Miss Prism

725 at Wolverhampton Low Level, 11 June 1910. I wouldn't fancy trying to apply lining behind those springs.

A couple of 806s (two pics of 824, the later one having an extended smokebox). I'm not too sure what the difference between the 806s and 717s are (must dig out my Jim Russell book!), and earlier 806s seem to have radically different outside frames (or were these rebuilt 439s? or was there another renumbering stage?)  (see Mikkel's pic of 718 above)

The first (Teignmouth) view is c 1902.

A couple of 717s, apparently a bit larger than the 481s.

The tender behind 717 has a very low body - compare with the one behind 720. Rear 3/4 views are very rare, and I wonder whether the 720 picture is by Peter Darke, whose book has other rear 3/4 views. He took a lot of pictures at Taplow, but I can't correlate that to the picture here.

106, apparently ex-Birkenhead Railway. Date unknown. I don't know the original loco builder - the body as updated here is mostly Wolverhampton, but the frame front end is a dead ringer for a Metro tank. The tender is strange, springs mounted low, Swindon-looking toolboxes, but it doesn't have a footplate. Looks more LNWR than GWR to me. (The Birkenhead Railway had feet in both camps.) 

I'm always struck by how much variation there was in Armstrong tenders. (Lengths, wheelbases, frame styles, tanks, body heights.)

For something ultra-modern, here's 489 at Worcester Foregate Street in 1905:

591 looks the same as 482 !!!  (Edit: Although its tender is a longer one.)

This incarnation of 482 is I suspect far too modern for what Jim is looking for:

Interesting thread on in-phase and out-of-phase driving wheel cranks.

Reamer = fancy broach (and considerably more expensive!)

Neil is using the right tool for that job - cutting broaches can be surprisingly precise.

I cannot add any fact, but merely observe that, locomotive-wise, it seems to me the notion of a 'GWR' in a corporate sense didn't begin to exist until at least 1865, and I think this is borne out by MikeOxon's perspective. Before that, locos were ordered from a succession of subcontractors, and we don't know what level of detail was specified by Gooch. It wasn't until Armstrong took up post at Swindon that a company standardised approach (and style) began to slowly emerge. In the meantime, Armstrong inherited a ragbag of broad gauge oddities to maintain and improve, not to mention being responsible for all rolling stock, both passenger and goods. So I think your surmise that injectors arrived from an independent (Beyer) is very plausible.

As Mike says, heady times indeed.
 

1 minute ago, Compound2632 said:

There's no coupling rods flailing around. Is there a 2-4-0 class with the same cylinders and boiler that one compare weight diagrams with?

Good question, and one @JimC is probably good at answering. The big singles seem to have a driver weight of c 16.5 tons, whereas on the biggest of the 2-4-0s, this reduced to c 15 1/4 tons and in many cases down to c 14 tons for the smaller-boilered locos. Presumably crankpin throws made a difference to hammer blow - on the singles this was 26" I think, and came down to 24" on the 2-4-0s.

I think what probably did for the 2-2-2s was their high middle axle weight (and the hammer blow).

The one on the right is a real sweetie.

Here's the strangely-shaped safety valve casing (on early 2-4-0ST 346):

I can't help wondering if there is a dome in the lower half. Wolverhampton had such fixtures on its R6 boilers.

32 minutes ago, wappinghigh said:

Well sure... R3127 and R3226 are the only ones  pictured with a raised (curved) footplate but a search of Hattons and ebay ..to take a look at the actual model draws a blank.

Are you sure about that?

https://www.rmweb.co.uk/topic/49734-gwr-8-coupled-tank-engines/

(there's a few other 42xx/52xx/72xx threads as well)

http://www.gwr.org.uk/no-heavy-tanks.html

32 minutes ago, wappinghigh said:

^ Because @Miss Prism all the actual 72XX Hornby models I can find (even though some are depicted on the box with the raised foot plate)  ..the actual (physical) model doesn't have a raised footplate! 

gwrrob's one has a raised footplate. R3127 is the Hornby number I believe.

10 minutes ago, wappinghigh said:

And the 72XX Hornby (for that matter).. which *is* a large prairie then.. and which doesn't even bother with the raised footplate? 

The Hornby 72xx is not a large prairie. A large prairie is 2-6-2T. A 72xx is 2-8-2T. Here is a Hornby raised footplate 72xx:

@wappinghigh - this is not a Large Prairie (which is a 2-6-2T).

192 at Newport, 1930. The smokebox and boiler originate at Swindon, but the remainder is substantially Andrew Barclay. The injector looks a bit Swindon. It was ex-31 of the Alexandra (Newport and South Wales) Docks and Railway, and lasted until 1946, so must have been regarded as worth keeping.

I agree the John Daniel spreadsheet needs re-writing before it is user-friendly! (I don't like the entries under the diagrams column - very confusing.)  Also, although I haven't studied the matter, there do seem to be different locos carrying the same number. (Not a fault of that spreadsheet of course.)

Your last comment strongly hints at the march of AI. What I am increasingly struck by is how stupid Google is. Despite having the gwr.org page to pour over for the last 13 years, it's only recently it has been able to define 'pannier tank' (a passable but slightly amusing effort), and it thinks they only came into existence in 1929. Hmmm.

I'm not so keen on the year-by-year analysis, because some years are chock-a-block full of revolution, and many other years are very quiet.

I think any choice of theme will inevitably be compromised because there's many ways of cutting the cake. The dominant theme of Nock's books for example is  'performance', which is fascinating in its own way, but as a result tends to concentrate on the larger express locos - the smaller lesser types are not considered to be particularly noteworthy.

The treatment slicing can also be non-era/non-class, e.g. 'brakes', 'boilers', 'superheating', 'maintenance'. For very nerdy niches, how about wheels and chimneys!

Modellers and prototype enthusiasts have different perspectives, and the 'era' slicing probably represents the best compromise and the most popular demand.

1 hour ago, Mikkel said:

In extension of that, I wonder if it's all that interesting to go through detail of how many locos were built in a particular class each year? That would perhaps be better presented in a table format? 

That's do-able, on a DIY basis, from John Daniel's spreadsheet:

http://www.gwr.org.uk/notes/Loco_lot_nos.xls

Also, BRDatabase is very useful for drilling down on a date basis.