Report of the chief inspector of safety appliances covering the investigation of an accident which occurred on the New Y...


Material Information

Report of the chief inspector of safety appliances covering the investigation of an accident which occurred on the New York, New Haven & Hartford Railroad near Westerly, R.I., on October 25, 1913
Physical Description:
32 p. : ill. ; 23 cm.
United States -- Interstate Commerce Commission
Belnap, H. W ( Hiram W )
Place of Publication:
Publication Date:


Subjects / Keywords:
Railroad accidents -- Rhode Island -- Westerly   ( lcsh )
federal government publication   ( marcgt )
non-fiction   ( marcgt )


General Note:
Dated April 24, 1914.
General Note:
Submitted by H. W. Belnap Chief Inspector of Safety Appliances.
General Note:
At head of title: Interstate Commerce Commission.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 004952900
oclc - 52166791
System ID:

This item is only available as the following downloads:

Full Text

U d L..
UNIV. 0F !-L Lt V *


To the Commission:
On October 25, 1913, there was a derailment of a passenger train
on the New York, New Haven & Hartford Railroad near Westerly,
R. I., which resulted in the injury of 74 passengers and 3 em-
ployees. Investigation of this accident was had in conjunction with
the Public Utilities Commission of the State of Connecticut, and a
public hearing was held at Providence, R. I., on October 31, 1913.
As a result of the investigation of this accident I beg to submit the
following report:
The derailed train was eastbound train No. 26, en route from New
York, N. Y., to Boston, Mass. It consisted of three Pullman cars,
all equipped with steel underframes, one smoking car, and two
coaches, all of wooden construction, hauled by locomotive No. 1309.
The train was in charge of Conductor Taber and Engineman Smith.
Train No. 26 left Westerly at 9.25 p. m., 14 minutes late, and at
about 9.30 p. m. was derailed at a point 1.6 miles east of Westerly
while running at a speed estimated to have been between 30 and 35
miles per hour. Neither the engine nor the tender were derailed.
With the exception of one wheel on the north rail, all of the wheels
under the first Pullman car were derailed, while all the other cars in
the train were derailed and came to rest on the south side of the track,
some of them extending partly over the embankment.
The train broke in two between the second and third cars, the
four rear cars being separated from the forward portion of the train
a distance of about 150 feet. Illustration No. 1 is a view looking in
a westerly direction, and shows the position of the last four cars
after the derailment.
This part of the New York, New Haven & Hartford Railroad is a
double-track line, and trains are operated under the controlled-
manual block-signal system. Approaching the point of derailment
from the west there are about 2,000 feet of tangent, all on a descend-
ing grade of about one-half of 1 per cent. The track is laid with
100-pound steel rails, 33 feet in length, single spiked to 18 or 19

f ,l '


APRIL 24, 1914.


untreated chestnut, oak, and pine ties, no tie-plates being used on
straight track. At the point of derailment the track is on a 12-foot
fill, chiefly conillposd of gravel. The ballast is of gravel varying from



12 to 16 inches in depth. Examination showed this track to be in
good conditions. It wa.i, rainiingI at tlh time of the derailment.
Ex!ii;iilnatiiinl of the eqiiiplinlt t of the derailed train showed nothing
which in any way could have contributed to the derailment. Exami-
ii:ttion of the track :-hIi\ed that the first indication of anything wrong


was a broken rail on the .oitlh side of the track. West of this bro kei
rail there were no marks of any kind upon the rails or tie-, while
east of the same the ties had been cut and broken by derailed wheels,
the track being torn up for a distance of abiit. 600 feet. East of the
initial point of derailnmelt the north rail \was torn out of aliLlgn-iiint
for a distance of about 12 rail lengths, while 12 successive rails on
the .-outh side were also torn up. Four of these rails on the south
side were separated from each other, the bolts at the rail joints having
been sheared off.
The crew of an eastbound passenger train which lpa-ed over this
track less than an hour previous to the derailment, testified that
they felt no'unevenne-s in the track, and that they did not notice
anything which would indicate that there was anything wrong with
it. Engineman Smith, of train No. 26, stated that the first thing he
noticed was a slight jar or yank. He at once applied the air brakes.
and on looking back saw fire flying from underneath the cars. After
the accident no defects or damage of any kind were found to exist
with respect to the locomotive, and he operated it through to Boston.
Fireman Muirphy testified that at the time of the derailment he was
putting coal on the fire. He did not notice any jar from the driving
wheels, being of the opinion that it came from behind the engine.
The testimony of the other members of the crew shed no light as
to the cause of the accident, their first in6timation that there was
anything wrong being the shock occasioned by the cars being derailed,
coupled with the application of the air brakes.
This accident was caused by a broken rail. The investigation to
determine the reason for the failure of this rail was conducted by
Mr. James E. Howard, engineer physicist, whose report immediately

The broken rail which caused the derailment of train No. 26 was
a C rail, 100 pounds section, open-hearth steel, of the New York.
New Haven & Hartford Railroad's design, manufactured by the
Bethlehem Steel Co., April, 1910, heat J-1391. It was laid in th,
track June 5, 1910, and therefore had been in service for a period
of three years and four months at the time of derailment.
The rail had the following dimensions:
Height-------------- --------- 6 inches.
Width of head 2 ------------ inches.
Width of base______-----_ ____ 5i inches.
Thickness of web ------------------ inch.
Length ------ ------- 33 feet.
Moment of inertia... .---_.--__- .-__.-- 47.18


The specifications for chemical composition governing its manu-
facture and the composition reported as having been furnished were:

Chemical con- Called for by Said to have
stituents the specifi- been
stituencations. furnished.

Carbon........... 0.70to0.83 0.80
Manganese....... .60to .90 .76
Silicon........... .20 .183
Phosphorus...... .04 .033
Sulphur.......... ............... .046

The specifications called for a drop test in which a 2,000-pound
tup should be dropped from a height of 15 feet, the rail resting upon
supports 3 feet apart. It was required that the rail should deflect not
more than 1.45 inches on the first blow, nor upon fracture display
less than 6 per cent elongation in 1 inch, or 5 per cent in 2 consecu-
tive inches. The drop test made on this heat of steel was reported
as having shown a deflection of 0.9 inch.
This rail showed very little wear as the result of its service in the
track. The head retained its shape, and externally the appearance of
the rail was good.
It fractured in three places at the time of derailment, at distances
of 1 foot 10 inches, 4 feet 10 inches, and 6 feet 7 inches, respec-
tively, from the leaving end. At the first and third of these places
transverse fissures were disclosed measuring in diameter about 1
inches each. The initial line of fracture was probably that which
occurred 6 feet 7 inches from the leaving end. The intermediate
fracture, believed to have been a secondary one, did not have a trans-
verse fissure.
Photograph, figure No. 2, shows the relative positions of these lines
of fracture as they were viewed from the gauge side of the rail.
The movement of the train on the rail was from right to left. Line
of rupture CC was the first to occur, it is thought.
The train left the track southerly through the opening made by
the three fragments shown on figure No. 2 and the opening made by
succeeding rails east of this point. To the west of the line of rupture
CC the track remained intact.
Subsequent to the derailment an additional fracture was made
when removing the rail from the track, at a place about 54 feet west
of line of rupture CC. At this place a transverse fissure 1- inches in
diameter was displayed. The rail was then shipped to Providence,
where two more transverse fissures were disclosed upon raising one
end of the rail and allowing it to fall upon a concrete walk from a
height of about 6 feet. These fissures measured 11 inches and five-
eighths inch in di; meter, respectively.


In all, five transverse
fissures were displayed
in the rail, each of
which was located on
the gauge side of the
head. In resume these
fissures were located at
the following distances
from the leaving end of
the rail : 1 foot 10 inches,
6 feet 7 inches, 12 feet
1 inch, 17 feet 8 inches.
and 21 feet.
A second rail, from the
same heat as the above,
was removed from the
track and its structural
condition examined.
This rail, branded
"Bethlehem Open
Hearth 100B IIII 10,"
was taken from the track
adjacent to or near the
broken rail. Both were
laid at the same time,
and each was exposed to
the same conditions of
These rails were tested
in part at the Bureau of
Standards, while con-
tributory work was done.
at the Washington Navy
Yard and by the New
Haven Railroad at its
New Haven laboratory
and at the works of the
Bethlehem Steel Co..
followed by a metallo-
graphic examination by
Mr. Wirt Tassin.
The report of the Bu-
reau of Standards upon
the chemical composi-

I X)


49 aS/LL~A



tion of the steel, slag determination, metallogi aphic exaniiination, and
tensile tests follows:


In Table I are shown the results of chemical analysis: of rails 1
and 2 taken at three places.

Location ar- Sul- Phos- Man- Si- Nickel. Oxad Chro-
bon. phur. phorus.: ganese. con. alag mium.

Rail 1:
Near running surface of head..... 0.83 0.039 0.063 0.78 0.166 0.081 0.11 0.05
Junction of web and head........ 82 .040 .061 .79 .164 049 .14 .04
.26 .......
Flange of base................... .84 .043 .063 .79 .166 .069 05
Rail 2:
Near running surface of head.... .83 .040 .058 .79 .157 .29 .07 .02
Junction of web and head........ .85 .041 .060 .79 .147 .27 .04 .03
Flange of base................... .84 .039 .059 .80 .152 .29 .07 .02

It will be noted that the two rails, barring slag, are practically of
identical composition.
The agreement between the three positions in each rail is also good,
except with to nickel and slag in No. 1, showing no apprecia-
ble segregation, if any, of the chemical constituents. The difference
as to nickel may be due to errors of analysis where such small amounts
are concerned, and are probably without significance.
Attention should be called to the values reported for slag and
the question of slag and oxide analysis in general. The methods
used for both are very unsatisfactory, in that we have no real knowl-
edge that they are reliable, but a good deal of reason to believe that
they fail to tell us what they purport to tell. That is to say, in the
ase of slag we do not know if all slag is obtained by the method
used, i. e., insolubility in iodine, or how much of what may be re-
ported as slag is such. For instance, in the present case, the silica
percentage in the 1 anI found in No. 1 does not exceed 20 per cent,
by actual test of several samples. This means le.- than 50 per cent
of silicate slag, if all the silica comes from that; but if any iron sili-
cide is included in the silica found, the slag percentage should be
lowered by an indeterminate amount.
Again, this slag (ignited) carries about 9.8 per cent P,O,,
which we may suppose to have belonged to iron phlosphile. If so,
and if the composition of the phosphide is FeP, and if again this
became converted during ignition to FeO, and PO,, we must deduct
the oxygen corresponding to this change, which in the present in-
stance would be 14.5 per cent.


Still again, the slag containedl a little chromium in unknown con-
dition. Allowing for the iimaximIll amount of real (.ilicat,-) slag
permissible as deduced from the silica percentage and of iron silicide
and phosphide, there remains a large probable deficitlicy, which may
perhaps be made up by oxide of iron or some oxide other than one of
manganese, which element is not preenit in the slag from either rail
(calcium is also absent). If the slag carried carbide and silicide
of iron, the iron and silicon of these would be left after ignition as
FeO, and SiO,. The variations in slag noted for rail 1 may be due
to actual local variations in -lag content or perhaps in part to uncer-
tain analysis.
It is prolballe that the slag analyses are comparable for the two
rails, since these are otherwise of very exactly the same composition.
It is of interest to note that the rail No. 1 which failed in service had
three or four times as much slag as the other rail from the same
heat in the same t rack, suggesting a greater inherent weakness refer-
le to this cause.
A section of rail No. 1 was cut 5 inches 1, a-k of break, 12 feet from
receiving end, polished and etched electrolytically by being made the
anode in a bath of ammonium chloride. By this treatment the areas
of segregation are shown by dark spots and streaks. Figure No. 3
shows the appearance of the section after this treatment. The web
shows a considerable amount of segregation, but the metal of head
and base is not -eriously affected. A section of rail No. 2 treated in
the same way shows a structure nearly identical with that of rail No.
1, as shown on the same figure. The amount of segregation shown by
these two sections may be regarded as typical and appears to bear
no intimate relation to the formation of the transverse fissures found
in rail No. 1.
Sections for microscopic examination were taken from head, web
and base of each of the two rails. The metal in the head was ex-
amined both from the gauge side and opposite portion. Except for
an increase of grain size in the head and occasional slag threads, the
structure is very uniform throughout. As near as can be judged, the
microstructure of the two rails is identical. The metal consists of an
intimate mixture of pearlite crystals, i. e., saturated or eutectoid steel.
The method of etching used, 2 per cent nitric acid in alcohol, darkens
the pearlite; the lighter appearance of many of the crystals is due to
the different reflection of the light caused by the orientation of the
In the interstices between many of the crystals are areas of very
coarse pearlite. Such areas are numerous and are found scattered
uniformly throughout the whole mass. In these areas the two con-
stituents of pearlite (ferrite or pure iron and cementite or carbide of


iron) are in particles of sufficient size (i. e., plates) that the weaken-
ing effect upon the metal as a whole must be appreciable.
These areas can not be well represented in a photomicrograph of 100
diameters magnification. They appear as small light-colored grains.
After annealing, these interstitial pearlite areas are more pro-
nounced and distinct.


0 0

End views A and B in figu re No. 4 show the typical appearance of
the transverse fissures abundant in the head of rail No. 1. No fis-
sures were found other than on gauge side of the rail. The metal im-
mediately adjacent to such fissures was examined in detail. Sections
cB I-

i' ~b.to0 F
f- ,s

mnediqteh' adjacent to such fissulres Wa'..~ examined in (let'111. Sections








o '

~ )o

o '

o- 0


W- 0


41 S07-14---2


were taken perpendicular to the face of the fissure and the metal im-
mediately back of the break examined. It was not found to differ
materially in structure or constituents from that of the rail as a
whole. Photomicrographs C and D show the structure of the metal
just k1 ak of fissures at breaks 12 feet and 15 feet 4 inches, respectively,
from the receiving end of the rail. It consists of the same mixture of
pearlite crystals as is found throughout the body of the rail. There
appears to be no unusual segregation of slag or foreign inclusions
to be found here.
Specimnens were taken from each of the two rails after annealing
a section of the head of each. The occurrence and distribution of in-
tei-rtiail areas of coaurely laminated pearlite are here more evident.
Both rails show considerable free cementite, which has coalesced as
boundaries of the grains as a result of the annealing process. Cement-
ite is the carbide of iron, FC, and is the hard, brittle constituent
of high-carbon steels (nlnenled) and cast irons. It occurs here, not
uniformly distributed throughout, but is more or less segregated. A
sample was taken from rail No. 1 after annealing just back of the
face of one of the transv'erse fissures. The specimen shows consid-
erable quantities of free cementite. The estimated carbon content of
such spots is over 1 per cent. The amount of cementite found in
specimens from rail No. 1 was considerably more than in the samples
from rail No. 2. This, however, may be only fortuitous. Before an-
nealing, such free cemientite can not be detected with certainty be-
cause it exists as isolated particles which, during annealing, coalesce
to form the grain boundaries and also because the metal contains in-
clusions of various natures which, under the conditions, can not be dif-
ferentiated from the cementite particles with certainty.
The chemicall analysis indicates a carbon content of (0.82) or very
slightly different from the eutectoid composition(0.85). The occur-
rence of so much free cementite may be attributed to the restraining
action of the manganese content. The carbon is retained in the con-
dition normally characteristic of a steel of higher carbon content.
Such steel will have properties approximating those of the steel whose
true carbon content is equal to the apparent content of the steel under
Tensile samples which had been cut from rail No. 2 were submitted
after having received the heat treatment stated in the tabulation of
the tensile tests. The structure, so far as can be distinguished under
the microscope, is the same in all. They are all in the indefinite stage,
sorbite, preceding the revolution into pearlite, which takes place in
the critical range. The tempering temperatures were not chosen at
wide enough intervals to show any decided structural change.
The critical or recalescence point of this steel was determined to be
about 12500 F. The very coarse-grained structure of the rail is due
to rolling at a temperature very much higher than the critical range.


The weakening effect of .slig a:.I-,'i:;ted with
have been lessened by rolling and finishing at

coarse strlucture would
lower tempelat lrcs.

No. 5.-A. View of one of the coarse interstitial pearlite areas, rail No. 1: magnification 250 diame-
ters; etched with 2 per cent nitric acid. Such areas are very numerous throughout the body of the
rail and must exert an appreciable weakening effect upon the whole.
B. Section from head of rail No. 1, after annealing; magnification 250 diameters: etching, hot sodium
picrate. In the unannealed -. il'cni'r- the free cementite does not occur in the form of definite cell
boundaries. These are caused by the coalescing of smaller particles during the annnealing process.
The indefinite dark circular spots are the cut ends of the slag threads and not free cementite.
C. Photomicrograph from the base of rail No. 1, section parallel to the rolling; magnification 100
diameters; etching, 2 per cent alcohol solution of nitric acid. The structure is a jumbled mass of
pearlite crystals. The long dark streaks are slag threads. There are coarse interstitial pearlite areas.
(The phutomiro i-rulphs of the head and web showed similar structure.)
D. Photomicrograph of tensile specimen from head of rail No. 2, after heat treatment, quenched in
oil from temperature of 1,4000 F. and drawn at 1,250 F. Magnification 250 diameters.
All five heat-treated specimens have sorbitic structure.


1. No unusual segregation of impurities is found.
2. The structure throughout the section of the rail is very uniform.


3. The metal inlllnediately. adjacent to the "transverse fissures"
appears to be of the same nature as throughout the rest of the rail.
4. No unusual segregation of impurities or slag can be associated
with the transverse fissures; slag streaks," however, are present here
in the usual numbers as are found throughout the rest of the head.
5. The occurrence of the interstitial areas of very coarse pearlite
must have an appreciable weakening effect upon the metal as a whole.
6. The occurrence of free cementite, which in itself is a weak, brittle
constituent, is a serious defect and may have a direct bearing on the
formation of the "tra nsverse fissures developed in rails of such com-
Tensile tests.


Description. Diameter. Sectional Elastic Tensile Eonga- on remarks.
area, limit, strength. tion. area.ion of Remarks.

Pounds Pounds
Inches. Sq. in. per sq. in. per sq. in. Per cent. Per cent.
Head of rail near initial frac- 1.129 1.00 .......... 60,000 () (1) Granular.
Head of rail, outside half from 1.129 1.00 60,000 78,600 (1) (1) Do.
west end.
Do...................... 1.129 1.Q00 55,000 77,300 (1) (1) Do.
Head of rail, from west end, .505 .20 60,000 135,500 4.5 4.5 Do.
near center of head.
Head of rail, from west end, .505 .20 65,000 91,000 (1) (1) Do.
outside of head.
Web of rail, from west end ... .505 .20 65,000 145,800 11.0 17.0 Do.
Do...................... .505 .20 65,000 145,500 12.0 17.0 Do.
Base of rail, from west end.... .505 .20 65,000 146,300 10.0 9.0 Do.
Do..................... .505 .20 65,000 148,700 11.0 15.0 Do.
Do...................... .505 .20 65,000 146,000 11.0 15.0 Do.
Do....................... .505 .20 65,000 142,000 10.0. 14.5 Do.


Head of rail, middle of....... 1.129 1.00 .......... 49,800 (1) (1) Granular.
Gauge side of head........... .505 .20 65,000 144,400 9.0 12.5 Do.
Center of head................ .505 .20 65,000 77,500 1.0 .5 Do.
Outside of head.............. .505 .20 70,000 140,800 3.0 3.5 Do.
Web of rail................... .505 .20 65,000 147,000 10.5 13.5 Do.
Base of rail................... .505 .20 70,000 149,000 11.0 15.5 Do.
Do...................... .505 .20 70,000 150,400 10.0 15.0 Do.
Head of rail, middle of, an- 1. 128 1.00 60,000 66,300 (1) (1) (2)
Head forged down and an- 1.129 1.00 .......... 95,400 1.6 (1) (a)
Head forged down to 1-inch
diameter, annealed, re-
heated to 1,400* F., and
'1iiunchplr oi il,thendrawn
as described:
Drawn at 1,000 F........ .505 .20 105,000 150,800 15.5 37.5 Silky.
1,080 F........ .505 .20 90,000 148,000 17.0 42.0 Do.
1,1500 F........ .505 .20 95,000 148,900 17.0 41.0 Do.
1,2000 F....... .505 .20 95,000 150,850 18.0 38.0 Do.
1,2500 F........ .505 .20 100,000 144.000 16.0 37.0 Do.

1 Inappreciable. 2 Granular blue-black spot I by inch. 3 Granular; broke in head at root of thread.


Drop tests were made at the works of the Maryland Steel Co. on
one piece of rail No. 1 and two pieces of rail No. 2. Prior to these
tests the second rail was broken in several places by bending loads
applied in the testing machine at the Bureau of Stanidards, but none
of the frnctiured surfaces showed transverse fi~lures. The drop test-
were made with the rail sect ions head up, supports 3 feet apart, 2,000
pounds tup, height of fall 15 feet. The rails were tested at a tem-
perature of 90 degrees.
Rail No. 1 sustained the first blow without rupture, showing a de-
flection of 0.84 inch. It broke on the second blow, with an extension
of the metal of 6 per cent. The sections from the second rail each
broke on the first blow, neither developing appreciable extension.
The appearance of these pieces is shown by figure No. 6. The upper
rail in the cut represents No. 1. The middle and lower sections were
those from rail No. 2.
It will be noted that rail No. 1, which failed in the track and caused
the derailment of train No. 26, successfully passed the prescribed drop
test, displaying an extension of 6 per cent as required, while the sec-
ond rail, which failed to meet the drop test, did not fail in the track.
The fracture near the bolt holes of one section of rail No. 2, was
secondary, occurring when this fragment struck the bed of the drop
testing machine succeeding the blow of the tup.
The section of rail No. 2, represented by the lower figure of the cut,
did not fracture under the place directly struck by the tup, but
sheared out a fragment 22 inches long, symmetrical with the supports.
In prescribing only 6 per cent extension of the metal under the
drop test, as an index of the ultimate ductility of the steel, it may be
said that such extension, developed as it is under transverse stress, is
not far above the zero limit of ductility. In the milder grades of
steel the extension under transverse stresses commonly exceeds many
times that witnessed in the tensile tests of the metal. In these two
rails the reverse was true. The 2-inch tensile specimens from the
bases of these rails showed 10 and 11 per cent extension, which, equated
for specimens of greater length of uniform section, would still be
more than the extension in the drop test of rail No. 1, which was
6 per cent, while No. 2 failed with zero extension.
The mechanical work required to rupture steels of high elastic limit
but incapable of permanent set is small compared with the work re-
quired to rupture mild steels which display the extension usual in
structural steels. If the indications of the present tests are confirmed
and it is found that rails normally of limited extension tend to fail
under the effect of rapidly applied loads without appreciable set,










i I I I I c



I 01


i E'0





i -a I
~-: I I 6
s vz


this feature in the use of hard steels will demand early\ considera-
tion, in which striking velocities and temperatures of the rail should
be included.
The striking velocity of the tup with 15 feet height of drop, as
prescribed in current specifications, is a low velocity com'ipareil with
ordinary train speeds.

The New York, New Haven & Hartford Railroad Co. conducted an
examination, which comprised tensile tests, chemical analyses, mietal-
lographic examination, and drop tests, the later at the works. of the
Bethlehem Steel Co., other parts at its New Haven laboraory. In
addition to supplying data confirming the information from other
sources, this examination resulted in showing the presence of a con-
siderable number of incipient fissures in the head of rail No. 1. at
places where the ordinary manifestations of transverse fissures were
not in evidence. That is, the fissures had not reached the advanced
stage in which they would ordinarily be detected by visual inspection.
Deep etching, with nitric and hydrochloric acids, developed short
transverse cracks or incipient fissures which were thus rendered
plainly visible to the eye. They ranged in length from a few hun-
dredths of an inch to three-eighths inch. So far as could be judged
the zone of greatest structural disturbance was in the head over the
gauge side of the webl and toward the gauge side. The etching being
very deep not unlikely brought into view fissures which originally
were less easily discerned.
A group of these incipient fissures is shown in figure No. 7, (a) and
(b). The fissures are here shown about natural size, the same group
appearing in both (a) and (b). In the latter they are partially ob-
literated by rough polishing the surface while removing some longi-
tudinal scratches and introducing others. The location of the zone in
which these fissures were found leads to the inference that we are here
examining the same class of phenomenon witnessed in the larger
transverse fissures, which ultimately result in the complete fracture
of the rail. If such is the case, structurally, a more general disinte-
grating effect has been brought about than indicated by the display of
transver.,e fissures in the rail-that is, influences which tend toward
the formation of transverse fissures are not localized at those places
only where fissures have reached an advanced stage of development.




o w



f a


Other rails were taken from service by the New York, New Haven
& Hartford Railroad Co., and used in this examination. One, an E
rail, designated as No. 10, was taken from the track under the sup-
position that it belonged to the same heat of steel as No. 1, which
caused the derailment. Analysis, however, showed that it came from
another heat. Check analysis at the steel works showed the follow-
ing composition:
Carbon, 0.83. Manganese, 0.49. PhorpdIrui. 0.020. Sulphur, 0.052.
This rail did not break, as others had, when lifted with a magnet
and dropped bodily from a height of 9 feet. When ruptured under
the regular drop test, head up, three breaks were made, each of
which were reported as having shown clean metal, free from trans-
verse fissures. A test of the metal from the gauge side of the head
showed a tensile strength of 122,000 pounds per squad re inch, with a
contraction of area of 20 per cent.
The Bethlehem Steel Co. in reporting upon the examination of the
material pertaining to this inquiry state:
As in all other cases of rails developing transverse fissures in the head, no
segregation of any kind was found in the rails to account for this defect. There
was no difference found between the microstructure of the core of the fissures,
the bright parts of the fissure, or any other part of the rail. All fissures
occurred on the gauge side of the head of the rails.
A rail rolled in the same year and month and branded the same
as rail No. 1, but which had not been used in the track, was taken
from a spare rail post and cut up for examination. The chemical
composition of this rail was found to be:
Carbon, 0.84. Matn gainie. 0.87. Phosphorus, 0.037. Sulphur, 0.025.
A tensile specimen of 1 square inch sectional area and 10 inches
long, takeni from, the center of the head, gave the following results:
ElI.-tic limit, 60,000 pounds per square inch.
Tensile strength, 148,000 pounds per square inch.
Elongation, 6.3 per cent.
Contraction of area, 6.9 per cent.
Appearance of fracture, granular.

Sections of several rails included in this inquiry were cut out for
photographic purposes, metallographic examination and check de-
termination of the carbon at the Washington Navy Yard. Metallo-
graphic examination was made by Mr. Wirt Tassin, with the assist-
ance of Mr. Paul E. McKinney.



The material examined included two specimens identified as
" Westerly Rail No. 1, which failed in the track," with five specimens
identified as Rail No. 10, which had been in service, but did not fail
in the track."
All work done was metallographic.


Figures 8a, 8b, and 8c show macroscopic transverse fissures in the
rail head, as seen at a magnification of 8, on a specimen taken from
the west end" of the rail. They are plainly visible to the unaided
eye and are not associated with sing, sulphide, or oxide areas, nor
are they accompanied by segregations of any kind.

8b 8c
No. 8.-Macroscopic transverse fissures in the head of rail No. 1, near its west end. Magnifi-
cation 8 diameters. Fissures are not associated with slag, sulphide, or oxide areas, nor
accompanied by segregations of any kind.


Figure 9a, at a magnification of 315, using a B. and L. 8 mm.
objective and a 15x eyepiece, shi,\ a transveris micr ,s'-,pic fissu'r
as seen on a longitudinal section cut from the ea-t end" of the
rail and located at about the upper center of the head. The fissure
could be readily traced for a di.taii-'c of 4 mun., 0.15748 inch.
Figure 9b, at a similar magnification, is another traniiver- fisiire
whose length could be traced for 11 mm., 0.43307 inch.
Figure 9c, at a magnification of 315, shows a typii-al area of
incipient fissuires.

9b 9c
No. 9. Microscopic transverse fissures from head of rail No. 1, near its east end; magnification 315
9a was traced for a length of 0.157 inch.
9b for a length of 0.433 inch.
9c shows a typical area of incipient fissures. These fissures are not associated with and have no rela- any areas of sulphide, slag, or other inclusions.


In each instance it will be noted that these fissures are not asso-
ciated with and have no relation to any areas of sulphide, slag, and
other inclusions. This is further shown in figures 10a, 10b, and 10c,
which are at the same magnification and show seams of such inclu-
sions with a complete freedom from microscopic or incil)ient. fissures.
The general structure of the rail is sorbitic. There is little or
no lamellar pearlite, no free cementite or ferrite. The sulphide, slag,
and oxide areas are small and sparingly distributed. There are no
seglrega t ions.

10b 10l
No. 10. Showing slag and sulphide seams in head of rail No. 1. Magnification 315 diameters. At these
seams there was complete freedom from microscopic or incipient fissures.



Figure 11 is a -ketili showing the location in the rail of the sections

Gauge side

r ecf/ ron-A-
f/Ture //
/rrcyu/ar oS/mies /'ndica/j f/'ssuredArLea.

No. 11.-Cross section of rail No. 10, showing manner of cutting up for examination, and location of
fissured areas found therein.


Figure 12a, magnification 315, shows incipient transverse fissures
in the center of the head of section A.
FiguI re 12b, magnification and real field as in figure 12a, is from
the lower center of the head of section A and shows incipient fissures.

12b .
No. 12. Incipient fissures found in the head of rail No. 10.
12a rtpre'Lntl fissures found in the center of the head, in section of rail marked "A.'
12b, fissures found in the lower center of the head. Magnification, 315 diameters.


Figures 13a and 13b, magnification 315, show the smallness of the
seams of slag and sulphide as follun in section A.

No. 13. From head of rail No. 10, section marked "A," showing the smallness of the seams of slag and
sulphide. .Magnitii alian 315 diameters.


Figures 14a, 14b, and 14c, magnification 315 diameters, show typical
incipient fissures as found in specimen B. Figures 15a and 15b show
analogous fields in -ec(tioin D.

14b 14C

No. 14. Typical incipient fissures found in head of rail No. 10, section marked "B." Magnification,
315 diameters.

Figure 15c shows one of the largest of the sulphide-slag areas and
illustrates the smallness of these inclusions.
Specimens C and E (see fig. 11) show no incipient fissures and no
abnormalities of structure.
Plotting the fissured areas as seen in figure 11, it will be noted
that they are practically limited to the gauge side at or near the


center of the head and are in the regions immediately affected by
the wheel loads.

15b 15c
No. 15. Typical incipient fissures found in the head of rail No. 10, section marked "D," represented by
15a and 15b. One of the largest sulphide-slag areas of rail No. 10, lII1l-ii.ltmn the smallness of these inclu-
sions is represented by 15c. MI'niti' .,riln. 315 diameters.

The opinion is advanced that these fissures are set up by these
loads, and in support of this the statement is made that new rails
which had not been in the track and which have as high and even
higher carbon content do not show under the microscope similar
fissured areas.

The Westerly rail No. 1 shows nothing in its structure to indicate
any abnormalities in the mill practice. The rail is comparatively free
from slag, sulphide, and oxide areas. None of the areas showing


macroscopic or microscopic fissures can be correlated with areas con-
taining sulphide, slag, oxide, or segregations.
Rail No. 10 shows nothing in its structure to indicate any abnor-
malities in the mill practice. Slag, sulphide, and oxide areas are
very sparingly distributed and are relatively small. Numerous in-
cipilent fi-slures are present and are located in the head on the gauge
The general structure of the two rails is such as to warrant the
statement that they are of a carbon content that will not afford a
toughness and ductility comparable with that of a properly treated
rail of a lower carbon content, hence the transverse fissures.
To remedy this condition two courses are open-fix the upper
carbon limit and prescribe the mill treatment which will insure the
maximum toughlne.-s and ductility with a sufficient strength, or reduce
the wheel loads. The former plan is to be preferred.

The metallographic examination by Mr. Tassin completed the
present examination of the material pertaining to this derailment.
Concerning the pre.'alelnce of transverse fissures in steel rails, of a
macroscopic order, not referring at this time to those microscopic, of
such dimensions as are mennicing to the safety of railway travel, they
are believed to be numerous. Forty-six transverse fissures of recent
occurrence have been reported in 32 rails. Instances have been re-
ported in which five transverse fissures have been found in the same
rail within the limits of 3 feet. They are present in both open-hearth
and Bessemer steels. They are not confined to the product of the rail
mills of one section of the country. They occur over ties and between
ties, near the ends of the rails and along the middle of their lengths,
on tangents and in the upper as well as the lower rails on curves.
But one general remark can be made-they persistently appear on the
gauge side of the head.
In their maximum state of development they have been witnessed
in 100-pound rails having attained a superficial area of 3.3 square
inches, leaving practically only the web and the base intact. Our
investigation shows without question that these hidden fissures in
some rails reach such a state of development before discovery as to
destroy nearly the entire head of the rail, therefore it is not reassulr-
ing that other rails of similar composition, working under similar
conditions of service, are not free from these interior defects. The
continuance of conditions which have resulted in derailments, at-
tended with loss of life and injury to passengers and employees,
places a great responsibility upon all who can in any manner aid
in the inauguration of measures which will tend immediately to
ameliorate these grave conditions.


Reference has been mn;ald( in ':irlier deraiilmnent reports to causes
which are believed to be contributory to the formation and develop-
ment of transver-e fissures. Data have since been gathered, some of
which are embodied herein, illustrating the prlobale sequence, of
events which attend their formation, from the results of which the
contrilbutory cal.-es more clearly admit of re'c ignition. Tr;iinwg the
fissures in the reverse order of their development, they are follow e
down from the larger ones of rails broken in the tru ;k having dark-
ened oxidized surfaces to those of an earlier stage of development, of
smaller diameter with bright silvery surfaces, which have not yet
extended to the outer surface of the rail and therefore retain the
brightness of freshly fractured surfaces.
Next earlier appear fine cracks, not easily discernible to the eye
until the steel has been etched, preceded by other cracks still finer of
a microscopic order, but of measurable length to the eye if the
cracks were of sufficient width to be seen, eventually leading to the
detection of microfissures and the fragmentary or partial separation
of the microconstituents. These indications follow each other in
such order that they appeal r to belong together, representing different
stages in the loss of structural integrity and destruction of the metal
of the rail.
Such are the indications which pertain to rails which have been
in service. They have been looked for, but not found, in steel which
has not been in service, of similar composition and concurrent manu-
facture. As the evidence stands the formation of such fissures seems
attributable to track conditions and to those stresses which reach a
maximum in the metal of the rail at the head.
Steel of the composition of these rails in inherently a strong steel.
Music wire is drawn from steel of 0.85 carbon and in the form of
fine wire displays great strength. Records of such wire show a
tensile strength exceeding 4.-0,000 pounds per square inch. This
grade of steel, in the form of hot-rolled, cylindrical bars, also, will
endure repeated stresses applied many million times, as high as
40,000 pounds per square inch. Furthermore, in these rails the steel
has displayed, in specimens from unaffected par ts of the cross section,
tensile strength in round numbers ranging from 140,000 to 150,000
pounds per square inch. The strength possessed by steel of this
carbon content shows that in dealing with rails of such composition
we have a metal which in its original state has high physical
The spontaneous rupture of steels from internal causes occasion-
ally occurs with metal hardened by heating and quenching. Intense
internal strains are set up by sudden quenching which, if not amelio-
rated by drawing the temper, may occasion spontaneous rupture.
Coils of hard-drawn wire have shown fractured ends developed a few


hours after drawing. Castings also have failed spontaneously.
These examples. however, refer to other classes of material than rails,
while internal tendencies, if any of special account exist in steel
rails due to casting or rolling conditions, have not so far as known
materially contributed toward rupture. Internal strains are left in
steel rails as in all hot rolled and naturally cooled steel, modified
by shape, weight of section, and rate of cooling, and higher strains
may exist in hard steels over those of lower grades in composition.
These features are all known to exist, and while in a strict sense they
are not negligible factors in the use of steels, nevertheless so far as
pertains to the formation of transverse fissures in steel rails they
are not held to be vital.
In reference to the relative strength of steel at the interior and
exterior of the shapes in the different passes of the bloom and
rail mills: The metal at the interior has been found lower in tensile
strength than that near the surface, in the early passes. However,
when the finished rail is reached these differences are inconspicuous,
but with the thinner sections of the web and base commonly showing
higher strength than the heavier section of the head. In this inquiry
the strength of metal at the center of the head of an unused rail, one
taken from a spare rail post, was found to be 148,000 pounds per
square inch. This indicated normal strength, showing no inherent
weakness in the metal of that portion of the head in which the weak-
ened steel of rails, taken from the track, has been found.
Gagging is an operation common in all rail mills. Necessarily
the elastic limit of the metal must be exceeded in this operation,
otherwise no permanent straightening would result, and since the
elastic limit must be exceeded, it follows that the higher the elastic
limit of the steel the greater will be the overstraining force required
to effect the straightening.
The actual bending force required for the purpose obviously may
be modified by regulating the distance between the supports of the
gagging press, but the resultant longitudinal strain in the rail must
in any event be sufficient to balance the internal strains which tend
to cause the rail to return to its original bent shape. These internal
strains will be greater in steels of high physical properties than in
mild grades of metal. In prescribing hard steel for rails these
severe internal conditions will necessarily be encountered.
Significance would attach to these several features in searching
for causes leading to the development of transverse fissures if inti-
mate relations were found to exist between them and such defects.
While not entirely disassociated, their influence and mutual rela-
tions seem remotely connected. Evidence connecting the formation
of transverse fissures with these primitive conditions is less direct
than that which attaches to track conditions.


In reviewing the results of this inquiry it will be noted that the
tensile strength of a rail of substantially the grade of metal repre-
sented in rail No. 1 was 148,000 pounds per square inch. This
strength was found at the center of the head of a rail which had not
been in service, but one which was rolled in the same year and month
and by the same miunfiif;tiurer as rail No. 1. Microscopically, no
fissures were found in this unused rail.
Rail No. 1 and its companion No. 2 showed substantially the same
strength in each of their webs and bases, closely approached by the
metal in some parts of their heads. Metal from the head of No. 1,
forged down and heat treated, gave correspondingly high results. The
normal strength of the steel, as indicated in these tests, ranged from
140,000 to 150,000 pounds per square inch. In general, however, the
metal in the. heads of both rails, taken in the condition the rail came
from the track, showed impaired strength. A tensile specimen from
the head of rail No. 1 fractured at 60,000 pounds per square inch,
while a corresponding specimen from No. 2 failed at 49,800 pounds
per square inch. An annealed specimen from No. 2 had a tensile
strength of 66,300 pounds per square inch. The latter ruptured at
a preexisting defect, an interior fissure of small size, which was
shown by an oxidize'd spot at the circumference of the test piece.
Tests conducted by the New Haven Railroad on these features of the
case furnished corroborative results. The metal in the head of each
rail was in a weakened condition.
The metallographic examination of No. 1 showed the presence of
incipient fissures and a condition of structural unsoundness existing
in different parts of the length of the rail. This unsoundness affected
the metal in the center of the head and on the gauge side. A gener-
ally shattered state characterized the metal of certain parts of the
head, while in other parts of the cross section of the rail the steel was
Although rail No. 1 had developed transverse fissures and had
broken in the track, while No. 2 had not experienced such extreme
vicissitudes, nevertheless the weakened condition of the metal in the
head of each was about the same. Had rail No. 2 been kept in the
track it is believed that it would eventually have developed the same
kind of fissures as shown by No. 1. The zones in which this state of
weakness was found where located in the same places in each rail and
closely coincided with the zone of disturbed metal which had been ex-
perimentally made in reciprocating tests of rails under high wheel
Another rail of this series was microscopically examined, one taken
from the track after the same period of service as rail No. 1, from
which it differed in having lower phosphorus and manganese. Its


phto.plio lorus content was 0.020; manganese, 0.49. This rail showed
similar incipient fissures to those of No. 1.
Loss in tensile strength, the display of macroscopic and microscopic
fissures, and the development of transverse fissures, all seem to be
ass.ciaited phenomena, and from the location of the affected metal
seem traceable to the action of wheel loads for their development.
No other cause is recognized as being present to which their formation
could be ascribed.
The general condition of the steel in these tests outside of the
affected zone was good and possessed of normal strength. The me-
tallographic structure was the same at the fissures as in other por-
tions of the rail. There were no slag inclusions or indications of the
presence of slag at the transverse fissures where the rail fractured,
nor was slag found at the microfissures, of which there were many
examples in the different rails. Slag inclusions did not, so far as can
be ascertained, locate either the places of incipient fissures or the
larger transverse fissures of the broken rail. Since such inclusions
would be aciculnir in shape, in size not materially detracting from the
sectional area of the steel, and drawn out parallel to the length of the
rail, they would not be expected to exert any appreciable influence on
the strength of the rail against longitudinal stresses. A number of
slag streaks were specially examined for the purpose of ascertaining
whether incipient fissures had their origin at such streaks, but no
fissures were found associated with them. Metallographically nor-
mal mill practice is indicated.
In conclusion it appears-
That the derailment of train No. 26 was due to a broken rail.
That the rail fractured under this train by reason of the presence
of transverse fissures in its head, one of which was located 6 feet 7
inches from the leaving end and is believed to have been the place of
initial fracture.
That five transverse fii.sures were found in the rail ranging in diam-
eter from five-eighths inch to 11 inches.
That fissures of lesser extent were present in both this rail and a
companion rail of the same heat, each of which had had the same term
of service in the track.
That the metal in the heads of these rails was in a weakened state
in certain affected zones, the affected zones being located near the
center of the hend and toward the gauge side.
That the weakened and structurally impaired metal shown in the
tensile tests was confirmed by tht metallographic examination.
That the steel in other parts of the rail was structurally sound and
possessed of normal strength.
That no slag inclusions were associated with the transverse fissures
nor with the microfissures.


That fissures were present in different stages of developmentiit,
associated with each other apparently as to a common, cauiist, the
inicroscopic examination indicating that such fissures were located
in metal otherwise structurally sound.
That microscopic fissures were present in certain other used rails,
which had not yet developed full-.ized tr;insverse fissu-res, furnishing
evidence that such rails were approaching a state in which full sized
fissures would eventually be formed.
That the proximl;ate causes to which the transverse fissurei in
broken rail No. 1 are ascribed are high wheel loads with their attend-
ing strains, evidence of other causes not having been found.
That testimony is to the effect that a considerable number of rail
failures have recently occurred by reason of the presencll of transverse
That conditions which were precursors to the formation of trans-
verse fissures in broken rail No. 1 exist in other rails now in service.
That the presence in the track and continued use of rails of the
same or similar composition to rail No. 1 and exposed to the same
service conditions is a source of danger.
That evidence acquired indicates that transverse fissures may be
and are formed through the action of high wheel loads upon hard
steel rails.

It is manifestly evident from the above report and conclusions of
Mr. Howard, which are concurred in, that in this type of rail failures
there is presented a serious situation. Rail failures of other types
have been the cause of many accidents.
The figures contained in the following table are taken from the
monthly accident reports made to the Commission by the railroads-
and show the number of derailments caused by broken rails which
have occurred yearly since July 1, 1901, together with the (caii';ltic'
and monetary loss resulting therefrom:

Number including
Year. of acci- Killed. Injured. cost of
dents. el:-jrirn

1902.......... 78 5 207 $128,769
1903.......... 150 12 204 166,140
1904.......... 176 9 139 157,682
1905......... 201 4 465 257,519
1906......... 220 7 635 25 1;2
1907.......... 308 12 699 t1, 7.7
1908.......... 238 16 433 296,327
1909.......... 196 5 498 191,842
1910.......... 243 24 369 293,899
1911.......... 249 12 463 292,749
1912.......... 363 52 1,065 511,778
1913.......... 340 17 827 401,551
Total.. 2,762 175 6,004 3,237,793

1 INIii i11111 II l UlhII III ulIIII IIlli IIl
32 INTERSTATE COMMERCE CON 3 1262 08856 1989

This enumeration of casualties refers to a period the greater part
of which elapsed before rail failures by transverse fissures were
known or had become so prevalent as they now appear. The tabu-
lated results, however, emphasize the importance which the subject
of safety in rails has assumed. To those elements of danger existing
in the past is now added this type of failure shown in the develop-
ment of interior fissures. On account of the insidious character of
these fissures and the fact that they are progressive in their devel-
opment, and so far as is known no system of inspection has been
found that will detect them until they have reached the surface of the
rail, make it extremely difficult to suggest any preventive against
future accidents of this character.
Although as noted in previous reports dealing with rails failing on
account of transverse fissures, it seems apparent that a remedy lies in
the diminishing of wheel pressures and the lowering of direct com-
pressive bending and shearing stresses.
From the constant increase in rail breakages occurring from this
new type of failure, it would appear that the danger zone in the
use of steel rails as at present manufactured has been reached, since
the study of the rails here under discussion appears to indicate that
transverse fissures are the direct result of high-wheel pressures act-
ing upon hard steel. A complete investigation of rail, track, and
wheel load conditions for the purpose of determining the effect
thereon of the recent types of locomotives and cars, with their greatly
increased wheel loads, should be undertaken for the purpose of sci-
entifically determining this matter and ascertaining a remedy.
Respectfully submitted.
Chief Inspector of Safety Appliances.



Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EN2AZ59W9_FSNR2N INGEST_TIME 2012-10-23T14:45:34Z PACKAGE AA00012211_00001