Sexes, Ages, and Condition as to Marriage

Next Selection Previous Selection



Sex Proportions. —Of the 36,070,492 persons enumerated in England and Wales in 1911, 17,445,608 were males and 18,624,884 were females. These numbers give an excess of 1,179,276 females over males, which would, however, be somewhat reduced if we could include in the reckoning the English and Welsh members of the Army, Navy and Merchant Service and mercantile community temporarily absent abroad and also the numbers of fishermen absent at sea on the night of the census.

The number of females to 1,000 males enumerated at each successive census was as follows:—

Year. Number of Females
to 1,000 Males.
Year. Number of Females
to 1,000 Males.
1801 1,057 1861 1,053
1811 1,054 1871 1,054
1821 1,036 1881 1,055
1831 1,040 1891 1,064
1841 1,046 1901 1,068
1851 1,042 1911 1,068

The decrease in female preponderance from 1,057 in 1801, and 1,054 in 1811 to 1,036. In 1821, was no doubt due to diminution after the war of the large forces which had been serving outside the country at the two earlier dates. Since 1821 this preponderance has on the whole steadily increased.

The proportion in 1911 was the same as in 1901, viz., 1,068 to 1,000, but when due allowance is made for the exceptional number of males absent on military service in South Africa in the last-named year there is no doubt that the real proportion of females to males in the population was somewhat lower in 1901 than in 1911. The alteration in proportion may best be gauged by comparing 1911 with 1891.

Although more boys are born than girls (the average proportion in England and Wales during the last 75 years was 1,040 boys to 1,000 girls), this initial predominance of the male sex is soon lost, and the relative proportions of the sexes transposed, as the mortality of males is greater than that of females in infancy and in adult life, This natural excess of females is further accentuated by the temporary absence of men abroad serving as soldiers or seamen, &c., and by the great excess of the emigration of males over that of females.

Local Variations in Sex Proportions. —The sex proportions of the population vary widely in different parts of the country, these local variations being determined in the main by social and industrial conditions. The following is a list of Administrative Counties (together with their associated County Boroughs) in which proportions of females to 1,000 males were lowest and highest respectively:—

Lowest. Highest.
Monmouthshire 908 Sussex, East 1,256
Glamorganshire 925 Cardiganshire 1,224
Brecknockshire 952 Surrey 1,166
Rutlandshire 973 Cornwall 1,164
Isle of Ely 980 Somersetshire 1,153
Durham 984 Middlesex 1,144
Radnorshire 992 Gloucestershire 1,142
Derbyshire 993 Isle of Wight 1,130
Lincolnshire— London 1,127
   Lindsey 1,001 Sussex, West 1,117
   Kesteven 1,004 Westmorland 1,112
Wiltshire 1,004 Carnarvonshire 1,106
Carmarthenshire 1,004 Devonshire 1,102
Denbighshire 1,005 Worcestershire 1,101
Yorkshire, North Riding 1,008 Cheshire 1,100
Northumberland 1,010    

It will be observed that most of the counties in which the female proportion was at its lowest were mining counties, that industry having the tendency to draw large numbers of single men into its ranks; while conspicuous among those in which the female sex was most predominant were Sussex, Surrey, Middlesex, London and the Isle of Wight, which may be described as residential counties in which large numbers of domestic servants are employed, and to which the unoccupied classes (preponderantly female) resort for residence. The urban populations, other than those of garrison towns, containing the largest male element, are those of mining and other industrial towns whose characteristic industry does not make much demand upon female labour. Where this latter condition prevails, as in the textile towns of the north, the female sex may be in considerable excess. The extreme positions at this end of the scale, however, are all occupied by residential towns, a number of them suburbs of London or other of the largest cities. The highest female proportions of all are met with in certain health resorts. This relationship is examined in greater detail, in conjunction with the age distribution of these populations on pages 79 and 80. In the rural districts as a whole there were only 1,001 females to 1,000 males, while in the urban districts (including London and the county boroughs) the proportion was 1,087 to 1,000; in the smaller urban districts it was 1,075 to 1,000 in the county boroughs 1,086, and in London 1,127 to 1,000.

Diagram IX shows the proportion of males and females in 1,000 persons of both sexes in each Administrative County (with its associated County Boroughs); and Diagram X shows this proportion in the County Boroughs, and in other towns of over 50,000 population.

DIAGRAM IX.—Sex Proportions in Administrative County (together with associated County Boroughs), 1911.

(The dotted line indicates equal proportions of the two sexes.)

sex proportions in 1911, by county

DIAGRAM X.—Sex Proportions in County Boroughs and other Urban Districts of which the population exceeded 50,000 persons in 1911.

(The dotted line indicates equal proportions of the two sexes.)

sex proportions in 1911, by large town

Sex Proportions in other Countries. —The numbers of males and females enumerated at the recent censuses in the populations of the three divisions of the United Kingdom and in a number of other countries are shown in the following statement, from which it will be seen how widely the sex proportions vary:—

Country Number of females
to 1,000 males.
Country Number of females
to 1,000 males.
Portugal 1,107 Belgium 1,017
Norway 1,099 Ireland 1,004
England and Wales 1,068 Japan 979
Scotland 1,063 India 953
Denmark 1,061 Union of South Africa 946
Sweden 1,046 United States 943
Italy 1,037 Australian Commonwealth 926
Austria 1,036    
France 1,034 New Zealand 895
Switzerland 1,034 Ceylon 887
German Empire 1,026 Dominion of Canada 886
The Netherlands 1,021    


Tabulation in single years. —until the present occasion the ages of the population have not been tabulated in single years throughout the whole period of life. In 1841 the figures ware necessarily grouped in periods of five years; this procedure was followed exactly in 1851, and, with the slight exception that ages of children at each year under five were shown from 1861 onwards, was continued until 1891. Additional detail was given in the report on the census of 1901, where the numbers at each age from 13 to 21 years were tabulated; and, with a view to testing the approximate accuracy of the returns, the results of an experimental tabulation by single years of the ages of about half a million of the population were shown. These results, valuable in themselves as indicating roughly what would probably be revealed by a similar analysis of the figures for the whole country, emphasised the need for greater precision in this branch of the census inquiry. The adoption of electrical tabulating machinery at this census has rendered practicable the elucidation of statistical facts in far greater detail than could be hoped for under the system hitherto followed, and a table frequently asked for, viz., a detailed record of the ages of the population as stated in the census schedules, has now been published for the first time in this country, and will, if is hoped, prove of great interest to the actuarial profession.

The detailed results for England and Wales as a whole and for certain large aggregates, e.g., the totals of the Urban and Rural Districts, are given separately in Tables 29 34 of the Summary Volume. Similarly detailed figures have been recorded for each administrative area, but considerations of space prevent their publication.

Uses of statistics of ages. —Statistics of the age-distribution of the population are second in importance only to the fundamental record of the actual number of persons living; for the age-constitution is the key which alone can furnish an explanation of many of the phenomena associated with different populations. Thus, for example, appreciation of the meaning of vital statistics depends on a just estimate of the relation of births and deaths (especially the latter) to population at the several ages. A further illustration of the desirability of knowing the ages of a population is seen in the number of legislative enactments which are applicable only to persons within certain prescribed age-limits; and administrative work in connexion with such Acts is necessarily hampered when, in the absence of definite knowledge on the subject, the population to be dealt with has to be estimated, generally from more or less unsatisfactory data. For educational purposes it is of paramount importance to know the number of children of school-age in various districts; and for many economic reasons the numbers living at successive ages are required. Other examples will suggest themselves, but enough has been said to show that a census which did not take account of the ages of the population would be lacking in one of its most essential features.

Comparison by age-groups. —Although for certain purposes it is desirable to know the precise ages as returned on the schedules, it will be found necessary in comparing the present with preceding censuses to group the ages into the quinquennial periods shown in Table 8, Vol. VII, which relates to the administrative counties and to all urban and rural districts. To facilitate study of their significance the numbers relating to administrative counties and the larger towns are expressed in Table 38 of the Summary Volume as proportions per 100,000 of the total population in each case, while Table 39 in the same volume states the ratio of each of these proportions to the corresponding figure for England and Wales. It may be remarked here that although the tabulation was not performed directly from the original returns until the present census, but from copies made by the local enumerators, we have no reason to suppose that comparison is seriously prejudiced by the change of method.

Sex and age-constitution in 1911. —Dealing first with the 1911 figures only, it will be seen that more than one-tenth of the total population is under five years of age, more than one-fifth under ten years and nearly two-fifths under twenty; ages 20-55 include nearly one-half the total, while the survivors at ages over 55 number but little more than one-tenth of the whole. Comparing the sexes age by age, Table 42 of the Summary Volume shows that the excess of females at all ages is maintained, though very unequally, at each period in the table except the first. Under five years of age the preponderance of male births (which average about 1,040 to 1,000 female births) is reflected in the population figures, and is, of course, most marked in the first year of life. But the greater mortality of males under five years of age (1,171 male to 1,000 female deaths in equal numbers living during the period 1838-1910) reduces that sex to a minority in the following quinquennium. At succeeding age-groups the sex proportions are subject to other influences; although the mortality of females at adult ages is generally below that of males, this becomes for a time a subsidiary factor, resuming its importance gradually as the more advanced ages are reached, when the excess of females becomes very great. In early adult life the excess of females is greater than would result from the excess of mortality amongst males and is largely due to the fact that there is a net loss to the population by migration, chiefly at the younger adult ages and affecting males more than females.

The sudden fall at age 20-25 as compared with 15-20 in the proportion of males shown in the last columns of Table XIV on the next page, and in Diagram XVI, and which may be traced in Table 42 of the Summary Volume at each census back to 1851, suggests as its most obvious explanation the absence on foreign service during early manhood of soldiers and sailors, whose return to England a few years later causes a temporary arrest in the otherwise steadily progressive increase in the proportion of females. A rough estimate of the numbers withdrawn from the country in this way shows that they approximately suffice to fill up the depression at 20-25 in the curve showing age-distribution of males in England and Wales in Diagram XV. It is, moreover, shown by Diagram XXVII to be probable that immigration of young women as domestic servants contributes to the sudden increase in the proportion of females at 20-25, while the mis-statement of age characteristic of this section of the population (see page 87) would tend to produce the same effect.

But examination of the number of males, not in quinquennial groups, but at single years of life as shown in Mr. King's graduated table (page 93), suggests a doubt as to the adequacy of the explanation of the facts by withdrawal of males on foreign service. It will be seen from the table referred to that if the differences between the graduated number at each age and that at the age one year greater be considered, these differences gradually increase from 13-14 to a maximum at 18-19, and then gradually return to about their former magnitude, which is reached at 23-24. Mortality at this time of life is low, and changes but little from year to year; so it cannot account for this rapid decrease in the numbers living, which seems therefore to imply a balance of emigration for males, setting in about 14 years of age, gradually reaching a maximum at about 19, and then gradually declining to disappearance or nearly so at about 23. The ages in question appear to be somewhat young for the foreign service explanation of the phenomenon, which, apart from this consideration, seems satisfactory.

The numbers as enumerated, on the other hand, fit in well with the foreign service hypothesis. It may be seen from Diagram XXVI that the errors of "round" and "even" numbers, discussed on pages 85 and 87, which cause great irregularity in the curve showing ages of males as stated in the returns, do not set in till age 30 is approached. The curve runs fairly smoothly between ages 2 and 26, suggesting approximate accuracy of statement, but there is a sudden increase in its steepness between ages 18-19 and 22-23, which if it represented the facts would harmonise well with departure for service abroad. Any such sudden fluctuation is, of course, eliminated in a graduated curve, and it seems possible that in this instance irregularity represents the facts better than smoothness.


Comparison with previous censuses. —When the present age-distribution of the population is compared with those at previous censuses, Tables 41 and 42 of the summary volume, and especially Diagram XI, show that striking changes have occurred. The proportion of children in the population has decreased to an extraordinary extent since the birth-rate commenced to fall in the year 1877. For each million at all ages in 1881 there were 135,551 children under five years of age, and this proportion had been fairly constant at previous censuses. In 1891, however, the fall in the birth-rate (which did not affect the 1881 population, because the mean recorded rate was as high for the five years 1876-80 as for 1866-70) had lowered this proportion to 122,523, and since then to 114,262 in 1901 and to 106,857 in 1911. The proportions of older children have also been lessened, but in their case the change has naturally set in later, as the generation born since 1876 has grown up. It is at the youngest ages that the diminution in the proportion of children is greatest, since, owing to the progressive nature of the fall in the birth-rate, the older children represent the survivors of births whose frequency differed less from that of former years than is now the case. The diminished proportion of children has entailed as a result an increased proportion of persons at all other ages.

DIAGRAM XI.—Population of England and Wales.—Persons of Both Sexes.

changes in proportions of each sex at different ages

These changes may be followed in Diagram XI, which compares the age-distribution of the population in 1911 (dotted line) with those of the four preceding censuses. The general similarity between the populations of 1871 and 1881 is evident. In 1891 the fall in the birth-rate had reduced the numbers living at the first two age-periods, causing a hump, so to speak, in the curve at 10-15. In 1901 this reduction affects the first four age-periods, and the "hump" has moved on to 20-25, while in 1911 it centres at about 30-35, having at the same time become more spread out and less marked at any one age-period. It is evident from this diagram that if the fall in the birth-rate continues, a further decrease may be expected in childhood and early adult life, with a compensating further increase in the proportion living at higher ages. The future age-constitution of our population, in fact, may be expected to approximate more or less to that of France in 1906 (the most recent date for which the figures can be obtained),1 the graphic representation of which will be found in Diagrams XIII and XIV.

DIAGRAM XII.—Population of England and Wales.—Persons of Both Sexes.

Ratio of the proportional Numbers shown in Diagram XI for previous Censuses to the Numbers at corresponding Ages in 1911.

age structure in earlier censuses compared with 1911

The same facts are represented from a different point of view in Table XV and in Diagram XII, which corresponds to it. The curve for 1871 has been omitted in this diagram for the sake of clearness, as it is practically the same in type as that for 1881. The figures in the table represent the proportional extent to which the numbers for other censuses in Table 42 of the Summary Volume, exceed or fall short of those for 1911. In 1881 the proportion to total population of children under five years of age was as much as 27 per cent, higher than in 1911. This excess is seen from the diagram to diminish with increasing age till it disappears at about (probably a little under) age 25. The same general statement holds good of the curve showing proportions in 1891, but the effect of the fall in the birth-rate which had occurred meanwhile is clearly shown by the smaller proportion living in the first two age-groups. Both curves continue to fall steadily till middle life is reached—ages 30-40—after which they become somewhat irregular with advancing years, but without any very definite general tendency to increase or decrease, fluctuating around the 90 per cent, level. In 1901, owing to the diminution in proportional number living at the earlier ages, with consequent increase in later life, the type of age-distribution approximates much more closely to that of 1911 than it did at any previous census. This approximation is also clearly shown by Diagram XL The effect of the sudden diminution in the birth-rate, dating from the year 1877, is again clearly marked in the 1901 curve (Diagram XII), the maximum excess over 1911 being reached at age 20-25 as against 10-15 in 1891. Comparing each census with its predecessor, Table XV shows that the proportions living at the first two quinquennial age-groups fell in 1891, at the first four ten years later, and at the first six in 1911, these groups in each case consisting of persons born after the decline in the birth-rate had set in.


It may be pointed out that though the effect of the fall in the birth-rate has hitherto been in a sense temporarily advantageous in that it has increased the proportions living at the working ages, a tendency to the reversal of this effect has already set in and may be expected to develop as time goes on. Thus the first effect of the changes which have occurred on the age-group 20-25 was to increase considerably the proportion living, as may be seen both in Diagram XII and from the " hump" already referred to in the 1901 curve in Diagram XI, but this temporary advantage has now been lost, the proportion now being' a trifle lower than in 1881 (Diagram XII). Similarly the first effect on the next age-group, 25-30, was to raise its proportion in 1891 and especially in 1901, but decline from the latter position has already set in and may be expected to continue (Diagram XII). These fluctuations are not without importance in their bearing upon the probable future proportion in our population of workers at the most economically efficient ages and especially upon the supply of recruits to the military services in future years.

The fluctuations shown in Diagram XII in middle and later life by the curves relating to the earlier populations are apparently to a large extent factitious. It will be noticed that the peaks all occur at the age-groups liable to be swollen by the error of statement of age in round numbers, discussed on page 85, especially the groups including age 50, age 60, and age 80. Reference to Diagram XI shows that the 1911 curve runs much more smoothly at this period of life, the absence of heaping up at age 60-65, marked in all the other curves, being especially noteworthy. The result of this of course, is to bring about a smaller deficiency in Diagram XII in the curves of the earlier populations at age 60—than at 55—or 65—with resulting irregularity of this portion of the curves. The improvement which has taken place in this respect since last census in the statement of age is discussed in Mr. King's report on the graduation of ages (see Volume -VII, pages xlvii and xlviii).

In the foregoing remarks the fall in the birth-rate alone has been considered as causative of the changes discussed. In reality, of course, the case is not so simple; the fall in the death-rate, applying unequally to different ages, and fluctuations in migration have also had their effects. These effects, however, are neither so dramatic nor so followed as that of the fall in birth-rate, and the degree to which the latter has the dominant influence may be gauged by the extent to which it is by itself capable of explaining what has occurred.

DIAGRAM XIII.—Standard Populations in Various Countries.

age structures of various countries compared with England and Wales

Comparison with other Countries. —As was pointed out in the Registrar-General's Annual Report for 1909, the age-distribution of the population in this country in 1901 exceptionally favourable to a low rate of mortality, owing largely to the high proportion in it of young adults, with mortalities below the average for persons of all ages, together with a very moderate proportion of children under five years of age and of old people. Tables XVI and XVII and Diagrams XIII and XIV show that, generally speaking, this statement still holds good.

DIAGRAM XIV.—Age Distribution of Populations in Various Countries:—Percentage deviation of the proportions in the several age-groups from the corresponding proportions in England and Wales.

age structures of various countries compared with England and Wales

These diagrams are constructed on the same plan as Diagrams XI and XII respectively, comparison in all cases being made between the other populations dealt with and that, of England and Wales in 1911. The most recent census is shown for each country except France, in which case the age-statistics for 1906 are given as those for 1911 are not yet available;2 the figures for Denmark and Norway are included in the tables, which show that the age-distribution of the population in these countries is similar to that in Sweden, and they have therefore, for convenience of space, been omitted from the diagrams.



It will be seen that the relative numbers living at ages under 20 were lower in England and Wales than in any European country except France, where the numbers of children and young persons were conspicuously lower than in any other country; in Ireland the proportion of children under 10 was lower than in England and Wales, while in Scotland and all the other parts of the British Empire the reverse was the case, the proportions being especially high in the Indian Empire and in the Union of South Africa.

At ages between 20 and 50 there was a higher proportion in England and Wales than was found generally in other countries; in the earlier part of this period, however, viz., at 20-25 and 25-30, there was an excess, as compared with England and Wales, in Australia, New Zealand, Canada, South Africa, and the United States, presumably as an effect of immigration. The general resemblance of the curves for these immigrant countries, especially that of New Zealand, with those figured in Diagram XXI for the "dormitory" or immigrant districts around London, may be noted. From 35-40 the proportion was greater in this country than in any other, without exception, while at ages 40-45 and 45-50, the only exception was France, where the excess over the English proportions, which is generally observable at the advanced ages, is shown earlier than in any other country.

At ages above 50 the proportions were lower in England and Wales than in the European countries generally, the German Empire forming the only marked exception, while in the Overseas Dominions and in the United States, people of these ages were proportionally less numerous than in this country, although at the extreme period (85 and upwards) Canada, South Africa and the United States show some excess.

The general characteristics of the figures indicate very clearly the effects of the long continued decline in the birth-rate of this country, and show by the example of France, the type of age-distribution which a further continuance of the decline is likely to produce. The present age-distribution of the English population is still favourable to low death-rates, but is becoming less so than it was in 1901. The movement along the curve of the point of maximum heaping up of population, referred to on page 61, has shifted this from age 20-25 to a period ten years later, when mortality is appreciably higher. The result of the changes that have occurred may be measured by the fact that the recorded death-rate in 1911, 14.6 per 1,000, has to be diminished to 14.3 to permit of fair comparison with 1901, the difference representing the effect of the change in age-distribution meanwhile.

Among all the populations illustrated in the diagrams that of Scotland approximates most nearly to the age-distribution in this country. Ireland resembles France in having a small proportion of young children and many old persons, herein showing the effect of long-continued drain by emigration (see page 67), but its distribution is peculiar and anomalous, as indeed it was also at the previous census. One of the minor points brought out by Diagram XIV is the prevalence of peaks at the age-groups including the "round numbers" 30, 40, 50, 60 and 70, thus indicating that the form of mis-statement of age referred to on page 85, as the "error of round numbers," is more common in the countries showing this feature than in this one; this heaping-up at the round numbers is especially pronounced in India where, owing to the general ignorance of precise age, the ages of as many as 31 per cent, of the persons aged 10 and upwards were shown as multiples of 10.

Urban and Rural Districts. —The age and sex-constitution of the population characteristic of the towns and of the country districts differs considerably, and if the larger are considered apart from the smaller towns these are found to have well-marked differences among themselves. In Tables XVIII and XIX and in Diagrams XV and XVI the distribution of population is shown in the rural districts and in the urban districts as a whole, and the latter are also dealt with in three divisions, London, county boroughs and other urban districts, representing towns of varying size.


The contrasts between the populations of these areas are seen to be very marked, but the tables and diagrams rather understate them. For although the rural districts have necessarily been taken as representing the rural portions of the country, and the urban districts the towns, it must be remembered that many rural districts include considerable populations living under urban conditions, especially in the mining districts, while the wide boundaries of a number of towns embrace considerable areas of distinctly rural type. The rural districts, in fact, are not wholly rural, and to a less extent the urban districts are not wholly urban, and the admixture must result in an understatement of the differences between the two. In distinguishing size of towns also, although the county boroughs are taken as representative of the larger provincial towns, a number of them are smaller than a few of the other urban districts. These considerations, however, merely tend to strengthen the distinctions which can be drawn between the different types of areas contrasted.

DIAGRAM XV.—Populations of England and Wales and of constituent Groups of Administrative Areas in 1911.

Age Distribution of Males and Females in each Population compared with that of Persons in England and Wales (Males and Females in 100,000 Population, and Persons in 50,000).

age structures of types of administrative area

Age. —Diagram XV shows how widely the age-constitution of the urban population as a whole differs from that of the rural. In each section of this diagram two curves, representing males and females in 100,000 persons with the age and sex-distribution of one of the contrasted populations, are compared with a third representing 50,000 persons of undistinguished sex, but in age-distribution typical of the country at large. The rural population naturally starts with a slight disadvantage in numbers living at age 0-5 in consequence of the smaller proportion of potential parents it contains. In the next age-period, however, the deficiency in rural births is more than compensated for by the smaller mortality of the children born, and the proportion of the population aged 5-10 becomes appreciably larger than in the urban districts.


This continues until the attraction of the town is felt by the country dwellers at age 15-25, when the relative position is reversed. The diagram shows how much the movement in question is concentrated upon these ages. The proportions living in the rural districts remain below the mean for the country at large till age 45 is reached, after which, in consequence of the much higher mortality in the towns, aided perhaps by the removal of elderly people to the country after active life is over, they are in excess, and as old age advances, in very great excess, of the corresponding proportions in the towns. Moreover, the depression in the rural curves caused by the continued exodus of young adults tends reflexly to raise the proportions at other ages, since the sum of 100,000 living must be made up in all cases.

DIAGRAM XVI.—Populations of England and Wales and of constituent Groups of Administrative Areas in 1911.

Ratio of the proportional Number shewn in Diagram XV for Males and Females to those for Persons in each Population.

age structures of types of administrative area

These facts are especially clearly brought out in Diagram XVI, where the curves represent, not the proportional numbers living at the different age-periods, but the percentage by which these numbers exceed or fall short of the common standard. The degree of excess or deficiency is equally well brought out in this diagram, whether the numbers compared be large or small. It shows, as Diagram XV cannot do, the enormous excess of aged persons in the rural districts, and incidentally throws light upon the excess of old age pensioners in Ireland, which may be compared with the English rural districts in regard to the emigration of adolescents and young adults. Ireland, in fact, illustrates particularly well the manner in which this excess of aged persons comes about, as its case is simplified by the fact that emigration on a large scale commenced at a definite date, the time of the great famine. Had the Irish population of this period lost only young persons from emigration and excessive mortality its proportion of aged persons would have increased even more rapidly immediately after the famine than was actually the case. But the loss of young persons at that time would long before now have affected all ages the Irish population, since those who were lost neither grew old nor had children in Ireland It is the continued loss by emigration in later years which accounts for the present excess of aged persons in Ireland, for as these later emigrants are not yet old enough to have increased the aged population had they remained in Ireland, their loss has decreased the number of young adults and children, but not of aged persons in the Irish population. In the case of the English rural population the growth of industrialism may be compared to the great famine in Ireland. The earlier migration to the towns has doubtless reduced the numbers now living at all ages, but the continued drain of recent years has reduced the number of young adults and of children who would have been born to them in the country districts without as yet affecting the number of the aged.

The urban type of population is illustrated in its extremest form by the case of London, where the influx of young women especially distorts the curve at age 20-30 years to an extraordinary extent. This feature of the urban curves becomes progressively less marked with decreasing size of the towns dealt with, and in this, as in other respects, the distribution of population in the smaller urban districts is very similar to that in the country at large. It may perhaps be inferred that the population of the smaller towns is at present less affected by migration than that of any of the other areas considered. Its constitution at least seems to be less disturbed by this cause, though it has increased more since 1901 than any of the other populations here considered.

The London distribution of population is the most dissimilar from the rural also in early life, for the numbers of children are especially low at all ages, although the proportion of males of fertile ages is slightly and of females greatly in excess. As well as being low throughout childhood, the proportions living in London fall more rapidly at this time of life than in either the county boroughs or the urban districts (Diagram XVI). This can scarcely be accounted for by the mortality to which they are exposed, which in 1911-13, for which years alone the information is at present available, was considerably less than that prevailing in the county boroughs. The surmise may be hazarded that this feature of the London curve represents movement of increasing families from the county to the more spacious surroundings of the outer ring, where house room is cheaper. If these considerations to any extent determine the movement to the suburbs, a certain amount of emigration of children from the County of London is presumably going on, which would account for this peculiarity of the London figures (see page 77).

The urban type of age-distribution is most strongly marked in old age not in London, but in the county boroughs. Diagram XVI shows that the proportions of old people in London and the smaller urban districts are very similar to those in the country at large, but that in the county boroughs this proportion is greatly below the mean. To a large extent, at least, this is accounted for by the fact that the mortality of the aged in these boroughs is much higher than in London or any of the other areas dealt with (Annual Reports of the Registrar-General for 1911-1913).

The divergence of the urban types and of the rural type of age-distribution from the mean for the whole country is illustrated in another form in Table XX, and in Diagram XVII, which shows the proportion of the population at several ages in the principal groups of administrative areas to the total population at the same ages in England and Wales.

DIAGRAM XVII.—Proportion of the population at several ages in Groups of Administrative Areas to the total population at the same ages in England and Wales, 1911.

age structures of types of administrative area, for males and females
age structures of types of administrative area, for persons

As a result of the differences in age-distribution between urban and rural populations it follows that in the country at large persons of different ages are very differently distributed amongst urban and rural areas. Thus Diagram XVII shows that while in early adult life less than 20 per cent, of the population was enumerated in the rural districts, this proportion rose with advancing age to nearly 35 per cent, at ages 85 and upwards. The Comity Boroughs, on the other hand, contained over 30 per cent, of the population at all ages up to 40-45, but beyond that age their proportion of the total falls to reach 20 per cent at 85 and upwards. The proportions of the total population contained in London and in the smaller towns vary with age very much less than those in the county boroughs and rural districts; and in the latter the variations are much greater for males than for females, the stress of mortality at the later ages in the large towns falling especially upon males.

It may be noted that the early adult ages, at which the distribution by type of area is least favourable (i.e. , at which the largest proportion of the population is subjected to the conditions of city life) are precisely those at which mortality in England and Wales compares most favourably with that of other countries.


Sex —The differences in sex-proportions at the various ages in the populations under consideration have necessarily been referred to in dealing with ages and are so clearly expressed in Diagram XVI that they require little description. The portion of this diagram referring to England and Wales shows clearly the sudden drop in proportion of males at age 20-25, followed by partial recovery for a time, and then by a much greater drop in old age, already referred to on page 60. This holds good to a large extent also of the county boroughs and smaller urban districts, but in London and the rural districts the effects of the causes determining the sex-proportions in the country at large during the working period of life are swamped by inward and outward migration respectively. The excess of male children and adolescents in the rural districts contrasts markedly with their deficiency, except at the first age-period (when the excess of male births still tells) in all the urban populations dealt with. Migration can account for the difference in regard to adolescents—Diagram XVI plainly shows that the exodus of females from the rural districts precedes that of males—but is not obviously capable of doing so at the earlier ages. The difference does not depend upon any peculiarity of sex mortality in rural children, for the ratio of male to female mortality is much the same in the country as in the towns, though "both are, of course, considerably lower in the country. As males are subject in early childhood to a heavy excess of mortality in the rural as well as in the urban districts it is not easy to see why the excess at age 0-5 which they have in common with the males of the towns does not give way, as in all the urban districts, to deficiency in the next age-period. Comparing the curves for the rural with, those for the total urban districts in Diagram XVI, it will be seen that the chief influx of females into the towns appears to occur five years later than their chief exodus from the country. That the former is not entirely due to the latter is shown also by the curves for England and Wales in Diagram XV. These show that the heaping up at age 25-35 in the curve for persons of both sexes, referred to on page 61, is due entirely to a much greater heaping up of females at these ages, an actual depression occurring here in the curve for males. Taking urban and rural females together then, the rural depression is submerged "by the urban elevation, and could account only for a portion of it even if the ages corresponded. This peculiarity of the England and Wales curve is apparently to some extent due to immigration of young women from other parts of the United Kingdom, who would naturally leave their homes later when seeking employment at such a distance than if it were merely a case of finding work in a neighbouring town. This is shown by the fact that the heaping up is somewhat less marked in a curve (Diagram XXVII) representing the age-distribution of females in the United Kingdom. Even in this curve, however, there is more of this heaping up than can well be accounted for by immigration of young women from abroad, and the probability remains, as was pointed out in the report on the Census of 1901 after detailed study of the figures, that the apparent number of females at this age is considerably increased by wilful mis-statement of age.

Populations of individual areas. —In Table 38 of the Summary Volume the age- and sex-distribution of the population of Administrative Counties, County Boroughs, Metropolitan Boroughs, and of Urban Districts containing over 50,000 inhabitants is set forth in the form of a statement of the number in each age and sex-group in proportion to 100,000 total population; and in Table 39 these proportions are compared with those for the country at large, the males in each age-group being expressed as so much per cent, of the proportion of males in England and Wales at the same age, and similarly for females. It is hoped that these tables, by facilitating comparisons of the distribution of population in the areas dealt with, may prove of service to students of their economic circumstances. This field of study is, of course, restricted in some degree by the labour of converting the tabulated ages to a readily comparable form, and it must be understood that the following remarks are restricted to the areas included in Tables 38 and 39. If smaller areas could be dealt with as well, many more examples of the different types of population discussed would be met with, and the features of these would probably be more pronounced, since the populations of large areas must necessarily tend more than those of small ones to mixture of type. The material in the tables, however, suffices to bring out several points of interest.

On glancing over Table 39 the eye readily distinguishes excesses, as compared with England and Wales, from deficiencies, for the former are denoted by three figures as being over 100 per cent., and the latter by two only. Study of the table soon reveals a certain number of characteristic arrangements of these excesses and deficiencies, which are repeated over and over again, and which prove to express certain typical distributions of population. Thus there is a type in which both males and females are in excess of average for the first three or four age-periods, and in deficiency, with few exceptions, throughout the rest of life. The groups of populations dealt with have been assembled solely on the ground of their conformity to one or other of the types to be described, no account being taken of the class of area or nature of occupation chiefly pursued. It will be found, however, that groups of populations similar to each other in regard to age-and sex-distribution are, as a rule, very similar also in regard to class and occupation. The method of grouping described was followed as affording the most satisfactory test of this correspondence, since there is much less room for bias as well as less need for knowledge in the selection of areas on this system than if the converse method had been followed of selecting certain areas as agricultural, mining, residential and the like and then ascertaining to what extent the distribution of population is similar in the members of each group.

If the populations had been grouped by their supposed economic circumstances, the selection would have been much more difficult; and when made would have left more room for differences of opinion as to its justification. Moreover, a number of populations display peculiarities in Table 39 which might not be foreseen when making the selection, even though readily explicable, as by the presence of institutions, etc., when attention is drawn to them. To reject such populations from a grouping by economic circumstances might savour of illegitimate manipulation, whereas under the procedure adopted they fall out automatically.

An additional reason in favour of the course adopted is that it avoids the danger presented by the alternative method that a faulty choice might be made of the features most likely to lead to similarity in age and sex-distribution.

Although, however, the method employed seems to present more advantages than the other, both would probably be found useful in a more complete investigation.

Proceeding on the plan indicated, we have been able to assign 119 out of the 192 populations in Table 39 to one or other of the six types to be described, or about three-fifths of the whole. Some of the remainder can be recognised as of mixed type, intermediate between two of the groups described, but in a considerable number we have not succeeded in recognising conformity to any particular type. Our attention, however, has been exclusively directed to the easily recognisable and outstanding types of distribution., and it may well be that others, probably of less importance, are to be found amongst the unclassified residue of populations. In Tables XXI to XXVI, which show the sex and age-proportions in the typical populations compared with the corresponding proportions in England and Wales, the excesses are printed in heavy type and the deficiencies in light type.

Group 1. —These are populations showing excess of both males and females during childhood and deficiency at nearly all adult ages. The first twelve areas in Table XXI may be regarded as pure examples of this type, while the remaining five are to some extent intermediate between it and the next group. The curves representing the group in Diagram XVIII, which depict the distribution of the combined populations, have therefore been constructed to refer to the first twelve areas only, though they are not seriously modified by inclusion of the remaining five. The first two sections of the diagram compare the proportion of males and of females at each age-period shown for the group in Table XXI with the corresponding figure for England and Wales. They show the number of young children in very considerable excess, which rapidly diminishes with increasing age till at 20-25 in both sexes the proportions fall below those for the country at large, remaining in deficiency, greater on the whole in the case of females than of males, throughout the rest of life. The curves for the two sexes approximate closely to each other in childhood and middle age, males being fewer than females in early adult life and old age. This may be seen in the third section of Diagram XVIII, which shows the. percentage excess or deficiency of males and females at the various ages in 100,000 of the group population as compared with 50,000 persons of both sexes in England and Wales on the plan followed in Diagram XVI.

Table XXI shows that the group consists of ports and manufacturing towns. The circumstances which have led to its particular type of distribution appear to be relatively high birth and death-rates, absence of attraction for female labour in the form of domestic service or otherwise, and want of attraction also for male labour, at least during recent years. We shall return to these points later, proceeding meanwhile to a short description of the other types distinguished.

Group 2. —The type of distribution here is similar to that shown by Group 1 except in one respect. Demand for male labour is shown by the fact that the proportions of males remain above the average for the country till about age 45 (Table XXII and Diagram XIX). Otherwise the curves are very similar to those for Group 1, there being the same excess of children and deficiency of old people. The distribution of females in fact is practically the same as in Group 1, but with some general reduction in their numbers owing to the larger proportion of males during the working period of life. The list of populations included in the group shows that they are largely those of mining areas, with some ports and manufacturing towns. Evidently the factors determining this type of distribution are relatively high birth and death-rates, absence of attraction for female labour, and a type of industrial activity which has led to continued immigration of young men. Glamorgan, the West Riding of Yorkshire and Durham, though conforming to this type of distribution, are not included in the figures for the group given in Table XXII, upon which the diagrams are founded. The reason is that towns belonging to these administrative counties are included in the group, so to admit the counties would involve counting populations twice over. It has been thought better to prefer the towns to the counties as being naturally purer examples of the type, but the counties, though excluded, require mention as conforming to the type. This course has been followed also where similar circumstances apply to other groups.

Group 1.—Excesses of Males and Females in Childhood and Deficiencies throughout Adult Life.


Proportions of males and females at different ages: Group 1

Group 2.—Females as in Group 1. Excesses of Males up to Age 40-50.


Proportions of males and females at different ages: Group 2

Group 3.—Deficiencies of Males and Females in Childhood and Old Age, and Excesses at Intermediate Ages.


Proportions of males and females at different ages: Group 3

Group 3 presents a third type of industrial population totally different from Groups 1 and 2, and of considerable interest. Its outstanding features, as shown by Table XXIII and Diagram XX, are excess both of males and females throughout the working period of life, and deficiency in childhood and old age. Blackpool and four Central London areas present these features also (Table XXIII), and perhaps strict adhesion to the plan, of selection described above would involve their inclusion in the group for all purposes. Their similarity to the nine textile towns in the group, however, proves on further examination to be to some extent superficial rather than real, for their distribution curves show them to be as much akin to Group 5 (residential areas) as to Group 3. For this reason, as well as because it seemed of interest to show the distribution in these textile towns without admixture with populations of a very different nature, the curves relate to the textile towns alone.

The low birth-rate of these towns, notwithstanding their large proportions of young and middle-aged adults, is reflected by the low proportions of children they contain, which are about as much below the average as those of Groups 1 and 2 are above it. The proportions are soon brought above the average, and sooner in the case of females than of males, presumably by immigration of workers from outside areas, attracted by the employment offered in the mills. This movement is, as might be expected from the nature of the prevailing industry, much more marked in the case of females, and the rise, for reasons given on page 85, does not appear to be entirely due to immigration. The most striking feature of this group is the low proportion of children, along with the high proportions of adults of fertile ages, the latter not being, as in the case of females in the residential areas, largely cut off from married life within the areas in question. The conclusion seems inevitable that the abundance of industrial employment for women in these towns leads to their low birth-rates, whereas in Groups 1 and 2 the opposite state of affairs prevails. Recent investigations into this point in individual areas seem to show that the gainful employment of married women does not necessarily lower their fertility, but in these textile towns, where their employment is outside their homes, it would seem that at the present day it must have this effect. Of course, the facts may differ considerably in the case of the other textile towns in Tables 38 and 39, none of which conform, to any types now dealt with, but the association of this type with the textile industry is none the less apparent.

Group 4. —This is a small group of suburban districts in and around London, with a very distinctive age-distribution of population. No doubt this type is to be met with 011 the outskirts of other large cities as well, but it is only in the case of London that portions of a city population of distinctive type appear separately in Tables 38 and 39, so, as the type cannot be looked for elsewhere than in the growing suburbs of a large town, it is only represented in Table XXIV by districts of Greater London.

The type is distinguished by excess over the general average during the first three quinquennia of life, succeeded by deficiencies during the fourth and fifth. Next follows a second period of excess from about age 25 to 45, followed by a second period of deficiency throughout the rest of life (Diagram XXI). These statements apply equally to both sexes, bat the second period of excess commences a little earlier with women than with men.

The explanation of this curious in and out age-distribution is readily apparent when attention is directed to the nature of the areas comprising the group. They are all more or less of what may be described as the "workman's dormitory" type. This type of distribution can only be looked for in the working-class districts, for elsewhere immigration of females as domestic servants, just at the age when in this group both sexes are in deficiency, completely alters the distribution in that sex. Apparently there is a tendency for couples who marry in London to move into these outer suburbs as their families begin to increase, in quest no doubt of increased houseroom and of purer air and more open space for their children (see page 68). If this is so, the proportion of children would naturally be high, their presence being the reason for the move. The population of these districts is, in fact, a selectedly fertile one, and so the proportions of children are above average. As the children grow of age to leave the parents' homes the proportions fall below the-mean. This may result from emigration and also to some extent from the rapid growth of many of the districts, the number of newcomers being sufficient to impress the population with their own type of distribution, that of parents with, families of young children. The second period of excess commences when the age at which the immigrant parents enter is reached. Naturally the wives are a little younger on an average than their husbands, and so the rise occurs a little earlier in the female sex, but domestic service may also contribute to this feature. The deficiency in old age is probably due, like that at 15-25, to the fact that the districts are increasing ones and the movement constantly going on, so that the balance is being constantly disturbed in favour of excess at the two periods of life when excess is characteristic of the group. These are, in fact, areas of immigration, and the numbers of elderly persons are low for the converse reasons of those which cause them to be high in districts and countries of emigration (see page 67). For some reason this deficiency is more marked in the case of males. (Diagram XXI, Section 3.)

Group 4.—Excesses of Males and Females in Childhood and at Ages 25-45; Deficiencies at 15-25 and over 45.


Proportions of males and females at different ages: Group 4

Group 5. —The age-distribution in this group is most distinctive in the female sex. Table XXV shows how uniformly the proportions of females are below the mean during the first three quinquennia of life and above it afterwards. In this respect the conditions met with are almost exactly the converse of those obtaining in Groups 1 and 2. The proportion of children is very low, but after the age of 15 is reached the curve showing distribution of females in the third section of Diagram XXII suddenly rises to a remarkable height, and thereafter remains much above the mean throughout the rest of life. It follows that this group is characterised by great excess of females over males. In view of the character of the areas included in it, there can be no doubt that the violent disturbance of the curve for females after the age of 15 is due to importation of young maidservants from areas of other types. The excess soon diminishes somewhat as these marry and move away, but always remains very large and becomes extreme in old age. Doubtless as life advances the excess of females in these residential districts is due less to the maids and more to the mistresses, far more women than men being free to select a residence in such favoured localities.

The males of this group present two types of age-distribution. In both the proportion in childhood is low, and thereafter tends to increase ; but whereas in one type, represented by the first section of Table XXV, this increase continues throughout life, so that from. 50-60 onwards the numbers rise above the mean, in the other they fall again as old age approaches, the mean never being reached at any period of life. The reason for this difference is obvious when we consider the nature of the areas included in the two sections of the group. The first section is made up entirely of places to which men retire after their life's work is done (this applies to Chelsea because of the military pensioners included in its population), and consequently the increase in proportional numbers of males living continues throughout life. The second section is made up entirely of "residential" districts in and around London for the reason already advanced in the case of Group 4, namely, that populations of this special type large enough for inclusion In Table 38 cannot be looked for except in the metropolitan area. The type is doubtless widely represented in the suburbs of other towns, but to single it out, _small areas, wards in most cases, would have to be dealt with. This is the class of district inhabited by the middle-class worker and from which in old age he frequently removes to areas of the other section included in the group or elsewhere, so it is natural that in old age the relative numbers of men should fall in such districts. _The rise in old age of the female curve is, as might be expected, much less extreme in this than in the other section.

Other minor differences between the two sections are shown by Diagrams XXIII and XXIV. The curve showing distribution of males in the working areas is characterised by a rapid fall in childhood, followed by a considerable rise in early manhood, which is not found in the case of the other section. This rise apparently represents immigration of young men, a natural feature in such areas. It is tempting to compare this curve with that of males in Group 4, as the form is somewhat similar, and Group 4 after all represents the residential districts of the working classes. The importation of domestic labour, which makes the female curves for the two groups so utterly dissimilar, affects the male curve under discussion comparatively little, except indirectly in so far as by raising the proportion of females it lowers that of males at all ages. For these reasons parallelism might be looked for, and to some extent it exists, but it seems doubtful whether the rise at age 20-30 means the same thing in the two cases. There is no reason to suppose in the case of Group 5 as in that of Group 4 that the Immigrants are married men. Probably most of them are not, so the similarity between the two is perhaps more apparent than real. The rapid fall during childhood is confined exclusively to the working section of Group 5. This may be due to the presence of boarding schools in the areas comprising the other section. On comparing the curves for females in the two sections, it is seen that the army of maidservants is much more formidable in the London districts.

Group 5.—Deficiencies of Males and Females in Childhood; Excess of Females in all other Age-Groups and at all Ages.


Proportions of males and females at different ages: Group 5, overall


Proportions of males and females at different ages: Group 5, retirement areas


Proportions of males and females at different ages: Group 5, London suburbs

Group 6.—Deficiency of Young Children and Young Adults; Tendency to Excess in later Childhood and great Excess in Old Age in both Sexes.


Proportions of males and females at different ages: Group 6

Group 6. —This is the largest of the groups described, comprising 35 different areas or about 18 per cent, of the total included in Tables 38 and 39. The age-distribution is in the main similar in the two sexes. The proportional numbers are below the mean, generally speaking, up to about 40-50 years of age, and thereafter in increasing excess. o this general statement one exception has to be made, namely, that in both sexes, but specially in the case of males, the numbers in later childhood and until the migration age is reached tend to exceed the mean. It has been pointed out (page 72) that migration of females appears to occur earlier than that of males, and accordingly we find that the excesses referred to chiefly affect the third quinquennium of life in the case of females, but apply quite as much to the fourth in that of males. These characteristics of the group are readily apparent on examination of Table XXVI, which forms a large sample of it, as well as from Diagram XXV, which refers to the population of the whole group. It will be noticed that the curves in this diagram are substantially the same as those shown for the rural districts in Diagrams XV and XVI, and it is evident from the list of areas included that they are of predominantly rural type. Long-continued emigration of young adults and low rates of mortality appear to be the conditions of the age-distribution characterising this group, in which only four towns are included, Canterbury, Gloucester, Great Yarmouth, and Ipswich. The rate of growth of the three former has been so low that they fairly compare with rural counties in this respect, but Ipswich has shown fairly substantial increase. Its population, however, is one of the least typical included in the group.

Comparing the rural districts in Diagrams XV and XVI with Diagram XXV it is seen that their agreement is closer for males than for females. In both cases emigration of females occurs earlier and is less in volume than that of males, but the difference in amount is much greater in the case of Group 6- than of the rural districts. This may be due to absorption of a considerable proportion of the females emigrating from rural districts in Group 6 as domestic servants in towns of the same counties. The drop in the proportions of males from age 15-20 to 25-30 is exactly the same in Diagrams XVI and XXV. The rural districts include a number of areas, especially mining districts and camps, to which there is an inflow of young men. This must tend to counterbalance for males in the rural districts the movement within Group 6 from rural to urban districts, and so may account for this difference between the sexes in the diagrams in question. The inclusion in the rural districts of mimng populations with comparatively high mortality also helps to account for the fact that the proportions of aged persons, especially women, are considerably larger in Group 6 than in the rural districts, but the migration of old people on retiring from active life must also affect the numbers living at these ages.

Birth and Death-rates in typical populations. —The types of population distinguished having now been described, it becomes of interest to examine how far the records of their vital statistics can explain the various characteristics noted and justify the deductions drawn from them. Materials for this inquiry are contained in Table XXVII, which summarises for each group the facts of chief importance from this point of view.

Table XXVII.—Birth-Rates and Death-Rates, 1901 and 1911, and Intercensal Rates of Increase, 1881—1911, in Typical Populations.

It should be mentioned that the figures included in this table are in a number of cases only approximations. For several of the towns dealt with the births and deaths are not available for the whole period 1901-10, changes of boundary have occurred in other cases which prevent exact comparison being made, and where counties are concerned the table necessarily refers to registration instead of administrative counties prior to 1911, and has therefore been made to relate to these throughout.

Group 1 is seen to be distinguished by high birth-rate, high death-rate, and high rate of natural increase. The actual increase, however, during the last ten years has been small, so there is at present shown a large balance of outward migration from these towns, though this was not the case in previous decades. The high birth-rate explains the large proportion of children in this group, but the numbers fall rapidly, owing to high mortality and emigration. The fall is specially rapid amongst males in adolescence and early adult life, the usual age for migration, in consequence no doubt of the comparative stagnation of these towns in recent years having led to emigration of young men in search of work elsewhere. After age 25 the fall in proportions of males shown in the third section of Diagram XVIII ceases for a time, and there is even a tendency to rise. This seems to point to a previous period when these towns attracted males aged 15-25, who, remaining in them, are now enumerated at later ages and prevent the curve in Diagram XVIII from falling, although the death-rate is high. This explanation is supported by the large increases in the population of this group shown in 1901 and especially in 1891. Confirmation of the supposition that it is only of late years that this considerable exodus of young men has been in progress is afforded by the fact that the proportion of females to 1,000 males has risen from 1,017 in 1901 to 1,032 in 1911, increase of females having exceeded that of males by 39 per cent. The loss of males at a time of life when the majority are unmarried does not involve, at first, so much loss of female population as it will in a few years time, if conditions remain unaltered.

The population of Edmonton, which increased by 38 per cent, from 1901-11, seems strangely placed in this group, and a distinct tendency to approximate to the distribution characteristic of Group 4, its presumably more natural position, may be noted in Table XXI. It is remarkable that so few anomalies of the kind occur under the system of grouping adopted.

Group 2 is seen from Table XXVII to be characterised by a birth-rate, death-rate, and rate of natural increase all somewhat lower than those of Group 1, but still distinctly above average. Its record in Table XXVII differs from that of Group 1 chiefly in the fact that the actual increase during 1901-11 was very nearly equal to the excess of births over deaths. This fact, of course, confirms the conclusion drawn from Diagram XIX as to demand for male labour in these areas.. It would seem from the curves and from Table XXII that immigration of males must have approximately balanced the large emigration of females from these areas, in which males at all ages exceed females by 2.2 per cent. "Unfortunately our records do not permit of analysis by sex of the rates of natural increase shown in Table XXVII, though this analysis will be possible in future, and it is therefore impossible to state the migration of the sexes separately. It may be confidently assumed, however, from the fact, that although females are in a minority in these areas their actual increase was not greater than that of the opposite sex, that emigration of females exceeded that of males. For where the distribution of the sexes is abnormal, nature constantly strives to redress the balance by providing larger natural increase of the sex in the minority. The unequal distribution affects the sex-proportion of deaths, but not that of births; and, as natural increase is the excess of births over deaths, preponderance of either sex in a population, by increasing the proportion of its deaths, tends to lessen its natural increase.

Four areas in this group, Bermondsey, Southwark, Stockton-on-Tees, and Dudley, show-either very slight increase or actual decrease during 1901-11. It is of course possible for special circumstances, for example, in London large common lodging-houses, to attract young and middle-aged men to a district which is not increasing, but such cases are exceptional.

Group 3 is marked in Table XXVII by low birth-rate and high death-rate with resultant low rate of natural increase, which, however, is somewhat higher than the actual increase which has occurred. It follows that notwithstanding the indications in Diagram XX of an influx of young workers, especially females, the balance of migration has been outward. This outward migration is not clearly impressed upon the curves, as is that occurring inwards, so presumably it is distributed over a much wider range of ages than the latter. The suggestion of inward migration in Diagram XX is to some extent increased by the fall in the birth-rate, which has been very heavy in these towns. Their native inhabitants aged 20-30 represent a higher annual number of births than that of 1901-10. The birth-rate for the group was 31.5 in 1881-90 and 27.3 in 1891-1900, so, as the increase in population during the last twenty years has been small, the number of births taking place twenty years ago was larger than it is at the present time. Decrease with decreasing age, owing to this cause, in the number of children and young persons living, no doubt adds to the suggestion of immigration conveyed by the curves.

Group 4.— Table XXVII shows that the inward balance of migration in these areas, previously very large indeed, was of comparatively moderate dimensions in 1901-10. Presumably the peculiarities of its age-distribution were much more marked at earlier censuses. Comparison in the case of a single typical area, East Ham, shows that this is the case. Moreover, in 1901, after the phenomenally rapid immigration into this town of the preceding decade, the age of maximum excess was earlier both for parents and for children than at present. Probably the heavy fall in the birth-rate of this group in 1911 compared with the decennial average rate is partly dependent upon lessening immigration of young parents. It seems doubtful whether Battersea, which decreased in population during 1901-11, is a district of the same character now as the others in this group.

Groups 5 and 6. —The figures in Table XXVII relating to these well-defined and familiar types of area are very much what might be expected, and call for little comment. It is interesting to note that the natural increase in the residential areas is almost equal to that in the rural group. This does not imply anything approaching to equal fertility in the two cases, for when the proportion of females at fertile ages is taken into account, the natural increase of Group 5 is very low in comparison to that of Group 6. The actual increase of Group 6 does not show the recovery during 1901-11 which characterised the rural districts as a whole (see page 38). Probably this is due to the fact that the type of distribution characterising Group 6 is incompatible with any considerable increase of population, so that the rural areas whose increase has been greatest are largely excluded from this group.

The emigration from Group 6 has been much less than that from Group 1, though the latter consists of large towns such as are generally looked on as attracting labour. It is probable that some of the emigration from these industrial towns is due to movement into suburban districts, but Table XXVII shows that these towns, as at present defined, have actually been the greatest exporters of labour of all the types dealt with.

Mis-statement of age. —Diagram XXIX shows the number of persons actually returned at each year of age compared with the number at the same age as calculated by the actuarial methods explained in Mr. King's report (see Volume VII, pages xxxix to xlviii). The differences between the two curves indicate to what extent the returns are vitiated by such mis-statements of age as are susceptible of correction by the mathematical methods employed. These mis-statements include at least three forms of error, which may be described briefly as (1) the error of "round numbers," (2) the "age next birthday" error, and (3) the error of "even numbers."

Error of "round numbers." —Although in the census schedule the "age last birthday" is asked for, there are many thousands of cases in which persons are returned as 30, 40, 50, etc., though they are actually a few years above or below one of those ages. Diagram XXIX shows that this error does not become apparent before age 30. It has already been shown, however (see page 62). that this source of error in the returns is less fruitful now than in earlier census years, and this conclusion is confirmed by Mr. King's report. Not only is there probably less suspicion that the results of the census inquiry may be used for other than their legitimate object, so that the majority of persons are more willing to impart accurate information; but they are also better able to do so owing to the spread of education, and to the fact that but a small proportion were born before the establishment of civil registration.

The extent to which this error of "round numbers" has prevailed at former English censuses cannot be fully determined; but that even at, recent censuses it operated much more than in 1911 is evident from Diagram XI, in which the gradual disappearance of the elevations in the curve at the alternate quinquennia containing the round numbers of age—40, 50, and 60—can be clearly seen. In the Report on the Census of 1901 a calculation based on a sample of the population was made, some comparisons with other records of ages were drawn, and the hope was expressed that this form of error was diminishing in England. A precise statement of the extent to which improvement as compared with 1901, appears to have taken place in the case of the selected area will be found in Mr. King's report (Volume VII, pages xlvii and xlviii).

Age next birthday. The form of error described under this heading has been assumed to exist from the fact that at all English censuses the numbers enumerated in the second year of life show an apparent deficit (see Diagram XXVI). In 1911 however at least, there are grounds for suspecting that this deficit extends to the first year of age as well, though here it is less obvious, as it cannot be deduced from the curve of age distribution. It becomes a question, therefore, how far an explanation which can apply to the second year of life only is true even for that year.

DIAGRAM XXVI: England & Wales:
Enumerated Population at each Year of Age: 1911.

age structure by individual year of age, in 1911

It has been assumed that in some cases the age next birthday is given on the schedule instead of the age last birthday. This error, if spread evenly over all the ages, would not appreciably affect any year except the first, for the gain from the preceding would approximately balance the loss to the succeeding year of life. There is, however, a check on this tendency as regards the first year of life in the instruction on the schedule that the age of children under one year is to be shown in months. Presumably, therefore, mis-statement of age is at a minimum in the first year of life, so the error in question first comes into play in the second year, from which it transfers children, without compensation, to the third. The slight effect of migration may be disregarded in this connexion, and the mortality in the second year of life so far exceeds that in the third, that the number living in the latter should be appreciably less than that in the former. The numbers actually returned, however, in the first five years of life were as follows:—


From these figures it has been assumed that there is a transference from the second to the third year of life; and examination of the figures for the several administrative counties and for their urban and rural aggregates in 1911 indicates that the same feature is met with throughout the country. The determination of its comparative frequency in the several areas shown in the tables is perhaps hardly of sufficient importance to justify the labour it would involve, but the following comparison with the European countries for which the figures are available may be of interest; in all the countries, except Denmark, the ages are tabulated according to the calendar year of birth, and as the censuses (with the exception of those in Holland and Sweden) were not taken at the end of a year, the figures represent only approximately the children in the second and third years of life.


These figures show that the excess of children in the third year of life over those in the second observed in the English census returns is not apparent in the majority of the eases included in the above statement.

While the figures as enumerated in this country point clearly to understatement in the second year of life and afford reason for believing that this is not shared by the immediately succeeding years, neither the numbers returned for the first year nor the reasons for believing age at that epoch to be correctly stated prove that those numbers are even approximately correct. And the considerations advanced by Mr. King in his report undoubtedly suggest that the understatement in 1911 may not be limited to the second year, but includes the first also. If this really is the fact it becomes a question how far the depression at the second year in Diagram XXVI is due to the explanation advanced above, and how far to omission of young children from the schedules.

It may well be that both causes of mis-statement are at work, for mere omission of young children could hardly stop short abruptly at the end of the second year, as it must be supposed to do in order to explain a discrepancy definitely limited to these two years. On the other hand, the "age next birthday" hypothesis is not only inherently probable, but does afford an explanation for the abrupt cessation of the shortage in the enumerated numbers after the second year,

Error of "even numbers." —The third irregularity in the figures referred to on page 85 has not been commented upon in previous reports It will be seen from the curve relating to enumerated population in Diagram XXVI that not only is the heaping-up at the multiples of 10 followed by depression at the next age, but that, speaking generally the frequency of all the "odd" years of life (multiples of five excepted) is less than that of the "even" years between the ages 25 and 70.

The elevation is greatest for some reason in the case of years ending in 8. It is natural, of course, that it should be greater at 8 and 2 than at 4 and 6, because the odd numbers 9 and 1 are particularly depressed by the overstatement at 0. If this were the sole explanation, however, the peaks at ages ending in 2 should exceed those at 8, since the excess at 0 has been shown to be drawn more from the age above than from that below (that is, more from 1 than from 9), whereas in fact the excess over the graduated curve is much smaller at 2, the peak at this digit in some cases actually falling below the graduated curve. It may be that the elevations at 2 are made less than they otherwise would be by transference of persons properly belonging there to the adjacent 0.

The rise in the curve at the years represented by the odd multiples of five may be regarded as analogous to the "round number" error.

The inquiry naturally arises whether the substitution of the question as to the date of birth in place of the age last birthday would be attended with any marked increase of accuracy in the replies. It is true that the returns at single years of age in some other countries present a more regular series than has been found at this census in England and "Wales, but it is at least possible that some part of their advantage may be due to the closer acquaintance of their inhabitants with official records than is common irx this country, and it may reasonably be supposed that some mistakes would be made in deducing the date of birth from the known age.

Wilful mis-statement of age. —In addition to the three forms of error so far discussed, arising generally through ignorance or carelessness, there are undoubtedly many cases of wilfully false statement of age, some persons being returned as many years older and others as many years younger than their true age. The general effect of such mis-statement appears to be inconsiderable, except among females in early adult life. Diagrams XV and XXVI indicate that the number of females aged 20-25 and 25-30 is disproportionately high. This may be due either to immigration or to wilful mis-statement of age, the tendency to which, in this period of female life, has frequently been commented upon. Diagram XXVII gives an indication of the extent to which immigration suffices to account for the phenomenon; it shows that the exaggeration at the ages in question is appreciably less in the United Kingdom than in England and Wales alone, from which it may be inferred that a part at least of the excess is due to immigration from other portions of the United Kingdom. If the effect of immigration from abroad could be included as well, doubtless the aggregation in the curve at these ages would be further accounted for. Even after making full allowance for immigration, however, it would appear that there must have been a certain amount of transference from the later ages, and the calculations given on pages 59 to 61 of the Report on the Census of 1901 indicate the probable extent-of wilful mis-statement of age by this section of the population.


Age distribution of females in UK and in England and Wales

Ages not stated. —In some cases no statement of age has been furnished on the schedule, and in other cases, generally where the return has not been made personally some indefinite statement is inserted, the age being shown, for example, as "between 30 and 40." The unstated ages in a given area have been distributed proportionally among the stated ages, and the indefinite ages have been assigned in accordance with the information furnished. The number of such cases at previous censuses is not on record, but for 1911 the numbers in the several administrative counties and county boroughs are shown in Table 17, Vol. VII. The following table shows how the number of defective age-returns in this country compares with those in some other European States:—

Country. Date of Census. Total
Numbers of
Ages not
England and Wales 3rd April, 1911 36,070,492 13,167
Belgium 31st December, 1900 6,693,548 8
Denmark 1st February, 1901 2,449,540 6,090
France 24th March, 1901 38,450,788 116,772
Holland 31st December, 1909 5,858,175 84
Italy 10th February, 1901 32,475,253 1,442
Prussia 1st December, 1900 34,472,509 6,741
Spain 31st December, 1900 18,618,086 20,698
United States 15th April, 1910 91,972,266 169,055


Of the 36,070,492 persons enumerated in England and Wales in 1911, 20,963,807 were returned as unmarried, 13,126,070 as married, and 1,980,615 as widowed. Among the unmarried there were 1,029 females to 1,000 males; the number of wives enumerated in the country exceeded the number of husbands by 134,498, and were in the proportion of 1,021 wives to 1,000 husbands; and the number of widows was 1,364,804 as compared with 615,811 widowers. In 1901 the ratio of unmarried males to unmarried females of all ages was 1,028 to 1,000; the wives exceeded the husbands by 106,156 and were in the proportion of 1,019 to 1,000; and the widows numbered 1,246,407, against 550,330 widowers.

The average annual marriage-rate was rather less in the ten years 1901-10 than it had been in 1891-1900, while the mean age at marriage slightly increased; it might, therefore, be assumed that the proportion of married persons in the population was less in 1911 than at the preceding census. There were, however, 372 married males and 356 married females per 1,000 of each sex according to the recent census, against 357 males and 340 females in 1901, and these latter numbers had shown a similar, but rather less marked, increase over the proportions in 1891. The explanation is to be found in the fact that owing to the changes in the age-constitution of the population a much larger proportion of the total is now living at the marriageable ages (see page 61). The general decline up to 1901 in the proportion of married persons in the adult population is shown, in the following table.


Tables 45 and 46 of the Summary Volume give comparative figures for males and females respectively from 1851 onwards, and show their distribution over the several age-groups. Thus, for example, the 633 married males in 1901 included 29 aged 20-25, 177 aged 25-35, 177 aged 35-45, and 129 aged 45-55, while in 1911 the numbers at the corresponding ages were 21, 167, 184, and 135 respectively. The married females showed a similar decline at ages under 35 years. If the proportion of married persons is calculated separately for each age-group it will be seen that there has been a general decrease among both sexes at each age-group since 1881.

The decrease in the proportion of married is most apparent at the younger age-groups, no doubt as a consequence of the modern tendency to postponement of marriage, which chiefly affects these ages. Moreover, the diminution at the later ages must be due in part to the fall in the death-rate, in consequence of which the proportion of persons becoming widowed in early and middle life has lessened. The effect of the latter factor is indicated by the proportions of widowed persons, which, in spite of the diminished frequency of re-marriages, both of widowers and widows, and of the increased expectation of life of the widowed as well as of other persons, have steadily declined since 1891.



Proportions of unmarried, married and widowed at different ages

Divorced persons were not asked to state that fact on the schedule, and have, therefore, been included as unmarried, married, or widowed according to the manner in which they were returned. In abstracting the particulars from the schedules written in Welsh it has been found that the word "gweddw" has been used in some cases instead of "sengl" to indicate the unmarried condition; this may have led to a slight overstatement of the numbers of widowed persons in the Welsh counties, but comparison with the figures for 1901, when these schedules were translated into English by the enumerators and any similar error was presumably corrected, indicates that the number of such cases is not very considerable.

The number of cases in which no statement was made as to the marital condition are shown in Table 17, Vol. VII, and these have been distributed among the stated conditions.


It has already been shown (pages 85 to 87) that the returns of ages as given in the census schedules are subject to error arising from various kinds of mis-statements. For actuarial purposes it is desirable, instead of accepting the returns as they stand, to correct them as far as possible by calculating from all the available data the probable number living at each year of age. For this important work we were so fortunate as to obtain the services of Mr. Geo. King, Fellow of the Institute of Actuaries, and the valuable tables prepared by him, together with his report, will be found in Volume VII of the Census Report (Cd. 6610), from which are reproduced the following diagrams showing the discrepancies between the enumerated and the calculated numbers of persons, and the tables showing, according to marital condition, the numbers of males and females as calculated by Mr. King's method of graduation.


Enumerated population compared with graduated population at different ages



1 See note on page 63.

2 The figures for the French Census of 1911 have been received while this volume was passing through the press, and too late to be utilised in the preparation of Diagrams XIII and XIV, which show the age-distribution in France in 1906. Generally speaking, the later figures show diminished proportions under 30 years and increased proportions above that age.

Next Selection Previous Selection