By Paul Holmes


One of the enjoyable aspects about building a fictional ‘might-have-been’ model railway layout is the ability to look into known historical fact, and railway Acts of Parliament in particular, and then rewriting history to suit ones own purposes.

On July 25th 1865 The Croesor and Portmadoc Railway was granted an Act of Parliament obtaining powers to extend their line South from the wharves at Portmadoc to Borth-y-Gest and North to Beddgelert, thus regularising the privately owned Croesor Tramway. At the same time powers to run passenger trains and run steam locomotives were granted. Exactly 7 years later, on the same date, the Gorseddau Junction and Portmadoc Railway were granted powers to run steam locomotives too, whilst the old Gorseddau Railway was re-laid to 2’ gauge from its original 3’.

The North Wales Narrow Gauge Railways, meanwhile, built their line from Dinas Junction, 4 miles southwest of Caernarfon in the 1870s. The initial line terminated at Rhyd Ddu latterly renamed [South] Snowdon station, with a branch to the Bryngwyn quarries. Again numerous plans were laid and Acts passed granting powers that were never exercised!! The NWNGR had plans to create a network of 1’11½” lines around North Wales. Branches would have stretched across to Betws-y-Coed and down to Beddgelert where connection and probable amalgamation with the Croesor Tramway would have been made. Most readers, unless they are well versed in the early schemes that burgeoned in and around the North Wales slate industry will never have heard of these schemes that were prevalent in the mid to late 19th century, but this, if built, would have resulted in the line that was eventually built in 1923 – and called the Welsh Highland.

My interest in the Festiniog and North Wales lines goes back, as seems to be so often the case, to family holidays in the 1960s. We witnessed the push towards Tan-y-Bwlch and Dduallt as avid junior members of the FR Society. We rode in the 4 wheeled bug-box carriages behind the newly acquired Linda, Blanche and the Alco, and of course were always most pleased when the loco was one of the double Fairlies. My teenage layout was a combination of OO and OO9 but narrow gauge interests waned and I progressed to EM and then P4 – a model of Shipston-on-Stour, GW in the late 1920s.

However the desire to build something Festiniog remained and a house move 4 years ago provided the impetus to pickup the soldering iron again and look at a fresh project. Shortly after leaving University I had started to build a 4mm scale model of the ‘Little Giant’, one of the later batch of FR England locomotives. It was a first serious attempt at scratchbuilding in nickel and the body and tender were completed; but the chassis never ran, as in 1980 I was unable to find a supply of gears. The loco was nevertheless carefully stored, complete with a set of lathe turned drivers and the Faulhaber coreless motor, awaiting the day of completion. My first step was to get this engine out of its box and take a fresh look at getting something to run.

The intention is to build a model featuring two of the Festiniog ‘mainline’ stations, but first a little test piece was planned. As is so often the case, this layout is that test piece and the main layout is only now being started. I needed a small layout to try out new ideas. There are very few 1’11 ½” gauge layouts built to P4 standard in existence and various things were untried. I did not really know if I could work to such silly dimensions, and a set of workable standards needed to be formulated and tried. As a reader of the 009 News, you may be wondering why I am bothering to work to such standards – so do I sometimes!! Seriously, as with standard gauge P4, there are all sorts of advantages to working to scale dimensions. Clearances around outside pistons and crossheads are easier with the narrower gauge, but the distance between the frames in an inside framed loco becomes uncomfortably tight – 5.4mm to be precise with 10thou mainframes!!!

Thoughts turned to a suitable prototype to model, and I remembered reading about the early Croesor Tramway in the Boyd Festiniog books…..Just suppose that the extension to Borth-y-Gest had been built. All the 2’ gauge lines that eventually formed the Welsh Highland ended up merging with the FR on the harbour at Portmadoc, and if the Croesor had built the line around the headland from Portmadoc harbour to Borth-y-Gest, then both the FR and the Gorseddau lines would have had running powers into the station. Borth-y-Gest was, by the latter 19th century, already a minor centre for shipbuilding. A small number of sailing vessels, mostly schooners, were built there for the local slate trade.

Thus was born the ‘might have been’ that became the layout of Borth-y-Gest. The date is 1888 – a time when the newer locos like the FR’s final double Fairlie, ‘Livingstone Thompson’ had just been introduced, and two of the early 0-4-0 England engines had just been rebuilt as saddle tanks – but others, like ‘The Prince’ and ‘The Princess’ had not.

In my scheme, we suppose that the Croesor had built their extension, and the Gorseddau Slate Co Ltd had their main slate wharf on the site [thus justifying a model of their sole steam engine, a diminutive de Winton vertical boilered locomotive]. This was, until late 2001 where my history ended, but I have now [2003] rewritten the history books once more.

If we assume the lines to Beddgelert had been built as planned, then the NWNGR would have had their line from Dinas to Portmadoc by 1888, and the natural terminus would have been the newly developed port at Borth-y-Gest. There would then have been a triangular junction on the wharves at Portmadoc, west of the Britannia Bridge, where the line from Borth-y-Gest might have merged with the line exiting the Cob, running on to Croesor and Dinas beyond. I can thus justify building some NWNGR stock in addition to the FR stock running!!


My first attempt at devising a set of workable standards used the P4 figures with 11mm knocked off the track gauge, and everything else remaining the same. However this is wrong. As the narrow gauge prototype is in effect a scaled down version of standard gauge practice, scale narrow gauge is of necessity scaled down too. After building the first loco, ‘Livingstone Thompson’ to P4 standards with Gibson P4 wheels, and building points to match, it soon became evident that all was not right. I then happened on the 4mm scale narrow gauge group of the Scalefour Society and discussed the matter with Lynden Emery, Editor of the newsletter. New standards were thereafter produced, based on those used by the 2mm association boys. Table 1 is the result. I now use 2mm Association wheels, regauged to a back to back of 6.9mm, still running on 7.83mm track. Suitable gauges have been made, although I rely more heavily on my digital micrometer than the track gauge.

Gauge	Back to back	Tyre width	Flange width	Flangeway	Crossing clearance
7.83mm		6.9mm			1.3mm		0.25mm			0.5mm		7.33mm


The harbour at Borth-y-Gest lies in front of the village, and is situated just around the headland from Portmadoc harbour. I have imagined that the line ran out from in front of the cliffs but behind the slate warehouse of the Gorseddau Slate Co. The station is built on a wharf that runs across the bay, creating a breakwater for the harbour behind. The wharf was formed from the ballast dropped by incoming slate ships, before loading up with Welsh slate, bound for the rest of Britain and also Northern Europe. Some thought was given to the aesthetics of different shaped wharves and a sweeping curve seemed a good idea, with tracks sweeping through – so much more interesting and [hopefully] prototypical than a boring rectangle. There is a spur running back from the mainline along the coast to the loco shed and to the remote gunpowder store. Gunpowder was brought in also by ship in barrels and stored separately – to be loaded into special timber lined iron closed wagons; these were the only private owner wagons on the FR.

I personally find that layouts look much better if presented as a diorama, and this is no exception. The layout is now 8’ by 2’ in size and sits at around 3’ 9” from floor level. The backscene rises some 15” above track level, so it is difficult to look over from the back, but I find this gives a better overall effect. The layout is then framed with a pelmet containing the lighting units. The idea of using a proscenium arch is of course lifted from the stage, but after all, we are aiming to present a bit of theatre. The backscene is painted with an emulsion sky scene and the harbour and village were painted in acrylics.


Construction is of the all ply school, with a mix of 4mm and 6mm ordinary ply, with joints strengthened with 20mm square planed softwood. All panels are glued and pinned – virtually no screws have been used. As the track level is some 50mm above the ‘sea level’ the construction took a little forward planning but is not particularly complex. The backscene forms an integral part of the construction.


Festiniog rail has evolved over the years, from very early lightweight rail fixed to stone sleepers, up to the double-headed rail in use by the 1880s. The rail in common use by this era had a rail height of 4.25” and a width across the rail running surface of 2”, I did look at using 2mm association rail, but the running surface is too narrow, and the rail height at 1mm is too low. Although the rail I use – Peco Z gauge code 60 flat bottom is strictly speaking wrong as it is flat bottom in section, this does not matter because throughout the Victorian period photographic evidence confirms the fact that the track was totally ballasted over. Both the sleepers and the rail base are complexly covered in stone, and the rail running surface is the only part visible. The Code 60 rail has a good match in this dimension.

Track is all laid on the usual base of cork sheet – not as quiet as the Rice style flexible base – but much more rigid. Standard gauge copperclad sleepers are cut in half; totally out of scale but eventually invisible, and glued down with Evostick. Rail is soldered as usual and points were made separately before hand. Problems were encountered with tiebars, which were eventually made from 1/32 ply and fixed below sleeper level. This has actually proved unsatisfactory and the technique will not be repeated. It is difficult to find a way of making tiebars that are

1. Robust
2. Invisible
3. Hold point blades only 6mm apart
4. Will be buried in ballast

I should welcome suggestions. In addition examination of photos shows that the sleepers are nowhere in evidence, even on pointwork – but 4mm stock will not run through gravely track like the prototype does. Initial experiments with home made point motors using solenoids from Farnell were not very successful either, so most points now use ‘tortoise’ motors.

Electromagnets are buried in the track work to operate the DG couplings.

Ballasting was a difficult task, left until recently because I hate doing it at the best of times. Fine granite ballast was used, but after laying it looked wrong, both in colour and in texture. I realised from looking at photos that the ballast should really be stone of around 1” and down i.e. stones down to dust size particles are present in the mix. Model ballast tends to one size graded chippings – it is the difference, in terms of building materials, between building sand, which has variable particle size and coarse sand where the particles are all the same size. After applying the stone carefully, the track was flooded with PVA diluted with water and a little washing up liquid. After it was dry the flangeways were cleared using a jeweller’s screwdriver and than all the ballast scrubbed with a larger screwdriver to give the impression of that worn powdery effect.

The entire track was then painted over the course of several evenings with washes of dilute acrylic paints to lighten the ballast and give a varied colour finish. The lighting in the pelmet is with standard fluorescent tubes; I know a lot is talked about colour temperature and these tubes, but my feeling is that as long as the painting is done in the same light and not under light bulbs, or daylight, then the effect will be right. Problems can occur if models are painted under different lighting conditions. I have noticed the greatest effect on crimson lake red, which appears darker and more purple under fluorescent strip than incandescent bulbs, under which it appears a brighter red.


Control is with a pair of Gaugemaster controllers wired through a common return system to the section switches and point switches – the inbuilt switches in the Tortoises switch the common crossing polarity. A colourful track diagram contains bicolour LEDs indicating track power and point selection, a fun-wiring situation. More recently a Pentroller has been used – the difference in smoothness of operation is quite remarkable. This unit was used exclusively at the recent 2002 Scaleforum in Leatherhead and performed faultlessly.


Structures are all scratchbuilt from card or plastic. The station building is based on the station at Tan-y-Griseau from mounting card, treated with shellac in the form of button polish. The goods shed; slate warehouse and water tower are from Wills scenic plastic sheets, as is the dockside. The loco shed is card covered in Das clay. I have used a variety of materials in order to experiment for the full layout to follow. I have found the Wills sheets good, but the plastic is so thick it is difficult to cut. My preference has always been for card, and when used carefully Das gives a good impression of both cut and random stone. One has to take care that the stones are cut carefully into the wet clay and that after partial drying any heaped up edges are pushed back and the faces slightly roughened to give a hewn appearance. I also have a long standing hatred of plasticard and its tendency to warp. This seems to happen either immediately after construction or even months or years later. Some people don’t seem to have the problem – but I am still mistrustful. The exquisite North Eastern coal hopper wagons, to diagram P4 in the recent MRJ, are a case in point. The original carriage built for this layout, one of the two bogie carriages Nos. 15/16 was constructed in plasticard back on 1980. It has warped around the underframe area below the end balconies – it has to be placed on the track the right way round!!

I wanted to experiment with dry stone walls, in order to develop a technique for the full size layout. The technique used was an adaptation of others and I feel works quite well. A piece of ‘Das’ was rolled out with a rolling pin onto a kitchen chopping board to around 1.5 to 2mm thickness [take care to ask the domestic authorities or hide your work]. Before it sets, scribe heavily through with a Stanley knife at 3 – 4mm intervals – no need to be too careful, except with your fingers! When almost dry, probably the next morning, again smooth off the rolled up edges. When fully set snap the piece into strips. Individual ‘stones’ were then snapped off the strip and the wall built up using a mortar of cheap PVA glue with a little black acrylic paint added. A short length of wall can be built up surprisingly quickly, as the wall can only be seen from the front, the slight variation in depth does not matter. The top course of stones was set as upright pieces. When thoroughly dry, the wall was painted with thin washes of acrylics in a black/white/yellow ochre mix, with added greens and blues to give a weathered lichenified effect. Sparing use of dark mortar means that there are no nasty lines appearing where the paint does not soak in and the stones are picked out in subtle shades of varying colour. The stones are probably a little overscale, but I do not think that matters, as the effect is definitely one of a dry stone finish rather than a coursed wall.


The great thing about modelling a small prototype like this or any of the light railways is that there is a finite number of models that can be built. In 1888 the Festiniog possessed five 0-4-0 England engines, one 0-4-4 Single Fairlie [Taliesin] and three 0-4-4-0 double Fairlies. Once they have all been constructed, there are no more to build! Similarly, they possessed six bogie coaches, built to three diagrams, and around twelve 4 wheeled carriages. Goods stock is again limited in number, although there are a great number of very similar slate trucks. On top of this, as mentioned the Gorseddau had one de Winton vertical boilered loco.

So far five locos are running and two are ‘under repair in shops’.

‘Little Giant’ Loco modelled as running after 1888 rebuild with new enclosed cab and new saddle tank, but original frames and wheels. Body scratchbuilt in nickel. Scratchbuilt chassis with lathe turned drivers from mild steel. Split frame pickup using 2mm Assoc system with tufnol block frame spacers, 2mm Assoc gears and axle muffs. Faulhaber 1616 motor.

‘Palmerston’ This was the first England engine to be rebuilt in 1887. Originally a side tank the rebuild created the mould that all the other small engines were to follow. Body started out as a white metal kit produced for the Festiniog Portmadoc shop in the 1960s – a few were available from the 009 Society last year in unbuilt state. The kit had tanks 3mm too long, which were rectified with the piercing saw. The total lack of any rivet detail was overcome by making a tank overlay from 5 thou nickel sheet with embossed rivets. This was glued in place with cyano. Again chassis scratchbuilt as per ‘Little Giant’. Motor Faulhaber 1219.

‘Livingstone Thompson’ The final double Fairlie built for the railway in their own workshops at Boston Lodge in 1885-6, as a replacement for the worn out ‘Little Wonder’. Model built from Backwoods Miniature brass kit, bodywork much as supplied. Much fun and games was had with the bogie chassis – this is a tale in its own right. Frames had to be respaced due to the 7.83mm gauge in place of 9mm gauge. Wheels are Gibson P4 profile. This loco is still not running correctly and awaits another chassis rebuild.

‘Merddin Emrys’ Loco was first all new build at Boston Lodge in 1879. Essentially an enlarged version of the earlier ‘James Spooner’. Model is a modified Backwoods Fairlie. Tank sides are 9” shorter. Under cab tank is longer. Different chimneys and faring on boiler front all characterise this engine. This one works!!! Has slimmed down Gibson P4 wheels with correct 6.9mm back to back. Fitted with specially ordered Maxon 19 × 13mm coreless motor with double ended shafts. Easily ordered from the UK suppliers, but had to wait as it was a special from Switzerland!!

de Winton Built in 1870s in Caernarvon as small shunting locos for quarries – essentially a steam mechanical horse. Model from the kit by Saltford Models, but heavily modified to represent the Gorseddau Quarry loco. Tiny Mashima motor mounted vertically in vertical boiler!! Dreadfully low gearing so has now been rebuilt with watch gears at about 100 to 1 ratio – but still doesn’t really run like I want it to!!

‘Snowdon Ranger’ Just finished in time for Scaleforum 2002. Based on the rather primitive whitemetal kit from Chivers Finelines, but heavily adapted with 5 thou nickel overlays to incorporate rivet detail and a reprofiled cab side with proper open cab doorways – the kit has no openings!!!! Power is from a 1610 Faulhaber driving through a carden shaft to the power bogie. Wheels are 2mm Association with turned nickel overlays hiding the spokes.

‘Taliesin’ The Festiniog’s only single Fairlie, quoted as being the enginemen’s favourite locomotive, being fairly powerful, yet faster than the double engines. The model is again adapted Backwoods etchings – this kit is an accurate portrayal of the locomotive in service now –rebuilt Boston Lodge 1999. But I needed the loco as built originally in 1876!! Major surgery has again been needed with lower tanks, totally scratchbuilt open half cab and old style sandboxes. Faulhaber 1610 motor driving through carden shaft to power bogie. 2mm Association wheels with split frame chassis and compensated trailing bogie for extra pick up


The usual mix of kits and bits has been assembled. There are still woefully few slate wagons built – it is so boring building dozens of identical items!! I must set to again. Fortunately the Parkside plastic kits of both 2 and 3-ton wagons are accurate and relatively quick to build. I carefully shave off about 0.25mm from the inside of the axlebox, then drill out the axlebox to 0.95mm dia hole and insert a 2mm Assoc pin point bearing with a dab of cyano. The 2-ton wagon is then assembled without floor and a new floor then cut from 1mm thick lead sheet and fixed in place of the plastic floor, again with cyano. The size of this piece is critical as if slightly under size then it falls out, but if oversize then it spreads the axleboxes and the wheelsets fall out. With care, a set of 2mm Assoc wheels on 12.8mm axles will run nicely. I currently am using their standard 6mm diameter wagon wheels – not quite correct but I feel that the spokes are so small and virtually hidden that the quality of the wheel makes up for the extra 2 spokes. The lead floor gives vital weight without upsetting the appearance – there is certainly no room for added weight anywhere on the wagon.

Other wagons have been scratchbuilt in wood and metal. Several low-sided wagons and small coal trucks have been made with 1/32” ply bodies and brass chassis, using thick paper as strapping. The F R 6 wheel Cleminson coal truck has just been finished in nickel and one of only two bogie ballast wagons – again from ply. One gunpowder van, the only private owner wagons to run on the line, has been built in brass.

4 rakes of coaching stock are now running. The passenger trains comprise a mix of scratchbuilt [plasticard] bogie composite, 4 wheelers from Colin Ashby and the curly roofed van from Worsley Works. The quarrymens’ train is from the Parkside stable. More recent carriages are all Worsley Works. Early on in construction, the carriages were compensated using the rocking solebar method – I now know from experience that this is totally unnecessary and more recent stock is built with a rigid wheelbase. Much easier to make it all square and certainly these short wheelbase items of stock seem to run as well in uncompensated state.


As a test bed for ideas, the railway has been very useful. P4 narrow gauge can work and I think the extra effort is worthwhile. I have found a few problems that have been reconsidered as I commence work on the big layout. I am not happy with the system I have developed for point tiebars – I have now started to build pointwork for the new layout and have rethought the tiebar problem. More of this anon, but the new layout uses code 60 bullhead rail from the 3mm Society stores, coupled with C & L copper clad fibreglass and plastic sleepering. I have found the Tortoise point motors reliable and will continue to use them, despite the expense.

I also have a hankering to build a schooner to sit in Borth-y-Gest harbour – perhaps the ‘Cadwalader Jones’ an 80’ 2 masted vessel actually built at Borth-y-Gest in 1878. So plenty of future projects to keep the brain active, or maybe a live steam loco in P4 – 7.83mm gauge of course. ‘Beddgelert’, the largest of the NWNGR locos would make a possible project. It was a hefty 0-6-4 saddle tank from Hunslett in Leeds built in 1878… only joking!! Work has actually started on the big project – Minffordd with Tan-y-Bwlch. The top (Blaenau) fiddle yard and Creuau bank are now taking shape and I will report back as work progresses.