M D UC-NRLF r2 ^ ill iiiiiii M#73 lillllllllll SB 4b lMI^f^ $B Mb fiSM T. DOCUMENTS DEPT. Special Report Series, No. 16. NATIONAL HEALTH INSURANCE. MEDICAL RESEARCH COMMITTEE. A Report on the Causes of Wastage of Labour in Munitions Factories employing Women. L O N D N : Pl'IlI. ISJIKD BY HIS MAJESTY'S STATIONARY OFFICE. To be purchased through any Bookseller or directly from H.M. STATIONERY OFFICE at the following addresses : IMPERIAL HOUSB, KINGSWAY, CONDON, W.C. 2, and 28, ABINGDON STREET, LONDON, S.W. 1 37, PETER STREET, MANCHESTER ; 1, ST. ANDREW'S CRESCENT, CARDIFF ; I'ORTH STREET, EDINBURGH ; or from 1C. PON - D., 116, (VRAFTON STUKKT, DUBLIN. 1918. Price 15. (W. net. NATIONAL HEALTH INSURANCE. MEDICAL RESEARCH COMMITTEE. The following publications relating to the work of the Medical Research Committee can be purchased through any bookseller, or directly from H.M. Stationery Office, at the following addresses : Imperial House, Kingsway, London, W.C. 2, and 28, Abingdon Street, London, S.W. 1 ; 37, Peter Street, Manchester; 1, St. Andrew's Crescent, Cardiff; 23, Forth 'street, Edinburgh ; or from E. Ponsonby, Ltd., 116, Graf ton Street, Dublin. First Annual Report of the Medical Research Committee, 1914-1915. [Cd. 8101.] Price 3d., post free 4d. Second Annual Report of the Medical Research Committee, 1915-1916. [Cd. 8399.] Price 3|d., post free 5d. Third Annual Report of the Medical Research Committee. [Cd. 8825.] Price 6d. t post free Id. Special Report Series. No. 1 . First Report of the Special Investigation Committee upon the Incidence of Phthisis in relation to Occupations. The Boot and Shoe Trade. Price 3d., post free 4d. No. 2. Report of the Special Advisory Committee upon Bacteriological Studies of Cerebro-spinal Fever during the Epidemic of 1915. Price 6d., post free Id. No. 3. Bacteriological Studies in the Pathology and Preventive Control of Cerebro- Spinal Fever among the Forces during 1915 and 1916. Price Is. 6d., post free Is. 8d. No. 4, Reports upon Investigations in the United Kingdom of Dysentery cases received from the Eastern Mediterranean. I. Amoebic Dysentery and the Proto zoological Investigation of Cases and Carriers. Price Is., post free Is. 2d. No. 5. II. Report upon 878 Cases of Bacillary Enteritis. Price Is. 6d., post free Is. Sd. No. 6. III. Report upon Recovered Cases of Intestinal Disease in the Royal Naval Hospital, Haslar, 191 .-16. IV. Report upon combined Clinical and Bacteriological Studies of Dysentery Cases from the Mediterranean. Price Is. 6d. t post free Is. 8d. No. 7. V. Report upon 2,360 Enteritis " Convalescents " received at Liverpool from various Expeditionary Forces. Price 2s., post free 2s. 2d. No. 8. Report upon Soldiers returned as Cases of " Disordered Action of the Heart" (D.A.H.), or "Valvular Disease of the Heart" (V.D.H.). Price Is., post free Is. Id. No. 9. A Report upon the use of Atropine as a Diagnostic Agent in Typhoid Infections. Price 1 s. , post free 1 s. Id. No. 10. The Mortalities of Birth, Infancy, and Childhood. Price Is. 6d., post free Is. Sd. No. 11. The Causation and Prevention of Tri-nitro-toluene (T.N.T.) Poisoning. Price Is., post free Is. 2d. No. 12. The Classification and Study of the Anaerobic Bacteria of War Wounds. Price 2s., post free 2s. 2d. No. 13. An Enquiry into the Composition of Dietaries, with Special Reference to the Dietaries of Munition Workers. Price 3d., post free 4d. No. 14. Reports of the Special Committee upon the Standardisation of Pathological Methods. No. I. The Wassermann Test. Price 3d., post free 4d. No. 15. A Study of 1 ,300 Convalescent Cases of Dysentery from Home Hospitals : with special reference to the Incidence and Treatment of Amoebic Dysentery Carriers. Price ., post free . No. 16. A Report on the Causes of Wastage of Labour in Munitions Factories Employing Women. Price Is. Qd., post free Is. 8d. The following is published under the direction of the Medical Research Committee by Messrs. Longmans, Green & Co., 39, Paternoster Row, London, E.C/4. Milk and its Hygienic Relations, by Janet E. Lane-Claypon, M.D., D.Sc. (Lond.). Price 7s. 6d. net. NATIONAL HEALTH INSUKANCE, MEDICAL RESEARCH COMMITTEE, A Report on the Causes of Wastage of Labour in Munitions Factories employing Women. (Approved for publication by the Medical Research Committee, 7th December, 1917.) Medical Research Committee. (National Health Insurance.) The Hon. WALDORF ASTOR, M.P. (Chairman). The Right Hon. CHRISTOPHER ADDISON, M.D., M.P. The VISCOUNT GOSCHEN, C.B.E. (Treasurer). C. J. BOND, C.M.G., F.R.C.S. (Hon. Colonel). A. K. CHALMERS, M.D., RP.H. (Lieut, -Colonel R.A.M.C. (T.) ). Professor WILLIAM BULLOCH, M.D., F.R.8. Professor F. G. HOPKINS, D.Sc., F.R.S. Colonel SIR WILLIAM LEISHMAN, C.B., F.R.S. Professor G. R. MURRAY, M.D., F.R.C.P. (Major R.A.M.C. (T.)J. SIR WALTER M. FLETCHER, K.B.E., M.D., F.R.S. (Secretary). 15, BUCKINGHAM SfR STRAND, W.( '. 2. (B2604 rB2367/4/'V 18 ) \Vt.l3357--66. 1000. 7/18. Sir J. C. & S. Gp. 32. INTRODUCTIOX. To the enquiry of which the results are given in the present Report the Medical Research Committee offered assistance as part of their co-operation in the investigations of various kinds conducted on behalf of the Health of Munition Workers' Com- mittee. The employment of women upon an immense scale in munition factories and workshops seemed to provide a unique opportunity for determining the results, so to speak, of a gigantic experiment, in which women were exposed by the national necessities to an unfamiliar stress of factory life. The simul- taneous and systematic organisation of this female labour in many parts of the country at once, under central control by the Ministry of Munitions, allowed statistical data to be secured without an undue expenditure of clerical labour and in numbers sufficient to give the results real significance. The proper mobilisation of women for national tasks, like that of men, is fundamentally a medical problem, and the measure of the success achieved in it, expressed in terms of the rate of disappearance of recruits unsuitably enlisted, seemed likely to yield information of a new and valuable kind. But the results have shown that the rates of loss by the falling out of w T omen from the employed ranks at given tasks were too high to be accounted for in terms of physical unfitness alone. As Captain Greenwood points out in his report, it must be taken as being definitely established that the " existing rates of loss in manufactories are unnecessarily high, and that so long as they continue the effective mobilisation of labour for national service has not even been approximately realised. When expressed in terms of the whole munition making female population, the avoidable losses must amount to many thousands." Avoidable loss means here in the main a loss not due to physical incapacity or ill-health. The causes at work in producing it must receive their explanation elsewhere, and must indeed remain unexplained until they can be expressed by further analysis in terms of social and economic factors through the organised studies of welfare w r orkers and the systematic " following up " of absentees. This side of the problem, however, lies outside the sphere of medical research. Nevertheless, some broad results of medical interest emerge clearly from Captain Greenwood's work. The collected facts show that the general strain of factory life in itself is not borne worse by elder than by younger women, but that when the general wear and tear of factory conditions is combined with physical labour of more than a light or medium degree, women over 23 years of age cannot as a whole stand the strain so well a>i their juniors. We are speaking here only of women engaged during physical maturity, and not in the decline of life, and yet the difference between the junior and senior members of this young adult class is well marked. As' Captain Greenwood concludes, " it is distinct enough to make an attempt at systematic recruiting of the older women for the physically lighter labour and of the younger women for the more strenuous tasks worthy of practical attention." There seems little doubt that the ex- planation is to be found, as he suggests, in the fact that the senior women will in larger proportion be married and have young children than their juniors, and will obviously be less likely to have the same reserves of strength and energy for the double demand made upon them at home and in the factory. The thoroughness with which Captain Greenwood has availed himself of the exceptional opportunities offered to him for this enquiry has appeared to the Committee to justify, even at the present time, the full presentation of his numerical data. These may long serve as a standard for reference when subsequent work of a similar kind is undertaken in the future. CONTENTS. PAGE. I. INTRODUCTION ... ... ... ... ... ... ... 5 II. PREVIOUS INVESTIGATIONS ... ... ... ... ... 7 III. STATISTICAL METHODS USED .T. ... ... ... ... 8 IV. PRELIMINARY EXAMINATION OF DATA ... ... ... 10 V. DETAILED EXAMINATION OF A LONDON FACTORY WITH FULL RECORDS ... II (a) Nature of Records ... ... ... ... 11 (b) Classification by Causes of Leaving ... ... 12 (c) Discussion of Findings ... ... ... ... 13 (d) Losses through 111 Health ... (e) Comparison of Peace Time and War Time Results 15 (/) Comparison of Age Groups ... ... ... ... 16 (g) Comparison by Occupations ... ... ... 16 (h) Summary ... ... ... ... ... ... 16 VI. STUDY OF A COUNTRY FACTORY ... ... ... ... 17 VII. STUDY OF A MIDLAND FACTORY ... ... ... ... 20 VIII. STUDY OF FACTORIES ON HEAVY WORK IX. GENERAL COMPARISON OF LIGHT AND HEAVY FACTORIES 24 (a) Introductory (b) Sources of Error 24 (c) Influence of Age (d) Influence of Type of Work (e) Discussion of Losses through 111 Health ... (/) Local Variations ... ... (g) Hours of Labour and Wastage (h) Relation of Wastage to Previous Occupation X. GENERAL CONCLUSIONS Appendix I. Tabular Matter ... Appendix II. Note on Errors of Sampling ... A REPORT OX THE CAUSES OF WASTAGE OF LABOUR IX MtMTIOXS FACTORIES EMPLOYING WOMEN. Made to the Health of Munition Workers' Committee Ministry of Munitions of War] MAJOR GREENWOOD (WELFARE AND HEALTH SECTION", MINISTRY OF MUNITIONS) Statistician to the Lister Institute of Preventive Medicine ; Reader in Medical Statistics in the University of London ; Captain 1st London Sanitary Company R.A.M.C. (T.). I. INTRODUCTION. The Health of Munition Workers' Committee have discussed various problems which offer themselves for solution when the object proposed is the greatest possible output of munitions pur- chased at the least possible cost of money, health and time. They have shown that certain hygienic standards ought to be estab- lished for the material conditions of the factories, that the hours of labour ought not to exceed a maximum which is different in different classes of work, that the exact organisation of day and night work is of moment and the system, as distinct from the amount, of wages remuneration by no means a matter of indifference. It has also been made probable that waste of time may be diminished by a heedful organisation of working periods. On the basis of these results various improvements have been suggested, some of which have now been carried into effect. One very important matter has not, however, been discussed : viz., the regularity with which employees in munition factories continue to perform their tasks. Although it is true that a majority of the operations carried out in munitions factories do not require a high degree of manipulative dexterity, hardly any are such that practice does not lead to increased precision of work and therefore rapidity of output. Thus in the manu- facture of cartridge cases, perhaps the simplest type of opera- tions, all the Committee's investigators agree that an employee cannot attain her maximum degree of efficiency until she has been engaged some weeks on her task. It follows, therefore, that a factory the operatives in which are constantly leaving and being replaced by new workers, cannot attain the same level of output as one in which the working population is stationary, and it becomes of importance to obtain exact information as to the extent to which munition workers do really change their employment. In obedience to the in- structions of the Committee that I should collect and submit to them data for a consideration of this problem, a preliminary report was completed in May, 1917. Since that time, my transfer to the headquarters of the Welfare and Health Section has enabled me to amass further material, and the report is less incomplete than would otherwise have been the case. Discontinuity and intermittency of employment are a familiar topic to economists, the able study contained in the report of the Minority of the Royal Commission on the Poor Laws being well known. Some of the large factors influencing the variations of employment in peace-time industries are, however, inappli- cable to munition work in war time, and pro tanto the problem is simplified. We have not, for instance, to consider changes of demand the demand for munitions is constant and practically infinite neither have we to allow for the desire, or alleged desire, of certain employers to maintain a reservoir, " a stagnant pool," of intermittently employed hands. But there are other complications which if not peculiar to are at least intensified by war conditions. One of these applies directly to men and indirectly to women, viz., the increasing demand of the military authorities for recruits. This must tend to shorten the average expectation of industrial life among the men directly and among the women indirectly, since an increasing fraction of the latter will be drawn upon to fulfil necessary services outside the walls of munition factories when the men in unreserved but still vital occupations are recruited. The second complication more immediately concerns the women. In the flush of patriotic enthusiasm many women, and doubtless not a few men, crowded into occupations of national importance, anxious to do their "bit." Some of these were unaccustomed to long hours of monotonous labour ; they had not learned or perhaps had forgotten, the bitter lesson of experience, viz., that enthusiasm is no substitute for dogged endurance. Instead of attributing their discomforts to the true cause, they blamed their health or the hygienic surroundings of the factory, and optimistically concluded that some other form of national work would be more congenial and better suited to their abilities. This motive of change which in normal times is held in check by the fear that when a job is lost no other may be found, has free play in the case of persons not forced to earn their bread, and even operates upon necessitated wage earners when the demand for most sorts of labour is much greater than the supply. II. PREVIOUS INVESTIGATIONS. It is a priori impossible to say whether the conditions special to peace time work and absent from war work effectively balance those special to war work and not normally in action, but it seems desirable at the outset to ascertain how labour does fluctuate in any peace time industry, the conditions of which are tolerably stable. The most recent contribution to this subject comes from America, and is contained in a report of an investigation by Mr. Joseph H. Willits, carried out in Philadelphia and published under the title " Steadying Employ- ment " in the Annals of the American Academy of Political and Social Science for May, 1916. Many of the facts recorded in this interesting document are not relevant to the Committee's task, but the following are note- worthy. A particular shop in a carpet mill was intensively studied, it was chosen because the kind of work done to some extent avoided the variability of seasonal trades, fashion being a relatively unimportant element and slack seasons being utilised to manufacture for stock. The dates of engagement and discharge of the employees were tabulated for the period 1907-15, and the following results emerged : Of males, employees hired between 1907-15, 4 per cent, remained in employment over 5 years, 3 per cent. 4-5 years, 4 per cent. 3-4 years, 5 per cent. 2-3 years, 9 per cent. 1-2 years, and 75 per cent, less than one year. Forty-eight per cent, of the men and 37 per cent, of the women remained in the service of the firm less than ten weeks. A wider but less detailed study made by Mr. Alexander of the General Electric Company, of a large number of factories of all sizes in the United States and Europe showed that on January 1st, 1912, 38,668 persons were employed in the factories studied, and on December 31st, 46,796, or an increase of 8,128. But the records also showed that in the interim 44,365 persons had been engaged, so that 36,237 had dropped out of employment during the year, or about 5| times as many people had to be engaged as constituted the increase of forces at the end of the period. These figures, as published, are not amenable to the analysis to which I have subjected certain munitions factory data to be discussed later on in this report, but it w r ill be of interest to con- trast them with some material which is also not suitable for detailed study. I obtained from a large munitions factory the following figures : In effect, the number of women employed at the end of June, 1916, was 4,571, and at the end of October 6,237, an increase of 1,666. During this period 2,791 were engaged and 1,125 left The number of engagements being 167-5 per cent, of the net increase in four months. Here also we see suggested a very large amount of discontinuity, and it is to be remarked that while in these months, together with the previous month (June), the firm lost 1,516 female employees, in only 663 cases, or 43-7 per cent., was the cause of leaving known to the firm, the balance of 56-3 per cent, having simply vanished. These statistics are quite inadequate to permit us to unravel the causes of changing employment, but they illustrate its importance and show that the problem merits further investigation. III. STATISTICAL METHODS USED. Before detailing such investigations, it will be appropriate to describe the methods of analysis which seem suitable to the case. The form in which we desire to express our statistics is analogous to that of an ordinary life table. We wish to know how many out of some basis number, say 1,000, starting to work will still be employed at the end of one month, two months, three months, &c. Mr. Willits has expressed his results in this form, but it is not clear from the summarised statement how the rates were computed and whether " exposed to risk " totals were formed and rates of loss ascertained, or how the entrants still at work were treated ; in a word, whether the summary relates exclusively to completed industrial lives. Assuming that we should proceed by the life table method, there are two distinct possibilities. The first involves the preparation of a separate card or slip for each entrant during the period of exposure studied. Suppose for instance that we are investigating all the employees of a certain factory who entered between 1st June and 31st December, 1915, and had left by, or were still at work on, 1st January, 1916. We shall (supposing for the sake of illustration that we do not use a finer unit of time than one month) prepare a series of cards upon each of which is entered the number of months each employee had remained in service, neglecting odd weeks, in the case of those who had left, or the nearest number, reckoning a duration of less than two weeks as zero, a duration of two weeks to one month as one month, &c., in the case of those still employed. From these cards we can build up the population exposed to risk of leaving for each month of service. Thus, the population exposed to the risk of leaving in the first month of employment will consist of : (1) All those still at work after a month's service. (2) All those who had left after having served at least a month. (3) All those who left in the first month. So that the rate of loss in the first month will be (3) divided by (1) plus (2) plus (3). Similarly for the next month we obtain the exposed to risk by subtracting the number leaving in the first month and the number still at work, but only having been at work for a single month (and therefore not exposed to the risk of leaving in the second month). This is a method often used in assurance offices for preparing the crude or unadjusted rates of mortality among their policy holders, and the assumptions and approximations involved will be found discussed in works on actuarial technique.* The second, sometimes called the census method, is less accurate, but requires also less minuteness of treatment. Suppose we wish to study the industrial life table, as we may call it, for the period of 1916, which we will suppose divided into thirteen lunar months of four weeks. Suppose that we knojw the number of entries and the number of losses during each month of 1915 and 1916. Suppose, further, that the losses of each month are classified into those who had been employed not more than one month, those that had been employed from one to two months, from two to three months, &c. Then, on the assumption of losses and gains being uniformly distributed throughout the month to which each figure relates, it is easy to determine the rate of loss for each period of employ- ment as experienced during 1916. The numbers leaving with less than one month's service during each month of 1916, divided by the sum of the entrants in each month, except the last of 1916, with the addition of half the entrants in the last month of 1915, and half the entrants of the last month of 1916 will give the rate of loss for the first month of service. The " exposures " in the second month of service will be the sum of half the entrants in the last month but one of 1915, all the entrants from the last month of 1915 to the last month but two of 1916 and half the entrants in the last month but one. From these are to be subtracted the losses in the first month of service from the last month of 1915 to the last month but one of 1916 (inclusive). The losses in the second month of service during 1916 divided by this sum will give the rate of loss in the second month of service ; this method can be carried back to the eleventh month of service if data for two years are available, and is practically the ordinary process used in preparing the data for a general life table, in the case of the early years of life, except that we have omitted the adjustment of the exposed to risk by comparison with the census population, that is the actual mean numbers of employees at work during the year 1916, and the subsequent graduation of the rates. * A very clear non-technical account is contained in Elderton and Fifford's " Construction of Mortality and Sickness Tables." (Black) 1914. More complete information can be found in Blaschke's " Vorlesungen iiber Mathematische Statistik." (Teubner) 1906. 10 IV. PRELIMINARY EXAMINATION OF DATA. One or other of these methods has been used in reducing the munition data. The first set of data obtained was not complete enough to allow either of these processes to be employed in its entirety, and consisted of the particulars set out in Tables 2-4. They come from the cordite department of a large munitions factory. Taking Table 4 first, for the full year from August 28th, 1915, to August 28th, 1916, we find that an increase of 624 has been the result of 1,031 new engagements among men, while 914 accessions have been obtained by 1,527 new entrants among women, percentages of 165 and 167 respectively, which compare favourably with the almost identical ratio of 167 per cent, in the case of the munition firm data quoted above, and relating to a period of only five months and also favourably with the peace- time ratio given in the American report. The second method of reduction outlined in the previous description cannot, however, be satisfactorily applied to these data. Not only is the period for which losses are classified by length of service restricted to seven months, but in addition the total losses for these months do not agree with the detailed losses set out in Tables 2 and 3. In the case of the women/ as the total losses of Table 4 for the seven months are greater (476 against 416) than the losses classified by length of service, the presumable explanation is that some girls left without- particulars of their length of service being recorded. But in the case of the men the totals from the detailed table is larger than that of the summarising table, so that some confusion with employees in another depart- ment of the works must have occurred. Lastly, the time intervals are unequal, being calendar months, or, in the case of the first period, five weeks. For these reasons it is futile to prepare averages purporting to bear comparison with exact rates of losses computed by any actuarial method. As a matter of curiosity, however, a calculation was made on the assumption that the intervals of time were really equal and that the rates of loss for different periods of service shown for the seven available months still applied to earlier experience. With these assumptions, it appeared that the rates of loss were approximately as shown in Table 5, from which we deduce that among the girls some 36 per cent, of entrants will have left by the end of the seventh month from entry, and a rather smaller percentage of the men. This, although comparing favourably with the Carpet Mill returns obtained by Mr. Willits, shows a considerable absolute loss ; but the basis of computation is so conjectural that it would be quite improper to attach more than the slightest degree of importance to the figures until controlled by more precise data. 11 V. DETAILED EXAMINATION OF A LONDON FACTORY, WITH COMPLETE RECORDS. (A) Nature of Records. The next set of data merits closer attention, and is worthy of completer analysis. It is discussed in full detail, both to illustrate the method and because its extent and accuracy are greater than attained by any other set studied. The data were collected in a munitions factory of high standing, the investigators being afforded every facility in consulting and tabulating the very full records kept. The factory is at present organised in two main departments,' one producing cases and bullets the other carrying out cartridge and grenade filling operations. The distinction is not absolute, but it is roughly true to say that the former does not handle explosives while the latter does ; hereafter the two are denoted by L.I and L.2 respectively. The firm has for some years past maintained a card index system, and the name, age, previous occupation, character, and reason for leaving has been recorded in nearly every case since 1910. Particulars of entry are indeed available back to 1907, but it does not appear that details as to discharge were universally recorded prior to 1911. This investiga- tion is therefore confined to the period irom January 1st, 1911, to December 1st, 1916. The whole of the cards were copied each upon a separate slip (I have to thank Mr. S. H. Burchell for undertaking this tedious work), and the duration of employment of each girl computed in the way described above. The computed durations are not strictly accurate as the differences of length of the calendar months have been ignored ; the practical error so introduced is quite unimportant in the present connection. The data were classified into two main groups : (1) covers the entrants from 1st January, 1911, to 4th August, 1914, and the rates of loss are deduced from the numbers leaving within the same period out of the entrants ; (2) deals with entrants from 4th August, 1914, to 1st December, 1916, the rates of loss similarly depending on the experience of that period. An analysis was also made of the rates of loss since the war among employees, who having been engaged before the war, but since 1st December 1911, were still at work for varying periods after the 4th August, 1914. The numbers in this category were, however, quite inadequate to provide useful tabulations (451 girls and women), and are not here discussed. In each of the two main groups, four sub-groups were formed consisting respectively of girls and women aged 13-17 inclusive, 18-22, 23-27 and 28 or over at the time of engagement. The actual number of separate engagements 12 dealt with was 8,818.* This is larger, of course, than the number of separate girls as there were numerous re-engagements. Evidently, however, for our purposes, the re -engagements should be the fundamental data. These different groups were then analysed by months from to 24 (i.e. those leaving after twenty- four months were counted as " existing at the close of the period "), it having been found on preliminary trial that the data were insufficient to provide rates for later months ; indeed, as it is, the rates for months beyond the sixth or seventh are evidently subject to large errors of sampling, and have little value. (B) Classification by Causes of Leaving. The causes of leaving were classified into three categories : (1) Those who left on account of ill health. (2) Those who left for some sufficient reason other than that of health. (3) Those who left for no reason or for an insufficient one. The meaning and limitations of this classification which has been used in all subsequent analyses must be fully grasped. (1) needs little comment save to remark that the diagnosis of ill health has not necessarily been made by a medical man, but is, in an unascertainable proportion of cases, merely the girl's statement, and that girls who left after an accident are included. Under (2) were comprised (a) girls who left to be married ; (b) girls who were moving from the district ; (c) girls required at home on account of illness there, or, in the case of married women, to keep house ; (d) girls who found the wages earned insufficient ; (e) girls discharged by the firm not for misconduct, but because they were unsuitable or because work was slack ; (/) girls considered by the factory surgeon to be physically unfit, although not complaining of illness. Under (3) were grouped : (a) girls who left without giving any reason at all ; (b) girls discharged for misconduct ; (c) girls leaving for other employment without assigned reason, girls wanting a change, or " dissatisfied," or * Sixty-five more slips were sorted into occupational groups than were used for the tables of losses. Nearly all these are accounted for by the number of girls still at work who had been engaged less than two weeks before the conclusion of the period of observation. An excess cf ten observations in the group of ages 13-17 in the first sorting is not explicable in this way, and is due either to a mistake in counting or sorting. It having been found that no sensible difference would be made in the rates by the discrepancy, the tabulation was not re-checked. In addition to the slips used for tabulation thirty-nine other entries were found which did not record all the particulars. No use could be made of these. The register also contained the names of 865 girls and women, engaged before 1st January, 1911, who were still at work or had been discharged between 191 1-17. For the reasons stated in the text these data were not available for the investigation. 13 stating the work to be " unsuitable." It may be mentioned that a large majority of (3) consisted of the girls giving no reason at all, whom we may colloquially term the vanishing girls. (c) Discussion of Findings. Tables 6-21 were prepared in the manner above described, and Table 22 gives the distribution of the peace time and war time entrants by occupations prior to entry. In these, and subsequent tables, the first column gives the month of service, e.g., the figure one means the second month of service, while the second column records the number which was exposed to risk of leaving in that month. A feature common to all the tables is the heavy rate of loss in the first two or three months' service. It is to be especially remarked that the rate of loss in the first month is a minimum figure. In addition to the girls registered in the firm's card index, 1,420 girls are known to have worked in the factory between 1910 and 1917 for periods varying from a few minutes to a few days. These cases are really industrial still-births never deemed to have a separate industrial existence, or, to vary the metaphor, they may be regarded as visitors, not residents. Clearly, their inclusion would suffice to double the rate of loss in the first month. Since the criterion adopted by different firms as to what girls shall be registered as definitely engaged may vary, rates of loss during the first month can only be compared when due allowance has been made for any such difference of system. Rates of loss in the second and subsequent months are not in general subject to this criticism.* A second characteristic is the large proportion of losses for which no satisfactory reason can be assigned. Taking the general table (7) as an illustration, it appears that of 1,000 girls entering the factory 132 will not outstay the second month. Even had there been no loss from sickness or from other reasonable cause the 1,000 would still have been reduced by 10 per cent, for no sufficient reason. (See also Table 23.) As this is a minimum figure (still-births excluded), and is the experience of a firm of high standing thoroughly familiar with the industry, which they have not entered upon for the first time since the war, it is evident that the absolute losses of hands occurring in munition factories are serious. Workers leaving with less than two months' service can, on the average, have made but very small contributions to the output of munitions. While, on the debit side of the account, we must enter the trouble to which they have put the officials of labour bureaux and the adminis- trative staff of the factory, as well as, in some cases, the demands made upon the transport facilities of the country. In the light * It is, however, important to remember that in some factories girls are not struck off the books as having definitely left until after the expiration of several weeks. This will sometimes explain a sudden rise in the rate of loss during the second or third service months. 14 of our new tables we may fitly return to the case of Table 1 and estimate whether the factory showing those results was losing hands at a greater rate then the one here specially considered. First, as to the numbers of those said to have started in each month who would still be present at the end of October. Using the appropriate rates from Table 7, we conclude that of the 3,413 entrants less than 542 would have been lost by the end of October. We have no information as to the distribution by length of service of the 4,340 already at work on June 1st, but assuming 14 per cent, to have been in their first month, 16 per cent, in their second month, 18 per cent, in their third, 20 per cent, in their fourth, and the remainder in the fifth month of service, we should expect 659 to be lost or the total losses to have been 1,201. Actually they were 26 per cent, more, viz., 1,516. Owing, however, to the difficulty about the system of registration we cannot be certain that this is really a worse result than would have been found in the other factory. Turning to the data of Tables 2-5, the proper comparison is with Table 8, as the factory providing Tables 2-5 handled cordite. The rates of losses from all causes are in very fair agreement, but the cordite factory lost many more for reasons of health. This may point to a real difference, owing to the nature of the work being different in the two cases ; but more extensive material is necessary to enable one to pronounce a definite opinion. We can now compare the three tables as wholes. It must be noted that the total rates entered at the feet of the tables are to be regarded with suspicion ; comparisons of them are open to an objection similar to that involved in the use of crude death-rates. The numbers at risk at different ages are not in the same proportion in the different populations, for instance, 8-8 per cent, of the total exposed to risk is in the service group for Table 6, but 10-4 per cent, in Table 7 and 14-6 per cent, in Table 8. As the rate of loss in every table is much higher in the first few months than in the later months, this causes the comparison of the crude averages to be unduly favourable to the pre-war populations because the latter include proportionally more girls in the later months of service. A better general idea is got from a study of diagrams. In Diagram 1 we have the curve of rates of loss for all causes, and it is seen that on the whole the peace-time rates run higher than the war-time rates for the case and bullet department, and are not very markedly different from the war-time filling department. The losses of the first month are higher in each war-time experience than among the peace-time employees, but this is set off by higher rates among the latter in subsequent months. But an important modification is necessitated by the fact that, before the war, periods of slack trade led to the discharge of girls whose services were not required, and no such factor has been in operation at this factory since the war. Consequently, Diagram 2 is more instructive. From this (paying of course very CO UJ a o OT m 5 o w d z 2 ^ 2 ^ CJ <0 W d P S " ~ < Of S I \ \ o ao S3SSOT PERCENTAGE OF LOSSES _ N Z ? 0) o _ z * ^ A . X C 1 z> eg JS- 0,0 o o CD r rn O) rn (p IS) 15 little attention to the erratic percentages for the later months) wi i may conclude that the trend of events is much the same in the peace time and war time industrial populations. But the suggest ion will be made that the class of the population in factories is now very different from what it was before the war, so that an identical rate of loss may mean really a greater steadiness owing to the gross rate being increased by the presence of a percentage of girls who are really unfit, by virtue of temperament or nurture, for the strenuous life of a factory. Some light is shed on this by Table 22, giving the occupations of the employees before engage- ment. So far as the girls under 18 are concerned, the occupational distributions are similar in peace time and war time experiences. This remark also applies to the group 18-22. No useful com- parison can be made in other classes owing to the smallness of the corresponding peace-time populations. Leaving on one side the question whether such differences as do appear are significant, in the statistician's sense of the term, it certainly cannot be urged that they are of much practical importance, or that the war-time population is very unlike the peace-time population in respect of the occupational classes from which it has been recruited. We may now briefly consider points suggested by the detailed tables. Such detailed comparisons will not be pressed far ; they involve a proper evaluation of the " probable errors " of the constants compared, and, in view of the fact that most of the differences are absolutely small, it is unprofitable to devote much labour to the rather intricate calculations which such a study involves. (D) Losses through III Health. Less than 300 girls quitted their employment on account of ill health or as the result of some accident during the whole period. The distribution of losses for different lengths of service cannot be satisfactorily elucidated. It does not appear that the differences are significant when one age group is compared with another (Table 24) ; the crude rate is highest in the filling depart- ments and lowest in the peace-time case and bullet section. Expressed as yearly rates the range from 33 to 61 per 1,000 is considerable, and the amount would, if confirmed on larger numbers, imply that sickness is a somewhat important cause of leaving. But, for the reasons given, I do not think that any such definite inferences should be drawn. The subject will come up for fuller discussion when we consider the whole of the data collected. (E) Comparison of Peace Time and War Time Results. The diagrams and Tables 6, 7, 8, supra, have suggested that, when the losses from all causes are considered, the war time 16 experience is somewhat more favourable than the peace time results. This is borne out by a more exact test when attention is paid to age. The group chosen was that for ages 18-22, and it was found that the probability of the observed difference being a chance event was small. A good deal of the discrepancy was due to the much higher rates of loss in the months beyond the tenth among peace time workers, but even without that group's contribution to the value of ^ 2 the fit is not good. It has been pointed out above that the advantage of the war time workers is decidedly less when only losses without reasonable cause are considered. (F) Comparison of Age Groups. The test has been applied to the age groups 13-17 and 18-22 of the war-time case hands, and the result is indecisive (Table 24). It is of course possible that some differentiation will be found to exist when we have the advantage of dealing with larger numbers, and more data will be found in subsequent paragraphs. At present it does not appear that any difference which is practically important can be demonstrated. (G) Comparison by Occupations. The general tables, together with the data provided by the factory manufacturing cordite, suggested that the rates of loss in departments handling explosives and chemical substances generally were higher than among the machine hands of the case and bullet departments. A comparison of workers at ages beyond 28 years in the two departments (this age group was chosen owing to its size in the case of the filling hands) neither confirms nor refutes this suggestion. All that can be said is that the differences is not so great that it could not be regarded as a coincidence. Here, again, the data are inadequate for pro- nouncing a final judgment. (H) Summary. The chief result of the analysis so far described is to establish the fact that the loss of effectives in a well-managed munition factory due to theoretically avoidable causes is great. It cannot, however, be taken as proved that the whole of the girls who left without giving any reason for so doing had in fact no sufficient reason. The loss of effectives, amounting to not less than 10 per cent, and probably to more than 10 per cent, of the original strength, after a period of employment so short that little useful service has been rendered must be contributed to by some or all of the following causes : (1) The default of the employee herself, such as caprice, unwillingness to submit to discipline, love of the excitement of change for its own sake. (2) The fault of the employer as displaying want of judg- ment in the selection of candidates for employment, the conditions of which are unknown to the applicant but known to the employer. (3) Similar lack of judgment on the part of the officials of labour bureaux. No national system of organisation can eliminate these causes of wastage, but it may be possible to diminish them. So far as the default of the employee is concerned, the State has endeavoured to meet the evil by the introduction of the permit system. I find, however, that difference of opinion exists among factory managers as to the efficiency of this system. It has been urged that since the penal mechanism must be set in motion by the original employer (i.e., if another firm engages the girl without permit, this firm is only penalised when the first employer takes action) it is impossible to take effective action without over-straining the administrative staff. The other side of the case, viz., the improper selection of applicants, merits further attention. Factory administrators and the officials of labour bureaux indulge in mutual recriminations. On the one hand it has been complained that in a certain national factory the engagement of boys is carried out by the supervisor of the Welfare Department, and that this officer is imperfectly acquainted with the technical conditions and requirements of the shops needing fresh hands, with the consequence that the labour exchange is put to needless trouble owing to the incomplete- ness of the particulars furnished. On the other hand, cases are said to have occurred in which a labour exchange has sent quite unsuitable candidates to factories situated at long distances from the exchange, with the consequence that both time and railway accommodation have been squandered. It is not, however, the province of this report so much to deal with causes as to contribute to knowledge of effects which stand in need of explana- tion. VI. STUDY OF A COUNTRY FACTORY. We have now studied the conditions observed in a factory located in the suburban area and will pass to one in very different surroundings. Factory L.7 is in the open country, six miles by rail from the nearest town and still further from any large city. L.7 is indeed very isolated. There is no village in its immediate neighbourhood, no sleeping accommodation and not even a picture palace. A very small number of the workers 18 can lodge in a village two miles away, but the vast majority must travel at least six miles to and from their work. At present the women work in two eleven hour shifts with weekly change. Forty -eight hours elapse between coming off one shift aftd entering upon another. There is no Sunday work. All labour is time work, and the rates paid suggest that the total earnings are on the average greater than in L.I and L.2 although the hourly remuneration is not. The work is divided into that of the cordite and that of the guncotton department ; it is therefore roughly similar to that of L.2. Ample provision is made for the care of workers' health and comfort. In these respects, factory L.7 is probably unsurpassed by any other in the country. The arrangements are controlled by a matron, who is a trained nurse, having under her four or five forewomen specially appointed (never promoted from among the workers), who have charge of the mess rooms, kitchens and changing rooms. Special attention is directed to the catering arrangements, the governing principle being that an ample supply of nourishing, well-cooked food must be provided for women at work. Employees are obliged to take advantage of these arrangements, nobody being permitted to bring food into the factory. All the workers receive dinner and. tea (which latter meal is of a substantial character) for lOd. per diem, which sum is deducted from wages. Special trains run to the factory to convey workers to and from their places of employment ; in these the women travel free of charge. Engagements are effected through the labour exchanges of the two towns least remote from L.7, and no direct applications for work are entertained. Any woman absent for more than two days without communicating with the management is auto- matically discharged, but in practice all these cases are followed up and the sick visited. The rate of wages has not been increased (except in the case of charge-women) since L.7 was opened, roughly twelve months ago (written in February, 1917). At present (i.e. when the data were collected) there are approximately 900 women at work. Tables 25-28 contain the analytical results which are strictly comparable with those above, and may be contrasted with the tables relating to L.2. For women over 28 the rates of loss from all causes in the first month of service are much the same in the two factories, but those for no sufficient reason are half as many again in L.2. In all other months at this age, and for all months at other ages, the rates are very much lower in L.7 than in L.I or L.2. As will be seen from Table 29, if the rates of loss found in L.2 had applied to the present factory, there would have been 19 227 losses ; actually there were 79, or little more than a third of the expected number. Table 30 gives the distribution of the employees by previous occupation, and brings them into comparison with those of L.2 (war time employees). It will be seen that the distributions are by no means identical. Unfortunately the age groups over 28 cannot be properly compared owing to the large proportion of women from L.2 the previous occupations of whom are unknown. In age group 23-27 the domestic servants, laundry hands and shop assistants are in considerable excess at L.7, and the factory girls in defect. The same remark applies to the under 22 group where, also, the girls not previously engaged in directly paid work (that is, those said to have done " household work ") are only half as numerous as in L.2. This is a very interesting fact, and the question arises whether (a) the women who have had no previous experience of factory life are really less likely to remain steadily at work than old factory hands, or (6) whether they are less likely to remain ; but the conditions of employment are so favourable at L.7 as more than to counter-balance the initial instability of the population. Evidently the ideal plan statistically is to tabulate losses by previous occupation, but even in L.I and L.2 the data are hardly numerous enough to allow of this being done with any hope of reaching averages which would bear statistical scrutiny, and they are of course much too scanty in L.7. This process must therefore be deferred until further data have accumulated. It is prudent to suspend judgment on the point. The investigation does, however, disclose facts of practical importance at the present time. We see that the losses of a factory, which is by no means in the rear rank of existing munition factories, are nearly three times those of another factory the conditions of which, although good, can hardly be considered an ideal unattain- able elsewhere. Now we cannot say definitely that these better conditions are the cause of the advantage ; it is arguable, in view of the previous occupational distributions, that what has really happened is that the country factory has drawn its labour from country people, and that these are less, prone to change than urban girls. But, if this is the case, it follows that special efforts should be made to recruit the country girls for munition factories, and, wherever possible, to extend the factories operating in provincial towns of the second or third rank, or in purely country districts-. It should, on this hypothesis, be considered whether the superior transport facilities of the great cities really compensate for the waste of time and effort involved in numerous short term engagements and re -engagements. Of course, if this alternative explanation should prove on further examination to be unneces- sary, then a rapid levelling up of the hygienic conditions to the standard established in the factory under notice will suffice to steady employment. 20 One point must at least be regarded as definitely established, viz., that existing rates of loss in many factories are unnecessarily high, and that so long as they continue the effective mobilisation of labour for national service has not even been approximately realised. When expressed in terms of the whole munition making female population the avoidable losses must amount to many thousands. Thus far we have examined in detail two factories, the external characteristics of which contrast ; we must now study one the circumstances of which are, in some measure, intermediate between two extremes. VII. STUDY or A MIDLAND FACTORY. The requisite data were obtained from a factory in a midland industrial city, manufacturing and filling fuses for 18-pdr. shells. The establishment is organised into tw^o branches, one manufacturing the fuses the other loading them, hereafter called L.4 and L.5. In both sections the general hygienic condition of the shops is excellent. The immediate surroundings are of the usual type one associates with the inner suburban zone of a manufacturing town. The factory is at some distance from the centre of the town, the neighbourhood is crowded, and there is much difficulty in obtaining lodgings. Cheap restaurants or tea shops are not numerous, perhaps less so than in corresponding metropolitan areas. To meet these inconveniences, the firm has provided canteens and residential hostels. The former are well constructed, and the arrangements for serving and preparing meals cleanly and efficient. The proportion of women who actually purchased their meals in the canteens was not, so far as simple observation revealed, very large, and there seemed to be an impression that the accommodation was insufficient. The residential hostels, which housed a larger proportion of the women employed in L.5 than of those in L.4, were provided with separate canteens. Here also, passing over certain com- plaints which refer to temporary occurrences, the smallness of the accommodation seems to have been the principal defect. This factory employs the whole-time services of a qualified medical woman, and all women admitted to work in the filling section are subjected to a preliminary medical examination. For purposes of inquiry the records kept proved to be similar to those made in L.2, L.I and L.7, with the rather unfortunate exception that no particulars were available as to the previous occupations of the women. In a general way, it may properly be inferred that the welfare system was inferior to that of L.7 but superior to that of factories L.I and L.2. As to material differences between L.4 and L.5, in addition to the points above noted with reference to preliminary medical examination and the fact that a larger proportion of 21 the women in L.5 resided in the hostels, it is to be mentioned that the rate of payment was higher in its shops. The period statistically investigated covers engagements effected between 1st January, 1916, and 1st December, 1916, and refers to 2,584 engagements in the filling and 2,424 in the making shops. The data have been handled precisely as those discussed above, with one exception. In L.I and L.2 the (very small) number of women who left on grounds of medical unfit ness, as distinct from actual illness, were classified under the general heading "left for 'sufficient' reason." In this factory the numbers, being relatively numerous, were scheduled separately. The nine tables contain the full analytical results. We will consider first L.4 (Tables 31-35). The differences between the rates for different age groups, whether statistically significant or not, are absolutely small and need not detain us ; we may compare the table for all ages with that of L.I. So far as the rate of loss from all causes is concerned the difference is not at all striking. The present factory has the advantage in the first month but loses it afterwards, with the result that the numbers lost by the conclusion of the fifth month from entrance are not very different in the two cases. The detailed distribution of losses is not, however, the same, the other factory had con- sistently higher losses for " no known reason," the present lost more for " health " reasons, even apart from the elimination of workers as medically unfit. As there is no reason to suppose that this factory is in any way more unhealthy than the former one, the explanation must be, either that a large proportion of those who left in the other factory without giving any reason did so for reasons of health, or, alternatively, that the health losses in the present case were really losses without adequate reason, the bad health being merely an excuse. The latter alternative is perhaps the more probable, because, in view of the presence of a whole-time medical officer, it seems likely that most of the genuine cases of illness will fall under the medical unfitness column which, it will be noted, dees not differ much in relative magnitude from the health column of the earlier table. If, however, this explanation is not accepted, it would seem that the health losses are rather serious, amounting to nearly 1 per cent, per mensem throughout the table, over and above the medical rejection of nearly J per cent, per mensem. Looking at the results generally we may properly assimilate the experience with that of the first factory, and contrast it with that of factory L.7. Passing now to L.5 (Tables 36-40) the rates are seen to be consistently much lower than in L.4. This difference is seen particularly in the sub -section of losses for no known reason, the health losses not being greatly "less than in the other factory. At first this factory seems to yield better results than even L.7 V but this first impression is not altogether borne out by further study. So far indeed as the first month is concerned this factory 22 seems better, but its rates of loss in subsequent months are consistently high, higher than in the other establishment, with the result that the numbers who have left by the end of the fifth month form 12-6 per cent, of original entrants as compared with 10 per cent, in the previous case. On the basis of the numbers at risk, it is not practicable to assert any superiority of one over the other. The evident superiority of the L.5 C. over L.4, which is too distinct to be properly attributable to chance errors, needs explanation. Three factors which have probably contributed are (1) the preliminary medical examination of entrants, which weeds out many of the unfit ; (2) the higher proportion of women arrived from long distances and specially housed under some supervision ; (3) the higher rate of payment. Some further light would be thrown upon the relative importance of these factors inter se if we had a knowledge of the occupational distribution of employees before engagement, and if we had accurate data as to net earnings for the two groups. Neither desideratum being available we must leave the possible explanations without comment. But the general result, viz., that the factory as a whole (i.e. L.5 and L.4 together) is better off than L.I and L.2 and worse off than factory L.7 in the matter of wastage, while it occupies a similarly intermediate position between the two factories in welfare efficiency, is to be remembered when the general welfare account comes to be balanced. VIII. STUDY OF FACTORIES ON HEAVY WORK. I now turn to a group of three factories, also situated in a provincial manufacturing city but doing rather different work from that which has so far come under our notice. These three factories are all manufacturing projectiles. H.3 produces 4-6 inch shells, H.2, 6-9 inch shells, and H.I, 9-15 inch projectiles. The H.I workers are on time wages, the remainder on piece-work. Up to August, 1916, eight hour shifts were worked in all three ; since then 'H.3 has used twelve hour shifts. Women employed in these factories are for the most part recruited in the district, and no residential hostels have been provided. The factory canteens are good, while two of the factories themselves are well constructed. The third (H.2) has not been considered altogether satisfactory from the hygienic standpoint. In Tables 41-53 I present statistical analyses, which are comparable with those prepared for the other factories. Table 54 similarly provides information respecting previous occupation. A consideration of Tables 41-53 reveals the fact that H.I, H.2, and H.3 do not, with the exception of the age group 18-22, compare favouiably with any other factory so far discussed in thip report. Great caution must be exercised in drawing con- clusions, because the absolute numbers of employees are small, and on this account I abstain here from any attempt to differen- tiate H.I, H.2, and H.3 one from another. That the rates of 23 loss at ages above 22 are significantly greater than in the first age group (no statistics are provided for girls under 18 as the number was quite insignificant) is I think established. I have applied the test mentioned above (p. 16) to the comparison of rates of loss in the age groups 18-22 and 23-27 for H.I, H.2 and H.3, grouping together the exposed to risk in month 9 and subsequent months. The probability that the observed difference is a mere chance is, by this test, very small ; less than 6 in 10,000. This test is not, as remarked in the appendix, altogether satis- factory, having been devised for a slightly different kind of statistics, viz., death-rate at ages of ordinary vital statistics, but is acceptable as a rough and ready criterion. If the separate tables are carefully examined it will appear that the unusually heavy losses are seriously contributed to by the " ill health " and " other sufficient reasons " columns, a fact which suggests that we are really dealing with an evil for which the employees themselves are not responsible. I am of course aware that the accuracy of statements purporting to describe the grounds for leaving is not likely to be of a high order. But, simply regarded as indices, these classifications are I believe of real value. It would be a gratuitous assumption to suggest that the proportion of untruthful persons in one factory is likely to differ sensibly from the percentage in another. I think, therefore, that these three factories offer a problem in need of an objective solution. The record of previous occupations does not furnish this. Unfortunately a large proportion of the women have had no entry made upon their cards as to the nature of the previous employment, but, so far as appears (Table 54), the percentage of industrially employed women is greater than in L.I, L.2 or L.7. We should expect this to be the case in a large manufacturing city using local labour, and it is possible that this contributes to frequent changing of employment, there being good opportunities to obtain another place. Further than this we can hardly go. The next fact to be considered is that, while the women undergo no preliminary medical selection, their work is appreciably heavier than in the fuse and cartridge factories, which provided so much of our other material. The large percentages of women, especially older women, who gave up work on alleged physical grounds are clearly consistent with the view that we have to deal with an occupation which is on the whole unsuitable for such women, so that both a preliminary medical examination and also perhaps a re -distribution of female labour as between factories of different types might solve the problem. I do not, however, think that any such far-reaching conclusion can properly be based upon data obtained from a single group of factories. It is necessary to widen the scope of this inquiry, and to look upon the results here communicated as suggestive of what such a wider inquiry might contribute to our practical knowledge. 24 IX. GENERAL COMPARISON OF LIGHT AND HEAVY FACTORIES. (A) Introductory. The reader has now been brought to the point at which I arrived on May 2nd, 1917 ; there seemed to be a considerable difference between the wastage of women workers in factories on relatively light physical work and in those engaged upon more laborious tasks, but this conclusion rested upon a comparison between several " Light " and only a single group of " Heavy " factories. It was accordingly decided to investigate several more " heavy " factories, and also one or two " light " factories. The " heavy " group comprises one very large controlled establishment making projectiles, eight national projectile fac- tories and one national shell factory ; the representatives of the " light " class were an explosives supply factory and a national cartridge shop. ^*&lj The data have been distributed into 55 tables which, owing to curtailment of paper supplies and expense of setting, cannot be reproduced here. In making comparisons, certain fallacies exist which cannot be completely guarded against but may be minimised . (B) Sources of Error. All the establishments previously considered were such and, from the point of view of time period investigated,* so chosen that variations of production or programme did not complicate the comparisons. When we so extend our data as to include as many factories as can furnish statistics (which is what was done with respect to national projectile factories) we cannot pick and choose, and may be dealing with a period during which, owing to changes of constructional programme, large numbers of hands were discharged. This does not of course affect the rates of loss for insufficient reasons, provided the records are properly kept, because women discharged on account of changes of programme are entered under the heading " left with sufficient reason." * A point in the method of reduction which may seem to the reader arbitrary, is that we restrict the analysis to persons who entered during the period to which the losses relate. Thus when we took, say, the losses of 1916, we only used those relating to entrants during that year, so that we could have no persons exposed to risk for more than a year. It may be asked why did we not tabulate all persons, irrespective of their date of engagement, who were exposed to risk during the chosen period, bringing for instance a person, engaged a month before the period started, into the " exposed to risk " column of the second month (not, of course, into the first month) and so on ? The advantage would be to increase the numbers upon which the rates of leaving for later months are based ; the disadvantage is that the survivors from an earlier period of service may (under the rapidly changing industrial conditions) have been subject to a process of selection different from that in operation during the period chosen. This disadvantage, combined with practical difficulties of consulting older records of engagement, has determiRed the actual procedure. 25 But it might greatly affect the total rate of leaving, and a compari- son of such rates might therefore be gravely misleading. Thus, the fact of one factory having a higher rate of loss than another might only mean that the constructional programme of the former had been varied during the period under review. As mentioned, the comparison of rates of leaving for no known reason or those due to ill health would be less affected by this disturbance were the records accurate. But no ex post facto investigations, as these necessarily are, can proceed on the assumption that the list of those who left without recorded reason does not include a proportion of Josses consequent upon a change of programme. In one or two of the factories (e.g., H.4) it was possible to select a period in which no changes of programme or difficulty in securing supplies of raw material complicated the analysis, and I always endeavoured to secure suitable periods. This endeavour may not, however, have been always successful. Another theoretical source of error is the fact that the period studied has not been the same in all factories. Consequently, in combining the data into a general summary we are bringing together observations which are not strictly comparable, for they do not relate to contemporaneous events. This error was unavoidable, but, as the periods available are not in fact widely separated, in particular, there has not been a systematic choice of more recent periods for H. factories as a whole, it is improbable that the admitted heterogeneity introduces a practically important error. The next point arising is the broad classification into L. and H. groups. In all but two instances there is no doubt as to the appropriate class ; the L.'s were engaged in rifle cartridge, fuse making or filling work, while the H.'s were on heavy projectile manufacture. The borderland cases are L.6 and H.5, where smaller shells were made. Hence both might have been grouped under H. or L. The considerations deter- mining the actual classification of these intermediates were that L.6 conformed in several other respects to the circumstances of L.8, the classification of which was free from doubt, while similarly H.5 had points of resemblance to H.6. In any event, the classification being doubtful, an allocation of one factory to each group is appropriate. On the whole there is no doubt that the actual amount of muscular effort required by a working day of the same length is greater in factories of class H. than in those of class L. Of course variations exist as between different factories within the same class, but the general difference is of the kind indicated. We can now review the influence of different factors upon the rates of loss. (c) Influence of Age. This matter has been discussed in detail above in connection with individual factories. A general survey is provided by Table 74 (in this table the under 18 group is not given because 26 few of the H. factories employed sensible numbers of girls under 18). The figures prefixed with the signs ^ are the probable errors of sampling computed by a method described in the Appendix. These probable errors are to be regarded as mere rough approxi- mations ; differences falling within three times the limits of these errors may be disregarded ; on the other hand greater deviations cannot be decisively asserted to lie outside the real limits of feasible sampling errors. The survival rates which are compared are the percentages of entrants still at work at the conclusion of the third month from entrance (i.e., after the losses of months 0, 1 and 2 have been allowed for). This limit has been taken as covering reasonably probable variations of registration, and also as corresponding to the extreme period of novitiate within which the average entrant has not attained her best level of productive power. It will be seen that the survival rates diminish with age almost invariably among the H. factories, while this is not so in the L. group. Taking the combined data of the two groups we reach Table 75. We find that the age group 18-22 is in a more favourable position than either of the later age groups among H. employees, while the differences observable in the L. group are almost within the sampling errors of the determination and not practically important. The older women differ from the younger, not only in age but also in that a larger proportion of them are married. Un- fortunately we have not statistics of civil state for the whole of the factories, but the data of Table 78, relating to three factories, sufficiently illustrate the point. It appears that the double handicap of age and home duties is insufficient to diminish appreciably the resisting powers of the older women, so far as light physical labour in factories is concerned ; when we pass to a heavier type of work this ceases to be true. The general analysis confirms the surmises based upon the particular compari- son of factories discussed above. We may conclude that, from the point of view of wastage, the recruitment of women over 22 for H. factories is relatively disadvantageous. (D) Influence of Type of Work. In continuation of the above comparison we may bring beside one another the same age classes of different factories. It will be seen that the wastages are on the whole smaller throughout the L. group at each age than in, the H. group (see Tables 61-72 and Tables 74-5). At ages 18-22 the difference is, however, small. At ages 23-27 and over 28 it is sufficiently obvious. At the same time I draw conclusions with some reserve. One at least of the H. factories (H.8), showing a very heavy rate of loss at all ages, did not furnish data which I regard as entirely satisfactory. On the whole, however, the general impression is confirmed by the increase of wastage with age among nearly all H. factories, and the failure of L. factories to exhibit this. 27 <** A specially interesting comparison is between H.6 and L.3, both of which are situated in the same part of the country and conform to the highest standards of general management. L.3 is a national filling factory handling T.N.T. ; H.6 is a national projectile factory. The three months' wastages are not significantly different at ages 18-22 nor at ages 23-27. At ages over 28 the filling factory loses appreciably fewer women than the projectile factory. In this comparison all factors are constant save the type of work, and the inference seems legitimate that labour in projectile factories is really less suitable for older women. (E) Discussion of Losses from III Health. I cannot attach much importance to the recorded incidence of sickness for reasons set out above ; the data are not attested by medical evidence and may be no measure of the true incidence of sickness upon munitions workers. Still, it is necessary to call attention to the results. From general Table 59 we learn that among nearly 37,000 women under observation for at least a month, illness or medical un fitness is given as the cause of leaving 1,651 times. A survival table showing the wastage which would have occurred had sickness alone been responsible for leaving is provided in Table 76. It appears that the losses from ill health are practically the same at all ages in L. factories, while they increase but not uniformly ' after age 22 in the H. factories. The girls under 23 have fewer sickness losses in the H. than in the L. factories. An explanation which suggests itself is the differential incidence of toxic jaundice and allied conditions upon younger women in filling factories. This is not, however, the real explanation, for it is unsubstantiated by a comparison of Tables 79 and 80, which record the particulars for age groups 18-22 in the most accurately comparable filling and projectile factories. In those tables it will be seen that the sickness losses of the filling factory are not higher than those of the projectile factory. (F) Local Variations. I have already called attention to the wide local divergencies in commenting on special factories in detail. I have suggested that some part of the explanation is to be found in differences in the efficiency of welfare supervision. I cannot say that this explanation is demonstrably correct, but the contrasting condi- tions of L.7 and L.5 on the one hand with those of L.I, L.2, L.4 and L.8 on the other are of interest. It might not unreasonably be suggested that any factory which loses more than 20 per cent, of its entrants within three months should institute a special enquiry into the reasons. Naturally, in some cases, it will be found that quite sufficient reasons exist. Thus, the heavy losses of L.6 are explicable on the ground of great difficulties in securing housing accommodation, but this explanation will not cover all cases. 28 (G) Hours of Labour and Wastage. Table 77 contains particulars of hours of labour brought into comparison with rates of wastage at ages 28 and over (these latter rates being, I think, sensitive indices of change). It is evident that the two variables are not closely connected. It is true that in the two factories with the longest working weeks the survival rates are low, but the highest survival rates are not found in the factories working the shortest hours ; ex- pressed as a correlation we have r -05 i -16, to which no significance attaches. The comparison throws no new light upon the connection between industrial wastage and length of hours. Of course, it does not demonstrate that no important connection exists. The data are unsuitable for an exact analysis. (H) Relation of Wastage to Previous Occupation. The facts elicited are shown in the various tables, and it would be of interest to determine whether wastage varies among recruits from different classes of industry. I am of opinion that the available data are inadequate for such an inquiry. The returns of no single factory are numerous enough to allow of the necessary sub -division being undertaken with a real hope of arriving at statistically significant results ; a combination of the totals is impracticable, since the proportions of employees whose previous industrial history was unrecorded vary from factory to factory. Altogether, apart from this, it is improper to combine the data because geographical factors necessarily influence the frequency with which particular occupations are drawn upon. These geographical factors, to which I alluded above, are perhaps of as much importance as previous occupa- tion. All the factories studied in this report, except L.7, are situated in purely industrial districts. I think it would be well to examine the records of some wartime factories which "have been erected in towns not normally industrial in character. X. GENERAL CONCLUSIONS. The primary object of this inquiry was to ascertain the actual extent of industrial wastage among women employed in munitions factories, and this object may, I think, be deemed to have been attained. It is true that the analysis covers but a small proportion of the total number employed in munitions industries, but the actual scope of the inquiry relating to nearly 40,000 employees in 18 representative factories is sufficient to entitle the sample to be regarded as typical of existing conditions. The general sense of the later observations is concord ent with that of the fully analysed material provided by factories discussed in the earlier sections of this report and confirms the inferences provisionally drawn from that material. Some of the minor conclusions are, as noted in the text, subject to a 29 measure of uncertainty and are therefore provisional, but broader conclusions cannot be contested. In the first place, it has been proved that the magnitude of the problem of wastage is considerable, that there is much dis- continuity of employment, even in the best -managed factories, which must prejudicially affect output. In the second place, it has been shown that much of this wastage is inexplicable or at least unexplained. The greatest proportion of the losses is unaccounted for. Of a total number of 11,005 women who left, 6,697 gave no reason for so doing or a frivolous reason. This is not a satisfactory state of affairs. In the existing shortage we can ill afford to allow the ranks to be depleted without sound reason being rendered. There is no panacea for the disease, but there is one promising remedy : this is organised welfare work. We have already seen that in factories with well -organised welfare departments the wastage unaccounted for is below the average, even if we must hesitate to bring the two things into the simple relation of cause and effect. The experience of such factories as L.7 is instructive in this respect. No claim is set up that the statistical inquiry has brought to light an unknown truth in this connection ; but it has suggested an approximate measure of the magnitude of the effect. When we compare the varying rates of loss in different factories doing similar or identical work and remember what a large proportion of these losses is debited to the unexplained group, we realise the extent of the field open to welfare work. What is needed is no mere paper system of stringent inter- rogatories administered to the women who leave. Such a system might diminish the number of women falling within my N.S.R. class, but whether it would do more than stimulate the operative's ingenuity in inventing plausible excuses is doubtful. Intelligent and sympathetic ''following up" of absentees is more likely to produce good results than any addition to the mass of questionnaires and forms. which already encumber the tables and burden the lives of factory officials. In the third place, it appears that the general strain of factory life in itself is not worse borne by elder than by younger women, even in factories engaged upon chemical work which has often been associated with specific danger to the health of certain individuals. On the other hand, when the general strain of factory conditions is combined with more than a very moderate degree of actual physical work, women over 23 come less well through the ordeal. The difference is not indeed catastrophic, but it is distinct enough to make an attempt at systematic recruiting of the older women for the physically lighter labour and of the younger women for the more strenuous tasks worthy of practical attention. It is to be noted that the nature of the distinction is not the mere truism which a hasty reader might suppose. No statistics are necessary to prove that women really in the decline of life are unsuitable for heavy labour. The 30 change occurs during physical maturity, and its most plausible explanation is that the older women, of whom a larger proportion will be married and have young children than of the girls under 23, have not sufficient reserves of energy to meet the double demand made upon them. I consider that the above inferences are fully established by the evidence submitted. The more detailed hypotheses and suggestions as to various points are, I believe, worthy of con- sideration and are on that ground submitted to the reader's judgment. Finally, it is my pleasant duty to acknowledge the valuable assistance rendered by my co-workers, Miss Thompson, Miss Willcox and Miss Woods, upon whom has fallen nearly all the labour of collecting and much of the labour of analysing the data. I have also gratefully to acknowledge the financial assist- ance of the Medical Research Committee which enabled me to secure clerical help at the beginning of the inquiry. APPENDIX I. TABULAR MATTER TABLE 1. Losses and Gains in Munition Factory A.I. (1916). Month. Number employed. Number started. Number left for various reasons. Number left without reason. Number employed at end of month. June... 4,340 622 138 253 4,571 July... 4,571 695 48 87 5,131 August 5,131 464 67 220 5,308 September ... 5,308 578 174 45 5,667 October 5,667 1,054 236 248 6,237 31 7 ci i d IT; 10 ^ s 8 8 a i 1 C CD CO ^ - ~ o o s ii 8" ptj M CO ^ - CN CO - - 1C 3 . 1* o c^ * 1C - CO ~ - CO CM c c 8 ""3 i- 00 o CO - - - s S^ X QO 55 <" M w M * CD Tf w - * - CM o o CM 3x a i s d d C5 cc p* - 3 -^ CM Is QO W M ? - CO CM c ^ c c s |i d d CM cc X - " CM o o c 6 o X Is M CO - CM - c - o o p- CO 3 . II 1 O ; CM | S W M - CO - p c c c ll d d - - O * p p c o z II W I-H c CO - Tt- O - o c -. -f O co c o 1 c 11 . . . . 1 IH o 5 2 T I 1C 1 3 1C 1 cr X r*** I 1 ? \ X | o 1 afc T c CM T = 1 33 TABLE 4. Cordite Factory. Men. Women. . x enod. Joined. Left. Joined. Left. 1915. July 19th Aug. 27th ... 16 59 162 45 Aug. 28th Sept. 27th ... 44 18 153 33 Sept. 28th Oct. 27th ... 66 16 49 22 Oct. 28th Nov. 27th ... 75 41 10 30 Nov. 28th Dec. 27th ... 123 26 69 15 Dec. 28th Jan. 27th ... 108 37 247 43 1916. Jan. 28th Feb. 27th ... 114 37 369 117 Feb. 28th Mar. 27th ... 131 28 127 44 Mar. 28th April 27th ... 89 31 62 49 April 28th May 27th ... 82 49 182 70 May , 28th June 27th . . . 33 44 .16 54 June* 28th July 27th ... July;" 28th Aug. 27th ... 94 72 48 32 144 99 48 88 Aug. 28th Sept. 27th ... 65 32 84 113 Totals 1,112 498 1,773 771 (B2604 r B2367) 34 TABLE o. Rates of Loss (Percentages), Cordite Factory. Period of Service (completed months). 1 o 3 4 5 6 All Causes. Health. A. H. A. H. A. H. A. H. A. H. A. H. Women 10-8 4.4 6-7 2-4 6-5 2-7 5-8 1-8 5-7 1-8 4-9 1-3 3-2 1-5 Men ... 11-7 4-7 7-5 4-3 5-5 0-9 5-9 2-0 4-2 1-0 3-6 0-6 4-0 0-7 FACTORY L.I. TABLE 6.* All Ages (1st Jan., 19114^ Aug., 1914). T * n ~4-lt Numbers Left. Percentages. Juengtn ! No No of Number Other reason, Other reason. time, in months. employed. Bad suffi- or health, cient insuffi- reason. cient Total. Bad health. suffi- cient reason. or insuffi- cient TotaL reason. reason. o 1,116 7 32 49 88 627 2-867 4-391 7-885 1 976 5 32 33 70 512 3-279 3-381 7-172 2 869 3 18 40 61 345 2-071 4-603 7-020 3 801 3 14 29 46 375 1-748 3-620 5-743. 4 748 3 25 18 46 401 3-342 2-406 | 6-150 5 687 3 22 19 44 437 3-202 2-766 6-404 6 626 28 13 41 4-473 2-077 6-550 7 571 4 12 11 27 701 2-102 1-926 4-729 8 541 11 6 17 2-033 1-109 3-142 9 518 2 7 8 17 386 1-351 1-544 3-282 10 501 2 14 10 26 399 2-794 1-996 5-190 11 472 2 14 10 26 424 2-966 2-119 5-508 12 442 1 12 7 20 226 2-715 1-584 4-525 13 421 11 9 20 2-613 2-138 4-751 14 401 11 5 16 2-743 1-247 3-990 15 380 1 9 7 17 263 2-368 1-842 4-474 16 359 1 10 7 18 279 2-786 1-950 5-014 17 336 8 2 10 2-381 595 2-976 18 322 1 506 311 1-553 i-sea 19 314 1 5 5 11 318 1-592 1-592 3-503 20 300 2 8 4 14 667 2-667 1-333 4-667 21 284 1 506 352 1-761 2-113 22 272 12 4 16 4-412 1-471 5-882 23 253 7 4 11 2-767 1-581 4-348 24 232 10 5 15 4-310 2-155 6-466 Totals 12,742 42 342 305 689 330 2-684 2-394 5-408 * In this and subsequent tables, the column headed " Number Employed " gives the number of women " exposed to risk " of leaving for each month of service, e.g., 1,116 passed through the first month or left during the first month. 35 FACTORY L.I. TABLE 7. All Ages (4th Aug., 191 1st Dec., 1916). T U Numbers Left. Percentages. .Length No No of Number Other reason, Other reason. time, in months. employed. Bad suffi- health. cient reason. or insuffi- cient Total. Bad health. sum- cient reason. or insuffi- cient Total. reason. reason. , 4,479 26 59 289 374 580 1-317 6-452 8-350 1 3,989 13 43 154 210 326 1-078 3-861 5-264 2 3,710 18 48 113 179 485 1-294 3-046 4-825 3 3,412 14 35 99 148 410 1-026 2-902 4-338 4 3,084 11 34 86 131 357 1-102 2-789 4-248 5 2,774 10 33 56 99 360 1-190 2-019 3-569 6 2,520 10 14 47 71 397 556 1-865 2-817 7 2,297 2 10 33 45 087 435 1-437 1-959 8 2,150 4 11 24 39 186 512 1-116 1-814 9 1,967 11 13 24 48 559 660 1-220 2-440 10 ,791 4 10 18 32 223 558 1-005 1-787 11 ,615 2 15 8 25 124 929 495 1-548 12 ,421 3 6 9 18 211 422 633 1-267 13 ,292 3 7 8 18 232 542 619 1-393 14 ,126 4 9 13 355 799 1-155 15 995 1 5 12 18 101 503 1-206 1-809 16 834 2 6 8 240 719 959 17 749 1 3 2 6 134 401 267 801 18 648 1 3 6 10 154 468 926 1-543 19 583 1 6 7 172 1-029 1-201 20 514 1 1 2 4 195 195 389 778 21 420 3 5 8 714' 1-191 1-905 22 317 1 1 4 6 315 315 1-262 1-893 23 266 1 1 2 376 376 752 24 225 1 2 3 444 889 1-333 43,178 137 364 1,021 1,522, 317 -843 1 2-366 3-526 FACTORY L.I. TABLE 8. All Ages (4th Aug., 1914 1^ Dec., 1916). Numbers Left. Percentages. Xj6ngtii No No of Number Other reason, Other reason, time, in months. employed. Bad health. suffi- cient or insuffi- Total. Bad health. suffi- cient or insuffi- Total. reason. cient reason. cient reason. reason. 2,727 15 48 198 261 550 1-761 ' 7-261 9-571 1 2,264 24 43 97 164 1-060 1-899 4-285 7-244 2 2,003 15 26 101 142 749 1-298 5-042 7-089 3 1,663 4 25 65 94 241 1-503 3-909 5-652 4 1,372 6 10 44 60 437 729 3-207 4-373 5 1,134 7 16 29 52 617 1-411 2-557 4-586 6 987 1 9 18 28 101 912 1-824 2-837 7 898 3 10 12 25 334 1-114 1-336 2-784 8 785 3 5 11 19 382 637 1-401 2-420 9 673 3 4 10 17 446 594 1-486 2-526 10 625 2 3 5 10 320 480 800 1-600 11 542 5 11 16 923 2-030 2-952 12 456 2 2 8 12 439 439 1-754 2-632 13 393 1 1 3 5 254 254 763 1-272 14 349 2 2 4 573 573 1-146 15 330 2 1 1 4 606 303 303 1-212 16 289 4 4 8 1-384 1-384 2-767 17 262 18 242 1 3 3 7 413 1-240 1-240 2-893 19 211 1 2 3 474 948 1-422 20 154 1 1 649 649 21 111 1 1 901 901 22 85 4 4 4-706 4-706 23 78 1 2 3 1-282 2-564 3-846 24 67 1 1 1-493 1-493 Totals 18,700 95 217 629 941 508 1-160 1 3-363 5-031 36 FACTORY L.I. TABLE 9. Ages 13-17 (1st Jan., 19114^ Aug., 1914). T - . ,41. Numbers Left. Percentages. Length No No I of Number Other reason, Other reason. time, employed. Bad suffi- or Total. Bad i suffi- or Total. in months. health. cient reason. insuffi- cient health. cient reason. insuffi- cient reason. reason. 588 5 24 20 49 850 4-082 3-401 8-333 1 510 1 15 15 31 196 2-941 2-941 6-078 2 466 8 24 32 1-717 5-150 6-867 3 431 2 8 14 24 464 1-856 3-248 5-568 4 405 3 10 12 25 741 2-469 2-963 i 6-173 5 373 2 9 12 23 536 2-413 3-217 6-166 6 342 14 10 24 4-094 2-924 7-018 7 317 2 6 8 16 631 1-893 2-524 | 5-047 8 301 8 4 12 2-658 1-329 ! 3-987 9 286 1 4 5 10 350 1-399 1-748 i 3-497 10 276 2 11 7 20 725 3-986 2-536 1 7-246 11 254 1 7 7 15 394 2-756 2-756 5-906 12 236 5 3 8 2-119 1-271 3-390 13 227 3 4 7 1-322 1-762 3-084 14 220 6 1 7 2-727 455 3-182 15 209 1 6 3 10 478 2-871 1-435 4-785 16 196 1 5 3 9 510 2-551 1-531 4-592 17 184 4 4 2-174 2-174 18 178 19 176 1 1 4 6 568 568 2-273 3-409 20 168 1 o 3 4 595 1-786 2-381 21 163 3 3 1-840 . 1-840 22 154 5 2 7 3-247 1-299 4-545 23 144 3 3 6 2-083 2-083 4-167 34 130 7 4 11 5-384 3-077 8-462 Totals 6.934 23 172 168 363 332 2-481 2-423 5-235 FACTORY L.I. TABLE 10. Ages 18-22 (1st Jan., 19114^ Aug., 1914). Numbers Left. Percentages. Length No ; No of Number Other reason, Other reason, time, in months. employed. Bad health. suffi- cient reason. or insuffi- cient Total. Bad suffi- health. cient reason. or insuffi- cient Total. reason. [ reason. 459 1 6 26 33 218 ! 1-307 5-664 7-190 407 4 16 15 35 983 3-931 3-686 8-600 2 353 2 7 11 20 567 '' 1-983 3-116 5-666 3 330 4 14 19 303 1-212 4-242 5-758 4 308 12 6 18 3-896 1-948 5-844 5 284 10 6 17 352 3-521 2-113 5-986 6 262 11 3 14 4-198 1-145 5-344 7 236 2 5 3 10 847 2-119 1-271 4-237 8 223 2 2 4 -897 897 1-794 9 216 1 3 3 7 463 i 1-389 1-389 3-241 10 209 3 3 6 1-435 1-435 2-871 11 203 1 5 3 9 493 2-463 1-478 4-433 12 193 6 4 10 3-109 2-073 5-181 13 183 7 - 4 11 3-825 2-186 6-011 14 172 5 4 9 ! 2-907 2-326 5-233 15 162 3 4 7 1-852 2-469 4-321 16 154 4 4 8 ' 2-597 2-597 5-195 17 144 4 1 5 i 2-778 694 3-472 18 137 1 4 5 703 2-919 3-649 19 132 4 1 5 3-030 758 3-788 20 126 1 7 1 9 794 1 5-556 794 7-143 21 116 1 2 3 862 1-724 2-586 22 113 7 2 9 6-195 1-770 7-965 23 104 4 1 5 3-846 962 4-808 24 97 3 1 4 3-093 1-031 4-124 Totals 5,323 16 144 122 282 301 2-705 2-292 5-298 37 TABLE 11 FACTORY L.I. -Ages 23-27 (1st Jan., 19114^ Aug.. 1914). Numbers Left. Percentages. Length of time, in months. Number employed. No Other reason, Bad suffi- or health, cient insuffi- reason. cient Total. Other Bad Biiffi- health. cient reason . No reason, or insuffi- cient Total. reason. n-asoii. 61 1 2 3 6 1-639 3-279 4-918 9-836 1 53 1 3 4 1-887 5-660 7-547 2 46 1 3 4 8 2-174 6-522 8-696 17-391 3 38 2 1 3 5-263 2-632 7-895 4 33 3 3 9-091 9-091 5 29 3 1 4 10-345 3-448 13-793 6 21 3 3 14-286 14-286 7 17 1 1 5-882 5-882 8 16 1 1 6-250 6-250 9 15 10 15 11 15 2 2 . 13-333 13-333 12 13 1 1 2 7-692 7-692 15-385 13 11 1 1 2 . 9-091 9-091 18-182 14 9 15 9 . 16 9 1 1 11-111 11-111 17 8 1 1 12-500 12-500 18 7 1 1 14-286 14-286 19 6 . 20 6 1 1 16-667 16-667 21 5 22 5 . 23 5 24 5 ; Totals 457 3 26 14 43 656 5-689 3-063 9-409 FACTORY L.I. TABLE 12. Ages 13-17 (th Aug., 1914 1st Dec., 1916). T Avirvtl* Numbers Left. Percentages. l>engtn of time, in months. Number employed. Other Bad suffi- health. cient reason. No reason, or insuffi- cient Total. Bad health. No Other reason, suffi- or cient insuffi- ; reason, cient Total. reason. ; reason. ,732 11 19 93 123 635 1-097 5-370 I 7-102 1 ,555 1 13 67 81 064 836 4-309 ! 5-209 2 ,435 4 13 47 64 279 906 3-275 | 4-460 3 ,326 9 10 47 66 679 754 3-544 I 4-977 4 ,197 5 11 37 53 418 919 3-091 4-427 5 ,084 7 9 20 36 646 830 1-845 3-321 6 970 1 3 25 29 103 309 2-577 2-990 7 881 5 11 16 568 1-249 1-816 8 829 1 2 12 15 121 241 1-448 1-809 9 774 5 4 14 23 646 517 1-809 2-927 10 721 5 5 693 -693 11 672 4 4 8 595 595 1-190 12 605 1 2 3 165 331 496 13 558 1 3 6 10. 179 538 1-075 1-792 14 487 1 8 9 205 1-643 , 1-848 15 425 1 4 5 235 941 1-176 16 359 2 4 6 557 1-114 ; 1-671 17 319 1 2 1 4 313 627 313 1-254 18 265 1 2 3 377 755 I 1-132 19 241 3 3 1-245 1-245 20 207 1 1 o 483 483 966 21 172 o 4 6 1-163 2-362 3-488 22 129 1 1 775 775 23 103 1 1 2 971 971 1-942 24 82 1 1 1-220 1-220 Totals 17,128 50 106 418 574 292 619 2-440 3-351 38 FACTORY L.I. TABLE 13. Ages 18-22 (4th Aug., 1914 Is* Dec., 1916). Numbers Left. Percentages. Length No No of Number Other reason, Other reason, time, employed. Bad suffi- or Total. Bad suffi- j or Total. in mont is. health. cient insuffi- health. cient : insuffi- reason. cient ' reason. cient reason. reason. 1,444 7 15 119 141 485 1-039 8-241 | 9-765 1 1,273 6 15 49 70 471 1-178 3-849 5-499 2 1,188 6 17 32 55 505 1-431 2-693 4-630 3 1,098 5 12 31 48 455 1-093 2-823 i 4-372 4 995 3 9 25 37 302 ! -905 2-513 3-719 5 910 IJ 19 26 769 2-088 2-857 6 846 1 7 13 21 118 827 1-537 2-482 7 773 2 5 15 22 259 647 1-940 2-846 8 729 2 5 7 14 274 686 -960 . -920 9 664 3 3 5 11 452 452 i -753 [ -657 10 617 3 8 10 21 486 1-297 1-621 -404 11 556 1 8 1 10 180 1-439 180 799 12 494 1 4 4 9 202 810 810 822 13 445 2 2 2 6 494 449 449 348 14 398 2 1 3 503 251 754 15 363 ! 2 6 8 551 1-653 2-204 16 309 o ! o 1 1 324 324 17 285 1 1 351 351 18 262 1 3 4 382 1-145 1-527 19 234 b 1 3 4 427 1-282 1-709 20 208 21 170 1 1 .ggg i 588 22 127 1 4 5 787 ! 3-150 3-937 23 108 24 93 2 2 2-151 2-151 Totals 14,589 42 127 351 520 288 | -871 2-407 3-567 FACTORY L.I. TABLE 14. Ages 23-27 (4th Aug., 19141** Dec., 1916). Numbers Left. Percentages. Length No No of Number Other reason, Other reason, time, in months. employed. Bad health. suffi- cient reason. or insuffi- cient Total. Bad health. suffi- cient reason. or Total, insuffi- cient reason. reason. 609 I 7 40 48 164 1-149 6-568 7-882 1 546 4 9 23 36 733 1-648 4-212 6-593 2 502 5 7 17 29 996 1-394 3-386 5-777 3 458 7 12 19 1-528 2-620 4-148 4 420 10 16 26 2-381 3-810 6-190 5 357 1 6 Q 16 280 1-681 2-521 4-482 6 324 4 1 1 6 1-235 309 309 1-852 7 300 1 1 333 -333 8 288 4 2 6 1-389 694 j 2-083 9 267 2 3 4 9 749 1-124 1-498 3-371 10 231 1 2 3 433 866 1 1-299 11 203 1 2 493 493 -985 12 183 2 3 6 1-093 546 1-639 3-279 13 168 1 595 595 14 141 1 709 709 15 121 1 2 826 826 1-653 16 100 17 88 1 1-136 1-136 18 69 1 1-449 1-449 19 64 20 61 1 1 1-639 1-639 21 47 1 1 2-128 22 32 . 23 30 24 26 1 Totals 5,635 21 64 130 215 373 1-136 2-307 3-815 39 TABLE- 15. Ages 28 FACTORY L.I. tint I over (4th Aug., 1914- 1st Dec., 1916), Numbers Left. Percentages. Length No No of Number Other reason, Other reason, time, in months. employed. Bad health. BUffi- cient reason. or insuffi- cient Total. Bad suffi- health. cient reason . or insuffi- cient Total. reason. reason. 694 7 18 37 62 1-009 2-594 ! 5-331 8-934 1 615 2, 6 15 23 325 976 ! 2-439 3-740 585 3 11 17 31 513 1-880 : 2-906 5-299 3 530 6 9 15 1-132 | 1-698 2-830 4 472 3 4 8 15 636 847 1-695 3-178 5 423 Q 11 8 21 473 2-600 1-891 ' 4-965 6 380 4 3 8 15 1-053 789 2-105 i 3-947 7 343 6 6 1-747 1-749 8 304 1 3 4 329 987 1-316 9 262 1 3 1 5 382 1-145 382 1-908 10 222 3 3 1-351 1-351 11 184 1 2 2 5 543 1-087 1-087 2-717 12 139 13 121 1 1 826 826 14 100 15 86 1 2 3 1-163 2-326 3-488 16 66 o ! i 1 1-515 1-515 17 57 . 18 52 1 1 2 1-923 1-923 3-846 19 44 20 38 1 1 ' 2-632 2-632 21 31 22 29 23 25 . 24 24 Totals 5,826 24 67 122 213 412 1-150 2094 3-656 FACTORY L.2. TABLE 16. Ages 13-17 (1st Jan., 19114^ Aug., 1914). T onrrfVi Numbers Left. Percentages. JLengtn of Number No Other reason, Other No reason, time, in months. employed. Bad health. suffi- or cient insuffi- reason. cient Total. Bad health. suffi- cient reason. or insuffi- cient Total. reason. | reason. 149 14 11 25 9-396 7-383 16-779 1 119 5 9 14 4-202 7-563 11-765 2 103 2 3 4 9 1-942 2-913 3-883 8-738 3 93 T 5 7 1-075 ! 1-075 5-376 7-527 4 86 2 3 1-163 2-326 3-488 5 82 ( 1 1 1-220 1-220 6 78 1 2 1-282 1-282 2-564 7 76 4 5 1-316 5-263 6-579 8 71 2 3 1-408 2-817 4-225 9 67 i 10 66 1 1 1-515 1-515 11 64 ' 12 63 1 3 4 1-587 4-762 6-349 13 59 1 1 2 4 1-695 i 1-695 3-390 6-780 14 52 15 52 1 1 1-923 1-923 16 51 17 51 1 1 2 1-961 1-961 3-922 18 46 19 45 1 1 2-222 2-222 20 43 _ 21 42 1 1 2-381 2-381 22 40 1 1 2 2-500 2-500 5-000 23 38 1 1 2-632 2-632 24 36 4 4 11-111 11-111 Totals 1,672 5 37 | 48 90 299 2-213 2-871 5-383 40 FACTORY L.2. TABLE 17. Ages 18-22 (1st Jan., 19114^ Aug., 1914). Numbers Left. Percentages. Length No No of Number Other reason, Other reason, time, in months. employed. Bad health. suffi- cient reason. or insuffi- cient Total. Bad health. suffi- cient reason. or insuffi- cient Total. reason. reason. 298 1 8 11 20 336 2-685 3-691 6-771 1 262 2 10 9 21 763, 3-817 3-435 8-015 2 213 4 6 10 1-878 2-817 4-695 3 203 1 7 8 16 493 3-448 3-941 7-882 4 187 2 2 4 i 1-070 1-070 2-139 5 181 1 2 2 5 552 1-105 1-105 2-762 6 169 1 1 592 592 7 166 1 4 1 6 602 2-410 662 3-614 8 159 2 2 1-258 1-258 9 157 5 2 7 3-185 1-274 4-459 10 150 1 1 , 667 667 11 149 4 3 7 2-685 2-013 4-698 12 141 2 5 7 1-418 3-546 4-965 13 133 6 2 8 4-511 1-504 6-015 14 122 3 1 4 2-460 820 3-279 15 114 1 5 1 7 887 4-386 877 6-140 16 103 3 1 4 2-913 971 3-883 17 99 18 97 4 4 4-124 4-124 19 91 3 2 5 3-297 2-198 5-495 20 84 ; _ 21 83 1 2* 3 1-205 2-410 3-614 22 77 1 2 1 4 1-299 2-597 1-299 5-195 23 73 2 2 4 2-740 2-740 fS-479 24 67 1 1 1-493 1-493 Totals 3,578 9 81 61 151 252 2-264 1-705 4-220 FACTORY L.2. TABLE 18. Ages 13-17 (4th Aug., 1914 1st Dec., 1916). Numbers Left. Percentages. Length No No of Number Other reason. Other reason time, in months. employed. Bad health. suffi- cient reason. or Total, insuffi- cient Bad suffi- health. cient reason. or insuffi- cient Total. reason. | reason. 211 1 3 12 16 474 1-422 5-687 7-583 1 182 1 5 4 10 549 2-747 2-198 5-495 2 167 2 13 15 1-197 7-783 8-981 3 142 1 2 4 7 704 1-408 2-187 4-930 4 128 2 3 5 1-563 2-344 : 3-906 5 116 1 6 7 862 5-172 6-034 6 97 2 1 3 2-062 1-031 3-092 7 90 1 1 1-111 1-111 8 83 9 74 3 3 4-054 4-054 10 67 o 2 2 2-985 2-985 11 61 12 58 1 1 2 1-724 1-724 3-448 13 56 1 1 1-785 1-785 14 50 1 1 2-000 2-000 15 48 1 o 1 2-083 2-083 16 43 1 1 2-326 2-326 17 41 18 39 19 35 . . 20 28 21 23 . 22 16 __ . 23 15 1 1 6-667 6-667 24 13 Totals 1,883 4 22 50 76 212 1-168 2-655 4-036 41 TABLE 19.- FACTORY L.2. -Ages 18-22 (4th Aug., 1914 Is* Dec., 1916). Numbers Loft. Percentages. Length | No No of Number Other reason, Other reason. time, in employed. Bad health. suffi- or cient insuffi- Total. Bad sum- or health, cient insuffi- otah months. reason. ( cient reason, cient 1 reason. 1 reason. 804 5 15 67 87 622 1-866 8-333 10-821 1 668 7 16 33 56 1-047 2-395 4-940 8-383 2 588 5 4 25 34 850 -680 4-252 5-782 2 512 2 8 21 31 391 1-563 4-102 6-054 4 450 1 3 18 22 222 -667 4-000 ! 4-889 5 385 3 9 12 24 779 2-338 3-377 6-234 6 338 4 5 9 1-183 1-479 2-663 7 313 1 2 6 9 319 -639 1-917 2-875 8 292 1 2 8 11 342 -685 2-740 3-767 9 259 1 3 4 386 - 1-158 1-544 10 249 4 4 1-606 1-606 11 225 3 6 9 1-333 2-667 4-000 12 191 1 1 5 7 524 524 2-619 3-667 13 172 1 2 3 581 1-163 1-744 14 159 2 2 1-259 . 1-259 15 149 1 1 2 671 671 1-342 16 131 3 3 6 2-290 2-290 4-580 17 116 ' 18 110 3 3 6 2-727 2-727 5-454 19 92 1 1 1-087 1-087 20 71 1 1 1-408 1-408 21 48 1 1 2-083 2-083 22 39 o 3 3 7-692 7-692 23 36 1 1 2 2-778 2-778 5-556 24 30 1 1 3-333 3-333 Totals 6,427 32 76 227 335 498 1-183 3-632 6-213 FACTORY L.2. TABLE 20. Ages 23-27 (4th Aug., 19141^ Dec., 1916). Numbers Left. Percentages. Length No I No of Number Other reason, 1 Other reason, time, in months. employed. Bad sum- health, cient reason. or insuffi- cient Total. Bad suffi- health. cient reason. or insuffi- cient Total. reason. reason. 515 2 ! 19 46 67 388 3-690 8-932 13-010 1 404 5 9 20 34 248 2-228 4-950 8-416 2 355 1 13 21 35 282 3-662 5-915 9-859 3 275 3 16 19 ! 1-090 5-818 6-909 4 228 1 1 10 12 439 -439 4-386 5-263 5 186 3 ! 1 2 6 1-613 -538 1-075 3-226 6 165 1 5 6 606 3-030 3-636 7 146 1 3 1 5 685 2-055 i -685 3-425 8 123 1 1 2 813 813 1-626 9 108 1 3 4 926 2-778 3-704 10 97 1 1 2 1-031 1-031 2-062 11 81 2 4 6 2-469 4-938 7-407 12 66 1 1 1-515 1-515 13 58 14 51 16 47 16 44 1 1 __ 2-273 2-273 17 39 1 18 34 1 I 1 2-941 2-941 19 30 20 22 21 16 . 22 10 , 23 10 24 8 Totals 3,118 18 54 129 201 577 1-732 4-137 6-446 42 FACTORY L.2. TABLE 21. Ages 28 and over (4th Aug., 19141^ Dec., 1916), Numbers Left. Percentages. l^WMfwIl No No of Number Other reason Other mason time, in months. employed. Bad health. suffi- cient reason. or insuffi- Total - cient Bad health. suffi- cient reason. insuffi. Total, cient reason. I reason. 1,197 7 11 73 91 585 919 6-099 7-602 1 1,010 11 13 40 64 1-089 1-287 i 3-960 6-337 2 893 9 7 42 58 1-008 784 4-703 6-495 3 734 1 12 24 37 136 1-635 3-270 5-041 4 566 4 4 13 21 707 707 2-297 3-710 5 447 6 9 15 1-342 2-013 3-356 6 387 1 2 '7 10 258 ; -517 1-808 2-584 7 349 1 4 5 10 287 ; 1-146 1-433 2-865 8 287 2 2 2 6 697 -697 697 2-091 9 232 1 1 4 6 431 -431 1-724 2-586 10 212 1 1 2 471 -471 943 11 175 1 1 571 571 12 141 1 1 2 709 709 1-418 13 107 1 1 935 935 14 89 1 1 1-124 1-124 15 86 1 1 1-163 1-163 16 71 17 66 18 59 19 54 1 1 2 1-852 1-852 3-704 20 33 21 24 22 20 1 1 5-000 5-000 23 17 i 24 16 Totals 7,272 41 65 223 329 564 894 3-067 4-524 FACTORIES L.I AND L.2. TABLE 22. Previous Occupations of Employees (Percentages). Group. Factory Hands, not Muni- tions. Muni- tion Wor- kers. Domes- tic Ser- vants. Shop Assis- tants. Miscel- laneous. House- work and no previous employ- ment. Unre- corded. Peace time. Ages 13-17 (736) ... 21-33 5-43 8-83 6-11 1-63 54-21 2-44 War time. Ages 13-17 (1,933) 22-61 9-47 8-95 3-78 1-71 51-58 1-91 Peace time. Ages 18-22 (779) ... 29-53 16-05 18-87 5-26 3-85 25-41 1-03 War time. Ages 18-22 (2,265) 27-24 13-47 21-02 5-03 3-31 26-63 3-31 Peace time. Ages 23-27 (110) ... 20-91 30-00 22-73 5-45 91 16-36 3-64 ' W^r time. Ages 23-27 (1,143) 17-59 22-75 21-61 3-24 2-36 18-37 14-09 War time. Ages 28- (1,905) 12-86 19-11 12-18 1-57 2-89 12-65 38-74 43 TABLE 23. Numbers remaining at Work out of 1,000 Entrants in accordance with Various Rates of Loss. Remain- ing after months Cordite Factory (War time). L.2 Factory (War' time). L.I Factory (War time). L.I Factory (Peace time). L.I Factory (Peace time, ages 18-22). L.I Factory (War time, ages 18-22). L.I L.I Factory Factory (War (War time, time, ages ages 13-17). 23-27). L.I Factory (War time, ages 28-). 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1 892 904 917 921 928 902 929 921 911 2 832 839 868 855 848 853 881 860 877 3 778 779 826 795 800 813 841 811 830 4 733 735 791 749 754 778 799 777 807 5 691 703 757 703 710 749 764 729 781 6 657 671 730 658 668 727 739 696 742 TABLE 24. Tests of Consistency. (See Appendix.] Samples compared. Groupings used. Q. X 2 - P. Ages 13-1 7 and 18- 22. War time. L.I. Rates of Sickness. Oand 1,2 and 3, 4 and 5, 6 and over. 4-9605 4-9539 2928 Age group 1 8-22 1 2 3 4 5, 6 89-7939 97-0416 Very Peace-time and War -time ex- periences. 7, 8, 9-13, 14 and over. small. Age group 28- . L.I and L.2. War time. 0, 1, 2, 3, 4, 5, 6, 7, 8, 9-13, 14 and over. 154186 15-0892 1788 Age groups 13-17 and 18-22. L.I. War time. All losses. 0, 1,2,3,4,5,6, 7. 8, 9-13, 14 and over. 16-2453 16-2331 1324 44 FACTORY L.7. TABLE 25. Ages under 22 (1st Jan., 19161^ Feb., 1917). Numbers Left. Percentages. Length of ser- vice. Number employed. Other Bad Buffi- health, cient No . reason, or insuffi- Total. Bad health. No Other reason, suffi- ! or cient insuffi- Total. ~ reason. cient reason. cient reason. reason. 399 1 4 6 11 251 1-003 1-504 2-757 1 369 1 4 5 271 1-084 1-355 2 165 1 1 606 606 3 139 1 . 4 107 5 74 1 2 3 1-351 2-703 4-054 6 60 1 . 1 1-667 1-667 7 36 9 10 11 12 Over 12 . | Total... 1,354 4 10 7 21 295 739 517 1-551 FACTORY L.7. TABLE 26. Ages 23-27 (1st Jan., 1916 1st Feb., 1917). Numbers Left. Percentages. Length of ser- vice. Number employed. Other Bad suffi- health. cient No reason, or insuffi- Total. Bad health. Other suffi- cient No reason, or insuffi- Total. reason. cient reason. cient reason. reason. 202 2 5 1 5 12 990 2-475 2-475 5-941 1 180 3 1 4 1-667 556 2-222 2 88 1 1 1-136 1-136 3 75 1 1 1-333 1-333 4 62 I 1 1-613 1-613 5 43 _. 6 38 7 28 8 2 . 9 1 10 1 11 __ 12 Over 12 Total... 720 ' 4 9 6 , 19 | -556 1-250 833 2-639 45 FACTORY L.7. TABLE 27. Ages over 28 (1st Jan., 1916 1st Feb., 1917). Numbers Left. Percentages. Length of ser- vice. Number employed. Other Bad j suffi- health. ; cient No reason, or insuffi- Total. Bad health. Other suffi- cient No reason, or insuffi- Total. reason, cient reason. cient reason. ! reason. 300 9 13 22 3-000 4-333 7-333 1 269 352 10 1-115 1-859 -743 3-717 2 124 1 1 2 806 -806 1-613 3 98 2 2 2-041 2-041 4 78 i 5 61 2 1 3 3-279 1-639 4-918 6 45 . 7 26 8 2 . 9 1 | 10 1 11 1 12 Over 12 Total... 1,006 5 1 18 16 39 497 1-789 1-590 3-877 FACTORY L.7. TABLE 28. All ages (1st Jan., 19161^ Feb., 1917). Numbers Left. Percentages. Length of ser- vice. Number employed. Bad health. Other suffi- cient No reason, ; insuffi- Tota1 ' Bad health. Other suffi- cient No reason, or insuffi - Total. reason, cient reason, cient reason. ! reason. 901 3 18 24 45 333 1-998 2-664 4-994 1 818 4 12 3 19 489 1-467 367 2-323 2 377 2 1 1 4 531 265 265 1-061 3 312 1 o 3 321 641 962 4 247 1 1 405 405 6 178 3 3 6 1-685 1-685 3-371 6 143 1 1 699 699 7 90 8 5 9 3 10 3 11 2 12 1 j_ ; Over 12 Total... 3,080 13 37 29 79 422 1-201 942 2-565 46 TABLE 29. Observed Losses of L.I compared with those expected from the Experience of Factory L.I. Under 22. Ages 23-27. Over 28. All ages. L.7. L.I. L.7. L.I. L.7. L.I. L.7. L.I. 11 43-18 12 26-28 22 22-81 45 92-72 1 5 30-93 1 4 15-15 1 10 17-05 1 19 63-13 2 1 9-54 2 1 8-68 2 2 8-05 2 4 26-27 3 8-42 3 1 5-18 3 2 4-94 3 3 18-54 4 5-23 4 1 3-26 4 2-89 4 1 11-38 5 3 4-61 5 1-39 5 3 2-05 5 6 8-05 6 1 1-60 6 1-38 6 1-16 6 1 4-14 7 1-04 7 96 7 74 7 2-74 8 04 8 03 8 -04 8 11 9 02 9 04 q 03 9 09 10 02 10 02 10 10 04 11 04 11 11 11 04 12 04 12 12 12 04 Over Over Over Over 12 12 12 12 Total 21 104-71 Total 19 62-37 Total 39 59-76 Total 79 226-84 TABLE 30. Under 22. L.7 L.2 14-49 28-91 36-45 23-33 11-92 ! 5-61 37-47 1-49 16-59 4-71 5-37 1-61 9-58 2-48 428 806 L.7 19-16 39-72 Age 5-61 s 23-27. 7-94 13-53 2-80 11-21 214 L.2 18-56 21-97 35-42 76 4-17 1-33 17-80 528 Over 28. L.7 ... 37-12 27-30 6-75 5-21 5-52 3-68 14-42 326 L.2 13-60 11-60 29-55 1-92 1-25 1-00 41-07 1,198 47 3 n-xx COCD co ^ l^ M X X rr 1 c-i c-i -r c; CT; 01 co c 1 s CD-OiCOiOiOOCN CDi iO5iOr^CDCDt^-l^iOCN t> CO lO Si CO 1- CO Tf >"4'CNCOCNC^ 1/5 ' C4 |jy X ' CO 5 CN fo X 1 1 C CD X 05 X CO 1 1 2 co icor^i>t^io o t> < ^I>XCNXCDCN OCD^H XO^tOiXiCO IO^^H CO O S p^ *5 M l^ i > ' CM CN CO CM CM CM ^ < CO i CN Tf i c ! sl CO O CO CD O> X X 05 t^ 1 CM 1 T 1 T. T. -r CD 1 Tf 1 CO 1 o oicDr^SxcoS IS 1 CO i 5 ^l 02^ ~ ~ ~ ^ 11 o" i-H t " H O5 PH c o3 O .2*5 T3 u t^ CD CD SSS 11 I 1 1 t CO da COIOCDO5O5CM t>O rcocDt^^cNo |v- 10 I i GO CO CM CO CD -r 1 05 CM 1 I cr CO '. ^* <3 -Ki SO 3 A ~ &\ t^ *r> \ \ O5 1 1 X CO CO lO 1 1 CD 1 1 0) lO i-H Tf i C5 pq X 5^^^ CM^ ~r~ rH -* 5e ^ . 52 ^S 1 tO(NC^ '"0 CN ^ ^ ro 2 i CO ^ 1-1 2 ^-- ? t> 3 e . ^M i . fi CO X 10 10 CM ^ 1 9 g g g 2 , 5 g CO CO rf CO 1 ~ \ ~ lO (^ r-Xl005CMTfCM^ IT-. 1 CO CO 8 of ^ 11 CO "" "t w 00^ g J W H fc ll -3 o i -M || 111 CD H PQ -< HxcOCMt^CM | 1| 1 CN T* K 1* HJ h^ 9 4J w | |e. | X ^ O5 m w fe fe ! ? OiCDXOi-^OCNCO^t^CN i005X-H^050^05 t>. X CO | "cL CO CO CM -H r-H CO OiXr^iococM' i i 05 * 1 CO i EH 1 CO ca O 'CNCO'tlOCDI>XO5O 4J O CMCOTt'iOCDt^XO5O 43 | m H o H 48 co 05 1 I 8 t^ cq ci CM CO CO PQ ^ h4 & <5 fe i |5|SSS28|| H t>COlC^ti 11 iC^l^fUOC^ 1C | lg lOcor^LCc^ -^tot^c^i CD ^o co ^o co CD r**^ oo o^ i IO CTi CD "^ UO ict> ICOOO 1 1 8 lp CN CN r-4 T-I i 1 CO ii. |.| 1 ill II 1 1 II CD CO c^ict^-H c-*or> '-OiCDOS t^CDOOO 1 COCOCNO t^OCOOO 1 (N c c H C^J ^ ^ ^ ^H ^ 11 1 ^ t > NC^ Oi t> ic i i fi i 3'E. C Tf CO Ol i i f < i i > :| (N ( i j H J3 Crt 4 ^D ^^ C^ CO ^* ^C CD t^ 00 O^ OO COCD-* i-icoai CO ^O lOTfC^^^^C^^ 00 U2 rf C^ ^H ^ CO C^ CS i J ^O5COCNI rf CN - ^ CD t^ C^N O Tf *- I-H 00 O Tf co op 10 p ^r ~ CM ~ CM "T 00 O 10 p ~ CM 8 I I I I I I I I i 55 I I I I I I I I *> .. t* ?-. We* 1 . II to ' I I I I "" I I I I I H I I I I I I I M 2i - C1 I II I I I I I E . a o s ~~ ,3 CO o^c^co^ioccit^ooaj I 50 CNCNO-.^-OCDCDCNOClC tN i QcSo^^O'i^^S^CO O o H r-CNi-nCNCNCNCOCNC^O CN 1 l ^COCDCNOOOOCD- COt^ CD p$ oSfi co r^* ^~< c^ r i GO oc I ^"^ oo C"** r- ,-, cq CD ^H M t 49 tiO ** I-H to "^t* i < t& o^ cr^ CN CN ^ ,^ [>x ^i c^ Q^I ^} cNI CO I 00 PI "5 e Tt-^i ^ 1C CN CD CO CD CN CN 1C Oi C7i 1 1 Tt< CD O lO CO t^ lO 1 1 -*t 8 *l r-H ^ CO ll 00 00 CN Tf 05 CO CD t^ O O CTi Oi iO 1 1 Tf Tt CN CD iO CN CD 1 1 Tf H '""' 1 00"*C7iCT5iOI>'* l CNi i CO CN CO r-< i 1 i- 2 1 la (2s|| iCO)i-^^tCDCO-< Ir-CN CN 1 H* a s lg 2 S of J CDCD iCNC^--<^-iCN 1 1 CN 3 ii O CO iO O5 rf CN 1 1 ^H 5 s i TJ'S ii t^ CD CN Tf CO ^ CN 1 1 CN w i c H 1C f 3 J2 r fi O5CDCDCDOCOCNt^^CN ^t CN CT) CD lO CO i i 0> i 1 CO D C> CN CO "^f 1 ^O CD t^ OO O^ S eg o H s^ F H ^f O> CN CN "t CN -< CO CN >O O> ^ O> J> O I rf CO 00 co CN co ir> p ^H 1 en co op i-H CN CN - " CN CO 1/5 00 CO CO 1-1 (X CD CO CO 00 S I F^ CO 00 " "t O I "t CO 00 rll CO ^ CO 1C r* co ^f i o ^-^ t> oo co i 11 i i O

i-l O r-l 00 CD CO rf co rt CD r^ 2 S os co CO 00 1-1 O O < TT ^ I 1 t^ CO l> 1C 1 a CO . co 1 1 I rr cp ^t 1 cp ~ 1 1 1 CO 05 02^ Medical Rejection. t^ CO X 1 1 1 ?cp 1 1*3-1 ~ CO 1C CM . CO O^ CO Oi 1C If T-C T- ~ 1 1 U5 III t^ 1C CD 1 1 1C III w 3 CMXXCMCOOC ^ \ ~ H ~ ~ . 1 cc ijf COCDX^CMCO^ | | | ^ >c i r 1 Other ifficient eaaon. CD 1C .-r I CO ^ 1 1 | 05 a K 3 if 1 1 1 I-" 1 1- 1 1C CD Q ' 1 II 1 1 1 t W j i Employed. ^cococNCM^ 05 ^ ^ 1 " CO < if | [1 i j 3 Jg "-I O ( H w H CDCOt^COXTfiCCDiCOO O5CMCM" o^ ^t* r^ oo ^~^ CCG^CCCC' < CD iO ^ l> CO O O5 -- -*g CTJOXiCt^CT^r^CO ^CDO:COt^CDC75a5 1 1 COCOCOCMCDCD^X 1 1 CM 1C CO ^H 05 s" oo r^ cji CD i^^* co '^ oo CD I COCDCOXCDO5 TfXiO 1 O5 R $ 11 1 CO ^ X CM Tf CM 05 O "^ CD O5 X O O5 1 1 1 1 ic ic ic co rr rf ic 1 1 1 1 | OJ | O5CDXF > ^- < t > O5^* < CO'^'' ' 2 fr 1C rr CM CO CM 8 1 2 O>-iTtTtT^co^<^^ XCMCCt^CO'^fCOCM'-HCD S iCCMt^CMOt^COCM i CO CM CM >-i ^ O O^CMCO"tiOCDr^XO5O 1 ^^ H 52 FACTORY H.3. TABLE 41. Ages 18-22. (1st March, 1916 IsZ March, 1917.) Length Numbers Left. Percentages. of time, in months. Number employed. Ill health. Ill health. Other suffi- cient Insuffi- cient Total. Other suffi- cient Insuffi- cient Total. reason. reason ' i reason. reason. 379 12 14 26 3-166 3-694 6-860 1 291 4 14 6 24 1-375 4-811 2-062 8-247 2 237 2 9 12 23 844 j 3'797 5-063 9-705 3 181 2 6 9 17 1-105 3-315 4-972 9-392 4 143 2 1 4 7 1-399 699 2-797 4-895 5 132 2 6 8 1-515 4-545 6-061 6 123 5 5 4-065 ! 4-065 7 106 1 1 4 6 943 943 3-774 5-660 8 67 2 1 2 5 2-985 1-493 2-985 7-463 9 40 10 36 11 14 j 12 ; Totals 1,749 13 46 62 121 743 2-630 3-545 6-913 FACTORY H.3. TABLE 42. Ages 23-27. (1st March, 1916 1st March, 1917.) Numbers Left. Percentages. Length of time, in months. Number employed. Ill health. Other suffi- cient Insuffi- cient Total. Ill health. Other suffi- cient Insuffi- cient Total. reason. reason. reason. reason. 340 5 13 15 33 1-471 3-824 4-412 9-706 1 247 6 14 8 28 2-429 5-668 3-239 | 11-336 2 206 2 10 9 21 971 4-854 4-369 10-194 3 169 3 6 13 22 1-775 3-550 7-692 13-018 4 135 4 7 11 2-963 5-185 8-148 5 123 3 1 5 9 2-439 813 4-065 7-317 6 109 3 2 10 15 2-752 1-835 s 9-174 13-761 7 86 2 2 2 6 2-326 2-326 2-326 6-977 8 49* 2 2 4-082 4-082 9 41 . 1 1 2-439 2-439 10 33 1 1 .4 3-030 3-030 11 16 j 12 i Totals 1,554 24 52 73 149 1-544 3-346 4-698 9-588 FACTORY H.3. TABLE 43. Over 28. (1st March, IQIQlst March, 1917.) Length Numbers Left. Percentages. of time, in months. Number employed. Ill ?uffi r ;insuffi - health. d cient Total " Other 111 suffi- . health, cient Insuffi- cient Total. reason. reason ' reason. reason. - 834 10 30 36 76 1-199 3-597 4-317 9-113 1 611 13 42 20 75 2-128 6-874 3-273 12-275 2 506 6 23 18 47 1-186 4-545 3-557 9-289 3 410 4 20 25 49 976 4-878 6-098 11-951 4 320 4 10 15 29 1-250 3-125 4-688 9-063 5 286 1 6 14 21 350 2-098 4-895 7-343 6 248 5 2 9 16 2-016 806 3-629 6-452 7 203 3 7 10 1-478 3-448 4-926 8 140 6 6 4-286 4-286 9 99 1 1 2 1-010 1-010 2-020 10 91 3 3 3-297 3-297 11 47 2 2 4-255 4255 12 Totals 3,795 43 137 156 336 1-133 3-610 4-111 8-854 53 FACTORY H.2. TABLE 44. Ages 18-22. (1st July, 1916 1^ March, 1917.) Length of time, in months. Number employed. Numbers Left. Percentages. Ill health. Other suffi- cient reason. Insuffi- cient reason. Total. Ill health. Other suffi- cient reason. Insuffi- cient reason. Total. 1 2 3 4 5 6 7 8 Totals 216 150 108 74 53 49 46 21 o 6 5 9 6 2 4 2 2 4 2 17 13 4 4 2 2 4 2 926 1-333 1 852 2-778 3-333 4-167 4-000 1-852 5-405 3-774 4-082 8-696 9-524 7-870 8-667 3-704 5 405 3 774 4-082 8-fi96 9-524 717 6 11 3.1 48 837 1-534 4-324 6-695 FACTORY H.2. TABLE 45. Ages 23-27. (1st July, 1916 Is* March, 1917.) Length Numbers Left. Percentages. of time, in months. Number employed. Ill health. Other suffi- cient Insuffi- cient Total. Ill health. Other suffi- cient Insuffi- cient Total. reason. reason. , reason. reason. 192 6 11 18 35 3-125 5-729 9-375 18-229 1 115 1 2 3 870 1-739 2 609 2 97 1 1 4 6 1-031 1-031 4-124 6-186 3 63 1 2 3 1-587 3-175 4-762 4 40 1 1 2-500 2-500 5 35 6 33 . 1 3 4 3-030 9-091 12-121 7 14 1 1 . 7-143 7-143 8 - Totals 589 8 14 31 53 1 -358 2-377 5-263 8-998 FACTORY H.2. TABLE 46. Over 28. (1st July, 19161^ March, 1917.) Length Numbers Left. Percentages. of time, in months. Number employed. Other 111 suffi- health. cient Insuffi- cient Total. Other 111 suffi- health. cient Insuffi- cient Total. reason. reason. reason. reason. 464 21 33 31 85 4-526 7-112 66*1 18-319 1 332 5 9 21 35 1-506 2711 6-325 10-542 2 258 2 6 15 23 775 2-326 5-814 8-915 3 195 1 8 9 513 4-103 4-615 4 152 2 3 5 10 1-316 1-974 3-289 6-579 5 137 , j 4 5 -730 2-920 3-650 6 123 1 1 6 8 813 -813 4-878 6-504 7 62 3 3 4-839 4-839 8 3 9 Totals 1,726 32 53 93 178 1-854 3-071 5-388 10-313 54 FACTORY H.I. TABLE 47. Ages 18-22. (1st Aug., 19161^ March, 1917.) Length Numbers Left. Percentages. of time, in months. Number employed. Other , 111 sum- In ?u ffi - | health. ! cient cient TotaL Ill health. Other l_ sum- In ? uffi " cient cient Total. reason. i reason - 1 reason. [ reason - 297 1 11 10 22 337 | 3-704 3-367 7-407 1 208 1 4 9 14 481 1-923 4-327 6-731 2 170 1 3 4 8 588 1-765 2-353 4-706 3 125 . 4 93 2 1 3 2-151 1-075 3-226 5 60 1 1 2 1-667 1-667 3-333 6 11 7 ! Totals 964 3 | 21 25 49 311 2-178 | 2-593 5-083 FACTORY H.I. TABLE 48. Ages 23-27. (1st Aug., 1916 1st March, 1917.) Length Numbers Left. Percentages. in months. employed. Ill health. Other suffi- cient Insuffi- cient Total. Ill health. Other suffi- cient Insuffi- cient Total. reason. reason. reason. reason. 291 2 20 18 40 687 6-873 6-186 13-746 1 193 2 10 8 20 1-036 5-181 4-145 10-363 2 157 2 4 9 1-274 2-548 1-911 5-732 3 128 1 2 5 8 781 1-563 3-906 6-250 4 88 3 2 3 8 3-409 2-273 3-409 ! 9-091 5 67 3 1 1 5 5-263 1-754 1-754 8-772 6 5 ! ; 7 Totals 919 13 39 38 90 1-415 4-244 4-135 9-793 FACTORY H.I. TABLE 49. Over 28. (1st Aug., 1910 1st March, 1917.) Length Numbers Left. Percentages. in months. employed. Ill health. Other suffi- cient Insuffi- cient Total. Ill health. Other suffi- cient Insuffi- cient Total. reason. reason. reason. reason. 720 10 52 46 108 1-389 7-222 6-389 15-000 1 486 6 15 27 48 1-235 3-086 5-556 9-877 2 400 4 11 13 28 I -000 2-750 3-250 7-000 3 316 3 4 10 17 949 1-266 3-165 I 5-380 4 236 4 5 9 1-695 2-119 3-814 5 158 2 o 8 12 1-266 1-266 5-063 7-596 6 11 y Totals 2,327 25 88 109 222 1-074 3-782 4-684 9-540 55 FACTORIES H.I, H.2 AND H.3. TABLE 50. Ages 18-22. Length Numbers Left. Percentages. r*f t irrm XT K*v* in months. iN umoer employed. Ill health. S " Total. Ill health. Other suffi- cient Insuffi- cient Total. reason.! 8011 - reason. reason. 892 3 29 33 65 336 3-251 3-700 7-287 1 649 7 23 21 51 1-078 3-544 3-236 7-858 2 515 5 12 18 35 971 2-330 3-495 6-796 3 380 2 6 13 21 526 1-579 3-421 5-526 4 289 2 3 7 12 692 1-038 2-422 4-152 5 241 3 9 12 1-245 3-734 4-979 6 180 9 9 5-000 5-000 7 127 1 1 6 8 787 787 4-724 6-299 a 67 2 1 2 5 2-985 1-493 2-985 7-463 9 40 10 36 11 14 i 12 - Totals 3,430 22 78 118 218 | -641 2-915 4-082 6-997 FACTORIES H.I, H.2 AND H.3. TABLE 51. Ages 23-27. Length Numbers Left. Percentages. of time, in months. Number employed. Ill health. Other suffi- cient Insuffi- cient Total. Ill health. Other suffi- cient Insuffi- cient Total. reason. reason. reason. reason. 823 13 44 51 108~ 1-580 5-346 6-197 13-123 1 555 8 25 18 51 1-441 4-506 3-243 9-189 2 460 5 15 16 36 1-087 3-261 3-478 7-826 3 360 5 8 20 33 1-389 2-222 5-556 9-167 4 263 3 6 11 20 1-141 2-281 4-183 7-605 5 215 6 2 6 14 2 791 930 2-791 6-512 6 147 3 3 13 19 2-041 2-041 8-844 12-925 7 100 2 2 3 7 2-000 2-000 3-000 7-000 8 49 2 2 4-082 4-082 9 41 1 1 2-439 2-439 10 33 1 1 3-030 3-030 11 16 12 Totals 3,062 45 105 142 292 1-470 3-429 4-637 9-536 FACTORIES H.I, H.2 AND H.3. TABLE 52. Over 28. Length Numbers Left. Percentages. of time, in months. Number employed. Other 111 suffi- health. cient Insuffi- cient Total. Other 111 suffi- health. i cient Insuffi- cient i Total. ! reason. reason. / reason. reason. i 2,018 41 115 113 269 2-032 5-699 5-600 13-330 1 1,429 24 68 68 158 1-679 4-619 4-759 11-057 2 1,164 12 40 46 98 1-031 3-436 3-952 8-419 3 921 8 24 43 75 869 i 2-606 4-669 8-143 4 708 6 17 25 48 847 ! 2-401 3-531 6-780 5 581 3 9 26 38 516 1-549 i 4-475 6-540 6 382 6 3 15 24 1-571 -785 3-927 6-283 7 265 . 3 10 13 1 1-132 3-774 4-906 8 143 6 6 4-196 1 4-196 9 99 1 1 2 1-010 1-010 2-020 10 91 3 3 3-297 3-297 11 47 2 2 4-255 4-255 12 - 1 - Totals 7,843 100 278 ! 358 '736 1-274 3-542 ; 4-562 ! 9-378 56 FACTORIES H.I, H.2 AND H.3. TABLE 53. All Ages. Length Numbers Left. Percentages. of time, in months. Number employed. Other i_ 111 sum- In ? uffi - health. cient cient Total. i Other L 111 sum- Insuffi - health. cient cient Total. reason. reason - reason. reason ' 3.733 67 188 197 442 1 1-527 5-036 5-277 11-840 1 2,633 39 114 107 260 1-481 4-330 4-064 9-875 2 2,139 22 67 I 80 169 1-029 3-132 3740 7-901 3 1,661 15 38 ; 76 129 903 2-288 4-576 7-760 4 1,262 11 26 43 80 873 2-063 3-413 6-349 6 1,037 9 14 41 64 868 1-350 3-954 6-172 6 709 9 6 37 52 1-269 -846 5-219 7-334 7 492 3 6 19 28 610 1-220 | 3-862 5-691 8 259 2 1 10 13 772 -386 3-861 5-019 9 180 1 2 3 -556 i 1-111 1-667 10 160 4 4 2-500 2-500 11 77 2 2 2-597 2-597 12 Totals 14,340 167 | 461 | 618 1,246 1-165 3-215 4-310 8-689 PROJECTILES. TABLE 54. Ages 18 22. Fac- Muni- Shop House- Un- Ser- Laun- Trans- Miscel- tory. tions. Assis- hold known. vants. dry. port. laneous. tants. Work. H.3 ... 41-35 15-14 8-17 6-97 14-90 9-13 2-64 1-68 H.2 ... 43-28 15-55 8-40 6-30 10-50 10-08 3-36 2-10 42 H.I ... 25-86 34-89 4-98 4-05 21-50 4-67 62 1-87 1-56 Ages 23 27. Fac- tory. Muni- tions. Shop Assis- tants. House- hold Work. Un- known^ Ser- vants. Laun- dry. Trans- port. Miscel- laneous. H.3 ... 33-97 11-96 11-41 10-33 18-48 9-51 1-63 1-09 1-63 H.2 ... 31-25 13-94 4-33 10-58 18-75 12-02 4-33 1-92 2-88 H.I ... 18-83 29-55 3-57 13-31 26-95 4-87 1-30 1-62 Over 28. Fac- tory. Muni- tions. Shop Assis- tants. House- hold Work. Un- known. Ser- vants. Laun- dry. Trans- port. Miscel- laneous. H.3 ... 23-97 12-42 11-11 16-23 23-64 6-97 3-59 33 1-74 H.2 ... 30-69 13-21 4-07 19-51 19-92 8-13 3-46 41 61 H.I ... 17-94 31-48 2-97 12-77 30-19 3-10 26 1-03 26 57 ALL FACTORIES. TABLE 55. Ages 13 17. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other suffi- cient No suffi- cient Total. reason. reason. reason . reason. 2,937 17 26 155 198 579 885 5-277 6-742 1 2,515 8 23 91 122 318 915 3-618 4-851 2 2,237 6 23 76 105 268 1-028 3-397 ' 4-694 3 ,936 12 19 68 99 620 981 3-512 5-114 4 ,666 6 19 52 77 360 1-140 3-121 4-622 5 ,466 8 10 38 56 546 682 2-592 3-820 6 ,251 4 5 31 40 320 400 2-478 3-198 7 ,090 2 7 14 23 183 642 1-284 2-110 8 989 1 2 13 16 101 -202 1-314 1-618 9 900 5 5 19 29 556 -556 2-111 3-223 10 800 1 2 5 8 125 250 625 1-000 11 735 4 4 8 544 544 1-088 12 663 2 3 5 -302 452 754 13 614 1 3 7 11 163 -489 1-140 1-792 14 537 1 9 10 186 1-676 1-862 15 473 2 4 6 -423 846 1-268 16 402 3 4 7 746 995 1-741 17 360 1 2 1 4 278 ! -556 278 1-112 18 304 1 2 3 329 657 987 19 270 3 3 741 741 20 235 1 1 2 426 426 851 21 195 2 4 6 1-026 2-051 3-077 22 145 1 1 690 690 23 118 1 1 . 2 847 847 1-794 24 95 1 1 1-053 1-053 Totals 22,933 76 161 605 842 331 702 2-638 3-672 ALL FACTORIES. TABLE 56. Ages 18 22. Numbers Left. Percentages. Months N"unifoBr of Service. Employed. Other 111 suffi- health. cient No suffi- cient Tota1 ' ! Other 111 suffi- health. cient 3 l Tota,. reason. reason. reason. reason. - 14,395 135 165 542 842 938 1-146 3-765 fi-849 1 12,388 128 147 388 663 1-033 1-187 3-132 5-352 2 10,690 72 120 365 557 674 1-123 3-414 5-210 3 9,056 75 117 300 492 828 -292 3-313 5-433 4 7,291 58 1 78 234 370 796 -070 3-209 5-075 5 6.869 43 85 172 300 733 -448 2-931 5-112 6 4,775 31 62 151 244 649 -298 3-162 5-110 7 3.937 12 48 116 176 305 -219 2-946 4-470 8 3,204 14 31 75 120 437 -968 2-341 ! 3-746 9 2,449 16 16 42 74 653 -653 1-716 i 3-022 10 1,959 10 15 41 66 510 -766 2-093 3-369 11 1,289 7 12 18 37 543 -931 1-396 2-870 12 722 2 5 10 17 259 -648 1-295 2-202 13 617 3 2 4 9 483 -324 648 | 1-459 14 557 4 1 5 718 180 i -898 15 512 1 2 7 10 195 ! -391 1-367 i 1-953 16 440 3 4 7 682 909 ! 1-591 17 401 i 1 1 249 i -249 18 372 , . 6 10 995 1-493 2-488 19 326 2 3 5 613 920 1-534 20 279 1 1 358 -358 21 218 2 2 -917 -917 22 166 j 7 8 602 4-217 4-819 23 144 1 1 2 694 694 1-389 24 123 3 3 2-439 2-439 Totals 82,179 607 925 2,489 4,021 739 1-126 3-029 4-893 58 ALL FACTORIES. TABLE 57. Ages 23 27. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other suffi- cient No suffi- cient Total. reason. reason. reason. reason. 7,564 87 170 342 599 1-150 2-247 1 4-521 7-919 1 ' 6,317 70 114 225 409 1-108 1-805 1 3-562 6-475 2 6,336 45 109 215 369 843 2-043 i 4-029 6-915 3 4.441 34 68 175 277 766 531 3-941 6-238 4 3,577 30 57 136 223 839 594 3-802 ! 6-234 5 2,798 32 46 86 164 1-144 644 3-074 5-861 6 2,251 19 23 72 114 844 022 3-199 5-064 7 1,836 14 26 57 97 763 416 3-105 6-283 8 1,431 8 17 37 62 559 188 2-586 4-333 9 1,050 10 12 23 45 952 143 2-190 4-286 10 792 5 5 9 19 631 631 1-136 2-399 11 509 2 3 14 19 393 589 2-750 3-733 12 267 2 5 8 749 375 1-873 2-996 13 226 1 442 442 14 192 1 521 . 521 15 168 1 2 595 595 1-190 16 144 , 1 1 694 694 17 127 1 787 ' 787 18 103 1 2 971 971 1-942 19 94 ; . 20 83 1 1 1-205 1-205 21 63 1 1 1-587 1-587 22 42 23 40 24 34 Totals 39,485 360 657 1,398 2,415 912 | 1-664 3-541 6-117 ALL FACTORIES. TABLE 58. Ages 28 and over. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other suffi- cient No suffi- cient Total. reason. reason. reason. reason. 11,840 168 272 593 1,033 1-419 2-297 5-008 8-725 1 9,516 112 165 361 638 1-177 -734 3-794 6-704 2 8,119 93 135 319 547 1-145 -663 3-929 6-737 3 6,821 55 104 260 419 806 -525 3-812 6-143 4 5,461 56 66 177 299 1-025 -209 3-241 6-475 5 4,336 43 59 155 257 992 361 3-575 5-927 6 3,521 34 45 114 193 966 278 3-238 5-481 7 2,847 15 23 74 112 527 808 2-599 3-934 8 2,232 10 13 54 77 448 582 2-419 3-450 9 1,692 9 16 41 66 532 946 2-423 3-901 10 1,312 3 . 8 28 39 229 610 2-134 2-973 11 776 6 3 17 26 773 387 2-191 3-351 12 304 1 6 7 329 1-974 2-303 13 228 2 2 877 877 14 189 1 1 529 529 15 172 1 1 2 4 581 581 1-163 2-326 16 137 1 1 730 730 17 123 18 111 j 1 2 813 813 1-626 19 98 1 1 2 1-020 1-020 2-041 20 71 1 1 1-408 1-408 21 55 22 49 1 1 , 2-041 2-041 23 42 24 40 I -Tctah 60,092 607 914 2,206 3,727 1-010 1-521 3-671 6-202 - 59 ALL FACTORIES. TABLE 59. Combined Ages. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. 1 Other 111 sum- health, cient No suffi- cient Total. 1 reason. reason. reason. reason. 36,736 407 633 1,632 2,672 1-108 1-723 4-443 7-274 1 30,736 318 449 1,065 1,832 1-035 1-461 3-465 5-960 2 26,382 216 387 975 1,578 819 1-467 3-696 5-982 3 22,254 176 308 803 1,287 791 i 1-384 3-608 5-783 4 17,995 150 220 599 969 834 1-223 3-329 5-385 5 14,469 126 200 451 111 871 1-382 3-117 5-370 6 11.798 88 135 368 591 746 1-144 3-119 5-009 7 9,710 43 104 261 408 443 1-071 2-688 4-202 8 7,856 33 63 179 275 420 802 2-279 3-501 9 6,091 40 49 125 214 657 804 2-052 3-513 10 4,863 19 30 83 132 581 917 2-537 4-034 11 3,309 15 22 53 90 453 665 1-602 2-720 l| 1,956 6 8 24 38, 307 409 1-227 1-943 13 1,685 4 8 11 23 237 475 653 1-365 14 1,475 6 11 17 407 746 1-153 15 1,325 3 6 11 20 226 453 830 1-509 16 1,123 6 10 16 534 890 1-425 17 1,011 1 3 2 6 099 297 i -198 593 18 890 2 6 9 17 224 674 1-011 1-910 19 788 1 3 6 10 127 381 761 1-269 20 668 1 1 3 5 150 150 449 749 21 531 4 5 9 753 941 1-695 22 402 1 1 8 10 249 249 1-990 2-488 23 344 2 3 5 581 871 1-453 24 292 1 3 4 342 1-027 1-370 Totals 204,689 1,651 2,657 6,697 11,005 807 1-298 3-266 5-376 60 C 1 CO CN *-< o o Eg 05 ~ ^Pn -OCO CM ii CD ^ Tf 1C o sj CM 1-1 dj c * CM ~ O5 "* P I V.- CD 1 E CO 00 co Q tl ^ CM CM "t o 9 !| CM ^H It CO 00 1C ; 1 1C ii CM O5 ,3 d, CM CD CM TF CM J ^ CO W 11 j CO CD CD 1 Tf r- N W x C^GO O5 CD li CM CO CM 1 il CO- ~ >> CD CM co H CM I> CD" CM" s ' Cfl sf W) -M .2 c "^^ CO t^ CM coco" s 8 O5 1C -f i o I l^ CM W ,. O C CM GO CM" 1 CO 05 05 O5 o i-i CM 00 l> ! CD CO ^r CM 3 KS 9 CD 1C CD r-. 05 1C CO 1C 00^ CD S CD CM 9 CM 1C 1C C a CD *t CM 10 CO I-H CD t^ CM S C^ 1C CD O CD O5 CO CD B.S CD 1C CO GO if> CO 9 ob CO O5 10 ^ O5 CO CO CM r>i co O CO 1C T-l O CD O CO CM CM oo r* s ob 1 00 CM ^- CD rf CO CO CO CD X CD 'CD CO CO 05 CM 6 CM ^ CO (M OC CO GO CD ^"t* ^^ CM CM CM GO CM 1C CM O5 CO CM O} <75 ^ O5 . o 6 CM CM CO CO CD CD g 100 CD CM O CM 00 S 6 IO O i CO CO 8 H T- ^ C r- 5 -*-< O CM o CO. 4 - CM Tf 2 . GO ij CD CM CD CO CO CD CO CO CD CD ^t GO rt O CD CM S CM t^ co" CO *f ~ 1 ro i II i i i i 1 1 CM ~ o o -a 00 10 a CD -* 10 CM CD ^ CD 9 CO CM CO CM CO "^ CO CD GO -*^> 00 10 CM CM ^t CM CD CM 10 S ! 2' lO rp 10 CD CO CM O IO O 00 lO CM ! CO* CO i E 1 i '~ CD "t 8 CD CD CO CM O CO 10 CM H PQ . \ j ""* |i I 21 IO C<1 CD CO cc g r>CM CD CO IO . CM 2 ^ s *- 1 ' CM - ' T^ PH ^ Laundry CM CM IO CD CM CM |Wf CO ^ CM ^L CD CD ,1, ,2 s co IO CO 2S iO CD ^ GO CM ^ ^ CO 00 CM "^ CM CO 1-H 3 1 CS CD 10 CO CO CM GO 4- CD CD 1 GO I> CO CO CD O I> CM ! a rf CD IO c^ r^ "f IO IO CT* ^H CD CD CD r- 10 1 ^ CD t [ ^ r^oo t^ CM CD o 10 I V *>. g? 4- 1-H 00^ CD i w CD CO ^ IO CD lO CO CO GO 10 CM 1-1 CD O r^ CM IO CN l> 1-1 CD^CD 10 CD i i la 11 Tf 00 t^ CM CD IO GO -rf CM ^ 00 CD 10 CM ^H CD CO CD F* IO IO CD CM CO ^' CO CO o b A ^H 00 00 CO "-i CM i CD 00 CD t>. i i IO 13 lO 'f CD CO CM CO CM CD CD CM CD CM CD 00 CD t^ CM" I . . en . . m . X . M 1) : : 8 s 1 * * bJD I 1 i tc d 1 -t-> d % pf -g 8 g rjr^ i n . C/2 ^j Q 1; -j 3 ( 1 .25 3 W h-3 H^l 65 ALL H. FACTORIES. TABLE 67. Ages 13-17. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other suffi- cient No suffi- cient Total. reason. reason. reason. reason. 477 1 1 21 23 210 210 4-403 4-823 351 3 13 16 855 3-704 4-558 2 260 1 2 7 10 385 769 2-692 3-846 3 210 1 1 16 18 476 476 7-619 8-571 4 166 1 4 10 15 602 2-410 6-024 9-036 5 135 8 8 9-926 9-926 6 113 3 4 7 2-655 3-540 6-195 7 75 3 3 4-000 4-000 8 64 1 1 1-852 1-853 9 42 1 2 3 2-381 4-762 7-143 10 9 1 1 11-111 11-111 11 2 12 Totals 1,894 11 9 85 105 581 475 4-488 5-544 ALL H. FACTORIES. TABLE 68. Ages 23-27. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other I No suffi. ; suffi- cient cient Total. El health. Other suffi. cient No suffi- cient Total. reason, reason. reason. reason. 4,418 58 114 207 379 1-313 2-580 4-685 8-579 1 3,649 44 81 124 249 1-206 2-220 3-398 6-824 2 3,140 30 75 151 256 955 2-389 4-809 8-153 3 2,663 27 48 132 207 1-014 1-802 4-957 7-773 4 2,105 22 38 94 154 1-045 1-805 4-466 7-316 5 1,598 20 31 58 109 1-252 1-940 3-630 6-822 6 1,274 14 20 48 82 1-099 1-570 3-768 6-436 7 991 11 19 47 77 1-110 1-917 4-743 7-770 8 744 5 9 18 32 672 1-210 2-419 4-301 9 503 4 6 16 26 795 1-193 3-181 5-169 10 344 3 3 8 14 872 872 2-326 4-070 11 162 2 9 11 1-235 5-556 6-790 12 18 1 1 5-556 6-556 Totals 21,599 238 446 913 1,597 1-102 2-065 4-227 7-394 ALL H. FACTORIES. TABLE 69. Unknown Ages. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other suffi- cient No suffi- cient Total. reason. reason. reason. reason. 155 22 22 14-194 14-194 1 128 1 3 5 9 781 2-344 3-906 7-031 2 113 2 3 11 16 1-770 2-655 9-735 6-310 3 93 4 10 14 4-301 10-753 15-054 4 73 2 7 9 2-740 9-589 12-329 5 61 6 6 9-836 9-836 6 52 4 4 7-692 7-692 7 44 1 5 6 2-273 11-364 13-636 8 36 2 2 5-556 6-556 9 28 1 1 1 3 3-571 3-571 3-571 10-713 10 13 _ 11 4 12 Totals 800 6 12 ' 73 91 750 1-500 9-125 11-375 (B2604- r B2367) 66 ALL H. FACTORIES. TABLE 70. Ages 18-22. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other | No suffi- ; sum- cient cient Total. reason. ! reason. i reason, reason. 7,831 70 Ill 251 432. 894 1-417 3-205 6-517 1 6,843 60 94 230 384 877 374 3-361 6-612 2 6,065 32 84 243 359 528 385 4-007 5-919 3 5,249 39 81 206 326 743 543 3-925 6-211 4 4,210 37 55 161 253 879 306 ! 3-824 6-010 5 3,288 19 53 105 177 578 612 j 3-193 6-383 6 2,720 17 39 103 159 603 -383 3-652 6-638 7 2,174 6 37 85 128 276 702 3-910 5-888 8 1,668 6 23 45 74 360 379 i 2-698 4-436 9 1,165 9 12 26 47 773 1-030 2-232 4-034 10 829 6 7 22 34 603 844 2-654 4-101 11 378 2 4 9 15 529 1-058 2-381 3-968 12 36 1 1 2 2-778 2-778 6-556 Totals 42,456 303 600 1,487 ; 2,390 697 1-381 3-422 5-500 ALL H. FACTORIES. TABLE 71. Age 28 and over. Numbers Left. Percentages. Months of Service. Number Employed. Ill health. Other suffi- cient No suffi- cient Total. Ill health. Other suffi- cient No suffi- cient Total. reason, t reason. reason. reason. 1 7,115 115 218 390 723 1-616 3-064 6-481 10-161 1 6,656 76 125 242 443 1-344 2-210 4-279 7-832 2 4,785 60 95 223 378 1-254 -985 4-660 7-900 3 4,006 43 73 194 310 1-073 -822 4-843 7-738 4 3,152 35 44 120 199 1-110 i -396 3-807 6-313 5 2,393 26 29 95 150 1-087 212 3-970 6-268 6 1,892 23 28 83 134 1-216 ! -480 4-387 7-082 7 1,394 8 17 45 70 574 220 3-228 6-022 8 1,123 6 6 31 43 534 534 2-760 3-808 9 770 4 10 24 38 519 1-299 3-117 4-93* 10 561 1 6 22 29 178 1-070 3-922 6-169 11 270 3 1 9 13 1-111 370 3-333 4-815 12 23 4 4 17-391 17-391 Totals 33,140 400 655 1,482 2,537 1-207 1-976 4-458 | 7-655 ALL H. FACTORIES. TABLE 72. All Ages. Numbers Left. Percentages. Months of Service. Number Employed Ill health. Other suffi- cient No suffi- cient Total. I Other 111 | sum- health. [ cient No suffi- cient Total. reason. reason. reason. reason. 19,996 244 444 891 1,579 1-220 2-220 4-456 7-897 1 16,627 184 303 614 1,101 1-107 1-822 3-693 6-623 2 14,363 125 259 635 1,019 870 1-803 4-421 7-095 3 12,221 110 207 558 875 900 1-694 4-566 7-160 4 9,706 97 141 392 630 999 1-453 4-039 6-491 5 7,475 65 113 272 450 870 1-512 3-639 6-020 6 6,051 67 87 242 386 942 1-438 3-999 6-379< 7 4,678 25 74 185 284 534 1-582 3-955 6-071 8 3,625 17 41 97 155 469 1-131 2-676 4-276 9 2,508 18 30 69 117 718 i 1-196 2-751 4-665 10 1,756 10 16 52 78 569 911 2-961 4-442 11 816 5 7 27 39 613 858 3-308 4-779 12 77 1 6 7 1-299 7-792 9-091 Totals 99,899 958 1,722 4,040 6,720 959 1-724 4-044 6-727 GO I CO pa H 4 ij ^ CO cc cp 05 il o r>. CN 00 l-H a 1 05 IO II 05 (0 CO 8 CO j CO 05 It - Laundry. r> - 00 r- i CQ I Servants. ^ I r I PI O5 t^ 10 *3 o CN CN CN 05 IO CN 05 CN 5 CO CO 00 1 1 05 CO CO l-H i op 10 1-1 CN t"> 10 r-H s 00 | 00 co CN I co 05 uo CO CN 6 O5 [> CO CN CN 55 CN LO CO CO CO Existing Left Percentages. . . . CM I 05 10 o IO o ^ CO i CN CO 05 CN CN ~ 10 ? 00 ~ i 10 . o O5 CO CO CO CN CN CN Cl 05 ! o s CN 8 CO 1-H CO 05 IO 10 cs CO CO CO cs V 4 CO 00 CO S Existing Left Percentages 4 CN CD O CD CN CO CN II 03 ii! O II O O3 co sg CO CO CD 1 CN O3 CO T-l CN it CO T* CN T-I i i CD CN CN CD CD CN i CO 03 IO CD CN r-< CD CD Servants. IO IO CD CN IN CO PI T-I CO 1-1 CO CN IO T < Hf CN II CN IO CD O CD CO CN CO Factory. O3 O3 i i CO CO CN tuO S to _|j 'x 73 W M Percentages . . CN CD CD CO O3 CD CO IO - CN IO CO rH CN * CN r-H ! CN CO 9 CO i 1 CN CD co co ? T < ? CD CO IO O T I 2 1 < ? O CO CO 10 s 1 . CO CD T-I CN CO T-l CN CD o CO 1 ; co ^^ Tt* CO * 1 l> O IO CD CN" CD IO 10 O3 CO CN O3 CD CO O3 CN i-T 10 CO T-l CD O CN CO 10 co^ co 03 O CO l> CN 4- t^ CO IO O 00 CD^ co" i-T CN _( ^_) W i-4 Percentages 1 t W i-4 Percentages . . 69- TABLE 74. Three Months' Survival Rates. Factory. Ages 18-22. Ages 23-27. Ages 28- . L.4 84-16 -83 81-08 1-21 83-95 1-06 L.5 93-60 -74 94-12 -71 89-50 MO H.4 88-32 -92 84-62 1-81 81-65 1-89 H.1 82-30 1-67 72-88 1-92 71-24 1-23 H.8 66-27 2-04 66-34 2-86 66-93 db 2-74 H.9 90-73 db 1*40 86-81 2-01 83-84 db 1'56 L.3 85-75 '73 81-50 db '98 86-07 -67 H.2 81-03 2-03 74-71 2-00 66-56 1-57 H.3 77-16 1-53 64-76 1-66 72-32 db M4 H.5 87-89 2-29 88-67 1-74 H.10 85-71 -52 82-70 -86 81-01 '78 L.1 81-32 db -71 81-07 1-08 83-02 -97 H.6 87-91 -61 78-80 -95 78-50 -71 L.2 77-45 db 1-05 71-81 1-16 80-92 -80 H.7 82-65 -59 77-55 db -70 75-65 '92 L.8 74-70 2-06 L.6 70-29 2-77 64-70 3-68 71-87 2-26 L.7 95-34 -78 90-92 db 1-46 87-78 1-38 TABLE 75. Months' Survival Rates. Age Group. L. H Difference. 18-22 23-27 28- 67-55 '52 65-15 -i -76 66-59 -t -59 59.64 -49 51-46 db '68 52-45 i -54 7-91 -71 13-69 1-02 14-14 _j- .go TABLE 76. Nine Months' Illness Survival Rates. Factories H. Factories L. Difference. Ages 18-22 Ages 23-27 Over 28 94-40 db -24 90-72 -45 90-60 -36 93-04 -30 93-21 -45 92-92 -35 4- 1-36 -38 - 2-49 -63 - 2-32 -50 (B2604 r B2367) o2 70 TABLE 77. Average Weekly Hours Three Months' Survival Factory. (day-shift), including Overtime. Rates. (Ages 28- .) L.1 43-5 83-02 L.2 43-5 80-92 L.3 52-5 86-07 L.8 55-3 L.7 60-5 87-78 L.4 61-8 83-95 L.5 61-8 89-50 L.6 63-5 71-87 H.6 ... 44-0 78-50 H.9 ... 45-0 83-84 H.10 49-0 81-01 H.1 53-0 71-24 H2 53-0 66-56 H.5 53-5 88-67 H.3 54-0 72-32 H.8 56-3 66-93 H.4 62-0 81-65 H.7 66-0 75-65 FACTORIES H.4, H.10, L.3. TABLE 78. Civil state. Married. Single. Widows. Total. Percentage Married. Under 22 243 3,856 4 4,103 5-92 23-27 ... 536 1,358 15 1,909 28-08 Over 28... 1,497 1,358 104 2,959 50-59 71 I CO ' J ^; Bi- ll ^ ^i oi>co OJ> GO CO U5 C* - .-- ** 73 APPENDIX II. NOTE ON ERRORS OF SAMPLING. When we desire to know whether a set of observations either (a) effectively obeys some assumed law or, (b), merely gives results in sensible agreement with those of some other set of observations, we must ascertain whether any divergence between theory and practice in the former case, or between the two practical experiences in the latter case, is or is not greater than might be expected often to occur by virtue of those chance fluctuations which occur when samples small, relatively to the populations from which they come, are our only means of information. Problem (a) has been solved by Professor Karl Pearson in a series of admirable researches, dating from 1900. If we have a distribution into classes : Group, Observed Frequency. Theoretically Expected Frequency. A 5 :: A n % . m n Pearson showed that the probability of chance fluctuations leading to any observed divergence between theory and observation was a function of ' where :-' = m What function depends upon the form of the group standard deviations and error correlations ? Pearson has supplied the necessary modifications for most schemata of statistical importance. Problem (a) is accordingly solved. Problem (b), however, is of more frequent occurrence in ordinary statistical practice. No theoretical law can be denned, but the significance of a divergence between two samples has to be gauged. For this purpose the above-mentioned method is widely utilised. For instance, in two recent papers (Biomeinka, 1916, XI, pp. 146-158 and 159-184) Pearson and Tocher employ this criterion to determine whether two samples of population, for each of which the numbers exposed at ages and the deaths at ages are, respectively, &c., d l &c., a' &c. f d' Q , d &c., differ significantly with regard to death-rates. It is assumed that the " theoretical " death-rate of persons in the s th age class is : d s -f d' a a, + a', i.e., that it is the value obtained by combining the two sets of data. Making this assumption it can be shown that : *, - ' . (N + N') a N and N' being the numbers of observations. 74 It is also pointed out that the fraction usually differ much from unity, so that : v a s -f a 'J* N N will not (d. + < is approximately equal to yf. The probability that the two sets of observations have been drawn from a common source is then deduced by entering x 1 with n = the number of age groups plus one, in the published tables, for testing goodness of fit and finding the appropriate value of P (see Tables for Statisticians and Biometricians, pp. 26-28). The weak point is the assimilation of the actual condition (6) to the a priori condition (a), a difficulty noted in the memoir from which the above results are cited. Pearson and Tocher (op. cit., pp. 173-4) show that, if quantities like d s and d' s are approximately normal in distribution then p t = - ~~ is a close approximation to that value of p t which will make x* as small and consequently P as large as possible. As Yule and the writer have pointed out (Proc. Roy. Soc. Med., 1915, VIII, Sect. Epidem., pp. 117 et seq.}, a practical consequence is that, although samples compared by this test and, on its finding, deemed to be differentiated may certainly be so regarded, the converse is not true. The theoretical considerations involved are of considerable interest, and will be developed at length in a mathematical paper by Yule and the writer, part of which is completed. In the present connection other difficulties are of more importance. Thus, although the use made of Pearson and Tocher's method in Table 23 of this report is quite defensible, there are certain objections. The fundamental data are : Ex- posure Time Interval Number at Bisk. "Deaths." Survival Bate. "Death" Bate. '. . PQ **o ~ Po _ Qo !-Q, = P, W < Pi n, _ p o _ ko _ Pl-Q 1 Qj = Pj w o - Po - k * S=l 5=1 M - Sp -Sft - P, *. n s P, s = o s = o 1 - Q, = P, S = l S=l n -Sp - Sk 1 . s=o s = o Variations of survival (or " death ") rates in individual periods are influenced by the system of writing off absentees. Were all absentees struck off the list within half a time unit, then "deaths " are properly credited. If a latent period elapses this is not so. For instance, let the state of affairs in two factories be as follows : Each engaged 1 ,000, of whom 100 left within a few days ; 50 others left after a month's service 75 and before completing the second month. Suppose one factory writes off its losses at once, we have : n. p. Q. P. t . . .. 1,000 100 -900 -100 /! .. ..900 50 -944 -056 Suppose the other factory does not write off any losses until the absentee's name has been on the books at least a month, its record will run : n. p. Q. P. t . . .. 1,000 1-000 -000 fj .. .. 1,000 150 -850 -150 The number of survivors at the end of the second month is in each the same ; the rates of loss for the separate months different. But what is really interesting is precisely the number of survivors at the end of a period such as three months, because this gives us the proportion of entrants actually reaching a stage of normal productivity. Differences in the method of reaching this proportion of survivors, which would be emphasised by the x a test, may very often have no real significance, be mere peculiarities of book-keeping. Of course, in the particular examples chosen in the text to exemplify the method, this source of disturbance was known to have been absent, while under no circumstances is it likely to involve such extreme consequences as indicated in the imaginary case. Indeed the rule, where any is found, is not that losses are only struck out after a definite period from entry, but that the absentee is only removed from the list when her absence exceeds a certain number of weeks. The effect is, in theory, merely to cause a uniform transfer of " deaths " from one group to the next. In practice, however, the transfer is much concentrated upon the early, especially upon the first, exposure group. Unexplained absences prior to leaving become less and less common with length of service. As our data chiefly relate to early months of service, I have thought it well to make use of the survival method for comparative purposes, and now pass to consider its probable error. We have : 's = Qo Qi Q.-I s s being the proportion surviving at the end of the s th exposure. Taking logarithms, and differentiating, we have : >* = SQo 4. *Qi , . Qo Qi Q.-I Hence : _ = -_3* -u Q-' 4. 2 s GO GO / rr\