Preview

INTERNATIONAL LABOUR OFFICE

STUDIES AND REPORTS
Series M (Statistics) No. 15

METHODS OF COMPILING
STATISTICS OF RAILWAY
ACCIDENTS

GENEVA
1929
Published in the United Kingdom
For the INTERNATIONAL LABOUR OFFICE (LEAGUE OF NATIONS)

B y P . S. KING & SON, Ltd.
Orchard House, 14 Great Smith Street, Westminster, London, S.W^^;.
<<*>^ BI3UUÍHEU!''

¡Ci ^ -

•' ' ' •

•*J '

1R|!!!Í¿1QO-J

VJJ^

LAa i U i ' '

CONTENTS

Page
INTRODUCTION

CHAPTER
I. — DEMARCATION OF RAILWAY RISK . . . .
The Statistical Units
The Industrial Risk
The Railway Risk Groups
The Employment Risk Groups

3
4
7
13
17

CHAPTER II. — RAILWAY ACCIDENT EXPOSURE
The Average Number Employed or Man-Hours Worked . . . .
The Mileage Rim

23
23
26

CHAPTER I I I . — CLASSIFICATION BY CAUSES
The Classification of Accidents
The Classification of Casualties
Proposed Classification

32
33
38
41

APPENDIX

NOTES ON THE STATISTICS OF RAILWAY
IN D I F F E R E N T COUNTRIES

ACCIDENTS
49

INTRODUCTION

The following Report is the second of the series dealing with
accident statistics among special groups of workers. Coal-mining
accident statistics have been treated in a separate report 1 ; in
the present Report statistics of railway accidents are considered.
Special interest has always been shown in the risks connected
with railway transport. On the one hand, the movement of trains
contains particular dangers ; on the other hand, not only
the workers b u t also passengers and the community are exposed
to railway risks. These facts have led in most countries to the
compilation of special railway accident statistics, in addition to
statistics of compensation of industrial accidents. The present
Report is confined to the industrial risks of railway employees only.
Differences in the methods employed in compiling the statistics
of railway accidents hamper comparisons of accident risk in different countries. I t is the task of this Report to analyse these
differences and to enquire by what means and in which directions
comparability may be attained, since at present incomparability is
so great as to render any international comparison almost futile. As
with coal-mining accident statistics, the main sources of incomparability appear to be variations in the methods of defining reportable
accidents and casualties, of demarcating industrial railway risk, of
calculating exposure, and of classifying the classes by causes.
These points are therefore specially considered in this Report.

INTERNATIONAL LABOUR, OFFICE : Methods of Compiling Statistics of

Coal-Mining Accidents. Studies and Reports, Series N (Statistics), No. 14.
Geneva, 1929. 90 pp.

CHAPTER I
DEMARCATION OF RAILWAY RISK

The notion "railway transport" usually implies the transport
of goods and persons by rail on a road or track specially reserved
for this purpose. Railway transport may be taken therefore not to
include tramways. There are, however, doubtful cases, in which
the railroad is combined with a general road which serves also
for the regular transport of goods and persons between different
places by trains, i.e. by a number of connected carriages driven by
an engine or motor. The risks connected with the running of such
street-trains will differ more or less from the risks of railways strictly
speaking, since in order to reduce the risk involved in running the
train on a public road the speed has to be reduced. The speed, as
well as the intensitj', of the traffic will be, on the other hand,
considerably influenced by the character of the permanent way.
Narrow-gauge railways will not as a rule bear as dense and intensive
a traffic as standard-gauge railways. Where narrow-gauge railways are included in the statistics, therefore, the risk should bo
separately assessed for standard-gauge and narrow-gauge railways.
A second important technical factor which may have an influence
on risk is the mode of traction. It is evident t h a t electric traction
entails certain risks which do not exist in the case of steam traction
and vice versa. Some countries — for instance, the United States
of America — publish accident statistics separately for electric and
steam railways, but very summarily only for the former. Other
countries have hitherto confined their statistics to steam railways,
in view of the small significance of the electric railways. .In
Switzerland, both are given together, also in Great Britain. I t
would seem advisable to give certain kinds of accidents separately
for steam and electric traction. Accidents on cog and rope railways should be separately recorded.

— 4 —

T H E STATISTICAL U N I T S

The risk of railway transport derives its special nature from the
movement of the trains. On the one hand, disturbances in the
regular operation of the railway, due for instance to defects in the
permanent way and in the rolling stock, or to negligence and
mistakes on the part of the persons engaged in working the railway,
are likely to have particularly serious consequences owing to the
rapid movement and weight of the trains; on the other hand,
other strictly personal events occurring during the regular
operation of the railway, e.g. falling off a train in motion, may
have serious results on account of the movement of the trains.
The movement factor therefore enhances the dangers connected
with failures or unforeseen events interfering with the regular
operation of the railway, and consequently gives special importance
to such events. The reporting of disturbances in the operation of
railways, whether they have resulted in injury to persons or not,
has therefore been considered desirable in many countries which
compile special statistics of railway risk. In these statistics interest
has largely centred on this so-called "accident", t h a t is, the event
which has, or might have, caused injury to life, or which has caused
damage (as distinct from the casualty, which refers to the personal case of injury). While the casualty — t h a t is, the fact
of one person being injured by unexpected circumstances — is
naturally the only unit on which compensation statistics are
based, the safety and prevention points of view have led to the
distinction in the special statistics of the two units : accident
and casualty.
The essential feature of risk is evidently the unexpected,
unforeseen, and unpremeditated event. Considering the cases of
injury, the qualification "unexpected" refers to the person injured :
the event which resulted in injury was unforeseen by the person
injured. The case is different, however, if events arejdistinguished
which have or might have resulted in injury or in damage, as is
done in most railway statistics : the qualification "unexpected"
or "unforeseen" can then only be considered in relation to the
regular operation of the railway and refer to an event not forming
a link in the chain of operations regularly planned for the running
of trains or connected therewith. The casualty will always be
connected from the point of view of the injured with an event
unexpected by him : the individually unexpected event. This
event, however, may be, and frequently is, also an event unforeseen

5 —

from the technical point of view of the regular working of the railway, i.e. an event disturbing the operation. This latter kind of
event connected with a disturbance may result in several injuries,
and thus become a statistical unit independent of the casualty.
The term "accident" is applied in British railway statistics
exclusively to the disturbance, irrespective of whether the event
resulted in injury or damage. Wherever such an event cannot
be stated, only casualties are registered. In Bulgarian statistics,
the term "accident" is also confined to disturbances in operation.
In other statistics, however, the term "accident" is generally
taken to indicate the event as distinct from the injury, and covers
not only events which are disturbances of operation, but also the
personal events 1 where no disturbance is involved. In the United
States, for instance, accidents are separately registered in all cases
of injury occurring in connection with the operation or movement
of trains, in addition to those cases where disturbing events occur.
These latter are reported provided they result in injury or damage
of more than 150 dollars. In German statistics the "accidents"
(Betriebsunfälle) registered as separate units are the disturbing
events and personal events which are connected with the movement
of trains. Disturbing events not resulting in injury are registered
if they occur during or as a result of the movement of trains, for
instance derailments and collisions, and certain other disturbances
involving the train in motion, and generally such as result in damage
of more than 1,000 marks.
Most other countries have similar methods : events not resulting
in injury are registered if they are unexpected disturbances, a limit
being usually fixed for the reporting of certain cases by stating
the amount of damage which must result in order to make the event
reportable. This limit varies from country to country, being, for
instance, 1,000 francs in Switzerland, 10 kroner in Norway, 1,000
kronor in Sweden, etc.
To sum up, therefore, the unit "accident"
(Betriebsunfall,
accident) refers in all statistics to events constituting unforeseen
disturbances in the operation of the railway; in most countries,
however, it covers also events which are only personal and have
resulted in injury.
As to the disturbances, the definition of reportable accidents
varies with regard to the minimum amount of damage making an
1
It is interesting to note that the term "personal accident" (accident
individuel) is found in French statistics.

— 6 —

event reportable and with regard to the specific kinds of events
reported. The number of accidents reported therefore varies
greatly from country to country and cannot be generally compared,
though certain categories, such as derailments, collisions, etc.,
may be comparable. Distinction is not even always made between
accidents which have resulted in injury and other accidents.
Though disturbing events are certainly of interest for preventive
purposes, the risk of the persons concerned appears only from
the number of casualties. If all unexpected disturbing events are
reported as separate units, without special tabulation of those
resulting in casualties, the frequency of accidents resulting in injury
cannot be stated, since accidents may result in injury to few or
many persons.
As regards the personal risks, i.e. casualties, the methods of
reporting differ considerably. Even with so-called fatal casualties,
the meaning of "fatal" differs, as it depends largely upon the
method by which the causal relation between the event and death
is established. While in Great Britain all casualties leading to
death up to the date of the Annual Report are classified as fatal,
many countries consider as fatal only casualties resulting in death
within twenty-four hours from the occurrence of the accident,
for instance, Switzerland, Germany, the United States, Austria,
and Norway.
In the United States Railway Accident Report for 1924, a table
was introduced for "subsequent fatalities", i.e. those casualties
which resulted in death after the expiration of twenty-four hours
and which were consequently included among non-fatal casualties
in the regular tables. The effect of the twenty-four-hour limitation
in the regular tables may thus be illustrated as follows :

EMPLOYEES ON DUTY KILLED AND INJURED IN TRAIN AND
TRAIN SERVICE ACCIDENTS
Men killed

First method
Second m e t h o d

Men injured

Number

Rato per mil:; on locomotive miles

Number

Rate per million locomotive miles

1,192
1,339

0.69
0.78

32,174
32,027

18.64
18.55

For explanation of train and train service, see below.

7 —

One hundred and forty-seven persons injured died after
twenty-four hours from the time of the accident and were, therefore, included among the "injured" by the usual method (first
method). If these persons are included among the "killed", the
fatal casualty rate per million locomotive miles rises from 0.69 to
0.78, and the non-fatal casualty rate falls from 18.64 to 18.55.
While, therefore, the non-fatal rate is not considerably modified
by the exclusion of persons dying more than twenty-four hours
after the accident, the fatal rate is 11.3 per cent, higher by the
second method.
The differences, however, in the limits of reportable casualties
are far greater in regard to non-fatal casualties reported : in the
United Kingdom, for instance, casualties which cause the railway
servant to be absent for at least one whole day from his ordinary
work are required to be reported; the limit of incapacity to work
is fourteen days in Germany, Sweden, and Switzerland, twenty
days in France, two weeks in Austria, etc., whilst in the United
States the casualty must have led to three days of incapacity
within the ten days following the accident. In Norway the casualty
must have led to medical treatment in order to be reportable.
Complete incomparabihty thus results for non-fatal railway
casualties.
T H E INDUSTRIAL R I S K

The primary aim of this Report is an analysis of the statistics
with a view to the determination of the railway industrial risk
— i.e. the risk of employment in railway service. Contrary to the
position in most other industries, however, accident risk on railways
is not confined to the workers engaged in railway service. Accidents
may also happen to passengers and other persons. I t is therefore
necessary to distinguish : (1) an industrial, and (2) a non-industrial
or public railway risk. Public safety and safety of employment
in railway work are the two points of interest for special railway
accident statistics. The public risk is of interest mainly in connection with the running of trains ; in other words, it is the risk t o
the public resulting from unforeseen disturbances of the regular
operation of railways. Casualties of passengers connected otherwise with the movement of trains are also usually considered as
railway risks. The industrial risk, on the other hand — the risk
incurred by railway workers during their employment — is made
up of all casualties occurring during their work, whether arising

— 8 —

out of disturbances in the operation of the railway or by or during
the movement of trains, or otherwise.
The scope of the special railway statistics in some countries has
been determined mainly from the point of view of public safety
in regard to the risk peculiar to the running of trains, that is to
say, from the operative point of view. In other countries, however,
the scope of the statistics has been extended so as to cover all risks
of employment of workers engaged in railway working, the industrial
point of view predominating.
The fact that railway accident statistics are not solely compiled
with a view to determining the industrial risks of railway service
has given rise to considerable differences in the demarcation of the
statistics.
(1) The notion of "railway risk" may be conceived in a narrow
sense. According to this, the public safety point of view prevails
and only risks arising from the actual process of running trains
in the narrowest sense are covered. The criterion here is the movement of trains.
(2) Where the industrial risk of the workers is considered of
main interest for the special statistics, the wider conception of
railway transport determines the scope of the statistics. In this
case the criterion may be either :
(a) a technical one, the operation of the railway — i.e. the processes during or by which the accidents happen — or
(6) a local one — i.e. the premises serving for the operations of
working the railway and on which the accidents happen.
(3) A still wider conception of railway risk prevails where the
financial criterion is accepted, i.e. the enterprise on the premises of
which, or to the employees of which, the accidents happen.
I. — In German, Swedish, Bulgarian, Australian and Swiss
statistics, the movement criterion is applied 1 . Accidents are only
reported if they happen by or during the movement of trains.
Thus, on the one hand, disturbing events in the operation are not
registered as accidents if they happen while the train is not in
1
In Switzerland, the statistics moreover include casualties due to
electric current.

motion 1 . On the other hand, events which are not connected with
disturbances but lead to injury of persons are only noted if they
occur in connection with the movement of trains. Coupling and
uncoupling accidents, falls from trains, accidents whilst attending
to locomotives, etc., are not reported if the train is stationary,
nor are accidents registered which occur on the railway premises
but not on the train or by train movements ; accidents while crossing
the rails, setting points, signalling, etc., are then excluded.
I t is to be borne in mind t h a t in most of these statistics the
reported "accidents" extend also to the personal events resulting
in injury, and not merely to disturbances in the operation of trains.
As to casualties, the method of adopting movement as the
criterion does not primarily take account of the status of the persons
injured. The persons covered by the statistics are those injured
in the accidents reported. The casualties reported are consequently
those resulting from disturbing events or from personal events
arising by or in connection with the movement of trains. The
demarcation of the industrial risk of the workers employed in railway service is therefore made subsequently by classifying the persons injured in reportable accidents. The notion "railwaymen"
however is not necessarily a clearly defined conception in this case.
Workers exposed to railway risks in the above sense of movement
risks may be :
(1) Railway servants properly so-called, who comprise :
(a) the more or less permanent employees of the railway companies or administration engaged in the regular train and
station service, or
(6) workers engaged in auxiliary establishments, such as
workshops and repairing sheds, or workers carrying on
auxiliary work on the lines, such as permanent-way men,
where these are employed by the railway company, provided
they are injured in a movement accident.
(2) Workers temporarily engaged in work on or near the railway,
but employed by contractors or other employers.
(3) Workers whose duties are regularly performed on trains,
e.g. postal officials, customs officers, etc., who, although they are
1
Pires in stationary trains or at stations, for instance, are not reported
in the German statistics. On the other hand, all boiler explosions are included in Sweden and Germany.

— 10 —
not in any way engaged in the operation of the railway, may nevertheless be injured in movement accidents.
Though casualties of all these categories may be included in the
statistics, their employment risk is not fully reported, not even
that of railwaymen properly so called.
The persons covered are classified in the German statistics as
"railway employees and railway workers on d u t y " and "postal,
customs, telegraph, police and other employees on duty". In the
Swedish statistics there is only a general group "railwaymen on
duty". These classifications however do not demarcate the employment risk of these persons, but only certain specific risks incurred
by them. This conception of railway risk, which is entirely
inspired by the peculiar "movement nature" of railway risk,
appears clearly from the following explanatory note in the German
report :
I t is of no account for purposes of accident statistics whether a worker
who is working for a railway in operation is paid directly by the administration (of the railway) or whether he is employed and paid by an employer
(contractor). In the latter case, the man, if injured during his work in
consequence of the working of the railway, is also considered a railway
worker on duty '.

I I . — A wider conception of railway working, which provides
for a more complete reporting of the industrial risks of railway
workers, pi'evails in, for instance, the British, Noruiecjian, French
and Belgian statistics.
(a) The operations required for working the railway are the
criteria in Norway, France, and other countries.
In Norway, for instance, the accidents reported are the disturbances occurring in the running of trains engaged in ordinary and
additional traffic and in the shunting required for this traffic,
the so-called "accidents in railway transport properly so called",
together with personal events occurring in this exploitation. The
resulting casualties are tabulated as "casualties" and are classified
as casualties of "railway servants on duty in train and shunting
service" — t h a t is, persons engaged in running or shunting trains
and employees working at stations or on the line in connection with
such operations — and of "(other) railway servants on duty" — t h a t
is, those not engaged in the running and shunting of trains, but
1
Statistik der im Betriebe befindlichen Eisenbahnen Deutschlands, No. 45,
p. 306; Berlin, 1924. Accidents in railway shops are altogether excluded.

— 11 —
working on the railway and injured in or near a train, or during
shunting operations. I t is evident t h a t in those cases where the
injury is not connected with a disturbance of the operation, the
answer to the question whether the corresponding event is to be
included among "railway transport accidents" will depend on the
nature of the operation in the course of which the worker was
injured. In addition to these railway transport casualties, the
casualties occurring in operations required for the working of the
line, but not directly connected with the running of trains, are
registered separately — for instance, accidents in workyards,
while working on the maintenance of the line, during construction
work, etc. By this method the employment risk of railway workers
is represented by three groups : casualties of railway servants
on duty in train and shunting service ; casualties of railway servants
on duty in other railway transport services properly so-called ; and
casualties of railway servants in other than exploitation services.
(b) The place where the accident happens and the operation
are the criteria applied in British statistics.
The notion of accident, which, as will be remembered, is limited
to disturbances of operation, covers all events, failures, disturbances,
and defects involving the train, the rolling stock, and the permanent
way — t h a t is, all disturbances on the premises or defects of the
premises on which the operations for working the railway are
carried on.
Casualties are : (i) those resulting from these accidents, and
therefore from the operation of the railway, and, in addition,
(ii) other casualties happening on the railway premises, whether
caused by the movement of trains or not. "Railway premises"
are taken to be all premises used for the working of the railway;
the working of the railway being taken in the sense of all operations
directly serving the transport of goods and persons, or designed for
securing this service. Stations, permanent way, goods yards,
sidings, etc., are thus included, while workshops, factories, warehouses, repairing sheds, hotels and other premises only indirectly
connected with railway transport, are excluded.
The employment risk is then isolated by grouping all railway
servants of the companies and of contractors as distinguished from
passengers and other persons. "Other persons" include those on
business at stations and in sidings.
By this method, therefore, the total employment risk of railwaymen, t h a t is, of persons directly engaged in railway transport
service or its proper maintenance, is covered. "Railway servants",

— 12 —
in British statistics, however, include all workers and officers
employed by the companies, i.e. by the enterprise, even auxiliary
staff, as canal staff, dock staff, hotel staff, as well as mechanics,
artisans and the like who are not mainly or always working on
railway premises. These workers incur casualties which are not
covered by the statistics of railway casualties, and their employment risk is therefore not fully represented.
III. — A much wider conception of railway operation determines the scope of the United States railway accident statistics,
namely, the railway enterprise. The operation of a railway is
meant to include all operations "performed under the supervision
of the carrier by its employees", including operations for the maintenance of equipment, roadway, buildings, etc., and for the construction of "additions, betterments, and new lines, when performed by
its regular forces of employees ordinarily engaged in the construction, operation or maintenance of existing lines". In addition to the
operation of trains, locomotives and cars, "facilities incidental to
their operation, such as tracks, railway stations, depots, freight
houses, warehouses, shops, coaling stations, wharves and ferries, etc."
are covered 1.
As to accidents in the sense of unexpected events, only those
occurring in connection with the operation or movement of trains,
cars or locomotives, are reported 2 , including disturbing events
causing damage of more than 150 dollars, as well as other disturbances and personal events resulting in injury.
Casualties, however, cover all casualties of railway employees,
including those not connected with the movement or operation of
trains, cars or locomotives, and, in addition, all casualties of other
employees of the railway company engaged not in railway working
but in industrial work connected with the railway.
A detailed classification by occupation and occupational
groups demarcates the employment risks of the various categories
of workers.
It thus appears that comparability of accident statistics is
quite impaired by differences in the methods of demarcation.
1

INTERSTATE

COMMERCE

COMMISSION,

BUREAU

STATISTICS :

Rules Governing Monthly Reports oj Railway Accidents, 1922, p . 11. Revision,
Washington, 1922.
2
"Operation" does not mean operation of the railway generally, only
of the trains.

— 13 —
Unification in this respect would therefore be the first and essential
requirement in order to render the statistics comparable internationally, and it appears from the foregoing analysis that the most
appropriate demarcation of the risks of working a railway would be
achieved by applying the criterion "premises serving for the working
of the railway and its maintenance". There may still be doubtful
cases as to what operations are to be considered as forming part of
the working of the railway — for instance, whether workshops
and repairing sheds should be included. As these are on the whole
industrial premises, it would seem more appropriate to exclude
them.
By this method all disturbances in the operation of the railway
would be covered, while all casualties happening on the premises
would be included. The total industrial risk of the workers engaged
in the working of the railway and its maintenance, as distinct from
the risk of other workers, could then be assessed by means of a
classification which would combine all such workers together in a
separate group.
I t is evident t h a t the risks of railway workers differ considerably
according as to whether their work brings them more closely into
touch with the actual operation of trains, cars and locomotives, or
whether they are mainly engaged in work on stations or in other
auxiliary jobs. These differences of risk may be viewed from two
standpoints : (1) the risks peculiar to the running of trains may be
distinguished from the risks of railway working generally, which
are partly of an industrial character; (2) the risks of the several
occupational groups may be distinguished.
T H E RAILWAY R I S K GROUPS

The risk of railway working may be specified according to its
direct or indirect connection with the running of trains, and in
particular with the movement of railway rolling stock.
In the railway statistics of Great Britain, only events
which constitute disturbances in the regular working of the railway
are summarised under the heading "train accidents", personal
events not being registered as accidents. Disturbances are subdivided into "accidents to trains", which comprise mainly movement
accidents, and "accidents to or failure of rolling stock or permanent
way". These train accidents constitute, in fact, the risk peculiar
to railway working.
The casualties resulting from these accidents are termed "train
casualties". This group, therefore, represents the risk of railway
2

14 —

workers due to improper or defective working of the railway.
Other casualties occurring on the railway premises, but not connected with disturbing events, are then specified in regard to their
connection with the movement of trains, and are classified as
"movement" and "non-movement" casualties. Casualties of a
similar nature m a y be tabulated in the one or the other group
according as to whether they are connected or not connected with
the movement of trains.
The complete classification is, therefore, as follows :
(1) Accidents :
(a) train;
(b) rolling stock and permanent way.
(2) Casualties :
(a) train;
(b) movement ;
(c) non-movement.

Train and movement casualties are then combined as together
constituting the more essential railway risk.
I n the United States steam railway statistics the three groups,
"train", "train service", and "non-train" accidents, do not correspond entirely with the three British groups. As to accidents,
the group "train accidents" covers disturbances in the regular
working of the railway, but only those which result in damage in
excess of 150 dollars to equipment or other railway property and
which arise in connection with the operation or movement of trains,
locomotives or cars. The group "train service accidents" comprises
those disturbing events which result in injury, but which do not
cause damage in excess of 150 dollars, together with personal events
resulting in injury, if these events arise in connection with the
movement or operation of trains, locomotives or cars.
Casualties are grouped as those resulting from train accidents.
on the one and, hand as "train service casualties" on the other.
The latter group refers mainly to casualties resulting from personal
events connected with the movement of trains and includes casualties resulting from disturbing events in as far as the damage involved
is not more t h a n 150 dollars. Moreover, it comprises casualties
not connected with the movement of trains — "non-movement"
casualties — provided they occur during work on or near trains
or cars or locomotives, i.e. in their "operation".
The third group of casualties, the "non-train casualties",.
comprises all other casualties which occur during work carried
on by the employees of the carrier, and therefore covers the non-

— 15 —
movement casualties of railway workers arising during the working
of the railway and not included in the group of train service accidents 1.
In Norway a group of "train accidents" is distinguished, including
all disturbing events happening in the running of trains for ordinary traffic which result or might have resulted in injury to persons or damage.
The group "other accidents" includes accidents in shunting
(events which result in damage or injury, whether disturbances.
or personal events) and other personal events which occur in the
running of trains.
Casualties are grouped into train casualties — t h a t is, those
resulting from train accidents — and into other movement and
non-movement casualties occurring in the running or shunting
of trains (exploitation) ; casualties other than in railway transport
properly so called are given separately.
In Canadian statistics, which report casualties only, casualties
resulting from the movement of trains and casualties resulting
from causes other than the movement of trains are distinguished.
I n those statistics which include movement accidents and
casualties only, accidents are usually classified as collisions, derailments, and other accidents, the two former groups being evidently
considered as comprising the main train accidents. The third group
covers all disturbing events reported which are not included in
the former groups, together with personal events connected with
movement and resulting in injury.
With regard to casualties, the same main groups as for accidents
are employed for casualties in Bulgarian, Polish, Sivedish and other
statistics, and were adopted in the German statistics up to 1921,
while from 1922 onwards casualties have only been grouped as
casualties from train accidents and casualties from other accidents,
the former group comprising derailments and collisions. The
latter practice is also adopted in Switzerland.
Other countries again, such as Portugal, the Netherlands,
Hungary, Czechoslovakia, France, etc., distinguish derailments,
collisions and other accidents with regard to accidents, but do
not extend this classification to casualties. I n some of these
countries, however — for instance, in France, Czechoslovakia, and
Italy — the third group "other accidents" is not of the same scope
1

"Non-train" casualties, moreover, comprise also various kinds of
industrial casualties which occur in establishments not belonging to the
railway premises properly so-called.

— 16 —
as in the German or Swedish statistics, as it includes not only
movement accidents but also certain non-movement accidents.
The French statistics 1 distinguish casualties resulting from train
accidents, casualties resulting from other accidents occurring in
the working of the railway (de faits d'exploitation autres que les
accidents de train) and casualties caused by imprudence or misadventure.
I n Belgium, which confines accidents to derailments and collisions, casualties resulting from derailments and collisions are given
in the same group, while casualties resulting from other accidents
reported form a separate group.
To conclude, it would appear t h a t a clear demarcation between
the more specific railway risks and the more general railway risks
might be based on a distinction between disturbing events (accidents) on the one hand, and casualties (cases of injury) on the other.
Accidents might with advantage be confined to disturbances
in the working of railways generally. I n cases where the event
is strictly personal the accident is identical with the case of injury,
and not really of interest as a separate unit, as the number of such
accidents is the same as the number of casualties. The purpose
of prevention, therefore, seems adequately served by registering
only the disturbing events, and their comparison with the resulting
casualties will show the frequency of collective accidents. I t
would seem, however, t h a t some definition as to reportable "disturbing events" should be agreed upon in order t h a t the statistics may
be comparable from country to country. The registration of
any disturbing event in the regular working of railways would
leave too much scope for discretion.
I n principle, all disturbances due to derailments and collisions
might be registered. Otherwise a reportable accident might be
defined internationally as an event occurring on railway premises
which disturbs the operation of the railway, provided t h a t it results
in injury to a person or persons or damage to thé railway property
of at least, say, £50 or 1,000 marks or 200 dollars.
These accidents could then be termed railway operation accidents, a group which would differ from the British group, as the
latter includes also events not causing damage of more than
£50 or injury to persons, and from the United States group, as the
latter includes only events involving trains, cars, or locomotives.
1

Statistique des Chemins de fer français.

— 17 —
I t would seem advisable to subdivide these accidents into :
(1) movement and (2) non-movement accidents.
Movement accidents would include all disturbing events involving
the train in motion or cars or locomotives in motion. Non-movement accidents would include all other accidents in the working of
the railway occurring on the premises.
The casualties would be : (1) railway operation casualties,
i.e. those resulting from railway operation accidents; (2) other
casualties, i.e. those occurring in the working of railways, but not
resulting from accidents in the sense defined.
Both groups would then be subdivided into : (a) movement,
and (6) non-movement casualties, the former group (railway
operation casualties) according to the character of the accident
from which they result, and the latter (other casualties) according
to whether or not the personal event was connected with the movement of trains or cars in a causal sense. The non-movement
casualties of the second group would comprise all other casualties
on the premises not resulting from accidents 1.
Group 1 (a) and 2 (a) could then be combined as representing
the typical railway risk, i.e. movement casualties.
T H E EMPLOYMENT R I S K GROUTS

By grouping, on the lines indicated above, the specific and
general risks of railway working, no account was taken of the
industrial risks of the different categories of workers exposed.
We have seen t h a t railwaymen properly so-called, i.e. those
actually employed in the working of the railway, are sometimes
isolated in the special statistics from other workers exposed to
railway risks, whereby the question whether they are employed
by the railway company or administration or by a contractor is
considered of secondary importance.
These railway servants properly so-called, however, are unequally exposed according to their occupation to the different
kinds of railway risks, viz. railway operation casualties, or movement
and non-movement casualties. Those mainly engaged on the
train will be most exposed to movement casualties — the most
characteristic of railway risks. Of those working on the line and
on stations, the exposure to movement risk will be the greater the
more their duties cause them to be on the line. Railway clerks,
1
I t should be noted that the groups "railway operation casualties" and
"other casualties" do not correspond to the two groups "train accidents"
and "train service accidents" in the United States statistics.

— 18 —

for instance, will be hardly exposed at all to movement risks, nor
will station-masters or carmen be much, exposed, and even signalmen
are not greatly endangered by the movement of trains. As a rule,
the total risk of any of these occupations is higher according as the
workers are more exposed to movement risks, though this is not
necessarily so. The following table taken from the United States
railway statistics illustrates the risks of various selected occupations
classified according to the risk groups 1.
CASUALTIES OF BAILWAYMEN IN THE UNITED STATES. 1 9 2 4 *
(CLASS I ROADS)
Train
accidents

Train service
accidents

Non-train
accidents

Casualties
per million
man-hours

Total

Occupation
TÍ

u
Track and roadway
section labourers . .
Signalmen a n d signal
maintainers . . . .

-d
CD

Tí

's?

.|
'a

Tí

¡3

Tí

_§_
"5°

577

639

0.03

1.97

15,183

233

15,706

0.47

31.67

7
166

4
12

372
7,033

5
71

380
7,179

0.24
0.33

18.19
33.02

5
3
3
5

240
303
432
304

0.26
0.37
0.15
0.31

12.63
36.96
21.83
18.88

180

513

—

1
10

1
56

Yardmasters a n d asOutside hostlers . .
Inside hostlers . . .
Switch tenders . . .
Road passenger conR o a d freight conductors (through freight)
Road freight brakemen and flagmen (local and way freight)
Yard brakemen and
yard helpers
. . .
Road passenger engineers and motormen
R o a d freight engineers
and motormen
(through freight)
Road passenger firemen and helpers . .
Road freight firemen
and helpers (through

* INTEKSTATE

—
—
—
—

—

5
2
2
5

115
55
11
151

—

5
1
2

—

125
243
420
151

225

—

275

0.16

10.68

102

1,256

—

1,387

0.69

33.15

3,557

—

125

3,757

1.12

50.80

189

8,242

200

197

8,528

1.53

66.34

103

514

—

648

1.12

22.66

824

—

951

0.52

17.64

103

914

.—

1,034

1.14

38.17

111

1,841

—

1,998

0.54

36.97

COMMERCE COMMISSION, B U R E A U

Calendar year 1924, p p . 34 et seq.

1
1

of STATISTICS ; Accident

Bulletin,

No. 93,

Washington, 1925.

1
I t s h o u l d b e b o r n e i n m i n d t h a t t h e U n i t e d S t a t e s classification into
t r a i n , t r a i n service a n d n o n - t r a i n service casualties does n o t correspond t o
t h e p r o p o s e d classification i n t o casualties i n r a i l w a y o p e r a t i o n a c c i d e n t s
a n d o t h e r m o v e m e n t a n d n o n - m o v e m e n t casualties.

— 19 —
I t appears from the table t h a t the highest fatal casualty rates
are those of railwaymen engaged on the trains, such as freight
conductors, brakesmen, yard brakesmen, passenger engineers and
motormen, etc. Very low fatal rates are found among clerks.
Non-fatal casualty rates do not always vary in similar manner
to fatal rates, but the highest fatal rate and the highest non-fatal
rate are both those of yard brakesmen, while the next highest rates
are those of road freight brakesmen and flagmen (local and way
freight).
Employment risk on railways is, in fact, so unevenly distributed
t h a t a general rate for all railway workers would appear of little
value. A classification at least by categories of workers according
to their main jobs seems indispensable. Nevertheless, many
countries have not classified casualties by occupations or even
occupational groups. A very summary grouping into "railway
employees and auxiliary workers" and "railway workers" was made
in German statistics up to 1921. In Norway, railway workers
are classified, as has been shown, into "railway servants engaged
in train and shunting service" and " (other )railway servants on duty"
— comprising the other railway workers — but both groups refer
only to "exploitation accidents" (running and shunting of trains).
In the British statistics, thirty-five occupations of railway
companies' servants are distinguished, both for train and movement
casualties together, and for non-movement casualties. The workers
engaged on the most characteristic railway work are classified in
some detail, whilst those less exposed to railway risks, and not
really railway workers in the sense of workers constantly engaged
on railway premises, are more summarily grouped. The final group
"miscellaneous" comprised in 1925 no less than 76,115 adults,
who had only 9 fatal and 146 non-fatal train and movement casualties while the total figures for this class of casualties were : railway
employees 703,532; fatal casualties 251, non-fatal casualties 3,659.
Non-movement casualties for the two groups were as follows :
miscellaneous employees : fatal casualties 1, non-fatal casualties
801; all employees : fatal casualties 37, non-fatal casualties 17,556.
I t will be remembered t h a t the "miscellaneous" group comprises
various kinds of workers employed by the railway companies but
engaged in work which may be only remotely connected with
railway operation.
In the United States railway accident statistics, casualties are
subdivided into no fewer than 148 occupations which include,
however, all workers employed by the railway companies, not only

— 20 —
railway workers. These occupations are summarised into seven
groups : (i) executives, officials and staff assistants ; (ii) professional,
clerical and general ; (iii) maintenance of the way and structure ;
(iv) maintenance of equipment and stores ; (v) transportation (other
than train, engine and yard) ; (vi) (a) transportation (yard masters,
switch tenders, hostlers); (vi) (6) transportation (train and engine).
The latter group is also termed "trainmen" for whom special statistics are given. Another report, Statistics of Industrial
Accidents,
published by the Bureau of Labour Statistics, giving a summary
of railway statistics, classifies railway workers into six groups :
trainmen, shopmen, stationmen, trackmen, bridge and building
men, and other employees.
In Belgium, workers in stations and shops, workers engaged in
coupling and uncoupling, gatekeepers, watchmen, ticket controllers,
trainmen (guards, enginemen and firemen), workers repairing the
permanent way and buildings, and "persons employed in other
departments", are distinguished 1 .
A classification of occupations suitable for international comparison should primarily group those workers exposed to similar
risks. Above all, the number of workers engaged on the trains
or locomotives should be given separately. Among these, distinction should be drawn between trainmen properly so-called, such as
enginemen, motormen, firemen, brakesmen and guards, engaged
in running the train, and auxiliary workers travelling on the train,
such as waiters, cooks, stewards, and postal workers permanently
engaged for train service.
Secondly, railway servants engaged in work on the stations or
on the line may be grouped together : among these, however, a
distinction should be made between workers engaged in shunting
and those in other train service, and among the latter again between
those directly engaged in train service and slightly exposed to
train movement risks, e.g. station masters and porters on the one
hand, and office workers (clerks) mainly working inside buildings
and hardly exposed to specific railway risks on the other. A third
category would be workers engaged in maintaining the permanent
way, the structure, buildings, etc., those repairing the track as
well as those repairing or rebuilding bridges, station buildings,
etc., and exposed to railway risks during their work on the premises.
From these, however, should be distinguished artisans, mechanics,
1
"Servants not on d u t y " form a group corresponding to those mentioned above.

— 21 —
and other labourers who do not usually work on railway premises
but in shops, goods yards, repairing sheds, etc., and who may be
victims of railway accidents or casualties as a result of occasional
presence on railway premises. The employment risk of these
workers is not fully represented by their accidents on railway
premises, and if their risk is to be calculated casualties which occur
during their employment on other premises should also be taken
into account.
The following classification is thus suggested. The classification
does not include workers travelling on the train or present on railway premises in pursuance of duties in other than railway service,
such as postal, customs, telegraph and other workers, so far as these
are not permanently or temporarily working for the railway.
I. — TRAIN SERVANTS

A. — Trainmen
1. Engine-drivers and motormen.
(i) Passenger trains.
(ii) Freight trains.
2. Firemen and assistants.
(i) Passenger trains.
(ii) Freight trams.
3. Guards (conductors), ticket collectors, brakesmen, flagmen, etc.
(i) Passenger trains.
(ii) Freight trains.
B. — Other Train

Servants

1. Cooks, waiters, stewards, etc., on dining and sleeping cars.
2. Postal, telegraph workers, etc., engaged in train service.
3. Others.
II.

— STATIONMEN AND LINEMEN

A. — Train and Shunting

Service

(a) Shunting Service Men :
1. Shunters.
2. Pointsmen.
3. Other men exclusively engaged in shunting operations.
(h) Other Men :
4. Signalmen.
5. Capstanmen, cranemen.
6. Porters :
(i) Passenger trains.
(ii) Freight trains.
7. Stationmasters and inspectors.
8. Other stationmen and linemen.
9. Clerks.

— 22 —
III.

— MAINTENANCE OF W A Y AND STRUCTURE M E N WORKING
ON RAILWAY PREMISES (EXCLUSIVE or

1.
2.
3.
4.
5.
6.
7.

WORKSHOPS)

Trackmen.
Signal fitters and telegraph wiremen.
Bridge and building men.
Electrical workers.
Foremen, supervising inspectors, etc.
Other mechanics and artisans working on railway premises.
Others.
IV.

— RAILWAY SERVANTS NOT EXCLUSIVELY ENGAGED

ON RAILWAY PREMISES (INCLUDING WORKSHOP STAFF)

Railway workers may have casualties whilst not on duty but
present on railway premises for some reason or other. Such
casualties are usually shown separately, or sometimes included
among casualties of persons other than railway servants and passengers. It would seem advisable to report them separately, as
they complete the picture of the risks incurred by railway workers
owing to their occupation. Railway casualties of workers whilst
not on duty are also peculiar to railway service. Railway workers
may be present on the premises after their hours of duty have been
completed.

CHAPTER I I
RAILWAY ACCIDENT EXPOSURE

In order to measure the industrial risk of railway workers and
to compare their risks in different countries, the number of casualties must be related to some units of exposure.
Two kinds of measures of exposure are found in railway accident
statistics :
1. Those representing the exposure to all kinds of risks involved
in railway service, the measures being :
(a) the average number of railwaymen,
(6) the number of man-hours worked.
2. Those applicable only to the particular risks connected with
the movement of trains, the mileage run being taken as special
measure of exposure.
Rates per mileage m a y be based on three measures :
(a) the train mileage ;
(6) the locomotive mileage;
(c) the car (axle) mileage.
T H E AVERAGE N U M B E R EMPLOYED OR M A N - H O U R S W O R K E D

The staff of a railway differs in two respects from the working
force in other industries. On the one hand, the railway servants
proper, i.e. those employed by the railway company or administration for the regular train service, are somewhat in the position of
State employees. They constitute largely an almost stable permanent staff, and labour turnover is small. By the nature of
the service, these railway servants are required to be regularly
on duty. Days, therefore, will, as a rule, be lost on account of
leave or sickness only. For this staff, consequently, t h e average
number, as compiled from the employment lists of the railway
companies or administrations, will fairly accurately represent
the relative exposure on a given railway, as the staff is not only
permanent, but railway servants on ordinary leave or sick leave
will remain recorded in the pay-books.
On the other hand, work for the railway (other than that required
to regulate traffic) may be carried on partly by workers employed

— 24 —

by the railway company or administration temporarily or by
another employer (contractor) undertaking work for the railway.
Building and repairing, for example, may be carried on by such
temporary workers in so far as it is not performed by the regular
staff.
Moreover, the permanent staff of the railway companies will
comprise workers, such as artisans and mechanics, who are to a
large extent engaged in work on other t h a n railway premises, for
instance, in workshops and repairing sheds, and who are in consequence only occasionally exposed to railway accidents. The
average number of these workers therefore cannot be taken as
representing the exposure to railway risk. The fact t h a t they are
employed by the railway company or administration does not make
them "railway" workers.
As regards workers temporarily engaged in railway work, their
exposure can only be measured by adding up the time during which
they work for the railway in a given period.
I n the case of the permanent staff of a railway mainly or partly
employed otherwise than on railway premises, it will be difficult
to measure their exposure to railway risks as the time which they
actually spend on the railway premises can hardly be determined.
If the proportion of these workers as compared with all workers
and the proportion of their time spent on railway premises as
compared with their total time worked were the same in all cases
and countries compared, their inclusion in the exposure would not
affect the comparability of the rates. This, however, cannot be
assumed to be the case. I n calculating summary rates for railway
workers, these auxiliary workers should not be included, nor would
it seem of any great value to calculate separate rates for this
auxiliary staff, since other non-railway casualties occurring during
their employment would have to be given elsewhere than in railway accident statistics. Their railway casualties might be given
separately.
Owing to these various difficulties of calculating the exposure
of railway workers, and owing also t o the fact t h a t many statistics
do not cover all railway risks (i.e. all casualties on railway premises),
but only movement risks or those in certain operations, rates per
average number of railwaymen are not often calculated.
In Sweden, where the statistics cover movement accidents and
casualties only, rates of fatal and non-fatal casualties per thousand railwaymen are calculated. The average number of workers

— 25 —
is computed from monthly returns in the district chief's reports,
and includes not only the ordinary and temporary staff engaged
by the railway for the performance of service, b u t any workers
carrying out work for the railway, whether they are paid directly
by the railway or by contractors.
In the French statistics (Annales des Ponts et Chaussées and
Annales des Mines), fatal and non-fatal casualty rates per 1,000
railway employees are calculated.
In the United States statistics, account is taken of differences
in the risks of various categories of railwaymen. The number of
employees in service per employee killed and injured and separate
rates for trainmen only are calculated. Moreover, rates per
thousand men are published for the different categories of trainmen
separately. The numbers of railwaymen and trainmen in service
are computed from monthly statements and include all employees
on duty as well as "employees under pay on vacation or sick
leave"; employees, however, who are not subject to call for duty,
"such as employees not under pay, absent on definite leave or under
suspension, and pensioners not bound to render service" are
excluded.
Certain difficulties may arise with regard to the classification of
exposure t o risk of the workers by occupation. This classification
must, of course, correspond to the grouping of casualties by
occupation. Some workers however may do certain jobs in turn,
as driving freight and passenger trains. The American rules
provide that, in calculating the exposure, employees whose duties
are such as to make them liable t o inclusion in two or more
divisions should be included in t h a t division indicated by the
greater part of their time during the month *. For the tabulation
of casualties, however, the character of the work at the time of
the accident is decisive.
The measurement of exposure by the average number of workers
employed is, however, open to the objection t h a t the individual
days worked and the hours actually worked by the employees may
differ.
Although the man-days worked or spent on duty, and therefore
the hours worked, by an equal number of employees of a given
railway in two periods may be fairly comparable, yet the days and
1

INTERSTATE COMMERCE COMMISSION : Rules Governing the Classification

of Railway Employees and Reports of their Service and, Compensation :
Effective on 1 July 1921, p . 11. Washington, 1921.

— 26 —
hours of work may differ considerably from country to country,
a,nd even as between the different occupations of a railway (cf.
page 29).
In order to render statistics internationally comparable, rates
should therefore be calculated per man-hours worked; account is
thus taken of differences in man-days worked and in the length
of the working day.
Rates per man-hours worked are calculated in the United
States statistics for each category of workers.

T H E MILEAGE

RUN

Far more frequent, however, than rates per average number of
workers employed or per hours worked are rates per mileage
applied to movement risks.
Rates per million train kilometres and per million car kilometres
were calculated in German statistics for railwaymen and railway
employees on duty. In Swiss statistics, rates per hundred thousand locomotive kilometres and per million car kilometres are
published for railway employees. I n Sweden rates per train kilometres and per car kilometres are published, in addition to those
per thousand employees. I n the United States statistics, fatal
and non-fatal casualty rates per million locomotive miles are calculated for train and train service accidents of all employees.
Moreover, rates per ten million freight miles are calculated for freight
trainmen and similar rates for passenger trainmen. I n Norway
train and shunting casualties are related to one million train kilometres. In Bulgarian statistics rates are given per one million train
kilometres and per ten million axle kilometres, etc. Mileage rates are
usually calculated for those casualties only which result from the
running of trains, for instance movement risks in Germany, Sweden
and Switzerland, and movement risks together with other risks
in the operation of trains, cars, etc., e.g. in the United States and
Norway.
By referring the number of casualties to the mileage run by
trains, locomotives or cars, the risk connected with a certain
amount of output is no doubt measured in so far as mileage can
be considered as output, irrespective of the number of passengers
and the quantity of goods carried. If, however, the mileage run
is to be considered as a measure of exposure, allowing a comparison
of the risks of railwaymen on different railways or at different
dates or for different categories of workers, it would have to be

— 27 —

proved either : (1) t h a t the mileage is a direct factor of exposure,
i.e. t h a t the number of casualties varies in proportion with the
mileage run; or, if this cannot be proved, (2) t h a t the mileage run
is an indirect measure of the time worked or the average number
of workers, i.e. these two factors must vary in proportion with
the mileage run.
The second alternative may be considered first. Is the mileage
run an indirect measure of time worked or of the numbers
employed ?
(a) Within a country and within certain limits the relation
between the numbers employed and the mileage run may be fairly
constant, but experience shows that a considerable increase in
traffic and mileage run can be effected without a proportionate
increase in the number of, or the time worked by, men not directly
engaged on the train. Moreover, the quicker the trains the fewer
will be the trainmen or trainmen-hours required to run a certain
mileage.
If these objections can be raised as regards any one country,
they will be even more valid in regard to comparison of risks in
different countries. The number of trainmen required t o run a
certain mileage may differ according to the average length of the
trains as well as the average speed. Further, the general staff, other
than trainmen, required per mileage run will evidently be greater
the more numerous the stations as compared with the length of the
road. The density of stations is, of course, a function of the density
of the population. Thus, the train mileage run per employee was
700 miles (1,126 kilometres) in the United States (1925) and 734
kilometres in Germany (1925); that is, about one and a half times
greater in the United States 1. This is no doubt largely due to the
greater distance between stations in the United States. I t may
also be due partly to a greater average speed in t h a t country.
The mileage run cannot therefore be considered as an indirect
measure of the time worked or the average number of employees.
(6) In order t h a t mileage rates should provide a satisfactory
measure of the risk of railwaymen, it would be necessary to show
1

United States: Freight, passenger and work train miles: 1,220,845,788;
number of employees : 1,744,311.
Qermany : Train kilometres : 549,946,274 ; number of employees :
749,638.
The difference would be still greater if in the United States only railwaymen properly so called were taken into account.

— 28 —

t h a t the number of casualties varies in proportion with the mileage
run ; this would mean t h a t the risk does not depend on the number
of railwaymen or the time worked, nor on the time taken by the
train to run a certain mileage, i.e. on the speed.
Data based on mileage rates are available only for the United
States. The tables below give the number of trainmen employed,
of train miles run, of casualties, and the risks of freight trainmen
and passenger trainmen. The number of freight trainmen is more
than twice as great as the number of passenger trainmen, while the
mileage run is only slightly less in the case of passenger trains. A
greater number of freight trainmen is necessarily required to run
the same mileage, as freight trains are much slower and therefore
more freight trains or more time respectively are necessary to run
the same total mileage. I t is a striking fact t h a t the casualty rate
per ten million miles is much higher for freight trainmen than for
passenger trainmen.
The proportion between the casualties of freight trainmen and
the casualties of passenger trainmen far exceeds the proportion
between the mileage run b y freight trains and the mileage run b y
passenger trains. The great differences in the mileage rates therefore are most likely due mainly to differences in the number of
men employed or man-time worked.
However, the risk per men employed also is greater for freight
trainmen than for passenger trainmen, and this difference appears
throughout the years of observation. I t suggests the presence
of another factor working to the detriment of freight trainmen.
I t might be assumed t h a t speed has some influence on the rates,
as the speed of the trains is t h e main difference in regard t o freight
and passenger trains. Statistics of fatal casualties in train and
train service accidents of trainmen in the United States from 1916
t o 1924 and of fatal casualties in train and train service
accidents by occupations in 1924 are given in the following
tables.

— 29 —
UNITED STATES. —

CLASS I ROADS : FATAL CASUALTIES I N TRAIN

AND TRAIN SERVICE ACCIDENTS OF TRAINMEN, 1 9 1 6 TO 1 9 2 4 1
Train
Number
of
miles
casualties (in 1,000)

Year

Trains'1

Freight
1916
1917
1918
1919
1920
1921
1922
1923
1924

681
760
847
509
605
302
328
436
295

693,671
705,167
690,046
614,095
675,788
561,372
587,950
683,481
631,256

Passenger
1916
1917
1918
1919
1920
1921
1922
1923
1924
1

113
136
150
128
147
94
97
109
81

INTERSTATE

COMMERCE

Number R a t e per R a t e per
of
10 million
1,000
trainmen
miles
trainmen

154,027
165,953
169,819
151,015
163,774
137,852
141,879
163,292
149,764

9.8
10.8
12.27
8.28
8.9
5.4
5.6
6.38
4.67

4.42
4.58
4.99
3.37
3.69
2.19
2.31
2.67
1.97

1,96
2.36
2.83
2.37
2.62
1.69
1.79
1.94
1.46

1.96
2.37
2.71
2.32
2.54
1.64
1.71
1.88
1.41

Trains

576,094
575,500
529,444
539,803
561,633
554,805
541,275
560,980
553,253

57,611
57,435
55,366
55,282
57,858
57,304
56,660
57,981
57,596

COMMISSION, B U R E A U OF STATISTICS :

Accident

Bulletin , N ° 9 3 , pp. 7, 8, 112. Mileage rates calculated by Office so as to include
Class I roads only.
2
Freight, mixed and work train miles.
UNITED STATES. — CLASS I ROADS : FATAL CASUALTIES I N TRAIN
AND TRAIN SERVICE ACCIDENTS BY OCCUPATIONS IN 1 9 2 4
Occupation
Freight trainmen :
Engineers a n d
motormen
F i r e m e n a n d helpers
Conductors . . . .
Brakesmen and
flagmen
Passenger trainmen :
Engineers a n d
motormen
F i r e m e n a n d helpers
Conductors . . . .
Brakesmen and
flagmen

R a t e per
Rates per
Man-hours
Number of Number of
thousand
(in 1,000) million
trainmen casualties
trainmen
man-hours

31,015
33,346
24,864

37
43
47

1.19
1.29
1.89

83,345
83,243
71,087

0.445
0.518
0.662

60,539

168

2.78

163,665

1.026

12,977
12,674
11,730

32
31
4

2.47
2.45
0.34

28,599
27,086
28,767

1.12
1.14
0.14

14,360

0.90

33,726

0.39

— 30 —
Though the total risk of freight trainmen is higher than the
total risk of passenger trainmen, this is not the case for each
category of trainmen. On the contrary, the rates of passenger
enginemen and motormen and of passenger firemen are about
double those of freight enginemen and firemen respectively. B u t
the risk of freight conductors is about five times greater than t h a t
of passenger conductors and the risk of freight brakesmen three
times t h a t of passenger brakesmen.
An attempt may be made to explain these facts.
Enginemen
and firemen are engaged on the locomotive. The speed of a passenger train being much greater, any accident involving the passenger
train is likely to have more serious effects than a similar accident
to a freight train for those who are mainly exposed, viz. the locomotive men. The greater speed would mean for these workers
greater exposure to risk. This assumption seems to be confirmed
by the following facts :
(a) If rates are calculated separately for through freight and
local and way freight enginemen and firemen, it appears t h a t rates
for through freight trainmen are higher than those for local freight
trainmen (rates per man-hours).
Through freight engineers and motormen
Local and way freight engineers and motormen
Through freight firemen and helpers
Local and way freight firemen and helpers

0.52
0.31
0.54
0.48

(6) The higher rates for passenger enginemen and firemen are
mainly due to the greater number of casualties in railway operation
accidents and among these particularly derailments. There may
be two explanations of the greater number of these casualties :
either passenger trains have more train accidents or derailments
respectively than freight trains, or the accidents have more serious
consequences for passenger trainmen on account of the speed.
The latter explanation seems the more valid as a greater number
of derailments is found for freight trains, even if it is taken into
account t h a t a much greater number of freight trains is required
to run the same mileage.
The case is very different in regard to conductors and brakesmen.
The rates per man-hours for these trainmen are very much higher
with freight trainmen than with passenger trainmen, though the
speed of passenger trains is much higher. I t would seem, therefore,
t h a t in this case personal events are the predominant source of
danger, or the so-called "casualties arising from train service acci-

— 31 —

dents" of the American statistics, which correspond to movement
and non-movement casualties in the sense adopted by the present
study. The risk consequently appears to be rather diminished
by speed as the slow and frequently stopping train evidently
requires far more operations entailing train service risks, such as
coupling, braking, switching, shunting, etc. In fact it is found
t h a t the higher rates for freight brakesmen and conductors are
mainly due to train service casualties, as the following figures
show :
FATAL TRAIN SERVICE CASUALTIES OP ROAD TRAINMEN
PER MILLION MAN-HOURS
Passenger conductors
Freight conductors (through freight)
Freight conductors (local and way freight)
Passenger brakesmen
Freight brakesmen (through freight)
Freight brakesmen (local and way freight)

0.115
0.526
0.581
0.267
0.803
0.946

I t appears moreover t h a t the rates vary inversely as the speed,
being highest for local and way freight brakesmen. I t must therefore be concluded that speed is an important factor of risk.
To sum up, it seems t h a t mileage run is not a suitable measure
of the rates of comparative risks of railwaymen. I t is, however,
useful for the calculation of accident rates, measuring the frequency
of disturbing events in relation to the "output".

CHAPTER HICLASSIFICATION BY CAUSES

An accident is usually due to a complexity of causes. The
unexpected coincidence of contributory circumstances — events,
acts, defects —• causes injury to life or damage to property, or both.
For the purposes of accident prevention, it is less important to
determine the causal relation than to specify those known circumstances which indicate how the accident might have been prevented.
The classification of accidents and casualties by causes is therefore
made from this point of view, and where a diversity of circumstances contribute to the accident, it may be necessary to introduce
a collective notion indicating not the details of the occurrence, but
the coincidence of different circumstances. Thus, on the railways
the disturbing event in the operation which results in injury or
damage, and which in the majority of cases might have been prevented, will be the main criterion. Different causes, for instance,
of an accident or casualty may be concealed under the heading
"derailment". I t is then the task of a further tabulation to specify
the contributory circumstances which led to the derailment — and
which may be termed secondary causes with reference to the
casualty. The disturbance may have resulted from events outside
the control of the persons engaged in regulating the railway service
or from events or circumstances due to personal negligence. Acts,
or failures to act, of persons or inadequate maintenance of the
premises or equipment may be particularly frequent on railways.
When injury is not the result of a disturbing event, it is often
more difficult to speak of a cause. The injury is the result of
coinciding circumstances which may, for the purpose of classification by causes, be termed "special circumstances" and are distinct
from the nature of the work in which the person injured was engaged
and from the railway operation during which the casualty happened.
The special circumstances (for instance, run over by a train, slipping,
struck by objects, etc.), though causes in the sense t h a t they have
resulted in injury, do not always supply sufficient information for
preventive purposes. The classification is therefore often completed by, or even confined to, an indication of the (1) nature of

— 33 —
the work, (2) the railway operation or (3) certain local circumstances
which had a bearing on the casualty.
The nature of the work is here meant to refer to the occupational
activity in which the worker was engaged when the accident
happened (not his occupation), while operation is meant to refer
to the working of the railway in a technical sense. The nature of
the work and the operation are of course not causes strictly speaking, and only by the coincidence of several factors, such as negligence, fatigue or haste on the part of the injured or on the p a r t
of third persons, may they give rise to the injury.
Finally, local circumstances may be of importance both for
the causation of accidents and of casualties; for instance, the
presence of a person on the track, the fact t h a t he is getting on or
off an engine or car, etc.
As in railway statistics the accidents (mainly disturbances of
operation) are always recorded separately as statistical units,
different classifications for accidents and for casualties are sometimes found. I t is better therefore to discuss first the classification
of accidents by causes, then the corresponding classification of
casualties, and finally to show how they are and might be combined 1 .

T H E CLASSIFICATION OF ACCIDENTS

The separate classification of accidents has a meaning only in
regard to the disturbing events. In the other cases where the
personal event is registered as an accident, the classification
coincides, to a large extent, with that of casualties. We therefore
discuss here the classification of railway operation accidents
(disturbances) only.
Accidents are always classified according to the nature of the
disturbing event. In most statistics we find such main groups
as derailments, collisions, boiler explosions and fires in trains. As
noted above, in a number of countries — especially those which
publish movement accidents only — three main groups of accidents
are distinguished : derailments, collisions and other accidents. The
group "other accidents" usually contains also the personal events
registered as accidents. The accidents in the third group are in
some countries further subdivided according to the nature of the
event. In Swedish statistics, for instance, the group contains
the subdivisions : running into vehicles, fire on trains, boiler
1

Some typical classifications will be found in the Appendix.

— 34 —
explosions and other accidents. These groups, however, are based
on the same criterion as the classification of the main groups
"derailments" and "collisions", viz. the nature of the accident.
On the other hand, the subdivision of the main groups "derailments" and "collisions" is one by causes as above defined.
The British classification, which includes all the disturbances
on railway premises, not only those involving trains, is based on
the distinction between accidents to trains and accidents to rolling
stock and permanent way. The former group is summarily divided
into five subgroups : derailments, collisions, running into obstructions, fires, and miscellaneous. The second group (accidents to
rolling stock, permanent way, etc.) is classified either by the rolling
stock or part of permanent way which was defective (for instance,
failure of tyres, axles, tunnels, boilers or machinery) or by certain
events such as fires in stations or on bridges, floodings, or slips.
These groups are further specified either according to the technical
nature of the rolling stock or to certain specifications of the defects
or failures, etc., e.g. failure of tyres : tyres of engines, of tenders,
etc., broken at screw or bolt holes, etc.
I n the United States, collisions, derailments, locomotive boiler
explosions and other locomotive accidents and miscellaneous
accidents are distinguished as main groups.
These main groups are generally subdivided according to various
criteria :
1. I n the statistics of various countries derailments and collisions are often classified according to the rolling stock involved,
this classification being sometimes extended to other accidents
also. I n German statistics, for instance, up to 1921 collisions and
derailments of passenger trains, freight trains and single vehicles
or parts of trains engaged in shunting were distinguished. In
Belgian statistics, derailments and collisions are given separately
for passenger trains and for goods trains and light engines. I n
British statistics, five groups of collisions and two groups of derailments are distinguished, according to the trains involved. I n the
United States statistics, each group of railway operation accidents
is divided into twelve classes according to the kind of trains or
train compositions involved. I n the Swedish statistics, the distinction between derailments and collisions in train service and in
shunting probably serves somewhat the same purpose, since it
distinguishes the operation in which the train was engaged.
2. Subdivisions according to the place or rather to the local
circumstances in which the accident occurred are frequently given,

— 35 —
either in combination with the kind of rolling stock involved or
separately.
Derailments and collisions on the road or in stations are
distinguished in German, Swiss, Norwegian, Swedish, French,
Italian and other statistics. Some indications of place are given
in the British classification of collisions. I n the United States
statistics, no classification by place is made, b u t the subdivision
of collisions by so-called classes contains some references to place
(see page 36). I n pre-war Belgian statistics 1, six groups of collisions were distinguished according to the place where the accident
happened or the rolling stock involved. Derailments at stopping stations and derailments on the road or at stations where the
train was not scheduled to stop were distinguished.
3. The subdivision of the main groups of accidents according
to causes properly so called, t h a t is, by contributory circumstances,
is usually based either on the defects of rolling stock or permanent
way or on the acts, or failures to act, of employees t h a t caused
the disturbing event. As in railway operation fault in the sense
of negligence or wrong performance of duty can frequently be
determined, the latter criterion is often applied — for instance,
wrong setting of points, careless shunting or defective signalling.
Not all countries give such subsidiary classification of accidents
by causes. In British statistics the main groups of accidents arising
in railway operation are only subdivided according to the rolling
stock involved with but one reference to faults of the personnel
(collisions of trains with buffer stops in consequence of too high a
speed) 2 . I n the United States statistics two classifications of the
main groups of railway operation accidents are found. A general
classification is applied to all the main groups, containing very
detailed subdivisions. The four summary groups of the general
classification are as follows :
(1)
(2)
(3)
(4)

Negligence of employees;
Defects in or failure of equipment;
Defects in or improper maintenance of way and structure ;
Miscellaneous.

These groups are divided in great detail into several hundred
subgroups. Moreover, each main group of accidents is classified
1
Post-war statistics were much reduced and this classification discontinued.
2
See classification of casualties in train accidents, Appendix, table IV.

— 36 —
separately according to its particular nature. The classification of
collisions is not one by causes, but is made either according to the
place where the accident occurred or t o the circumstances in which
the trains collided. The special classifications of the other main
groups are by causes and partly summarise such detailed groups
of the general classification as are of special importance for the main
group in question. The classification reads as follows :
UNITED STATES : CLASSIFICATION OF TBAIN ACCIDENTS

Collisions :
Rear-end.
Head-on.
Broken-train.
Side or raking.
Collisions at railway crossings :
On private rights of way.
On publie streets, etc.
Trains with cars not in trains.
Switching.
Not elsewhere classifiable.
Derailments due to :
Defects in or failure of power-brake apparatus, hose, etc.
Defects in or failures of couplers.
Other defects in or failures of locomotives or cars.
Defects in track, bridges, switches, and signals, or other defects in roadway.
Accidental obstructions or defects in track due to fires, landslides,
floods, etc.
Obstructions at public highway crossings.
Negligence, mistake, or misconduct of trainmen or other employees.
Mistake or misconduct of persons other than employees.
Ascertained causes not classifiable above.
Unknown causes.
Locomotive-Boiler Accidents :
Shell explosions.
Crown-sheet explosions, low water, no contributing causes.
Crown-sheet explosions, low water, contributing causes.
Other explosions.
Tubing or pipes subject to steam pressure, defects in or failures of.
Other boiler accidents.
Other Locomotive Accidents :
Failures or defects in :
Cylinders.
Driving gear and machinery.
Wheels and axles of locomotives.
Tenders.
Miscellaneous.

— 37 —
Miscellaneous Train Accidents :
Colliding with trolly-oars, automobiles, etc., at public highway crossings.
Other miscellaneous train accidents.

As to train service accidents, their classification coincides with
the classification of train service casualties.
In German statistics, up to 1921 collisions and derailments were
classified either according to defects or t o wrong or negligent acts
of employees. Swedish statistics have a classification by causes
similar to t h a t of the former German statistics. I t reads as follows :
SWEDISH CLASSIFICATION OF TRAIN AND MOVEMENT ACCIDENTS

A. — Derailments :
(1) Obstacle on line.
(2) Defect in superstructure :
(a) Breaking of rails.
(b) Other defects.
(3) Defect in rolling stock :
(a) Breaking of axle.
(b) Breaking of felloe.
(c) Other defects.
(4) Excessive speed.
(5) Fault in shunting.
(6) Defective state of turntable.
(7) Other causes.
B. — Collisions :
(1)
(2)
(3)
(4)
(5)
(6)
(7)

Defective signalling or inattention to signals.
Parting of train (broken train).
Train running into stations at too high a speed.
Starting vehicles at wrong time.
Faulty shunting.
Vehicles side-tracked without properly closing line or other carelessness.
Other causes.

C. — Other Accidents :
(1) Running into vehicles.
(2) Fire on trains.
(3) Locomotive boiler explosion.
(4) Other accidents.

In France train accidents are also classified by defects or failures
of rolling stock or permanent way or by negligence, etc., of
employees.
In Norwegian statistics the classification of derailments and
collisions is based either on acts or failures to act of employees or

— 38 —
defects of equipment or premises and on certain extraordinary
events such as falls of earth or stone, avalanches, etc. Other
railway working accidents are not further specified in respect of
causes except for "fires in train".
I n Belgian pre-war statistics three main subdivisions were given
for derailments and collisions : defects of permanent way, defects
of rolling stock, and accidents due to faulty execution of service.
These were again subdivided either according to the nature of the
defects or failures of equipment and structure as regards the two
former groups, and according to acts or failures to act of employees
as regards the third group.

T H E CLASSIFICATION OF CASUALTIES

With regard to the classification of casualties, a distinction
must be made between casualties resulting from accidents arising
in railway operation in the sense of disturbing events and casualties
not resulting from such railway operation accidents. I t must be
noted t h a t this Report is concerned with the risks of railwaymen
only. Classifications applied to all persons killed or injured are
therefore not considered here.
(a) In many statistics casualties resulting from accidents in
railway operation are classified first of all into the main groups
of accidents.
I n British statistics, for instance, casualties in train accidents
are grouped much in the same way as these accidents themselves,
though somewhat more summarily as regards the casualties in
accidents to rolling stock and permanent way. I n the United
States statistics the main groups of train accidents, and also the
subdivisions of these groups, the general classification and the special
classifications, are applied to casualties. Certain railway operation
casualties, it will be remembered — namely, those resulting from
events causing less than 150 dollars damage to property and injury
or only injury to persons — are included among train service
casualties.
I n the statistics, however, in which railway operation accidents
are not distinguished from personal events, casualties in derailments
and collisions are often grouped separately, while casualties resulting from other railway operation accidents are given together with
casualties not resulting from such accidents. This is done, for
instance, in the German, Swedish, Swiss, French, and other statistics. The subdivisions of the classifications of accidents are not

— 39 —
usually applied to casualties, except, e.g., in the Bulgarian statistics.
In most other countries all casualties — railway operation casualties
and others — are separately classified in a scheme adapted to casualties not resulting from railway operation accidents. In Belgium,
casualties in derailments and collisions are given separately from
all other casualties.
(b) Casualties not resulting from railway operation accidents —
those not connected with disturbing events — are classified either
by the special contributory circumstances, by the nature of the
work or by the local circumstances in which they occur.
In regard to movement casualties, groups based on the criterion
"nature of the work in which the injured was engaged", include
casualties while coupling or uncoupling, moving wagons, shunting
(in the sense of the personal work and not in the sense of an operation). Groups combining casualties according to local circumstances include casualties while walking on the line, or getting on
or off engines and cars. Finally, groups determined by the special
circumstances resulting in injury include falls from trains, caught
between cars, run over by trains, striking against obstacles, etc.
The German classification, which the Swedish, Swiss, Norwegian
and some other systems resemble, contained six groups based either
on the nature of the work, carelessness or "local circumstances".
Distinction is drawn between casualties resulting from :
(1)
(2)
(3)
(4)
(5)
(6)

train accidents;
carelessness in getting on or off cars in motion, or in cars, etc. ;
shunting and moving cars, etc. ;
coupling or uncoupling;
being on line or passing line not a t proper time;
other carelessness in service.

In some countries — for instance, in Norway — certain groups
are summarised under the heading "shunting accidents".
In Great Britain, movement casualties are classified first by
the operation during which they happen — shunting accidents and
other accidents — and then subdivided in some detail by various
criteria. Some groups, which indicate the nature of the work
(e.g. attending to machinery), are co-ordinate with groups based
on the criteria of the special contributory circumstances (e.g.
being caught between vehicles), while others indicate the local
circumstances in which the accident happened (walking or standing
on line).

— 40 —
In the United States statistics, train service casualties are subdivided according to groups of causes. The same grouping, incidentally, is applied to train service accidents, of which only a few
are disturbing events. Some of the groups are based on the
criterion of the nature of the work (coupling of locomotives
and cars, coupling of air or steam hose and safety chains, operating
locomotives, operating hand brakes, operating switches), while
others are based on the criterion of special contributory circumstances (e.g. coming in contact with fixed structures while on moving
cars or locomotives ; struck or run over by cars or locomotives not
classified otherwise), and one group indicates the local circumstances (getting on or off cars or locomotives). Finally, there is
one group comprising level crossing accidents and casualties, which
is mainly a group of railway operation accidents and casualties 1 .
These groups of casualties are then divided into several hundred
subgroups which specify the nature of the work, the special
contributory circumstances, or the local circumstances.
In Belgian statistics, movement and certain non-movement
casualties constituting one group are subdivided according to a
detailed classification which combines a grouping by occupation
with a grouping by causes. Casualties connected with the movement of trains, etc., are always separately mentioned.
In some statistics, therefore, non-movement casualties not resulting from railway operation accidents are combined with movement
casualties and classified with them, as, for instance, in the United
States and in Belgian statistics. In British statistics non-movement
casualties are separately classified. The groups are based mainly
on the criterion of special contributory circumstances, frequently
combined with an indication of the place or the local circumstances.
Some groups, however, refer to the nature of the work.
In the United States statistics, the non-movement casualties
which are termed "non-train casualties" comprise, in addition to
non-movement casualties of railway workers 1 , casualties of
employees not engaged in railway operations properly so-called
which are similar to casualties in manufacturing industries or other
services; the classification of non-train casualties is of a general
industrial character and not specially related to railway risks.

1
Train service casualties include, as mentioned above, non-movement
casualties connected with the operation of trains, cars or locomotives. A
great part of the non-movement casualties of railway workers are thus
included among train service casualties.

— 41 —
To conclude, it has been seen t h a t the underlying principles of
the classifications are much the same in most countries. Specific
groups, however, do not coincide sufficiently to allow of a satisfactory comparison. I t should not be difficult to secure agreement
upon a uniform scheme if the general grouping into railway operation accidents and the resulting casualties on the one hand, and
casualties in railway working not resulting from operation accidents
on the other hand, and the corresponding definitions, were adopted.

PROPOSED CLASSIFICATION

I t would appear from the instances cited t h a t co-ordinated
groups are frequently based on different criteria which may be
an obstacle not only to comparison, but also to a clear demarcation.
For each subdivision only one criterion ought to be applied.
A tentative proposal for a uniform classification may be given.
The scheme is the result of the analysis of the various classifications,
and is based in the main on those criteria which have been found at
the basis of most classifications.
(1) Accidents : railway operation accidents might be grouped
according to the nature of the disturbance (e.g. derailments, collisions) whereby movement and non-movement accidents might be
distinguished.
The main groups might be subdivided into accidents due t o
defects of equipment or of premises or to the acts or failures to
act of employees. A further subdivision according to the rolling
stock involved and the place or local circumstances under which the
accident occurred might be introduced.
(2) Casualties : casualties resulting from railway operation
accidents should be classified according to the same scheme as the
accidents themselves and could be included in the same tables.
Casualties not resulting from railway operation accidents might
be tabulated as movement and non-movement casualties. I n
order to ensure a clear demarcation of the groups and the greatest
possible comparability, one criterion only should be applied in
determining the main groups of movement and non-movement
casualties.
The nature of the work in which the injured was engaged does
not appear to furnish the best criterion, as it is not always of interest

— 42 —
for preventive purposes. Moreover, a classification by occupations,
which is altogether indispensable, will in many cases specify the
nature of the work during which the accident occurred.
The special contributory circumstances are more important
from the point of view of prevention : if the classification by special
circumstances is combined with t h a t of occupation, it will be shown
how the risk inherent in the work actually resulted in injury.
I t may, however, be necessary in order to obtain a clear idea
of the origin of the accident to take into account in a subdivision
either the railway operation, or the nature of the work or personal
job. The workers in a given occupation may be engaged part of
their time in particularly dangerous operations, e.g. technical
processes such as shunting ; also the operation may require different
jobs entailing unequal risks : coupling, for instance, may be
especially dangerous.
A co-ordinated subdivision by these two criteria (operation and
nature of work) may therefore be desirable in order to bring out
the inherent risk of the work.
Finally, the casualty may happen not in any particular operation
or job, but under certain local circumstances entailing special
degrees of exposure. I n such cases this criterion might be applied
in the subdivision.
The following classification is based on the principles set out
above.

— 43 —

I. — Accidents Arising Out of Railway Operation and the Resulting Casualties
1. MOVEMENT ACCIDENTS BY MAIN CAUSES

Non-fatal
Number

Rate per
1,000 railwaymen
or 100,000
man-hours

Fatal
Number

Rate per
million
train miles

Number

Rate per
million
train miles

Number

Causes of accidents

Casualties

Resulting in
casualties to
railwaymen

Rate per
1,000 railwaymen
or 100,000
man-hours

Accidents
Involving
material
damage only

A. Derailments (see 2
and 3A) :
(a) on line . . . .
(b) in station, shunting yard, etc.
B. Collisions (see 2 and
3B) :
(a) on line . . . .
(b) in station, shunting yard, etc.
C. Locomotive boiler
explosions (see 3C)
D. Running into obstacles a t crossing
(see 3C) :
(a) guarded crossing
(b) unguarded crossE. Fires in trains (see
3C)
F. Miscellaneous (see
3C)

2. DERAILMENTS AND COLLISIONS BY CLASS OF ROLLING STOCK
Cause and kind of rolling stock involved
Derailments :
(3) Parts
Collisions :
(a) with passenger trains . . .
(b) with freight trains . . . .
(c) with parts of trains or single
(2) Freight trains :
(a) with freight trains . . . .
(b) with parts of trains or single
(3) Parts of trains with parts of
trains or single vehicles . . . .

Steam traction
(subdivided as
in I, 1)

Electric traction
(subdivided as
in I, 1)

— 44 —
3. MOVEMENT ACCIDENTS BY DETAILED CAUSES
In station or
On line (subdivi- shunting
yard (subded as in I, 1) divided as
in I, 1)

Causes of accidents

A. — Derailments
(1) Fault of service :
(a) Wrong or improper running of
(b) Defective signalling
(c) Disregard of signals
. . . .
(d) Defective shunting
(e) Other fault of service . . . .
(2) Defects in rolling stock :
(a) Failure of tyres

(e) Defects or failure of locomotive
(f) Defects or failure of cars or
(3) Defects in structures and permanent way :
(a) Defects in rails or joints . .

(e) Defects of other fixed structures
(f) Other defects
(4) Obstacles on track or line :
(a) Vehicles at crossings
. . . .
(b) Snow, ice or avalanche . . .

B. — Collisions
(1) Fault of service:
(a) Wrong order of station staff
or failure to transmit or obey
(c) Disregard of signal or other
(d) Wrong setting of points . . .
(e) Failure in handling brakes . .
(f) Other carelessness in shunting
(g) Running at too high a speed
(h) Other fault of service . . . .
(2) Defects or failure of rolling stock :
(a) Failure of couplers
(b) Failure of brakes
(c) Other defects

— 45 —

3. MOVEMENT ACCIDENTS BY DETAILED CAUSES
Causes of accidents

(continued)

(Subdivided as in I, 1)

C. — Accidents Other than Derailments and Collisions
(1) Locomotive boiler explosions :

(2) Running into obstacles at level
crossing (other than 3 A, sub-

(3) Fire in train :
(b) Fault of passengers

4. NON-MOVEMENT ACCIDENTS BY CAUSES
(Subdivided as in I, 1)

Causes of accidents
( 1 ) Fire in stations :
(b) Sparks from train

. . , . .

(2) Fire on line :

(3) Floodings
(4) Miscellaneous :
(a) Slips, etc., of embankment . .
(b) Snow, ice, avalanches (other
than 3 A, subheading 4 (b)) .

— 46 —

II. — Casualties Not Resulting from Accidents Arising
Out of Railway Operation
1. MOVEMENT CASTJALTIES BY CAUSES
Fatal casualties
Causes of casualties
Number
A. Run over by trains, ears, locomotives
(1) Shunting operations :
(a) Coupling or uncoupling . .
(b) Moving cars, wagons, etc. .
(c) Braking or chocking wheels
(d) Attending to ground points
(e) Other operations
(2) Working on line
(3) While passing over or standing
or walking on line
(4) While getting on or off engines,
cars, etc., in motion
(5) Otherwise
B. Caught between trains or cars :
(1) Shunting operations :
(a) Coupling or uncoupling . .
(b) Moving cars, wagons, etc. .
(c) Braking or chocking wheels
(d) Attending to ground points
(e) Otherwise
(2) In other operations
C. Falling off trains, cars, locomotives
in motion :
(1) Shunting operations
(2) While on train in road service .
D. Struck by or striking against structures or other obstacles :
(1) Shunting operations
(2) While on train in road service
E. Otherwise injured :
(1) Shunting operations
(2) While on train in road service .
(3) While getting on or off engines,
cars, etc., in motion
(4) Otherwise

Non-fatal
casualties

Rate Number

Rate

47 —

2. NON-MOVEMENT CASUALTIES BY CAUSES

Fatal casualties
Causes of casualties
Number
A. Falling off ladders, platforms,
trains, etc. :
(1) Shunting operations
(2) While getting on or off trains or
cars at rest
B. Falling or slipping :
(1) While walking on line or passing
(2) During transport of luggage or

C. Struck or otherwise injured by
falling objects (wagon doors, rails,
D. Otherwise injured :
(1) By flying particles
(2) By electric current
(3) By animals

Rate

Non-fatal
casualties
Number

Rate

APPENDIX 1
NOTES ON THE STATISTICS OF RAILWAY ACCIDENTS
IN DIFFERENT COUNTRIES
AUSTRALIA
Railway accident statistics are published on a uniform plan by
the State railways of the several States in their yearly reports to
Parliament and by the Commonwealth Railways for the federal
lines.
Scope
State and Federal Railways only are covered; private railways
are not covered.
Accidents in the sense of this Report are not reported.
Casualties reported are those which result from accidents which
occurred through or in connection with the movement of rolling
stock. All non-movement casualties (railway operation as well as
other non-movement casualties) are thus excluded.
Fatal and non-fatal casualties are, since 1924-1925, classified
by the following groups : train accidents; accidents on line other
than train-accidents; shunting accidents; proceeding to or from
duty within the railway boundary.
Up to 1923-1924, the classification was as follows : while in
the execution of their d u t y ; proceeding to or from duty within
the railway boundary. The former group was subdivided as
follows : causes beyond own control ; contributory negligence ;
solely own action or negligence.
Casualties reported are those occurring in the working of the
railway or on railway premises caAising the employee to be absent
for at least one whole day from his ordinary work.
AUSTRIA
Summary statistics of railway accidents are published annually
in the financial report of the Austrian Federal Railways {Geschäftsbericht der Unternehmung "Oesterr eichische Bundesbahnen"). These
are much reduced as compared with the statistics published before
the war in the report entitled Oesterreichische
Eisenbahnstatistik (Austrian Railway Statistics). Formerly Austria also
collaborated in the publication of the statistics of the "Verein
1

tion.

Tables are only given in order to illustrate the chief types of classificaFigures are not comparable as between countries.

— 50 —
deutscher Eisenbahnverwaltungen". Only the more extensive
pre-war statistics are described here 1 .
Scope
The statistics covered State and private railways, steam as well
as electrical traction. Statistics for the main lines and local lines
were more elaborate than those for minor railways (narrow-gauge
lines) and cog railways. Statistics of accidents in ancillary undertakings, buildings for the administration, etc., classified by causes
and nature of accidents, were compiled for administrative purposes,
b u t were not published.
Main and local lines. Accidents were classified according to the
nature of the accidents and by causes.
The accidents reported were derailments, collisions, trains running into vehicles, personal accidents, and miscellaneous. Each
group was subdivided into train and shunting accidents. Personal
events not resulting from disturbances of the operation were thus
included.
With the exception of boiler explosions, which were always
reported, only movement accidents were reported, t h a t is, those
arising by or during the movement of trains and rolling stock.
Accidents in shunting were reported only if they had resulted in
injury t o persons or material damage of more t h a n 1,000 gold
kronen. Miscellaneous accidents comprised all other events arising
in the course of the movement of rolling stock if they resulted in
injury to persons. The classification by causes was as follows j
defects in equipment, etc., defective execution of service (both
groups being further subdivided), disregard of regulations, running
at too high a speed, natural phenomena, misadventure, obstacles
on line, criminal acts (Bahnfrevel), and other causes.
Casualties. — Fatal casualties reported were those leading to
death either immediately or within twenty-four hours after the
accident occurred. Non-fatal casualties were registered if they
resulted in incapacity to work for more than fourteen days.
Casualties were classified as those due to own fault and those
due to fault of third persons.
Railway employees, in the sense of the statistics, were all persons
working for the railway concerned, even if they were not directly
remunerated by the railway administration concerned. Postal
servants were comprised among "other persons".
Statistics of accidents on minor railways were less elaborate
than those for main lines and local lines. As a fatal casualty was
considered any casualty leading to death, no period being fixed
within which death had to ensue. A non-fatal casualty was one
entailing incapacity to work for more than three days.
Measures of Risk
Accidents per 1,000,000 train km. (exclusive of personal events).
Casualties per 1,000,000 train km. were calculated for casualties
of all persons only, not for railwaymen separately.
1

Republication was planned for 1927.

— 51 —
BELGIUM
The statistics of accidents on the railways are published in the
form of a "Report on Operations" which appears annually in the
reports presented to the Legislative Chambers by the Minister of
Railways, etc.. entitled Chemin de fer et Service de l'Electricité.
From 1920 the statistics furnished have been considerably reduced.
The earlier and more detailed statistics are described here.
Scope
The statistics refer to State railways ; as regards private railways,
only a few figures are given. Mention is made of events in railway
transport properly so-called, and its maintenance, as well as in the upkeep of the permanent way, and of movement casualties in workshops.
Statistics refer to the following accidents (disturbing events) :
(a) Those which have caused serious material damage to trains,
with resulting death or serious injury to passengers or railway
employees, or which gave rise to an obstruction of the main roads for
more t h a n an hour, and caused the suspension (suppression) or
delay (détournement) of trains or displacement of passengers.
(6) Those which have similar consequences but of less importance, or which might have caused a serious accident (detached
cars, etc.).
Accidents. — These are classified as collisions and derailments,
and are therefore only accidents arising from train movements. The
apportionment is made according to the place in which the accidents
occurred. For collisions a distinction is, moreover, made according
to whether they are head-on collisions, or collisions between trains
running in the same direction, etc. The aggregate of accidents is
subdivided as follows :
(1) According to causes :
(a) defects of permanent way ;
(b) defects in rolling stock;
(c) faulty execution of service.
(2) According to whether they occur to passenger or to goods
trains, light engines, etc.
Each of these classifications is combined with that by place, etc.
In addition, the number of collisions and derailments which caused
injuries is indicated separately.
Casualties. — Mention is made of the number of employees
killed, injured or suffering from contusions. A fatal casualty is
one entailing instantaneous death or subsequent death in consequence of the injury, before the Commission of Enquiry charged to
examine the causes of the accident has terminated its report (i.e.
two to three days at most after the accident). Casualties are classified in two main groups according to whether they resulted from
accidents due to collisions and derailments of trains or from other
causes. Cases falling within the first of these groups are apportioned according to the place of the accident and the nature of
the collisions (see Accidents).

— 52 —
Other casualties happening to employees appear in a combined
table. The cases referred to here include those arising in connection with the movement of engines and cars or while attending to
apparatus (appareils) serving for the railway working (exploitation)
and for the running of trains, such as points, signals, appliances for
lighting of cars, for loading of goods, etc. The division into killed,
injured and contused is made according to responsibility on one
side, and according to the various duties in the course of which
the victim of the accident was killed or injured and by causes on
the other.
Employees. — In the case of collisions and derailments, account
is taken only of employees of the railway administration ; in the
case of other accidents, there are included, in addition, employees
of other administrations (postal servants, etc.), accidents to whom
appear separately. Employees repairing the permanent way and
buildings are counted among those of the railway administration.
On the other hand, among persons not connected with the service,
other than passengers, is found a group relating to casualties resulting from loading operations or assisting in shunting.
BULGARIA

The statistics of accidents on railways form part of the 'Statistics of Railways, etc.", published annually by the General Directorate of Railways and Ports, which forms part of the Ministry
of Railways, Posts and Telegraphs.
Scope
Account is only taken of normal-gauge lines, and movement
accidents occurring during railway transport, properly so called,
and its maintenance.
Accidents. — The classification resembles t h a t of the German
statistics. Mention is made of derailments, collisions, and other
accidents, apportioned in each case according to the kind of trains,
the causes and the places of the accidents. The group "other
accidents" only applies to accidents arising from movements of
trains and rolling stock. I t is, however, to be observed t h a t the
German statistics also include deaths and injuries which occur
even if the train service is not disturbed ; on the contrary, the Bulgarian statistics keep exclusively to accidents in the sense of this
Report, i.e. events disturbing the operation of the railway technically. Moreover, the Bulgarian statistics are more detailed than the
German in the sense t h a t they combine the kind of trains and the
places of accidents with the causes of the latter in the same table.
Casualties. — Persons who die of their injuries within twentyfour hours are included in the deaths. Persons suffering from contusions are included among the injured. The victims of operating
accidents due to train movements are grouped according to the
causes (as for the accidents proper). In addition, figures are given
for occupational risk during shunting, occupational risk in the train

— 53 —
service, and other causes. Here then, are cases not arising from
railway operation accidents, but resulting directly from train
movements or shunting operations. Lastly, the classification of
casualties is made according to the various groups of the railway
personnel. The personnel includes employees in the train service,
those engaged in the maintenance of the permanent way and employees of the administration, but does not include postal servants.
Measures of Risk
Accidents :
(a) Derailments 1 ( 1 ) per
(6) Collisions
J (2) per
Casualties :
(a) Killed
1 (1) per
(2) per
(6) Injured
j (3) per

1,000,000 train km.
10,000,000 axle km. of all wagons.
1,000 members of the personnel.
1,000,000 train km.
10,000,000 axle km. of all wagons..

CANADA
The Dominion Bureau of Statistics, Transportation Branch,
publishes annually railway accident statistics in the Reports on
Statistics of Steam Railways of Canada and Statistics of Electric
Railways of Canada. The reports are made from returns made by
the railways in accordance with the provisions of the Railway
Act, 1919, and the Statistics Act, 1918, certified by the controllers
of the railway companies or other officers. Up to 1921, railway
statistics were published for steam and electric railways together.
Accident statistics of electric railways are confined to a summary
of the employees killed and injured. The following analysis refers
to statistics of steam railways only.
Scope
Statistics cover all State and private railways and all kinds of
gauges.
Accidents in the sense of a material event causing death or injury
to several persons are not reported.
Casualties are reported as fatal if they result in death within
twentj'-four hours.
Non-fatal casualties to employees are only reported if preventing them from performing their service for more than three days
in the aggregate during the ten days immediately following the
event.
Casualties, both fatal and non-fatal, are classified as "accidents
resulting from the movement of trains" and "accidents resulting
from other causes than the movement of trains". The former
group (movement casualties) are subdivided by causes (coupling
and uncoupling, collisions, derailments, parting of trains, locomotives and cars breaking down, falling from trains or cars, jumping
on or off, struck by trains, etc., overhead obstructions, other

— 54 —
causes); and also by classes of employees (trainmen, trackmen,
switch tenders, etc.; stationmen, shopmen, telegraph employees,
other employees). Casualties not resulting from the movement
of trains are only classified by classes of employees. The classification differs from t h a t above quoted and runs as follows : trackmen,
stationmen, shopmen, other employees, postal clerks, etc. Finally,
accidents at highway grade crossings are given separately for trainmen and passengers in common.
Employees in the sense of the statistics are all persons employed
by the railway companies, including postal and telegraph clerks,
and freight handlers.
CZECHOSLOVAKIA
The Statistika Ceskoslovenskych Drah, published annually by
t h e Ministry of Railways, contains accident statistics.
Scope
The statistics include State railways and private railways, and
cover exclusively railway transport service properly so-called and
its maintenance.
Accidents. — A distinction is made between collisions and
derailments on the one side, and "other extraordinary events" on
the other. Collisions, grouped according to the place in which they
happened, are classified as collisions of trains, collisions of vehicles
with trains; then the cases are classified according to their cause.
Derailments are also classified by place and cause. The "other
extraordinary events" include irregularities in the movement of
trains which did not entail injury or material damage, e.g. cases
due t o the following causes : passing signals, running on wrong
line, collisions on same track, etc. All other extraordinary events
are grouped separately from those just mentioned, whether they
are accidents in train operating (movement), such as trains running
into vehicles at level crossings, or accidents which do not arise
from train movements, such as fire in buildings; criminal acts
against the railways; damage by flood; breaking of axles, wheels
and tyres; or damage preventing the subsequent running of trains.
Casualties. — Deaths of and injuries to railway employees are
classified according to the nature of the work done by the persons
injured and while walking on line. A special column, however,
relates to casualties resulting from railway operation accidents.
Statistics cover trainmen and train service employees. Figures
concerning deaths and injuries of postal servants in the course of
d u t y are furnished in other special columns.
There is no official definition of "fatal casualty"; in practice,
however, fatal casualties are those resulting in death within four
weeks from the date of the accident or personal event.
Measures of Risk
(1) Accidents and extraordinary occurrences : (a) per 100 km.
of line worked; (b) per 1,000,000 axle km.

— 55 —
(2) Deaths
train
(3) Deaths
train

and injuries of railway employees : per 1,000,000
km.
and injuries of postal servants : per 1,000,000
km.
FRANCE

Before the war French railway accident statistics were published
annually by the Ministry of Public Works aud of Transport,
Railway Department (Ministère des Travaux publics et des Transports, Direction des Chemins de fer), in the report entitled Statistique des Chemins de fer. These statistics were republished in
1928. Since 1924, the Director of the Railways Technical Operation Inspectorate in the Ministry of Public Works (Contrôle de
l'exploitation technique des chemins de fer au Ministère des Travaux publics) publishes an "Annual Report on the general
position with regard to accidents on the great French railways"
(Rapport annuel sur la situation générale des grands réseaux de
chemins de fer français au regard des accidents) in the two reports
Annales des Mines and Annales des Ponts et Chaussées.
^'ANNALES DES M I N E S " AND " A N N A L E S DES PONTS ET CHAUSSÉES"

Scope
The statistics cover the great railway systems of general interest
only, whether State or private railways.
Accidents. — Only accidents in railway transport properly
so called are registered, exclusive of accidents in workshops or
stores (dépôts). Serious cases only are reported, viz. :
(1) Train accidents resulting in death of persons or injury to
persons of whatever severity.
(2) Other accidents arising from railway operations properly
so called, if they resulted in injury entailing incapacity to
work of more than twenty days.
Accidents are classified as follows :
(1) Train accidents resulting in death or injury to persons.
(2) Accidents at level crossings.
(3) Other accidents arising from the operation of the railway,
such as train accidents involving serious damage to rolling
stock, etc. (matériel), coupling accidents, shunting accidents,
switching accidents, accidents while loading or unloading
wagons, and other personal events entailing incapacity to
work of more than twenty days, etc.
To sum up, statistics do not cover all accidents resulting in
injury, but include also personal events not connected with or
resulting from a disturbance of train operations.
In a combined table, train accidents are classified by their
nature into five groups and by seventeen main groups of causes.
The five former groups are as follows :

—

o e -

il) Side way collisions.
(2) Collisions, shocks, and trains or parts of trains running into
each other during shunting (tamponnements en manœuvre).
(3) Collisions of trains running in the same direction (tamponnements par l'arrière;, rattrapage de trains).
(4) Derailments.
(5) Miscellaneous (trains running away or going adrift) (emballements de trains ou dérives).
The classification by causes groups accidents according to fault
of railway servants and according to failure or defects of material.
Accidents at level crossings are given separately for guarded
crossings and crossings without gates (barrières) on the one hand,
and are grouped as accidents to persons on foot and to vehicles
on the other.
Casualties. — As noted above, all casualties arising from train
accidents and all other casualties entailing incapacity to work for
more than twenty days are reported '. Fatal and non-fatal casualties
are classified by the same groups as accidents, and casualties resulting from train accidents are also grouped according to the nature
and causes of the accidents. Both classifications, however, apply to
all persons injured only, not to railway servants separately. Casualties not resulting from disturbances of the operation of trains are
included without further distinction in the group "other railway
operation accidents".
Railway servants in the sense of the statistics include trainmen
and train service employees, but not office and workshop employees,
nor porters. As casualties of railway servants are counted only
those happening to servants on duty.
Measures of Risk
Rates of fatal and non-fatal casualties of railwaymen per
thousand railwaymen are calculated.
An attempt is made in the Annales to compare the results of
the French statistics with the corresponding statistics for the
United States, Great Britain and Germany.
" STATISTIQUE DES CHEMINS D E EER "

Scope
These statistics include the State Railways and private companies, both main and secondary lines.
The accidents are subdivided into : derailments, collisions, and
miscellaneous accidents. I n each of these groups the accidents
are classified according to the place at which they occurred.
The casualties are classified as follows : casualties in train accidents, in accidents of railway working other than train accidents,
and casualties from imprudence or misadventure. Each of these
categories is grouped under killed and injured.
1

A casualty is classified as "fatal" if death ensued immediately, or at
most a few days after the accident.

— 57 —
Measures of Risk
Casualties from train accidents :
Killed per 1,000,000 train km.
Injured per 1.000,000 train km.
TABLE I . — NUMBER OF FATAL AND NON-FATAL CASUALTIES OF
RAILWAYMEN IN TRAIN ACCIDENTS AND OTHER ACCIDENTS OF
RAILWAY WORKING, BY CAUSES, NUMBER OF RAILWAYMEN AND
RATES PER 1,000 RAILWAYMEN, IN FRANCE, 1 9 2 2 - 1 9 2 6 '
Casualties in

Year

Train
accidents

Accidents
at level
crossings

Other
accidents
of railway
working

All
accidents

Casualties
per 1,000
railwaymen

Number
of railwaymen

NonNonNonNonFatal fatal Fatal fatal Fatal fatal Fatal fatal
1922
1923
1924
1925
1926

20
8
15
15
12

84
49
68
79
107

—
1

—-

403
420
495
411
414

631
740
792
708
595

423
428
510
426
426

715
789
860
787
702

438,230
485,764
504,100
512,881
503,506

Fatal

Nonfatal

0.964
0.880
1.018
0.831
0.846

1.618
1.624
1.706
1.534
1.495

Annales des Mines, 1928, p p . 120-121.

TABLE I I . —
NUMBER OF FATAL AND NON-FATAL CASUALTIES OF
RAILWAYMEN IN TRAIN ACCIDENTS AND OTHER CASUALTIES IN
RAILWAY WORKING, AND CASUALTIES DUE TO CARELESSNESS
OR MISADVENTURE, IN FRANCE, 1 9 1 1 - 1 9 1 3 AND 1 9 2 6 L
Casualties in
Casualties due
accidents of railto carelessness
way working
R a t e per 1,000,000 other than train (imprudence)
or misadventure
train km.
accidents

Casualties in train accidents

Year

1911
1912
1913
1926

Number

Fatal

Nonfatal

35
15
16
14

162
139
113
121
1

Fatal

0.0875
0.0366
0.0387
0.0317

Non-fatal
0.4050
0.3393
0.2732
0.2740

Fatal

Nonfatal

Fatal

Nonfatal

6
14
5
2

21
31
21
76

366
352
393
413

562
586
598
673

Statistique des chemins de fer français.

GERMANY
Railway accident statistics are published annually by the
Ministry of Transport of the Reich in the report entitled Statistik
der im Betriebe befindlichen Eisenbahnen Deutschlands.

— 58 —
Scope
Statistics relate to all standard-gauge railways of the State
(federal) as well as the private lines. Both steam and electric
railwaj's are included.
Only "railway accidents and casualties" so called are comprised,
i.e. cases only which occurred by or during the movement of trains.
All non-movement cases are consequently excluded.
The scope of the statistics was greatly reduced in 1922. As to
accidents, all derailments and collisions involving trains are reported,
whether personal or material damage of property resulted or not.
Collisions and derailments of parts of trains are only reported if
either personal injury or damage of more than 1,000 gold marks
results. Explosions of locomotive boilers are reported in any case ;
fires in trains, if personal injury or damage to railway property or
to goods of more than 1,000 gold marks resulted. Trains running
into vehicles are reported where persons are injured, animals killed
or street vehicles damaged. All other events are only reported
if persons are injured.
Casualties reported are all fatal casualties resulting in death
or leading to death within twenty-four hours and all non-fatal
casualties disabling the worker for more than fourteen days or
leading to death after twenty-four hours.
Data and
Up to
collisions
the place

Sources
1921, accidents were classified by causes; derailments and
moreover according to the kind of trains involved and
where they occurred — at stations or on the line.

Casualties are classified :
(a) By categories of persons, viz. (1) railway employees and
railway workers on duty; (2) postal, telegraph, police, and
other employees on duty; (3) other persons, including railway employees and workers not on duty.
(6) Casualties of railway employees and railway workers on
duty were separately classified by causes u p to 1921. The
classification was different from t h a t of accidents (see
table III).
Casualties are now classified as those in derailments and collisions (combined) and "others".
All workers, whether working for the railway company itself or
for a contractor, provided they are injured during their work by
the working of the railway, are considered as "railway workers".
Workers in workshops are altogether excluded.
Measures of Bisk
Up t o 1921, the following rates were calculated :
(1) Accidents per 100 km. of track.
(2) Accidents per 1,000,000 car axle km.

— 59 —
(3) Casualties of railway employees and workers per 1,000,000
train km. : (a) fatal, (b) non-fatal.
(4) Casualties of railway employees and workers per 1,000,000
car axle km. : (a) fatal, (6) non-fatal.
(5) Casualties of all persons (passengers, railway employees and
workers, postal, telegraph, etc., employees and other persons
including railway employees and workers not on duty) per
1,000,000 train km. and per 1,000,000 car axle km.
These rates have now been discontinued and only accident rates
per 1,000,000 car axle km. are calculated.
F O R M B B CLASSIFICATION OF ACCIDENTS

1. — Derailments
A. — By Place and Rolling Stock
Passenger trains :
(a) on road,
(b) in station.
Goods trains :
(a) on road,
(6) in station.
Parts of trains engaged in shunting or single ears, etc. :
(a) on road,
(6) in station.
Total :
(a) on road,
(6) in station.
JS. — By Causes
Interruptions or obstacles on line.
Wrong train service.
Defects of rolling stock :
(a) breaking of axles,
(b) breaking of tyres,
(c) other defects.
Defects of superstructure.
Wrong or defective :
(a) setting of points,
(6) handling of other movable structures.
Other causes.
2. — Collisions
A. — By Place and Rolling Stock
Passenger trains 1 with other trains, parts of trains in shunting, and
Goods trains
} single cars, etc. :
(a) on road,
(6) in station.
Parts of trains in shunting and single cars :
(a) on road,
(fc) in station.
B. — By Cotises
Wrong orders of station service.
Wrong setting of points.
Defective signalling or disregard of signals.
Running into station at too high a speed.
Careless shunting or wrong placing of cars, etc.
Untimely moving of cars.
Uncoupling (broken train)
Other causes.

— 60 —
3. — Other
A.

Accidents

— By Place
(a) o n r o a d ,
(6) i n s t a t i o n .

B. — By Nature

Accident

R u n n i n g i n t o vehicles.
Fire in train.
Boiler explosions.
O t h e r a c c i d e n t s involving d e a t h or injury t o p e r s o n s .
TABLE i n . — NUMBER OF FATAL AND NON-FATAL CASUALTIES O F
RAILWAY EMPLOYEES AND RAILWAY WORKERS BY CAUSES,
AND RATES PER 1 , 0 0 0 , 0 0 0 TRAIN KM. AND P E R 1 , 0 0 0 , 0 0 0
AXLE KM., AND NUMBER OF FATAL AND NON-FATAL CASUALTIES OF POSTAL, TELEGRAPH, POLICE AND OTHER EMPLOYEES
BY CAUSES, I N GERMANY, 1 9 1 1 TO 1 9 1 3 AND 1 9 1 9 TO 1 9 2 1

a
CO

'S
<s
ö
'S
u

-p
tí
M

"0>

a
§1
OC®

•tím
-p u
tí c3
3 «
-G
œ

Whilst on line or passing line
not at proper time

Year

Carelessness in getting on or off
ears, etc., in motion or in cars, etc.

m
-p

_G
"ft
3
0
0

g
u
0

ft
3

-p
03

O
ü
-p
jß

-tí

Casualties
per

à0

u
a>
m *
_G
en
co
G
CD

a
S

a
"3

-p

^
cS
0
ki

<D
PH

-p

»-g
8 ¿s

©•

0
©

tí

-tí
-p

Carelessness in ascending or descending from cars, etc., in motion
or on railway premises

Postal,
telegraph, police,
and other employees on d u t y

Railway employees and railway workers

"3
-p
0

O
Fatal
1911
1912
1913
1919
1920
1921

36
24
30
67
48
27

64
82
92
110
77
66

55
75
71
109
68
121

98
120
120
180
109
81

Casualties

243
295
349
209
189
203

67
86
85
86
98
40

Non-Fatal
1911
1912
1913
1919
1920
1921

179
181
218
390
273
266

280
277
280
210
190
198

263
272
239
313
248
380

173
217
170
235
188
151

235
280
273
152
114
136

213
230
226
177
196
152

563
682
747
761
589
538

0.77
0.89
0.95
1.87
1.30
1.05

0.02
0.02
0.02
0.04
0.03
0.02

0.04
0.05
0.04
0.07
0.05
0.05

1
1
4
1
1

12
14
13
23
9
10

13
15
13
27
10
11

56
50
58
42
25
33

14
19
16
27
13
17

70
69
74
69
38
50

Casualties
1,343
1,457
1,406
1,477
1,209
1,283

1.83
1.91
1.79
3.64
2.68
2.50

— 61 —

GREAT BRITAIN
Railway accident statistics formerly published by the Board
of Trade are now published by the Ministry of Transport, in
annual Returns of Accidents and Casualties. A Report to the Minister
of Transport upon Accidents that have Occurred on the Railways of
Great Britain summarises and analyses this information.
Scope
Accident statistics cover all public railways in the United
Kingdom up to 1922, and in Great Britain only since 1922, including
London Tubes and other electric railways.
Statistics cover all accidents and casualties on "railway premises", viz. "stations, permanent way, goods yards and all other
premises used for working the railway, but not factories, workshops,
buildings used exclusively for warehousing goods, repairing sheds,
hotels and other similar premises". Both movement and nonmovement accidents are therefore included.
As to accidents, statistics include :
(1) Any accident attended with loss of life or personal injury.
(2) Any collision where one train is a passenger train.
(3) Any passenger train or part of a passenger train leaving
the rails.
(4) Any accident of a kind not comprised in the foregoing
description, but which is of such a kind as to have caused
or to be likely to cause loss of life or personal injury and which
may be specified in t h a t behalf by any Order to be made
from time to time by the Board of Trade.
As to casualties, fatal casualties are those in which the worker
is killed or so severely injured "as to die at any subsequent date
after the accident, but prior to the date of this report". All nonfatal casualties disabling the worker at least one whole day from
doing his ordinary work, are reported 1. Casualties to servants of
railway contractors are included.
Data and Methods of Computation
Accidents have to be reported, under section 6 of the Railway
Regulation Act, 1871, by the companies to the Ministry of Transport.
Accidents are only reported for the group "train accidents",
which are those involving trains, rolling stock or permanent way.
They are grouped as : (1) accidents to trains, and (2) accidents to
or failure of rolling stock or permanent way, and are classified by
"causes" and the kind of trains involved.
Casualties are grouped as those resulting from : (a) "train
accidents", (6) "movement accidents", and (c) "non-movement
accidents". Each group is classified by causes and railway
company (see tables IV, V, and VI).
1
Casualties of passengers or other persons have to be reported, however
alight the injuries may be.

— 62 —

Train casualties, movement casualties, and
non-movement
casualties to railway servants are, moreover, classified according
to the occupation of the worker involved and according to the
nature and location of the injury.
The railway employees whose railway casualties are included
comprise all persons employed by the railway companies, including
those employed on the Manchester Ship Canal, and not simply
workers on "railway premises" as defined above.
The Reports give summary figures for several years and periods.
A classification by so called "primary causes" is made for movement
casualties.
TABLE IV. — NUMBER OF FATAL AND NON-FATAL CASUALTIES OF
SERVANTS OF COMPANIES AND CONTRACTORS IN TRAIN ACCIDENTS, BY CAUSES, IN GREAT BRITAIN, 1 9 2 2 TO 1 9 2 7 1

MINISTRY OF TRANSPORT : Returns of Accidents and

Casualties.

^ CO ^ 4*- CC 09
M CO H CTI © M
rf^. C i co b * O i t—'

tf^ CO b« CTI C5 i—*

i—t CO CD < 1 OS CO

M ^ C O t O C f t Ol
CO Or ~ J CTt OD h-1

CD CO CD CO CO CO

to io to to io to
-a es o» ^ co bo

O H -

i—

Î
J t—i CO r£- b» CO

00 O l CD CD CD 00
bit t * 00 CO CO CTI

I to to I

CO < I C i CO C i CO

co Ui «-arf^—
i < £t

Coming in contact, whi
on vehicles, with other
etc., standing on adjace

w t o I co I 1

H(«»-aMCo

co co *>•#*.ri*»co

Coupling or uncoup
vehicles

*•• Cn cc Cn Cn Oi

CO CO >f»> Ot rf^ CO
GO CO 00 i—' CD 00

co to co co co to
h-• <I CO Ol Ci Ol

Year

CO CO CO CD CD CD
t-O t:-3 tO IO tO t>S

Passing over, under or
upon buffers
Getting on or off, or fa
engines, wagons, e
Braking, spragging or c
wheels

Attending to ground

Ol 03 t o 05 tO Ol

Moving vehicles by h
capstans, turntables

HHKJtôbSH
-a os os os o -J

Other shunting accid

1—' H-' h-• H-< * - H-'

to to to to to to
-a osCTif» co iss

HIHI

Year

H-» I—» t—' h -J 1—* h - '

co os to >— to to

Falling off trains in mo

H-fcOrf^GO OS bO

H- h-i h-i h-1
<£> O O 1— 1—'<£!

iss iss iss co Iss iss
0 0 CO CO If». *>. 00

-a co zo iss co H1 1 I IsS (t- If»H H M H H H
If». CO OSCTCTCT

oi e» co os œ tss

{—« *—i 1

Û 5 H H

££S-g3g
1 H - l - 1 CO

ssssgg

co if* #. ^ oi co
CO CO < I O SO ~ 3

co iss co oi co iss
© en o co oi iss

if»- o iss - j co co

ot co >f*. ^ tfk co

OO l— tsS O l Os

1—* 1—' t—' H*

co -q co -j oi o

OSTINO CO CO^ÇO^sS
"M O O " C T ' - J V OO
• J GO OS Ci CO OS
- J tsS h— If». 1— If»

II1§!1

Working on the perm
nent way, sidings, et
Attending to gates a
level crossings
Walking or standing on
Being caught betwee
vehicles
Falling or being caug
between trains a n d
platforms, etc.
Miscellaneous
All
Causes

O CO CD COCTCO
If». CO i— tsS CT O

Fatal Casualties

ÇC OS CO Î C CO * J

Non-Fatal Casualties

CT If* O l OS CTI If*
© 0 0 - J |f>. CO ISS
CO O l CO <D - J OS

Getting on or off engi
vans, etc., during runn
of trains
Coming in contact wi
bridges or erections
sides of line
Attending t o machine
etc., of engines in mot

— 65 —
TABLE VI. — NUMBER OP FATAL AND NON-FATAL CASUALTIES OF
SERVANTS OF COMPANIES AND CONTRACTORS IN NON-MOVEMENT
ACCIDENTS, BY CAUSES, IN GREAT BRITAIN, 1 9 2 2 TO 1 9 2 7
O «
fi O
O H

M> CO

Yeai

11
m
Ta

fi-5

fi §

8.2
SD-g

.3 "
>

°a

•* 2
cTO

ME
tí<2
© -^>

!» O

"-i

.3 s
'S S

Fatal

h .g

Casualties

1922
1923
1924
1925
1926
1927

10
8
S

9
11
13
Non-Fatal

1922 2,097
1923 2,528
1924 2,268
1925 1,963
1926 1,588
1927 1,535

o £

Total

1°

bo tí
0
CS

,- fi
S T3
-fi -S

45 fi

•s-a
fr

tí «l.

B< ©
t+4

" Pt
*'

O M

a '
=> s
•3
es
o
i-l

.,
co

g O

«'.3
"Ss

O
60

o
O

,¿4

1,036
1,328
1,379
1,387
1,207
1,099

33
38
35
42
37
44

Casualties

205 580 1,894 1,020 390 402 763
303 709 1,990 1,412 410 443 918
323 1,752 2,050 1,695 447 494 956
362 2,187 1,880 1,732 528 461 1,005
290 2,720 1,189 1,504 472 423 786
341 2,752 1,410 1,675 439 497 970

35
52
53
48
75
48

816 1,977 1,806 13,021
6842 352 2,330 15,467
855 2,677 2,241 17,195
934 2,7532,395 17,635
881 2,190 2,164 15,493
1,054 2,3632,447 16,631

HUNGARY
The Statistical Year Book contains statistics of accidents on
railways. Formerly, Hungary also collaborated in the publication
of the statistics of the "Verein deutscher Eisenbahnverwaltungen".
Scope
Statistics cover State railways as well as private railways. Main
lines and local lines are included. Only accidents in railway
transport properly so called are reported. Workshops, therefore,
are altogether excluded from the statistics, and accidents in work
on new buildings, etc., are included only if the accident occurred
in connection with the movement of a train or parts of a train.

— 68 —

Accidents. — Only accidents in railway working are taken into
account. They are classified by the nature of the accident, viz.
derailments, head-on and side-on collisions and miscellaneous accidents. The latter group contains also accidents not arising in
connection with movement of trains or rolling stock, the group
being made up as follows : running into vehicles, fire in trains,
boiler explosions and other events. The last group comprises also
personal events not connected with disturbances of train operations,
such as persons being run over, etc.
Casualties. — Only casualties, fatal and non-fatal, resulting
from accidents reported, are recorded. A casualty is classified as
fatal if death results before the obligatory report has been submitted,
or if death will doubtless ensue in a very short time.
Railwaymen. —• Only employees engaged on trains and in train
service are included.
Measures of Risk
The number of employees per (a) fatal and (b) non-fatal casualty
is calculated.
INDIA
Railway accident statistics are published annually by the Railway Department of the Government of India in the Report by the
Railway Board on Indian Railways.
Since 1924-1925 statistics are much the same as those published
for the United Kingdom. The main groups are as follows :
(1) Accidents to trains, rolling stock, permanent way, etc.
(2) Accidents caused by movements of trains and railway
vehicles, exclusive of train accidents.
(3) Accidents on railway premises in which the movement of
trains, vehicles, etc., was not concerned.
The first group, therefore, comprises railway operation accidents and casualties, while the second and third group contain
casualties not resulting from railway operation accidents. These
are grouped into movement and non-movement casualties. All
three groups are classified by detailed catises.
Movement casualties of railway servants on Class I railways
are moreover classified according to acts, or failures to act, of railway servants, or according to failures or defects of rolling stock, etc.
The classification reads as follows :
Misadventure and accidental; want of caution or misconduct by the
injured person; want of caution or breach of rules, etc., of other servants;
defective systems of working; dangerous places and conditions of work;
want of rules or systems of working; defective apparatus, appliances, etc.,
or want of safeguards, etc.

Percentages of all movement casualties by these causes are
moreover calculated.
No measures of risk are given.

— 67 —

ITALY
The State Railways administration furnishes material for two
different sets of statistics. One is published by the Ministry of
Transport under the title Relazione per l'anno finanziario (Report
for the Financial Year). The other, emanating from the Central
Statistical Office, is entitled Statistica dell'esercizio (Statistics for
the Civil Year), of which a summary is given in the Statistical
Annual.
Scope
The statistics of accidents are limited to railway transport
properly so called and its maintenance.
Accidents (Relazione). — These are classified under : derailments,
collisions and miscellaneous accidents; collisions and derailments
are subdivided in their turn according to whether they occur on
the line or in stations. In the case of collisions, a further distinction is made according to whether they occur at a guarded level
crossing or an unguarded one. Collisions include encountering and
overthrowing obstacles. The miscellaneous accidents comprise,
apart from those due to shunting and to outbreak of fire (along the
line and in stations), personal events without damage to property,
i.e. falls from trains, outrages and suicides.
The classification adopted in the second publication {Statistica)
is much more detailed. Principal groups : causes and effects of
accidents, the second group referring evidently to the nature of
the accident. The "causes" are divided into defects in locomotives,
wagons and permanent way; faults in train and shunting service,
and various causes such as negligence, etc., outrages, atmospheric
influences, and unknown causes. The "effects" are classified as
material damage to rolling stock or fixed property, obstruction of
the permanent way, encountering and overthrowing obstacles,
fires (in trains and in stations), derailments, collisions, falls from
trains. Each of these groups is subdivided in its turn into several
categories. Accidents reported are thus mainly train movement
accidents, but certain cases not arising from train movements
are included, such as accidents due to railway operations, properly
so called, and also a number of personal cases not arising from
railway operations.
Derailments and collisions during shunting operations are only
mentioned if they result in :
(1) a fatality or a non-fatal casualty;
(2) important material damage ;
(3) an interruption of the service of at least thirty minutes in
the case of passenger trains, and of at least an hour for goods
trains, independently of delays due to other causes.
Casualties (Relazione). — Deaths of, or injuries to, employees are
grouped according to the causes of the death or injury : accidents,
carelessness of the personnel, criminal acts (attentali e aggressioni),

— 68 —
other causes. Moreover, insurance statistics of casualties are
published, containing cases of injury and the days of sickness.
Cases are grouped according to severity.
(Statistica.) Injuries which result in death within twentyfour hours are counted as deaths. An injury is counted as such if
it is not healed within five days.
The classification of casualties in the Statistica is the same as
t h a t in the Relazione. Casualties resulting from accidents (causa
degli accidenti) are given only in so far as the accident is a "cause
arising directly from duty" (causa diretta del servizio).
Employees (Statistica). — Employees who are killed or injured
other than in the course of duty are counted in the statistics as
persons foreign to the service. Statistics cover trainmen and train
service employees (on line and in station), as well as office employees
and permanent way men and inspectors.
Measures of Risk
- (Statistica.)
Individual cases (killed and injured) in groups
according to causes are classified :
per 100 km. of line;
per 100,000 train km.;
per 100,000 locomotive km. ;
per 100,000 axle km.
TABLE VII. — NUMBER OP FATAL AND NON-FATAL CASUALTIES OI
RAILWAYMEN, BY CAUSES, IN ITALY, 1 9 2 2 - 1 9 2 3 TO 1 9 2 7 - 1 9 2 8

Year

Railway
accidents

Carelessness Ox
persor-nel

Criminal
acts

Other
causes

All
causes

NonNonNonNonNonFatal fatal Fata] fatal Fatal fatal Fatal fatal Fatal fatal
1922-1923
1923-1924
1924-1925
1925-1926
1926-1927
1927-1928

22
10
19
15
16
26

221
131
161
165
153
63

37
34
45
37
39
25

199
195
172
180
262
183

1
2
1

26
14
24
12
12
4

19
15
16
26
28
38

137
193
180
185
200
208

79
61
80
79
83
89

583
533
537
542
627
458

JAPAN
The Statistical Summary of the Empire furnishes annually
certain statistics concerning railway accidents, namely, the number
of accidents and the number of employees, workers, etc., killed
and injured.
Scope
The statistics refer to the State Railways and private lines.
Accidents. — These are mentioned in so far as they take place
on railways and in connection with train traction. The figures
seem therefore to extend to accidents following from the movement
of trains and those happening apart from the movement of trains.

— 69 —
NETHERLANDS
The Statistical Annual furnishes statistics of railway accidents
and indicates the reports of the "Bikstoezicht op de Spoorwegen"
as its source. These reports have been stopped since 1921. Before
the war the Netherlands were included in the statistics of the
"Verein deutscher Eisenbahn Verwaltungen".
Scope
The statistics cover the State Eailways and other public lines,
main as well as local lines. They refer to railway transport service
properly so called, but not to workshops, building of new lines, etc.
Only accidents and casualties resulting from the movement of
trains and rolling stock are included.
Accidents. — Accidents are classified as derailments, collisions
and other accidents, comprising yard-shunting accidents resulting
in bodily injury, serious material damage or serious disturbances
of service.
Casualties (individual cases). — Accidents to employees are
classified as follows : "fault of service", "own fault", and "during
shunting in stations", and are grouped in killed and injured. Fatal
casualties are those resulting in death before the report of the
enquiry into the causes of the accident is made.
NORWAY
Railway accident statistics are reported in the yearbook Norges
Jernbaner, published by the "Hovedstryret for Statsbanene" in the
series entitled Norges O/fisielle Statistihk.
They are for the years
beginning 1 July and ending 30 June.
Scope
The statistics cover all accidents and casualties arising from
operations connected with railways, and the conception does not
only extend to movement and non-movement accidents on railway
premises properly so called, as in Great Britain, b u t also to annexed
premises and operations, as shunting in gravel pits and work yards,
work in gravel pits, work in shops, cleaning of material, loading
and unloading goods in warehouses, etc. Accidents and casualties
are then divided into two groups :
(1) Accidents and casualties in the running of trains, i.e. in
railway transport properly so called, including train service
and shunting service 1.
(2) Casualties in other railway work.
This division does not entirely correspond with the division into
movement and non-movement risks, as the first group only includes
accidents and the resulting casualties occurring in the running of
trains engaged in ordinary traffic or additional trains, and the
1

"Under den egentlige jernbanedrift."

— 70 —
accidents in the shunting of such trains. The second group (other
railway work) therefore comprises casualties occurring in the running of work trains, trains engaged in ballasting, driving snow
ploughs, etc., as well as casualties in the maintenance of the lines
and the like, which may also be movement casualties.
Accidents are grouped into "accidents to trains" and "other
accidents", covering together the "accidents in railway transport
properly so called" of group (1). Accidents to trains are subdivided
into derailments, collisions, and miscellaneous accidents. "Accidents to trains" include only such accidents "as have occasioned or
by their nature might have occasioned danger to persons or material,
and accordingly every derailment, breaking of axles, etc., but not
accidents which only caused delay, such as boiler-tube explosions . . . etc.", except where they result in material damage
or personal injury. Damage of less than 10 kroner is not taken
into consideration. Serious damage is damage of at least one-tenth
of the cost of the material.
"Other (than train) accidents" comprise accidents in train and
shunting service not included among train accidents, if they result
in personal injury or, as regards shunting accidents, in material
damage.
Casualties are classified as fatal and non-fatal, a casualty being
fatal if it results in death within twenty-four hours, while non-fatal
casualties include only those in which the injury required medical
treatment.
Casualties in the "running of trains" (railway transport properly
so called) and in other work are kept apart.
Data and Methods of Computation
Train accidents are classified by causes and place. Other
(than train) accidents in railway transport properly so called are
grouped as "accidents in the running of trains involving persons"
and "accidents in shunting service".
Casualties in railway transport properly so called, i.e. in train
and shunting service, which are those resulting from the accidents
of group (1), are given separately for railway servants in train and
shunting service, which include only persons engaged in running
or shunting trains and employees working in stations or on the line
in connection with such operations on the one hand, and "servants
on d u t y " on the other hand, the latter comprising railway employees
"travelling or being otherwise on the railway in connection with
service but not engaged in running or shunting trains", such as
persons engaged in inspection, if killed or injured in or near a train
or during shunting operations. The former group (casualties of
railway employees in train service and shunting service) is classified
by causes (see table).
Employees not on duty and customs officials travelling free,
if injured in railway transport accidents, are included among
"other persons and employees not on duty".
Casualties happening in other railway work, that is, outside
railway transport properly so called, are given for all railway

— 71 —
servants and workers on duty together, and are not furti;er subdivided.
Measures of Risk
Casualties of railwaymen in train and shunting service, per
1,000,000 train km. : (a) fatal, (6) non-fatal.
TABLE VIII. — NUMBER OP FATAL AND NON-FATAL CASUALTIES OF
RAILWAYMEN IN TRAIN AND SHUNTING SERVICE, BY CAUSES,
AND RATES PER 1 , 0 0 0 , 0 0 0 TRAIN KILOMETRES, AND NUMBER
OF FATAL AND NON-FATAL CASUALTIES OF SERVANTS ON DUTY,
IN TRAIN ACCIDENTS AND OTHER ACCIDENTS OF RAILWAY
TRANSPORT, IN NORWAY, 1 9 2 2 - 1 9 2 3 TO 1 9 2 6 - 1 9 2 7

—
—

—

—•

—
—
—
—

1
3
1

—

O +a

o) ,2

co
ob

—

3
1

—
1

—
—

4
2

0.33
0.16

—

—•

0.16

5
7
4
2
2

14
15
12
15
12

1.16
1.24
0.97
1.19
0.95

3
3
1
1
4

Casualties

Non-Fatal
1922-1923
1923-1924
1924-1925
1925-1926
1926-1927

Fatal
1922-1923 |
1923-1924
1924-1925
1925-1926
1926-1927

on duty

Other causes

auses

3
Mfi

Other causes

ilst sta riding on
passuig line

i—i

ther ca uses

fi°.2

Sh2

couplin g cars

o ¡î

ncoupli ixg cars

t-i

r off trai
hilst pas

n train accident

Year

o
o

alties p
train ki

In
m fi shunting service

-S»

Casualties

i train accidents

Casualties to railwaymen in train service and shunting
service

—
—
—
—

4
1
2
6
5

Casual ties
4
4
2
10
6

—
2
1

—

—
1

—
—

1
2
1
7
3

POLAND

The Ministry of Railways has published, since 1922, a Statistical Annual of the Polish State Railways for the Administrative
Fear, which contains statistics of accidents.
Scope
The statistics are established separately for normal-gauge and
for narrow-gauge lines, and include accidents and casualties due
to movement of trains and rolling stock.

— 72 —
Accidents. — These are classified under : collisions, derailments,
miscellaneous accidents, and subdivided in each case into accidents
happening along the line and in stations.
Casualties are divided as such into : collisions, derailments, collisions daring shunting, and miscellaneous accidents. Information
is also furnished concerning the amount of property damaged.
Measures of Risk
Accidents :
(a) On the line : per km. of line.
(6) In stations : per station.
Casualties :
per 0 0 0 0 0 0 train km

(6) £"urtd 1

Property damaged : per 1,000,000 train km.
PORTUGAL
The Statistical Annual publishes statistics of accidents on railways.
Scope
The statistics cover State and private railways.
Accidents. — Movement accidents only are reported, while
casualties include also those of workers in workshops, repairing
sheds, etc. Accidents are classified as derailments, collisions and
miscellaneous accidents.
The derailments and collisions include those of wagons or carsengaged in shunting.
"Miscellaneous accidents" comprise all accidents arising through
the movement of trains, whether railway operation accidents as
defined in this Report or personal events involving injury to passengers, railway servants or other persons only (accidents while
ascending or descending moving trains, falling from trains, hit by
wagon doors, run over or caught by trains in stations, at level
crossings or on lines, etc.).
Casualties. — Fatal and non-fatal casualties reported are those
resulting from the accidents registered, t h a t is, casualties in railway
operation movement accidents and others occurring during the
movement of trains. They comprise therefore all movement
casualties. Moreover, casualties in workshops and repairing sheds
are reported separately.
Casualties are tabulated in three groups :
(1) Casualties in workshops and repairing sheds (Nos Officinas) ;
(2) Casualties (resulting from railway operation accidents) in
other services;
(3) Persons injured during the running of trains (on or by trains
in motion).
The two latter groups therefore comprise the casualties resulting
from the accidents reported in the three groups above-mentioned
and no additional cases.

— 73 —

The workers whose casualties are reported are therefore all
railway workers whether in the train service, working on permanent
way, etc., on the one hand, and all workers in workshops and repairing sheds, such as metal workers, smiths, carpenters, painters,
plumbers, etc., on the other.
RUMANIA
Rumania was formerly included in the statistics of the "Verein
deutscher EisenbahnVerwaltungen". The Rumanian Statistical
Year Book furnishes summary statistics of accidents on railways.
Since 1925, more detailed statistics have been published.
Scope
The statistics include the State Railways, which have acquired
most of the private lines.
Accidents. — In the number of accidents are also included cases
of death and injury not arising from an event causing an interruption in the service (disturbance of operation).
Accidents are classified, first of all in three kinds of collisions
— collisions with stationary objects, head-on collisions, side-on
collisions — and in derailments. These cases seem to represent
the accidents arising from the nature of railway work, as they are
added together.
Mention is also made of other kinds of accidents, which are distinguished as : wagons moved by wind, breaking of trains, breaking
of axles, unforeseen non-stops, unforeseen stops, trains running
into miscellaneous objects, falls from trains, fires, covering of snow,
other accidents, floods. These accidents are combined :
(1) with the place a t which they happen, either along the line,
or in a station, and in the case of the latter (a) on the arrival
and departure of the train; (b) during shunting;
(2) with the causes, according to whether they arise from defects
in the line or in rolling stock, irregularity in operating, causes
foreign to the railways (atmospheric accidents, force majeure,
own fault, criminal acts), unknown causes.
Casualties. — Casualties are classified by the nature of the accidents from which they result (same classification as for accidents)
and according to whether they are the results of faults in railway
working or other causes. The two classifications are combined.
Each of these two groups is subdivided into deaths and injuries.
SWEDEN
Railway statistics in Sweden are published by the administration of the Swedish Government Railways in the annual reports
Statens Järnvägar (State Railways) and Allmän
Järnvägsstatistik
(General Railway Statistics, containing figures for the different
railway systems of the country, with a French translation of the
table headings). Both reports are published in the series Sveriges
offideila Statistik (Swedish Official Statistics).

— 74 —

Scope
Swedish statistics cover only "movement" accidents and casualties, i.e. those occurring in connection with the running of trains
and shunting. Not included, therefore, are casualties occurring in
repairing sheds, in coupling or uncoupling cars at rest, in getting
on or off vehicles at rest, in handling turntables, signals, etc., while
loading or unloading, etc., where no movement of rolling stock
takes place.
Accidents. — Accidents are classified as derailments, collisions,
and other accidents. All derailments involving locomotives or
cars intended for train service are reported under this heading,
but not those of bogies, trollies, etc., and single cars derailed b u t
replaced on rail without serious damage. In shunting service, only
derailments which cause serious damage to persons or material
are reported.
Collisions are such between vehicles intended for train service
only; collisions between such vehicles and bogies, trollies, etc., or
other objects are therefore not included. Collisions in shunting
service are reported only where they result in serious personal or
material damage.
"Other accidents" include all locomotive boiler explosions;
trains running into road vehicles, if serious injury results, animals
are killed or vehicles destroyed ; fire in trains if resulting in serious
damage to persons or material and all other accidents resulting in
serious personal injury : derailments and collisions not reported
under these headings; running over persons; accidents while stepping on or off moving vehicles, and others resulting in injury.
Serious material damage means damage of at least 1,000 kronor.
Casualties reported are those resulting from movement accidents,
provided they result in death or injury incapacitating for work for
at least fourteen days.
Data and Methods of Calculation
Accidents and casualties have to be reported by the district
officers of the State Railways and by the managers of the private
railway companies.
Accidents, i.e. the three groups collisions, derailments and "other
accidents", are further subdivided by causes as well as by place
(on road or at station), and "class" of service (train and shunting
service). The resulting casualties are given for all persons involved,
not for railwaymen separately.
The classification of accidents by causes is given in the text
(page 37).
Casualties are classified separately by causes. The classification is given in table I X .
The railwaymen covered by the statistics include not only the
ordinary and temporary staff employed by the railway, but all
workers carrying out any work for the railway, whether directly
paid by the railway or by contractors.

— 75 —
Measures of Risk
(1) Accident rates per 100 km. line for all accidents.
(2) Accidents per 10,000,000 car (axle) km., for all accidents.
(3) Casualties per 1,000 railwaymen (exclusive of workers in
shops, etc.) :
(a) fatal,
(6) non-fatal.
(4) Casualties per 1,000,000 train km. :
(a) fatal,
(6) non-fatal.
(5) Casualties per 10,000,000 car (axle) km. :
(a) fatal,
(6) non-fatal.

TABLE IX — NUMBER OF FATAL AND NON-FATAL CASUALTIES OF
RAILWAYMEN IN TRAIN AND OTHER MOVEMENT ACCIDENTS, BY
CAUSES, AND RATES PER 1 , 0 0 0 RAILWAYMEN, PER 1 , 0 0 0 , 0 0 0
TRAIN KM. AND PER 1 0 , 0 0 0 , 0 0 0 CAR AXLE KM. IN SWEDEN,

1920 TO 1926

1
1
1

—
1

1
1
2
1
4

4
2
2
1

—
—
—
—
—

—

2
1

Non-Fatal
1920
1921
1922
1923
1924
1925
1926

2
4
2
3
1
2
3

23
16
10
10
13
13
22

46
25
46
62
51
59
42

26
31
32
19
29
27
33

Otherwise durin
service on duty

Falls from train

Walking on the
line carelessly

In coupling car
1
4

Per 10,000,000
car axle km.

—

Casualties

Per 1,000,000
train km.

1
—

—

Fatal

¡so .

Per 1,000
railwaymen

—
—
—
—

All causes

1920
1921
1922
1923
1924
1925
1926

In shunting car
or trains

Year

In derailments a
collisions

In getting on or off
vehicles in motion

Casualties

3
3
1
4
4
4
4

18
10
5
7
7
8
10

0.57
0.36
0.19
0.28
0.27
0.32
0.39

0.75
0.48
0.23
0.31
0.29
0.32
0.38

0.20
0.13
0.06
0.08
0.08
0.09
0.10

38
19
24
17
10
21
25

142
99
115
113
109
123
131

4.50
3.59
4.48
4.49
4.26
4.91
5.15

5.88
4.78
5.26
4.96
4.55
4.91
4.94

1.58
1.33
1.42
1.35
1.23
1.33
1.32

3
1

—
—
—
—
—
Casualties

5
1

—

1
3
1
3

2
3
1
1
2

—
3

— 76 —

SWITZERLAND
Railway accident statistics are published in three different
reports :
(1) The Federal Post and Railway Department publishes
annually a report entitled Schweizerische
Eisenbahnstatistik.
(2) The financial report (Geschäftsbericht) of the Federal Council
(Bundesrat) contains accident statistics.
(3) The Tableaux statistiques, annexes au rapport de gestion de
la Direction générale pour l'année . . . , published by the
Federal Post and Railway Department, contain some statistics on accidents on the Federal Railways.
Scope
The first report refers, as far as the subject of our Report is
concerned, to all railways, standard gauge and narrow gauge,
and cog railways —• federal as well as private lines, whether with
«lectric or steam traction, being included.
The financial report combines accident statistics for railways
as above defined, as well as tramways and rope railways.
The statistics cover movement accidents and casualties, i.e. those
connected with the movement of trains or cars, as well as casualties
due to electricity (electric current).
Accidents reported are those arising in the running of trains
properly so called.
As derailments are registered all derailments of passenger
trains; derailments of other trains and trains engaged in shunting
are registered only if they result in serious injury to persons or
serious damage.
Collisions registered are all collisions of passenger trains with
other trains or single cars, and those collisions of other than passenger trains and of trains engaged in shunting which result in
serious injury to persons or serious damage.
"Other accidents" are all events involving trains or single cars
resulting in serious injury to persons or serious damage, including
personal events arising in connection with the movement of trains
or cars, such as accidents while crossing the line, etc., and in addition those due to contact with electric wiring (Fahrleitungen).
Casualties. — The casualties registered are those arising in connection with the movement of trains or due to electric current.
F a t a l casualties reported are those injuries resulting in death
immediately or within twenty-four hours after the occurrence of
the accident.
Non-fatal casualties are casualties entailing incapacity to work
for more than fourteen days (up to 1925, more than six days).
Data and Methods of Computation
Schweizerische Eisenbahnstatistik. — Accidents are classified as
derailments, collisions and "other accidents" : the former two groups

— 77 —

are further subdivided as those "in stations" and those "on the line ".
"Other accidents" are distinguished as those resulting in injury to
persons and other accidents. Serious damage means damage of
more than 1,000 francs.
Casualties of railwaymen are classified as casualties in derailments and collisions, on the one hand, and casualties in other
accidents, on the other.
The railway employees covered are all persons employed by the
railway administration.
Measures of Bisk
(1) Casualties of railway employees per 100,000 locomotive km.:
(a) fatal,
(b) non-fatal.
(2) Casualties of railway employees per 1,000,000 axle km. :
(a) fatal,
(b) non-fatal.
Geschäftsbericht des Bundesrates. — Accidents are classified as
above. Moreover, derailments are classified as under :
(1) Defects in the permanent way {Bahnanlage).
(2) Obstacles on line.
(3) Wrong setting of points.
(4) Defects in rolling stock.
(5) Other and unknown causes.
Collisions are subdivided as follows :
(1) Defective orders, communication, or wrong acts of staff.
(2) Defective signalling or setting of points.
(3) Disregard of regulations in shunting and carelessness of
trainmen.
(4) Wrong placement of vehicles.
(5) Other causes.
Moreover, accidents to moving trains, and in shunting, are
distinguished.
Casualties are classified as above. Moreover, casualties from
accidents other than derailments and collisions are subdivided by
causes as follows :
(1) Collision with motor-cars or street vehicles.
(2) Slipping on moving cars, etc.
(3) Getting on or off moving cars, etc.
(4) Walking on or crossing line.
(5) Carelessness in moving train and during shunting movements.
(6) Misadventure in shunting service.
(7) While coupling or uncoupling moving cars, etc.
(8) Forbidden or wrongly executed shunting movements.
(9) Opening and closing of wagon doors.
(10) Injury to eyes by (flying) particles, etc.
(11) Injury due to electricity in service.
(12) Other causes.
6

— 78 —
In the "Statistical Tables" of the Federal Railways, we find
as to accidents, that derailments and collisions are classified by the
kind of trains involved, and by secondary causes.
For casualties, the following rates are calculated :
(1) Rates per 1,000,000 train km.
(2) Rates per 1,000,000 locomotive km.
(3) Rates per 1,000,000 axle (car) km.
(4) Rate per 100 of total staff engaged in railway service properly so called.
UNITED STATES
Statistics of accidents on steam railways are published annually
by the Interstate Commerce Commission, Bureau of Statistics, in
the Accident Bulletin, and monthly in a Summary of Accidents
Reported by Steam Railways.
Moreover, the United States Department of Labour publishes summary statistics of railway accidents —
together with those of other industries — in the Statistics of Industrial Accidents, which appears irregularly. A report entitled Statistics of Railways, 1906-1916, published by the Bureau of Railway
Economics in 1918, contained a summary of accidents.
Statistics of accidents on electric railways are registered in the
report of the Department of Commerce, Bureau of the Census,
Electric Railways, which contains very summary accident statistics
only.
In the following résumé reference is only made to the Accident
Bulletin as providing the main regular reports and sources of
information summarised elsewhere.
Scope
Accident statistics refer to all steam railways of the United
States. The more detailed data, however, are reported for Class I
railroads only, viz. those having annual operating revenues above
1,000,000 dollars.
As to accidents, the report in the Bulletin covers all accidents
arising from the operation of a railway, in connection with the
operation or movement of trains, cars and locomotives, which result
either :
(1) in damage to railway property in excess of 150 dollars;
(2) in death or injury to persons as defined.
As to casualties, a fatal casualty is one resulting in death
within twenty-four hours after the accident occurred. Non-fatal
casualties therefore include casualties leading to death later than
twenty-four hours after the accident and those casualties which
entail injury to the employee disabling for more t h a n three days
within the ten days following the accident. The report publishes
all fatal and reportable non-fatal casualties resulting from the
"operation of the railway", not only those arising in connection with
the operation or movement of trains, cars or locomotives. The
"operation of the railway" covers all operations performed "under

— 79 —
the supervision of the carrier by its employees". Statistics include,
therefore, accidents and casualties on tracks, in stations, depots,
freight houses, warehouses, shops, coaling stations, wharves and
ferries as well as accidents in the maintenance of equipments,
roadway, track, buildings, etc., and the construction of "additions,
betterments, and new lines, if this construction is performed by the
carriers' regular forces of employees ordinarily engaged in the construction, operation or maintenance of existing lines".
Statistics, therefore, cover movement and non-movement
casualties, the latter including also those of workers not directly
engaged in the working of the railway.
Data and Methods of Computation,
Accidents have to be reported by the railway companies under
the Accident Reports Act of 1910 to the Interstate Commerce
Commission.
Accidents. — Accidents are grouped as train and train service
accidents :
(1) Train accidents are accidents with or without casualties,
arising in connection with the operation or movement of
trains, locomotives or cars and resulting in damage to equipment of other railway property in excess of 150 dollars.
(2) Train service accidents are those arising in connection with
the operation or movement of trains, locomotives or cars
and resulting in reportable casualties to persons, but not in
damage in excess of 150 dollars.
Train accidents are classified by the nature of the accident (see
table X), and the main groups of this classification are subdivided
by negligence of employees, defects in or failure of equipment,
defects in or improper maintenance of way and structures, and
miscellaneous causes, as well as by kind of train involved, all three
classifications being combined. Each of the groups of causes
enumerated above is subdivided by very detailed specific causes.
Train accidents, by nature and causes, are given for districts.
Train service accidents are classified by groups of causes. Highway
grade crossing accidents are classified by the kind of protection
afforded at crossing and the kind of vehicles involved.
Train and train service accidents are specified by kinds of trains
involved.
Casualties are grouped as casualties in train accidents, train
service accidents, and as non-train casualties. Non-train casualties
are those not caused directly by the operation or movement of
trains, locomotives or cars, resulting in reportable injuries or death
to persons.
Casualties in train accidents are grouped by the nature of the
accidents and by the same causes as train accidents. The classifications are given for trainmen and other employees separately. The
detailed classifications by specific causes are also given for trainmen
and other employees, but not again specified by the nature of the
accident (collision, etc.).

— 80 —
Train casualties are, moreover, grouped by the nature of the
injury received, by the nature of the accident and by certain occupations of trainmen, in a combined table.
Train service casualties are classified like train service accidents
for trainmen and other employees separately. The detailed specific
classifications are subdivided by those occupations of trainmen
which are mainly involved in the specified group of casualties.
Train service casualties are also classified by nature of injury
in combination with the main groups of causes and certain occupations.
Non-train casualties are grouped by causes.
All casualties, train, train service and non-train casualties, are,
for each group separately, classified in greatest detail by occupations, in combination with main causes. Main groups of occupations are moreover classified by roads, for train and train service
casualties combined, and for non-train casualties, by districts.
The number of employees in service is taken from the monthly
reports of the Service and Compensation of Steam Railway
Employees, as is the number of man-hours worked.
(a) The number of employees is based on monthly statements
(middle of month), counting the employees usually on duty on the
day of the count as well as employees under pay on vacation or
sick leave and "extra" men in train and engine service. In the
classification of occupations and groups of occupations, employees
whose duties are such as to make them capable of being included
in two or more "reporting divisions" are included in t h a t division
indicated by the greater part of their time during the month.
(6) The man-hours are reported in the same way (they include
normal hours and overtime hours). In computing the man-hours
for each occupation, however, the time of each employee is distributed among the occupations he may be alternatively engaged in,
each occupation being credited with the hours the worker actually
spends in t h a t work.
Measures of Bisk
(1) Casualties per 1,000,000 locomotive miles; in all train and
train service accidents :
(a) fatal,
(6) non-fatal.
(2) Casualties per 1,000,000 man-hours (all causes) by occupations :
(a) fatal,
(6) non-fatal.
(3) Casualties per 1,000,000 man-hours (all causes) by roads and
occupational groups :
(a) fatal,
(6) non-fatal.

— 81 —
(4) Number of employees per employee :
(a) killed,
(b) injured.
(5) Number of trainmen per trainman :
(a) killed,
(b) injured.
(6) Trainmen killed per 1,000 trainmen in train and train service
accidents by :
(a) yard
]
(6) road freight
[ service.
(c) road passenger J
(7) Train accidents per 1,000,000 locomotive miles by causes.
(8) Casualties to trainmen per 10,000,000 train miles in train
and train service accidents :
(a) fatal,
(6) non-fatal,
for freight trainmen and passenger trainmen separately.

TABLE-X. —
NUMBER OF FATAL AND NON-FATAL CASUALTIES OF
EAIWAY EMPLOYEES ON DUTY IN TRAIN ACCIDENTS, BY CAUSES,
IN THE UNITED STATES OF AMERICA, 1 9 2 3 TO 1 9 2 7

Collisions
Year
Fatal

1923
1924
1925
1926
1927

112
85
84
104
65

Derailments

Locomotivo boiler
accidents

Other locomotivo
accidents

Miscellaneous

Total

NonNonNonNon- Fatal NonNonFatal
Fatal
Fatal
Fatal
fatal
fatal
fatal
fatal
fatal
fatal
940
709
696
840
547

115
97
121
64
82

839
652
616
596
513

42
24
15
14
25

57
45
42
34
21

15
13
17
17
13

6
10
11
8
22

88 275 1,939
58 216 1,477
112 232 1,483
102 190 1,589
117 194 1,211

— 82 —

TABLE X I . - - NUMBER OF FATAL AND NON-FATAL CASUALTIES OF
RAILWAY EMPLOYEES ON DUTY IN TRAIN SERVICE ACCIDENTS,
BY CAUSES, IN THE UNITED STATES OF AMERICA, 1 9 2 3 TO

Miscellaneous

Struc k or run-over, not
classi fiable in preceding
columns

High way grade-crossing
accidents

Gett ing on or off cars
c>r locomotives

!-t

Coming in contact with
fixed structures while on
movin g cars or locomotives

a,
O

ating hand-brakes

erating switches

Opei ating locomotives

Coup ling or uncoupling
air hose

Year

Coup ling or uncoupling
loc omotives or cars

1927

Total

403
310
331
325
306

1,288
976
996
1,102
972

o
Fatal Casualties

1923
1924
1925
1926
1927

103
72
64
64
48

27
21
19
20
21

31
20
14
11
22

1
3
3

—
1

24
34
33
29
32

47
36
38
45
47

29
30
29
48
27

522
388
400
498
400

8,096 100
6,564 97
6,680 96
7,165 104
5,665 89

793
707
709
768
629

101
62
65
62
68

Non-Fat il Casualties
1923
1924
1925
1926
1927

1,954
1.592
1,538
1,591
1,325

520
430
452
455
398

8,043 1,084 2,571
5,877 1,050 2,042
5,458 930 2,229
5,697 1,142 2,474
4,467 802 2,158

859
730
724
665
590

13,517 37,537
11,608 30,697
11,946 30,762
12,276 32,337
10,592 26,715