This is the part I have really been looking forward to: laying individual prototypical rail lengths and bolting them together with fishplates... The CMER (as a reminder) was originally laid with a mixture of 12lb/yd and 25lb/yd rails, which are 2" and 2¾" high respectively. The main and branch lines were gradually all relaid with the 25lb/yd rail in 15' lengths, while the 9' lengths of the lighter rail were used in sidings. The original 36" sleeper spacing was retained in sidings, however as rails frequently distorted, subsequently during maintenance the main line and the Clachbeg extension were relaid with a seven sleepers per 15' rail spaced at 12½"–2'7"–2'7"–2'7"–2'7"–2'7"–12½" centres and on bridges eight sleepers spaced at 1'–2'2"–2'2"–2'2"–2'2"–2'2"–2'2"–1' centres; turnout sleepering is also spaced at around 24" centres.

Scale lumber stacked up: 6" × 4" sleeper timber in the foreground and 8" × 4" turnout timber in the background. This is lovely basswood from the USA, from Kappler Scale Lumber.

I started with the template of the track from the fiddle yard turntable through the cutting to the bridge. three full rail lengths plus a very short piece at the left end of the picture. I cut the requisite sleepers and roughly sanded off the edges at their ends. As these ones will not be seen I have not distressed them at all.

I used the Fasttracks rail bender to bend the Code 205 rail to fit the Templot-generated template. I had to order this rail from the US as no-one in the UK stocked it; to save on shipping costs I arranged with Microengineering that they cut their stock 6' lengths into five 14 13/32" lengths, which works out as 16'5" scale: ideal for me as I would be cutting each length anyway.

The short rails at the turntable end of the template and the joint. I left a ½" gap between rails (handily the thickness of a piece of card stock). The short rails are straight and about 3' long: they form the transition between the 33' radius curves of the fiddle yard tracks and the cutting track.

In my impatience I decided to glue the sleepers in place at this stage... it would have been better to stain them first! Still, easily-enough remedied later.

Drilling the four holes in a fishplate, using a very rudimentary jig and simple scribed measurements.

Platelaying ironmongery: 500 bolts and 500 nuts as well as two nut spinners for the M1 nuts.

Fishplates!

For the Code 205 rail I use 30 mm long pieces of 2.5 mm × 1 mm brass strip from Eileen’s Emporium. This is a prototypical size and fits snugly into the web of the rail; I learnt that for curved rail these are quite stiff, and use 0.5 mm thick pieces on the inside of the rails, which also has the effect of ensuring that deep flanges do not strike the bolt heads as they are slightly more recessed.

The bolts are little square-headed marvels from miha in Germany, 1mm wide and 6mm long, with hexagonal nuts. I drill the fishplates with a 1 mm drill and the rails with a 1.5 mm drill.

Fitted... the first ones took nearly thirty minutes, as I grew used to the tiny nuts and bolts. I attempted unsuccessfully to use the pin chuck as a spanner.

The fishplate here is as-is; I learnt to stroke the end edges against a needle file to remove burrs.

The bolts are rather too long and will be shortened later with side-cutters—the length is useful when assembling the joint. The nuts are not tightened much (as with the prototype: they only align the rails), so I add a dab of superglue to each to hold them in place. I like the look of the fishplates inside the rails in this along-the-track view; once you start looking, this is often a feature of narrow-gauge track, in which the fishplates are relatively massive.

The second length of rail attached to the initial stubs. The fiddle yard-side sleepers are unstained. This shows a Sierra Valley wheel set on the track; its flanges easily clear the medium spikes I use with this rail size (I use small spikes with the lighter rail). The proportions of the sleepers and rail are absolutely prototypical and are very pleasing.

Close-up of the rail bender in action. Although rated up to Code 148 it performs perfectly well on Code 205 rail. It cannot bend the last ½" or so of each rail, so I mark this and measure from the mark, later cutting off the tail if the rail is to curve to its end.

A completed rail joint showing the ½" expansion gap. Note too the slightly closer sleeper spacing at the joint, as well as the spiking pattern.