Census of GWR/WR standard gauge 6-coupled saddle and pannier tanks, 1864-1966

[Miss Prism]

Build dates, pannier conversion dates and withdrawal dates are from RCTS Part Five and BRdatabase. In keeping with the 'overview' nature of this census, some minor complications, e.g. a few locos reverting to saddle tanks and then back again to panniers, and the locos originating with side tanks for a short while (a few Buffalos and the 1813 class), have been ignored. A total of approx 2380 standard gauge saddle and pannier tanks were built by the GWR/WR. The total population (saddles plus panniers) peaked in June/July 1955 at 1285 locos. Absorbed locos are not included in this survey.

 

(The current notation of the graph axes is 'as given' by Excel, and I will be exploring ways of improving it.)
 

 

 

Saddle to pannier conversion dates indicate the benefit of panniers was considered important enough for the programme to continue throughout WW1, and many of the conversions in the 1920s and 1930s were done in the knowledge that the locos would last only a few years longer, or at least until the Collett pannier designs began to come on stream in significant numbers from 1929 onward. Of the approx 1100 saddle tanks built, only 73 never received pannier tanks. (Off the scale of the graphic are the early 1813 class panniers of 1903-6 and 2021 class loco 2048, which was converted as late as 1948.)
 

 

 

 

 

Edited September 16, 2023 by Miss Prism
images re-instated (again)

[The Johnster]

56 minutes ago, Miss Prism said:

the benefit of panniers was considered important enough

As a water storage system, pannier tanks have no particular advantage over saddles, and their use by the GW was in connection with the introduction of belpaire fireboxes on replacement boilers, which is what was considered of benefit; the belpaires could not fit beneath saddle tanks, and thus the saddles needed to be replaced.  As locos which had previously been built as saddle tanks still required access to greasing points between the frames which would have been blocked by side tanks, panniers became the solution, and iconic of the GW in the 20th century. 

 

 

 

Edited January 1, 2021 by The Johnster

[melmerby]

Just to think all those matchboxes derived from No. 1490 built in 1898, complete with a Belpaire firebox, but with a 4-4-0 wheel configuration.

Edited January 1, 2021 by melmerby

[melmerby]

The only other locos in the UK with Panniers 'style' tanks, that I can recall were some LNWR ones.

There were 2 Ramsbottom "Special" STs which had squared saddles originally to do with condensing apparatus (later removed) and the batch of Webb Coal Engines that were rebuilt with a similar style square saddle. Neither were true panniers, as they still straddled the boiler, rather than hung off it.

 

Were there any true panniers that were not GWR? Industrial maybe?

[Miss Prism]

1 hour ago, The Johnster said:

As a water storage system, pannier tanks have no particular advantage over saddles

 

I beg to differ. I think the centre of gravity for a full saddle is slightly higher than that of a full pannier, so the pannier has more lateral stability. RCTS notes saddles "... were also unsuitable for running at speed", and a Buffalo saddle derailment in 1904 was considered partly due to the surging of water in the saddle tank.

 

That aside, wasn't the basic argument in favour of panniers being that the various bits and pieces (the tanks and the boiler) could be disassembled and re-assembled quicker than a saddle tank? Reducing turnround time in the works was a key strategy.
 

Quote

the belpaires could not fit beneath saddle tanks

 

Generally yes, but the wide diversity and progression of designs led to some exceptions - the 2101 class were built with Belpaire boxes and saddle tanks. (Conversely, some 850s had panniers and roundtop boilers!)
 

[The Johnster]

There were industrial pannier tanks provided by RSH to the NCB in the early 50s (for the Philadelphia system? somewhere in the NE of England anyway), the last ever built of the ‘Stephenson Long Boiler’ type; I submit that these fulfil the requirements of ‘true’ panniers.  The GW rebuilt some absorbed/constituent saddles as panniers, including some Cardiff Railway 0-6-0s and the Rhymney Railway K class, an outside framed 0-6-2.  
 

I’m not sure that the change to panniers from saddles made a lot of difference to centre of gravity or stability at speed.  The  Buffalo that derailed in 1904, on Loughor bridge in South Wales, was a leading pilot and the incident led to the rule that such assisting locos be coupled inside the train loco on the GW.  This was never applied to assisting locos at Severn Tunnel, but 0-6-0s were not used here; the norm was 31xx (later 5101) and 3150 large prairies.  Large prairies were used as pilots on the South Devon banks as well, in accordance with the rule.  Not sure why a loco that had to be coupled inside the train loco between Newton Abbott and Dainton was considered more stable between Severn Tunnel and Badminton. 
 

I believe the Lougher derailment was put down to the Buffalo not having a leading bogie or pony rather than water surging in the saddle tank.  Water surging in tanks must have been a perennial problem and presumably still is on steam railways but has not been cited as a cause of a derailment since Sevenoaks in 1927, and the condition of the track was a factor in that accident.  

So, how was surging water coped with?  Was it damped by baffles?  
 

In any case, the problem must have been similar in saddle and pannier tanks, and I would have thought if anything worse in panniers, which are rectangular.  In general, whatever problems surging caused, they were within safe limits and coped with adequately, or we’d have been waist deep in derailed locos!

 

[ikcdab]

Great spreadsheet, by the way. Having seen your excel query a couple of days ago, I now see where you were going.

Brilliant.

[Penrhos1920]

Very interesting, excellent.  I’m really surprised to see that it was 1921 when there were more panniers than saddles. If you’d asked me I would have said it was a few years later.  So you’ve justified me having an equal number of both types.  

 

Do you have the data arranged so that you can do separate graphs for the small and large classes?

[melmerby]

2 hours ago, The Johnster said:

There were industrial pannier tanks provided by RSH to the NCB in the early 50s (for the Philadelphia system? somewhere in the NE of England anyway), the last ever built of the ‘Stephenson Long Boiler’ type; I submit that these fulfil the requirements of ‘true’ panniers.  The GW rebuilt some absorbed/constituent saddles as panniers, including some Cardiff Railway 0-6-0s and the Rhymney Railway K class, an outside framed 0-6-2.  
 

 

 

Hordern Colliery had an Andrew Barclay saddle tank "Hordern" with square tanks (currently being restored at Tanfield.)

Not a true pannier and more like the LNWR locos.

 

I can remember seeing a photo of a loco (not GWR, probably industrial) that had what I would call pannier tanks, the join between the tank and the boiler had a radius to it, unlike the Western panniers.

 

EDIT this one:

 

Probably the long boiler you were thinking of MR Johnster?

 

Also found a Bagnall 0-6-0PT "Fife" from 1940 which was at Rosyth dockyard, it had typical dock shunter short wheelbase and O/S Walschaerts gear.

It was scrapped in 1968 and the only photo I can find is on blogspot which can't be linked to.

 

Edited January 1, 2021 by melmerby

[Nick Holliday]

I suspect this one doesn't really count, as it seems to have tank supports off the running plate front and aft, but it is another early design.

One of a handful of LBSCR Craven shunters with pannier style tanks dating from 1866. Three were sold to the Alexandra (Newport & South Wales) Docks & Railway Company, one at least surviving until 1906. Not quite GWR, but close!

[JimC]

17 hours ago, melmerby said:

. Neither were true panniers, as they still straddled the boiler, rather than hung off it.

I think this is something of a non-existent distinction. If you look at the GWR pannier tanks the straps across the boiler are not large at all, but the support brackets under the tanks are very substantial. I suspect few engineers would pick hanging tanks in tension over sitting them on solid structures in compression. 

[Mikkel]

18 hours ago, Miss Prism said:

 

 

 

Thanks Miss P, quite interesting. And right at the center of the "X" is the immediate pre-grouping period, again highlighting how these often overlooked years have interesting potential for modellers looking for diversity.

[Andy Kirkham]

17 hours ago, melmerby said:

 

Were there any true panniers that were not GWR? Industrial maybe?

There were some narrow gauge ones:

 

[JimC]

Excellent piece of work. The numbers may seem dry, but they tell us a lot we simply cannot find out any other way. 

 

[melmerby]

59 minutes ago, JimC said:

I think this is something of a non-existent distinction. If you look at the GWR pannier tanks the straps across the boiler are not large at all, but the support brackets under the tanks are very substantial. I suspect few engineers would pick hanging tanks in tension over sitting them on solid structures in compression. 

The straps, I assume, are there to help keep the tanks in position, rather than hang them off the boiler.

Saddle tanks are however sitting on and therefore I assume supported by the boiler to a great extent.

[The Stationmaster]

18 hours ago, The Johnster said:

 

I’m not sure that the change to panniers from saddles made a lot of difference to centre of gravity or stability at speed.  The  Buffalo that derailed in 1904, on Loughor bridge in South Wales, was a leading pilot and the incident led to the rule that such assisting locos be coupled inside the train loco on the GW.  This was never applied to assisting locos at Severn Tunnel, but 0-6-0s were not used here; the norm was 31xx (later 5101) and 3150 large prairies.  Large prairies were used as pilots on the South Devon banks as well, in accordance with the rule.  Not sure why a loco that had to be coupled inside the train loco between Newton Abbott and Dainton was considered more stable between Severn Tunnel and Badminton. 

 

I'm sorry but having posted the detail at various times and in various threads in the past the above is not entirely accurate, in fact in many respects it is the opposite of accurate.  The alleged 'rule of thumb' that the assisting loco had to go inside the train engine on the GWR was never a rule of thumb and has regrettably been misquoted ad infinitum over the years to the extent that lots of people have come to treat it as gospel when it couldn't be further from the truth!.

And not getting at you ' Johnster' because your comment is typical of far too many.  It is regrettably also the case that those who observed or even worked with  steam on the WR should also know  better.  There was never a 'rule of thumb' although some seem to quote it  as some sort of tablet of stone handed down from Swindon and it definitely was not.

 

For example for assisting from the bottom to the top of an incline any engine with a driving wheel diameter greater that 4ft 6" could be used. (e.g. a saddle or pannier tank of a number of classes could therefore be used) and unless assistance in rear was specifically permitted over that section the assistant engine had to go on the front of the train engine.  

The situation on level track and falling gradients was different but basically any 4-6-0 or 4-4-0 was authorised to assist front of the train engine.  Only engines which were not a 4-4-0 or a 4-6-0 had to go inside the train engine.  But even that was not consistent because 2-6-0s (see Note 1) and large wheeled 2-6-2Ts were permitted to assist front of the train engine in various places (including the Severn Tunnel and in South Devon).

The Instructions regarding engines with leading pony trucks assisting leading bogie engines were further revised in 1948

Instructions were different in respect of assisting 'Kings'

 

Note 1.  When I typed 2-6-0s I forgot to mention the wheel diamter hence it actually means 43XX and variants thereof,

 

Sorry to go OT Miss P and I'm not taking a particular swipe at Johnster on this occasion but he alas wheeled out some oft spouted incorrect comments about assisting engines on the GWR/WR. 

Edited January 2, 2021 by The Stationmaster
Afdd Note i.

[Miss Prism]

16 hours ago, Penrhos1920 said:

Do you have the data arranged so that you can do separate graphs for the small and large classes?

 

Not as such, although it could be done by duplicating the file and splitting, but I wonder whether that would show anything significantly different. Besides which, my personal perception is the classes falling into 'small', 'medium' and 'large':

 

- 'small': 850/1901, 2021/2101, crane tanks, 1361, 1366, 1600
- 'medium': 302, 119, 322*, 1016, 1076*, 1661
- 'large': 645, 655, 1813, 1854, 2721, 5700, 5400, 6400, 7400, 9400, 1500

 

* some locos transitioned later to 'large'
 

[The Johnster]

16 hours ago, melmerby said:

Hordern Colliery had an Andrew Barclay saddle tank "Hordern" with square tanks (currently being restored at Tanfield.)

Not a true pannier and more like the LNWR locos.

 

I can remember seeing a photo of a loco (not GWR, probably industrial) that had what I would call pannier tanks, the join between the tank and the boiler had a radius to it, unlike the Western panniers.

 

EDIT this one:

 

Probably the long boiler you were thinking of MR Johnster?

 

Also found a Bagnall 0-6-0PT "Fife" from 1940 which was at Rosyth dockyard, it had typical dock shunter short wheelbase and O/S Walschaerts gear.

It was scrapped in 1968 and the only photo I can find is on blogspot which can't be linked to.

 

If that’s not the loco I was thinking of, then it is very similar.  Glad to see it’s still with us!

[JimC]

59 minutes ago, Miss Prism said:

. Besides which, my personal perception is the classes falling into 'small', 'medium' and 'large':

Any number of ways to skin the cat, my usual division is small wheel v large wheel, but with 54s, 64s and 74s honorary members of the small wheel club. It is tricky to draw firm lines, especially on the basis of power, because Collett's 'small' classes, esp the 74, whilst clearly of the 850 line of development were actually more powerful than earlier members of the 'large' 57 lineage. 

[The Johnster]

You have to establish boundaries somewhere or (perhaps the best idea) abandon distinction by size.  My interpretation, which I do not claim to be exclusively correct or even correct at all, it’s just a system that works for me, is that the 4’1” driving wheel lineage is ‘small’, the ‘2021’ lineage including 54/64/74xx is ‘medium’ and anything where the boiler protrudes significantly above the pannier tanks is ‘large’, clearly inadequate delineation for saddle tanks but those are out of period for me anyway.  94xx and 15xx seem to me to be in a different category altogether; though locomen seem to have regarded 94xx as equal to 57xx/8750, they were red RA and D power class.  BR rated them 4F and the 57xx/8750 as 3F.  I think of these classes as ‘heavy’ panniers.  

[melmerby]

25 minutes ago, The Johnster said:

 94xx and 15xx seem to me to be in a different category altogether; though locomen seem to have regarded 94xx as equal to 57xx/8750, they were red RA and D power class.  BR rated them 4F and the 57xx/8750 as 3F.  I think of these classes as ‘heavy’ panniers.  

The 94XX has the same "machinery" as a 57XX to produce the same TE, presumably BR thought that having a larger taper boiler would allow harder use from the machine.

With ther extra mass the loco braking effort should have also be better.

 

The 15XX has the some commonality with the 94XX & 57XX but has piston valves instead of slide, but the O/S Walschaerts, short wheelbase and frame style are the stand out differences.

 

 

Edited January 2, 2021 by melmerby

[JimC]

A study of numbers indicates that the 94s were built to replace the last of the pre group 0-6-2T and 0-6-0T, which were withdrawn over the period the 94s were built. So a cheaper alternative to more 56xx.

 

And yes, more boiler capacity and brake power than the 57s.

Edited January 2, 2021 by JimC

[Steamport Southport]

The Doxford Crane Tanks were also Panniers.

 

 

 

Jason

[JimC]

It's one of the amusing quirks of steam enthusiasts in the UK that folk wish to deny certain locomotives have pannier tanks. I suppose because the pannier tank configuration is so strongly associated with the GWR. So I observe enthusiasts and or owners of industrials and others that clearly have pannier tanks ardently denying it. Even more amusing when you consider locomotives with no footplate and outside motion, where the distinction is at best blurred, and verging on non existent on some locomotives. 

[The Johnster]

TTBOMK, the term 'pannier tank' was coined by the GW and is thus associated with that railway.  GW locos, whether saddles converted to the form or built to it, were pannier tanks, and, arguably (not that I particularly want to argue it) other locos that had tanks that were clear of the footplate but that did not stretch across the top of the boilers were usually referred to by their owners and crews as tank locos.  Not, you will note, as side tank locos, that inferring a different  thing altogether.  Then you have the adaptation of the Whyte notation as suffix to the wheel arrangement; T, ST, PT, WT, BT, CT and any others I can't think of now just.

 

On horses, where I believe the terms are rooted (and I am not claiming to be any sort of equestrian authority here), saddle bags sit over the horse's back while panniers hang from the sides connected by straps across it's back, the term being transferred to baskets and bags carried by humans, particularly milkmaids carrying milk in pannier buckets.  To describe the tanks of a steam locomotive as panniers is not, therefore, strictly correct as the weight of them, which can be considerable when they are full, is not bourne by the top of the boiler.  There are straps across the boiler top connecting the tanks and no doubt steadying them against the flexing of the loco as it moves, but the main load is carried by brackets attached to the footplate where the loading is carried through to the frames.  A saddle tank usually has such brackets as well but more of the load is carried by the boiler top.  On side tanks, none of the load is carried by the boiler whatsoever.

 

It's all a bit academic though;  by and large we know what we mean by the terms.  If somebody described a narrow gauge Baldwin 4-6-0 or ALCO prairie as a pannier tank to me I'd understand what was meant and would not object to the usage, but similarly if somebody wants to describe any non-GW loco with that form of tankage as a tank loco I would have no objection to that either.  I would probable accept a description of any GW pannier as a tank loco.

 

Then there are the 97xx, usually described as panniers but they weren't really.