Three gauges, one railway

For most of the twentieth century, a passenger who wanted to leave a trunk route and continue on a branch line had to change trains at the junction — not because the branch was a different service, but because the two tracks were a different width. The branch was metre gauge, 1000 mm between the rails. The trunk was broad gauge, 1676 mm. No locomotive, coach or wagon could pass from one to the other. Passengers walked across the platform; freight was unloaded by hand, basket by basket, and reloaded.

That is usually filed under railway history. It is really a case of how infrastructure accumulates. A gauge is not a preference that can be revised later; it is a number already cast into embankments, bridge girders, tunnel bores, platform edges and several thousand vehicles. India's network carries three of those numbers because it was built by different people, in different decades, with different amounts of money — and the bill for reconciling them has been running since 1992.

Three widths, three decisions

The first line, Bombay to Thane in 1853, was broad gauge: 1676 mm, or 5 ft 6 in. The choice was argued the following year by Lord Dalhousie, then governor-general, in a minute that recommended 1676 mm as the standard for the whole country. The reasoning was mechanical: a wider track allows a wider, taller vehicle that carries more for the same train length, and leaves room for a bigger boiler. In a country with no network yet, the extra width is paid for once, in earthworks and bridges.

Then the arithmetic changed. Most early lines were built by private companies whose shareholders were guaranteed a fixed return by the government, and from the 1860s the government wanted to connect more districts with less capital. In 1871 it adopted metre gauge — 1000 mm — for lines where traffic was expected to stay light: district headquarters, cotton towns, pilgrim routes. Narrow gauges of 762 mm and 610 mm followed where the alignment was the constraint rather than the traffic: hill sections, tea gardens, sugar mill sidings.

The result was a network that looked unified on a map and was three networks in the metal. Trunk corridors were broad gauge. The feeders that connected district towns to the trunk were often metre gauge. Hill and industrial lines were narrow. Every one of those decisions was locally sensible, which is exactly why the pattern is hard to undo: nobody chose a three-gauge railway. Each administration chose a gauge for its own line.

What a gauge actually fixes

The distance between the rails is the visible part and the least of it. Once that number is set, it propagates: sleepers long enough to hold the rails, a formation — the width of the embankment or cutting — wide enough for sleepers plus shoulders, bridges wide enough to span it, tunnels bored to clear the resulting vehicle, a platform edge at a fixed distance and height above the rail head.

Weight is the part that is easiest to miss. Broad gauge in India is designed around an axle load of about 22.5 tonnes; metre gauge around 13. A bridge designed for 13-tonne axles is not a bridge that will take 22.5-tonne ones, however good the new rails are, nor is the culvert beside it. Formation width is not a detail either: a broad-gauge single line wants roughly seven metres of embankment where metre gauge managed with about five, and on a hillside those extra metres mean cutting further into the slope or building a retaining wall.

Then there is the fleet. Coaches, wagons and locomotives are built to a gauge: bogies, underframe, brake rigging, buffers. Re-gauging a locomotive is not a workshop job anyone does at scale, so the stock is written off and replaced. When a line changes gauge, the track is the cheap part. The expensive part is everything that was dimensioned to the old number.

Metre gauge was not a mistake

It is tempting to read the metre gauge as penny-wise folly; at the time it was not. A metre-gauge line cost substantially less per mile: lighter rail, smaller earthworks, cheaper bridges, sharper curves, less land acquired. Under the guarantee system, where the state underwrote the investors' return, a cheaper line meant a smaller annual payment for the same length of railway. Given the choice of one broad-gauge line or two metre-gauge lines through two districts, metre gauge connects twice as many people for the same money. For the traffic it was built for — a cotton mill, a market town, a pilgrim season — it was the correct answer, and it stayed correct for about seventy years.

It stopped being correct when the traffic grew. A metre-gauge trunk route carrying coal, cement, foodgrain and people runs out of capacity long before a broad-gauge one does, and the junction where the gauge changes becomes a bottleneck with a workforce attached to it: goods are handled twice, by hand. By the 1980s the argument for a single gauge was not aesthetic. It was a capacity argument.

Unigauge, from 1992

Project Unigauge was announced in the railway budget of 1992, with the aim of converting metre and narrow gauge lines to broad gauge instead of running three separate fleets. At that point broad gauge was roughly half the route length. It is now over nine-tenths, and what remains is a short tail, part of it kept on purpose.

On the ground, conversion is a sequence rather than an operation. The formation is widened. Bridges are rebuilt or re-decked for the heavier loading. Tunnels are opened out or re-bored. Track centres are re-spaced, which means platforms are re-cut, because the platform edge sits at a fixed distance from the rail and a wider vehicle needs more clearance. Then the line is relaid with heavier rail on concrete sleepers, resignalled, and re-equipped with stock that has to be newly built or cascaded in.

Because it all happens on a railway still running trains, the work goes in patches, a few hundred kilometres a year across the whole system, fitted around new line construction and electrification, which compete for the same capital budget. Conversion is therefore not a project with a completion date. It is a standing item of expenditure that has now run for more than thirty years.

What is left

Some lines were never converted. Where the traffic never arrived, the metre gauge was simply closed, and the embankment is now a road or a strip of weeds. The hill railways were kept narrow deliberately: Darjeeling Himalayan and Matheran at 610 mm, Kalka-Simla at 762 mm, Nilgiri on metre gauge. A wider vehicle would not clear those curves, and no freight case justifies rebuilding.

What conversion leaves behind is harder to photograph: withdrawn stock, retooled workshops, fitters whose specialty has no future. Standardising a network costs the new work and writes off the old.

The general lesson is worth more than the railway. A gauge is a decision that gets copied into everything downstream of it, and once copied it cannot be argued with. India spent a century running three widths because three generations of engineers each made a defensible choice with the money they had, and it will spend decades more reconciling them. That is what infrastructure debt looks like when it is made of embankments instead of bonds: nothing appears on a balance sheet until somebody decides to fix it.

This is a piece in Everyday Life, part of About India.