The American railroad system was, in the words of Jessamyn West, “a big iron needle stitching the country together.” The “needle,” I suppose, is the locomotive, but it doesn’t accomplish much without the “thread”—the track.

In this article, I will be covering railroads in pre-RoF Grantville, the immediate post-RoF rail supply, route planning, track design and construction, rolling stock, and some safety issues.

Railroads in Pre-Ring of Fire Grantville

Canon does not specifically describe the rail network in the vicinity of Grantville. However, since Grantville is based on real-life Mannington, West Virginia, it is worth taking a look at Mannington’s situation. In 2005, CSX owned a line which extends approximately 17.51 miles from Milepost BS 306.32 near Barrackville to Milepost 319.48 near Mannington. This line closely parallels Buffalo Creek. The right-of-way (ROW), which varies from 30 to 90 feet wide, was originally acquired in the 1850s by the Baltimore & Ohio (B&O) Railroad (which became part of first the Chessie System, and later, CSX). (It presumably is part of the original B&O main line out to Wheeling.) CSX also owns, in the same area, the 4.35-mile Dents Run Spur between Milepost BSB 0.00 and Milepost 4.35. I am not sure exactly how much of this track would be within the Ring of Fire, but certainly some of it is.

As of 2005, there were still rails, cross-ties, and ballast on the line, at least in some sections. We know this because CSX sought government approval to abandon the line and salvage that material. We don’t know the weight of the rail, but for reasons explained later, I would guess it to be at least 100 pounds a yard on the part alongside Buffalo Creek, and at least seventy pounds a yard on the Dents Run spur.

The line also includes some old bridges, “built from 1904 to 1912. They are all steel/concrete or steel ‘I’ beam structures.”

Barflies visiting Mannington have documented the existence of track serving various abandoned coal mines or lumber camps in the area. In the 1632 Dead Horses FAQ, their findings are summarized as follows:

* 45 narrow gauge rail cars were found in the ROF mostly in abandoned mines. There was a mix of 2 ft and 2.5 ft gauge as well as cars that could operate on either . . . .

* 60 miles of 20lb track were found on lumber trails and in doghole mines—of this about 45 miles is serviceable.

A rail line went into Grantville for servicing the power station. So that track, as well as some rolling stock and even a kind of a small car workshop, came with us through the Ring of Fire. Regrettably, there was not even a single engine at this time in Grantville.

Besides the track, there is also a brick, ex-B&O railroad station in Mannington, built in 1896, and now used to house a business ( Arrowhead Resources LLC, 104 West Railroad St).

Rail Supply

The USE starts with both the old B&O standard track, and the narrow (2 or 2.5 foot) gauge lines servicing the lumberjacks and miners. Some of the latter is being used by Pitre’s railway battalion. The story doesn’t say anything about the rest of that track, but it seems reasonable to expect authors to follow the guidance of the FAQ, which says that “the unserviceable track . . . was turned over to USE Steel for recycling into bar rail.”

The initial supply of steel rail is whatever salvaged up-time rail is not used to make armor for the four ironclads (Weber and Flint, 1634: The Baltic War, Chapter 52; only three were referred to in Weber, “In the Navy,” Ring of Fire). We know that the navy is going to need “miles of track.” How many? Until we have the specs of the ironclads—the thickness of the armor, whether it is just flank armor or also deck armor, and the length, height above water, and width of the ships—we cannot determine the quantities involved. However, that didn’t stop me from making a crude estimate.

Eddie Cantrell at one point makes reference to the Benton and the Tennessee, as being bigger than what the USE Navy will be building.

The first CSS Tennessee would probably have displaced 800 tons like her sister ship CSS Arkansas. She was burnt before completion. The second was 1273 tons, 209 feet long with a 48 foot beam. She was captured in 1864 and became the USS Tennessee. Both were larger than the Benton.

The USS Benton was a 1033-ton Civil War ironclad, the largest in the western flotilla. It was 72 feet wide and 202 feet long. The sides were slanted, probably at a 45 degree angle, and bore armor 3.5 inches thick. The wheelhouse and stern had 2.5 inch thick iron. I am not sure of the height above water, but for the “city class” ironclads like USS St. Louis, it was twelve feet. If the stated beam is at the waterline, then each flank had 990 cu. ft. of armor, and the bow and stern each had 252 cu. ft. So we need around 2,500 cu. ft. to armor all four sides. (I am ignoring deck plating, if any.) One cubic foot of wrought iron weighs 480 pounds, so that is about 600 tons.

Is that a good estimate? By way of comparison, an 1862 Navy Department specification for an iron-clad steam battery, with a 465 foot water line, side plating mostly 4.25 inches thick, 1.5 inch deck plates, and two armored conning towers, said that the estimated weight of the armor was 691.6 tons. (Baxter, The Introduction of the Iron Clad Warship, Appendix B) Clearly, I am in the ballpark.

Steel weighs about 500 pounds per cubic foot. So, to get 2,500 cubic feet out of rails weighing 40 pounds a yard, we would need 31,250 yards (17.8 miles). For three such ironclads, we would need over 53 miles of 40 pound rail, which means over 26 miles of track.

However, the USE ironclads are designed to resist seventeenth century, not nineteenth century, cannon, and therefore probably carry thinner armor. Moreover, since they are designed to traverse rivers considerably shallower than the Mississippi, they are likely to be smaller, too. So their salvaged rails are likely to be quite a bit less than 26 miles of track.

Also, while we know that some of the rail is going to be used for the ironclad armor, canon does not insist that all of the armor be salvaged track. Some of it could be down-time iron, rather than up-time steel. The Civil War ironclads in fact used wrought iron armor.

John Zeek, one of the authors of the TacRAIL stories, stated on 1632 Slush Comments that TacRAIL used only 20 miles of the serviceable track, the other 25 miles worth having been turned over to the government. I would think it likely to have been used to make “strap rail” for the civilian railroad, rather than melted down.

Route Planning

It should be appreciated that it usually will not be possible to run the line on a straight, level path, except for short distances. If you encounter a hill, your choices will be to go over it (so the train must climb and then descend a grade), to curve around it, or to cut or tunnel through it. If you meet a river, you can cross it with a bridge (or a tunnel), or transfer the train cars (or their cargo) by ferry. An approach curve may allow a river to be crossed perpendicularly rather than obliquely.

In nineteenth-century America, where the railroad was penetrating thinly settled areas, it was not unusual to construct, at relatively little expense, a line with sharp curves and steep grades, and later, after the railroad had recouped its initial investment, reroute. New towns and businesses sprang up along the line. In Victorian Britain, on the other hand, the railroad tended to go to places where people, factories, and mines were already located, and it was easier to raise the money to finance expensive bridges, tunnels and the like. The USE of the 1630s is likely to take a middle course.


Grade (Slope). Grade is expressed either as the change in height relative to horizontal distance (e.g., 1 in 10, or 10%), or as the degree of the slope (5.7 degrees being the slope of a 1 in 10 grade).

If you are going uphill, the locomotive has to overcome gravitational force as well as rolling friction. This grade resistance is roughly 20 pounds per ton of load, for every 1% of slope. (Armstrong, 20) If the locomotive isn’t powerful enough to pull the entire train up a bill; it can double up, that is, take half the train up to the top, then head back down to retrieve the second half.

As it progresses, the locomotive is doing work, converting chemical energy into potential energy. To lift one ton a distance of 100 feet will cost 0.1 hp-hour (about 75 watt-hours). It doesn’t matter if you lift it slowly or quickly, or up a gentle slope or a steep one. The energy bill is the same.

Lifting is costly, too. The energy needed to lift one ton by that distance might be enough to propel the train 10.5 miles, at 15 mph, on level track.

You will recover that energy when you go downhill if the slope is gentle enough so you can safely coast down. If you have to brake, to avoid losing control, then some of the energy is irrevocably lost.

You see the Catch-22 here. If you economize on road cutting by letting the track go up and down hills, you will need to pay the piper by building more powerful locomotives, or running shorter (lighter) trains.

In early nineteenth-century England, great care was taken to limit grades to a maximum of 1 in 330. This of course entailed a great capital investment in road bed construction. In America, economy of materials and cash was triumphant, and much steeper grades were deemed acceptable. (NOCK/L, 34). In 1911, at least in British practice, a grade of 1 in 400 was considered easy; 1 in 200, moderate; and 1 in 100, heavy.

On a normal railway, the maximum possible grade depends on the adhesion of smooth wheels to smooth rails, which in turn depends on climate, the adhesion being strongest when the rails are dry. According to EB11, the theoretical limit is around 1 in 16 to 1 in 20,and the practical one at that time was more like 1 in 22.5.

Steeper grades can be climbed by means of a rack railway, in which a cog wheel on the train engages a tooth rack on the rail. According to EB11, rack railways can have a gradient of 1 in 4 or even 1 in 2. If that isn’t good enough, one must resort to a cable railway, in which the train is pulled up.


In the U.S., about one-seventh of all track is on a curve. (Henry, 79) Curves force the train to reduce speed (so it doesn’t derail), and also result in an effective increase in resistance. The sharper the curve, the greater the problems presented. (Armstrong, 26).

There are two ways of expressing the sharpness of a curve. First, you can calculate the angle through which the track turns in 100 feet, say, one degree. Or you can imagine the curve to be an arc of a hypothetical circle, and state the radius of that circle (for a 1 degree curve, it is 5,729 feet).

When you go around a curve, you experience a centrifugal force. The train wants to keep going straight, the tracks want it to turn. If the centrifugal force is too great, the train wins, and it leaves the tracks. Not good.

One engineering trick for dealing with this problem is called super-elevation; the outer rail is raised a few inches, tipping the train, and causing the combined effect of gravitational force and centrifugal force to keep the train on the track. Six inches of super-elevation is perfect if you are going 45 mph around a five degree curve.

Unfortunately, if you run slower trains on the same track, the wheel flanges bear down heavily on the inner rail, resulting in excessive wear. So the degree of super-elevation is usually a compromise on lines carrying both fast passenger trains and slow bulk freight.

Another remedy is to provide a check rail on the inner side of the inner rail. As the wheels try to slide horizontally outward, they meet the check rail.

A third solution is to “tweak” the shape of a railcar wheel. If you stand in front of a train (hopefully one which is stationary!), you will see that wheels are not perfect cylinders, they are slightly conical (standard “taper” is 1 in 20), with the narrowest diameter on the outside. As the train moves onto a curve, the wheels shift outward, so the outer wheel’s diameter at the point of contact increases, and that of the inner wheel decreases. That corrects for the curve.

If these expedients are insufficient to compensate for the centrifugal force, the train must slow down. The speed on a fifteen degree curve might be half that on a five degree one, and one-quarter that on a mild one degree veer.

There is also extra friction involved when you take a train around a curve. For each one degree of curve, figure increased resistance of about 0.8 pounds for each ton of load. That means that a 25 degree curve has the same effect on resistance as a 1 in 10 grade. Consequently, route designers usually find it better to curve around hills, than to traverse them. On the Horseshoe Curve of the Pennsylvania R.R., which is a nine degree curve, the grade is 1.8%. If the track were laid straight, the grade would have been 8.5% (Henry, 56). The most extreme examples of trading curves for grades is on certain mountain railways, which climb a hill by a series of switchbacks.

What kind of curves might be expected on the USE’s railroads? EB11 says that in Great Britain, a 15 degree curve (383 foot radius) is considered “very sharp, at least for main lines on the standard gauge.” It is very sharp, indeed; you probably would encounter such a curve only in a mountain region, or to get onto an industrial siding. (Henry, 57). Clarke (8), discussing late nineteenth-century practice, says that on main lines, most curves are of at least 1,000 foot radius in Europe, and 300 foot radius in America.

Traversing curves has an energy cost. First of all, there is the increased friction. When the curves add up to a full circle, you have used up about 0.014 hp-hour per ton, the equivalent of lifting a ton a distance of fourteen feet. Moreover, curves increase the effective distance traveled, and work equals force times distance.


Acceleration. To stop at a station, you must decelerate. Then, when you are ready to move again, you must accelerate until you reach “cruising speed.” Likewise, if you slow down to negotiate a curve, you will pay a price when you speed up again. You need extra tractive force (above and beyond that needed to overcome the base train resistance) whenever you want to accelerate; as Newton said, Force equals mass times acceleration. According to EB11, the acceleration resistance equals the weight of the train, multiplied by the ratio of the desired acceleration to gravitational acceleration.

Acceleration also has an energy (fuel) cost. According to Armstrong (24), the energy needed to accelerate a stopped train up to 30 mph is what would be needed to get it to roll three miles at a constant 30 mph. Decelerating doesn’t cost energy, but does result in brake wear.

Track Gauge

The track gauge is the distance between the inside edges of the two rails forming a single track. Wider gauges allow the railway to carry heavier loads (because it can use wider cars), but construction costs are higher, because everything must be scaled up: the weight of the rails, the length of the crossties, the width of the supporting bed. Narrow gauges not only reduce those costs, but also allow the track equipment to negotiate tighter curves.

In the twentieth century, gauges ran from one feet for certain industrial railways, up to nine feet for one of the Japanese funiculars. The so-called “standard” or Stephenson gauge is 4 ft. 8.5 in. Seven different gauges were in use in the United States in 1860.

It is very undesirable for there to be a mix of gauges within one country’s rail lines. If gauges are mismatched, then, at the point where the gauge is “broken,” passengers and freight must be unloaded from one train and reloaded on another. Or some other adjustment must be made, e.g., “cars with a sliding wheel base, hoists to lift cars from one wheel base to another, and, most commonly, a third rail.” (

On the other hand, a change of gauge at a national border makes it difficult for an invader to use the victim’s tracks, at least until it captures the latter’s rolling stock. The Finns deliberately used a different gauge than the Russians.

Since no other power has a railroad, the USE isn’t going to be much concerned about the defensive value of a change of gauge. It seems pretty likely that the old timeline “standard gauge” will be standard for civilian railroads in the new timeline, too. See Weber, “In the Navy” (Ring of Fire).

Loading Gauge

The loading gauge is the maximum height and width of the rolling stock which can use the track, and is defined by bridge, roadcut and tunnel clearances and, in some cases, limited by the stations through which the tracks pass. Modern British railways under a restrictive limitation of 9’0″ width and 12’11” height. In contrast, in America the rolling stock can be as wide as 10’10” and as high as 16’2″. (NOCK/RE, 208-9).

The tradeoff, here, is that a permissive loading gauge results in greater construction costs for bridge and tunnel work, while a restrictive one forces the use of more cars to carry the same load. I suspect the American standard will prevail in the post-RoF USE.

Bear in mind that having a large loading gauge is ineffectual if the track gauge is narrow. If you run big cars on a narrow gauge, they may tip over when the train tries to negotiate a curve.

Monorail Systems

Thus far, we have assumed that each track uses two rails. However, some short rail lines are monorail. A monorail system has some real advantages. You only need have half as much rail, obviously. And you don’t have to keep two rails level with each other, and at the correct separation.

So what’s the catch? A train with only one line of wheels, riding on top of a single rail, would tip over, just as a riderless bicycle would if you let go of it. The stability problem has been “solved” in several ways, three of which are mentioned in EB11: you can suspend the train from an overhead rail, you can straddle the cars over a somewhat lower rail, or you can equip the cars with gyroscopic stabilizers.

Suspended monorails have found some acceptance in mountainous areas; a good example is the “floating railway” (1901–present) of the Wupper Valley. The longest suspended monorail in the world is presently the Chiba Urban Monorail (15.2 km)(Wikipedia).

The problem with a suspended monorail line is that you have provide a tall structure from which the train can hang. If that structure is made of iron, you will probably need more metal for it than you would for a second rail. If it’s made of reinforced concrete, the ferrous demand will be less, but it’s unclear that the savings in iron will justify the investment which must be made in concrete formulation and construction.

A supported quasi-monorail line, designed by Lartigue, operated from 1888 to 1924 (Listowel to Ballybunion, in Ireland, 9.5 miles), and, at its peak, carried 1,400 passengers a day and 10,000 tons of freight a year. ( MBI 114) Its supports were waist-high, A-shaped wooden trestles, with the main rail on top. I call it a quasi-monorail becuse there were two additional lower rails; the cars had unpowered guide wheels which rode upon them. Its construction cost was 30,000 pounds; the line ran 9.5 miles. It had to use a custom “Siamese Twin” locomotive and custom cars; these were divided, so they “hung” over the monorail, like panniers on a camel. This of course lowered their center of gravity.

Most modern monorail systems use cars which straddle a 2-3 foot wide reinforced concrete beam. (Wikipedia).

A full size (40 foot) prototype of a gyroscopically balanced monorail railcar was demonstrated in 1909–10. It never attracted sufficient investment interest to progress further.

Track Design

The Grantvillers have plenty of up-time track—complete with steel rails, cross-ties, and cross-bed—to study. So they know what they want to build. Let’s take a closer look at the track components . . . .

Track Foundation

In early nineteenth-century Britain, and occasionally in America, the rails were fastened to square stone blocks, perhaps two feet to a side. These were cheap in England, and had the advantage that horses could walk freely between the rails, drawing the train forward. They were also durable. Unfortunately, they didn’t have any resiliency, so the ride was harsh, the rail and the wheels were subject to heavy wear, and the jar tended to shift the stone blocks out of position, leading to a variation in the gauge. (NOCK/D, 7-9, 112; Mills 210).

Hence, the stone blocks were replaced with wooden sleepers. This was usually a transverse sleeper (crosstie), a support which laid perpendicular to the rails, with both rails fastened to each sleeper. (Mills).

In 1889, there were still 1,000 miles of longitudinal sleepers (each under and parallel to a rail) in use. Light crossties were used together with the longitudinal sleepers to maintain the gauge. (SciAm Fig. 8).

Wood was preferred because it was resilient, and therefore easier on the rolling stock. In America, where timber was cheap and widespread, this also led to lower construction costs.(Stover 32).

Sleepers normally have a rectangular cross-section. However, to save money, a railroad can use a “half-round” sleeper; essentially, a log split in half (Mills 211). In fact, for a pioneer railroad, one can cut down a tree, plane down the sections where the rail would lie, and otherwise leave the log intact.

According to EB11, the British sleepers typically have a length of nine feet, a width of ten inches, and a thickness of five inches. The most common American crossties are eight feet long, eight inches wide, and six or seven inches thick. However, the width and thickness can be adjusted to the intensity of the traffic.

The sleepers are usually placed two to three feet apart (measured center to center)(EB11). Henry, addressing 1940’s practice, says that there are 3,000 ties/mile on the main lines, and 2,800 ties/mile on side tracks. (Henry 68).

Beam theory says that the maximum deflection between supports is proportional to the cube of the distance between them (Gordon, Structures, 382-3), so maintaining a spacing suitable for the expected loadings is important.

Sleepers deteriorate as a result of decay and abrasion. Baltic wood, impregnated with creosote preservative, will last twelve to eighteen years in America and only six or seven years untreated. I would expect similar performance in Europe. In the tropics, and in dry climates at high altitudes, creosote isn’t particularly effective; the sleepers rot within three or four years. (Mills 213-4).


The sleepers are most commonly made of wood, and in 1911, European railways had the wood pretreated to preserve it. EB11 suggests the use of three gallons of heated “dead oil of tar” (cresote) per sleeper. This is forced into the sleeper, which is also warmed. At that time, American railroads relied mostly on open air seasoning, but by the time of the RoF, they had changed their practice. Preservation treatment is most urgent if the rail is running through timberless country, as it is then less convenient to replace a defective sleeper.

The Boston and Lowell used solid granite longitudinal sleepers, and found that while they had a long life, they destroyed the rolling stock (Meyer 311; Bradlee 4).

Concrete and ferro-concrete sleepers have also been used, especially in twentieth-century Europe. Concrete doesn’t burn or get devoured by termites like wood, or rust like iron or steel.

Because concrete is not resilient, like wood, you have to provide cushioning pads. Because of its deficiencies in tensile and bending strength, concrete ties need pre-tensioning and steel reinforcement. Since you can’t spike concrete, you need to provide inserts to receive fasteners. (Armstrong 34). Concrete ties cannot easily be mixed with wooden ties because of the differences in mounting equipment.

In 1909, concrete sleepers cost 50% more than wooden ones, but offered twice the life expectancy. (CCE).

I doubt that concrete sleepers will be used on the initial USE tracks. However, once we are laying track in northern Germany and the Netherlands, where wood is scarce and expensive, concrete sleepers may be cost-effective.

Steel sleepers, introduced by 1875, are relatively lightweight, dimensionally accurate, and immune to biological attack, although of course they can corrode. You don’t want to use steel sleepers if the tracks are running over salty soils! In 1884, they had an expected life of perhaps 35 years (Mais 50). But steel is going to be so valuable in the first post-RoF decade that I doubt that the railroads will use it for sleepers (as opposed to high traffic rails). Wrought iron or cast iron have also been used in place of steel. (Vernon-Harcourt 252ff).

Recycled plastic and rubber, or composite, sleepers have resiliency similar to that of wood but of course don’t decay. These sleepers started to enter the market around 2004 and hence the up-timers will not be aware of them.


The sleepers themselves rest on a layer of ballast, which is sloped to provide drainage. For a double tracked line, the width at the top is 25 feet in British practice (EB11), and greater in America. The materials used are earth, gravel, broken stone and the like, sorted so that the coarser materials are on the bottom. Ballast also fills the gaps between the sleepers. EB11 recommends a depth of six inches to one foot, or more.


Overview. The rails (stringers) have two purposes. The first is to provide a low-friction surface over which a load can be dragged with relatively small force. The second is to provide a guideway so that one vehicle can lead a train of followers.

The first rails were made of a single type of wood. These were replaced by a composite of a relatively inexpensive softwood, overlaid with a hardwood as the wearing surface. The hardwood, in turn, was replaced with iron, resulting in so called “capped” or “strapped” rail. (There were also experiments with putting iron on top of stone, but it was more expensive as well as more time-consuming to lay. See Dilts, 128, 136.) The wood-and-iron composite in turn gave way to all metal rails: the cast iron rail, the wrought iron rail, and finally the steel rail.

Steel rail is made in a rolling mill. Rough stock, usually in the form of a rod, is heated and then shaped by one or more rollers. EB11 has an article on “Rolling Mills.”


Quentin Underwood criticizes the present (early 1634) rail line to Halle, because it is merely capped (strap) rail. But it is important to understand just what the use of capped rail implies.

So far as being able to pull a rolling load is concerned, capped rail is just as good as modern rail. The rolling resistance (more on that later) is dependent just on the surfaces which are in contact, so just a little pull is necessary to move a big load.

The hidden price of strap rails is that they are not as strong as all-metal rails, and so there is a limit on the axle weight of the cars. That is important if you are transporting heavy bulk freight. Maintenance costs will be higher, because it will be necessary from time to time to refasten the metal strips to the wood. The metal wearing surfaces are thin, and hence are also likely to fail more quickly than a modern steel rail would under the same circumstances.

Canon doesn’t say whether the rail surface on the Halle line is iron or steel. The strap rail used on the Baltimore & Ohio Railroad in 1829 was wrought iron, 0.625 inches thick, 2.5 inches wide, and fifteen feet long, with a weight of fifteen pounds a yard. It was imported from England ($55-60/ton), since domestic production would have been almost twice as expensive. The total cost of laying the track was $4,000 a mile. (Stover, 32).


The reason strap rails replaced wooden rails was not because they further reduced friction between rail and wheel, but because the older rails broke, wore away, or decayed too quickly. This changeover occurred even before the invention of the locomotive. Now, a curious fact—not likely to be known to the up-timers, but perhaps capable of rediscovery by someone like Dr. Gribbleflotz, who wouldn’t know not to try to do it—is that it may be possible to improve the qualities of wood by impregnating it with iron oxide and lime. This was reportedly done in the 1840s, and the wood rails cheerfully bore the equivalent of twelve trains a day for seven years. (MBI 113). In those parts of Germany which are still well-endowed with trees (see Cooper, “The Wooden Wonders of Grantville,” Grantville Gazette 13), treated wooden rails might be worth experimenting with.


There was also some early experimentation with stone rails. The “Granite Railway” (built 1826) carried granite three miles from Quincy to Milton. Initially, it had twelve inch wooden rails on granite crossties. When the wood decayed, they were replaced with stone rails. (GRCQM)

A modern engineer would think of concrete, rather than cut stone. Concrete rails have experienced a renaissance in the form of guided busways. Of course, rubber tires on concrete experience much greater rolling resistance than steel wheels on steel rail, mostly because of the deformability of the rubber.

Concrete rails have the disadvantage that once such rails are produced, they can’t be bent. In contrast, steel rails can be bent as needed to form curved sections.


Rail Composition. Rails can be made of iron or steel. Cast iron is hard, but too brittle. Wrought iron is tough, but a bit too soft. (Armstrong, 31). Nonetheless, wrought iron replaced cast iron in the 1820s, as it permitted use of longer rails, thereby minimizing the number of rail joints. (NOCK/D, 28).

Still, the best rail material is steel, which is harder, tougher, and stiffer. According to experiments in the 1860s, it had 2.4 times the compressive strength of wrought iron and 3.5 times that of cast iron. In tensile strength, it was about twice that of wrought iron and four times that of cast iron. (Flint) In general, it can carry three times the load borne by the same rail formed in cast iron (Henry, 73), and almost twice the load carried by a similar wrought iron rail (Clarke, 122).

Moreover, it is much more durable than iron. In an experiment in which trains ran over both iron and steel rails, eight iron rails were completely worn out, on both faces, while the one used face of the steel rail was worn down only one quarter of an inch. (Flint) Another authority says that steel rails have 5-6 times the working life of iron rails (Clarke, 37).

Carbon has a major influence on the hardness of steel. EB11 suggests that steel rails made by the Bessemer process should have a carbon content in the range of 0.55-0.65%, and not more than 0.1% phosphorus. If the steel is made by the more modern open hearth furnace, the carbon content is the same, but phosphorus is limited to 0.03%. The purpose of these specifications is to minimize rail failures.


Rail Geometry. The standard rail for the last one hundred fifty years is of a kind called “edge” rail, because it looks like the old plate rail turned on its edge, or “T” rail because it is fancied to look somewhat like an upside down “T”. Personally, I think it looks more like the letter “I”. However you visualize it, the base (foot) is wider than the top (head); the vertical connection is called the web. Modern rails are “parallel,” that is, they have a uniform width throughout the rail; certain nineteenth century rails had bases which were wider at the center (“fish-bellied”). This was designed to strengthen them against bending where they were least supported by cross-ties, but they tended to break near the “chairs” which held the junction points.

EB11 Fig. 13 shows the cross-section of American ninety pound rail, and its interaction with the rail joint. EB11 also comments that in the eighty five pound rail used by the Pennsylvania Railroad in 1908, 42.2% of the metal is in the head, 17.8% in the web, and 40% in the base. The Canadian Railway specifications for the same weight rail contemplate less metal in the head (36.77%), more in the web (22.21%), and slightly more in the base (41.02%).

A 132 pound rail has a cross-sectional area of thirteen square inches, and lighter rails would have less. (Armstrong 33).

Rail height controls stiffness; a 7 3/8 inch rail is about 14% stiffer than one which stands only 7 inches high. (Henry, 74).


Rail Weight. The wrought iron rails used by George Stephenson in the 1820s weighed 28 pounds per yard. By the mid-1830s, the new English rails were fifty pounds per yard, and by 1911, main line rails were more like one hundred pounds per yard. (EB11) The rails used by the Baltimore & Ohio in 1830 were about 35 pounds per yard (Dilts, 128).

By the late 1970s, new rail was 112-145 pounds a yard, and six to eight inches tall. (Armstrong, 31). The mainline standard in the USA, just before RoF, was 141 pounds per yard. In 1960, only five per cent of the rail used on Class I U.S. railroads were less than seventy pounds per yard. Even on Class III railroads, only about one-third was that low in weight. (NMRA) Nonetheless, we will probably need to content ourselves with even lighter rails.

The rail’s cross-sectional dimensions (head, base, and height) are scaled to its weight. The stiffness of a rail is proportional to the square of its weight (Bitzan), as well as to a measure of the intrinsic stiffness (Young’s modulus of elasticity) of the rail material. The deflection of the rail under load, which stresses it, is a function of its stiffness. Thus, a heavier rail can tolerate rolling stock with higher axle loads.


Rail Length. The length of the individual rail is a compromise between several factors. The shorter the rail, the more sections must be fitted into each mile, and hence the more joints which are created. Each joint is a point of weakness. The longer the rail, the harder the rail is to manufacture, transport to the line end, and lay to extend the track. In the United States, the standard rail length is 39 feet, because the standard car length was once 40 feet. In Britain, in 1911, the rail lengths were 30-60 feet.


Jointing and Welding. In early nineteenth century Britain, the ends of the rails, which had rounded bottoms, were wedged into cast iron chairs (see EB11, Fig. 10), which, in turn, were spiked or bolted down on wooden sleepers. Later, the joints were suspended between two chairs, and were connected by what is called a fish joint (See EB11, Figs. 7 and 11). American practice was to simply spike flat-bottomed rails directly to the supporting sleepers (see below), and to connect the joints by “angle bars” (EB11, Fig. 13).

Among the many subtleties of jointing is the practice of putting the head ends of the bolts alternately on either side of the rail.

Rails can be laid so that the junction of one rail of a track are aligned with those of the other, or so that they are staggered.

The effective length of the rail sections can be increased by welding sections together. This can be done in-place, if there is suitable portable equipment. Ideally, this is done on a hot day. (Armstrong, 32-3).

Road and Track

A single track, by itself, can only accommodate traffic in one direction. If two-way traffic is light, it can be handled by providing sidings. The train with priority will keep going; the yielding train will move onto the siding and stay there until the priority train passes.

The greater the number of signal blocks and sidings, the less two-way traffic is impeded.

Ultimately, the railroad company will want to double track the entire line, assuming it is a busy one (Formally speaking, this doubles the “miles of track,” but the “miles of road” remains unchanged.) Then the traffic moves in one direction on one track, and in the opposite direction on the other, without interference.

In the vicinity of major stations, and railroad service facilities, it is not unusual to need more than two parallel tracks.

Initially, we will be building one complete track, together with a few sidings. That means that we need to discuss block movement and signaling (see “Safety” below).

Track Construction

On the site you can find my “railroad construction spreadsheet.” The original version relied on Carsten Edelberger’s proposed routes (“Railroading in Germany,” Grantville Gazette 7) and allowed you to set the following for each line:

—initial construction date going “up the line” (e.g., Grantville to Magdeburg). (The date had to be expressed as 193x rather than 163x to sidestep an Excel limitation.)

—initial construction date for a second crew going “down the line.”

—a difficulty multiplier for each section (default value 1).

—a miles/week construction rate, independently, for each month from January 1632 on. (This rate applicable to all crews on all lines).

I have in the works an improved spreadsheet that lets you specify separate construction rates for each month, for each line, in each direction.

All completion date estimates given in this article were achieved at by making assumptions and plugging them into the original or the improved spreadsheet.


The nitty gritty of surveying, road preparation and track laying are explained by Evans, “Binding the Land With Steel” (Grantville Gazette 25). Evans describes how railroads were built in America after decades of experience. He describes a “living machine,” a relentlessly advancing assembly line.

After the first few decades of railroad building, the supervisors knew how to maximize efficiency, and there was a supporting infrastructure—suppliers of rails, crossties, food, and so forth.

Even if the up-timers have read about how the transcontinental railroads were built, I am doubtful that they will be able to take full advantage of this book knowledge. They will have to learn the hard way how to estimate the rate at which they prepare the road and lay the track, and how that rate is affected by weather, season (long or short days) and terrain. They don’t yet know how many man-days it takes to extend the line by a mile, or how most efficiently to allocate the laborers to clearing, grading, ballasting and track-laying.

None of the down-timers they are supervising will have any prior experience with laying track, and while they can cut trees and shovel dirt, their approach to road preparation may be crude by up-time standards.

Then there is an infrastructure problem. The railroad is competing with the military, and other industries, for the available iron. The down-timers don’t make steel on a large scale. So, to enjoy the benefits of steel rails, we have to reconstruct those techniques, too.

Also, there is a shortage of wood in some parts of seventeenth-century Germany (it was used as a fuel, as well as in making ships, furniture, etc.). In early nineteenth-century America, in contrast, while iron rails were imported, timber was often available locally.

Like an army, the railroad work force must be supplied with water, food and shelter, and that may not be easy in some part of war-torn Germany.

There can also be social problems. Labor unrest has disrupted the construction of several lines, including the B&O’s Washington line in 1835. In DeMarce, “Prince and Abbot” (1635: The Tangled Web), talk of seizing land for a railroad by eminent domain sparks a peasant revolt, of sorts. Contrariwise, towns which want to be served by railroad but fear that service will go to a rival town may protest violently, as the good people of Omaha did when they believed the Union Pacific would cross the Missouri at Bellevue.

Finally, it is quite possible for the promoters to run out of money, and the whole project grinds to a halt while they go through another round of financing. That’s what happened to the Pacific Railroad after July 1853; as a result, it took nineteen months to complete the next eighteen miles. (mopac).

Admittedly, it will be easier to interest investors in the new time line, when the benefits of railroad are plain to anyone reading the encyclopedias, but sources of capital are still more limited than they were in the nineteenth century.

The bottom line is that I think it is safer to look at the construction experience of the earlier railroads than that of the transcontinental companies.

Construction of the first public railway, the Stockton and Darlington, began May 23, 1822, and the 27-mile line opened on September 27, 1825. That corresponds, you’ll note, to a construction rate of about a mile a month.

The B&O’s second line, running Baltimore-Washington (32 miles) line was surveyed by September 1831, but it took another two years to get funding for construction. Construction began October 1833, and service didn’t begin until August 25, 1835. (Reynolds, 16). So it took two years to go 32 miles, even though they already had the experience of building the main line to Harper’s Ferry.

I am sure we can do better than that. However, not necessarily a lot better.

The Union Pacific, using the “living machine” methods I alluded to, laid eight miles of track in one ten-hour day (and its competitor, the Central Pacific, outdid it with ten). But if you look at the Union Pacific’s entire record, you get a somewhat different perspective.

The UP broke ground at Omaha on Dec. 2, 1863, and started laying track in the spring of 1864. It took one year to complete the first forty miles to Valley; it was difficult to get construction materials to the railhead; the persons in charge were not at that time experienced; funds were scarce, and there were disagreements regarding the route to be taken. (Bailey).

The next 60 miles took from December 31, 1865 to June 2, 1866; a rate of 2.75 miles/week. Over the next 147 miles to the 100th meridian, they picked up the pace, to 8.2 miles/week, arriving October 5, 1866. Over the 460 mile section from Laramie, Wyoming (May 9, 1868) to Ogden, Utah (March 8, 1869), the Union Pacific made 10.6 miles/week.

Overall, it took until April 28, 1869 to reach Promontory, Utah, 1086 miles from Omaha. That’s about 200 miles/year, or less than 4 miles/week. On the other hand, if we ignore the pathetic Omaha-Valley leg, it took from December 31, 1865 to April 28, 1869 to go 1046 miles—about 6 miles/week.


According to Camp (188), if rails are carried in a rail car, and ties are hauled ahead with teams, 56 laborers, three foremen and eleven teams can lay a mile of track in ten hours. (If ties are carried, you need 64 laborers, three foremen and two teams.) You can assemble enough men to lay four or more miles of track in a day, but the cost per mile will be higher. In August 1887, a force of 217 men averaged 4.27 miles/day. (189).

With the 1902 version of the Holman track laying machine, a crew of 30-33 men can lay 1.5 miles/day. (195). The track laying machine doesn’t actually lay the track; it positions the rails more efficiently than can be done by teams.

That’s just the labor for laying track on an already prepared roadbed. More is needed for clearing and grading. For example, a team with a wheel scraper can move 50-55 cubic yards of earth per day on a haul of 90 feet. (10).


In deciding what would be a suitable construction rate for the early post-RoF period, we also have the constraints of canon to consider.

Until the creation of the New United States, in the aftermath of the Croat Raid of August 1632, there was no reason to build a long-distance railroad. After the NUS was created, with Magdeburg earmarked to become the industrial base, there was good reason to create a Grantville-Madgeburg line, but the next step would be to survey a route. So figure that there wouldn’t be any construction until say November 1, 1632.

We know that rail service existed, but had not yet reached Halle, as of summer 1633, but had reached it by spring 1634. Lutz and Zeek, “Elizabeth” (Grantville Gazette 4). This led me to propose construction rates which began at 1.25 miles/week, and eventually increased first to 1.5 and then 1.75, with a stoppage in September-November 1633 and a reduced rate December 1633-February 1634 because of the diversion of steel to the ironclad project. Based on those calculations, in “The Chase” (Ring of Fire II) I made it canonical that there was service to Naumburg, but not significantly further, in July 1633. (Of course, there is no reason a writer can’t find an excuse to speed construction up, at least temporarily, after that, in order to meet a plot requirement.)


Carsten proposed a trunk line running from Grantville north to Jena (milepost 27.8), Naumberg (56), Halle (88.75), Stassfurt (134) and Magdeburg (158), and south to Kronach (milepost 44), Bamberg (78) and Nurnberg (117).

The northern line is running to Naumberg as of July 1633 (Cooper, “The Chase”, Ring of Fire II), and Halle as of spring 1634 (“Elizabeth,” supra).

In April 1634, Fischer stated that Reverend Chalker’s revival group was working its way up rail line to Magdeburg, and, if the rail company kept to its construction schedule, they would hold a big event in Magdeburg in late April or early May. Hughes, “Turn Your Radio On, Episode Five” (Grantville Gazette 23).

As of May 1634, supplies for the Halle-Magdeburg section had been found, but only after a “struggle” (DeMarce, “Prince and Abbot,” Grantville Gazette, Volume 8 and 1635: The Tangled Web). Consistent with this, Chalker and Fischer go to Magdeburg by RV, not train, in May. Indeed, Iona, going to Quedlinburg in June 1634, gets off the train at Halle. DeMarce, “Until We Meet Again,” (Grantville Gazette 4).

If (1) the railroad reached Halle in March, (2) there was no construction in April or May because of the dearth of supplies, but then (3) track was laid at 2 miles/week for June-October and 1.5 in the winter months, we would reach Magdeburg around February 11, 1635 (Take this exact date, and those that follow, with a very large grain of salt.) We could complete the line sooner if there were a second crew working down from Magdeburg at the same rates; they’d meet the second half of September, 1634, between Hettstadt and Stassfurt. Or if track was laid faster than the rates given above.


As early as October 1634, there were at least plans to push the railroad north of Magdeburg. Flint and DeMarce, 1635: The Dreeson Incident (TDI), Chapter 18). No details are given.

My advice (if Mike and Gustav are listening) is to build a line from Magdeburg to Celle on the Aller. The Aller feeds into the Weser, which runs down to Bremen and the North Sea, and up to the general vicinity of Kassel. Celle is near Wietze, and a pipeline could bring the oil to Celle. Another possibility is a line from Magdeburg to Braunschweig, Hannover (on the Leine) and perhaps Minden (on the Weser). There’s oil near Hannover, too, although nothing has been said in canon about it. In fact, there could be a branch line from Hannover or Braunschweig to Celle.

Gustav is probably more concerned with rail connection to the Baltic ports, that being his lifeline to Sweden.

However, given limited resources, it would make more sense to run steamships down the Elbe and either emulate the post-RoF Elbe-Lubeck canal (itself based on the pre-RoF Stecknitz Canal) or run a short rail line along the same route (Lauenberg-Lubeck).


The southern line had not reached Nurnberg as of September (Cooper, “First Impressions”, Grantville Gazette 19) or even October (Flint and DeMarce, 1635: The Dreeson Incident, Chapter 18) 1634. In fact, I doubt that it had gone beyond Saalfeld. But in late February, 1635, Bamberg is chosen by referendum as the capital of the State of Thuringia-Franconia. The effect is that “it’s suddenly become a real high priority on everybody’s list to push the railroad through Kronach and all the way down to Bamberg.” (TDI, Chapter 42).

With a single crew building down from Saalfeld beginning March 1, 1635, I would expect the track to reach Bamberg around February 7, 1636. That assumes a 1.75 miles/week rate most of the time, and 1.5 during the winter. If we could have a second crew working up from Bamberg at the same time, the construction time would be cut in half, with tracks running from Grantville to Bamberg by the end of July, 1635. To get the supplies to Bamberg, we’d probably have to ship them up the Rhine and then up the Main; Bamberg is a short distance from the confluence of the Main and the Regnitz. The necessary logistics would probably delay the northbound crew by 1–2 months, moving the completion date to the end of September, 1635. Alternatively, a single crew could finish by November, 1635 if the average construction rate overall for Saalfeld-Bamberg was about 2 miles/week. Which isn’t that much more than the 1.75/1.5 I assumed previously.


Carsten also advocated a cross-line, running from Jena both east to Gera and west to Weimar, Erfurt, Gotha, Eisenach and Horschel. It’s doubtful that any part of this line has been constructed as of October 1634, because Simon Jones then muses “it would be great when a spur went west. It would be worth a big detour not to have to travel from Erfurt [to Grantville] by horseback.” (TDI Chapter 18).

Unfortunately, there is a bit of an inconsistency in canon. In March 1634, Fischer says that after reaching Magdeburg in April or May, they would take a break, “then start working the other rail line out to Eisenach. By September, we should have covered the major points on both lines.” Hughes, “Turn Your Radio On” (Grantville Gazette 23). “TYRO” was written way before TDI, and the inconsistency no doubt slipped through for that reason.

There is also reference in canon to private interest in a line running from Erfurt-Eisenach to Hersfeld, Butzbach, Frankfurt am Main and Mainz; the route crosses the Fulda Gap. There were surveyors in Fulda in May, 1634, but construction had not yet begun. (DeMarce, “Prince and Abbot,” supra). Indeed, it wouldn’t make sense to begin it until the Jena-Eisenach connection was in service.

Passenger Cars

The first passenger cars were little more than stagecoach bodies with flanged wheels. On stagecoaches, the privileged rode inside, and the second class passengers on the outside. Since rail travel soon attained speeds at which the latter hazarded life and limb, the railway companies were forced to provided separate rail cars for their economy passengers. These vehicles were usually open-topped, and seatless.

In the 1830s, American engineers developed the modern four- or six-wheel passenger car, in which the axles, instead of being mounted directly on the car frame, were instead attached to independently pivotable “bogies” or “trucks.” Since the front wheels could turn before the back wheels, these American-style cars could negotiate the sharp curves of the North American rail lines with aplomb. Moreover, the cars could be longer. (NOCK/RE, 100-3, 134-5).

The passenger car bogie is actually quite complicated. Not only can it pivot, it can also move horizontally or vertically, independently of the carriage body, to some degree. So it can adjust to all sorts of track anomalies.

We can reconstruct this nifty piece of equipment after careful study of an up-time exemplar: the bogies mounted on the caboose which has fortuitously passed down through the RoF.

Anther technological problem which had to be confronted was vibration. Nineteenth-century American track was rather uneven (we didn’t believe in spending money on road beds) and hence some form of vertical springing was highly desirable. We also needed shock absorbers at both ends of the car, since the car came to a halt when it bumped into the one ahead. Rubber springs and buffers (see Cooper, “Bouncing Back,” Grantville Gazette 6) will help solve these problems.

Passenger cars were initially made of wood, and later of steel. A 1908 wood coach, carrying 70, weighed 41 tons. A 1920s steel passenger car, seating 90, weighed 70 tons. An eight car wood coach train might weigh (including the locomotive) 1725 pounds/seat, and a steel one, 2495. (Thompson 214). These would be a formidable load for the pickup trucks and horse teams alluded to by Quentin Underwood (1633, Chapter 34), and so it’s safe to assume that the first “cars” of the Grantville-Halle run are smaller and flimsier.

I cannot resist discussing a rather more unusual challenge. The Rottingdean Extension (1896) of Volk’s Electric Railway featured cars on stilts; they needed to be able to move through fifteen feet of water at high tide. These were, perhaps, the only railroad cars required to be equipped with a lifeboat. (Mike’s).

Freight Cars

Early cars were made of wood, later ones, of steel. The standard freight car has four axles, each pair mounted on a swiveling bogie. In the 1970s, while cars ranged from 30 to 175 tons in nominal capacity, most were 70 or 100 tonners. A 70-ton car had 33 inch diameter wheels; a 100-ton car, 36 inches. Wooden cars could carry their own weight, a steel car, two to four times as much. The exact load limit depended on the commodity for which the car was designed. Car lengths ranged up to 89 feet (not counting the couplers), with 40 feet being traditional. Freight car types include the box car, the covered or open hopper, the gondola, the flat car, and the tank car. (Armstrong, 65–7, 75, 129).


Traffic Control. Trains are controlled by “T&TO,” which stands for “timetables” and “train orders;” orders are given by telegraph and override the timetable. If central control is on-the-ball, and the trains follow their T&TO, then you won’t normally have two trains on a collision course.

T&TO is supplemented by a second layer of protection, which is the block system. The line is divided into blocks and only one train is allowed per track, per block.

In the manual block system, there were traffic guards at the junctions of the blocks, who communicated with each other. In the automatic block system, the presence of a train in a block is sensed electrically.

Blocks are equipped with signals which tell the trains whether to stop, proceed at full speed, or proceed with caution.

If the block is automatic, then in the absence of a train, current goes up one rail and down the other, energizing, through a relay, a “clear” signal. If a train is present, the steel wheels and axles create a short circuit, the relay opens, and the signal says “Stop!”

The signals themselves have taken a number of forms, including hoisted colored balls (giving us the term “highball”), semaphores, and colored lights. Perhaps the most important signaling principle, although it went unrecognized for a while, is that, if the equipment is faulty, it should not give an “all clear” signal. (NOCK/D, 7).

Brakes. It doesn’t do much good to receive a danger signal if you can’t respond to it; you never know when a villain has tied a pretty girl to the track ahead of you. Seriously, the train ahead of you may have made an unexpected stop. Or worse, is heading toward you on the same track. Or a bridge fails. Or a herd of animals decides to cross the track to see what is on the other side. And even if there is no emergency, the train must still be able to stop at its destination.

The kind of brake we want is, at a minimum, the Westinghouse air brake patented in 1869, and described in the EB11 “Brake” article. The brakeman in the locomotive turns a cock, so that compressed air flows out of a reservoir and into a pipe connecting all of the cars of the train. (The pipe in each car is connected to that of the adjacent ones by flexible hoses; ideally, these would be made from our dwindling supply of rubber, but if need be, a leather hose could be used.) The air presses brake shoes against the wheels, thus applying braking force to all wheels.

EB11 also describes several improvements in this brake. One is the automatic brake; the “Brake” essay provides a sectional view through the triple valve and brake cylinder. At least three further improvements were described, culminating in the Westinghouse “high speed” brake of 1894. A train of fifty empty cars, traveling 8 mph, could be stopped in 14.5 seconds, with a stopping distance of only 475 feet.

An alternative brake mechanism, the vacuum brake, is also described and illustrated.

The caboose in the museum once again comes in handy, as it should be equipped with even more modern brakes than those discussed in EB11.

Couplers. The original types of coupler were the chain-and-hook in Britain, and the link-and-pin in America, and had to be operated manually. Horrific accidents occurred when the railwayman was caught between the two cars. This led to the replacement of the manual couplers with the automatic “knuckle” or “Janney-type” coupler.

The efforts to design such a coupler began at least as early as 1874, but did not reach fruition until 1887. The Master Car Builders’ Association commented that it had been the “knottiest mechanical problem that had ever been presented to the railroad.”

It is therefore fortunate that the materials which passed through the Ring of Fire provide very detailed guidance as to how this solution can be duplicated. In EB11, Fig. 28 shows a perspective drawing of the coupler, and Fig. 29, is a plan view of the working faces.

Moreover, the caboose in the museum should have working couplers at both ends. (Yes, I know that cabooses were usually at the rear of the train. But I have checked several pictures of cabooses, and they can be coupled on either end.)


John Moody declared that “the United States as we know it today is largely the result of mechanical inventions, and in particular of agricultural machinery and the railroad.” The railroad will if anything play a greater role in creation of a prosperous and democratic society in post RoF 163x Europe.