While much progress has been made on the layout, there is little to say about it. Baseboard construction, track laying, and wiring, are all using well-established techniques which have been executed and written up far better elsewhere. As a result, I have decided to write a little about rolling stock, and specifically the question of how to choose what to buy and what to build yourself. Everybody's views on this will of course be different, but I will try to set out my approach.

The first thing to do is to set out the options. Firstly, I could choose to make something using traditional materials - cardboard, styrene sheet, etc. I often use this for buildings, where you generally want a bespoke item to fit the theme and the space available, but will only want one or two. I started out using the Wordsworth range of kits and cereal boxes - thicker card was obtained by using more layers - and have since moved on to designing some of my own buildings. Model Brick Yard is an invaluable resource for this. The drawback is that I find it difficult to be precise when drawing and cutting (a combination of impatience and poor coordination) so its not as good for small items or fine details.

However, this is where 3D printing has a role to play. While I only have a small and fairly basic FDM printer (the Creality Ender-2 Pro) I can design and print things which it would be a huge struggle to produce using traditional techniques. The drawbacks are well documented - the risk of warping on large pieces is a particularly important one, and I am also held back by the fact that my laptop isn't really powerful enough for very large, complex CAD files, such as those needed for brick buildings.

The final two options are kits and ready-made items, where the main downsides are price and the need to take whatever is already on offer.

The first thing to consider is locomotives. Here the choice is fairly simple. I could not make a model which runs as well or looks as good as a ready to run loco (and even some who are very good kitbuilders say the same), so unless I want to model a niche or fictional prototype, it just doesn't make sense. Everything I actually need is available ready to run (although some locos will need minor modifications) so purely ready-to-run locomotives are the way to go.

Coaches are more challenging. The Airfix autocoach is suitable for the era and readily available. A seemingly common train formation on the branch after nationalisation is a B-set with an extra coach to strengthen it. Whether this combination was used in the 1930s is difficult to establish due to a lack of photographs from the period. However, it's an interesting train formation and makes use of the Airfix coaches. More difficult is selecting stock for the longer Paddington services. The 1920s Collett corridor stock provided by Hornby is a good starting point, but 1940s photographs show a real mixture of stock, including occaisional clerestories and Centenaries. Concertinas show up more than once and toplights are very common. Of these, only the Centenaries (Airfix) are available Ready to Run (excluding the various flawed clerestories now made by Hornby), but these were purpose built for the Cornish Riviera and unlikely to have been used elsewhere before the war. Concertinas, even if they were available, are 70ft coaches and so take up much more space than the other options. A Sunshine coach, courtesy of Mainline, is perhaps the best RTR option to mix up the fleet.

Toplights seem like the ideal candidate, with 57ft examples available and being quite numerous. However, the only RTR model likely to appear soon is of the unusual Mainline & City set. There are metal kits available, but this is a medium I haven't worked with before and they cost more than I would be willing to risk given my lack of experience building and painting stock. Fortunately, basic drawings are available from The Great Western Archive and this opens up the 3D print/scratchbuild route.