(Only one of twelve subclasses of causes appears to have any move ... 16 416 1,918 20 II To show the number and percentage of casualties from those causes which may, in some instances, have been a result of length of trains, the data in the foregoing table are summarized in Table VI as follows: It appears, then, that of the total number of collisions and derailments in 1915, only 1,545, or 15 per cent, might conceivably have been due in some cases to the length of the trains. But the actual number of such accidents assignable, on any reasonable theory, to the length of trains can be only a fraction of this number, for not all of these collisions and derailments happened to long trains, because long trains constituted only a small portion of all trains. Most trains are either passenger trains or short freight trains. The average number of cars per freight train in 1914 was only 34.3, showing a large majority of trains to have had less than 50 cars,1 the usual arbitrary division point between long and short trains. Further, it is highly improbable that more than a part of such long trains as suffered collisions or derailments did so by reason of any cause due to their length. It is therefore not easy to see how length of trains could have been a determining cause of collisions or derailments in any but a very small percentage of the total cases. Considering casualties to trainmen from causes other than collisions and derailments, it appears that the total casualties from these other causes that could possibly, on any hypothesis, have been affected by the length of trains, were 109 killed and 5,070 injured, or 15.4 and 15.9 per cent, respectively, of all the casualties from such causes. As a matter of fact, however, these causes also operate independently of the length of trains; they have many more occasions to operate in connection with the much more numerous short trains; and the number of casualties from these causes which conceivably might have been due to long trains, if any at all, must be much larger than the number that actually were due to them. When the general arguments offered to support the claim that long trains, considered as operating units, augment accidents are thus examined in the light of an analysis of the causes of accidents, the potential influence of long trains on the number of casualties is seen to be, at most, only a negligible matter. The Human Factor in Accidents. It appears from other evidence that a large percentage of accidents to trains is due to error or failure on the part of employees. From this it is argued that an increase in the number of trains, and 'If all mathematical possibilities be considered, the fact that the average of a number of items is less than some stated amount does not necessarily show that the majority of all the items are less than that amount. But, in view of the high probability that the number of trains within successively lengthened limits does not uniformly increase as rapidly as those limits, the statement in the text can hardly be questioned. therefore, in the number of employees, beyond what experience has shown to be necessary for the movement of a given traffic, would be attended by an increase in accidents to trains as well as to employees. During the 13 years, 1902 to 1915, the Interstate Commerce Commission investigated directly through its own agents 1,635 train accidents which were responsible for the death of 4,062 persons, or about one-third the total number killed in train accidents during this period, and for the injury of 23,981 persons, or over one-eighth of the total number injured in train accidents. The Commission found that 1,150 of these accidents were due to negligence, mistakes or other failures of employees, 240 to mechanical defects of track, equipment, signal system or air-brake apparatus, and the remainder to weather conditions, outside interference, parting of trains, miscellaneous and unknown causes. Of the whole number investigated, therefore, 70 per cent were due to error or failure on the part of the personal element concerned in train operation. Only 152 of the accidents, or less than 10 per cent of the total, could reasonably be connected with the length of trains, namely, 110 cases of defective or weakened equipment (which may or may not have been due to train length), 31 cases of air-brake failure (which again may not have been due to train length), and II cases of train-parting. These 152 accidents represent virtually all that could reasonably be attributed to a large number of cars per train. That this is the case becomes even clearer when it is noted that the 1,635 accidents include those in which passenger as well as freight trains were involved. Table VII, following, gives in detail the causes to which these accidents were attributed by the Commission : |