Chapter VI: The Atmospheric System (1)

A.D. 1840--1848. ÆTATIS 35--43.

PRELIMINARY OBSERVATIONS--THE SOUTH DEVON RAILWAY--DESCRIPTION OF
THE ATMOSPHERIC SYSTEM--HISTORY OF ITS INTRODUCTION PRIOR TO
1844--REPORT BY MR. BRUNEL, RECOMMENDING ITS ADOPTION ON THE SOUTH
DEVON RAILWAY (AUGUST 19, 1844)--EXAMINATION OF THIS REPORT--MR.
BRUNEL’S EVIDENCE BEFORE THE SELECT COMMITTEE ON ATMOSPHERIC
RAILWAYS, 1845--HISTORY OF THE APPLICATION OF THE SYSTEM ON THE
SOUTH DEVON RAILWAY, 1844-1848--REPORT ON STATE OF WORKS (AUGUST
28, 1847)--REPORT ON CAUSES OF FAILURE (AUGUST 19,
1848)--ABANDONMENT OF THE SYSTEM, SEPTEMBER
1848--_NOTE_--COMPARISON OF STATIONARY AND LOCOMOTIVE POWER.

In the year 1844 Mr. Brunel recommended the adoption of the Atmospheric System of propulsion on the South Devon Railway, a line of 52 miles in length, which he was then constructing between Exeter and Plymouth. This system had, under the management of Messrs. Clegg and Samuda, been in operation with success on the Dalkey line for some time before Mr. Brunel adopted their apparatus on the South Devon Railway. After it had been in use on the South Devon for about twelve months, it was abandoned, and the railway worked throughout by locomotives.

* * * * *

It is therefore as important as it is interesting to examine the causes of the failure of the Atmospheric System, and to consider the reasons which induced Mr. Brunel in the first instance to adopt it, and afterwards to recommend its abandonment.

* * * * *

Up to about the year 1843, the cost of railways, which was in a great measure due to the conditions imposed by the limited capabilities of the locomotive, had prevented their construction, except in cases where they would secure a large traffic, and at the same time traverse what was then considered a practicable country.

A curvature of one mile radius was regarded as the maximum generally admissible on a line where high speeds were aimed at, and auxiliary locomotives were required to work heavy gradients.

Nevertheless, by the growing wants of the public and the growing boldness of engineers, the railway system was gradually being forced into districts hitherto regarded as unsuitable for it; and no country was held to be impracticable where the gradients could be surmounted by the inconvenient and costly expedient of auxiliary power.

* * * * *

The south of Devon had for several years demanded railway accommodation, and at the period now under review, Mr. Brunel projected what was called the coast line. This line, while it best accommodated the population of the district, passed through a very difficult country. If it was to be constructed at a moderate cost, curves of a quarter of a mile radius had to be admitted; and above 30 miles of its entire length traversed a district involving the adoption of gradients steeper than had been elsewhere used for such considerable distances. The Act for this railway was obtained in the Session of 1844.

* * * * *

The South Devon Railway, on leaving Exeter, crosses the flat country on the right bank of the river Exe, as far as Starcross, a village nearly opposite to Exmouth. From this point it runs down to the coast and along the sea-shore, by Dawlish, to Teignmouth; being protected by a sea-wall for the greater part of the distance, and passing through several headlands by short tunnels. Beyond Teignmouth it follows the left bank of the river Teign, which it crosses a short distance before reaching the station at Newton Abbott. The portion of the line from Exeter to Newton--21½ miles in length--is very nearly level, the steep inclines for which the railway is noted being west of Newton. Between Newton and Totness, for the first mile and a half the line is almost level, and in the next two miles it rises 200 feet, with gradients of 1 in 100, 1 in 60, and nearly a mile of 1 in 43. At the summit is a short tunnel; and thence the line descends 170 feet in a mile and three-quarters, with gradients of 1 in 40 and 1 in 43 for about three-quarters of a mile, and gradients of 1 in 57 and 1 in 88 for the rest of the incline. It then runs with more moderate gradients and about a mile and a half of level line to Totness.

From the valley of the Dart at Totness the line rises at once by a rapid ascent of 350 feet in four miles and a half, with gradients varying from 1 in 48 to 1 in 90, more than a mile and a half averaging 1 in 50. Thence it runs, with easy up and down gradients, for a distance of 12 miles along the skirts of Dartmoor, crossing by lofty viaducts the deep valleys which penetrate the moor. It then descends to Plympton, in the valley of the Plym, falling 273 feet in a little more than two miles, with a gradient of 1 in 42½. From Plympton the line for two miles is level, and then rises on an incline of 1 in 80 for a mile and a half, and descends by a similar gradient into the Plymouth station.

The main characteristics of the railway are that, while it traverses a very heavy country, its principal changes of level are concentrated into four long and steep inclines. These four inclines were intended to be worked by auxiliary power.

Hitherto on gradients of unusual steepness a stationary engine with rope traction had been generally regarded as the only available expedient; but the special difficulties by which this system was encumbered rendered it unsuitable for high-speed passenger traffic, and practically inapplicable to an extended line. It had, however, been very successfully employed by Mr. Robert Stephenson on the Blackwall Railway, a line of about 3¾ miles in length.

Messrs. Clegg and Samuda, the projectors of the Atmospheric System, which was another mode of using stationary power, had, previously to this period, laboured to attract the attention and win the favourable opinion of engineers and the general public.

* * * * *

It is desirable, before proceeding further, to give a brief description of this system of traction, upon the merits of which distinguished engineers entertained widely different opinions.

Between the two rails of the line of way was laid a cast-iron tube, which on the Croydon and Dalkey railways and the completed or level portion of the South Devon Railway was fifteen inches in diameter. On the inclines it was proposed to use a twenty-two inch tube.

At intervals of about three miles along the line were erected stationary engines, working large air-pumps, by means of which air could be exhausted from the tube, and a partial vacuum created within it. A close-fitting piston was placed in one end of the tube, and the air being exhausted from it, the pressure of the external air on the surface of the piston which was towards the open end of the tube forced the piston through the tube towards the end where the air-pumps were working; so that if the piston were connected with a carriage running on the rails, it would draw the carriage with it. The connection between the piston and the carriage was arranged by Messrs. Clegg and Samuda in the following way:[62] Along the top of the tube was a slit about 2½ inches wide; this slit was closed by a long flap of leather, which was strengthened with iron plates, and secured to the tube at one side of the slit. One edge of the leather thus formed a continuous hinge; the other edge, where it closed on the tube, was sealed with a composition of grease, to render it air-tight. This flap was known by the name of the longitudinal valve.

When the valve was closed, the air could be exhausted from the tube in front of the piston, and a partial vacuum formed. Behind the piston, the air being at atmospheric pressure both within and without the tube, there was no objection to opening the longitudinal valve; and a bar, extending downwards from the under side of the carriage, entered the slit obliquely under the opened valve, and was connected to the rear end of a frame about ten feet long, the front end of which carried the piston. To allow the bar to pass along the slit, the valve was opened on its hinge, being pressed upwards by a series of wheels carried by the moving piston-frame inside the tube. The valve closed again after the passage of the train; and the tube was ready to be exhausted in preparation for the passage of the next train.[63]

The Atmospheric System was first tried in 1840. An experimental tube was laid down at Wormwood Scrubs on part of the short line now incorporated into the West London Railway, and then known by the title of the Bristol, Birmingham, and Thames Junction Railway. Its working was the subject of eager discussion among engineers.

In 1842 Sir Frederic Smith, R.E., and Professor Barlow, under an order from the Board of Trade, reported so favourably on the system with reference to the proposal for its application on the Dalkey branch of the Dublin and Kingstown Railway, that it was adopted there. In 1843 Mr. (afterwards Sir William) Cubitt determined to employ it on the Croydon Railway; and about the same time Mr. Robert Stephenson was desired by the Directors of the Chester and Holyhead Railway to report on the propriety of introducing it on that line.

Mr. Stephenson’s report was based on a series of experiments on the working of the system at Dalkey. The view he took was adverse to its adoption, not only on the Chester and Holyhead, but on almost every railway whatsoever; and this on the ground that, though it was quite capable of being developed into a practical working system, yet on lines with ordinary gradients the atmospheric traction must be considerably more costly than locomotive traction, and on steep gradients than rope traction; in other words, that, as a mechanical appliance it was, though practicable, not economical.

Mr. Stephenson’s report had no sooner appeared than the correctness of his conclusions was disputed on his giving evidence before a Committee of the House of Commons, in the spring of 1844, on the Croydon and Epsom Railway Bill.

* * * * *

Mr. Brunel also was summoned as a witness. Previously to this time he had taken a great interest in the various attempts which had been made to introduce the Atmospheric System, and he had himself conducted experiments at Wormwood Scrubs and at Dalkey. As early as July 1840 he had considered its applicability to the Box Tunnel incline on the Great Western Railway. He had also considered it in reference to various projected lines; and in 1843 he recommended it for adoption in a long tunnel on one of the steep inclines of the Genoa and Turin Railway, the success of the system being then (he wrote) sufficient to justify its use on a part of the line protected from weather. It was not, however, applied on this railway.[64]

Mr. Brunel’s views at this time are indicated in the following letter:--

April 8, 1844.

Any part I have taken in examining into the system has been purely
from the desire which I always feel to forward good inventions; and
when I have formed a decided opinion, no fear of the consequences
ever prevents my expressing it. My great anxiety, however, is to
see a line of railway and all its appurtenances made expressly for
the Atmospheric System, and worked accordingly; until this is done
the results will be comparatively unsatisfactory.

Although unwilling to express general opinions, Mr. Brunel spoke strongly before the Croydon and Epsom Railway Committee in favour of the advantages of the Atmospheric System under certain circumstances, and approved of its use on that line.

A few months later Mr. Brunel recommended the Directors of the South Devon Railway Company to adopt the Atmospheric System, and they resolved to act on his advice.

His report was as follows:--

August 19, 1844.

I have given much consideration to the question referred to me by
you at your last meeting--namely, that of the advantage of the
application of the Atmospheric System to the South Devon Railway.

The question is not new to me, as I have foreseen the possibility
of its arising, and have frequently considered it.

I shall assume, and I am not aware that it is disputed by anybody,
that stationary power, if freed from the weight and friction of any
medium of communication, such as a rope, must be cheaper, is more
under command, and is susceptible of producing much higher speeds
than locomotive power; and when it is considered that for high
speeds, such as sixty miles per hour, the locomotive engine with
its tender cannot weigh much less than half of the gross weight of
the train, the advantage and economy of dispensing with the
necessity of putting this great weight also in motion will be
evident.

I must assume also that as a means of applying stationary power the
Atmospheric System has been successful, and that, unless where
under some very peculiar circumstances it is inapplicable, it is a
good economical mode of applying stationary power.

I am aware that this opinion is directly opposed to that of Mr.
Robert Stephenson, who has written and published an elaborate
statement of experiments and calculations founded upon them, the
results of which support his opinion.

It does not seem to me that we can obtain the minute data required
for the mathematical investigation of such a question, and that
such calculations, dependent as they are upon an unattained
precision in experiments, are as likely to lead you very far from
the truth as not.

By the same mode M. Mallet and other French engineers have proved
the success of the system; and by the same mode of investigation
Dr. Lardner arrived at all those results regarding steam
navigation and the speed to be attained on railways, which have
since proved so erroneous.

Experience has led me to prefer what some may consider a more
superficial, but what I should call a more general and broader
view, and more capable of embracing all the conditions of the
question--a practical view.

Having considered the subject for several years past, I have
cautiously, and without any cause for a favourable bias, formed an
opinion which subsequent experiments at Dalkey have fully proved to
be correct; viz. that the mere mechanical difficulties can be
overcome, and that the full effect of the partial vacuum produced
by an air-pump can be communicated, without any loss or friction
worth taking into consideration, to a piston attached to the train.

In this point of view the experiment at Dalkey has entirely
succeeded. A system of machinery which even at the first attempt
works without interruption and constantly for many months, may be
considered practically to be free from any mechanical objection.

No locomotive line that I have been connected with has been equally
free from accidents.

That which is true for one railway of two miles in length is
equally true for a second or third, although they may be placed the
one at the end of the other; the chances of an accident are only in
the proportion of the number, or in other words, the length, a
proportion which holds equally good with locomotives, except that a
locomotive may be affected by the distance it has previously run,
while a stationary engine and its pipes cannot in like manner be
affected by the previous working of the neighbouring engine and
pipes.

In my opinion the Atmospheric System is, so far as any stationary
power can be, as applicable to a great length of line as it is to a
short one.

Upon all these points I could advance many arguments and many
proofs, but I shall content myself with saying that, as a
professional man, I express a decided opinion that, as a mechanical
contrivance, the Atmospheric apparatus has succeeded perfectly as
an effective means of working trains by stationary power, whether
on long or short lines, at higher velocity and with less chance of
interruption than is now effected by locomotives.

I will now proceed to consider the question of the advantage of its
application to the South Devon Railway.

It will simplify the discussion of the question very much if it is
considered as a comparison between a double line worked by
locomotives in the usual manner, and a single line of railway
worked by stationary power, the only peculiarity of the present
case being that upon four separate portions of the whole 52 miles
stationary assistant power would under any circumstances have been
used, these four inclines forming together one-fifth of the whole
distance.

It is necessary to consider it as a question of a single line on
account of the expense, the cost of the pipe for each line being
about 3,500_l._ per mile.

An addition of 7,000_l._ per mile, or of about 330,000_l._, in the
first construction could not be counterbalanced by any adequate
advantage in the saving in the works on the South Devon Railway,
and probably not by any subsequent economy or advantage in the
working; but the system admits of the working with a single line,
without danger of collision, certainly with less than upon a double
locomotive line. And I believe also that, considering the absence
of most of the causes of accidents, there will even be less
liability of interruption and less delay in the average resulting
from accidents than in the ordinary double locomotive railway.

By the modification of the gradients and by reducing the curves to
1,000 feet radius where any great advantage can be gained by so
doing, and by constructing the cuttings, embankments, tunnels, and
viaducts for a single line, a considerable saving may be effected
in the first cost.

In the permanent way and ballasting, the reduction will be about
one-half. I should propose to make the rails about 52 lbs. weight
and the timber 12 × 6; the quantity of ballast would probably be
rather more than half, but at the present prices of iron and timber
the saving could not be less than 2,500_l._ per mile.

From a careful revision of the works generally, I consider that a
reduction may be effected in the following items, and to the amount
specified in each, viz., ballasting gradients and curves:--

£ s. d.

Reduction in earth work 16,500 0 0

" in length of principal tunnel 14,000 0 0
--------------
30,500 0 0

_Saving by single line._

Earthwork 25,000 0 0

Tunnels 11,500 0 0

Viaducts 15,000 0 0

_Permanent way and ballast._

To allow for sidings, say 50 miles, 2,500_l._ 125,000 0 0
---------------
207,000 0 0
===============

_Per contra._

£
Pipe on 41½ miles[65] 138,500
Increase on inclined planes, 10½ miles 6,500 £ s. d.
145,000 0 0
Engines for the 41½ miles 35,000 0 0
Patent right, say 10,000 0 0
---------------
190,000 0 0
===============
The difference in first cost therefore is 17,000 0 0
To this must however be added the cost of the
locomotive power, with its attendant expenses
of engine-houses, &c., which cannot, I think,
be put at less than 50,000 0 0
--------------
Making a saving of 67,000 0 0

I have not included in the expense of the Atmospheric apparatus
that of the telegraph, because at its present reduced cost of
160_l._ per mile I am convinced its use would repay the outlay in
either case.

It would appear, then, that the line can be constructed and
furnished with the moving power, in working order, on the
Atmospheric System, for something less than the construction only
of the railway fitted for the locomotive power, but without the
engines; and that taking into consideration the cost of locomotive
power, a saving in first outlay may be effected of upwards of
60,000_l._

But it is in the subsequent working that I believe the advantages
will be most sensible.

In the first place, with the gradients and curves of the South
Devon Railway between Newton and Plymouth, a speed of thirty miles
per hour would have been, for locomotives, a high speed, and under
unfavourable circumstances of weather and of load, it would
probably have been found difficult and expensive to have maintained
even this; with the Atmospheric, and with the dimensions of pipes I
have assumed, a speed of forty to fifty miles may certainly be
depended upon, and I have no doubt that from twenty-five to
thirty-five minutes may be saved in the journey.

Secondly, the cost of running a few additional trains so far as the
power is concerned is so small, the plant of engines, the
attendance of engine-men, &c., remaining the same, that it may
almost be neglected in the calculations; so that short trains, or
extra trains with more frequent departures, adapted in every
respect to the varying demands of the public, can be worked at a
very moderate cost. I have no doubt that a considerable
augmentation of the general traffic will be thus effected, by means
which with locomotive engines would be very expensive, and
frequently unattainable, particularly as regards one class of short
trains, whether for passenger or goods, which from the
inconvenience of working them by locomotives are hardly known--I
refer to trains between the intermediate stations.

By many means, which the easy command of a motive power at any
time, at every part of a line, must afford of accommodating the
public, I believe the traffic may be increased.

It appears to me also that the quality of the travelling will be
much improved; that we shall attain greater speed, less noise and
motion, and an absence of the coke dust, which is certainly still a
great nuisance; and an inducement will thus be held out to those
(the majority of travellers) who travel either solely for pleasure,
or at least not from necessity, and who are mainly influenced by
the degree of comfort with which they can go from place to place.

Lastly, the average cost of working the trains will be much less
than by locomotives.

With the gradients of the South Devon Railway, and assuming that
not less than eight trains, including mail and goods trains,
running the whole distance, and certainly one short train running
half the distance, be the least number that would suffice, I think
an annual saving of 8,000_l._ a year in locomotive expenses,
including allowance for depreciation of plant, may very safely be
relied upon.

For all the reasons above quoted, I have no hesitation in taking
upon myself the full and entire responsibility of recommending the
adoption of the Atmospheric System on the South Devon Railway, and
of recommending as a consequence that the line and works should be
constructed for a single line only.[66]

In this report Mr. Brunel rested his recommendation principally on two assumptions, which he held to be indisputable--(1) That stationary power, if freed from incumbrances such as the friction and dead weight of a rope, was superior to locomotive power; and (2) That the Atmospheric System of traction was theoretically a good and economical method of applying stationary power, and that it was also a practical and working system, as had been shown in its first and somewhat crude application at Dalkey.

The superiority of stationary as compared with locomotive power depends on two principles--(a) That a given amount of power may be supplied by a stationary engine at a less cost than if supplied by a locomotive. (b) That the dead weight of a locomotive forms a large proportion of the whole travelling load, and thus inherently involves a proportionate waste of power--a waste which is enhanced by the steepness of the gradients and the speed of the trains.

A detailed examination of these principles is given in the note to this chapter.

It is there shown that at the time referred to stationary power could be obtained for one farthing per horse-power per hour, while locomotive power cost more than one penny per indicated horse-power per hour, the cost of the locomotive power being more than four times that of the stationary power. On a level line at a speed of 60 miles per hour, for each horse-power usefully employed, a locomotive, in consequence of its own dead weight and the friction of its machinery, is obliged to expend more than one and a half horse-power; in this case, therefore, the useful work done costs nearly seven times as much as if it had been performed by a stationary engine. Again, on so moderate an ascent as one in 75, for each horse-power usefully employed, the locomotive has to expend at 40 miles per hour more than two horse-power, and at 60 miles per hour three horse-power; so that the useful work done costs in the one case nearly ten times, in the other thirteen times, as much as if it had been performed by a stationary engine.

This great advantage of stationary over locomotive power was a sufficient justification for introducing a system which promised to realise it to any considerable extent; although many difficulties might have to be encountered.

As has been already stated, Mr. Brunel had satisfied himself that the Atmospheric System was theoretically economical, and that its trial at Dalkey had shown that it was practically free from mechanical objections. Mr. Stephenson, indeed, had admitted that the mechanical details had been brought to a remarkable degree of perfection; but this admission was not any qualification of the radical difference of opinion which existed between him and Mr. Brunel, to which Mr. Brunel drew attention in his report of August 1844.

The subsequent abandonment of the Atmospheric System led many to believe that Mr. Brunel had been rash in rejecting the detailed investigations and conclusions of Mr. Stephenson, and that the adverse conclusions which Mr. Brunel had refused to entertain were subsequently established.

But, in fact, the failure of the Atmospheric System was due to failure in some of its mechanical details, and was not due to those inherent defects in its principle which Mr. Stephenson in his report considered that his experimental data had established.

Mr. Brunel’s opinions were again brought prominently into notice in the evidence given by him before a Select Committee of the House of Commons, which was appointed in the Session of 1845, in consequence of the number of projected lines which it was proposed to work by the Atmospheric System.

The following letter, written by Mr. Brunel to one of the members of the Committee, explains what he considered to be his position at this time:--

April 3, 1845.

I am summoned to attend your Committee on Friday, and as it is
known that I have expressed opinions favourable to the Atmospheric
System, and that I am actually applying it upon a line of some
length, it would be considered an absurd affectation, and would,
moreover, be useless to attempt, to avoid giving evidence when so
called upon; but I am a most unwilling witness. I think it rather
hard upon a professional man, who wishes to be cautious and
prudent, that he should be called upon to express general opinions,
which, if written even in the most studied and careful language,
cannot be so worded as to be applicable to every case that may
hereafter arise, or to be proof against the unfair and unscrupulous
attack of the paid writers on these controversies. I mention these
difficulties, which I feel that you may make some allowances for,
if my feelings should appear in my evidence, or if that evidence
should appear to fall very short of the opinions I am known to
entertain, and which I must entertain to induce me to apply the
system extensively, as I am doing. I find it difficult to define
the points upon which it would be desirable to examine engineering
witnesses, and I really believe that, entering freshly upon the
subject and feeling as one of the public, you are more likely to
elicit the useful points than one who, like myself, has been
turning his whole attention (lately, at least) solely to the
mechanical construction. However, I enclose a copy of a letter I
addressed to a party interested in the patent, which refers to my
opinions on the several points--opinions, however, expressed
without that caution to which I referred as so necessary.

(_Enclosure._)

March 31, 1845.

I object very much to giving evidence upon the abstract point of
the applicability of a particular system, and thus furnishing
general opinions which others are to apply as they may choose to
particular cases; and if I could, I would refuse to give evidence
at all before the present Committee, whatever might be the
consequence to the promoters of the Atmospheric System.
Circumstances, however, render such a refusal impossible; but I am
equally anxious not to be drawn into becoming an advocate of a
system.

When I gave evidence last year,[67] although it was then very much
against my inclination, it was in support of a particular case; and
it was only incidentally that my opinion was advanced as to a
system. The evidence is now avowedly sought in support of a system,
and I do not, as I before stated, intend to become an advocate of
this or any other system. I mention these my views to prevent
disappointment. If the following facts and opinions are likely to
be of use to the Committee, I can give evidence on them.

I made experiments upon the portion of railway laid on Wormwood
Scrubs. These experiments were made for my own private
satisfaction, and not made public in any way.

They satisfied me of the mechanical practicability of the system.

In 1843 and 1844 I made several experiments upon the Dalkey
railway.

The result of my observations and of those experiments is an
opinion that the mechanical difficulties attending the application
of this system may be overcome, and the whole as a machine made to
work in a very perfect manner; that is, as a mechanical power for
locomotion it will generally, but not in every case, be more
economical than what is strictly called locomotive power, that high
speeds may be more easily attained, that, from the absence of the
locomotive engine, the rails may be constructed and maintained in
more perfect order, and as a consequence the carriages may be
constructed and worked in a more perfect manner, and so as to run
more smoothly, and that in all respects the travelling may be
rendered more rapid, more luxurious, and more safe. As regards the
last, viz. safety, collisions may be rendered altogether
impossible, or most remotely possible; while all other sources of
danger, now very small, may be almost entirely removed by the
increased perfection of the rails.

As regards first cost, a single line may be made to answer all the
purposes of a double locomotive line for most railways, except main
lines in immediate connection with the metropolis, or forming trunk
lines for others with important branches not under the same
control; and a single Atmospheric line will generally cost as
little, often less, including the working power, than a double
locomotive line without the engines. I am now constructing a line
of 52 miles in length entirely for the Atmospheric System. I
already see many advantages to be attained in the setting out and
constructing of a railway, if originally designed for this system;
the principal advantages can only be attained, and, above all, the
principal difficulties in the system can only be properly provided
against, where the line is originally designed for the system; the
choice of gradients and curves and levels, the position, and, above
all, the arrangements of stations, will generally be totally
different in the two systems, and the difficulties to be avoided
will equally differ.

It is unnecessary to give any extracts from Mr. Brunel’s evidence,[68] as it only repeats the opinions and calculations embodied in his report of August 1844.

The Committee, while they allowed that experience could alone determine under what circumstances of traffic or of country the preference to the Atmospheric or the Locomotive System should be given, reported very strongly in favour of the general merits of the Atmospheric System.

As soon as it had been decided that the South Devon was to be constructed as an Atmospheric line, the dimensions of the cuttings, embankments, and tunnels were arranged for a single line, except on the long incline west of Totness, and on that east of Plympton. In the approaches to the principal summits the gradients were made somewhat steeper, so as to reduce the excavations.

In December 1844, Mr. Brunel prepared a specification and drawing of a steam-engine and vacuum pump, and a copy was sent to the most eminent engine-builders, together with a letter inviting tenders for six pairs of engines. The letter concluded as follows:--

Any party whose tenders may be accepted, shall, if required,
furnish forthwith a more detailed drawing and specification of the
engines as they propose to furnish them; the specification and
drawing now sent being expressly made very general, in order that
each manufacturer may, so far as is consistent with the general
requisites and conditions, adopt his own methods of construction,
or use any existing patterns.

The economical character of the engines which Mr. Brunel desired to obtain is sufficiently indicated by the requirement that they were to be high-pressure condensing engines, fitted with double-seated expansion valves, and having boilers proved to 100 lbs. per square inch, and guaranteed to work with safety valves loaded to 40 lbs. per square inch.

The tenders accepted were those of Messrs. Boulton and Watt, Messrs. Rennie, and Messrs. Maudslay and Field.

Mr. Brunel left it to the contractors to prepare their designs without any interference on his part, deeming it best to rely on their unfettered judgment.

The task of manufacturing and fitting the cast-iron tube was one of some difficulty, the longitudinal slit allowing of considerable distortion in the casting. This work was undertaken by Mr. George Hennett, by the aid of a set of very effective tools devised by Mr. T. R. Guppy.

The tubes were supplied at the rate of one mile per week, and by the middle of 1846 nearly the whole line was laid to Newton, and the valve was ready to be fixed.

In the autumn of 1846, Mr. Joseph Samuda went to Dawlish, taking with him a staff of assistants trained in the working of the system at Croydon; and every effort was made to advance the completion of the engines and the other parts of the apparatus.

Owing, however, to vexatious delays in the erection of the engine-houses and engines, it was not until the commencement of 1847 that a piston-carriage was able to traverse the first six miles out of Exeter. And, though repeated experimental trains continued to be run, no passengers had been conveyed by Atmospheric trains prior to the general meeting of the shareholders, at the end of August. Mr. Brunel’s report on this occasion was as follows:--

August 27, 1847.

It is a subject of great regret, and to no one more than to myself,
that we have as yet been unable to open any portion of the line to
the public with the Atmospheric apparatus, although a considerable
distance has for some months been in a state to admit of frequent
experiments being made upon it. This delay has arisen principally,
if not entirely, in that part of the whole system which it might
have been expected would have been the least exposed to it--namely,
the construction and completion of the steam-engines.

It is due to Mr. Samuda that I should say that, so far as regards
the mere pipe and valve, and other details which may be said to
constitute the Atmospheric apparatus, we might long since have
commenced. But the engines, although designed without any
interference with their plans, and furnished by the best makers of
the country, and although differing so slightly from the ordinary
construction of steam-engines, have proved sources of continued and
most vexatious delays, both in the unexpected length of time
occupied originally in their erection, and in subsequent correction
of defects in minor parts. While the engines were imperfect, it
would not only have been unwise to have commenced working the line,
even had it been practicable, but the frequent interruptions to the
continuous working of all the engines rendered it impossible to
complete and test the different portions of the Atmospheric
apparatus. There are still some defects to be remedied in one or
two of the engines, and I am using every endeavour, by persuasion
and by every other means in my power, to urge on the manufacturers
in their work of completion. Within the last week or two only have
we been able to work at all continuously between Exeter and
Teignmouth, so as to have the opportunity of trying the different
parts, and getting the various details requisite for actually
working trains tested and brought to sufficient perfection to
ensure efficiency and regularity.

Since the beginning of last week, however, four trains per day have
been run regularly, stopping at the stations, and keeping their
time as if working for traffic. The tube and valve appear in good
order, and the whole has worked well, but the running in this
manner can alone show the deficiencies which may still exist in the
details necessary for stopping, and starting quickly from the
stations, and all the other minor operations incidental to working
the traffic in the ordinary course; and, until all these
arrangements are completed, and the engines in more perfect order,
I think it would be much better to defer at least the substitution
of the Atmospheric for the locomotive working. Trains, in addition
to those now running may perhaps be advantageously worked for the
public, after a further short continuance of the present
practising.

The two engines completing the number to Newton are nearly ready
for trial, and it is to be presumed that, after the experience of
the past, the makers will be enabled to put them at once into an
efficient state.

The delays and difficulties attending the bringing into operation
the Atmospheric System upon this portion of the railway have been
beyond all anticipation, and beyond what any previous experience
would have justified anybody in anticipating. The difficulties have
all been seriously aggravated by the necessity (consequent,
certainly, upon the original delays) of working the line with
locomotives during the construction and completion of the
Atmospheric apparatus. Not only has the constant occupation of the
line interfered with the progress of the work, but it has been
necessary to devise all the arrangements so as to admit stations,
sidings, and line being worked either by locomotive or by
Atmospheric in succession, or even at the same time.

These difficulties, added to those always consequent upon the
introduction of any new system, have been most wearying and
incessant, and I am not surprised that the public and the
proprietors should have been impatient. I trust the ultimate result
will remove any grounds for disappointment.

The stress of personal anxiety and personal fatigue, experienced by Mr. Brunel and by all who were engaged in the work, was very severe, and continued so to the end. Not only was the progress in the completion of the work slow, but in spite of every exertion the results were incessantly marred by unfortunate contingencies which involved further delay, discouragement, and expenditure. Moreover, the reaction which followed the railway mania had set in; calls were ill responded to, and great difficulty was experienced in raising the money requisite for the completion of the line.

Under these circumstances it was resolved, on September 1, not to incur any new expenses in relation to the Atmospheric System beyond Totness, and to limit any expenditure already contracted for, until its working between Exeter and Totness had been fairly tried, except to provide assistant power up the two inclines.

On September 8 the Atmospheric trains began to take their share in the passenger duty of the line, four trains running each way daily; and, except when occasional mishaps caused delay, the new mode of traction was almost universally approved of. The motion of the train, relieved of the impulsive action of the locomotive, was singularly smooth and agreeable; and the passengers were freed from the annoyance of coke dust and the sulphureous smell from the engine chimney.

In other respects the record of progress is but a chequered one, and exhibits, in spite of great and able efforts and brightening intervals of occasional improvement, indications of growing difficulties deepening into ultimate defeat.

In examining the chronicle of events which correspondence and memoranda supply, it is inevitable that references to failure and disaster should be found relatively in far greater abundance than records of success; and this for the simple reason that there was at that time great use in taking note of the unfavourable incidents that occurred, almost none in mentioning successful work.

There is therefore some danger of falling into a mood of unjust depreciation, such as Mr. Brunel had in energetic terms urged the Directors to guard themselves against. He protested against their requiring (as they once intended to do)--

continuous and detailed reports--if true and honest, of course
containing nothing but accounts of mishaps--of a system which (he
says) we are struggling to render perfect. Why, a daily account of
our locomotive mishaps would ruin the locomotive system, if it were
new! I will undertake to say that the mishaps of yesterday or
to-day on the Great Western Railway were as great as that of
Tuesday on the South Devon.

The Atmospheric System was vaguely credited with every delay which a train had experienced in any part of its journey; though, in point of fact, a large proportion of these delays was really chargeable to that part of the journey which was performed with locomotives. It often happened that time thus lost was made up on the Atmospheric part of the line, as is shown by a record of the working, which is still extant. In the week, September 20-25, 1847, it appears that the Atmospheric trains are chargeable with a delay of 28 minutes in all; while delays due to the late arrival of the locomotive trains, amounting in all to 62 minutes, were made up by the extra speed attainable on the Atmospheric part of the line.

Not unfrequently, however, casualties occurred; due indeed to remediable causes, but yet of discouraging aspect in themselves, and deriving additional weight from the manner in which they reacted on the cost of working. Such, for instance, were the frequent and occasionally very serious breakages in essential parts of the pumping-engines. Again, the cupped leathers of the travelling-piston, which made it air-tight, were often destroyed while it was passing the various inlet and outlet valves. Improvements in the valves were introduced to meet this difficulty; but the remedy could not be applied at once throughout the line, and much inconvenience was thus experienced, and a considerable expense incurred. Another source of inconvenience was the water which at times accumulated within the tube.

* * * * *

In many respects the results which had been calculated on were realised, and the new arrangements necessary to the working of the system were successfully brought into operation.

The speed of the trains corresponded fully with the degree of vacuum obtained; that is to say, the train resistances proved to be what had been anticipated.[69]

After the trial of a great variety of air-pump valves, a form was adopted which was found to answer exceedingly well.[70]

In the Atmospheric tube, the system of self-acting inlet and outlet valves, by which the piston was enabled to leave the tube on approaching a station and enter it again on recommencing its journey, were, on the whole, successfully adapted to their duty.

Again, an arrangement for starting the train rapidly from the station, without the help of horses or of locomotives, had been brought practically into operation. This arrangement consisted of a short auxiliary vacuum tube containing a piston which could be connected with the train by means of a tow-rope, and thus draw it along till the piston of the piston carriage entered the main Atmospheric tube. Some accidents at first occurred in using this apparatus, but its defects were after a time removed; and it is hardly to be doubted that the various minor difficulties of the Atmospheric System could soon have been effectually mastered.

* * * * *

It now remains to show how it was that, in spite of much that was hopeful, a vigorously sustained contest ended in defeat, instead of being prolonged into victory.

In working the Atmospheric System on the South Devon Railway grave difficulties were throughout encountered, for which to the last only imperfect remedies could be found.

As regards the power consumed, the engines of each pumping station worked up to about 236 indicated horse-power, and their regulated duty for each train, including the anticipatory pumping, was equivalent to 5·5 minutes of work for every mile of the length of tube they had to exhaust. As the running speed averaged 40 miles per hour, or a mile in 1·5 minutes, the 236 horse-power during the 5·5 minutes of pumping must be regarded as equivalent to 865 horse-power during the actual passage of the train. Now, making full allowance for piston friction and extra friction on curves, for the power expended in getting up speed, for the excess of air-pump resistance due to the changes of temperature experienced by the air under exhaustion, and even for the very large actual amount of friction in the engines employed, the work done should have been represented by an expenditure of 240 horse-power during the passage of the train. If to this is added an allowance for leakage, such as the experiments at Dalkey indicated would be amply sufficient with the longitudinal valve in good condition, it may be said that Mr. Brunel had a right to expect that the duty would be performed with an expenditure of 300 horse-power; whereas it actually required 865 horse-power, or nearly three times the amount.

The explanation of this waste is simple.

Serious and unexpected causes of failure developed themselves in the longitudinal valve, and led to an excessive amount of leakage. A great part of the normal duty of the engines was, as has been stated, to exhaust the tube previous to the entry of the train; and when, owing to leakage, the amount of air to be so pumped out was greatly increased, it became necessary that the operation should be commenced much earlier. There was thus a longer time during which the leakage could take place, and a still greater amount of air to be pumped out. It therefore followed that a large increase of leakage involved waste of power in an enormously increased proportion.

The length of time occupied by the anticipatory pumping was often increased by the difficulty of arranging proper telegraphic communication on the South Devon Railway, and by the absence of it on the Bristol and Exeter Railway. The Electric Telegraph was in its infancy, and though Mr. Brunel had been the first to apply it in connection with railways, namely, between London and Slough on the Great Western Railway in the year 1839, it was not brought into perfect working order on the South Devon Railway till the Atmospheric System was on the point of being abandoned. The result of the defects in the telegraph was that, when a train was late, warning was not received at the several engine-houses;[71] and thus, when this was the case, the pumping-engines which had been started at the right interval of time before the train was due, had to be kept at work for a needlessly long period pumping out the air, which was all the while leaking in through the deteriorated valve. This inefficiency of the telegraphic apparatus would have been of trifling importance but for the defects of the valve. Had the valve been as perfect as it was expected to be, the vacuum, after it had been formed, could have been maintained by an expenditure of power very moderate in comparison with that which was actually required.

As regards the relative cost of the power consumed, it appears that, owing to imperfections in the engines, their expenditure of fuel per indicated horse-power was more than double that of the best of the Cornish pumping-engines, to which they were analogous; while the cost of working was more than three times as great.

The defects in the engines were for the most part such as might have been remedied; and this would have been done, had not the excessive duty imposed on them by the leakage of the longitudinal valve prevented their being stopped for repairs and alterations. In this way the defects of the different parts of the apparatus mutually aggravated each other.

It appears then that, chiefly owing to the defective longitudinal valve, the engines were expending nearly three times the power which they should have done for a given tractive duty, according to previous experience, and the results obtained on the Dalkey line, and that they cost per horse-power at least three times as much as was expected.

The cost of traction, nearly nine times as much as had been calculated on, was between two and a half and three times what it would have been with locomotive power; and this was on a level part of the line, where the comparative advantages of the Atmospheric System were not exhibited as they would have been in the part which had steep gradients.

The imperfections of the longitudinal valve have now to be described. By its condition the Atmospheric System had to stand or fall. With an efficient valve, the defects of the other parts of the apparatus would have been of minor importance, and time would have been given for remedying them. When the leakage became considerable, the defects of the telegraph and the defects of the engines alike assumed a formidable aspect.

The failure of the valve was due, partly to the composition which was used to seal the joint where it opened, and partly to the material of which the valve consisted. The difficulty of obtaining a suitable composition was the first which had to be encountered. On the South Devon a lime soap was eventually found to answer the purpose well. Its surface, however, from exposure to light and air, formed into a hard skin; and to remedy this a thinner and more fluid material, a compound of cod-oil and soap, was laid on to keep it soft. This answered satisfactorily, but it required frequent renewal, as it was apt to be drawn into the tube by the rush of air when the valve was opened. The renewal of the various compositions, and the careful examination and repair which the valve constantly required, was a cause of great anxiety and expense.

But in the materials of the valve lay the source of the more serious difficulty.

The ready affinity of leather for oil and grease, and its suppleness and closeness of grain when saturated with substances of that nature, had long been known and utilised. It had not been anticipated how readily, with air-pressure on one side and a partial vacuum on the other, the oily matters with which the leather was charged would escape from it, especially in the presence of water. Although, while the leather was saturated with water, the valve was remarkably air-tight; when frost supervened the water became frozen, and gave a fatal stubbornness of texture, which rendered the valve incapable of closing properly. Again, in long-continued drought the leather became intractable from its dryness; and the stiffening, whether from frost or drought, rendered it liable to be torn. An immediate application of seal-oil penetrated the leather, and relaxed its stiffness; but the remedy often could not be applied in time, and, moreover, was expensive.

A still more grave defect was all the while becoming matured, and was undermining every hope that a suitable dressing could be discovered, and that the longitudinal valve might be made perfect.

Under the joint action of water in the leather, and of the affinity of iron for tannin--and on the enduring presence of tannin within its texture the consistency of leather depends--a destructive decomposition had long been at work; the oxide, established in the iron plates of the valve by continued contact with damp leather, had been steadily abstracting the tannin; thus the leather had become converted into an ink-stained and comparatively decomposed tissue. Large portions of it became torn, and incurably pervious to air.

It was not until early in June 1848 that Mr. Brunel discovered the condition which the valve had assumed. He then instituted a careful examination throughout the line, and the extent of the disorder was realised.