Introduction

Introduction

I. The background and nature of the assessments

The object of this volume is to present an index to the names of individuals enumerated in the City (within the Walls) in connection with the 1694 Act (6 & 7 Wm. & M., c. 6). Such an index is clearly of the greatest value to historians, especially when used in conjunction with the original assessments and with other contemporary materials. But before discussing the uses and limitations of the information collected under the 1694 Act, it is necessary to explain what that Act prescribed. (fn. 1)

The official purpose of the 1694 Act was to provide revenue, for carrying on the war against France, by levying taxes upon burials, births and marriages and annual dues upon bachelors over 25 years of age and upon childless widowers. The Act came into force as from 1 May 1695, initially for a five- year period, but was later extended until 1 August 1706 (by 8 & 9 Wm. III, c. 20). How this particular means of raising revenue came to be proposed, is not known. Among Gregory King's papers there is a printed broadsheet entitled A probable calculation of the annual income to be raised by a tax on marriages, burials, and legacies, with a note in King's handwriting on the back — 'Fryth's Project of the Duty on Marriages, Births and Burials'. And after the Act had been passed, the Treasury received a petition from Richard Firth, claiming that he had (some six years earlier) suggested such taxes to the Duke of Shrewsbury, who had mentioned it to the King, and 'in this sessions it was accepted by the House of Commons; praying their bounty for his charge and pains'. The Treasury comment was: 'To be considered if there be any places to be disposed on this fond.' No other supporting evidence has so far been found. The references in the Journals of the House of Commons are entirely formal and give no indication of the proposers or of the arguments they had in mind. Various amendments to the Bill were made during its passage through the House. One was to exempt from the tax on bachelors fellows and students of Oxford and Cambridge, 'where, by the Statutes of their Colleges, they are to be displace, if they shall marry'. Another was concerned particularly with recording the deaths (and descent) of persons of quality. (fn. 2) But the Bill appears to have had an easy passage; introduced in the Committee on Ways and Means in February 1694/5, it was sent to the Lords on 8 April and their assent was notified on 12 April. Nevertheless, an aura of mystery still surrounds the Act. Three points arise, in particular.

First, the apparatus required for implementing the Act was quite formidable, involving nothing less than a complete enumeration of the population and a comprehensive system of vital registration. Because some of the taxes fell upon individuals with specified characteristics (childless widowers and bachelors above the age of 25 years) it was necessary to know the names of all such individuals. In addition, the taxes were graduated in accordance with the social status of the individual — and graduated in a complex way. For example, the tax on the burial of a 'common' person was 4 shillings. For a Duke the tax was £50 4s for himself, the same for a Duchess, £30 4s for the eldest son or his wife, and £25 4s for a younger son or his wife or for an unmarried daughter. And similar variations applied to the duties on births and marriages and to the annual taxes on bachelors and widowers (Table 1a and b). The Act came into force on 1 May 1695 and on or before that date the Assessors were supposed to furnish the Commissioners with complete lists of the population in their areas, specifying their names and surnames, estates, degrees, titles and qualifications and indicating the taxes and duties to which they were liable, or would be liable if a specified event occurred (that is, a marriage, birth or death). Moreover, these lists were to be brought up to date each year, being corrected in respect of the 'death change of quality or degree or removal of any person or persons or otherwise' (Sections XI and XVI). Lodgers and servants were to be included in the enumeration, for they were to be taxed at their place of residence (Section XXIV). A further complication was that in London the basis of the enumeration was the parish, in contrast to the more usual ward basis. This may well have involved substantial problems, especially in the recruitment of a large and separate body of Assessors and Collectors.

The parish basis was linked to the requirements for vital registration, for in order to prevent evasion it was necessary to improve the reliability and scope of the parochial system. A double check was provided. First, the scope of the parish register was widened to cover everyone married, buried, christened or born in a parish (Section XX), and stillbirths were to be notified (Section XXI). And persons in Holy Orders (and their substitutes) were instructed to keep accurate records of marriages and burials and of all persons 'christened or born', under penalty of a fine of £100 (Section XX). Secondly, a special responsibility was placed upon parents in respect of children born to them; it was their duty to notify the Collectors, within five days of the occurrence of the birth of a child, whether live or stillborn — and a stillbirth had to be attested by two or more persons (Section XXI). A similar duty was placed upon Quakers, Roman Catholics and Jews (and any other comparable individuals) to report their marriages to the Collectors within five days (Section LVII). And special arrangements were prescribed for 'the better preserving the Genealogies Descents and Alliances of the Nobility and Gentry'. Upon the death of any of them (anyone liable to a burial duty of 20 shillings or more), the person responsible for paying the duty had to provide the Collectors with a certificate showing the 'name surname quality office employment (if any) of such deceased person with the age time of death place of burial marriages and issue and the ages of such issue together with the names sirnames titles and qualities of the parents of such deceased persons. . .'. These certificates were to be sent to the Receiver-General or his deputies for transmission to the College of Arms, which institution was instructed to 'number schedule and digest the same in alphabetical order in Books to bee provided for that purpose' and to file the originals for public use (Section L).

It is clear from the Act that a great deal of thought had been given to creating a system which — at least in theory — would provide a full range of demographic statistics. Moreover, later amending legislation attempted to deal with the technical defects which became visible in practice. Thus it was not made completely explicit in the original Act that all births were to be covered by the parish register. Under 7 & 8 Wm. III, c. 35, therefore, it was made clear that the parents 'of every Child which shall be att any time born after the said Day and Yeare & during the continuance of the said Act' had to notify, within five days, the Rector, Vicar or other appropriate person (Section IV). And when individuals were buried in a parish different from that in which they had normally lived, the notification of death was to be sent to their parish of last residence (Section VI). A second amending Act (9 Wm. III, c. 32) endeavoured, among other provisions, to ensure more detailed and reliable entries in the registers (Section III). To prevent people from evading their taxes by moving away from the parish in which they were liable for payment, arrangements were made for the taxes to be levied in the parish to which they had migrated (Section V).

Provisions and stipulations of the kind do not, of course, mean that the Acts were effectively enforced. They were not. Evidence of defects will be discussed later and in any case the defects were recognised and in part condoned by an Act of Indemnity (4 & 5 Anne, c. 23). What is important to emphasise here, however, is that the legislation envisaged and attempted to create a statistical system without precedent in Britain. A complete enumeration was involved, with an indication of marital condition; no subsequent complete census was taken here until 1801 and marital condition was not included in British decennial censuses until 1851. Registration of births was to take place in five days (as compared with forty-two days under our present system), and deaths were to be transferred back to the last place of residence, a provision which was not included in our present system of civil registration until the present century, and the lack of which before that time complicates the problem of calculating local death-rates. As for the special register of the vital statistics of the nobility and gentry, nothing comparable was provided subsequently by official action; the various private publications, used by Holdings-worth in his studies of the demography of the British peerage, are neither as full nor as reliable as the proposed official register might have been. (fn. 3) Given all this, and given also the fact that the 1694 Act was initially implemented for only five years (and was renewed only for a further five years) there is every reason to ask whether the main purpose of the Act was fiscal, or whether there was at least a parallel objective of collecting reliable statistics on the population of the kingdom. Unfortunately, no clear and documented answer can now be given. The one direct reference so far found is in an eighteenth-century pamphlet, concerned with the 1753 census proposal. The anonymous author wrote: 'As you know I am now very aged, I remember when Numbering the People in King William's Time was attempted with the greatest Care imaginable. Mr. Lock, Sir Isaac Newton, and other great and curious Men were then alive. And to render the Thing effectual, it was thought that giving the Government a Poll-Tax would make the Matter be carefully executed.' (fn. 4) Though this is only a feather's weight of testimony, the statement is not implausible. (fn. 5) Political arithmetic was certainly in the air. Graunt's Observations had been given a very favourable reception; it had been reprinted several times and, even so, was reported by John Aubrey to be 'now very scarce'. The Royal Society had in the 1660s set on foot a direct survey of agricultural practice. A contemporary writer had argued that the population of England had probably been generally underestimated. As to the actual number, he proposed to reserve his judgment until 'some such accurate Survey hath been made thereof, as I have heard Sir W.P. that Mathematical States-man wish for'. But he nevertheless was prepared to assert as highly probable 'that the Total of the number of the people here will upon any actual view hereafter to be made by publick Authority, appear considerably greater than any cautious Calculators have made it'. (fn. 6) None of this is conclusive. But it does at least suggest that in England — as in France — there was at the least some expressed and informed interest in 'numbering the people'.

The second main point in the mystery — and this tends in some degree to reinforce the first — relates to the fiscal objective as such. How much revenue was it hoped to raise and how much was raised in reality? When the Bill was passing through the House of Commons a clause was proposed enabling the Bank of England to lend 'above' £1,200,000 on the anticipated proceeds of the Act. (fn. 7) But the clause was withdrawn, and the Act itself provided (in Section LI) that loans of up to £650,000 at not more than 8 per cent interest might be raised upon the credit of the Act. Gregory King estimated the probable yield at a lower figure. Allowing for 'omissions, frauds and insolvent', as well as for those 'excus'd by receiveing alms', and covering both the ordinary rates of tax and the surtaxes on persons of quality, he put the annual yield at about £81,050. (fn. 8) That is already a low figure both as compared with other taxes and in relation to the great apparatus set up under the Act. And in practice the Act proved to be still less profitable. That it was likely to be unpopular is implied in Davenant's attack on it, as a 'Fine upon the Marriage-Bed', involving 'a very grievous Burthen upon the poorer Sort, whose Numbers compose the Strength and Wealth of any Nation.' (fn. 9) Waters, writing about the 1694 Act assessments for the parish of Melbourne, Derbyshire, stated — though without providing evidence — that 'the tax was so unpopular that the local authorities took pains to destroy the machinery for levying it.' (fn. 10) But there is evidence in the Calendar of Treasury Books; of alleged fraud on the part of one collector; of the omission to send duplicate assessments for various places to the Receivers-General; and of the failure of various agents to make any payments in respect of the duties. (fn. 11) As late as 1698 there was a reprimand to the Receiver-General for London, Middlesex and Westminster, as concerning the 'great neglects. . . in not examining the produce and duplicates of the Duties . . . and in failing to afford the receipts of the same as directed ... and that thereby the amount of said Duties has fallen considerably short of what might be reasonably expected'. (fn. 12) And in the town of Newark, the initial assessment required by the Act was met with violent opposition. 'William Wynne, another of the said Surveyors, endeavouring to make a survey of the town on Tuesday the 12th inst., the common people of the town got together in a tumultuous and disorderly manner, and with stones, dirt, and other things foret the said officer from his duty...' (fn. 13) It is not surprising that the tax yield was low. In September 1696 the annual yield was estimated at £50,000. (fn. 14) The shortfall in general was acknowledged in the Deficiency Act of 1698 (8 & 9 Wm. III, c. 20). (fn. 15) And the final Act of Indemnity (4 & 5 Anne, c. 23) recognised that the clergy had not always kept accurate accounts of the vital events in their parishes — thus depriving the collectors of a reliable check on their own lists of persons from whom taxes were due.

The third point concerns the statistical surveys produced as a result of the 1694 Act. The mystery occurs here because so few of the returns appear to have survived — or, at least, to have been located and identified. Waters, in keeping with his views on the general antagonism to the Act, wrote — again without citing his evidence — that it was 'not to be wondered at if these lists of inhabitants were generally destroyed as soon as the Taxation Act expired, and the few which have been preserved are commonly found in the hands of laymen'. (fn. 16) Since Waters's time more returns have been located, mostly in county archives. The returns for London still constitute the largest set so far located; they cover eighty of the ninety-seven parishes within the Walls and thirteen parishes without the Walls. (fn. 17) But additional returns have been found for the following places: Derbyshire, township of Darley; Devon, fragments of a parish near Clyst St Mary (unconfirmed); Dorset, Lyme Regis; Kent, New Romney and Wingham division; Leicester, Rothley parish and three parishes of Leicester city; Lincolnshire, wapentake of Candleshoe; Shropshire, Shrewsbury; Southampton; Somerset, Bristol; Staffordshire, Bilston; Warwickshire, Fenny Compton; Westminster, St Margaret's parish (incomplete); Westmorland (group of parishes); and Wiltshire (group of parishes). (fn. 18) The number is small — remarkably so when it is remembered that several copies of the assessments were required, including copies for the central Government. Moreover, the assessments were supposed to be revised annually, thus increasing the number of copies. But even fewer of the revised assessments have been found. (fn. 19) For London, for example, the Guildhall collection contains no revised assessments; the only examples known were located by Mr P. E. Jones among the Additional Manuscripts in the British Museum — a copy in the handwriting of Gregory King of the 1696 assessments for the parishes of St Benet and St Peter, Paul's Wharf. (fn. 20) Yet though the total number of returns extant is small, their distribution over the country suggests that the assessments were, as the Act required, undertaken on a national scale, and it is by no means improbable that further copies will be found in future, especially because there is now a more widespread and informed interest in the demographic history of England. (fn. 21) It is, however, unlikely that any remains will be found of the special register for the nobility and the gentry prescribed by the Act. Inquiries at the College of Arms have revealed no trace and it is, in fact, possible that the register was never actually compiled. The heralds were given no incentive, for the compilation was supposed to be done without payment — scarcely encouraging to individuals who, if Gregory King's experience was at all typical, were heavily dependent upon outside fees to supplement their small official income.

One further question may also be raised here, namely that concerning the part played by Gregory King, both in relation to the 1694 Act itself and to the analysis of the statistics which it furnished. The question arises because of the wide use made of those statistics by King in his Observations and in his various manuscript journals and other notes, and also because his sole publication in the field of political arithmetic was a broadsheet which summarises the various rates of tax imposed by the Act. (fn. 22) However, there is no evidence which links him with the drafting or initiation of the Act. His only clear association lies in his appointment as one of the Commissioners for the City of London, but that is unlikely to be significant, since more than 300 Commissioners were appointed and it cannot have been too easy to find suitable people. (fn. 23) In any case, the 1694 Act simply took over the list of Commissioners who had been appointed under 6 & 7 Wm. & M., c. 3, an Act relating to the duties of tonnage and poundage. In the analysis of the resultant statistics, however, King's link was obviously much closer. Indeed, without access to those statistics — which were not published — his demographic studies could not have been carried out. Nevertheless, as has been shown elsewhere, King had only a very limited access to the statistics, and even for the City of London his plan for collecting data for every parish was not realised. Though he prepared a table covering all the parishes, the details were supplied for only thirteen of the ninety-seven parishes within the Walls, for five parishes without the Walls, and for seven parishes beyond, in the eastern districts of London. (fn. 24) Neither his data for London nor his collection of statistics for other towns and villages would suggest that he was empowered to demand information, but rather that he had useful friends and acquaintances in various parts of the country. Access to the assessments for the London parishes may have been obtained in that way, though his appointment as a Commissioner must presumably have been of some assistance. A purely personal element may also have been involved in the case of the assessments for the two London parishes, of which he himself made detailed copies, for one of them — St Benet, Paul's Wharf — is the parish in which the Herald's Office was situated, and he and his wife and their three servants appear in the enumeration. (fn. 25) In sum, therefore, though King made most effective use of the data he was able to obtain as a result of the 1694 Act, his association with that Act, except as one of very many Commissioners, remains unproven.

As for the statistics in general, those now in existence are both variable and imperfect. Quite substantial differences are found between the form and detail of the assessments in different parts of the country. In one case — Bristol — forms with printed headings were used. In another — New Romney — the nobility and gentry were listed first, as if in order of status, with the remainder of the population following. In some cases a clear separation of different houses is visible, one house being divided from another by a line drawn on the assessment. In other cases, the names follow each other without a break; families can still be distinguished, but not houses. The assessments for London are in general the most informative and complete and the houses, which are nearly always distinguishable, appear to be arranged in some street order. Two assessments actually note street names (nos. 17 and 57). Several give substantial information on occupations (e.g., nos. 15, 23, 39, 42, 45, 61, 66, 70, 71), that for St Mary le Bow (no. 62) listing the occupation of almost every inhabitant. Servants are almost invariably shown, and usually lodgers, too, while bachelors and widowers are shown separately in most assessments. The social status of individuals — that is, status defined in terms of the graduated taxes and duties — is specified and, in contrast to New Romney, those who were liable for surtax were not grouped together but enumerated as they occurred. The assessments took longer to prepare than was envisaged by the statutory requirement (not surprisingly, since not more than fourteen days were allowed) and the actual period of enumeration was not identical for the different parishes. (fn. 26) But underneath the variations there are strong common features, and this applies, though in somewhat lesser degree, to the returns for other parts of the country. That there were omissions is clear, and some of those for London will be examined later. Nevertheless, the impression left on studying the material as a whole is that no other demographic data of comparable detail and quality were collected by the government until well into the nineteenth century. (fn. 27)

II. The assessments for the City of London

A general description of the assessments for the City of London (within the Walls) has already been given. (fn. 28) The purpose of this part of the introduction is to comment on the ways in which the basic material can be used, to present a few examples of demographic analyses, and to examine some of the limitations of the original data.

The index itself is, of course, primarily a 'search' instrument. In essence it is a directory, covering eighty of the ninety-seven parishes within the Walls, giving the name (and the sex can almost invariably be inferred) of every person enumerated in 1695. In addition, the index records whether the individual was married, widowed, or unmarried and, if an unmarried man, whether he was a bachelor above the age of 25 years. Children are sometimes so designated and usually servants and apprentices. Because the 1694 Act provided for graduated taxes, the index records social status, at least in the sense that it shows whether a surtax was paid and in respect of which of the middle-or upper-class categories designated in Gregory King's broadsheet (Tables 1a and b). The index is so arranged as to show families. The wife and children of a man are listed next to his name. Thus, Joshua Bagshaw is shown as having a wife and two unmarried daughters. John Baber, with a wife and a son and daughter, is also shown to have had personal estate worth not less than £600. Households as such are not shown, save implicitly when they consist only of a nuclear family of parents (or parent) and children. How much more detail is reported depends upon the particular assessment in which the individuals originally appeared. Most of the surveyors did not go beyond the formal requirements of the Act, but in several cases occupations were ascertained. The assessment which is most complete in that respect is for St Mary le Bow, and an extract is shown in Table 2. The occupation of almost every relevant person is given in that assessment, and the schedule of duties and taxes states what sums would be due if the specified event were to occur. Husbands were responsible for payment of the tax on births if their wives bore children; bridegrooms were responsible for the marriage tax; surviving relatives were liable for the burial tax. (fn. 29) The extract reveals that the attorney at law was classed as a person with a personal estate of at least £600 or real estate valued at not less than £50 per year. He was not a Doctor of Law, or he would have been liable to a higher rate of tax, and he was neither a gentleman nor 'one so reputed'. No bachelor above the age of 25 years is recorded, and no childless widower. The entry for Edward Harris suggests that he was under 25, though of marriageable age, whereas Richard Peters was a child. But though that might be a reasonable inference here, it would not necessarily be equally reasonable in respect of other assessments. In some assessments all unmarried individuals were shown as liable to the marriage tax, whatever their age, and some so listed must have been children. Bachelors above the age of 25, however, appear to have been ascertained and are recorded as such. The term 'servant' requires some caution in interpretation. Often a more specific term is used, such as maidservant or footboy or nurse, and the domestic roles are then confirmed. But 'servant' as such may mean no more than that the head of the household stood in loco parentis; and the term might cover apprentices or assistants as well as servants in the modern sense. (fn. 30)

Used as a directory, the index is of unique service to historians of seventeenth-century London, providing a check on the location and status of almost 60,000 individuals. The value of the list would be still greater when used in combination with other contemporary materials, such as the London returns under the 1692 poll tax, which constitute a roster of the occupations of the London population. Matching is difficult, for the poll-tax returns are on a ward basis. But the task is not impossible, and the comparison of the occupations in the parish of St Mary le Bow with those for the appropriate precinct in the poll-tax return suggests a very close similarity in occupational description when the same individual can be identified in both documents. It also suggests that the poll-tax return records actual occupations and not simply formal guild designations. For many purposes the index will need to be used in conjunction with the original assessments under the 1694 Act. And for demographers and social historians it is the territorial and household groupings of individuals in those assessments which would be of the most immediate interest, and understandably so, since it is in such a form that the material would throw most light upon the structure of the largest urban community in seventeenth-century Britain.

The material was indeed employed in that way in the classic paper by P. E. Jones and A. V. Judges. (fn. 31) Its analysis provided a new estimate of the total population of the City of London in 1695 (almost 70,000 for the ninety-seven parishes within the Walls and almost 54,000 for the thirteen parishes without the Walls) and a test of the completeness of parochial registration in that community. The comparison of the 1695 population with that enumerated at the 1801 census suggests a relatively early centrifugal movement of population from the urban centre — a movement which became characteristic of urban communities during the nineteenth century. (fn. 32) The use of the material for the individual parishes is still more revealing. Leaving aside the estimated birth- and death-rates (for reasons which will be discussed later), particular interest lies in the relative 'social status' of the different parts of the City, indicated by calculating the proportion of 'substantial householders' in each parish. The results of some of the calculations are given in Table 3. 'Substantial householders' are defined as those liable to surtax under the 1694 Act — that is, individuals with a personal estate worth not less than £600 or real estate worth not less than £50 per year, together with certain strictly named categories of persons — as shown in Table 1a and b. The category so defined must have been very much the upper part of society in respect of income and status; for the eighty parishes, the proportion amounted to approximately 27 per cent. (fn. 33) Table 3 gives the results only for two groups of parishes — those with the highest and those with the lowest proportions of 'substantial householders', but the degree of differentiation within the whole group of eighty parishes is displayed in the accompanying maps. The table and maps taken together give a picture of quite considerable social articulation. Thus there are three areas with high proportions of substantial households — a major area (covering five contiguous parishes, namely, Allhallows, Bread Street; St John the Evangelist; St Martin, Ironmonger Lane; St Mary le Bow; St Matthew, Friday Street), and two secondary areas. Between the latter two areas is a band of apparently lower status. But generally the proportions of substantial households are higher near the centre and decline towards the periphery, being especially low in the parishes outside the Walls (with the exception of St Dunstan in the west). This kind of gradient is in contrast to that found in the modern city, in which the centrifugal movement of population has occurred particularly among the middle classes. The 1695 pattern is more characteristic of the pre-industrial city in Europe and also of the Latin American cities built or taken over by the Spanish conquistadors. The pattern broke down under the impact of industrialisation and with improved urban transport. In London the older pattern was still visible in 1831, though some outside areas (St Martin-in-the-Fields and Hanover Square, Westminster) were showing high values of real property per head of the population. By 1851 considerable changes had already occurred, the districts with the highest indices of socio-economic status being St James, Westminster; Hanover Square; Hampstead; Kensington; St Martin-in-the-Fields; and the Strand. (fn. 34)

It is, of course, true that the situation is complicated by the wide variation in the total populations of the individual parish units — from 152 people in St John the Evangelist (with high social status) to 2,782 in St Anne, Blackfriars (with low social status). But even so, it is of interest that in St John some forty-eight of the eighty households were 'substantial', whereas even with the very much larger population in St Anne there were only about four such households out of a total of 373. A relationship between social status and population per house is also suggested by the run of the data in Table 3. The areas with higher proportions of 'substantial householders' appear to have very slightly larger ratios. The relationship is far from close, but it appears more evident if St Anne, Blackfriars, is removed from the lower half of the table. And to do so may not be entirely unjustifiable, for that parish contains a number of extremely densely peopled houses — eighteen containing fifteen or more persons, out of a total of 380 households. An examination of the assessments for some of these very large groups shows that they contain several households. (fn. 35) They may be semi-institutional, but it has not so far been possible to confirm that from other contemporary information. A slightly different way of looking at social status variations will be mentioned subsequently.

In their analysis, Jones and Judges were concerned with certain general features of London structure and with some comparisons with the estimates of Gregory King. But the assessments also permit a more detailed study of the internal structure of the parishes, and the remainder of this section will concentrate mainly upon the results of such a study. The work involved is rather tedious and there was not sufficient time to cover more than a small sample of parishes. Later, it is hoped, the analysis will be extended to the eighty parishes for which assessments are available; but for the movement the discussion will be largely restricted to the results obtained for only ten parishes. These ten were not chosen to be truly representative of the total group, but rather to provide a comparison between parishes with differing proportions of 'substantial householders'. This criterion was used because it seemed possible that — as is often the case in present-day urban communities — a relationship between social status and other characteristics might be found. Hence of the ten parishes, three were chosen from the group with the lowest proportion of 'substantial households', three from the middle range, and four from the highest (with 60 per cent or more of such households). The information extracted for these parishes is summarised in Tables 4 and 5.

One point may be seen immediately in Table 4, namely the use of a slightly different index of 'social status' from that in the study by Jones and Judges. The index in column (3) in the table is based not only on householders but on all persons (householders, their kin and other relevant household members) liable to surtax. This index actually yields a narrower absolute range of difference, but one which is perhaps somewhat more realistic, for it takes into account the fact that though surtax households may be large, a sizeable element will consist of servants not liable to higher rates of tax. And it is clear from column (6) that, in general, parishes with higher proportions of surtax payers have more servants per household, whatever the meaning of servant may be. That the data on lodgers are poor may be seen from column (8); in very few of the parishes was there sufficient detail to present a consistent listing of lodgers. Before considering more closely the nature and variation of some of the indices and ratios in Table 4, it is desirable to see what evidence can be found in respect of the general reliability of the results.

It must be emphasised that there is little in the way of direct and incontrovertible evidence. Jones and Judges scrutinised the material and, where possible, compared it with assessments under the 1666 hearth tax, the 1673 eighteen months' tax and the 1678 poll. They concluded that 'the 1695 assessment was, almost throughout the City, conducted with more diligence and with fuller results than was usual in the period.' (fn. 36) This evaluation must carry considerable weight, though it would be useful to have it in quantitative form — for example, in terms of a comparison of the numbers of households noted in connection with each of the taxes. But so far as a direct check is concerned, only two examples have been found, namely those in Gregory King's copy (it is dated 1 May 1696) of the 1696 assessment for St Benet, Paul's Wharf, and St Peter, Paul's Wharf. In both cases there appears to have been something comparable to a 'post-enumeration survey', undertaken perhaps by King himself or at his instigation. Noted on his copy, in the margins and at the end, are the names of individuals or families who he believed might have been omitted from the original count. King himself attempted to assess the accuracy of the enumeration, but his estimates contain two errors. First, his totals were not quite correct (that is, accepting the data given in his copy); and secondly, his method of correcting the enumeration was based upon assumptions regarding the relative death-rate in the two parishes. (fn. 37) The specific indications of omissions and of 'double-counting' are summarised, together with the uncorrected assessment material, in Table 6. The data suggest a net undercount of about 5 per cent of persons in St Benet and of about 9 per cent in St Peter, Paul's Wharf. King's final 'adjusted' figures (total populations of 845 and 390 respectively) would yield an omission rate (in relation to the final population) of 20 and 7 per cent respectively, or about 15 per cent for the two parishes combined. To compare this with the allowance which King made in his Observations for London as a whole, (fn. 38) the omissions have to be expressed as a percentage of the initial enumeration total. The rate for the two parishes would then be almost 19 per cent, in contrast to the 10 per cent which King finally suggested for the area of London and the Bills of Mortality. The equivalent rate, based only on the initial indications of under – and overenumeration in King's copy of the assess – ments, would be about 7 per cent. King's ultimate estimate thus lies between the two extremes and is perhaps not too wild a guess at the overall situation. Equally, a rate of 10 per cent on the initial figure is far from being a 'wildly excessive' one for the first detailed enumeration conducted in this country. The 1801 census was certainly imperfect, and even now no ad hoc census is likely to achieve complete accuracy. Nevertheless, the omissions need to be taken into account and it would be worth making an attempt to draw upon other contemporary materials in testing the degree of incompleteness both in London and elsewhere.

To return to the assessments, one of the questions upon which they are informative is 'household structure'. But the information is not unequivocal, and various questions arise, to which definitive answers cannot be given at present. In the Jones and Judges summary for the eighty parishes within the Walls, the relevant statistics are described as relating to 'inhabited houses' and the listings in the assessments certainly appear to confirm this. (fn. 39) It is within the houses that the problems lie, in respect of the number, size and composition of households. For in present-day classification, a household is defined as consisting of individuals who share a common board, whether or not they are kin. A lodger who provides his own food and does not eat at the common table is treated as a 'single-person household', and this is justified, especially in relation to the estimation of housing needs. But the assessments naturally given no indication as to what constitutes a 'lodger'; he may be a boarder, but he may equally be a self-contained unit, of the kind often referred to in Victorian novels in accounts of middle-class bachelors. Further, servants are sometimes noted in association with a particular family group with a house. Not infrequently, however, they are simply put together in a group at the end of the listing of the inhabitants of a house, and it is then impossible to say how many belong to each of which household units within the house. 'Parish children' are sometimes treated in the same way, and 'inmates', too. And in the case of 'inmates' or 'poor women', they may presumably have been either lodgers or boarders. Analysis of household structure thus involves a certain amount of guessing — in some cases, a good deal. Nevertheless, some tentative results are of interest, and a few examples will be given.

It is easier to distinguish the initial nuclear household in a house than accurately to separate the other sub-units. To show what happens when this relatively simple analysis is done, the first fifty houses in the assessment for St Mary le Bow were examined. On the Jones and Judges list almost 54 per cent of the householders were 'substantial', and the assessment shows a good number of merchants — linen drapers, mercers, drugsters, bone-sellers, and unspecified merchants — as well as apothecaries, cutlers, corn-chandlers and professional people, such as attorneys, and also a 'surgeon of His Majesty's household', Thomas Gardiner. (fn. 40) In the fifty houses the total population amounted to 323, the numbers per house ranging from one to fourteen and the average number being 6.46. Taking the nuclear households only, the upper limit of size is reduced to nine and the average falls to 4.58 persons per household. The population in the sub-units is ninety-four — almost a third of the total — and contributes an average of 1.88 persons per house. There are some genuinely large households. For example, the household of nine persons consists of John Smith, a victualler, his wife, two sons, an apprentice, two boys (with unspecified roles), and two maidservants. There is, in addition, an 'inmate', Mary Tompson, and the size of the household would be ten if she boarded with John Smith's family. On the other hand, some of the houses contain several units, or what appear to be such. The house with the largest number of inhabitants — fourteen — is an example. The nuclear household appears to consist of John Bakewell, a book-keeper, his wife and three daughters. There is a maidservant in the house; it is not clear to whom she 'belongs', but it seems rather as if she is attached to the second unit, consisting of Joseph Lawrenson, a beltmaker, his wife, a son and an apprentice. The third unit contains James Rawlins, gentleman and widower, together with a nurse and her daughter — though the latter two may perhaps have been a separate unit. Finally, there is a pensioner, Elizabeth Howard, whose attachment to any unit is not specified. Thus the house may have contained three, four or five units, and the mean household size would have varied accordingly. The still more complicated case of the house in St Anne, Blackfriars, has already been mentioned. Containing twenty-seven people, it appears to consist of seven units, the largest being that of one of the 'lodgers' — a husband and wife and four children. (fn. 41)

In King's copy of the 1696 assessment for St Benet, Paul's Wharf, the apparent division into households greatly reduces the complexity of the structure. So far as some of the wealthier households are concerned, the division may have been prompted by the nature of these dwellings. In the page reproduced from the assessment (Plate II), the households in the Herald's Office have a simple structure — presumably because the families were listed as they were living, in separate apartments. This is also the case of the households in Doctor's Commons, recorded in the same assessment. Various legal officials were originally lodged there — including the Judge of the High Court of Admiralty, the Dean of the Arches, the Commissioners Delegate, the Vicar-General and the Master of the Prerogative Court of Canterbury. (fn. 42) When it was rebuilt after the Great Fire the building appears to have been used in a similar way and in 1696 it contained the following households:

The households, which appear to be relatively simple nuclear units, range in size from one to ten persons, with an average of 4.2 (excluding the single- person households, the average is five).

The frequency of the nuclear, two-generation household has recently been stressed by Peter Laslett as constituting the reality of seventeenth-century England, in contrast to romantic notions regarding the role of the extended family, whether of three generations or of parallel kin (several adult brothers and sisters, for example). (fn. 43) A certain amount of information on this question can be obtained from the assessments for London, though by no means as much as would be required to give a definitive answer. Two types of information can, however, be provided. First, there is the kind summarised in Table 8, which shows the cumulative frequency of the numbers of persons per house. In St Anne, Blackfriars, for example, 223 of the total of 380 houses had not more than seven persons per house. Houses with not more than six persons constituted almost half of all the houses in the parish. These figures relate to houses, and not to households; even so, they suggest that relatively small households must have been fairly common. Secondly, it is possible to estimate the frequency of nuclear households — even without attempting to split up the multiple households which are found in many houses. It will be assumed that wherever kin is specified in a house (other than in a direct two- generation relationship), there is some form of extended household, even if the kin were living in entirely separate and self-contained accommodation within the house. This would yield a low estimate of the frequency of nuclear households and would compensate for cases in which there were kin, though this was not evident from either the description or the surnames. (fn. 44) A degree of guesswork is still inevitable, but not a high degree, for the analysis is concerned only with the composition of the initial (main) household in a house and its relationship to other individuals in the same house, not with the number or structure of the other household units. The broad results of such an analysis for three parishes are summarised in Table 9, in which the nuclear household is defined as consisting of two generations, headed by both parents or by one parent, and regardless of the rest of the population in the house, provided only that it does not contain individuals who are related in some other degree (such as brothers, sisters, aunts, uncles) to the nuclear household. There is no doubt as to the dominance of that type of household in all three parishes, even though they cover the whole social status range (as may be seen in Table 4). It might be expected that in St Matthew, Friday Street, at the top of the status scale, households with other kin in addition to the two-generation kind would be more important than in either St Anne or St Mildred, and the run of the figures certainly suggests this. But the numbers of houses in St Matthew are too small for the results to be significant in any realistic sense. Apart from this kind of variation between parishes, the general dominance of the two-generation household (or the relative absence of three-generation households) is what would be expected from the probable demographic situation in the period. For given high mortality and a not very low age at marriage, together with a high fertility, three-generation households could scarcely constitute a very large proportion of all households, even supposing that other types of kin or non-kin households did not exist. Relatively few parents would live sufficiently long to become grandparents when the expectation of life at birth was only around twenty-five or thirty years. Mortality would have had to fall quite substantially before the ratio of surviving parents to their adult (and married) children became much more than a third. Theoretically, the possibility of three-generation households would be very much higher when, with low mortality, fertility had also fallen very considerably. But in societies with that combination of fertility and mortality, social and economic factors have intervened between demographic possibility and household reality. (fn. 45)

One point not so far discussed is the general population structure as it appears from the analysis of the assessment for the ten parishes. Unfortunately, relatively little can be said, for as Table 4 shows, the information collected under the 1694 Act does not permit any analysis of age structure. Even for males, it is not the division into 'under 25 years' and '25 years and over' which is specified, but only that division for the single male population. Marital condition is shown, and there appears to be a fairly systematic relationship between the proportion of 'never-married' persons (that is, single) and the social status of the parish; parishes with the largest proportions of 'substantial households' have the higher percentages of single men and women. This in turn is linked to the higher frequency of servants (and apprentices) in such parishes, for those designated as servants in the assessments are almost all single. The parish of St Mary le Bow may be taken as an example. According to Jones and Judges, its total population was 669 and almost 54 per cent of its households were substantial; it was thus fairly high up in the social-status scale. The assessment for the parish lists 144 servants (including eighteen male), of whom only two were married (both men). Similarly, seventy apprentices were listed (including two girls, apprenticed to a barber), all of whom were unmarried. There were thus eighty-six males and 128 females, only two of the total being married. Apprentices and servants together constituted almost 32 per cent of the total population of the parish, a very sizeable fraction. (fn. 46) Clearly, this affected the marital structure. Nevertheless, even the existence of large numbers of unmarried young workers did not appear to produce an 'abnormal' distribution in respect of marital condition, that is, not very different from that in London in the nineteenth century. Since the age structure is not given, it is impossible to select a strictly comparable population or to apply marriage rates to each age group to arrive at total numbers of married persons. But a perhaps not too grossly inappropriate comparison is attempted in the bottom row of Table 5, using the marital structure of the population of London in 1851 as the yardstick — chosen because both fertility and mortality were high at that time. (fn. 47) The proportions of single, married and widowed for London at that time appear to be at about the midway point in the sample of ten parishes — not startlingly different from the situation at the end of the seventeenth century, and generally in support of the view that the age at marriage then was already at a 'modern' level.

One further question will be considered in this introduction, namely the reliability of the birth- and burial-rates as derived from the Collectors' returns under the 1694 Act. Such rates — expressed in terms of births or burials per thousand total population — are given by Jones and Judges for each of the eighty parishes. For the sample of ten parishes examined here it would be possible to calculate marital fertility rates, by relating births to the numbers of women recorded as married in the assessments. There are, however, two reasons for not doing so at present. First, the populations in the ten parishes are small and the rates would therefore not be very meaningful. It would be more satisfactory to wait until the data for a much larger number of parishes are available. Secondly, the estimated numbers of births and burials used by Jones and Judges — derived from the Collectors' annual reurns — are subject to defects, some of which can be corrected by a more detailed study of the material. An indication of how this might be done, and with what results, may be obtained by using two parishes as examples: St Matthew, Friday Street, and St Mildred, Poultry.

In their study Jones and Judges compare the births and burials reported by the collectors under the 1694 Act with the christenings and burials published in the Bills of Mortality, of which John Graunt made such magnificent use. The comparison shows that the collectors report much larger numbers of births than the Bills of Mortality; but that the Bills report rather more burials than the collectors. Jones and Judges were not convinced that either source was fully reliable — and least of all the parish clerks' returns of christenings. The data for the two parishes referred to suggest that no single source is satisfactory, and that by using two sources a closer approximation to reality can be obtained.

The method used in examining the records was to work from the collectors' returns to the parish registers, (fn. 48) and similarly to work from the parish registers to the collectors' returns. Both records list names (and occasionally some other characteristics) and events and individuals in each record were 'matched' as far as the information permitted. Some doubtful cases remain, but in general this 'matching' could be done quite effectively. For each parish, and for births and burials separately, Table 10 shows the numbers of events reported in both sets of records; in the registers but not in the returns; and in the returns but not in the registers. It would not be unreasonable to argue that, at the moment, a much better approximation to the 'true' numbers would be obtained by taking the total of the three categories. But some allowance should also be made for the births and burials which completely escaped recording — that is, which were not reported in either the parish registers or in the collectors' returns. Since there is no independent evidence on this — at least at present — the allowance has to be made on a probability basis, assuming that the chances of escaping entry in the different sets of records are independent of each other. The allowance is shown in Table 10 under the rubric of 'double exclusions'. (fn. 49) The final estimates of 'true' numbers of births and burials are used in producing the ratios in Table 11 — ratios indicating the factor by which the numbers recorded in the parish registers or in the collectors' returns would have to be multiplied to allow for 'under-registration'. The basic numbers are, of course, extremely small and no reliance can be placed upon the precise levels of the results given in the table. 'Nevertheless, the ratios appear to confirm, for both parishes, the differential incompleteness of the two sets of records as between births and burials. The collectors' returns are more comprehensive than the parish registers so far as births are concerned, but very much less comprehensive in respect of burials. This may be the result of the provision that, though births to people receiving alms were exempt from tax (an entry in the collectors' returns would thus be desirable), tax had to be paid on all burials, the parish having to pay the tax for people receiving alms (the parish register would thus be the primary source of information on burials). How far the under-reporting suggested in Tables 10 and 11 applied to other parishes cannot be ascertained until the records for those parishes have been examined. And it would be worth extending the study to such other areas in the country as also have similar records, for it may then be possible to obtain a useful measure of the completeness of parish registration in England at the end of the seventeenth century.

A fuller exploration of the data for London itself would in any case be justifiable, for there is a wealth of ancillary materials. The use of these, together with the assessments, would yield a detailed picture of the social structure of a major urban community a generation or two before the beginnings of the new industrial era. The present introduction is only a preface to that further work — to be followed, it is hoped, by more substantial publications in the future.

III. Description of the assessments for the City of London (within the Walls) (fn. 50)

The eighty surviving assessments vary greatly in size (the largest, no. 77, measuring 19 1/2 × 14 1/2 in., and the smallest, no. 13, measuring 8 × 11 1/2 in.), and in the number of pages (no. 40 with three folios written on one side only and no. 17 with 131 pages). Each assessment is separately bound in a modern semi-stiff buckram cover. Some are stitched along the left-hand edge, others at the head. A number have recently been repaired. Apart from the assessments which are lacking altogether (nos. 10 and 82 – 97) (fn. 51) three are imperfect: no. 78 lacks f. 1 ; in nos. 37 and 41 the lower part of the last (?) folio has been cut off (f. 21 and f. 10 respectively), but the missing portions probably contained only the signatures of the Assessors and Commissioners and a note of the collectors appointed.

Who actually wrote the copies of the surviving assessments is not known, although the 1694 Act provided for payment to Commissioners' clerks for this work (Section XIV). It is, perhaps, worth nothing that several assessments are written in the same hand (e.g. nos. 3 and 5; nos, 4 and 24; nos. 6, 31, 36, and 56; nos. 14 and 43; nos. 15, 23, 35, and 66; nos. 18 and 60; nos. 25, 34, 46, 51, 52, and 64; nos. 27, 55, 75; nos. 44 and 65; nos. 57 and 78; nos. 67 and 76). The assessments are all signed (except nos. 37 and 41, which are imperfect) by the Assessors, usually, but not always inhabitants of the parish, and generally by three Commissioners as laid down by the Act.

The assessments purport to list the inhabitants of each parish on 1 May 1695. However, the Act left the surveyors little time for their work. The Commissioners were instructed by the Act (Section XI) to meet on or before 30 April 1695. The Assessors then appointed were to appear before the Commissioners within ten days so that the latter could explain how the assessments were to be made. The assessments were to be returned to the Commissioners before 13 May (Section XII), whereupon three Commissioners would issue warrants for collection. With so little time to carry out this work, it is not surprising to find that only one parish appears to have been assessed within the prescribed period (no. 32). Most of the assessments appear to have been completed later in May, June or July, while one (no. 80) was not signed by the Commissioners until 2 August. How far a serious attempt was made to ascertain the exact population of the parish on 1 May 1695 as opposed to the population at the date on which the assessments were made, is difficult to say. The preambles of most of the assessments state that they refer to the inhabitants on 1 May and in several instances persons are entered as deceased since that date, while in two parishes births and burials between 1 May and the date of making the assessment have been separately noted (nos. 11 and 59).