this problem. Other equally important matters have been dealt with by the Locomotive Control Compartment Committee, including the development of a performance requirement for an advanced locomotive cab seat that provides more support and more safety to the locomotive engineer. This comparative program has not involved substantial expenditures by the FRA or the AAR, but it has made it possible to bring into cohesive form the concerns of labor, the interests of the industry, and the responsibilities of government. It has been effective, and it has contribution to significant improvements in safety; and its work is continuing.

These examples indicate the critical importance of new knowledge in resolving some of the safety problems confronted by the industry. The analysis of information about safety travels, however, that there are no problems of such a character that there are no problems of such a character that their solution will make a revolutionary change in railroad safety. We recently completed an analysis of employee casualties by job classification, nature of injury, and what the employee was doing when the injury was sustained. The purpose was to assist in the development of safely program priortities based on accident frequently and severity. This analysis revealed that the complete elimination of any one of the highest-ranked safety problems, measured in terms of number of injuries or number of lost work days, would improve overall rail safety by only a few percentage points. This came as no real surprise. The profit incentive to reduce accident rates has already resulted in the correction and control of the major safety problems in the industry. The implication is clear, and the record supports that implication. There is no systematic, clear definition of narrow points of focus to alleviate fatality and injury problems. Accordingly, it cannot be expected that legislation or regulation will have a profound impact on safety. All that can be hoped for is fine tuning of the system. That fine tuning requires new information, new knowledge, new insights gained from research. It requires an aggressive but flexible environment for the application of that new knowledge. Legislation and regulation tend to freeze a system and deny the opportunity for flexible response as opportunities arise. Heavy-handed enforcement of regulations that can do very little for safety tend to deflect interest and attention away from new concepts that may emerge from research. Thus, there is much to be said for a more flexible approach toward safety provided there is adequate investment in research and an adequate commitment to the application of research by the railroad industry. That commitment exists, the research resources are modest but are being used effectively.

FRA needs continued funding support as does the AAR. Effective railroad safety is being pursued. Attempts are being made to find out by research methods which practices by what companies are important to improved safety. Once that information is developed, it will be made widely known to the industry; and others will be urged to adopt it. Thus, support of FRA's safety program is an important safety research program and in the long run is an essential element of improving safety by the railroad industry. That research program should be accompanied by an awareness of the flexible necessary to ensure effective utilization of advanced ideas. Another area of great concern, but which also has no simple solution, is the death toll at grade crossings. It can be predicted that about 1,000 persons will die at grade crossings this year. This number will represent substantial improvement over past years but it is the largest single cause of rail-related fatalities and as Appendix A shows-overshadows all others by a considerable margin.

Between 1967 and 1978, the actual exposure at grade crossings increased 40 percent. This is due mainly to an increase in motor vehicle miles. U.S. railroads are currently spending $100 million a year just for the maintenance of active traffic control devices. This is more that one-third of the total U.S. railroad industry income in 1977. As a result of this investment and cooperative programs funded by the FRA, the Federal Highway Administration, and other concerned parties, grade crossing fatalities have dropped 52 percent since 1967. (See Appendix B for detailed statistics.)

As I have said, however, there is no simple answer to reducing grade crossing deaths-unless government at various levels would undertake the massive expenditures that would be involved in physically separating all highways from the tracks they now cross.

The overwhelming majority of grade crossing deaths occur as a result of error on the part of an automobile driver. Difficult as it may be to believe, there are numerous cases in which drivers not only drove around physical barriers but actually drove into the side of a moving train.

Quite obviously, then, education rather than regulation is the greatest need. And that is what the railroads are undertaking-in cooperation with the National Safety Council and state governments-in their current programs.

While I have touched briefly on a number of the subjects in which the Subcommittee has expressed an interest, I would like to note again that my full statement as submitted for the record contains more detailed discussions of these and other issues.

EMPLOYEE SAFETY PROGRAMS

Perhaps the greatest impact on safety in the rail industry will come from a comprehensive study of the effects of human behavior on railroad safety which began three years ago in the Safety and Special Services Division of AAR.

Recognizing the limitations of mechanical solutions to problems which are largely the result of human behavior and systems errors, the industry began exploring better approaches to understanding and managing the human element-the prime contributor in over 38 percent of accidents which are occurring.

A number of industry efforts are focusing on different phases of this problem: (1) Employee Assistance Programs-Many railroads currently have or are installing programs to help employees cope with problems such as alcohol and drug abuse. Seven roads recently participated in a government sponsored study of program effectiveness which brought out the contributions these programs are making to employee safety.

(2) Study of Individual Safety Programs-Railroad safety programs generally consist of training and orientation, accident investigation and analysis, inspections and audits, personal protective equipment, work place environment monitoring, and employee participation and motivation. To determine which safety programs are effective in preventing injuries, the AAR is participating in a government survey which will identify effective programs. Results will be furnished to the industry for application on individual properties.

(3) Training-According to a recent government study, railroads now spend $40 million per year to prepare employees to work safely in the railroad environment. More effective safety training methods, which will address specific needs of both managers and employees, are currently being evaluated.

Many different methods are currently being used by the industry to provide training for employees. It is recognized that much of the "training" is a one-time activity and that attempts to measure quality in terms of hours spent in training can be very misleading. Current research is attempting to establish both the best methods which can be used to train employees to safely perform duties and productive methods which can be used to motivate them to use the procedures taught.

EFFECTIVENESS OF FEDERAL SAFETY REGULATIONS

We in the railroad industry have gone on record as supporting the findings by the Office of Technology in its report entitled "An Evaluation of Railroad Safety" that the regulatory activities of the Federal government have had no apparent effect on the railroad accident rate. But it is also clear that this accident rate does not portray a situation which is out of hand. The most important consideration is that the regulatory process is not the solution to improved railroad safety.

It is true that many of FRA's safety regulations have been counterproductive, but one of the more encouraging developments during the last two years has been the FRA's comprehensive review of the railroad safety regulations. Shortly after the promulgation of the 1970 Rail Safety Act, the FRA issued numerous regulations. Many were not related to safety. Many required premature maintenance and thus diverted limited rail resources from other essential programs. Most of the fines assessed by the FRA were issued for violations of regulations clearly unrelated to increased safety. In 1978, in response to President Carter's Executive Order 12044, the FRA began a comprehensive review of the old regulations. Numerous hearings have taken place, and we hope that the new regulations, which we expect to be issued during the next several months, will be more cost effective than the current regulations.

Preliminary cost studies indicate that the new freight car regulations will avoid the unnecessary expenditure of several million dollars annually. Similarly, it is anticipated that the new locomotive inspection standards will eliminate extremely expensive requirements which have not been necessary from a safety standpoint. We will continue to work with and encourage the FRA to revise the old regulations and develop new standards which truly address safety and which will not present unreasonable burdens on railroading.

We also anticipate that with a more analytical approach to the task of ordering priorities in the area of rail safety, FRA will not be encumbered with obligations to implement nonproductive programs which deflect resources from beneficial safety programs.

STATE INVOLVEMENT IN THE FEDERAL INSPECTION PROGRAM

Currently, the Federal program does not preempt State regulatory enforcement authority. There is a rather detailed regulatory scheme under which states may