Ballarat trains looping through Bacchus Marsh

One question I’ve seen people ask many a time is why the Ballarat line loops around after passing through Bacchus Marsh, instead of just running in a straight line. But the answer is simple – there is a bloody great big hill in the way.

Going for a ride

We start on the Melbourne side of Bacchus Marsh, looking down into the valley below.

Carriage set FSH25 trails P18 through Parwan, bound for Bacchus Marsh

Pass over Parwan Creek on a curved embankment.

Three car VLocity rounds the Parwan curves out of Bacchus Marsh

Then into another cutting.

Vlocity climbs out of Bacchus Marsh for the Parwan Curves

To finally arrive into Bacchus Marsh station.

VLocity VL41 departs Bacchus Marsh on the up

But that was only a taste of the journey ahead – we’ve got an even bigger hill to climb.

VLocity VL52 leads a down Ballarat service out of Bacchus Marsh

Trains get a short respite from the climb at Maddingley, where V/Line have a stabling yard and crossing loop.

N469 in the yard at Maddingley, having run around the carriage set

But five minutes later, you’ll see the same train again, but far above your head, and running in the other direction.

Five minutes and a horseshoe curve later, the VLocity train is still climbing Ingliston Bank

Having rounded the horseshoe curve.

VLocity VL24 rounds the horseshoe curve and climbs Ingliston Bank

Then hugging the hillside.

VLocity VL24 climbs Ingliston Bank bound for Bank Box loop

Don’t look down – the V/Line stabling yard is back at the bottom of the hill.

N467 stabled on carriage set LH33 at Maddingley

But trains still have further to climb.

Three car VLocity 3VL37 on the way up Ingliston Bank

Now running in a cutting hewn out of the rock.

Three car VLocity 3VL37 on the way up Ingliston Bank

The grades ease off again at ‘Bank Box’, where a crossing loop allows opposing trains to pass.

VLocity VL07 trails a down Ballarat service through Bank Box

Then it’s off into the rugged bushland of Werribee Gorge.

VLocity VL14 runs around the curves near Bank Box loop

A tall bridge crossing between the ridges.

VLocity crosses a bridge through the Werribee Gorge

Until the railway line finally rejoins level ground, and can leaves the curves behind.

VLocity VL04 and VL17 head towards Bank Box Loop for a cross

So why does it curve around so much?

I’ve written about the history of the Melbourne-Ballarat railway before – born as a line headed east from Ballarat towards Ballan in 1886, and a second branch west from Sunshine towards Bacchus Marsh in 1887 – the hills outside Bacchus Marsh presented a formidable barrier.

Stabled Sprinter consist beside a carriage set stabled for the weekend at Bacchus Marsh

Topographic maps showing the steep country around the Werribee Gorge.

Bacchus Marsh 1:50 000 topographic map, Geoscience Australia

A newspaper report from the period describing it as.

There still remains to be constructed that portion of the line extending from Bacchus Marsh station to Ballan, a distance of 17 miles.

The route of this section passes through most difficult country, there being a rise of over 1800ft in that distance The earthworks on this section alone will necessitate the removal of no less than 1¼ millions of cubic yards, cuttings for long distances being upwards of 40ft in depth while the embankment will be correspondingly high, in one instance more than 100ft.

These heavy earthworks result from the fact that the line has to be carried along the northern spurs of the great plain which extends from between the Werribee River and Little River to Port Phillip. The steepest gradient upon the line will be 1 in 48.

It will be seen from the figures given that the through railway is far from being an easy line to construct. The country through which it passes is so broken and difficult as to compel the provision of very large works, while the earthworks are upwards of 50 per cent greater than upon the line by way of Geelong.

And the benefit of the new line.

The present distance to be travelled by rail between Melbourne and Ballarat, by way of Geelong, is 100 miles, but when the direct line is completed that distance will be reduced to 74 miles, a saving of 26 miles been thus effected in the journey to Ballarat.

But despite all the massive earthworks, the railway still required a large horseshoe curve outside Bacchus Marsh to attack the hill.

Bacchus Marsh 1:50 000 topographic map, Geoscience Australia

But still the railway had one of the steepest grades on the Victorian rail network – a 1 in 48 climb all the way from Bacchus Marsh to just outside Ballan.

Victorian Railways grades and curves diagram

What about an alternate route?

I’ve seen it on Twitter, I’ve seen it on Reddit – but even when the railway was brand new, people asked why the railway descended into Bacchus Marsh, only to climb back out again.

There has been a question as to the wisdom of taking a fast passenger line down and up the long gradients of Bacchus Marsh to reach the township. It is alleged that the extra haulage required will be fatal to economy and fast service, and that the Ballarat and intercolonial passengers will gain only a few minutes in time, with an added risk of accident. It is contended that a route on a level could have been found encircling the Marsh, which would have secured every advantage.

It was asked again by the late W. Williams in his book “A History of Bacchus Marsh and its Pioneers“, and serialised in the Bacchus Marsh Express.

The horseshoe bend of the railway near the town is a puzzle to many. The question is asked : Why did the line undertake this acrobatic performance? Why did it not pursue its even course on the plateau?

Then follows an ominous shake of the head — “I suppose some job again,” and immaculate departmental purity suffers defilement at the hands of an undiscerning public. For a passing moment the pictorial style of diction is indulged in. I see the train like an elongated caterpillar crawling up yonder summit, with two engines before, and one behind to give a friendly help in time of need.

Straight across to the plateau is only two or three miles, and yet to reach that identical spot the concentrated procession of engines and cars has travelled a circuitous course of eight miles.
On the face of it some mistake appears to have been made, especially when it is remembered that the section is part of an interstate line, in connection with which time is the essence.
of the contract.

But taking a shorter route would have only given minor time savings.

The answer is that the exclusion of Bacchus Marsh by taking the outer route would not be compensated for by any practical gain; that the saving in distance would be only 29 chains, and that the gradients are not such as will seriously prejudice the traffic, and that of all the 17 surveys that have been made none have shown a route that did not join the selected route at the point where the question of gradient has any force.

Alternative routes having been surveyed.

The line as constructed between Parwan station and the Dog Trap reservoir is about 24 chains longer than the route surveyed via Ryan’s corner (near Parwan station) and Collie’s bridge (close to the Dog Trap reservoir), between the same points. The two lines rejoin at the road about midway between the crossing of the Parwan Creek and the Dog Trap Gully on the Ryan’s corner route; and therefore the long gradient of 1 in 48 from the Parwan Creek on towards Gordon must have been the same upon either line.

On the Ryan’s corner route the gradients would have been comparatively easy between Parwan station and the crossing of the Parwan Creek, where the steep ascending gradient of 1 in 48 commences, which is common to both routes. The length of ascending gradient of 1 in 48 upon that portion of the Ryan’s corner route between Parwan station and the Dog Trap reservoir would have been about 110 chains in length, against 240 chains of 1 in 48, and 27 chains of 1 in 50 ascending, and 90 chains of 1 in 49 and 45 chains of 1 in 50 descending gradient upon the line as constructed between the same points.

If the line had been constructed via Ryan’s corner it would have been necessary in order to accommodate Bacchus Marsh, to construct in addition a branch line to that place. And a viaduct would have been required over the Parwan Creek of similar construction to that over the Werribee River. The length of this would have, been about 300 feet, and the height about 120 feet.

A similar bend of about six miles occurs between Mitcham and Belair on the Adelaide end of the overland route. Those places are about two miles distant from each other the crow flies, but are eight
miles apart by railway line, the cause being that Mitcham is 241 feet above sea level and Belair, 1008 feet above the sea.

Now let the doubter for ever hold his peace. Abundant evidence is supplied that from an engineering point of view “the horseshoe” is valid performance, without the faintest approach to any “fishy” aroma.

The horseshoe even seen as an engineering wonder.

It was a great engineering feat to make that railway call at Bacchus Marsh, and then to scale the plateau and the glory involved in same is to be shared by three engineers, Mr Leo. Cussen, Mr. G. C. Darbyshire and Mr. W. C. Billings.

Versatile indeed was the genius of Mr. Leo Cussen, for how seldom in the one mind is there the conjunction of the literary and the mathematical faculty; but law now claims the erstwhile Engineer as one of its brightest ornaments, for in Judge Cussen we have one whose judgments are almost invariably beyond successful appeal.

But the operational cost of the line was high.

It is well known amongst railway employees that the engine drivers and guards who travel over the line dread the journey from Gordons to Bacchus Marsh, for although the distance between those stations is only 25 miles, yet there is a drop in that distance of 1536 feet, in negotiating which the brakes have to be used continuously whilst running down steep embankments and ugly looking curves.

Owing to the heavy pulls between Bacchus Marsh and Ingliston, where the line rises 1170 feet in 13 miles, it requires two engines to take the Adelaide express when heavily laden to Ballarat, and
when only only engine is used on a light train it is assisted as far as Ingliston by a “bull-dog” engine that is always stationed at Bacchus Marsh for that purpose.

Naturally the cost of haulage is unusually heavy, and will in a large measure explain the fact that last year the loss incurred in working the line was £1857.

And as trains grew bigger and heavier, the sharp curves and steep grades grew ever more limiting.

Pair of B class diesel-electric locomotives haul 1300 ton load up Ingliston Bank, 20 August 1952 (PROV image VPRS 12800/P1, item H 2545)
PROV image VPRS 12800/P1, item H 2545

So the idea of shortening the route was brought up from time to time, such as this 1976 report on upgrading the Melbourne to Serviceton railway.

Parwan to Horseshoe Creek Deviation

This scheme, costing $1.8 million, includes construction of an 8 km long deviation between Parwan and Horseshoe Creek, bypassing Bacchus Marsh to avoid the descent to Bacchus Marsh followed by a steep climb. It also includes keeping the existing track for use by commuter trains originating from or terminating at Bacchus Marsh. A 6 minute reduction in transit time is expected for both directions of travel if the deviation is introduced. Furthermore, the reduction in transit time would render one crossing loop unnecessary.

But it took the Regional Fast Rail project to finally do something about it.

Government propaganda sign at Deer Park spruiking the Regional Fast Rail project

Straightening the curves over Parwan Creek in the descent into Bacchus Marsh.

VLocity winds through the Parwan Curves descending into Bacchus Marsh

With the construction of new cuttings and embankments.

VicPlan map

And a 8.2 kilometre long deviation between Millsbrook and Dunnstown.

VLocity Melbourne bound crossing the Moorabool River on the Bungaree deviation on the Ballarat line

But with the measure of success for the project being the 60 minutes “Country Express Run Time” between Melbourne and Ballarat, the tracks through the horseshoe curve and up to Bank Box were left with timber sleepers – saving the Victorian Government $404,110.

Flogging upgrade at Dog Trap Gully

Only to end up being replaced with concrete sleepers a few years later anyway.

Passing track work near Ingliston

But since then focus has rightly moved from raw speed, to frequent and reliable trains – leading to the recently completed Ballarat Line Upgrade project, which delivered double track to Melton, and a second track and platform at Bacchus Marsh and Ballan stations.

VLocity VL60 and VL63 depart Bacchus Marsh on a down Ballarat service

That’s a change in focus that I can get behind.


Here is the driver’s view of the slow climb up from Bacchus Marsh to Ballan.

Liked it? Take a second to support Marcus Wong on Patreon!
Become a patron at Patreon!
You can leave a response, or trackback from your own site.

14 Responses to “Ballarat trains looping through Bacchus Marsh”

  1. Joe Healy says:

    Fascinating stuff. I love the history and perspective you provide on these topics. The pictures of the trains at each stage would have taken awhile to capture, but really add to the article.


    • Marcus Wong says:

      I’m glad you enjoyed the post – the photos are the product of 15 years out and about following trains along the line, I just had to sort through them all to find appropriate ones to suit the story.

  2. Andrew says:

    I believe the ‘hill’ is actually the ‘Bacchus Marsh Syncline’ which extends all the way from Geelong to Bacchus Marsh. It’s caused by the Rowsley fault.

    The tall bridge is known (officially?) as Ironbark Gully.

  3. Tom the first and best says:

    The Parwan-Horseshoe Bend bypass of Baccus Marsh made a lot of sense, when the line was a major freight route and had long distance passenger trains (Adelaide, Mildura and Horsham/Dimboola) running express through Bacchus Marsh. The diversion of some express services via North Shore, as part of the New Deal for Country Passengers, may not have occurred.

    After Kennett`s cuts and Adelaide standardisation via Cressy, it would be less useful (presuming its existence did not cause standardisation to go via this route), mostly used for peak expresses and transferring trains to Ballarat East for maintenance, if it had not been closed. RFR might have used it, instead of upgrading the Parwan curves into Bacchus Marsh, due to its peak express focus.

    • Marcus Wong says:

      Agreed on that point – with all passenger trains stopping at Bacchus Marsh and only a handful of off-peak freight services the need for a bypass doesn’t really exist anymore.

  4. cat says:

    Really interesting. Thanks

  5. peter flynn says:

    Easy to blast a tunnel throught that tiny rise just get a few Europeans who know hoW to build straight rail ; What a crock to say hill in the way easy to blow that away these days

    • Marcus Wong says:

      West of Bacchus Marsh it’s actually an escarpment not just a hill – so you’d have to have an uphill grade in the tunnel to get you back to the surface anytime before Ballarat.

  6. James Dixon says:

    Excellent post as always Marcus. This reminded me of a long lost article in Newsrail (which I found – February 1985) which was fairly speculative but suggested if a standard gauge line was to be built to Adelaide it should go via Ballan rather than Cressy. The author suggested that a number of diversions could be built to ease the grades and there were a number of diagrams provided. History went against the author (who isn’t named) but at least one of their suggestions – bypassing the loop through Gordon and Bungaree – came to pass.

    • Marcus Wong says:

      Thanks James! I remember reading that same Newsrail article years ago – many of the deviations were minor ones to avoid the foothills of Mount Langi Ghiran and the like.

  7. […] you’ve read my past items on the history of the Ballarat line and the looping section of track outside Bacchus Marsh you know the story by now – the line was born as two single track branch line serving towns […]

  8. […] time around it was ABC Radio Ballarat, who had seen my recent posts on the Ballarat line through Bacchus Marsh and curve easing for faster trains, and thought it would be of interest to their […]

Leave a Reply

Your email address will not be published. Required fields are marked *