Welded rail

Discussion relating to the D&H. For more information, please visit the Bridge Line Historical Society.

Moderator: MEC407

Welded rail

Postby Allen Hazen » Sat Nov 22, 2008 11:10 pm

My recollection is that the first installation of continuous welded rail in the U.S. was an experimental quarter mile somewhere on the D&H main line. Does anyone here remember when it was put in? Was this another Loree-era experiment, more successful in the long run than the high-pressure steam locomotives, or was it later?
Allen Hazen
Posts: 2424
Joined: Fri Mar 12, 2004 10:14 pm
Location: Edmonton, Canada (formerly Melbourne, Australia)

Re: Welded rail

Postby march hare » Tue Nov 25, 2008 3:52 pm

Try the Wikipedia listing for "welded rail". It has a couple of historical references documenting the D&H role in the 1930s.
march hare
Posts: 361
Joined: Thu Mar 25, 2004 12:44 pm
Location: upstate NY

Re: Welded rail

Postby ChiefTroll » Sun Nov 30, 2008 6:10 am

This is strictly from memory (second hand, of course). The first installation of continuous welded rail on a steam railroad main line in the US was on the D&H behind the General Office Building at Albany in 1932, during the Loree administration. It was about one quarter mile of 130 RE rail. A few years later, 1934 if my memory serves me well, they made a much larger installation, of 131 RE rail, on the Schenectady Main when the relocated line was built. I was told by some folks who worked on that that they had installed expansion joints to relieve thermal stresses, but the joints never moved and so they were taken out, with no adverse consequences. When I was the Track Supervisor at Plattsburgh in the late 1960's we had several long stretches of 131 RE welded rail, including most of the track through Willsboro Rocks, laid in 1937.

A major obstacle to the adoption of CWR was the matter of anchoring the rail to prevent longitudinal movement. Up to 1932, rail anchors were not very strong or reliable, but the D&H had begun to install a different type of rail fastening called "M&L Construction." The L was for Leonor Loree, the president, and I think the M was for Mulholland, one of his collaborators on that and other things, including the M&L tender booster engine for steam locomotives.

M&L Construction uses two spring clips on each tie plate to clamp the rail to the plate, and restrain it by friction from moving longitudinally. The plates are fastened to the tie by two, three or four lag bolts screwed into pre-drilled holes. The number of lag bolts is determined by the curvature of the track. That system works well, and it permitted the D&H to install the first CWR in the US.

After Loree left the D&H in 1937, they went back to conventional jointed rail until the late 1950's, when they started single-tracking segments of double track. The relayer rail released in those projects was welded at local semi-portable welding plants (I remember one at Whitehall in 1959) with the joint sections cropped off. The earlier CWR used gas welds, but in the 1950's they were using electric flash-butt welding.

Gordon Davids
Posts: 363
Joined: Sat Mar 13, 2004 6:51 pm
Location: Severna Park, MD

Re: Welded rail

Postby Steve Wagner » Thu Dec 11, 2008 3:58 pm

Gentlemen, I will attempt to reproduce here something I typed up for the Bridge Line Historical Society Bulletin (the monthly publication of the active group of Delaware & Hudson fans) a few years ago, including a short passage from Trains and a much longer piece prepared by a D&H official in 1938:

A comment about our favorite railroad’s pioneering in the use of welded rail by Steve Wagner in the October 2002 Bulletin prompted member Pat Patrick to send in what follows.

A feature on the most important railroad developments of the Twentieth Century in the January 2000 Trains included a photo of welded rail being unloaded from a freight car with this caption:

WELDED RAIL. The sound of “clickety-clack” might be romantic, but it’s also expensive. Steel wheels pounding on bolted joints of 39-foot track sections wreak havoc on the maintenance-of-way budget. That’s why railroads embraced continuous welded rail, or CWR. First employed by street railways in the 1920’s, it made its debut in heavy-duty trunkline service in 1937 along a 30-mile section of Delaware & Hudson in upstate New York. CWR paid immediate dividends, cutting maintenance costs even as it smoothed out the ride. The technology took off in the 1970’s to the point where, today, railroads proudly report that 104,000 miles of track are CWR.

And Railway & Locomotive Historical Society Bulletin No. 49 (1939) contained the following article. [Some punctuation has been changed in the interest of clarity, but unusual spellings have been left alone. For instance, what today is called thermite is referred to as Thermit.]

Continuous Rail

Illustrated lecture by H.S. Clarke, Engineer, Maintenance of Way, Delaware & Hudson Railroad, before the New York Chapter of the RAILWAY AND LOCOMOTIVE HISTORICAL SOCIETY in the Engineering Societies Building, 29 West 39th Street, New York, on Nov. 12, 1938.

Few of us realize how true it is that a railroad is never completed. Spectacular fast trains, mighty locomotives, and air conditioning are only steps in a process of improvement that is continually going forward.

So, with permanent way, it is necessary, in order to meet these ever increasing demands for higher speeds and heavier loads, under which the cost of track maintenance must rise, to devise more comprehensive measures of track construction and maintenance methods than those which we all used in the late past. Such improvements to the standard of track construction must be radical and far reaching.

There is an old axiom that one must spend money to save money, just as one must spend money to make money, and the railroads have been looking around for the places to spend money to save money. On the Delaware and Hudson Railroad, we have been singularly fortunate in having a management who recognized this trend for some years, and have provided us with every opportunity and backing to develop a track structure and maintenance methods to meet these conditions.

With labor the largest item in railroad maintenance, there are opportunities for spending money to save labor, reducing maintenance costs, not necessarily during the current year, but of large importance in the long run, over a period of years.

Like most railroads, we have been carrying on a program of improved drainage, ballasting with ballast of suitable type and to an adequate depth, installing hardwood creosoted ties and heavy rail, and we have continued to carry on these programmes on a somewhat reduced scale throughout the depression. With the exception of steel ties, most railroads have been using these means to modernize their permanent way.

The steel tie used on the Delaware and Hudson Railroad is made by welding a standard tie plate on two eight-foot lengths of scrap rail, the welding being done with an automatic electric welding machine developed by the General Electric Company. We have, at present, some 165,000 of these ties in yards, sidings and, to some extent, on branch lines, their use being limited to where we have no signal circuits.

However, to meet the problem of better track, which may be more economically maintained, these steps, it was thought, would not be sufficient.

By preventing decay, the creosoting of ties lengthens their life; however, the treatment is expensive and, to obtain full advantage of the investment, the tie must be protected against mechanical wear and spike killing. The rail joint is universally recognized as the weakest point in the track and it is recognized that over 45 per cent of track labor is now necessary in order to keep the joint in proper line and surface. More rails are renewed yearly and have their service life shortened by reason of battered and chipped ends than for any other cause, thus providing a major problem.

Here were two problems well worth considering.

To meet the first condition, a heavy double-shouldered tie plate was developed, canted and crowned, fastened to the tie by means of a compression screw spike, independent of the rail fastenings. Thus the tie plate became, in effect, a part of the tie and prevented mechanical wear of the tie.

It also was held necessary to get rid of the old destructive track spike and its spike-killing effect on the tie; therefore, it was felt necessary to develop a fastener to fasten the rail to the tie plate independently of the screw spike holding the tie plate to the tie and, at the same time, eliminate the inefficient anti-creepers generally in use.

Such fastenings must met certain conditions:

They must stand being strongly tightened and remain thus in order to prevent the rails from creeping on the ties. They must have some elasticity in order to dampen the effect of the stresses developed in the track and the movement of the tie in the ballast resulting from impact. It must be possible, at any time, even after a number of years, further to tighten the fastening to take up wear or renew part or all of the component parts. Could such a fastening be developed that would be effective instantly it was applied to resist the movement of rail longitudinally in either direction against movement from creeping, and also prevent the rail expanding or contracting?

After many experiments with various types of clip fastenings, the spring clip fastening now used on the Delaware and Hudson Railroad was adopted and used with the double-shouldered plate previously described, punched to take the fastening.

With these fastenings, the rail is held by means of two spring steel clips, bolted to the tie plate at the center of the crown and bearing on the base of the rail. Since the rail-to-tie-plate fastenings are independent of the tie-plate-to-tie fastenings, each group is free to perform only the work it is designed to do, while the rail, the tie plate and the tie are co-ordinated into one strong but elastic and flexible unit.

The fastening of the rail to the tie plate and through that to the tie by means of the spring steel clips greatly strengthens the track structure, leaves it sufficiently elastic to absorb the impacts and loads applied to it and greatly decreases the vertical motion of the rail and of the tie over that of the standard track, although the normal wave motion of the rail is permitted to pass freely.

The decreasing of this vertical motion means a lessened pounding on the tamped ballast beneath the tie. It is this incessant pounding or pumping of the tie that breaks down the interlocked ballast, and causes the track to go out of line and surface.

Spring-clip constructed track stays in line and surface for longer periods than standard track, and the relative motion of the rail and tie remains substantially the same in all tests, which is extremely indicative of the uniformity of this type of track structure. The pressure exerted by the spring clip permits the free wave motion of the rail, but does not permit it to progress longitudinally. So well is the rail restrained that changes in length due to expansion and contraction are prevented. It was this feature of spring clip construction, combined with several other facts that led to the belief that it was possible and practical to weld rails into long lengths, lengths limited only by the necessity for insulated joints for signal circuits, and location of switches and crossovers.

There are two distinct factors which must be taken care of in standard track: one is the creeping of track, and the other is expansion or contraction of the rails due to temperature changes. Combined, there are many indications that these factors cause large temperature stresses in a great mileage of standard track and sometimes result in what is known as “sun kinks.”

The tendency to movement by creeping is caused by the running of trains. When the movement takes place, it usually extends over comparatively long stretches of track. The principal cause is the wave motion of the rail which is set up by the moving train. There is usually a slight upward and then a downward motion of the rail and ties just preceding a moving locomotive or train, owing to the flexibility of the rail.

The ground below also springs for quite a depth underneath the track and for some distance each side, so that there results a wave motion in the rail of much greater amplitude than at first appears. In thirty-nine foot rails or other short lengths, the propagation of the undulating wave motion is, more or less, arrested at every joint, causing each length of rail to move ahead. Hence running or creeping of rail takes place successively by rail lengths, one rail at a time; or, at most, by a few lengths acting together as one. With continuous rail, there are no joints to interrupt the wave motion of the rail; therefore, the tendency for rail to creep or run is eliminated.

Experiments conducted some years ago with a five-hundred-foot length of rail developed that, over a range of temperature changes between minus 20 degrees Fahrenheit and plus 20 degrees Fahrenheit, the length of the rail was unchangeable as between these limits, and no movement was observed, but above 20 degrees Fahrenheit, the rail was quite sensitive to temperature changes.

Where rail is so fixed that it cannot expand into an extended position when heated and cannot move in any other way, as, for instance, by buckling, it must remain unaltered in size since the external forces holding it are greater than the compressive stress set up by the temperature rise. Alternatively, if rail cools below the temperature at which it is fixed, a tensile stress is created to be resisted by the influence that fixes the rail; that is, the rail is compelled to undergo the stresses set up by expansion of the metal without lengthwise extension and vice versa.

These stresses are so far below the elastic limit of the rail itself that thy can have no effect upon it and, therefore, are of no consideration in connection with the subject of continuous rail other than representing a force that tends to lengthen or shorten the rail.

Summarizing, we reach the following conclusions:

(1) In long rails, the weight of the rail, in itself, reduces the theoretical amount of
expansion or contraction of a rail.

(2) The double-shouldered tie plate has a tendency to bind the rail and further to
decrease the amount of expansion or contraction.

(3) The holding power of the spring clips against longitudinal movement in either
direction combined with the above facts out-balances the forces of expansion and contraction, so that there is no lengthwise movement, except a slight movement at the end rails, not much more than would be expected in 39-foot rails.

(4) In continuous rail, rail creepage does not enter into the problem, as it does not
exist to any appreciable extent.

(5) The ranges of temperature changes that actually affect changes in the length of
rail are not the total temperatures of the climate in question, but only from approximately plus 20 degrees Fahrenheit to plus 130 degrees Fahrenheit, or the hottest day that may be encountered.

(6) In heavy rock ballast, little attention has to be given the temperature at which
the rail is fixed, as it can be fixed within a range of, say, 65 degrees Fahrenheit to 90 degrees Fahrenheit without fear.

(7) In cinder or gravel ballast, which is lighter and has less holding power than rock ballast, the rail should b fixed within a range of say, from plus 80 degrees Fahrenheit to 95 degrees Fahrenheit.

(8) In ordinary track, there are numerous cases of sun kinks, due generally to poorly anchored track which has been creeping and closed the expansion gaps.

(9) Sun kinks in ordinary track always result from causes which easily may be determined and corrected.

(10) There should be far less fear of sun kinks in continuous than in ordinary track, as creepage does not enter the problem and the cause of sun kink is better controlled, may be noted and corrected quickly.

These facts were tested out in continuous rail laid in Albany in 1933, the longest rail being 2,750 feet; this rail has been in service four years without giving us any trouble. Further installations were made at Mechanicsville [sic – SW] in 1934, at Schenectady and Windsor in 1935, and at Comstock and Cohoes in 1936 and 1937; this year, all new rail laid was being welded. At the end of this year, we had 450,000 lineal feet of welded rail in service, or approximately forty-five track miles of continuous track.

This introduction of continuous rail brought up an important and exact condition that had to be met: namely, a sound, high-quality weld. Extensive studies and research were carried out, relative to the many phases of welding theory and practice, in both Europe and [the] United States of America. Exhaustive tests were made under different procedures, with the co-operation of expert consultants, and have resulted in the development of satisfactory welds.

Test welds have been subjected continuously, day and night, to severe rolling load fatigue tests, following a similar procedure of the extensive rail investigations by the University of Illinois and considered the nearest approach to actual service conditions in track. Many other well-known rail tests were also made.


The Thermit pressure weld is made by casting a steel collar around the base and web of the rail and then forcing the rail ends together, making a pressure weld in the head of the rail. This is accomplished by using a material known as Thermit. This material consists of powdered aluminum and iron oxide. At a high temperature, the aluminum unites with the oxygen in the iron oxide and there results aluminum oxide slag and free iron. In making a weld, this reaction is started with ignition powder, and in about twenty seconds the reaction is complete and the temperature has been raised to about 5,000 degrees Fahrenheit.

At this time, the molten mass is poured into molds previously clamped around the rail ends. The steel or iron, being heavier than the aluminum oxide slag, flows in first and unites with the base and web of the rail ends while the aluminum oxide slag lies in the mold around the head of the rail and raises the temperature of the head to a welding temperature in three minutes, when the rail ends are forced together with clamps, making the pressure weld.


The rail ends are clamped in the welding machine, one in fixed position and one movable. Current clamps are connected to a low volt transformer. Short circuiting between the two faces preheats the rail ends. No preparation of faces other than cleaning or flux is needed. Continuous short circuiting burns off the high spots on the faces and builds up a pressure of gases between the rail ends. Air cannot gain access to the faces. There is, therefore, no oxidation of the faces, which remain chemically clean. About two minutes after the application of current, the faces are brought together under proper pressure. All molten metal is squeezed out and the weld is made. After the rail has been removed from the welder, it is slow-cooled or normalized in an oil furnace and then passes on to the grinders, where the excess metal squeezed out by the welding pressure is removed and the head of the rail properly surfaced.

On the Delaware and Hudson Railroad, it has been our practice to do the welding by either the Thermit or flash method on cars in the yard, welding the rail into 800 to 1500 foot lengths and then, when twelve strings have been completed, to haul the train load out and unload it where it is to be installed.

Continuous track has many advantages over track laid with the standard length rails, some of these being as follows:

(1) Joint maintenance is eliminated, including maintenance of bolts and angle bars, surfacing and lining.

(2) Rail life increased, joint batter ended, less frequent rail laying.

(3) Rail creepage eliminated.

(4) Deterioration of the joint ties eliminated and not necessary to space ties.

(5) Elimination or replacement of joint bars and reconditioning rail ends.

(6) Smoother riding for passengers.

(7) Rolling stock maintenance reduced; no shock at joints.

(8) Signal circuit bonding eliminated.

(9) Less tractive power is required to haul trains over jointless track.

It will be of interest to you to know that on one railroad using electric trains, the “nosing” of the trains has been very serious. In most cases where they had this “nosing” trouble, it would start immediately when the rail was renewed, and, in a very short time, there were indications of the rail being very badly cut. An installation of a mile of continuous track with spring clip construction, put in open track about a year ago, thus far does not indicate the slightest sign of “nosing” or damage to the rail from this cause.

It is also of interest to note that there have been several installations of continuous rail on other roads, both in the open and in tunnels, and it appears that in the next year there will be many more.

I had the pleasure of conducting an English engineer over our installations, and he advised me that in the London Tubes, which are much smaller in size than those in New York, the contour of the tube giving little clearance outside the cars, the noise was so great, it was impossible to talk to a person sitting beside you without shouting. In an endeavor to overcome this, they laid some test stretches of welded track and, in addition, placed some sound insulation around the lower sides of the tube, resulting in an eighty percent reduction in the noise.

Apparently there are many advantages of welded or continuous rail that, as yet, we have not realized, and more installations, under various conditions, will bring these out.

While it is a little early to make any reasonably accurate estimate of the reduction in maintenance expenditures resulting from continuous rail and spring clip construction, there is already obvious a substantial saving. We have track of this type, put up several years or more ago, over which there is a fairly heavy traffic, on which there has since been no work and which is in as good line and surface as when originally put up. This track gives every indication of staying in good line and surface for many more years with little, if any, attention. Railroad men who have been studying continuous rail feel that this development opens up the possibility of securing far stronger, safer and better riding track with a reduction in future maintenance expenditures of 50% or more, plus a now unascertainable saving in maintenance of rolling stock.

[Three photographs accompanied the article. “A stretch of Main Line on the D. & H.” shows an immaculate stretch of straight double track with light-colored ballast in a rural area. “Carrying the continuous rail to place of installation” is an interior view of wood gondolas with their ends lowered (or perhaps flatcars with wood sideboards) loaded with five lengths of welded rail. “Close-up of a Thermit Pressure Weld” seems to show a very conspicuous buildup of metal below the top of the welded rails.]
Steve Wagner
Posts: 216
Joined: Wed May 26, 2004 10:22 pm
Location: Massachusetts

Re: Welded rail

Postby ChiefTroll » Thu Dec 11, 2008 5:03 pm

H. C. Clarke probably knew as much about continuous welded rail (CWR) as anyone in the world at the time. He is the guy who did it. A few of his assumptions are a bit off in the light of our later experience, but they worked for the D&H because the track was strong in the first place. Buckling from thermal expansion is a serious problem with CWR today, as it would have been in 1938 if the D&H had not had a very stringent specification for the ballast section, particularly wide (18-inch) shoulders. CWR also runs in waves under a train, the same as jointed rail, but the spring clip fastening design on the D&H kept that longitudinal movement under control.

That is a very worthwhile and complete history of early CWR on the D&H. Thanks, Steve.

Gordon Davids
Posts: 363
Joined: Sat Mar 13, 2004 6:51 pm
Location: Severna Park, MD

Re: Welded rail

Postby ricebrianrice » Fri Dec 12, 2008 8:26 pm

So what happened to the new style clips for holding the rails to the plates, and fasteners for holding the tie plates to the ties?
Nothing is impossible for the person that does not have to do it!
Posts: 489
Joined: Tue Jul 18, 2006 12:30 pm
Location: Binghamton, NY

Re: Welded rail

Postby ChiefTroll » Fri Dec 12, 2008 9:21 pm

ricebrianrice wrote:So what happened to the new style clips for holding the rails to the plates, and fasteners for holding the tie plates to the ties?

In short, World War II. The D&H and all other railroads were limited to buying one of two standard rail sections - 112 RE and 131 RE. Where the D&H needed to replace rail, the War Production Board limited them to 112 RE rail, and they had to use single-shoulder re-punched tie plates from 90 RE rail. CWR was out of the question, so most of the new rail at that time was limited to replacing 90 RE rail on the Susquehanna Division and the south end of the Champlain Division.

When the D&H began use of heavier rail, 132 RE and 115 RE, they made the decision to use jointed rail and standard cut spike construction, for reasons not known to me but probably related to the expense of M&L plates and clips. At the same time, rail anchor designs were improved so that several clip-on anchors became available. Those could be clamped onto the base of the rail, up against the face of the cross tie, and they gave suitable anchorage to prevent the rail from running longitudinally. The D&H adopted those rail anchors for later installations of CWR, which resumed around 1958, and they have continued in use on installations made since that time. If every alternate tie is box-anchored, that is, the tie has one anchor on each rail on each face of the tie, for a total of four anchors, the rail will be suitably restrained. The ends of the strings, near joints or fixed objects like road crossings or turnouts, should have every tie box-anchored for 195 feet, or five rail lengths of 39 feet. That works.

Tie renewal under M&L clip construction was a problem because of the labor required to remove the lag spikes and clips, until in 1968 the D&H acquired a new set of tie renewal equipment. I began the process of leaving the tie plates clipped to the rail, and sawing and pushing the old ties out from under the rail and plates. We could push a new tie into place, then tamp it up against the plate, drill holes for the lag screws, and drive the new lags with a mechanized spike driver. That process came too late to save the M&L system from eventual oblivion, but I think some of it still might exist somewhere on the D&H. I just don't have a current track chart to see if any pre-WW II 131 RE rail exists anywhere on the D&H. It doesn't have to be CWR. I suspect the most likely location would be on the Upper Hudson Railroad, or some part of the Adirondack Branch south of North Creek.

M&L Construction was more expensive than conventional cut spikes and punched tie plates, and the D&H just didn't generate the capital needed to continue its use.

Gordon Davids
Posts: 363
Joined: Sat Mar 13, 2004 6:51 pm
Location: Severna Park, MD

Re: Welded rail

Postby Engineer Spike » Sat Dec 13, 2008 6:37 pm

When I used to work the Palmer Falls, on the AD branch, most of it was cwr between AD Cabin and Corinth. They have been doing lots of rail replacement lately, so much of the old rail is being replaced. Some of the 131 may still be in on the sidings. Some of them have a real hodge-podge of rail. About 2 years ago they replaced the rail behind my house. It is just north of Ballston Spa siding. There was 112RE rolled in 1940 and 1942. Now it is 115# CWR. It is much quieter, but the only problem is that they left a joint between two strings right here. I have kidded the roadmaster about when he is going to weld that joint.
Engineer Spike
Posts: 1856
Joined: Tue Mar 16, 2004 3:24 pm

Re: Welded rail

Postby staustell92 » Sat Sep 17, 2011 11:04 pm

Just read this string (nearly three years after everyon eelse) and wanted to breathe a little more life into it:

1) I believe the New York State Museum has a 16 mm film of that first continuous welded rail installation behind the headquarters building in Albany. I found the film (along with others) in a trailer at Colonie during the bankruptcy and Belke gave me the okay to take it to the museum. As I recall, one of my friends (who was a news manager at a lcoal TV station) first copied it onto a VHS tape. I will ask him what the status of it is now.

2) In all the discussion about civil engineering advances during the Loree administration, I didn't notice anyone mention that L.F. Loree (the President) was himself a civil engineer. I don't know what engineering school he went to. I understand he took a personal interest in the design (and construction) of the Colonie Shop conplex and other facilities. At the time of its construction (about 1911), Colonie was state-of-the-art.
Posts: 12
Joined: Fri Feb 25, 2011 1:37 pm

Re: Welded rail

Postby march hare » Mon Sep 26, 2011 12:18 pm

Hope they post it on YouTube
march hare
Posts: 361
Joined: Thu Mar 25, 2004 12:44 pm
Location: upstate NY

Re: Welded rail

Postby ChiefTroll » Thu Mar 01, 2012 5:42 am

Leonor F. Loree was a Civil Engineering graduate of Rutgers College. From Wikipedia:

He obtained a Bachelor of Science degree from Rutgers College in 1877, a Master of Science from Rutgers in 1880, Civil Engineering degree from Rutgers in 1896 and a Doctor of Law degree from Rutgers in 1917. He also obtained a Doctor of Engineering degree from Rensselaer Polytechnic Institute in 1933. He was President of the Delaware & Hudson Railroad; had interests in Kansas City Southern, Baltimore and Ohio, New York Central, and the Rock Island Railroads. Was a Trustee at Rutgers University from 1909–1940 and was Chairmain of the Rutgers Board of Trustees Committee on New Jersey College for Women (now Douglass College) until 1938. He was the donor of the New Jersey College for Women Athletic Field (which is now Antilles Field). Rutgers has a building named after Leonor Fresnel Loree. The Loree Building was erected in 1963 and is on the Cook/Douglass campus.

- Gordon Davids
Posts: 363
Joined: Sat Mar 13, 2004 6:51 pm
Location: Severna Park, MD

Return to Delaware & Hudson

Who is online

Users browsing this forum: No registered users and 1 guest