In the first of these tables, of roads having NO passengers killed in train accidents, is represented a mileage and traffic greater than the combined mileage and traffic of Germany, France, Russia in Europe, and Italy; in the second is represented a mileage and traffic greater than that of the combined railways of Austria, Hungary and Spain; and in the third is represented a mileage and traffic greater than that of the combined railways of Great Britain, Germany, France, Russia in Europe, Austria, Norway and Sweden.

The true relation of railway accidents to passenger travel in the United States is partly disclosed in the following table of number of passengers carried one mile for one killed during the period for which statistics have been kept:

The vagaries of railway casualties to the opportunities for accident strikingly illustrated in the foregoing table caused the official statistician as early as 1891 to remark:

"These facts seem to indicate that density of traffic has little to do with rate of casualties, but that the character of the equipment and the degree of intelligence among employes are the most important factors."

In his report for 1892 he said:

"Independently of character of employment and the rules by which running trains are managed, railway accidents depend upon two facts, namely, quality of labor and density of business.''

The remarkable decrease in the ratio of killed to passengers carried one mile in 1895, was attributed to the dropping out of the "incompetent and inefficient" employes attending the retrenchment of the preceding year and to "the fact that the demands upon the passenger service during the present year have been somewhat decreased."

From 1897 to this day, these conditions have been reversed. The demands on the passenger service have become more severe and exacting both as to accommodations and speed; while the recruiting of new and unexperienced men into the service has proceeded at an average rate of over 73,000 men per year. The inevitable effect of such conditions can be traced in the foregoing table, which shows an increase in the ratio of fatalities to traffic in spite of the fact that 32,818 miles of line are wholly and 44,574 miles are partially protected by the block signals, whereas in 1891, the official statistician remarked that "At present a very small portion of the railway system of the United States is operated on the block system."

It is therefore evident that the block system, now installed on nearly all the roads with heavy traffic, is not the primary essential to safe railway operation that the Commission, and the inventors and manufacturers of patent safety appliances continually represent it to be, but that competent, conscientious and experienced employes are.

Statistics, schools and legislatures cannot provide these last. They come through careful selection and training. Experience cannot be taught outside of the school which confronts the inexperienced tyro with the responsibility to act in the inevitable emergency which comes unheralded.

The statistics of railway accidents record the coincidence of such emergencies with the failure of the men or the equipment to meet them. They fail to record the billions of opportunities for for disaster passed with safety.

The chief reason why the Commission and the official statistician have been unable to discuss railway accidents profitably has been their failure to recognize the predominating influence of freight traffic on the safe handling of passenger trains. An analysis of the 87 "prominent" collisions reported in the quarterly Accident Bulletins of the Commission for 1905, classifies them as follows:

It will be observed that there were more than twice as many collisions between passenger and freight trains as between passenger trains only, and that the fatalities were greater in spite of the fact that there were 24 passenger trains involved in the collisions of passenger trains to only 15 in the collisions with freight trains.

It will also be noted that the "prominent" collisions between freight trains outnumber those of the other classes combined, although happily the loss of life was not so great.

How unaccountable are the elements entering into the risk of railway travel may be judged from the following table of casualties to passengers in train collisions during the four quarters of 1905, the most fatal year in the passenger record:

Quarterly Bulletins.

For July, August and September, 1904. . .

For October, November and December, 1904.

For January, February and March, 1905.

For April, May and June, 1905 ..

Total for year

No previous first quarter of the year shows within 60 per cent of the fatalities here recorded for the first quarter of 1905; and no previous last quarter shows within 100 per cent as few.

The only conclusion to be drawn from such variations is that the conditions accounting for railway accidents lie for the most part outside the realm of statistics, and that it is not safe to generalize as to the causes of accidents from the distressing numbers of the killed and injured.

RAILWAY ACCIDENTS IN FOREIGN COUNTRIES.

As it is popular to make invidious comparisons between railway accidents in the United States and Europe, it may be pertinent to present here the latest available statistics from official sources on the subject. Omitting the injured, because the definition of what constitutes a reportable injury varies with every country, the figures from the British Blue Book, issued last October, are as follows:

(a) In accidents to trains only.

(b) Excluding suicides, but including passengers killed otherwise than in

accidents to trains.

(c) Including passengers and employes.

(d) Including employes.

In comparing these figures with American statistics of accidents, it should be remembered that our passenger mileage is about one-half and our ton mileage more than double that of all Europe, the average speed of our trains greater and the railway right of way over there is not used as a public highway.