by Robert Humm
(Originally published in the Journal of The Stephenson Locomotive Society for July-August 2015; there are some extra notes about Charles Rous-Marten and Cecil J Allen at the end of the article.)

3440 City of Truro seen at Bentley Heath Crossing, near Knowle & Dorridge, on 4 Sept 1960, with SLS special from Birmingham to Reading and Swindon. Unknown photographer.

While turning over a quantity of Cecil J Allen’s old papers the writer came across a clip of correspondence that throws some light upon the official Great Western attitude to high speed running. The papers were treated as highly confidential at the time (Allen was enjoined to show them to no-one) but after eighty years we may regard them as of historical interest.

It began with the March 1934 instalment of British Locomotive Practice And Performance in the Railway Magazine, where Allen commented upon the passing of G J Churchward. This prompted a Mr A H Holden of Purley to write at length to Allen about City Of Truro’s alleged 102.3mph dash down Whiteball to Wellington. Mr Holden produced a number of countervailing arguments, suggesting that the actual speed was more in the region of 90-92mph. The outcome was that Allen decided in his July 1934 piece to write at length about the City Of Truro exploit, locomotive performance on the Cheltenham Flyer and the general question of speeds above 90-100mph.

The draft clearly worried Railway Magazine editor W A Willox, who persuaded his boss John A Kay to submit it to C B Collett, Chief Mechanical Engineer of the GWR. In turn Collett consulted the General Manager, Sir James Milne :

Collett to Milne 10th May 1934
Dear Sir James,
I send you a letter dated the 9th instant I have received from Mr Kay, enclosing a copy of Mr Willox’s memorandum dated 7th May, and a draft of Mr Allen’s proposed article, from which you will gather that the main object of the article is to cast doubt upon the accuracy of Mr Rous-Marten’s record of the high speed of City Of Truro in May 1904, when it was said that the speed of 102.3 miles per hour was attained.

Neither I, or any of my colleagues are able to express any opinion on this matter, as the only information we have ever had is contained in the records made by Mr Rous-Marten which were published some few years ago.

The question which is continually being raised as to whether a steam locomotive with reciprocating pistons can really attain a speed of over 100 miles per hour can, I think, be readily settled by making a trial under plenty of expert observation.

We believe in the past that that we have many times exceeded this speed, but have never taken any special precaution for making the record, having merely been guided by the ordinary methods of using the stop-watch in conjunction with the mileposts.

Personally, I do not like these high speed tests and, whenever I have been concerned in making them, we have been careful not to expose any members of the public, Post Office, etc, to the unnecessary risk which, in my opinion, is always involved. One has only to review the continuous efforts that are made to set up records in motor-car racing, to realise what a dangerous business it becomes, as soon as the high speeds usually obtained in ordinary practice are passed, and recently making the final arrangements, we carried out some preliminary trials, the results of which, from one cause and another, decided us not to make the final test, which it was then proposed should be made with a mail train.

It is one thing for a racing motorist to set up a record in which, as a rule, no one else’s life is at risk but his own, but it is quite another thing to set up a record on a railway, where the lives of several people are bound to be at stake.

On the other hand, if for any good reason, it is considered necessary or desirable for something of this sort to be done, there would be no difficulty whatever in getting together sufficient staff to voluntarily carry out the tests.

You will no doubt appreciate that the risk with the locomotive is considerably greater than with most other modern types of engines, on account of the difficulty of satisfactorily maintaining proper lubrication at very high speeds of the valve gear and driving mechanism, which is entirely unprotected, it being impossible to encase it in an oil bath like other modern machines and, although we have successfully made high speed runs for short periods in the past, we have had very numerous failures in the course of our experiments in that direction and, as you will know, the failure of oil to any part of the valve gear leads to the complete smashing up of the mechanism in a very few seconds.

It is only in recent times that we have so improved our lubrication arrangements that we are able to run with regularity a high speed service like the Cheltenham Flyer and this, even with the same engines that are doing it today, could not have been carried out under the conditions of only a few years ago.

From this you will gather that we are in a better position now than ever to make a high speed test from a locomotive point of view, and our main concern of recent times has been connected with the track, one of the most difficult problems being that of proper super-elevation on the curves, which needs to be quite abnormal for the extra high speeds which would be equally unsuitable for the ordinary speeds.

I think you will find that the highest speed records of the past were never purposely obtained, but were, as a rule, the result of some misunderstanding.

Please do not think in this matter that I am opposed to to progress or I think it is undesirable for our speeds to be increased, but I have come to the conclusion that for speeds over 100 miles per hour it would be necessary for track, as well as the vehicles, to be specially prepared and maintained and, in our last effort in this direction, you will remember we had great difficulty in finding a suitable place to carry out such a test.

I have set out this matter so fully and think it is only right to do so, as one might be so easily accused of taking unnecessary risks merely for a stunt but, on the other hand, if it is particularly desired that such a test should be made, I should be only too happy to do all I can to carry one out to a successful issue.
Yours faithfully,
C B Collett

Sir James Milne’s response to this Sir Humphrey-like screed has not survived.

Collett to Kay, 4th June 1934
I am obliged for your letter of 30th May enclosing the proof of Mr Allen’s article, which you propose to publish in the July isse of The Railway Magazine.

I have read this through with great interest, and think it is excellent in every way, and have no modification to suggest.

There is, however, one point which I do not think is sufficiently appreciated by some of the advocates of very high speed runs, and that is the great importance of having the outer rail of the curves correctly super-elevated for the speeds. This, on any railway where varying services are run, has, of course, to be compromised, and is never really sufficient for the normal high speeds, and, (what I think is not appreciated) is quite inadequate for the extra high speeds; in this connection, the matter was brought home to me very vividly when we were recently carrying out a preliminary run for a high speed trip which was to have taken place with the mail train, and we found that in running down a grade of 1 in 300 between Badminton and Somerford, where there are several curves, but of a fairly flat radius, that the insufficiency of the super-elevation of the outer rail for the speed caused terrific flange friction, and this grinding against the outer rail absorbed in running at high speed down this grade as much power as pulling a heavy train up one of the steep banks in the West of England and, under these circumstances, the risk of mounting the rail is very serious.

I am enclosing you a diagram setting this matter out but, of course, do not wish it published; you may, however, like to refer to the subject.
The full line speed curve on the chart clearly shows how the grinding of the curves retarded in each case the acceleration which was taking place, as the throttle was not interfered with.

Curiously enough on the morning on which the actual trial was to have taken place, I received the following telegram from the driver of the up 10.30 train from North Road:

“I was stopped at Wrangaton and informed by the signalman, platelayers reported road out of order. I left Wrangaton at 11.12 with the platelayer to the obstruction, which was a very bad buckle in both rails. After straightening as well as possible I pulled the train very slowly in safety. Total delay 33 minutes.”

This on a road which the night before was considered perfect for a high speed test. Naturally we do not want this published either.

Your draft returned herewith.
Yours faithfully,

C B Collett

Collett’s strictures upon lubrication and super-elevation were worked into Allen’s published article. Like Collett, Allen was sceptical of Rous-Marten’s claims and noted that whilst there were a number of authenticated maxima of 92mph nothing so far had exceeded that figure. Perhaps 92mph was a ‘natural’ limit for steam, he mused. Yet within a year or so Allen was reporting the 100mph achieved by Flying Scotsman, the 108mph by Papyrus and the 112½mph by Silver Link. Clearly rail grinding was less of a problem on the East Coast.


Cecil J Allen (1886-1973) was a permanent way engineer employed by the Great Eastern and LNER.  He was also a prolific railway author, speaker and broadcaster. Between 1910 and 1959 he wrote the monthly “British Locomotive Practice And Performance” feature in the Railway Magazine, and later similar series in Trains Illustrated and Railway World.
 A biography of Allen by RH was published in the Railway Magazine for January 2015; we might serialise it in the blog in due course.
Charles Rous-Marten (1844-1908) was an English technical journalist.  At an early age he emigrated to New Zealand and became involved in railway matters. In 1887 he returned to this country to prepare a report for the NZ government on the latest railway technical developments.  Back in New Zealand he prepared similar reports on several other railway systems. He returned to England for a second time in 1893.  He made numerous contributions on locomotives to “The Engineer” and developed the art of train-timing. He originated the monthly “British Locomotive Practice & Performance” column in the Railway Magazine.
 Rous-Marten was held in high esteem in official and railway circles, though there is considerable doubt as to whether he possessed any formal engineering qualification.  His account of City Of Truro‘s apparent record-breaking run has always been controversial.
British Locomotive Practice & Performance” still features in the Railway Magazine, under the abbreviated heading “Practice And Performance”, and is probably a record run in its own right. It is currently manned by Keith Farr and John Heaton – indeed Mr Heaton revisited this very issue as recently as August 2015.