Stacking bodies on rails: limits coming from the strength of the rails

1. Background: From Sherman to Sobibor

During his campaign in Georgia, General William Tecumseh Sherman ordered the destruction of the railroads. The tracks were to be destroyed by heating them in a fire and then bending them beyond repair. The twisted rails came to be known as “Sherman’s neckties.”

Some bent rails:


Some rails ready to be “cooked”:


A drawing of necktie manufacture:


At Treblinka, Belzec, and Sobibor, we are told that the Germans burned some 1.5-2 million bodies on open-air roasts constructed out of railroad tracks, which supported as many as 3,000 bodies at a time. The question arises, whether rails would have been able to support such loads, given how readily Sherman’s men bent them.

To get an idea of what it takes to bend rails in a fire, I strongly recommend watching this video, in which some Civil War reenactors bend some modern rails.

Note that the rails are already bent when they come out of the fire.

It’s also worth watching the depiction of Sherman’s men bending rails from the movie The Horse Soldiers:

The problem of rail strength has been mentioned before in revisionist literature, but seems never to have been thoroughly discussed. It was mentioned on p. 148 of Mattogno and Graf’s Treblinka, and a brief allusion to this issue in One Third of the Holocaust provoked the following complaint from Roberto Muehlenkamp:

Just what kind of «specially built alloy beams» he has in mind Bud doesn’t specify, neither does he explain why one would need such beams when old railway tracks would do the job or, for that matter, why old railway tracks would not be sufficiently resistant «to hold a lot of weight in high heat conditions» as required.

1.1 A simple formulation of the argument

In the first video above, the rails were severely bent after around an hour in the fire. That degree of bending, if it occurred at one of the Reinhardt camps, would destroy the cremation facility. And these rails were not even loaded! Therefore rails could not have withstood the heat of the cremations.

There’s a natural objection to this argument. The rails in the video were bent in a direction that would be horizontal when the rails were in the ordinary orientation for rails. But a wisely set up cremation facility would rest the rails on their base, so that in an ideal world we would only have to worry about vertical bending. Rails are stronger when their profile looks kind of like an I-beam than when their profile looks kind of like an H-beam.

How much more resistant are rails to vertical bending than to horizontal bending? We’ll see in section 4 that they are about six times more resistant. In the video, rails bent under their own weight. But the holocaust story claims that thousands of bodies were stacked on the rails during the cremations – a load much more than six times the load of the rails’ own weight. Therefore the rails would not have had the strength to support the Reinhardt cremations, so the Reinhardt story cannot be true.

2. Witness descriptions of Treblinka cremation facilities

The key issues are

  • how long the cremation facility was
  • into how many spans it was divided
  • how long each rail was
  • what type (in particular, what weight) of rail was used
  • It’s quite difficult to get answers to these questions, and entering into a detailed analysis of the witness accounts would take us off topic. Suffice it to say that the current consensus has cremation facilities 25-30 meters long, and is quite vague on the number of supporting pillars. Arad claims there were three pillars (two spans); the Wiernik model has four pillars (three spans), as does the Treblinka museum’s model, and the USHMM’s map. The Düsseldorf court’s map, on the other hand, shows a single span.

    The models showing multiple spans seem to show a single set of rails running the entire length. On the other hand, Samuel Rajzman spoke of carrying rails which “had to be carried by a team of six or eight people.” He also emphasized how weak the carriers were. If the rails weighed 60 kg per meter, then 30 meter long rails would have weighed 1,800 kg, which certainly could not be carried by a team of six or eight people. Such a team of prisoners would probably not be able to carry rails any longer than 5 meters.

    Of course, lighter weight rails could have been used, but that would have make them weaker. One also has to wonder exactly how 30 meter long rails were to be delivered. Did the Germans even have rolling stock of that length? If, as some witnesses say, the rails were scavenged from the surrounding area, it’s unlikely that they were as long as 30 meters.

    I focus on Treblinka over the other camps because there are more explicit descriptions of its cremation facilities. However, the case at the other camps may be even worse. The Central Commission for the Investigation of German Crimes in Poland found the following:

    Large scale incinerations began in the winter of 1942/1943 and continued up to the liquidation of the camp. At first simple pyres were used, but eventually this system was replaced by the use of grates made from railway rails.

    Such an installation was very simple. Rails were mounted on two parallel rows of concrete blocks, layers of corpses were placed on them, and a fire was lit below. It is probable that easily flammable material was used.

    Thus at Sobibor there was only a single span!

    3. Formulas for the deflection of a loaded beam

    We need to know what the maximum deflection is in a uniformly loaded beam on a simple span, as well as deflections for double and triple spans. These are standard engineering formulas.



    The above charts didn’t give max deflection for double spans but rather deflection in the middle of the span. Here’s a chart to remedy that gap:


    Young’s modulus E for steel is a little over 200 GPa, the exact figure depending on the specific type of steel. We’ll use 210 for our calculations.

    A handy calculator for beam deflection can be found at

    4. Limits of the strength of unheated rails

    Using the above formulas we can calculate the deflection for whatever setup of rails we like. The only piece of information we still need is the (area) moment of inertia for rails. These tables give moments of inertia for the rails made by Tata steel:



    Here are moments of inertia for some historic rail profiles, given in in^4 rather than cm^4.

    These values are for new rails; worn rails will of course have a lower moment of inertia. This book gives the value of 1,440 cm^4 for the moment of inertia of worn rail.

    I will assume in the calculations that follow that 6 rails are used, and that the weight of the bodies bears equally on them. This is a best case scenario, as the number of rails used is variously described as 2, 4, 5, and 6, and the load on the central rails would likely exceed the load on the rails under the legs or head.

    4.1 Calculations for single spans

    We’ll use Young’s modulus = 210 GPa and moment of inertia = 3,000 cm^4 for all calculations in the following three subsections.

    Span length: 30 m
    load: 1226.25 N/m [500 bodies, 45 kg per body]
    max deflection: forget it! (more than 2 meters)

    Span length: 15 m
    load: 1226.25 N/m [250 bodies, 45 kg per body]
    max deflection: 12.83 cm

    Span length: 10 m
    load: 1226.25 N/m [166 2/3 bodies, 45 kg per body]
    max deflection: 2.53 cm

    Span length: 10 m
    load: 4905 N/m [666 2/3 bodies, 45 kg per body]
    max deflection: 10.14 cm

    Span length: 5 m
    load: 4905 N/m [333 1/3 bodies, 45 kg per body]
    max deflection: 0.63 cm

    4.2 Calculations for double spans

    These are for a single beam spanning two spans of the given length.

    Span length: 15 m
    load: 1226.25 N/m [250 bodies, 45 kg per body]
    max deflection: 5.33 cm

    Span length: 10 m
    load: 1226.25 N/m [166 2/3 bodies, 45 kg per body]
    max deflection: 1.05 cm

    Span length: 10 m
    load: 4905 N/m [666 2/3 bodies, 45 kg per body]
    max deflection: 4.21 cm

    Span length: 5 m
    load: 4905 N/m [333 1/3 bodies, 45 kg per body]
    max deflection: 0.26 cm

    4.3 Calculations for triple spans

    Span length: 10 m
    load: 1226.25 N/m [166 2/3 bodies, 45 kg per body]
    max deflection: 1.34 cm

    Span length: 10 m
    load: 4905 N/m [666 2/3 bodies, 45 kg per body]
    max deflection: 5.37 cm

    Span length: 5 m
    load: 4905 N/m [333 1/3 bodies, 45 kg per body]
    max deflection: 0.34 cm

    5. Behavior of steel at high temperatures

    So much for the behavior of loaded rails without a fire. Steel gets weaker at high temperatures. We’ll start with two subsections giving examples of this, then proceed to some more exact analysis.

    5.1 Temperatures used in blacksmithing

    Blacksmiths bend steel at temperatures which, as we will see, are attained in wood-burning fires:

    5.2 Collapse of bridges in fires

    There have been a number of cases in which a fire from the crash of a tanker truck has caused a bridge failure. One such incident occurred in 2004 in Bridgeport, Connecticut.

    ”When firefighters arrived, these 22-to 30-inch-thick beams were glowing bright orange and sagging 3 to 4 feet,” Chief Maglione said.

    Seen from below, the burning bridge looked like a hammock with a heavy person lying in it. a witness said.

    Another took place in Oakland in 2007.

    Yet another, in Detroit in 2009.

    Some people may be inclined to say that these collapses are possible because we are dealing with gasoline, and that wood couldn’t produce a similar effect. This is based on the incorrect idea that gasoline burns hotter than other fuels. As noted in The Analysis of Burned Human Remains, [p. 4]

    It was once thought that gasoline-fueled flames were much hotter than those of ordinary combustibles such as wood or plastics, and therefore any supposed ‘high temperature’ effect was due to such a petroleum product. This has now been demonstrated to be untrue. A flame from a burning pool of gasoline is indistinguishable in its average temperature from that of a pile of wood.

    5.3 Data on strength reduction

    Here’s a simple representation of the weakening of steel at high temperatures:


    Of course, one can ask for more detail – tracking Young’s modulus or the yield strength as a function of temperature, but this will do for now.

    5.4 A model

    We can apply data on the value of Young’s modulus as a function of temperature to predict deflections as we did in section 4, but this approach is misleading. Calculations of deflection based on Young’s modulus apply when the steel is behaving elastically and linearly. Neither of these will be the case at high temperatures.

    The paper On the Behaviour of Single-Span Steel Beams Under Uniform Heating gives us a useful look at how things work at high temperatures. Here’s the deflection in a beam graphed as a function of temperature at several load levels:


    The key feature is that at a certain temperature things, well, go very wrong. Note that at less than 750 degrees C a beam that had less than 2 cm deflection (with load ration 0.2) at room temperature has 80 cm of deflection – and it’s just getting started failing. In accordance with the data on the reduction of the strength of steel (more precisely: of Young’s modulus) at high temperatures, we would expect a deflection of maybe 10 cm for this beam at 750 degrees C – but reality is much worse than this. Rather than increasing ~6 fold, as a linear analysis would tell us, the deflection increased 50 fold! The reader should look back at the calculations in section 4 to see what this implies for the deflections of beams in the Reinhardt cremations.

    Another way to look at this example is to note that the room temperature deflection at load ratio 0.2 was only 1/1000th the length of the beam, which is 18 meters. Yet the beam failed at less than 750 C.

    In the linked paper, the critical temperature is defined at the last temperature given; it is not the point of inflection but the final point, with deflection ~80 cm in the above graph. The lower the load on a beam, the higher the critical temperature. This table shows the critical temperature as a function of the loading:


    Would the beams on the Reinhardt pyres reach these temperatures? We’ll see in section 6 that the answer is yes.

    Note: The following links are worth looking at:
    Most books on structural engineering and fire also contain at least a brief overview of the properties of steel at high temperatures.

    5.5 Fire tests
    Fire tests offer a useful base of examples of beam deflections in fires. See for instance the following paper:

    This thesis also has a good deal of useful information, such as the following chart of failure temperatures versus load levels:


    This report gives some details on the deflection of beams in an actual fire. The deflections are on the order of 1/30 the span:


    But the beams were only lightly loaded:


    Surprisingly, the temperatures of the steel in this fire barely exceeded 600 C:


    6. Temperatures attained in wood burning fires

    The final question we need to answer is how hot the beams would have gotten in the Reinhardt cremations.

    Here is a modern pyre cremation.


    Part b of this chart shows the temperatures in an experimental pyre cremation by the same author who conducted the pyre cremation in the above photo:


    The book Ceramics for the Archaeologist (p. 77ff) records a juniper-wood burning fire that reached 905 degrees C, and mentions once reaching 962 C in a similar wood-burning fire. This chart shows the temperature profile in the former fire. Also of interest is the temperature profile of a wood-fueled fire in a pit, which is directly relevant to the claims that in some camps rails were placed directly on top of a fire-pit.


    This article notes that pyres can reach 1200 C.

    This article notes that small campfires (using only 6 kg of fuel) reached 900-1000 C.

    Another example: the crib fire data of D. Gross. Here is one of his crib fires:

    And here is his temperature data. The fire in the picture above is 9.15-7-10. The first number, in this case 9.15, indicates the thickness of the fuel used in centimeters.


    The report of the results of Project Flambeau says this about temperatures in test crib fires:

    Fuel Zone Temperature
    Milled fuel beds used varied in size from 4- by 4-by 4-inches to 42-feet by 42-feet by 64-inches. Peak temperature within these fuel beds did not appear to vary with the size of the bed. Peak temperatures ranging from about 1,900 F to 2,500 F were measured in both the small and large fuel beds.

    None of these fires is the equal of the alleged Reinhardt cremations, even if one accepts the statements on fuel requirements made by Roberto Muehlenkamp. The Gross fires and experimental pyres are not their equal in scale; the Flambeau fires are not their equal in fuel load. Fires capable of cremating 2,000-3,000 bodies would have to be very intense. We can therefore be confident in asserting that in the Reinhardt cremation story, temperatures of over 900 C would have been reached in the rails, and that significantly higher temperatures are highly probable. This means that the rails would have lost 95% of their strength, if not more.

    6.1 Temperature of the fire versus temperature of the steel

    It might be objected that the above analysis overstates the strength reduction of rails in fire, because the temperature of the rails will not equal the temperature of the fire, so the rails will be stronger than might be thought based on the temperatures of wood-burning fires. This objection would be valid for a fire of short duration, but not for outdoor cremations, which need to last several hours at the least.

    The rate at which steel is heated in fires is analysed in terms of section factor = (perimeter of beam exposed to flames)/(cross sectional area). For a rail weighing 60 kg per meter, we have a cross sectional area of around 0.0075 square meters and a exposed perimeter of perhaps 70 cm; this gives a section factor of 93 m^-1. As the following chart shows, such a beam will be fully up to the ambient temperature within 90 minutes.


    One should also account for the fact that there will be some mass loss in the pile of bodies during the time it takes the rails to come up to temperature, but this is not going to be a large enough factor to overturn our basic conclusions.

    7. An experiment

    This video, which is part of a National Geographic series against the so-called truthers, contains an experiment highly relevant to the problem at hand. The part that concerns us runs from 2:15-7:20.

    The beam used is a W8x18, with moment of inertia 2,587 cm^4 – a perfectly reasonable approximation of a rail. The load is 3,000 pounds (although it is concentrated near the center of the beam rather than uniformly distributed). The beam is 20 feet long, but the span is probably only around 5 meters or so. The video gives temperatures, which rise to just over 2,000 degrees F, but these are gas temperatures, not steel temperatures. This fact is evident not just from calculations of the temperature rise of the beam, but also directly from the video, as we get an uninterrupted view of the beam just seconds before it collapses:


    It is not glowing. Here is a temperature/color chart for steel:


    The exact temperature at which redness becomes visible in steel depends on the lighting, but by the time steel reaches 600 degrees C it’s certainly perfectly evident; the beam is therefore cooler than that (probably quite a lot cooler).

    This low-temperature failure tells us that had there been time for the fire to raise the beam to a higher temperature, it would have failed even with a much lower load. Looking at our chart of the strength of steel as a function of temperature again, we see that a good pyre would reduce the beam’s strength to 10% or less of that which it possessed when it failed in this experiment. Remember that the load was only 3,000 pounds, or 30 bodies on a 5 meter span. With 6 rails, that’s 180 bodies – still less than was needed according to the witness accounts of Treblinka.

    Aside: Arad mentions (p. 175) that the rails at Treblinka glowed in the heat. I do not know from whose statements he derives this assertion. It would certainly have to be true – any cremation facility that places bodies on rails above a fire will have to heat the rails hot enough to glow if cremation is going to take place – but it does imply that the rails would have been quite weak. Small loads and short spans are the only way to make this kind of a system work. (Given how difficult it is to design a system of mass cremation in which the bodies rest on rails that won’t lead to the rails being destroyed each cremation, why would the Germans have chosen such a method to destroy millions of bodies?)

    This entry was posted in Uncategorized and tagged , , , , , , . Bookmark the permalink.

    3 Responses to Stacking bodies on rails: limits coming from the strength of the rails

    1. Pingback: Dimensions and structure of the Treblinka cremation facilities | holocausthistorychannel

    2. Pingback: Bending of rails in the Reinhardt cremation facilities: yield strength analysis | holocausthistorychannel

    3. Pingback: Lateral torsion buckling and the Reinhardt cremation facilities | holocausthistorychannel

    Leave a Reply

    Fill in your details below or click an icon to log in: Logo

    You are commenting using your account. Log Out /  Change )

    Google photo

    You are commenting using your Google account. Log Out /  Change )

    Twitter picture

    You are commenting using your Twitter account. Log Out /  Change )

    Facebook photo

    You are commenting using your Facebook account. Log Out /  Change )

    Connecting to %s