Southern Performance during The 1948 Locomotive Exchanges
By Jeremy Clarke

By the same token, when the Labour Government of 1945 came to consider the compensation due to shareholders of the Big Four the attitude towards them was comparable, meaning the agreements made would not be kept. In the ungracious words of the Chancellor of the Exchequer, Dr Hugh Dalton, during the Parliamentary debate concerning their takeover by the State, the Railways were described as ‘a very poor bag of assets’. Of course they were! They and their staffs had been worked into the ground to keep the supply of men and munitions flowing, often under dangerous and disorganised conditions due to the enormous volumes of passengers and freight transported, the ‘blackout’, and the many inconvenient interruptions by the Luftwaffe. They were, not to put too fine a point on it, exhausted and by the Government’s attitude disgracefully unappreciated. That was not new of course. As John Bright - a great Parliamentarian, a brilliant orator and the President of the Board of Trade at the end of the 1860s - wrote, “Railways have rendered more services and have received less gratitude than any other institution in the land”. Hear, hear! And unlike much of Continental infrastructure the British Government decided to invest very little of the millions of dollars received under the post-WW2 Marshall Plan in railway reconstruction.

So, to the delight of many and the concern of more than a few, British Railways came into being on 1st January 1948 though the last ‘takeovers’ were not completed until September that year. Not much changed at the beginning. Other than company logos being suppressed in favour of BRITISH RAILWAYS in full on tank or tender sides, the slow addition of ‘Regional’ prefixes to engine numbers and a rash of experimental liveries, the trains carried on much as before. But with a unified Motive Power authority in being plans were made to test various ex-‘Big Four’ engines over routes other than their own, ostensibly with a view to developing a ‘standard’ BR range from the results. Opinions on the reasons and the needs for the tests vary, and questions were asked about how expenditure in adding yet more locomotive designs to the many already in being could be justified in the light of existing economic conditions. That questioning was also asked as post-war orders for many engines, most of recent or updated ‘Big Four’ design, continued to be filled up to 1948 and indeed for some time beyond.

The most cynical answer perhaps is that the newly-appointed Chief Mechanical Engineer, R A Riddles, who had supervised design and construction of the ‘WD’ freight classes, was keen to leave a more obvious mark. In many respects however, continuing with steam motive power was sensible. The steam engine was relatively cheap to build and the country still had rich supplies of good steam coal though the best went for export to improve the Exchequer’s financial condition. Moreover, the brief flirtation with oil-firing in 1947 had shown how the Treasury soon ran out of the necessary dollars to purchase locomotive fuel in quantity, especially as demand and thus price rose. (Sterling, at $4.86/£1 in 1939, had been devalued by 17% to $4.03 under the Bretton Woods Agreement of 1940. By 1949 this had proved unsustainable and a further devaluation of no less than 30.5%, a condition of the loan from the USA, saw the rate pegged at $2.80. Further devaluation by another 14% in November 1967 took sterling to $2.40.)

At Nationalisation there were but forty-nine diesel shunters of various designs on Britain’s railways, and one mainline diesel-electric locomotive, the LMS no 10000. Major investment in this form of motive power was also out of the question for the time being, except for small shunting units, for the same reason as the abortive oil-firing of steam engines, though shipping was necessarily rather less affected by this. The investment made in 1947 in modifying locomotives to burn oil and purchasing and erecting the necessary equipment had had to be written off by the railway companies to their disadvantage. Only when the products emanating from the 1955 ‘Modernisation Plan’ started to come on stream in the later 1950s did the major shift to diesel power begin and in hindsight it is plain to see how poorly this was managed. But these reasons of themselves surely could not justify a whole range of new steam designs being proposed. On the other hand deferment of the most obvious course, widespread investment in electrification, the ideal motive power, was hardly surprising in a country bankrupted by war.

So steam it was to be, but complete unification of motive power was seen by the new British Transport Commission as the ideal. Thus, between April and September 1948 a series of tests was run with locomotives over several different routes in three categories, Express Passenger, General Purpose and Heavy Freight, with the view to cherry-picking the most desirable aspects of existing designs to form new builds. Competitive testing was not new of course, the earliest recorded being in 1870 between a LSWR Beattie ‘Centaur’ class 2-4-0 and a SER Cudworth ‘Single’, principally to determine the relative efficiency of burning coal in their respective ‘patent’ fireboxes. The results are elusive!

The most notable concerned the 1925 exchange between a Collett ‘Castle’ and a Gresley ‘A1’ following the British Empire Exhibition of 1924. The pioneer Caerphilly Castle, noticeably the smaller as it stood side-by-side with Flying Scotsman, was advertised by the GWR as the ‘most powerful engine’ in Britain. On the basis of that most deceptive of yardsticks, tractive effort, it was. Gresley challenged Collett to prove it*. He did! As a result alterations to lengthen the valve travel of the ‘A1’ reduced coal consumption from an average of 50lbs per mile to 38lbs, and taken together with the introduction of a higher pressure boiler and modified cylinders ushered in the immortal ‘A3’.

But I ask again, was all the ballyhoo of 1948 really necessary? I think not, for in 1937 the LMS and LNER had decided jointly to fund and build a testing plant at Rugby. Neither the GWR nor the Southern expressed an interest, the one already having its own testing facility at Swindon, the other concentrating on an ongoing electrification programme. But the Rugby proposal followed an earlier one made in 1927 by Gresley for a similar installation at Leeds: it was hoped some Government funding might be forthcoming. However, the financial situation that developed at the end of that decade saw Government support refused and the scheme dropped in 1930. But buoyed by the improving financial situation work at Rugby resumed late in 1938, the 1933-built testing station at Vitry in France, and the Pennsylvania Railroad’s Altoona plant both providing some design ideas. The outbreak of war halted progress, commissioning finally coming in October 1948, a month after the Locomotive Exchanges had been completed. It must have been clear even as the Big Four were drawing their last corporate collective breaths this plant would be ‘on line’ within the year. Scientifically-controlled testing there and properly conducted trials ‘on the road’ would prove in time to yield far more meaningful and useful data than ever came out of the Exchange Trials.

A further question about their necessity arises. It was soon clear that with former LMS men dominating BR’s motive power offices - Riddles himself was Crewe-trained and naturally, as the boss, picked many of his team - that company’s practices were most likely to predominate irrespective of the outcome of the Trials. Was the whole thing then simply a charade, Riddles considering it a sop to the Mechanical Engineers of the other companies who retained their posts pro tem knowing, or at least suspecting, their contributions would be minimal? Perhaps I do him a dis-service. But that was very nearly the ultimate result when the ‘standard’ designs were decided upon and produced. E S Cox – another LMS man who took up a premium apprenticeship at Horwich in 1917 - summed this up well when he wrote, “It so happens that in certain cases (my italics!) LM designs have the highest route availability for their power, and at the same time incorporate the latest advances in design from an operating and maintenance point of view”. Hm! Who could forget the original Ivatt 4MT’s whose steaming abilities were put to shame by the ordinary kitchen kettle? But then the nail is driven in,”…. and the latest SR engines do not provide the requirements of simplicity and ease of maintenance which is the cornerstone of the present scheme”. Hm again! But then did nobody in authority ever look at a later ‘Arthur’, in Derek Cross’s opinion the Southern’s answer to the LMS ‘Black 5’ and a better one too, or a Maunsell ‘mogul’? Neither was that modern I agree, but surely had other expressed qualifications! Apparently not. (Having said all that it was a former Southern Railway man, C S Cocks, who went from Brighton to Derby to take charge of design under Riddles and Cox.)

The Southern’s contribution to the tests consisted of Bulleid ‘pacifics’ of both types to compete – though that was a word never used in this context – in the Express Passenger and General Purpose categories. Having no designated locos in the heavy freight category the SR took no part there. (An ER 2-8-0 was later tried between Salisbury and Eastleigh, successfully perhaps, but ultimately unnecessarily.) The specified requirement was for engines that had run between 20,000 and 30,000 miles since the last general overhaul, which ensured they were well run in but still fresh enough to minimise the possibility of failure away from ‘home ground’. Crew selection also proved critical, at least so far as the Southern was concerned, and for some time beforehand senior enginemen found themselves coming under unusually frequent attention from Locomotive Inspectors. Two firemen and four drivers were ultimately chosen. Two of the four, Jack Gilmore and George Robjant, were to act as conductors for the ‘foreign’ crews, the latter also being ‘standby’ in the event of a selected ‘away’ driver being unable to take duty. The ‘away’ crews were Driver George James and Fireman George Reynolds, and Driver Jack Swain with Fireman A E ‘Bert’ Hooker. It may be noted all these men were based at Nine Elms. They were probably as well fitted for the job as any since neither the Eastern nor Central Divisions of the SR could readily muster journeys as long as those regularly undertaken by crews on the Western Division. Even a Kent ‘rounder’ did not form one continuous journey.

At this time the Southern double-manned many of its main line engines. Three of the men selected, James, Reynolds and Hooker, were all from 21C20, Bibby Line, though that engine did not feature in the tests. Those selected for the Express Passenger trials were 35017, Belgian Marine, and 35019, French Line CGT. These were from the second batch of ten built at Eastleigh to order HO1189, appearing in April and June 1945 respectively. James and Reynolds worked Paddington-Plymouth with 35019 and manned 35017 between Kings Cross and Leeds: Swain and Hooker took that engine over the Euston-Carlisle route. For the General Purpose tests the Southern selected ‘West Country’ class nos 34004, Yeovil, 34005, Barnstable and 34006, Bude. These also were from 1945, all three coming into service from Brighton (order no 2421) in July. James and Reynolds took 34005 between St Pancras and Manchester and 34006 on Bristol-Plymouth runs. Swain and Hooker had the latter engine between Marylebone and Manchester and 34004 over the former Highland Railway Perth-Inverness route. That engine had been booked to work on the Great Central line but did not do so. Following a light-engine trip two days before going to Neasden, mainly to see if any minor matters needed attention, some superheater flues were found to be leaking badly. It appeared they had not been ‘rolled’ properly following replacement and renewal of the elements. Bude was sent instead. Being built for the relatively short distances run on the Southern, the tenders were exchanged for the LMS ‘Black 5’-type holding nine tons of coal but only 4000 gallons of water. Use of the pick-up apparatus had thus to mastered. But Bert Hooker recalls he was disappointed to find the coal space was not self-trimming as in Bulleid tenders and many of those then attached to Maunsell engines. Moreover the shovelling plate was at least six inches lower than that of the ’MN’. The upshot was that for the first fifty or sixty miles of a trip the coal continually fell off the plate through the large gap under the doors and had thus to be fired off the floor. Things then became more manageable though again coal had to be lifted, this time from the shovelling plate to the firehole door. Finally, in the latter stages, perhaps for the last 80 miles or more of the run to Carlisle for example, the fireman had to enter the tender to pull coal forward, a rather tiring exercise after several hours work. (For tests on the Waterloo-Exeter route the ex-LMS engines were provided with ‘WD’-type eight-wheel 5000-gallon tenders.)

Those test engines selected from other Regions in the Express Passenger category were: ‘A4’ and ‘Duchess’ ‘pacifics’, and ‘King’ and rebuilt ‘Royal Scot’ 4-6-0s. General Purpose representatives were ‘Hall’, ‘Black 5’ and ‘B1’ 4-6-0s. In some respects the selections were questionable, the West Country’ for example being power classified ‘7P/5F’ against ‘5MT’ of the others in its category. And why did Doncaster go for ‘A4’s at least ten years old rather than the latest Peppercorn ‘A2’s, particularly knowing the potential weakness of the Gresley derived valve motion, twice a failure during the tests? (The first of the splendid Peppercorn ‘A1’s was not outshopped until August that year.)

I am still rather disappointed that what one might call ‘second rank’ engines did not appear, a ‘V2’, a ‘Castle’, a ‘Jubilee’ perhaps, and a ‘Lord Nelson’ for example. It would have been interesting to see how a ‘Nelson’ with Bulleid’s modifications would have fared given its doubtful reputation. But with an experienced crew and Stephen Townroe’s preparation who could tell? As it turned out the ‘Scot’ in the Express Passenger category was arguably the star of the show. Were Maunsell’s shades rather amused at the thought Fowler’s draughtsmen had laid out the original with a full set of ‘Nelson’ drawings at their elbows? In 1943 Stanier had started putting on new taper boilers, the three ‘Scots’ taking part in the tests being thus equipped. The trials might also be assessed as incomplete, loading gauge limits prohibiting GWR engines from running over Southern metals and the ‘King’ outside its own sphere only permitted between Kings Cross and Leeds. The ‘Hall’ was similarly limited to the Great Central route.

Various experienced travelling timekeepers ‘logged’ some of the test runs and that doyen of journalist/railwaymen, Cecil J Allen, published a book containing a fairly comprehensive though an inevitably incomplete collection of them, including his own. Obviously expecting every run to be timed by people other than the staff manning the dynamometer cars is an impossible ask since more than one test could be run on the same day and the majority of the ‘amateurs’ were in full-time employment. Allen also pointed out that when the book appeared in 1949 no data in the freight engine category had been released though his follow-up publication contained a resume of all the results published by the Railway Executive.

But we are more fortunate in having the human side of the event from the estimable Bert Hooker. In that regard Allen could be fallible as, for example, when commenting on a LCGB Special out of Waterloo headed by ‘A4’ no 60022, Mallard, in January 1963. Approaching Byfleet driver Hooker states he noticed the water had got very high in the glass. Wary lest priming should occur he partly closed the regulator until the level had fallen back a little. Allen noted in his record of the run published in Trains Illustrated, “The engine was unaccountably eased at Byfleet” though he obviously made no effort to find out why as there certainly was ‘account’. AEH leaves us to draw our own conclusions. As if to emphasise how his opinions should not always be taken at face value Allen provides a rather devastating account of a trip on a Billinton ‘Baltic’ 4-6-4 tank engine in 1922. Three miles out from Brighton, “the driver fixed his cut off at 35% and regulator at ? open. Thence he never touched either, uphill or down dale, until a slight signal check at Purley compelled action……about the most unenterprising bit of locomotive handling that it has been my lot to witness”. But as one commentator later said, would he have written so adversely had he timed the trip from the train rather than from the footplate? The 350-ton ‘Southern Belle’ was being run nicely to time until that check, and the long but relatively easy rises and falls of the Brighton Line did not require constant attention to the controls or, indeed, to the fire. As it was, signals checked the train until East Croydon but the driver still made Victoria on time. The fact Allen makes no comment lauding that again leaves one to draw one’s own conclusions. This is not to decry his presentation of the facts as known, nor his general assessments of the trips, merely that experience tells us some care ought perhaps to be taken in accepting some of his comments and opinions at face value.

Having said that he questions why a particular driver, someone he states he had almost always found very dull and unenterprising - that word again! - in his work was chosen to run the ‘A1’ ‘pacific’ on home ground in 1925 in competition with no 4079, Pendennis Castle’, in the hands of the imperturbable and very experienced driver Albert Young of Old Oak Common. Coincidentally he voices the same opinion on the nameless SR man whose running on home soil apparently compared poorly with that of his colleagues in the away team. And by the way, in his primary book on the Exchanges Allen deliberately shows no detailed records of any ‘home ground’ runs, justifying this by saying that such work could be seen every day anyway and it was thus pointless losing limited opportunities to travel with the ‘away’ crews, though he does occasionally draw comparisons.

Every crew was given three days in the week prior to the tests to gain at least some experience and familiarity with each route by working over it on normal service trains. They were also accompanied by an inspector from their own Region though he usually rode in the train rather than on the footplate. Allen comments that foreign engines working over the Southern probably put up the most consistent performances because only two men acted at pilots. From Hooker’s account it appears he and Jack Swain were usually in the care of the crew that would have taken the turn though the fireman would sometimes cede his place on the footplate to a member of the dynamometer car team and ride in the train. Obviously, in those circumstances, without the uniform style of driving imposed on French footplate crews for example, a driver’s particular way of doing things would impinge directly on his piloting advice which could also be dependent on how well he took to instructing others.

One other point to make clear. Bert Hooker is adamant that he and his colleagues received no advice or instruction on how their engines should be driven and fired. And he is equally adamant that Southern crews away from home were keen to show what their engines could do. This contrasts with the attitude of some other ‘contestants’ who appeared to favour fuel economy over timekeeping or, more particularly, making up any lost time. To a degree the attitude of the Southern men is a reflection of the SR’s vigorously-pursued ‘on time’ philosophy, essential when it ran the world’s most intensive electric suburban service. As a further illustration of this E S Beavor, sometime shedmaster at Exmouth Junction, maintained his family home in Exeter following a promotion that took him back to London. He comments that in 1960 and for part of 1961 he would regularly travel back to Devon on a Friday evening by the 5.30pm from Paddington, ‘The Mayflower’. That train rarely arrived in Exeter on time, averaging about 30 minutes late, sometimes twice that, usually without apparent cause. He writes, “It was my impression that either there was a deplorable lack of effort on the part of the enginemen or that some of them were repeatedly manufacturing overtime”. Though an LNER-trained man he took wholeheartedly to the Southern’s rigorous follow-up of ‘lost-time’ tickets which is clearly why he deplored the lackadaisical running on his Friday evening journeys by what the WR authorities obviously considered a ‘premier service’ though apparently failing to check it was being run like one. But then, for years, the SR circulated each month a sort of ‘league table’ showing comparative running performances of the various depots. To a degree it was competitive though Beavor admits Exmouth Junction, by virtue of much of its allocation being out-shedded at rural locations with comparatively sparse traffic and, in the main, somewhat undemanding timetables, usually came well up the league. This is not to imply the crews of Western engines taking part in the trials did not always do their best though to a degree they were handicapped by firing hard Yorkshire coal in line with all the other participants, rather than their favoured soft Welsh. I merely point out that Southern men had the necessity for good timekeeping imbued in them from their earliest days on the footplate.

Before getting into the ‘nitty-gritty’ it may be worth reminding ourselves how far the railways have advanced in terms of speed in the intervening years. Throughout the recording of events there is a sense of real achievement in point-to-point averages of, say, 55mph, or hill-climbing sustained at 45mph or even maxima into the 70s, such things being considered worthy of sometimes effusive praise. It may also be worth noting that the highest ‘equivalent drawbar horsepower’ recorded in the tests was 2,010 by none other than the Southern ‘West Country’ Bude heading south out of Leicester following a call from the dynamometer car to ‘give us a pull Jack’. In fact this engine recorded two of the three highest outputs noted for all the engines. But think too how much effort into gaining such output compares with what is available nowadays at the flick of a switch or the movement of a control handle. One final point: loads were substantial by today’s standards though not uncommon up to that time. So then, from the Hors d’oeuvre to the Entree!

From the operating point of view the West Coast main line between Euston and Carlisle was the one most plagued by speed restrictions for permanent way and other maintenance work and the consequent signal checks. Perhaps for this reason in his reporting of journeys over this route Allen tends mainly to summarise the test runs rather than fully log them. So as workings here are among the least documented it seems best to discuss their running first.

Bert Hooker describes the apprehension he and his colleague Jack Swain and, no less, the Southern’s travelling inspector, the popular and much respected Danny Knight who accompanied them, felt at tackling this route. After all, the distance of 299 miles was well over twice that of the longest they’d ever make on the Southern. In general at this time trains on the London-Weymouth run – 142¾ miles – were re-manned at Bournemouth – 108 miles - if the engine were working ‘through’. Nine Elms men on Exeter or Plymouth trains were relieved at Salisbury, 83½ miles from London, and if the ‘MN’ on the ‘Atlantic Coast Express’ for example were working right down to Exeter the distance is still only 171.4 miles.

Having then had their three days ‘practice’, on May 13th Jack Swain took no 35017 out of Euston at the head of the 530 tons gross ‘Royal Scot’. (As part of the test no banking assistance was permitted up to Camden.) Barely into motion Belgian Marine was brought to a stand by signals before entering Primrose Hill tunnel and was stopped again at Kilburn. Having then observed a 30mph pwr at Willesden the train was already ten minutes late passing Watford, only seventeen miles out. Further signal checks and pwr slowings meant the lateness had grown to 17 minutes by the time Belgian Marine drew to a stand at Rugby. However, despite further slowings for track repairs at Tamworth and Lichfield Swain kept the 56-minute allowance to Stafford before meeting another pwr on the climb to Whitmore. But a steady acceleration to 59mph at the head of the climb and a maximum of 70½ at Betley Road brought the train into Crewe in just over 83½ minutes from Rugby, 6½ minutes less than schedule despite five out-of-course slowings.

North of Crewe operating failures in addition to some severe speed restrictions conspired to make any attempt at timekeeping futile and invalidate much of the test recording. None of the ‘foreign’ engines involved kept the schedule of 183 minutes for the 141 miles from Crewe to Carlisle. Jack Swain made the most spirited run, perhaps because of the delays suffered to Crewe. Having arrived there 11 minutes late and stood for 13 minutes, which included taking water, Belgian Marine suffered further delays of 5¾ minutes by more maintenance restrictions. But Swain gained 2¾ minutes on the twelve uphill miles between Carnforth and Oxenholme before being summarily checked there by signals and thus losing all impetus for the seven miles at 1 in 124/131/106 up to Grayrigg. Allen reports the engine was not pressed too hard over this section which meant it had plenty in hand for the steeper climb to Shap, taking only 9½ minutes from Tebay to summit. He also comments that although speed at Tebay following the slack at Dillicar had risen only to 57½mph it had fallen back no further than to 26½ at Scout Green, halfway up the 1 in 75, but then increased to 29mph when the engine was ‘opened out’. In the course of the swift run onwards to Carlisle 35017 picked up 11 minutes though five of those were marked ‘recovery time’.

On a familiarisation run the previous week the engine had had to stop for water at Lancaster. The crew had already discovered the scoop picked up almost nothing at the troughs at Bushey and Castlethorpe. Water had been taken during stops at Rugby and Crewe though again pick up at Moore and Brock troughs had been minimal. Bert Hooker reports because of the slow flow it took seventeen minutes to fill the tank and there appeared to be any number of tiddlers thrashing about in it. In that instance Swain made a grand climb to Grayrigg, covering the 26.2 miles in only 34½ minutes and continued on to Carlisle without taking any more water. (It transpired a shunter moving the engine at Nine Elms had mistaken the operating handle for that of the tender handbrake, the lowered scoop striking a brick pathway across the track and becoming thoroughly distorted out of shape.)

Due to the slack-ridden state of the route and other test runs occurring at the same time over less troubled lines Allen made only one southbound run from Carlisle, on May 14th, and that behind Belgian Marine. But, as he put it, “I would not have missed this trip for worlds”. The previous evening, in the Upperby hostel, fireman Hooker had asked a top-link Crewe driver how fast a climb could be made from the Penrith start to Shap summit. On being told “22 or 23 minutes, not less”, Bert had surmised no 35017 could achieve it in 20 minutes. The succinct reply was to the effect there wasn’t an engine capable of doing that, if expressed perhaps in rather more colourful language.

The 525-ton gross load of the ‘Up Perth’ was taken over the 17.8 uphill miles from Citadel to Penrith in a minute over the 31 scheduled, though signals had twice all but stopped Belgian Marine in the approaches to the Cumbrian town. The first three miles out of Penrith are relatively easy, no 35017 attaining 46mph on the ¾-mile level before the seven miles up through the Eden Valley at 1 in 125. Here, Allen states,”….we settled down to a steady 41mph, which continued unvaryingly for mile after mile…. Before reaching the end of the 1 in 125 the speed was beginning to rise: on the [following 1¼ miles at] 1 in 142 we attained 46mph and the Shap level [station] carried the rate up to 51mph. By the end of the 1 in 125…. the engine was exerting fully 1700hp at the drawbar continuously. A brief drop to 46mph on the final 1¼ miles up at 1 in 106/130 and we were passing the Summit cabin in one second over 20½ minutes from the dead start at Penrith – a gain of 6½ minutes on schedule!” (Some hyperbole deleted!)

The train was still six minutes ahead of time on passing Preston despite two temporary restrictions en route, which Allen reckoned cost three minutes in running. By Springs Branch (Wigan) Belgian Marine was running 11½ minutes early but there were no fewer than eight permanent way and signals checks to contend with before Crewe. These included a stand outside the station though arrival there was still 2½ minutes before time, no 35017 drawing up alongside a ‘leave special’ carrying RAF personnel, the probable cause of adverse signals received in the approaches. But what followed nullified any of the following findings in the dynamometer car, for the ‘special’ was permitted to leave Crewe immediately ahead of Belgian Marine. The first signal check came after a mile, the second three miles further on. Allen comments very acidly that he saw no reason why the ‘special’ couldn’t have been moved to the slow road before Stafford as they passed nothing on it in the intervening thirteen miles. But not until Leighton Buzzard, 118 miles to the south, was that move made. No 35017 thus arrived at Euston 40 minutes late which, Bert Hooker writes, was apparently in line with the ‘Up Perth’s usual record.

But he adds another slant here. Having merely noted that ‘the operating authorities…..really slipped up there’, he adds that it crossed his mind the engine and crew perhaps did too well on their familiarisation runs, reason enough maybe to spoil their timely progress on this test trip. Does that perhaps lend weight to my earlier suggestion that LMR authorities already knew/surmised little would actually come of the tests simply because ‘their team’ was already in charge of locomotive development, so anything that might dent their certainty had to be discouraged? Or am I being unduly cynical?

On that rather sour note then to Kings Cross. Running on the East Coast line was handicapped by an overall speed restriction of 60mph out to Hatfield and 70mph thereafter in view of maintenance arrears still being made up though there were fewer speed restrictions in force than faced crews out of Euston.

Allen had no opportunity to ride behind Belgian Marine on northbound runs but he summarises one timed by a correspondent on Tuesday 25th May. Permanent way slowings to Peterborough were observed at Wood Green, right at the foot of the eight miles at 1 in 200 to the summit at Potters Bar, and at Connington at the bottom of the run down towards Peterborough from Abbotts Ripton. With a gross load estimated at 535 tons no 35017 recovered to a steady 37½mph after the Wood Green check to pass Hatfield only ½-minute late and Hitchin (31m 74ch) on time. Speed did not exceed 68mph before Peterborough, reached in 90¼ minutes or about 85 net. The really interesting part followed, the climb to Stoke summit. Here comparisons are made with the work of the ‘Royal Scot’, Queens Westminster Rifleman. With a similar load and in the hands of the LMR’s most enterprising driver, Brooker, the Rifleman started rather the faster, sustaining 61-62½ mph up the slightly undulating rise to Essendine followed by a minimum of 48mph at Little Bytham, before falling only to 47mph on the final three miles at 1 in 178 to Stoke from an intermediate recovery to 53mph at Corby Glen. By comparison, and despite the slower start, George James got to 65mph before Essendine, though passing through the station itself at 61mph and then accelerating to 63mph before falling back to 47mph at Little Bytham. As with the ‘Scot’, there was a recovery at Corby Glen, to 55mph, but then speed fell only to 50mph at the summit. Allen summarises the times as 6m 13s (4-6-0) and 6m 40s (4-6-2) to Werrington Junction (3.1m), 15m 17s and 15m 29s to Essendine (12.2m), 24m 52s and 25m 0s to Corby Glen (20.7m), and 28m 34s and 28m 26s to Stoke (23.7m). Rightly, in view of its smaller size, Allen balances this run in the Scot’s favour. Grantham incidentally (29.1m), was reached in 34m 49s and 34m 56s respectively, a gain of three minutes on schedule. For further comparison, City of Bradford ran to time in 37m 39s but King Henry VI, despite making a typically smart Western start, suffered two signal checks and exceeded the time allowance by three minutes though the net time was calculated at a minute less than schedule. There is no further mention of Belgian Marine’s work so one has to assume the correspondent left the train at Grantham.

35017 with dynamometer car backing into Kings Cross past the loco (shed).
Copyright Russell Coffin collection

Allen made a southward journey behind each of the four engines with the 7.50am Leeds-Kings Cross, 35017 featuring on 28th May. One particular oddity: Holbeck was the first stop, barely a half-mile from the start. Three minutes were allowed for that ½-mile but CJA remarks that with two minutes stand time there the departures were all 1-1½ minutes earlier than the public timetable. The normal nine-coach load out of Leeds was augmented by four more vehicles from Bradford at Wakefield, one from York at Doncaster and another from Lincoln at Grantham. In the case of Belgian Marine the gross tonnage increased from 320 to 455, 495 and 535. The ‘Scot’ on Allen’s trip was the most heavily laden of the four because the train was packed with supporters going to the Rugby League Cup Final at Wembley: from Grantham it hauled an estimated 545 tons.

The start from Holbeck is a difficult one, uphill at 1 in 50 for a half-mile followed by a short respite before the three miles at 1 in 100 to Ardsley. No 35017 tied with the ‘King’ in a few seconds over eleven minutes, cutting all but a minute off the allowance from Holbeck, but then drew slightly ahead before being stopped by signals outside Wakefield. This occurred on every one of Allen’s trips while the Bradford portion was drawn clear before attachment. He tartly remarks the 7.50 could have left Leeds ten minutes later and still picked up the existing schedule at Grantham without any real effort being demanded. As it was 35017 exceeded the twenty-one minute allowance to Wakefield by fifteen seconds, though Allen calculates the net at only 16½ minutes for the 9.9 miles.

Belgian Marine attained 48mph on the downhill 1.7 miles to Sandal and fell away only to 40mph at the head of the subsequent climb at 1 in 150, passing mp169¾ (6.1miles) in 9½ minutes, more than a minute faster than the ‘Scot’ and the ‘King’ and 1½ minutes quicker than the ‘Duchess’. It is virtually all downhill thereafter to Doncaster, 35017 attaining a maximum of 72½ mph before a signal check outside the station. Nevertheless, James cut the 26-minute allowance from Wakefield by 1¼ minutes, the only one of the four to better the schedule though Byford on the ‘Duchess’ may just have done so but for a similar check on the approaches.

Only the ‘Scot’ failed to keep the 24-minute booking to Retford, though of the four the ‘Duchess’ alone had an unchecked run, in 22 minutes. Allen, however, credits the ‘King’ with that time net, attained mainly because of driver Russell’s smart running over the latter part of the section. The ‘MN’ was assessed as taking ½-minute longer and the ‘Scot’ 24¼ minutes. The forty-four minutes allowed on to Grantham proved more than adequate though signals again checked three of the trains on the approach, the ‘King’ on this occasion being the beneficiary. The ‘Scot’ made the fastest time, of 40¼ minutes (net 38½), but only by dint of an excessive 77½ mph on the descent from Askham tunnel through The Dukeries, though Allen nevertheless comments that rain and mist made rail conditions difficult. Despite the check James, with 35017, cut the timing by 2¾ minutes though credited with the same net time as the ‘Scot’.

The train now is non-stop to Kings Cross on a schedule of 122 minutes for the 105½ miles. James made the fastest climb to Stoke in an outstanding time of 9 minutes 37 seconds for the 5.4 miles at 1 in 200 from the start, passing the summit box at 46mph. Allen comments that he believed the record had been set by the ‘King’ a week previously in a time of 10m 8s, and went on to say he had not discovered among his records such a fast start with a comparably loaded LNER engine as put up by the ‘MN’. (His further, more considered perusal shows that three ‘A4’s and a Gresley ‘A1’ had in fact bettered it but by a matter of seconds only. However, the rebuilt ’hush-hush’ engine beat it by ¾-minute!)

But then, having attained 70½mph by Corby Glen, James eased the engine for the rest of the run down Stoke bank until the speed had fallen away to about 55mph where normally trains would be travelling at 70. As there appeared to be no signalling responsible for this CJA wonders if it may have been due to caution by the pilotman, particularly as Russell had also eased King Henry VI through Essendine without apparent cause, possibly with the same man on his footplate. Nevertheless, 35017 was through Peterborough 1¾ minutes ahead of time. The ‘King’ was marginally quicker, having been a ½-minute behind at Stoke box but making a faster run along the level through Werrington Junction.

In typical Western fashion the ‘King’ made a smarter recovery from the 20mph limit through Peterborough North station but Belgian Marine climbed strongly to Abbotts Ripton to pass Huntingdon a minute early and 40 seconds ahead of King Henry VI. Thereafter, the two ran within seconds of each other until no 6018 was twice checked by signals, slight at Potters Bar - taking not quite a ½-minute more than 35017 from Hatfield - but a little more harshly at New Barnet on the run down into London. Thus James was through Wood Green more than three minutes ahead of Russell though both had to observe a pwr at Finsbury Park. It is apparent from the few intermediate timings shown both drivers had been intent on getting a little time in hand to compensate for this. No 6018 then had a clear run into Kings Cross but 35017 was checked to walking pace at Belle Isle though still arriving 1¾ minutes before time: the ‘King’ was a minute early.

Driver Brooker with the ‘Scot’ meantime averaged a mile-a-minute from Stoke to Stevenage after observing a slack at Corby Glen and the permanent slowing through Peterborough North, taken at 22mph. The Rifleman went through Huntingdon at 77½mph - the second time on the trip that speed had been reached - and took only 25½ minutes for the 27 miles thence to Hichin, by which time Brooker was back on schedule.

Neither LMS engine had had to observe the pwr at Finsbury Park, and though City of Bradford received a slight signal check at Belle Isle Byford arrived in 118¾ minutes from Grantham. Brooker also got in early, by 1¾ minutes, but Allen credits his engine with the fastest net time, 113¾ minutes. The ‘Duchess’ is accorded a net of 114¼, the ‘King’ 116 and Belgian Marine 117 minutes. I think this a little ungenerous considering the signal check at the close meant the last 2½ miles from Finsbury Park took no less than 7½ minutes, against the 3½ minutes of the ‘King’. I’d cut this ‘net’ figure by a half-minute and perhaps a bit more because James lost only a single minute on the 17 allowed from Potters Bar, the pwr to 31 mph and the severe signal check notwithstanding. (And Brooker twice grossly exceeded the overall line speed limit!)

This was an altogether happier event for James and Reynolds than Swain and Hooker experienced out of Euston though both crews acquitted themselves very well and certainly upheld the Southern tradition of attempting to run to time. Incidentally Bert Hooker tells of a conversation overheard by Inspector Knight between two sets of Crewe enginemen ‘travelling home’ who joined him at Rugby in the front brake van on one of the ‘Royal Scot’ trips. One told his mates to “get the cards out, this thing on the front [35017] will be an hour late into Crewe”. In fact they had a good onward run, unchecked by signals, and got to Crewe eight minutes early. He also tells amusingly of the instruction by a Polmadie pilotman coming aboard at Crewe to “fill up the firebox and fill the boiler” and who then got very agitated when Hooker, sitting quietly drinking a cup of tea, refused. “After all, I knew Belgian Marine – he didn’t!” He also had to restrain the fireman who had moved to pick up the shovel. The pilot crew was apparently astonished that after this slight disagreement the safety valves were sizzling as the train passed through Warrington. Then, having climbed Shap without trouble Jack Swain was moved to ask the Scotsman, “Where’s this Shap summit we’ve heard so much about?” The reply said it all: “Away! I’d give a week’s money to take this engine through to Glasgow!”

So, now to move west. The Paddington-Plymouth test provided as much of a challenge for the crews in the Express Passenger category as it had presented to driver Albert Pibworth on the LNER ‘A1’ Victor Wild in 1925, though by this time both Westbury and Frome had been by-passed. (On the test runs Westbury was actually served.) Other than the final three miles the westward climb to the summit at Savernake is not particularly steep, but the 35-mile long and debilitating hike up the Kennet valley from Southcote Junction, and the 17-mile eastward climb from Lavington both abound in curves. In its way the ‘Berks & Hants’ line is quite as difficult a proposition as the Devon banks. Other summits, at Brewham and particularly at Whiteball have steeper approaches, but all three climbs require skilled handling from both sides of the footplate.

SR Merchant Navy Class 35019 at Paddington, 1948.
Russell Coffin collection

James and Reynolds crewed French Line CGT between London and Plymouth, the three ‘pacifics’ being required to handle loads of between 505 and 525 tons gross to Newton Abbot on the 1.30pm down. The ‘Scot’, with the enterprising Brooker again in charge, had the benefit of one coach less, its train weighing 465 tons gross. It was the only engine Allen records to better the 45-minute schedule to Reading, though all four had exceeded the 23-minute booking to Slough, the ‘A4’ Seagull by 2¼ minutes. But with 22 minutes allowed for the next 17½ miles to the Berkshire town time might have been kept but for a pwr before Twyford that also affected 35019 on the detailed trip of 27th April. James got in 1¼ minutes late while Burgess with 60033 was a minute behind time. Byford and City of Bradford had made the best start but were still a ½-minute late at Slough. However, a signal check to 25mph before Maidenhead hampered the run into Reading, making the ‘Duchess’ 1½ minutes down also. Brooker meanwhile, untroubled after a pwr at Ealing Broadway came into Reading two minutes early.

Allen remarks that the seventy-minute allowance for the 59.6 miles to Westbury and particularly of only 42 minutes for the 34.1 miles to Savernake requires rather more effort than that very easily graded first 36 miles. But it is also rather puzzling the downhill 25½ miles thence to Westbury are allowed 28 minutes, curvature notwithstanding.

French Line CGT and Queens Westminster Rifleman made the best starts from Reading, both gaining more than two minutes to Savernake, no 35019 sustaining a very respectable minimum of 45½mph over the final three miles at 1 in 175/183/145/106 to the summit. But a bad permanent way slowing to 18mph after Pewsey, which affected none of the other engines, spoiled the run to Westbury though French Line had got up to 71½mph when passing Lavington at the foot of the decline. Allen reckons the net time for 35019, actually arriving ¾-minute to the good, to be about 4½ minutes less than schedule. Meanwhile Brooker, again with Queen’s Westminster Rifleman, seemed to suffer from an over-cautious pilotman, for not once on the descent to Westbury did he reach a mile-a-minute, but still got in 1¼ minutes early. By comparison Byford with the ‘Duchess’ made a steady but timely ascent to Savernake and then forgot himself by providing an uncharacteristic display of the engine’s capabilities, achieving 82mph at Lavington and arriving in 4¼ minutes less than the allowance. Seagull made a steady run without particular exertion going up or untoward speed coming down, contriving to arrive a minute under schedule.

James made the smartest start from Westbury but still dropped 1½ minutes on the impossible seven-minute timing for the 4¾ miles to Clink Road Junction at Frome. Having sustained 47½mph to Brewham French Line passed through Castle Cary ‘on the dot’ at the prescribed limit of 60mph. Brooker on the ‘Scot’ appeared still to be restrained by his pilotman for a pedestrian 65mph at Bruton was followed by a slowing right down to 48mph for the Castle Cary slack. It is tempting to suppose the Western man did not want the ‘foreigner’ to do too well. More charitably, maybe he found the riding to his distaste: the ‘Scots’ could get very rough as mileage built up but at the parameters set by the test that should not have been a factor here. Perhaps the man himself was overly cautious at the regulator which on his ‘home’ engine would have been on the right side of the footplate, in both senses of the word to him. As it was Byford on the ’Duchess’ upheld LMS prestige by repeating his showing on Savernake at Brewham, topping the summit at 44mph but then coming down at a sprint, touching 77½ at Bruton and being checked only to 65mph through Castle Cary. As a result he was all but 1¼ minutes early at Creech Junction. French Line and Seagull both took a half-minute longer but Brooker, who had had to observe a pwr after Curry Rivel following the descent from Somerton tunnel, was by now almost six minutes late passing the junction. All four engines were slowed by another pwr to 15mph before the stop at Taunton but no 35019 had managed to cut a bare ¼-minute off the 53 scheduled from Westbury. The ‘A4’ and ‘Duchess’ were both on time but the ‘Scot’s lateness had increased to 6½ minutes.

James was the only driver to keep the sharp 38-minute booking net from Taunton to Exeter. The climb from Norton Fitzwarren to Whiteball begins quite easily but the four miles between Wellington and the summit at Whiteball siding includes 2½ miles at 1 in 90/86/80 before the easing to 1 in 127 through the 50 chains of the tunnel itself. With 505 tons on the drawbar 35019 was making 50mph at Wellington (7.1m), entered the tunnel at 31mph and left it at exactly 30mph, covering the 10.9 miles from Taunton to summit in 17¼ minutes. The engine had accelerated to 75mph by Tiverton Junction, been eased for the curves through Cullompton and Hele and would have made Exeter in 36½ minutes but for an annoying slowing to walking pace by signals at Stoke Canon. Actual time was 40m 52s to a stand at St David’s. None of the other engines got to French Line’s pace at Wellington and all lost more than three minutes net to Exeter.

Even given the all-but level road the timings between Exeter and Newton Abbot are demanding with the loads being hauled. Brooker with the ‘Scot’ overturned things by being the only driver to better the sixteen minutes allowed to Dawlish, but that by rather overstepping the limit through Starcross. Allowing for delays demanded by pwr checks James was the one driver to keep time to Newton, indeed Allen credits him with a gain of ¾-minute. The load was reduced here to 335 tons gross following detachment of the Torbay portion. The climb to Dainton begins a little less than 1½ miles from Newton Abbot but on the run recorded by Allen no 35019, observing a pwr to 27mph right at the foot at Aller Junction, gained very little impetus for it. Thus the engine lost 1¼ minutes on the eight-minute allowance to Dainton, passed at 25mph though accelerating from a minimum of 21½ over the slightly easier last ¼-mile. Despite the curving and speed-restricted run down to and through Totnes James regained a half-minute before the nine-mile climb to Wrangaton. But for once French Line struggled, falling to 20½mph at Tigley (11½ miles from Newton) and topping Rattery a minute late. A further ½-minute was lost by Brent though Allen calculates James to be just on the right side of time given the delay at Aller and not allowing for departure from Newton Abbot 1½ minutes ahead of the public timetable. A ¾-minute gain to Hemerdon siding and a swift run down the 1 in 42 from there, with a maximum of 69mph before Plympton, helped alleviate the effects of a 17mph pwr at Lipson Junction, James bringing the train into North Road at 6.55pm, on the dot of the advertised arrival time.

Eastbound runs were hampered from the start by a succession of restrictions between Plymouth and Newton Abbot such that timekeeping was impossible, the first, at Plym Bridge, near Tavistock Junction, being about 1½ miles from the foot of the climb to Hemerdon. Having observed this at 20mph on the return on 28th April, French Line had recovered to 42½mph before tackling the 1½ miles at 1 in 42. This time James made a very fine climb, speed falling to a minimum of only 17½mph at Hemerdon “but not”, according to Allen, “appearing to be in any trouble at any point”. But for the check at Plym Bridge he estimates no 35019 would have passed the summit (6.7m) in fourteen minutes from the start rather than the 14½ minute schedule and the actual time of a few seconds over 16 minutes.

But this effort was surpassed by Seagull which had had to come down to 16mph for a relaying slack in Plympton station itself. Speed went up to 24mph on the short 1 in 111 immediately after it and had only fallen away to a very creditable 18½mph after the 1½ miles at 1 in 42 to Hemerdon, though by then 60033 was three minutes late. Only the ‘Scot’ escaped any other permanent way checks, no 35019 having two in the 9½ miles between Hemerdon and Brent, the latter station being passed four minutes late. A minute had been recouped by Totnes, passed at a restricted 40mph. But then followed another excellent climb, taking only 7¼ minutes for the 4.9 miles to Dainton with speed being 23½mph at the summit. James was moderately checked by signals at Aller Junction and so brought his 350-ton train into Newton Abbot just over 2¼ minutes behind time though credited with a two-minute net gain on the 51-minute schedule from Plymouth.

Allen remarks that the nine-minute allowance for attaching the Torbay portion ought to have been ample though it was exceeded in every case, French Line CGT being detained for no less than 13½ minutes. Departure with a load now of 525 tons gross was seven minutes late. But 35019 was the only engine to make Exeter in the 26 minutes allowed (actually four seconds over!). Then, barely away from St David’s, James was brought to a stand for four minutes by signals at Stoke Canon, losing all the impetus for the climb to Whiteball. Nevertheless, speed had fallen from 60mph at Cullompton only to 47½ at Burlescombe, almost at the head of the two miles at 1 in 115 that lead to the summit at Whiteball siding. Time recovery was hampered by yet another pw check just beyond the summit, arrival at Taunton thus being 4¼ minutes over the time allowed: James left Taunton nine minutes late.

Gradients are relatively easy for the next twenty-three miles or so but then comes the trying seven miles to Brewham finishing with 1¾ miles at 1 in 98/81. French Line was over the top in 38 minutes dead from the Taunton start at 35mph and gained 3½ minutes on the schedule to Westbury. Leaving there now only 6¼ minutes late, 35019 made an excellent climb to Savernake, passed at 55½mph, and thereby gained 2¾ minutes on the 37-minute allowance to Bedwyn. Signals slowed the train to 24mph before Kintbury but James was still through Newbury in a half-minute less than allowed.

The load was reduced by 40 tons on shedding the Reading slip coach, though passing through the platform loop there at 40mph provided some lively riding. The last scheduled check, to 36mph at Twyford, was followed by a sprint for the terminus with maxima of 71½ at Slough and Ealing Broadway. Taking 105m 10s from Westbury James had brought his train into Paddington 1½ minutes before time. Allen estimated the net at 101 minutes, or 12 minutes less than schedule. He also calculates French Line CGT could have come up non-stop from Exeter, without delays, in a shade over three hours for the 173½ miles, an average of about 57½ mph.

While this was an excellent performance against the odds it did not match that of the ’Scot’ over the latter section. For some reason Brooker had not done too well over the Devon banks or to Brewham and so got away from Westbury 14½ minutes behind time. Whatever his pilot driver may have advised Brooker was clearly determined to salvage the reputation of himself and his engine. The start was as a shot from a gun with Pewsey passed at 70½mph and Savernake topped in only 27½ minutes at 50, the speed reduced to that from 69mph at Burbage siding for the curves. Another acceleration to 70½ at Kintbury preceded a 30mph pwr before Newbury, but Brooker was still through the station five minutes under schedule. Speed had recovered to 71½ before the slowing at Southcote junction in preparation for Reading. Such was the custom here the train was forced to stop rather than ‘slip’, being held in Reading’s usual slothful fashion for all but three minutes, though arrival had been eight minutes under the scheduled passing time. But Brooker was away in almost a minute less from Westbury than it had taken James to pass through, and from the standing start the Rifleman had achieved all but ‘even time’ by Old Oak Common before a final pwr to 41mph at Westbourne Park. Allen puts the net time for this remarkable run at 96½ minutes, 16½ less than allowed though the train was actually into Paddington 6½ minutes down.

On that note let us turn to the working of Bude in the General Purpose category over much of the route just discussed. Nothing at all was recorded by Allen regarding any westbound workings from Bristol neither, he comments, did he receive any correspondence on this from other timers. This is a pity for it would have been very interesting to see how the GP engines would have handled the loads in this direction over Whiteball and the Devon banks. So we are restricted to looking at the Up times, which are, happily, well worth reiterating.

With a gross load of 260 tons from Plymouth Bude suffered the same 20mph pwr check before Plympton as all the Express engines, and was further checked through that station to 30mph at the foot of the climb to Hemerdon. As a result speed fell to 15½mph at the top of the 1 in 42 which, together with a further pwr just over the summit, put the engine 3¼ minutes late at Brent. Another engineering restriction beyond that station hampered the possibility of any real time recovery down the grade to Totnes. The three miles to Dainton summit were completed in 6¾ minutes, a minute quicker than either the ‘Black 5’ or the ‘B1’, though speed had fallen to 16mph at the top. Restrained running around the curves down to Aller Junction precluded any recovery in the 3-minute lateness being made before passing through Newton Abbot.

The story is taken up again with departure from Exeter St David’s, the load being increased here to 449 tons net or about 475 tons gross. No 34006 seemed not to notice, the 19.9 miles to Whiteball being reeled off in 22¼ minutes start-to-pass, an average of 54mph. Speed had reached 68mph before the 60mph slack through Cullompton and rose again to 66½ at Tiverton Junction for the final assault through Burlescombe, where speed had fallen only to 53mph. No note is made of the rate at the summit itself as the engine was already braking for another pwr north of Whiteball tunnel. But 34006 went through Wellington at 76mph and came to rest at Taunton in 34m 53s from St David’s, the only engine bar one of the seven timed over this section - the ‘Black 5’- to better the 38-minutes allowed. Allen calculates the net time as 33¾ minutes.

If this were a good showing the run onward to Bristol produced some real fireworks. Having got to 55mph in the 2½ miles to Creech Junction speed rose rapidly to 73mph after Bridgwater and continued so along the level past Dunball, Bude achieving ‘even time’ by Highbridge, 17.9 miles from Taunton in 17m 37s. The engine went over Flax Bourton summit at 70mph and then got to 75 through Long Ashton on the descent to Bristol. Such joie de vivre couldn’t last, the train facing several signal checks before being brought to a stand at Temple Meads West box no less than eleven minutes early. It eventually drew into Temple Meads in 50¾ minutes against the 53 allowed. From Exeter Bude had cut no less that 16½ minutes net off the 93-minute timing. It is only fair to add that both the smaller engines coped well with the load. Driver Smith and the ‘Black 5’ made an unchecked run from Taunton to Bristol a minute faster than Bude, though no 45253 had been four minutes behind only a mile before Temple Meads. Similarly, driver Ratter and ‘B1’ no 61251 arrived in 2¼ minutes less than schedule despite two signal checks in the approaches, though they had been five minutes behind Bude at Flax Bourton. But in neither case could the net time match that put up by the ‘pacific’. One oddity may be noted: the working timetable allowed two minutes more from Taunton than the public one.

So, from such rather tantalisingly scant but very satisfactory details from the Southern’s point of view, back to London and the two routes to Manchester. CJA contends with some justification that, considering the loads hauled, the work out of Marylebone in particular was the hardest set the General Purpose engines.

Only the ‘B1’, ‘Black 5’ and ‘West Country’ worked on the ex-Midland route. Like its West Coast counterpart the southern part of the line abounded in checks for track renewals. Allen himself did not make any runs out of St Pancras but reproduces those of another very experienced timer, though all start from and one finishes at Leicester. In all cases too the northbound workings set out late from Leicester due to the restrictions further south. The Southern’s entrant in this instance was 34005, Barnstable, again crewed by the Georges, James and Reynolds, and the only one of the three that consistently kept sectional timings. The recorded northbound trip was made on 22nd June with a gross load of 325 tons. The start from Leicester is down the gentle slope of the Soar valley as the river heads for the Trent so, on the face of it, the timings noted ought to present few if any problems. Sixteen minutes are allowed for the 12½ miles to Loughborough, George James taking full advantage of this favourable road to pick up two minutes on the booking with a maximum of 70mph at Barrow-on-Soar. A further 2½ minutes were gained to Derby, 72mph having been reached at Kegworth before the 50mph slack for the curves at Trent Junction where the climb over the Pennines begins. This is easy for almost all the first thirty miles or so, though James dropped a half-minute to Matlock, mainly by being over-cautious through the curve at Ambergate. Time was just kept to Millers Dale despite a pwr at Rowsley, right at the foot of the fourteen steep miles to Peak Forest. Except for three very short downhill sections the gradient is continuous, predominantly at 1 in 100, but it finishes with 3¼ miles at 1 in 90.

Following observation of the Rowsley restriction at 17mph, Barnstable had accelerated to 42½mph by Headstone tunnel (6¼ miles) and then to 53mph in the following ½-mile dip. Speed fell back on the rise before the next dip, to Millers Dale, but only to 42mph. Ten minutes are allowed for the 4.6 miles thence to Peak Forest, James cutting a quarter-minute off this and making 64½mph once over the summit. The curve at Chapel-en-le-Frith demanded a slowing to 55mph and another pwr before Chinley brought speed down to 18mph. Despite these slacks James cut two minutes net off the twenty scheduled to this point from Millers Dale.

The final 19.7 miles into Manchester are steeply downhill as far as Cheadle Heath, passed at 72mph on the dot of the 13-minute allowance, before the usual slow run in from Chorlton Junction. Barnstable reached the Central station a quarter-minute under the 25 scheduled from Chinley and dead on time, the only one of the three recorded engines to do this. Allen estimates the net gain from Leicester at 8¾ minutes.

Heading south the following morning Barnstable departed with a gross load also of 325 tons. Allen’s correspondent apparently did not join the train until its first stop, at Chinley, but some skeleton data were available. Cheadle Heath, right at the foot of the climb to Peak Forest, was passed a ½-minute early at 60mph, rather faster than either of the 4-6-0s, and the engine sustained a minimum of 33½mph on the 1 in 87/89/90 to Chinley. Arrival here was no less than 4½ minutes before time. Getting smartly away again no 34005 had accelerated to 33mph up the 1 in 90 to Dove Holes, cutting two minutes off the 14 allowed to Peak Forest. But the reduction in speed to 31½mph during passage of the near-1¾ miles of Dove Holes tunnel, which also rises at 1 in 90 for its entire length, suggests some easing of the regulator. The tunnel is driven through porous rock and is therefore wet though there is no indication the engine suffered from slipping: the easing may then have been precautionary.

Following a quick burst at 60mph Barnstable was slowed to 48 on the approach to Millers Dale but still came to rest there in a time of 17¾ minutes against a schedule of 21. On the run down to Matlock a reduction in speed to 40mph was made at Hassop, 20mph less than either of the 4-6-0s, though the cause is not shown. That and the pwr at Rowsley and another to 13mph after Darley Dale brought no 34005 into Matlock in 2¾ minutes more than the sixteen allowed. Signals spoiled a fast run thence into Derby though time was kept.

Getting smartly away from Derby the 72mph recorded at Sawley came to an abrupt end with another signal stop of 3¾ minutes duration at Trent Junction, Loughborough thus being passed 4¾ minutes late at 72mph. Despite 70½ at Barrow-on-Soar and 66½ at Syston Barnstable was still a minute behind time into Leicester. Allen, however, estimates a net time of no more than 30½ minutes for the 29.4 miles from Derby, a gain of 6½ minutes on schedule. Both 4-6-0s had had a clear run and thus no need to hurry, the ‘Black 5’ – its ‘homeground’ working unusually being recorded - arriving 2¾ minutes early and the ‘B1’ 1¼ minutes before time.

The continuation to St Pancras illustrates the effects of the engineering restrictions. No 34005 had barely left Leicester before a pwr reduced speed to 29mph. Although a further check to 28mph interfered with progress James passed Market Harborough only a quarter-minute down and then went over Desborough summit at 48mph after 3½ miles at 1 in 132. The run down to Wellingborough was spoiled by another pwr, this time before Kettering, to 26mph. Nevertheless, the train was through Wellingborough three minutes under schedule and James set about the climb to Sharnbrook with enough gusto to gain another minute at the summit, topping it at 47½ mph. PWRs either side of Bedford, to 17 and 30mph, provided incentive for another swift acceleration, up the long 1 in 200 gradient to Leagrave, cleared at 52½mph before a 68mph dash through Luton, six minutes ahead of time. With speed held in the high-sixties a further ½-minute had been acquired to St Albans despite the penultimate pwr of the trip, to 17mph on the approach. Hendon was passed at 74mph before the final slowing for engineering work, after Finchley Road, to 16mph. But a signal check at Kentish Town spoiled the run into St Pancras though Barnstable came to rest in 115¾ minutes from Leicester, 3¼ minutes early. Allen calculates no 34005 showed an overall net gain of no less than 29¾ minutes from Manchester, 16½ minutes of that from Leicester. It may be added the ‘B1’ acquitted itself very well: Allen calculated a net time of 111 minutes from Leicester against the schedule of 119, though it arrived ¾ minute late.

We move west along and beyond the Euston Road now to the quietest of all London termini, to take the ex-Great Central route to Manchester. As before the ‘West Country’, in this case no 34006, Bude, crewed by Swain and Hooker, was rivalled by a ‘Black 5’, no 45253 again, but this time a ‘Hall’ is included, the now-preserved no 6990, Witherslack Hall. Bude’s first recorded trip is northbound on 8th June with 380 tons gross. Allen notes the line out to Rickmansworth “was crammed with permanent way restrictions” and track rearrangement in progress at Harrow. Nevertheless, despite checks to 18mph at Wembley Park and 5mph on the steep Harrow-on-the-Hill approaches, Swain had dropped less than ¾-minute on arrival there, mainly by virtue of a very fast start, Neasden South Junction, five miles from Marylebone, having been passed in 9¼ minutes at no less than 66mph. The restriction continued on leaving Harrow but then no 34006 faced a succession of signal checks culminating in a dead stand at Northwood Hills. That and a further pwr to 30mph contributed to the train passing Rickmansworth three minutes down. Undeterred, Swain made an excellent climb up the 6½ miles to Dutchlands summit, constantly at 1 in 105 other than short levels through the stations at Chorleywood and Chalfont & Latimer, reaching 45mph at the top just beyond Amersham, and then falling away from 71½mph to only 60 on the subsequent four miles uphill through Great Missenden. There followed a 74mph dash through Wendover before another pwr, this time to 35mph, preceded arrival at Aylesbury 2¾ minutes over the allotted forty. Neither of the other engines came anywhere near this time though neither had suffered a signal stop and the ‘Hall’ had had a clear run this far.

There followed what Allen describes as a “joyous sprint” when Bude achieved all but ‘even-time’ over the 31.2 miles to Woodford. Having got to 76½mph after Quainton Road speed had to be reduced to 67 for Grendon Underwood Junction. Then, from 71½mph after Calvert the near-six mile long rise through Finmere was topped at 59, followed by another swift acceleration in the short dip thereafter that saw Helmdon passed at 64mph. A final 69mph at Culworth Junction was enough to provide a Woodford arrival in 31m 24s from Aylesbury against the schedule of 35 minutes. This works out at an average of 59.6mph start to stop over a generally rising road, most of those rises being at 1 in 176.

There was no ‘fat’ whatever in the 17-minute timing for the 14.1 miles to Rugby though Swain may just have made it but for a 30mph pwr before the station. The maximum on this section was 76½mph at Braunston, at the foot of the 6½ miles downgrade at 1 in 176 from the south end of Catesby tunnel. Allen estimates the net at 16 minutes against Bude’s actual 18½. The continuation on to Leicester saw Swain cut a few seconds short of three minutes off the 23 allowed for the 19.9 miles, another 59mph+ average, with a maximum of 75mph at Ashby Magna before the engine was eased at Whetstone. More significantly no 34006 had gained most of those three minutes on the uphill 1 in 176 grades either side of Lutterworth.

Bude kept exact time on the tight allowance of 12 minutes to Loughborough but despite a fine acceleration away from there at 53mph to Barnston tunnel, a long slowing to 35mph because of pitfall meant the 17-minute allowance to Nottingham Victoria was exceeded by two minutes. It is worth noting that Bert Hooker said he found the ex-GCR line “a lovely railway”, with sweeping curves and long gradients. Certainly the later-built section south of Annesley – the true ‘Great Central’ - could be so described but north of there the old MS&LR was into mining country and from Sheffield faced the long and steep climb over the Pennines.

Bude left Nottingham on time and proceeded to give yet another demonstration of the hill-climbing ability of the Bulleid ‘pacifics’. The gradient is rising almost continuously at 1 in 130/132 for eleven miles to Kirby South Junction, but it starts with the damp Mansfield Road and Sherwood Rise tunnels right off the platform end. Emerging from the latter at 24½mph no 34006 accelerated to 33mph before a level mile took speed up to 48. The engine maintained this up the following climb until observation of a relaying check to 20mph at Hucknall. But then came an exceptional acceleration, to no less than 50mph on the ensuing four miles at 1 in 132 to top the rise in 19¼ minutes from Nottingham. Mining subsidence now took its toll with four severe slowings before the call at Staveley Town though Swain took only 41¾ minutes against the 43 allowed. He also cut the 20-minute allowance on to Sheffield Victoria, two permanent way checks notwithstanding, arriving 1¼ minutes ahead of schedule. After a half-mile level there follows the nineteen-mile climb to Dunford Bridge and Woodhead tunnel. No 34006 had got to 40mph up the five miles at 1 in 120 from Oughty Bridge before a pwr check to 15mph through Thurgoland tunnel, and so lost ¾-minute on the 25-minute booking from Sheffield to Penistone.

Things did not improve, for having reached 41mph on the 1 in 124/135 above Penistone signals at Dunford Bridge brought the train down to walking pace. Taken with an uncharacteristically slow run through Woodhead tunnel, perhaps because the check was due to a train not too far ahead, and another pwr at Valehouse, Bude came into the stop at Guide Bridge all but five minutes late. Timekeeping matters weren’t helped by the train having to be drawn up at both Penistone and Guide Bridge but even with a final engineering check at Fairfield Bude was only four minutes late into Manchester (London Road).

Incidentally, on an earlier down trip the drop grate on Bude fell into the ashpan between Rugby and Leicester and could not be repositioned unless much of the fire were drawn. Being in no fit state to continue an ex-GC class ‘O4’ 2-8-0 was commandeered to haul Bude and its train to Leicester where the problem could be resolved over a pit. Bert Hooker relates that with much of the fire removed the bars went straight back into the proper position. But arrangements then had to be made for the engine to run ‘light’ to Gorton, not merely finding a path but a pilot driver to accompany the crew. As he relates “...…we did not arrive in Gorton loco until about 6.30 in the evening and as usual in Manchester it was raining”. He also makes the point the hostel there wasn’t a patch on the one at Upperby! The whole drop grate operating assembly was later redesigned and made more robust.

The return of Bude on 9th June was made with a train of 390 tons gross. The start is a difficult one, for the climb of nineteen miles to the east end of Woodhead tunnel begins almost off the Manchester platform end. In addition the train had to call at Guide Bridge five miles up the bank, the scheduled time of nine minutes appearing very demanding, particularly with a ‘cold’ engine. Moreover, a pwr imposed after Fairfield prevented any sort of recovery before the stop. Swain had got Bude up to 41½mph at Gorton where, it appeared, most of the Works staff had turned out to see the ‘stranger’ pass by. Having observed the pwr at 23mph Swain exceeded the allowance to Guide Bridge by almost 1¾ minutes. The ‘Hall’ was but seconds behind though as it took a half-minute less in from Fairfield the suggestion is driver Russell perhaps showed a little less regard for the restriction. The ‘Hall’ made the better start from here and was through Mottram (4.8 miles) a minute up on Bude. But the rapid acceleration that followed saw the ‘West Country’ on time at Valehouse before being slowed by another pwr, to 27mph. As the ‘Hall’ did not have to observe that restriction Russell was a minute ahead at Torside Crossing (10.3m). But then Swain poured on the power, sustaining 38mph up the five miles at 1 in 117 past Crowden and entering Woodhead tunnel at 37½mph, now 1½ minutes up on schedule and only a few seconds behind the ‘Hall’.

Although the gradient through the tunnel is at 1 in 201 Bude dropped back to 32½mph. Bert Hooker remarks that the tunnel was very wet and the engine was prone to the odd bout of slipping. Not surprisingly then they lost a half-minute on the five-minute booking to traverse it. The ‘Hall’ meanwhile, from being ahead at Woodhead, also lost ground and its lead in the tunnel, coming to rest at Penistone one minute ahead of the 42 allowed but a half-minute behind Bude. The ‘Black 5’ did not do too well on getting away from Guide Bridge and though not having to observe the pwr at Valehouse was all but 1½ minutes over time at Dunford Bridge. However, fast running down to Penistone brought no 45253 in ¾-minute to the good. Bude’s very restrained running down the hill combined with the check at Thurgoland tunnel caused a 2-minute late arrival at Sheffield Victoria and a further half-minute was dropped to Staveley Town by reason of pwr cautions at Woodhouse and Eckington. But the superb hillclimbing ability of the ‘West Country’ was again demonstrated on the near-four 1 in 100 uphill miles to Springwood tunnel and the next five rather easier miles to the summit at Pilsley, in which four minutes were gained on schedule. Thereafter pitfall slowings prevented any further gains, though Bude arrived at Nottingham Victoria in 2½ minutes less than the 43 allowed from Staveley Town.

Alone of the three no 34006 had to observe a pwr to 30mph soon after passing Arkwright Street but with a maximum of 67mph down the 1 in 176 from Barnston tunnel came into Loughborough only 2½ minutes late. Neither of the other two engines managed to keep the 16-minute timing and Bude alone maintained it over the two brisk snippets to Leicester and Rugby. The latter was achieved by an excellent climb up the 7¼ miles at 1 in 176 through Ashby Magna where speed had fallen only to 50mph over the 4½ miles from Whetstone, near the foot of the bank, passed at 61mph. But a pwr and signal check to 5mph conspired to lose 34006 two minutes on to Woodford though the tight 14-minute timing to Brackley was bettered by 1¾ minutes and Aylesbury reached another half-minute to the good.

Other than its immediate approaches and easier gradients through Stoke Mandeville and Wendover stations, the 6½ miles from Aylesbury to Dutchlands summit are almost entirely at 1 in 117. Allen describes Bude’s climb as “….one of the most startling exhibitions that I recorded during the whole series of tests and one to which I know of no parallel with any ex-LNER locomotive similarly loaded”. This time he was right! Perhaps the incentive was provided by Swain having to ‘set back’. On the 1¼ miles to Stoke Mandeville speed had risen to 37½mph and had increased to 48mph before Wendover. The brief 1 in 264 through the station saw it rise to 50mph, “…. with 395 tons of train (the passenger complement had increased considerably”), and finally to 51mph at the summit. Despite an annoying pwr on the descent at Great Missenden the ‘pacific’ cut 4½ minutes off the 32-minute allowance to Rickmansworth. But the operating thereafter went awry with a prolonged signal stop before Harrow to add to the pwr there, costing some seven minutes in running. Eventually Bude arrived at Marylebone four minutes late, a disappointing end to what had been a very spirited trip. Allen makes no attempt to calculate the net time from Aylesbury but I would put it at no more than 50½ minutes against the 57 scheduled. Of all the tests in which the ‘light pacifics’ were involved this is, I think, the most notable in revealing the true extent of their power and ability in the right hands.

That brings us then to the final route and a trip of 450 miles to reach the starting point at Perth. This was made in two stages, no 34004 double-heading a ‘Royal Scot’ on the 5.05pm Euston-Holyhead train as far as Crewe and the next day double-heading another ‘Scot’ on the 8.55am Crewe to Perth as far as Carlisle. There it became the train engine with a ‘Black 5’ piloting for the rest of the journey. One comment made by the pilotman north of Carlisle caused Bert Hooker some amusement. With 34004 running at 70+mph on the descent from Beattock, ‘riding as beautifully as only a ‘West Country’ could and the ‘Black 5’ in front of us fairly leaping about’, the man commented ‘…. in all seriousness, “if he’s going too fast for you put the brake on a little bit”’. Apparently the irony in Swain’s laconic reply, ‘We can stand it if he can’, was lost on him.

The first ‘learning’ trip to Inverness took place with the 4.0pm from Perth on Monday 5th July. Allen makes a comment in his record of events that bears repeating. Much of this route is single track with some quite severe curvature and, of course, passing loops with tablet catchers, another skill to be mastered. Being Nine Elms men the crew faced nothing like this in their everyday work. Allen contends with some justification that despite the very good showing Swain and Hooker ultimately made a crew from, say, Exmouth Junction or one of its more outlying sub-depots in Devon and Cornwall, who regularly worked over similar routes, might have been more comfortable with it, particularly in running swiftly downhill. How much the work and the results may have differed from those actually produced is impossible to know but it would have been fascinating to find out. (Had these tests been conducted, say, three years later, by which time four ‘light pacifics’ had been allocated to Bath Green Park shed, a Branksome crew would no doubt have relished the challenge of the Highland Line which differed little in substance from their own heavily graded and curvaceous route over The Mendips.)

In view of the particular difficulties the line posed Jack Swain had naturally to be very reliant on his pilotman and that raises another point. Both the ‘Black 5’ and the ‘B1’ with which Yeovil was ‘competing’ were manned by Scottish crews and in the case of the ‘B1’ at least accompanied by Scottish pilotmen. How easily understandable then was the necessary communication on Yeovil’s footplate between the Highland Scot and the South Londoner, another fascinating question?

The first northbound test run that Allen tabulates came on 13th July, also with the 4.0pm which called at almost all stations en route. Allen himself made no trips behind any of the engines but relied on information received from several correspondents. When compared to that of the ‘B1’ a week later this trip illustrates how the engine was restrained downhill, sometimes excessively so, surely primarily the responsibility of the pilotman. (Perhaps like the LMR man on the way north the Scot thought these strangers from the little railway in the south unused to fast travel!) The load of 380 tons gross was lightened at Aviemore by detachment of the portion routed to Inverness via Forres and along the coast through Nairn: the train to be taken over Slochd then weighed about 275 tons.

The brisk start from Perth saw the twelve-minute schedule to the start of the Highland Line at Stanley Junction cut by 1½ minutes, the train already being halfway up the seven-mile climb to mp 8½, more than a mile of this being at 1 in 93. Yeovil went over the top at 34½mph still 1½ minutes up. But then for the tablet exchange at Murthley Swain was brought down to 33mph against the 50 or so at which a regular crew would probably make it. By doing so he lost impetus for the 2½ miles up to Kingswood Crossing, much of it at 1 in 82. But the engine again showed the hillclimbing ability of the Bulleids with a minimum of 28mph at the top and now with three minutes in hand. However, a slow approach and a stand of 2¼ minutes outside Dunkeld while a ballast train got itself into the up loop saw all that time lost. Another relatively slow passage of the loop at Dalguise and a pwr to 20mph soon after conspired to bring no 34004 into the call at Ballinluig 1¼ minutes late. Another 1¾ minutes were lost on the eight minutes allowed to Pitlochry and a further minute to Blair Athol. Maybe these together put Swain on his mettle for the seventeen steep uphill miles to Druimuachdar summit, mostly at 1 in 70. As usual assistance was provided ‘up the hill’, but for some reason no 14501, a Pickersgill class ‘3P’ 4-4-0, banked Yeovil rather piloting as was apparently customary. The two cut the nine-minute allowance to Struan by 1¾ minutes and then the ‘West Country’ gave a startling illustration of power over the next 11¼ miles to Dalnaspidal. Instead of the thirty-one minutes scheduled the two engines came to a stand in 19m 23s, a start-to-stop average of all but 35mph. Whether the ancient at the back actually had been winded by the speed of ascent is true or myth is of no consequence. The fact is this was no less than 5¾ minutes faster in running time from Blair Athol than taken by the ‘5’ which, on its runs, did not call at Struan.

Once over the summit, in just under five minutes for the 2.1 miles from Dalnaspidal, time was lost consistently by cautious progress downhill and particularly at passing loops. The three-minute excess over time for the 7¾ miles to Dalwhinnie could not be solely attributed to the 20mph slack before the station. And the slowings for tablet exchange at Inchlea and Etteridge would account for the 1¼ minutes lost to Newtonmore. (The ‘5’ at these points was travelling at 50 and 47mph against Yeovil’s 36mph and 33.) Further losses, of ¾-minute to Kingussie, a half-minute to Kincraig and a full minute thence to Aviemore were due to rather slow acceleration away from the stations, possibly due to the weather.

This had been bad since leaving Perth but now worsened into a typical Highland storm with very heavy rain and a strong wind. Yet in these conditions Swain, with a train now of 275 tons only, apparently had little difficulty regaining his losses during the climb to Slochd. A pwr after Aviemore was strictly observed and Swain had got no 34004 to 45mph up the initial burst at 1 in 150, falling back only to 41mph after 1½ miles at 1 in 75 to mp88, thus recouping 1¼ minutes on the 13-minute booking to Carr Bridge. Fourteen minutes are allowed for the almost 5½ miles thence to Slochd, graded at 1 in 60/70 with a short intermission at 1 in 92. Swain cut two minutes off this time and retained that advantage down the winding 1 in 60 to the call at Tomatin. No less than 3½ minutes were gained on the 20-minute schedule to Culloden Moor though mainly because neither of the conditional stops, at Moy and Daviot, was called. There followed the one and only time a mile-a-minute was recorded on this trip before a signal check at Milburn Junction. But arrival at Inverness was still ¾-minute less than allowed.

Allen estimates the net gain on schedule by Yeovil and her crew to be 20½ minutes, “a first class performance”. But he also cites another run made on 7th July, during the familiarisation week, when no 34004 was saddled with a 13-coach train of some 415 tons gross for the first 23½ miles to Ballinluig. The start was slower than on the test trip and though on this occasion the tablet change at Murthly was achieved at 53mph the various engineering slacks were as bad. Nevertheless, Swain cut a half-minute off the 41-minute allowance. From Ballinluig the load was reduced to 380 tons by loss of the Aberfeldy coach but two minutes were lost to Blair Athol. No 15401 again provided assistance up the hill though on this occasion the train took a minute more to Struan than on the breathless run with the test train, and five minutes more onward to Dalnaspidal. But that aggregated still to seven minutes less than schedule. Point-to-point times were kept to Aviemore and with only 165 tons to be taken on to Inverness six minutes were gained to Slochd, where the 41mph speed had to be reduced to 30 for the loop, resulting in a net gain of fourteen minutes to Inverness.

One factor Bert Hooker mentions is leakage from the oil sump, specifically from small cracks in the casing. The opinion of Inspector Knight is that the engine “…. doesn’t like these sharp curves”. The problem was solved temporarily by pressing soap into the cracks. Yet I wonder. The coupled wheelbase of a ‘WC’ is three inches less than that of the ‘Black Fives’ that regularly worked the route, and eighteen inches less than that of the ‘B1’. The engine’s chassis would therefore theoretically suffer less stress than either of the others in trying to straighten out those curves and particularly as the layout of the frames over the axleboxes ought to have provided less distorting leverage. That being so I’d surmise it was not so much the fault of the engine per se but specifically that of the sump. The power for the engine’s flood lubrication system is based in this casing, the pumps being driven off the valve gear’s three-throw crankshaft and delivering oil through a piped system to all points in the oil bath needing it. Did the ‘light pacifics’ working on other routes noted for curvature, those of the further-flung parts of the ‘Withered Arm’ for example, or later on the Somerset & Dorset, suffer from similar leakage problems. Depositing oil on the track or within the boiler cladding occasionally leading to fire was not unknown. (See Peter Smith’s “Mendips Engineman” for an example.) So, were cracks in the sump or bath casing cause for the one and perhaps a contributor toward the other? If so it appears to be little recorded. Whatever the case Yeovil certainly suffered a small degree of leakage during her Highland stay though the soap seems to have much reduced it.

Returning to the records again, the start from Inverness with a ‘cold’ engine can be a real trial. Following a level mile the line climbs for the next 21¾ miles from sea level to 1315’ at Slochd summit. Other than ¾-mile down through Culloden Moor the gradient is 1 in 60 for almost three miles, 1 in 70 for the next three miles and, after a slight easing through Daviot station, another 1¾ miles at 1 in 60 to mp105¼ between Daviot and Moy. A downhill breather at 1 in 200 precedes six miles of relatively easy gradients to Tomatin where the 1 in 60 resumes for the final 3½ miles to the summit. The gradient averages 1 in 87.3 for those 21¾ miles.

Not only did Yeovil have to contend with weather conditions similar to those of the previous day but a fish train that preceded it managed to leave a trail of unwanted lubrication on the track. Nevertheless, on the 8.20am to Perth the following morning Swain had got his engine and its 260-ton train up to 27mph on the first length of 1 in 60, increasing to 33½ on the following 1 in 70. The conditional stop at Culloden Moor was not called, enabling no 34004 to reach 50mph in the dip before climbing resumed. Having attained 28mph Yeovil was already 6¾ minutes early passing Daviot and was thus eased for the rest of the climb, though making 41mph in the dip before Moy and arriving at Tomatin 5¼ minutes early. Swain made a fine getaway from this stop, covering the 3½ miles to Slochd in a ¼-minute less than schedule at no less than 37½mph. He attained no more than 56mph on the equally steep fall to Carrbridge but still arrived 2¾ minutes under the seventeen allowed. However, the nine minutes to Aviemore was exceeded by 2½ minutes due to the pwr in the approaches.

With the load now made up to twelve coaches of 375 tons gross, Yeovil kept time to the calls at Kingussie and Newtonmore, and then cut the eighteen minute allowance to the stop at Dalwhinnie, up the first 10¼ miles of the climb to Druimuachdar, by two minutes. Most of the rest of the route to the summit is at 1 in 80, no 34004 completing it in 10½ minutes, reaching 47½ mph on the slightly easier grades north of Balsporran and topping the climb at 44mph. With speed reduced to 41mph at Dalnaspidal to take up the single line tablet, and despite easy running thereafter, the engine was still 4½ minutes under schedule into Blair Athol. The ‘B1’ by comparison had kept none of the intermediate times, managed just 26mph at Druimuachdar and despite posting a speed of 61mph at Dalanraoch shaved just a ¼-minute off the 35 allowed from Dalwhinnie.

No 34004 again kept time to the second to Pitlochry and despite another pwr to 26mph at Guay made an untroubled and unhurried run into Perth – other than a spot of joie de vivre once south of Stanley Junction by reaching 70mph at Luncarty - cutting another 4¾ minutes off the schedule. Allen estimates the net gain at 21 minutes.

The success of the ‘West Country’ over the Highland Line later led to suggestions a number should be posted north to work it. With reason perhaps nothing came of this mainly because, being so far away from ‘home’, any problems arising with its unique features might be difficult to solve. As it was ‘Black Fives’ continued to hold sway until the end of steam in the Highlands though double-heading of loads in excess of 260 to 280 tons – depending on the weather? - had to be resorted to.

SR West Country Class 34004. Test trips all over at Perth loco, on the last trip from Inverness. 1948 Locomotive Exchange Trials.
Copyright Russell Coffin collection

The first part of Yeovil’s return trip to London proved interesting and involved double-heading the ‘Day Perth’. This was a load of seventeen bogies but with an apparent shortage of engines the crew was asked if they’d be prepared to work the train on their own. Having agreed to, at the last moment signals were put back to ‘danger’ and a very shabby ‘Black 5’ was coupled ahead for the initial phase, to Stirling. No 34004 took the train on to Carlisle alone where a ‘Scot’, which would work the service through to Euston, coupled up ’inside’ as Yeovil was due to come off at Crewe. The next day no 34004 was to double-head the Aberdeen ‘Night Sleeper’, due away from Crewe at 3.0am though that ran in 90 minutes late. This had repercussions at Euston because Yeovil came off the train at Willesden in view of the late-running to make a direct journey to Nine Elms. Unbeknown to the crew a small reception committee had convened at the terminus to welcome them back. So then on to the dessert, or it may be the aroma given off by a particularly ripe fromage. What sort of useful data did these six months of testing produce? In summing up Allen makes one blunt statement, that to a passenger the first priority in the operation of passenger trains – other than safety of course – is to run as implied in the timetable, to provide ‘exactly what it says on the tin’. In that light he makes no apology for looking favourably on crews who made timekeeping their main objective. With some dismay, not to say exasperation, he reiterates the lack of a common standard of driving, citing in particular the extraordinary variation between the two drivers from the LMR, Byford with the ‘Duchess’ and Brooker on the ‘Scot’. Why were the standards produced by these two men from the same depot poles apart, the one consistently failing to make up lost time, the other achieving something quite out of the ordinary?

As with driving so with conducting. As already stated the fact the Southern provided just two pilotmen brought about a consistency quite lacking in other Regions. He is also adamant that the performances put up by Southern engines “behind which it was a joy to travel”, was the most uniform. CJA also cites some exceptional power outputs on the part of the Bulleids – and, indeed, with the ‘Royal Scot’ - such as that magnificent southbound assault on Shap by Belgian Marine, and the work by the ‘light pacific’ on the ex-GCR route. If nothing else these confirm the tremendous steaming capability of the Bulleid boiler. One point noted with approval by E S Cox in his writings concerned the small drop in pressure between boiler and valve chest with a fully open regulator on the few occasions that actually occurred. (Equally, he disapproves of the majority of the work being done with part-open regulator, and not just by the Southern engines.)

There was a downside of course. Even before the test results were published anecdotal evidence suggested the Bulleids were rather heavier on coal and water than others on trial. Bert Hooker himself confirms the point though at no time does he feel the fire gets away from him, that keeping it fed is unduly arduous. As he states on several occasions he preferred to fire carefully placed half-shovelsful rather than scatter full ones. But then such heavy consumption must be set against the exceptional power outputs in such feats as outlined above. Also on the debit side were reports of much black smoke at times of coasting or running under light steam, an indication a proportion of the fuel went straight up the chimney unburnt. Had the engines been using soft and friable Kent coal that would not have surprised. But the hard Yorkshire coals fired throughout the tests, though obviously leaving some dust behind when being broken up, should not have produced an excessive display of smoke even with the engine working really hard though that also occurred, especially on the Highand Line. Was the fact the Bulleids were not fitted with dampers a contributory factor here?

At the time Allen published his first book the test results had not been publicised and it was not until 1950 his slim follow-up appeared with these details summarised and discussed. The first surprise he notes is the lack of detailed information displayed in the Report for each trip. For example, there is nothing about speeds attained and no point-to-point timings beyond the total sum for the journey, including stops. Signal checks and temporary speed restrictions were noted but there is nothing about the estimated time cost of these checks which, of course, can vary from a dead stand to, say, 40mph.

CJA cites in particular the journey of Belgian Marine when it trailed the desultory special train from Crewe to Leighton Buzzard. The engine is shown as having lost 27 minutes in running and to have experienced eighteen checks and two unbooked stops. Not one estimate appears in the record of how much time these cost or, worse, is any reason for the poor running noted, that is the seeming incompetence or lack of initiative of the LMR operators rather than, as implied, that of the offending engine and its crew. Allen’s record shows 27 checks, three of them between 2 and 5mph, a stop on Camden Bank of such duration as to cause the train to take 11½ minutes over the last mile into the terminus, and several other long stretches taken at a crawl approaching a signal at ‘danger’. One telling feature of this dreadful journey is that the weight of coal consumed went up from 3.61lbs per drawbar-horsepower-hour when working the same train two days previously, to 3.86lbs. Yet nothing in the accompanying statistics indicates why. Were the LMR motive power group at work here again?

The second strong point in this regard is the lack of any point-to-point timings. Why? That tells most strongly against crews who attempted to run to the working timetable so far as was possible but covered the sins of those that didn’t. In that respect Byford on the ‘Duchess’ was a particular winner, for he had a habit of dropping time uphill but on those occasions he attempted to gain it back, not always successfully, resorted to running freely - sometimes too freely - downhill. Had the crews been given instructions to run as closely to intermediate times as they could, and particularly making up lost time, and had the cost of delays in running as well as recording of the actual passing times at intermediate timing points been made, the overall results would surely have provided a truer reflection of performance and therefore been of some value.

The third and a most surprising omission from the official report is the estimated weight of passengers and luggage. It seems utterly beyond comprehension that tare weights were calculated so meticulously, down to two decimal places of tonnage at times, yet no attention is paid to train complements. It is as though the dynamometer car staff simply couldn’t be bothered to get out of their seats and do a headcount, or take a glimpse at the volume of luggage on racks and in vans, even if it were only a simple estimate. Almost every run of the ‘Atlantic Coast Express’ for example, irrespective of motive power, as well the Leeds-Kings Cross train of Queen’s Westminster Rifleman were so full the load must have added 10% or more to the tare weight. If the ‘amateur’ timers could calculate a reasonable figure for the gross weight of the train it was surely within the capability of the professionals conducting the trials to do so.

These seem to me to be serious omissions in a trial supposedly needed to decide the future design and construction of British Railways steam engines and, frankly, lends weight to the argument that, at base, the future had already been decided. Further, it appears to me the dynamometer car staff really had very little idea of exactly what was required of them, as though they had never received a full and comprehensive brief. Were they working to orders or, worse, was it plain incompetence when it came to knowing how a railway is or should be run, inconceivable surely? This is not to say everything done served no purpose at all but the results would have been more meaningful, and the statistics have provided information more reflective of engines in everyday service if a little more thought had been given to the basis of performance on which the data were gathered and measured. It is common knowledge now that the Bulleid’s did not come out too well from these tests. But I reiterate my view that alone of the crews involved, with the exception of the LMR’s Brooker who matched them, the Southern men were the only ones to whom timekeeping was a principal objective and especially in attempting to regain time lost.

So far as coal consumption is concerned the ‘MN’ and ‘WC’ came bottom in the league table, French Line CGT being the exception in one instance. While consuming 48.02lbs of coal per mile on the Western Region tests, (the ‘King’ ate slightly more), 35019’s per db-hp-hr output was 3.61lbs, better than either the ‘King’ or the ‘Scot’. ‘MN’ coal consumption throughout the tests worked out at 3.60lbs per db-hp-hr. But the variations between best and worst on each Region were quite small and I would suggest the consumption quite in keeping with regard to the work done.

The same applies to the ‘light pacifics’. If one takes the work on the London Midland Region for example, no 34005 is deemed to be using five pounds of coal per mile more than the ‘5MT’, yet the coal fired per db-hp-hr is only .09lbs more. Overall however these engines fared badly, using no less than 4.11lbs of coal per db-hp-hr compared with 3.57lbs by the ‘B1’ and 3.94lbs by the ‘Hall’.

In the record of coal consumed per lb of water evaporated however, the Bulleid boiler showed its worth. The ‘MN’ used one pound of coal to steam 8.45lbs of water, second only to the ‘Duchess’, and the ‘WC’ managed to get 7.94lbs of water steamed per pound of coal. But both classes fared badly when relating the water consumption to the output, again coming at the bottom of the list. The ‘MN’ used 30.43lbs per db-hp-hr while the ‘WC’ needed 32.64lbs. This lends much weight to the theory that weakness in the valve events of the engines meant, in general, the steam produced was not being used as efficiently as it might have been thanks, probably, to running with part-open regulator and generous cut-off position. It is pointless, if fascinating, to wonder how much improvement would have been made on these figures had Bulleid been able to install poppet valves and rotary gear as he originally intended, the necessary material being unavailable in wartime. The very short cut-offs at which the rejuvenated Duke of Gloucester is now capable of running ought to give a clue.

Happily there were positives. The Report noted the exceptionally free-steaming boiler, with the firemen having little trouble maintaining pressure even with the engine working hard, and the minimal drop between boiler and valve chest pressures with the regulator well open. The free-running and steadiness in riding were also noted as was the level of output gained and sustained when required. But maybe the most satisfying from the point of view of those taking part - and here we have to thank Bert Hooker again - was the almost universal and enthusiastic appreciation of the crews who piloted Southern men on ‘foreign’ soil. That is proof, if proof were needed, a Bulleid ‘pacific’ was an engineman’s engine, if not that of the pen pushers and unimaginative accountants – and consistency is the last resort of the unimaginative. However, from conversations with enginemen who knew them the general opinion seems to be that the rebuilding carried out from 1956 removed the ‘sparkle’ which marked out the engines as special. Perhaps they were right, for after controlled tests made with the rebuilt no 35020, Bibby Line, it was concluded “the engine gave one of the most predictable performances of all the locomotives that had been tested under the auspices of British Railways”.

Power outputs place the engines highly. No 35017, for example, registered a pull of 4.1 tons on the drawbar, or 1,260 drawbar horsepower on the climb from Corby Glen to Stoke on 25th May, (equivalent drawbar horsepower of 1,501**.) This was achieved at a speed of 51½mph with 25% cut-off and but 185psi showing in the steam chest. Coming up to London three days later the near-record climb from Grantham to Stoke saw the pull at 5.19 tons and the actual drawbar horsepower reach 1,411, (equivalent dhp 1,659). This again was at 25% cut-off but with 200lbs in the steam chest. These figures were surpassed only by the ‘Duchess’ when lifting its train up the 1 in 100 from Wakefield towards Ardsley. (Belgian Marine was exerting a pull of 6.14 tons, almost the same as the LMR engine, but at a slower pace which showed in the lower horsepower figures.) CJA also makes the point that the magnificent southbound climb to Shap he timed from the train on May 14th had been virtually duplicated by Jack Swain two days previously, an equivalent dhp figure of 1,920 being registered then compared with 1,929 on his run. The cut-off for these ascents was apparently 33% with 215-225psi of steam in the valve chests. Outputs as these levels were also recorded northbound: no other engine came close to them.

For their size the ‘West Country’s performed very well too, Bude registering the highest output of any engine on Wellington bank - express ones included - and more than holding its own over Rattery and Dainton. No 34006 also excelled on the ex-GCR route, and as noted, there set the highest recorded equivalent dhp figure in all the tests at 2,010. In this case, with the cut-off at 27% and the regulator nearly full open, pressure in the steam chest was only 20psi less than the 260 in the boiler in making a speed of 67.8mph up the 1 in 176 between Whetstone and Ashby Magna. (The ‘Duchess’ once registered 2,400dhp on the climb to Honiton tunnel though it was deemed transitory whereas Bude’s was sustained.) Little needs to be said about Yeovil’s Highland exploits beyond stating her power outputs at all stages recorded left the ‘5’ and the ‘B1’ far behind.

A final point on the tests concerned adhesion factors and the extent that slipping affected results. Both Bulleid designs had a bad reputation for picking up their wheels but were so ably handled that bouts of slipping on starting were far less than reportedly appeared normal in everyday service. Both have low adhesion factors, the ‘West Country’ just on ‘the right side’ of the supposed optimum of 4.0 at 4.06, the ‘MN’ much on the wrong side at 3.76. It is worth noting however, that both the engines notionally more powerful than the ‘MN’, the ’Duchess’ and the ‘King’, have factors of 3.75 and the latter in particular is noted for its surefootedness. On Rattery in the down direction, for example, no 35019 was slipping continuously for 2¼ miles. But with sand applied the drawbar pull rose from a maximum of 6.83 tons to 7.44 tons. However, Allen calls into question some of such recordings based on an extraordinary statement in the Report concerning the ‘Duchess’. On May 19th City of Bradford is recorded as slipping almost to a standstill on Dainton. But Allen and another timer were aboard the train that day and though the engine did slip badly on passing Dainton box the minimum there was 18mph, hardly a case of ‘almost a standstill’. One wonders if the professional and amateur recorders were actually on the same train!

One aspect of the largest ‘standard’ designs went very much against past practice. All five classes in the Express Passenger segment of the trials were multi-cylindered. Yet there was only one multi-cylindered ‘standard’ engine produced and that an afterthought to fill the gap left by the destruction of the LMS rebuilt ‘Turbomotive’ no 46202 at Harrow & Wealdstone on 8th October 1952. As a ‘one-off’, still incomplete when the decision was made by BR to go for diesel power, a major and particular flaw in 71000 were never eliminated. Coal consumption proved very heavy, in part because of the tendency of the firebed to deteriorate and fragment due to vibration, but mainly because, as a saving in first cost, the Kylchap exhaust arrangements specified with the Caprotti valve gear was not fitted. With the ability of the engine, as shown in tests at Swindon under the eminent Sam Ell, to work with the cut-off as low as 3%, need of this apparatus to clear the exhaust was essential. As E S Cox commented rather sadly, ‘….this engine could have been a world beater.’ Was he being prophetic? With the skill, time and opportunity as well as, most particularly, the will of the preservationists this grand engine has now fulfilled its potential. It may be noted that other than 71000 and the class ‘9’ freight engines, all the ‘standard’ classes were deemed ‘Mixed-Traffic’.

So what did the Bulleids provide to the new standard range? Not that much! But one was the driving wheel diameter of 6’2” in the larger engines. Cox notes that with a modern front end the ability to reach and maintain 90mph with wheels of this dimension had been decisively demonstrated by the Southern engines - but not during the trials be it said! Note, however, he does not relate this to piston speed. With a stroke of 28” that of the ‘Britannias’, the ‘Clans’, the ‘5MT’s and the sole ‘8P’ would be rather higher at any given rate than a Bulleid with its 24” stroke. Cox also notes the work of the ‘7P/5F’ ‘West Country’ class had been impressive enough to show there was scope for developing an engine of 18½-ton axleloading in this category - though ‘without the poor efficiency and mechanical complexities’ of the Bulleids. The outcome, the 6MT ‘Clan’ class, turned out to be an engine of so little advance on the very successful class ‘5MT’ orders subsequent to the first ten were all cancelled.

The ‘MN’ also provided the three-part ashpan to the ‘pacific’ designs and the layout of the frames. In the Bulleid engines the hornguides are welded to the frames and the frames mounted over the centre line of the axleboxes. In this way offset loading of the spring hangers is avoided. However, in the other designs, including the ‘9F’ with its wide firebox, the usual arrangement of offsetting the hornguides was resorted to because of the need to get the frames between the backs of the driving wheels and the maximum width of the narrow firebox. (The original layout for a heavy freight engine had a 2-8-2 wheel arrangement with a ’Britannia’ boiler. The trailing pony was eventually dispensed with and the coupled wheelbase went to five axles to maximize adhesion. A unique boiler design thus became necessary.) The Bulleid method of shrinking tyres to wheel rims was also adopted, there being no fixtures at all between the two. The pony truck, which gave its engine crews such a smooth ride, was also adopted for the ‘pacifics’ without change other than substituting laminated springs for coils. At the time breakage of coil springs was common, if much reduced though not entirely eliminated in later years by putting in stronger springs.

So far as the Southern is concerned many ‘standard’ engines found themselves roaming the Region from Cornwall to Kent, in particular 4-6-0s of both the 73xxx and 75xxx varieties and the 4MT 2-6-0s in the 76xxx series. Class ‘4MT’ and ‘2MT’ tank engines - including incidentally the Ivatt version of the latter - also appeared in numbers though none of these classes had any specifically ‘Bulleid’ features. It may be noted that for some months in 1948 two LMR Fairburn 4MT 2-6-4 tank engines, nos 42198/9, were tested over all three Sections of the Southern Region independently of the main trial. As a result Brighton began to turn out members of this class for local use, and latterly started construction alongside them an order for forty-four of the BR ‘standard’ class ‘4MT’ 2-6-4T engines, nos 80010-53, to be divided between the Southern, Scottish, North Eastern and London Midland Regions. These appeared in 1951. In time all the Brighton-built Fairburns were transferred north of the Thames, mainly to Scotland, the Southern receiving ‘standards’ in return. (Brighton completed 130 Standard 4MTT engines between 1951 and 1956.)

‘Britannia’s featured regularly only on the Eastern Division, nos 70004 and 70014 being allocated to Stewart’s Lane for a period in the 1950s, perhaps simply to ‘show the flag’ on Continental Boat trains. The many Bulleid ‘pacifics’ in traffic made it unnecessary to allocate more. But I wonder how many ‘light pacifics’ might have been transferred to the Great Eastern line had ‘BB’ Sir Archibald Sinclair’s brief foray out of Liverpool Street in 1949 been followed up. Or was that engine’s visit responsible for the posting of the early ‘Britannia’s to Stratford and Norwich with a revolutionary effect on the timetable? Whatever the case Richard Hardy writes, “How we enjoyed that machine!” (He was, of course, to have closer acquaintance with the Bulleids for almost 2½ years following his appointment as shedmaster at 73A in August 1952.)

On its final day on the GER line Sir Archibald was set to haul a thirteen-coach train from Norwich to London, a train that included the General Manager’s saloon with the formidable Leslie Preston Parker – incidentally a great supporter of the ‘Britannia’s from the design stage onward - among its passengers. All was set for a good run when, as Hardy puts it, “….. we climbed the 1 in 84 to Trowse Upper as if it were 1 in 300”. But only a few miles out of Norwich the steam pipe supplying the reverser broke, meaning the engine could neither notch up nor reverse. Nothing daunted, having advised the managerial members in the saloon what had happened, the train continued its journey. But as Richard Hardy asks, what other engine could have run the 107 miles to London in full gear with 60lbs or less in the steam chest and still kept time on a demanding schedule with a thirteen coach train? (But an awful lot of coal was consumed!)

Perhaps his words following sum up best these remarkable engines whose innovative design and characteristics were too much to bear for the cautious inhabitants of 222 Marylebone Road: “The Bulleid Pacifics, wayward, difficult, brilliant, fascinating, had that very human trait of rising to the occasion”. And those in preservation still do. Perhaps we should finish then by raising a glass of a favourite liqueur in a toast to a unique engineer and his masterpieces.

*It is quite as likely the challenge was made by, or at least emanated from the LNER Chairman, William Whitelaw, to his opposite number on the GWR, Sir Felix Pole who was never afraid to fight in the defence of his company.

**Horsepower is the rate at which work is done, 1 hp being 550 foot-pounds per second. (I have no idea how to translate that into non-imperial speak!) Drawbar Horsepower is the actual hp exerted by the engine between tender and train at specific points, being a measure of the power required to overcome the resistance of the train to motion and, where climbing, the downward pull of gravity. In locomotive tests drawbar hp figures relate to output on level track and calculations using well known and proven formulae have thus to be made to add to those figures the additional power necessary for the engine to move itself and its tender and train against the pull of gravity. The total figure is then expressed as ‘Equivalent Drawbar Horsepower’.

Bibliography

The Locomotive Exchanges, Cecil J Allen, Ian Allan Ltd., 1949.
New Light on the Locomotive Exchanges, Cecil J Allen Ltd., 1950.
Nine Elms Engineman, A E ‘Bert’ Hooker, Bradford Barton. (Undated).
Bert Hooker, Legendary Engineman. A E Hooker, Oxford Publishing Company, 1994.
Two Million Miles of Train Travel, Cecil J Allen, Ian Allan Ltd., 1965.
Bulleid, Last Giant of Steam, Sean Day-Lewis, George Allen & Unwin, 1964.
British Railways Standard Steam Locomotives, E S Cox, Ian Allan Ltd., 1966.
British Pacific Locomotives, Cecil J Allen, Ian Allan Ltd., 1962.
Locomotives Illustrated no 7, ‘King Arthurs’, Derek Cross, Ian Allan Ltd., 1976.
Locomotives Illustrated no 12, Bulleid Merchant Navies, Derek Cross, Ian Allan Ltd., 1977.
Locomotives Illustrated no 28, Bulleid Light Pacifics, Derek Cross, Ian Allan Ltd, 1982.
Loco Profile: Royal Scots, Profile Publications Ltd., 1971.
Steam in the Blood, R H N Hardy, Ian Allan Ltd., 1971.
Railways in the Blood, R H N Hardy, Ian Allan Ltd., 1985.
Steam was my Calling, E S Beavor, Ian Allan Ltd., 1974.
Mendips Engineman, P W Smith, Oxford Publishing Company, 1972.
Southern Region Steam Album, S C Nash, Ian Allan Ltd., 1974.
The Changing Southern Scene 1948-1981, Michael Baker, Ian Allan Ltd., 1981.
The Somerset & Dorset in the Fifties, vol 1, 1950-1954, Ivo Peters, Oxford Publishing Company, 1980.
Red for Danger, L T C Rolt, 3rd ed., David & Charles, 1976.
British Railways Pre-Grouping Atlas and Gazetteer, Ian Allan Ltd.,
Track Atlas of Great Britain, TRACKmaps, 2009.
Gradient of the British Rail Line Railways, Ian Allan Ltd, 2016.
Various editions of Ian Allan Ltd ABC’s, 1942 onwards.

Copyright 2018 © Jeremy Clarke - reproduced here with permission