-re , nw*, UJ^X,A» 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 REPORT OF THE CHIEF OF THE BUREAU OF SAFETY COVERING THE 
 INVESTIGATION OF AN ACCIDENT WHICH OCCURRED ON THE CHI- 
 CAGO GREAT WESTERN RAILWAY NEAR WYETH, MO., ON JANUARY 
 3, 1920. 
 
 April 29, 1921. 
 To the Commission : 
 
 On January 3, 1920, there was a derailment of a passenger train 
 on the Chicago Great Western Railway near Wyeth, Mo., which re- 
 sulted in the death of 1 passenger, and the injury of 79 passengers and 
 3 Pullman employees. As a result of the investigation of this acci- 
 dent I submit the following report: 
 
 The accident occurred on the. seventh district of the southern 
 division, a single-track line extending between Leavenworth and Con- 
 ception, Mo., a distance of 74.2 miles, over which trains are operated 
 by time-table and train orders. Proceeding westward from the 
 station at Wyeth there is 912 feet of tangent, followed by a 3-degree 
 curve to the left 846 feet in length, 1,678 feet of tangent, and a 2- 
 degree curve to the left, 679 feet in length. The accident occurred 
 on this latter curve at a point 427 feet west of its eastern end. The 
 grade from Wyeth is practically level for 1,750 feet, followed by a 
 descending grade to the point of accident, varying from 0.29 per cent 
 to 0.83 per cent. The track is laid with 85-pound rails, 33 feet in 
 length, single spiked to an average of 18 oak ties to the rail, ballasted 
 with stone and cinders to a depth of from 8 to 12 inches. Tie-plates 
 are used on curves. The rails were laid in the track in October, 1904. 
 The general condition of the track was fairly good, although many 
 of the rails were not in good condition. At the time of the accident 
 the weather was cold and partly cloudy. 
 
 The train involved in this accident was westbound passenger train 
 No. 3, en route from Minneapolis, Minn., to Kansas City, Mo. It 
 consisted of 1 mail car, 1 baggage car, 1 smoking car, 1 day 
 coach, and 8 Pullman sleeping cars, in the order given, hauled 
 by engines 923 and 286, and was in charge of Conductor Cavanaugh 
 and Enginemen Venn and Peavy. It passed Rea, 3.3 miles east of 
 Wyeth and the last open telegraph office, at 7.17 a. m., 3 hours and 
 8 minutes late, and at about 7.26 a. m. was derailed at a point about 
 three-quarters of a mile west of Wyeth while traveling at a speed 
 estimated to have been about 40 or 45 miles an hour. 
 
 The two engines and the first three cars came to a stop about 
 2,100 feet beyond the point of derailment, wit^" o1nlylJ^\|arjwa3'tT pq(r 
 
 of wheels of the rear truck of the baggage car derailed. All of the 
 remaining cars in the train were derailed to the left side of the 
 
 47603—21—1 
 
 U.S. 0EPO«TO«Y 
 
2 INTERSTATE COMMERCE COMMISSION. 
 
 track. The day coach was derailed just before passing over bridge 
 F-460, a 96-foot plate-girder structure, located 630 feet beyond the 
 initial point of derailment, and after passing over the bridge went 
 down a 20-foot embankment, coming to rest about 50 feet from the 
 track. The first and second sleeping cars went down the embankment 
 just before reaching the bridge, coming to rest on their sides at a point 
 about 50 feet from the track. The third sleeping car came to rest 
 in an upright position with its forward end down the embankment 
 and the rear end resting on the roadbed, being almost at right angles 
 to the track. The fourth sleeping car remained upright with its 
 forward end close to the rails and its rear end down the embank- 
 ment, 10 or 12 feet from the track. This car remained coupled to 
 the fifth sleeping car, which was also entirely derailed, with its for- 
 ward end down the embankment, while the rear end remained on 
 the roadbed close to the left rail, with the body of the car inclined 
 at an angle of about 45 degrees. The remaining three sleeping cars 
 came to rest in an upright position close to the rails. 
 
 Examination of the track showed that the first marks of derail- 
 ment weTe at a broken rail on the inside or left of the curve, at a 
 point about 150 feet back of the last car in the train. The receiving 
 end of this rail was intact for a distance of 23 feet 6 inches, and re- 
 mained upright and spiked in its proper position. The running sur- 
 face was slightly beveled at the break, indicating that a few wheels 
 had passed over it. The remainder of the rail was broken into many 
 pieces, 32 of which were recovered; these 32 pieces constituted only 
 a small portion of the leaving end. The receiving end of the next 
 rail on the west remained in place, with the rail joints and the end 
 badly battered by wheels passing over them. This rail was intact for 
 a distance of 30 feet 6 inches. The balance of the leaving end, to- 
 gether with the seven adjoining rails, was torn out of the track by 
 the derailed equipment. None of the rails on the outside of the curve 
 was disturbed in any way. The first flange marks on the ties were 
 on the left sides of both rails, and were first visible opposite the first 
 broken rail at the end of the receiving portion of 23 feet 6 inches. 
 Careful examination failed to disclose any marks indicating that any 
 portion of the train had been derailed before reaching this point. 
 
 Engineman Venn, of engine 923, stated that he shut off steam a 
 short distance beyond Rea, at which time the speed of the train was 
 about 45 miles an hour, and that he reduced this speed to about 35 
 miles an hour on the curve just east of Wyeth. The speed then in- 
 creased slightly on the descending grade and was about 40 miles an 
 hour at the time of the derailment. He did not notice any jar or 
 unusual motion of the engine when passing over the point of derail- 
 ment, and had started to work steam when the brakes were applied 
 from the rear: at this time his engine was at the eastern end of the 
 
ACCIDENT NEAR WYETH, MO. 3 
 
 bridge. He at once placed his brake valve in the lap position, and 
 then changed it to the emergency position. He stated that he did not 
 make any detailed examination of the track, but from his observa- 
 tions concluded that a broken rail was responsible. He considered 
 the track to be in fair condition and safe for the speed at which his 
 train was running. 
 
 Fireman Hutchings, of engine 923, said that he did not notice 
 any unusual motion of the engine. His first knowledge of the derail- 
 ment was when he felt the brakes applied. At this time he was rid- 
 ing on his seat and on looking back he saw the day coach going 
 down the embankment. He thought the speed of the train at the 
 time was about 35 or 40 miles an hour. 
 
 Engineman Peavy, of engine 286, stated that he shut off steam 
 about 4 miles east of the point of derailment. The speed was re- 
 duced between Rea and Wyeth by an application of the air brakes, 
 and again when passing Wyeth. The balance of his statements 
 corroborated those of Engineman Venn. The statements of Fireman 
 Wilcox, of engine 286, added nothing to those of Engineman Peavy. 
 
 Conductor Cavanaugh stated that he was riding in the forward 
 end of the smoking car when he felt a severe jar, as though the car 
 had been derailed, and he at once pulled the emergency cord for the 
 purpose of stopping the train. He estimated the speed to have been 
 from 40 to 45 miles an hour. Shortly after the accident he went 
 back to the rear of the train, but did not make a detailed examina- 
 tion of the track, concluding that a broken rail was responsible for 
 the accident. 
 
 P. A. Nolden, employed in the engineering department of the 
 railway, was a passenger on the train. Shortly after it was derailed 
 he went back to examine the track, and about 100 or 150 feet beyond 
 the rear of the train he found a broken rail, the leaving end of 
 which was broken into a number of pieces, while the receiving end 
 was intact. The next rail on the west was also broken on the leav- 
 ing end. 
 
 Track Supervisor Millett stated that he reached the scene of the 
 accident at about 10.30 a. m. Examination of the first broken rail 
 disclosed a flaw in the head. He saw nothing on the running surface 
 of the rail to indicate that it was defective. He also found a crack 
 in the head of the receiving portion of this rail which extended back 
 into the rail. In his opinion the defect in this rail was such that it 
 could not have been detected without lying down on the ground. 
 He thought that the break at the leaving end of the adjoining rail 
 was due to the derailed equipment. He also said that he had passed 
 over the track on the preceding day, and at that time did not notice 
 any indication of anything wrong. The superelevation on the curve 
 
4 INTERSTATE COMMERCE COMMISSION. 
 
 was 2 inches, and he considered it safe for a speed of 50 miles an 
 hour. 
 
 The investigation clearly developed that the accident was due to 
 a broken rail, which apparently had been in a defective condition 
 for a considerable length of time. The examination to determine 
 the reason for the failure of this rail was conducted by Mr. James 
 K. Howard, engineer-physicist, whose report immediately follows: 
 
 REPORT OF THE ENGINEER-PHYSICIST. 
 
 The derailment of train No. 3 was due to the failure of a rail, the 
 leaving end of which was broken into a number of small fragments. 
 The receiving end, for a length of 23 feet 6 inches, remained intact. 
 The type of rupture displayed was a split-head fracture. A ver- 
 tical plane of rupture was developed which nearly separated the 
 head into halves, which extended along the length of the rail for a 
 distance of 13 feet. Of this section, 9 feet 6 inches was broken into 
 small fragments, 32 being recovered. 
 
 The rail was 85 pounds weight, A. S. C. E. section, rolled in Sep- 
 tember, 1904, and laid in the track in October of that year. Its age 
 was therefore 15 years 3 months. It was made of Bessemer steel, 
 heat number 46664, and branded " Illinois Steel Co. So. Wks. IX 
 1904 8504." 
 
 This rail presented a common type of rupture, the characteristics 
 of which are illustrated in the several cuts herewith, reproduced from 
 photographs and sulphur prints. In rails with split heads a lon- 
 gitudinal, vertical plane of rupture is developed, located along the 
 middle of the width of the head. The origin of the plane of rupture 
 is an interior one, located about one-quarter of an inch, more or 
 less, below the running surface of the head. The shallow zone above 
 the origin of the rupture remains unbroken until the last stages of 
 failure are reached. In the development of the split head the plane 
 of rupture extends downward until abreast the junction of the head 
 and the web. Here it commonly bifurcates, the branches extending 
 right and left toward the fillets under the head. Final rupture 
 occurs by the complete separation of the halves of the head and 
 their detachment from the web. At the upper initial edge of the 
 plane of rupture a small v-shaped ridge of metal, attached to the 
 upper zone of metal, is frequently found. This acts as a wedge to 
 separate the walls of the fracture. 
 
 The incipient point of rupture, in respect to the length of the 
 rail, was somewhere in the 9-foot 6-inch section, which was broken 
 into small fragments. The characteristics of the metal of the rail 
 were similar along this portion of its length and in the unbroken 
 part adjacent thereto. The photographic cuts and sulphur prints 
 represent the adjacent metal. 
 
ACCIDENT NEAR WYETH, MO. 5 
 
 Figure No. 1 shows the appearance of the rail, in cross section, 
 near the fractured portion, after polishing and etching with tincture 
 of iodine. The upper edge of the plane of rupture is characterized 
 by the presence of markings on the end surface revealed by the tinc- 
 ture of iodine. These markings, here viewed on end, represent 
 longitudinal streaks in the steel. They are lines of structural weak- 
 ness, affecting the metal under crosswise strains. 
 
 The split in the head, shown at this stage of development, is much 
 wider at its upper edge than elsewhere. The wedge-shaped rib of 
 metal at its upper edge is forced by the successive wheel pressures 
 between the faces of rupture, thereby increasing the width of the 
 head. 
 
 The plane of rupture separated into two branches. One branch 
 extended and reached the periphery of the section at the fillet under 
 the head on the outside of the rail. The other branch extended to- 
 ward the gauge side of the rail and appears in a partially developed 
 stage. 
 
 Figure No. 2 is a side view of the rail showing the line of rupture 
 under the head which had reached the peripheral surface. Figure 
 No. 3 is an endwise view farther along the length of the rail, where 
 the split in the head was less developed. Fractures of this kind do 
 not always continue in an unbroken course, but deviate under the in- 
 fluence of contiguous streaks in the metal. The end surface shown 
 in this cut was polished and then etched with tincture of iodine. 
 
 Figure No. 4 is a sulphur print of the end surface shown by figure 
 No. 3. The markings are substantially the same as those revealed by 
 the use of iodine. Figures Nos. 5 and 6 are sulphur prints of longi- 
 tudinal surfaces of the head and base, respectively, each about one- 
 fourth inch below the peripheral surfaces. Other sulphur prints at 
 different depths showed similar longitudinal streaks, but not coin- 
 ciding with those on surfaces near by. 
 
 Figure No. 7 represents the rail in cross section at a place where 
 the head was intact. The iodine markings show the continuance of 
 the structural conditions which prevailed in other parts of the rail. 
 Figure No. 8 represents another part of the rail where the head was 
 intact. This section was pickled in hot hydrochloric acid. The same 
 characteristic markings appear on the cross sections, whether re- 
 vealed by polishing and etching with tincture of iodine, by means 
 of sulphur prints, or upon pickling in hot acid. 
 
 Photomicrographs of this rail will be presented and discussed in 
 a later part of the report. Illustrations are herewith presented of 
 other rails which have failed and caused derailments — failures which 
 were influenced by the structural state of the metal. 
 
 Figures Nos. 9 to 13, inclusive, represent a rail which failed on 
 March 30, 1920, causing the derailment of train No. Ill, near Savan, 
 
b INTERSTATE COMMERCE COMMISSION. 
 
 Pa., on the Indiana Branch of the Buffalo, Rochester & Pittsburgh 
 Railway Co. This was an 80-pound rail rolled by the Carnegie Steel 
 Co. It failed by the development of a fracture under the head, at 
 the junction of the web. Its development was progressive, having its 
 origin at a zone of streaked metal. The head was detached from the 
 web, followed by the fracture of the web and the base. 
 
 Two views of the principal surfaces of rupture are shown by fig- 
 ures Nos. 9 and 10. Figure No. 9 is looking up, at the under side of the 
 head. Figure No. 10 is looking down, on the upper edge of the 
 web. The surfaces of progressive fractures wherever found are 
 similar in appearance, leaving no doubt concerning their identity. 
 The characteristics of fractured surfaces commonly furnish reliable 
 evidence upon the manner of failure of steel members. 
 
 Opportunity was taken to further examine the Savan rail in quest 
 of surface seaminess — a condition of the metal of the base which has 
 led to many fractured rails. The results are shown by figures Nos. 
 11 and 12. Figure No. 11 represents the appearance of the base of 
 the rail, two fragments, as it appeared when removed from the track. 
 Spike-maul marks are shown on the left-hand figure of the cut. 
 Figure No. 12 shows the appearance of the surface of the base after 
 pickling in hot hydrochloric acid. Surface seaminess, not in evi- 
 dence on the rail as it came from the track, was revealed upon pick- 
 ling. 
 
 Two end surfaces of the Savan rail as they appeared after pickling 
 are shown by figure No. 13. 
 
 Figures Nos. 14 to 18, inclusive, represent a rail which failed on 
 January 19, 1921, causing the derailment of train No. 9, of the Erie 
 Railroad, near Friendship, N. Y. This was a 90-pound rail rolled 
 by the Carnegie Steel Co., and was branded " Carnegie 1909 E. T. 
 IIIIIIII 90 A." It illustrates a piped rail, in connection with which 
 a split-head fracture developed. 
 
 Split-head rails are often erroneously reported as piped rails. 
 The primary causes which lead to the failure of these two types are 
 distinctly different, and their origins are located in different parts 
 of the cross section of the rail. A split-head fracture has its origin 
 in the upper part of the head. A piped rail has a plane of separa- 
 tion in the web and lower part of the head. Split-head rails are of 
 frequent occurrence, while piped rails are not. A split-head fracture 
 may occur in conjunction with a piped rail as the present rail 
 shows — a matter not affecting their separate origins. 
 
 Figure No. 14 shows the hot sawed end of the Friendship rail. 
 The vertical line of separation in the web shows the characteristic 
 feature of a piped rail. This rail displayed a composite fracture 
 in which its piped condition was probably the leading cause. Asso- 
 ciated with the plane of rupture in the web, there was a horizontal 
 
ACCIDENT NEAR WYETH, MO. 7 
 
 shearing fracture in the head and also a split-head fracture. The 
 piped state of the rail was embraced in part and reenforced by the 
 splice bars. The support given the rail by the splice bars doubt- 
 less accounts for the display of the several types of rupture in imme- 
 diate association with each other. 
 
 Figure No. 15 is a side view of this rail showing the fractured sur- 
 face of the web between the bolt holes and partially exposed a short 
 distance beyond. Above this portion of its length the head of the 
 rail was broken in a crosswise direction in addition to a vertical 
 plane of rupture which nearly separated it into halves. Both the 
 pipe and the split-head fracture continued beyond the limits of this 
 photograph. 
 
 Figure No. 16 is a view in cross section of the Friendship rail 
 farther along its length, photographed after polishing and etching 
 with tincture of iodine. The pipe extended into this section. The 
 opposite faces were in close proximity to each other, hence the pipe 
 does not show in this cut. The split in the head had separated and 
 is clearly visible. 
 
 Figure No. 17 is a sulphur print of the Friendship rail at a place 
 where the pipe was clearly visible, and also where the split in the 
 head was in an advanced stage. A lateral branch of the split-head 
 fracture had nearly reached the peripheral surface of the rail at the 
 fillet under the head on the gauge side. The fracture of the rail at 
 this point is of interest in showing the dominating influence of the 
 wheel pressures in relation to fractures in the head. The formation 
 of a lateral branch of the split-head fracture doubtless resulted from 
 the wheel pressures exerting a spreading effect on the metal at the 
 top of the head. In the order of development the formation of this 
 lateral branch probably was the last part of the fracture to occur. 
 The walls of the split-head fracture were separated by the wedge 
 action of the cold-rolled metal of the top of the head. When the 
 medial line of pressure between the wheel and the rail occurs near 
 either the gauge side or the outside edge of the head there is an 
 overturning moment applied to the head. Such eccentric loading 
 leads to the bifurcation of the line of rupture in the case of a split 
 head. The different contours of the treads of wheels are responsible 
 for different elements along the running surface receiving the max- 
 imum impinging pressures and causing the alternate loading of one 
 side and then the other of the head of the rail. This alternate eccen- 
 tric loading of the head accounts for the bifurcation of the vertical 
 plane of rupture in a split-head rail abreast the junction of the head 
 and the web where bending stresses in crosswise direction under such 
 circumstances attain high limits. It will be inferred from the present 
 exhibit and the remarks which are submitted that relatively there 
 is a greater tendency, under the influence of track conditions, for 
 
8 INTERSTATE COMMERCE COMMISSION. 
 
 a rail to fail by the development of a split head than by reason of 
 the presence of a pipe. 
 
 Figure No. 18 shoiws the appearance of the Friendship rail after 
 pickling in hot acid at the cross section upon which the preceding 
 sulphur print was taken. Greater solubility of the metal takes place 
 at those places which are stained by the iodine, or marked on the 
 sulphur print, than on other parts of the cross section, resulting in 
 the close similarity of the illustrations furnished by these three 
 methods. 
 
 In the study of the failure of steel, attention centers upon one 
 principal feature and the relations of subordinate features to the 
 principal one. The principal feature relates to the stress or strain, 
 mutually related factors, which the metal is capable of enduring and 
 the manner in which limiting values in terms of stress or strain are 
 reached. In plainer language, it is desired to know what loads the 
 steel will carry ; what relations the properties of the metal which are 
 shown under test bear to the endurance of service conditions ; whether 
 any part of the elongation displayed in the tensile test or the con- 
 traction of area then displayed will be realized in service; whether 
 the ultimate tensile strength of the steel represents a particular value 
 to the rail when in service ; whether the ductility clause of the drop 
 test has any real significance; whether the drop test itself has any 
 definite relation to the serviceability of the rail. These and other 
 queries suggest themselves. Chemical composition and finishing tem- 
 peratures are factors which influence and control physical properties 
 in rolled shapes — established by the laws of nature and not by speci- 
 fications. 
 
 Opportunity occasionally offers to acquire data touching upon 
 some of the above queries. On the present occasion microscopic ob- 
 servations were made upon the structural appearance of the metal 
 of the Wyeth rail adjacent to the running surface of the head, at 
 the upper terminal of the split-head fracture, and at the lower ter- 
 minals of the bifurcated fracture. The observations were directed 
 to the distortion of the grain of the steel adjacent to the running 
 surface, where flattening and flow occur immediately, due to the 
 wheel pressures, and noting the undistorted shape of the grains at 
 the split-head fracture. 
 
 The metal of this rail was capable of displaying ductile flow. The 
 distorted shape of the grain next the running surface and the for- 
 mation of a fin along each edge of the head is evidence thereof. At 
 the terminals of the split-head fracture and the apex of the wedge- 
 shaped rib at the upper edge of the split-head fracture there was, 
 however, no appreciable distortion of the grain. At these places, 
 microscopically, there was absent any evidence of an appreciable per- 
 manent set of the steel having taken place. Ductile flow and brittle- 
 
ACCIDENT NEAR WYETH, MO. 9 
 
 ness in the same rail are here shown, the result of the manner in 
 which the stresses were received. The same is true of other carbon 
 steel rails. Whether any part of the elongation or contraction of 
 area of the tensile test is displayed in service depends upon the man- 
 ner in which those features are developed. 
 
 A series of four photomicrographs was taken along the upper zone 
 of the head of the Wyeth rail, showing the shape of the grain of the 
 steel within the zone directly affected by the wheel pressures. Each 
 represents the metal of the rail in cross section, and each at a mag- 
 nification of 100 diameters. Fig. No. 19 represents the shape of the 
 grain near the running surface of the head about in line with the 
 gauge side of the web, that is, not far from the middle of the width 
 of the head. The grains were slightly flattened along this element, 
 the distortion hot being very pronounced. This element appeared to 
 be within the neutral axis with respect to distortion and direction of 
 flow of surface metal. 
 
 Figure No. 20 shows the distortion of the grain at a place nearer 
 the gauge side of the head. The flattening of the grain here is very 
 pronounced, with a drift toward the gauge side. Figure No. 21 
 shows the distortion of the grain on the opposite side of the center 
 line of the head. The flattening of the grain and drift toward the 
 outside edge of the head is here also very pronounced. A fin was 
 formed along each edge of the head, the metal to form which neces- 
 sarily came from the upper part of the head. Under such circum- 
 stances it is quite evident there would be a neutral element on either 
 side of which the surface flow would take place in opposite direc- 
 tions, as these photomicrographs indicate. The flow of metal at the 
 extreme outside edge of the head showed a laminated state, as illus- 
 trated by Figure No. 22. The laminations were separated and in- 
 dividually broken. The dark Z-shaped lines on the cut represent 
 lines of fracture. 
 
 The depth of the zone of distorted grain was about five-hundredths 
 of an inch. Below this depth normal shape of grain prevailed. A 
 feature of interest is raised in this connection — namely, that micro- 
 scopic evidence is not presented showing a disturbance of the struc- 
 tural state of the metal so far down as the origins of split-head frac- 
 tures are located. Evidence of a state of internal compression in the 
 upper part of the head does not rest upon microscopic indications, 
 but upon the results of strain gauge measurements. In comparing 
 the results of strain gauge measurements with the indications of the 
 microscope it has not been made clear that the presence of internal 
 strains of either tension or compression may be recognized with the 
 aid of the microscope. If such was the case the most far-reaching 
 results would come from the use of the microscope in ascertaining 
 the existing states of strain in all kinds of engineering structures. 
 47603—21 2 
 
10 INTERSTATE COMMERCE COMMISSION. 
 
 Figure No. 23 represents the apex of the wedge-shaped rib at the 
 upper terminal of the split-head fracture shown by figure No. 1. 
 There was no distortion of the grain at the apex of this wedge. This 
 illustration touches upon another feature in the behavior of metals — 
 namely, that cubic compression, however great, and it has been 
 observed up to a pressure of 117,000 pounds per square inch, has no 
 permanent effect upon the structural state of the steel. The wedging 
 apart of the split-head fracture, therefore, does not demand there 
 should be of necessity a distortion of the grain of the wedging mem- 
 ber. 
 
 Figure No. 24 shows the termination of the shorter branch of the 
 bifurcated split-head fracture illustrated in figure No. 1. The crack 
 in the steel is indicated by the oblique irregular line which ap- 
 pears in the cut, darker than the ferrite boundaries of the grains and 
 lighter than the pearlite grains. This appearance of the crack is due 
 to its being filled with iron rust. 
 
 Figure No. 25 shows the termination of the split-head fracture 
 illustrated in figure No. 3. The plane of the fracture at this place 
 had deflected and approached nearer the fillet under the head than 
 shown by figure No. 1. The irregular black line extending obliquely 
 downward in this cut represents the termination of the crack. The 
 surfaces of the fracture were not oxidized in this case. 
 
 The last two photomicrographs illustrate this feature : That a 
 plane of rupture may pass into or through a steel member without 
 change in shape of the grain, and therefore without display of ap- 
 preciable ductility. This has been found true in different grades of 
 steel. The behavior of rails in service furnishes the basis for the 
 query concerning the value in itself of the ductility clause of specifica- 
 tions. A greater or less display of elongation will take place in 
 steels, depending upon their composition, when tested in such a man- 
 ner as to permit of the display of ductility; that is, certain steels 
 inherently possess such ability by reason of their chemical composi- 
 tion. Specifications can enumerate these numerical values, but with- 
 out changing the results. These remarks are made because the 
 failures of materials are often attributed to lack in meeting specifica- 
 tions, when as a matter of fact the relation between the specified 
 properties and the ability of the material to endure service stresses 
 has not been given consideration. This is a plea for a better under- 
 standing of the specific causes which lead to failures, in which those 
 of rails present notable opportunities. 
 
 In summation, the failure of the rail which caused the present de- 
 railment was due to the presence of a split-head fracture. Wheel 
 loads cause distortion of the grain of the steel and induce lateral flow 
 of the metal at the running surface of the rail, the tendency of such 
 loads being to spread the railheads. The successful resistance of 
 
ACCIDENT NEAR WYETH, MO. 11 
 
 such lateral forces depends upon the structural soundness of the 
 metal in the railhead. Longitudinal streaks are lines of weakness 
 which influence the formation of split-head fractures and locate 
 their incipient points of origin. Longitudinal streaks are due to 
 casting and mill conditions. Their elimination, or reduction in num- 
 bers and gravity of development, are matters for the steel makers to 
 consider. The ages at which split-head rails manifest themselves 
 indicate such fractures are of slow and progressive development. It 
 is a matter of conjecture, although having the appearance of proba- 
 bility, that split-head rails would be unknown if strictly seamless 
 steel was available for rails. The rail problem is intensified by rea- 
 son of the employment of high wheel pressures. Soft rails display 
 mashed heads. Hard rails furnish a large number of transverse 
 fissures. 
 
 There is a popular fallacy entertained that split-head rails do not 
 constitute a dangerous type of fracture, since at certain stages in 
 their progress of rupture they may be detected in the track. This 
 evidence, however, is presented at a late stage, after the necessary 
 margin in strength in the rail has been practically exhausted, and 
 not prior thereto. An element of danger has arisen when split-head 
 rails are detectable in the track. An economic question is involved 
 in the elimination of the causes of split-head failures, since many 
 rails are removed for this cause which are not otherwise unservice- 
 able. Finally, split-head failures should not be reported as piped 
 rails. 
 
 SUMMARY. 
 
 The cause of this accident is shown to have been due to the failure 
 of a rail which displayed a split-head fracture. The head of the 
 rail was separated into halves by a vertical plane of rupture; the 
 halves of the head broken into fragments, with lines of rupture sep- 
 arating the web and the base. A portion of the length of the rail 
 broke into a number of small fragments. 
 
 As illustrated and described by the engineer-physicist this type of 
 fracture has its origin in the upper part of the head, the incipient 
 point being located a short distance below the running surface. It 
 also appears from the best evidence on the subject that split-head 
 fractures are induced by the presence of certain longitudinal streaks 
 or seams in the metal, and that such seams represent the incipient 
 places from whence planes of rupture extend and destroy the rail in 
 the course of their development. 
 
 It appears to be well established that split-head fractures begin 
 in the upper part of the head and progressively extend downward 
 in substantially vertical planes. When the plane of rupture reaches 
 a depth which brings it abreast or nearly abreast the junction of the 
 
12 INTERSTATE COMMERCE COMMISSION. 
 
 head and the web, the plane of rupture commonly changes its course 
 or separates into two branches, one of which eventually reaches the 
 periphery of the rail at the fillet under the head. When this stage 
 of rupture has been reached the ultimate failure of the rail soon takes 
 place. 
 
 The width of the split is narrow in comparison with its depth of 
 penetration — a circumstance which renders the detection of a split- 
 head rail uncertain until the period of final rupture is close at hand. 
 This fact should dislodge a popular fallacy concerning split-head 
 rails — namely, that such fractures are easily detected in the track, 
 and therefore should not occasion anxiety, overlooking the fact that 
 when detectable the rail has reached a weakened condition and may 
 be on the verge of rupture. 
 
 Data in the report are presented on the extension of cracks of in- 
 terior origin, showing the absence of the display of ductility of the 
 metal in fractures of this class. Without the display of ductility, ex- 
 ternal and visible evidence of impending rupture is evidently want- 
 ing. Methods of test have been offered for the detection of interior 
 fractures in rails. The development of such apparatus does not ap- 
 pear to have reached a state in which its application to rails in the 
 track has been attained. As the case now stands, the early detection 
 of split-head rails in the track depends upon the vigilance of the 
 track supervisors and the section men. 
 
 The engineer-physicist has ventured the remark that split-head 
 rails would be nearly or quite unknown provided seamless steel was 
 found in rails. The relation which seaminess of metal bears to split- 
 head rails appears to furnish a basis for this remark. It commonly 
 takes years of service to develop split-head fractures, which gives 
 encouragement to the thought that an improvement in the struc- 
 tural state of the steel would measurably prolong the lives of cer- 
 tain rails. Elements of safety and economy would be subserved if 
 the primary cause in the formation of split-head failures was re- 
 moved or the influence of such cause measurably lessened. 
 
 No comprehensive consideration can be given the subject of rail 
 failures without taking into account the effects of high wheel loads, 
 effects which are destructive in their character. Regardless of 
 whether responsibility in the abstract attaches chiefly to the makers 
 or the users of rails, statistics show that a considerable number of 
 rails fail under present conditions of service. A reduction in the 
 number is highly desirable. In respect to the display of split-head 
 failures, promise of improvement appears to lie in the direction of 
 using steel of less seamy state. 
 
 Respectfully submitted. 
 
 W. P. Borland, 
 Chief, Bureau of Safety. 
 
ACCIDENT NEAR WYETH, MO. 
 
 13 
 
 Fia.A 
 
 F19.B 
 
 Fig. A Typical split head fracture. 
 Fia.B Typical piped rail. 
 
14 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 1. — End view of rail polished and etched with tincture of iodine, showing 
 relations of markings in upper part of head and origin of split head fracture. 
 
ACCIDENT NEAR WYKTH, MO. 
 
 15 
 
 Fig. 2. — Side view of rail, showing longitudinal crack at the fillet under the 
 head, where split head fracture had reached the peripheral surface of the 
 rail. 
 
16 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 3. — End view of rail, polished and etched with tincture of iodine. Cross 
 section of rail where split head fracture was less developed than where shown 
 by figure 1. 
 
ACCIDENT NEAE WYETH, MO. 
 
 17 
 
 Fig. 4. — Sulphur print of surface shown by figure 3. 
 47603—21 3 
 
18 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 
 . * _■ --. * 
 
 vtycs 
 
 «"— ' . — 
 
 r.-*n «— *-- HO i n .1 1 , I f. ■ > "" i*~«»«Vt^»i*— JtfPM 
 
 Fig. 5.— Sulphur print of longitudinal, horizontal surface of head of rail, V 
 planed off the running surface. 
 
ACCIDENT NEAR WYETH, MO. 
 
 19 
 
 
 
 m > ' g. i ^iir 'ff iiejj jfg frfj t m,'. I 'm'j^ t mm i ^ ,<y^-~ 
 
 W&IBm& u m fT m w **~i i r , &_¥* • * ' .' « »»m iwi«» : <«^- 
 
 
 «'«-*: ^s^sfe r .-^.*>?^rr-,-^;' 
 
 Fig. 6. — Sulphur print of longitudinal, horizontal surface of base of rail, \' 
 planed off the bottom of the base. 
 
20 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 7. — Appearance of cross section of rail, where head was intact, polished 
 and etched with tincture of iodine. 
 
ACCIDENT NEAR WYETH, MO 
 
 21 
 
 Mi 
 S 
 
 $* 
 
 ■ ■■'. S« ' 
 
 f 
 
 ..»s>sstf p 
 
 ■M&&* 
 
 " t ;;#>>v^ r. >K>v'. 
 
 '._''_' &i*^ 
 
 &4 
 
 Fig. 8. — Appearance of cross section of rail, where head was intact, after 
 pickling in hot hydrochloric acid. 
 
22 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 > -a 
 
 o S 
 
ACCIDENT NEAR WYETH, MO. 
 
 23 
 
24 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 
 SfflHH 
 
 Pig. 11. — 80-pound rail which failed near Savan, Pa. Appearance of base before 
 
 pickling in hot acid. 
 
ACCIDENT NEAR WYETH, MO. 
 
 25 
 
 Fig. 12. — 80-pound rail which failed near Savan, Pa. Appearance of base after 
 
 pickling in hot acid. 
 
26 
 
 INTERSTATE COMMERCE COMMISSION. 
 
ACCIDENT NEAR WYETH, MO. 
 
 27 
 
 Fig. 14. — 90-pound rail which failed near Friendship, N. Y., Erie Railroad. 
 View of hot sawed end showing piped fracture in web, and horizontal 
 shearing fracture in the head. 
 
28 
 
 INTERSTATE COMMERCE COMMISSION. 
 
ACCIDENT NEAR WYETH, MO. 
 
 29 
 
 Fig. 16. — 90-pound rail which failed near Friendship, N. Y. Cross section, at 
 place beyond the limits of figure 15. Pipe in web obscurely shown, split 
 in head visible. 
 
30 
 
 INTERSTATE COMMERCE COMMISSION". 
 
 Fig. 17.- — 90-pound rail which failed near Friendship, N. Y. Sulphur print at a 
 place where it exhibited a piped web and a split head fracture, a branch ex- 
 tending in the direction of the fillet on the gauge side of the web. 
 
ACCIDENT NEAR WYETH, MO. 
 
 31 
 
 Fig. 18. — 90-pound rail which failed near Friendship, N. Y. Appearance after 
 pickling in hot acid, same cross section as shown by figure 17. 
 
32 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 19. — Photomicrograph of Wyeth rail, cross section 
 shown by figure 7, just below running surface, about 
 in line with gauge side of web, showing flattening 
 of the grain by wheel pressures. Magnification 100 
 diameters. 
 
ACCIDENT NEAR WYETH, MO. 
 
 33 
 
 Fig. 20. — Wyeth rail, same surface as shown by figure 
 19. Showing flattening of the grain and flow toward 
 gauge side of head, at a place between gauge side of 
 head and surface shown by figure 19. Magnification 
 100 diameters. 
 
34 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 21. — Wyeth rail, same surface as shown by figure 19. 
 Showing flattening of the grain and flow toward the 
 outside of the head, at a place between the outside 
 of the head and surface shown by figure 19. Magnifica- 
 tion 100 diameters. 
 
ACCIDENT NEAR WYETH, MO. 
 
 35 
 
 Fig. 22. — Wyeth rail, same surface as shown by figure 19. 
 Showing laminated structure of the fin formed along 
 the outside edge of the head. Laminations separated 
 and individually broken. Z-shaped dark lines indicate 
 lines of rupture. Magnification 100 diameters. 
 
36 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 23. — Yyeth rail. Microstructure of apex of wedge- 
 shaped rib at upper terminal of split head fracture 
 Bhown by figure 1. Shapes of grains undisturbed. 
 Magnification 100 diameters. 
 
ACCIDENT NEAR WYETH, MO. 
 
 37 
 
 Pig. 24. — Wyeth rail. Mierostructure at shorter branch of 
 bifurcated head fracture shown by figure 1. Crack shown 
 by irregular oblique line intermediate in color between 
 the ferrite boundaries and the pearlite grains, penetrating 
 steel without distortion of the grain. Magnification 300 
 diameters. 
 
38 
 
 INTERSTATE COMMERCE COMMISSION. 
 
 Fig. 25. — Wyeth rail. Micro-structure at lower terminal of 
 split head fracture shown by figure 3. Crack shown 
 by irregular dark line extending obliquely downward, 
 penetrating the steel without distortion of the grain. 
 Magnification 300 diameters. 
 
 o 
 
UNIVERSITY OF FLORIDA 
 
 3 1262 08856 1542