John Graunt F.R.S. (1620-74): The founding father of human demography,epidemiology and vital statistics

Abstract

John Graunt, a largely self-educated London draper, can plausibly be regarded as the founding father of demography, epidemiology and vital statistics. In his only publication, based on a pioneering analysis of the London Bills of Mortality, he replaced guesswork with reasoned estimates of population sizes and the first accurate information on male:female ratios. He quantified the extent of immigration from countryside to city and his demonstration of the ‘dying out’ of a cohort paved the way for life table analysis. His comparison of London data with rural data provided the first recognition of the ‘urban penalty’. His use of the first known tabular aggregates of health data clarified distinctions between acute diseases, which were often epidemic, and chronic illnesses which were often endemic. He quantified the high infant mortality and attempted the calculation of a case fatality rate during an epidemic of fever. He was the first to document the phenomenon of ‘excess deaths’ during epidemics. He provided a template for numerical analysis of demographic and health data and initiated the concepts of statistical association, statistical inference and population sampling. By making a novel concept intelligible to a broad audience he influenced the thinking of doctors, demographers and mathematicians.

Biographical background

John Graunt was born in Birchin Lane, Cornhill, in the City of London to Henry Graunt (1592–1662) and his wife Mary in 1620. John was probably the eldest of at least seven children. His father came from Hampshire but had been apprenticed in London and was admitted to the Freedom of the Worshipful Company of Drapers in 1614.^1–4 The early Companies or Guilds of the City of London acted as mutual protection societies for their members but later the members gained financially valuable trading rights. The Drapers’ Company came to regulate and control the sale of cloth in London, and dealers were obliged to sell cloth only to Freemen of the Company. ⁵ Henry Graunt was therefore a man of some substance and importance by the time of John's birth. At the age of sixteen John was apprenticed to his father and was admitted by patrimony to the Freedom of the Drapers’ Company when twenty-one and granted the Livery when thirty eight. Like his father before him he held many civic offices in Cornhill Ward. He was a member of the Common Council of the City of London, a member of the New River Company (actually a canal to provide piped water from Hertfordshire to the City) and an officer in one of the “Trained Bands”, four regiments of which acted as a militia for London from 1572–1647. ⁶ Graunt was therefore a public spirited member of his community.

Little is known about John's own family. He married Mary Scott when he was twenty-one and she was seventeen. According to the biographer, antiquary and natural philosopher John Aubrey (1626–97) they had two children, a son who died in Persia and a daughter who became a nun in Belgium. ¹ The deaths of two other children are recorded in parish registers. ²

Aubrey describes Graunt as “my honoured and worthy friend” and as “A man generally beloved; a faythfull friend. Often chosen for his prudence and justness to be an arbitrator; and he was a great peacemaker. He had an excellent working head and was very facetious and fluent in his conversation.” ¹ According to the Oxford English Dictionary, at that time ‘facetious’ meant polished, agreeable and given to pleasantry. Aubrey also describes him as “a very ingeniose [sic] and studious person” who “rose very early in the morning to his study before shop-time. He understood Latin and French.” He also “wrote short-hand dextrously.” ¹ The lawyer and bibliophile Richard Smyth (1590–1675) described Graunt, whose name he spelt as Grant, as “an understanding man, of a quick wit and a pretty scholar, my old acquaintance.” ⁷ Graunt's character and learning took him into circles well beyond those of most successful tradesmen. Aside from Aubrey, the diarist Samuel Pepys (1633–1703) described his collection of prints as among the best he had ever seen “there being the prints of most of the greatest houses, churches, and antiquitys in Italy and France”. ⁸

He was also a friend of the portrait painters John Hayls (also Hales, 1600–79) and Samuel Cooper (1609–72), and the poet Benjamin Rudyerd (1572–1658), for whom Graunt funded a memorial in West Woodhay Church in Berkshire. In spite of Graunt's friendships with artists, no authenticated images of him are known.

An important friendship was with the physician, natural philosopher and administrator Sir William Petty (1623–87) M.D. Oxon (1650), F.R.C.P. Lond (1658) F.R.S. (1662). Graunt was instrumental in obtaining a professorship in music for Petty at Gresham College in 1651. ² In 1659 Graunt, Petty and John Martyn (c.1619–80) purchased land in Lothbury, an area which like Graunt's house in Cornhill, was destroyed in the Great Fire of 1666. ⁹ In 1662 Martyn published the first edition of Graunt's Natural and Political Observations, publishing all subsequent editions under the aegis of the Royal Society to which he became printer in 1663. ¹⁰ Martyn also published works by Thomas Willis, Marcello Malpighi, Robert Hooke, John Evelyn and other notable authors making him perhaps the most important scientific publisher of the era. ¹¹

It is evident that by no later than 1660, and probably earlier, Graunt had formed important contacts and friendships with people in trade, the arts and the newly developing sciences. Indeed Pelling sees him as a member, albeit a peripheral one, of the Hartlib Circle, the informal discussion network centred around Samuel Hartlib (c.1600–62) which sought to promote improvements in education, science and medicine and with which Petty and other members of the Royal Society were connected. ¹² Through Petty's involvement with the Experimental Philosophy Club which initially met in his rooms at Buckley Hall in Oxford from 1649. ¹³ Graunt may have met men like Robert Boyle, Christopher Wren and Robert Hooke even before he started work on his Observations. That Graunt was the son of a haberdasher, Petty the son of a clothier and Martyn the son of a tailor, all of them from what might now be regarded as non-establishment backgrounds, is an indication of the potential for social mobility at that time.

Graunt's natural and political observations … made upon the bills of mortality ¹⁰

Graunt's legacy rests on his only published work which was based on a systematic study of London's Bills of Mortality. ¹⁴ These had been published sporadically from at least 1532, then more regularly from 1592 and weekly from 1604. Causes of death other than plague were recorded from about 1604 but only appeared in the printed versions from 1629. Numbers of baptisms were recorded from the mid-1560s. The Bills were published weekly from 1603 but the sex of individuals was not included until 1629 and the age only from 1728. An annual summary of the weekly Bills was published from 1604.

Graunt observed how most of those who subscribed to the Bills looked only at changes in the numbers of deaths, especially in time of plague, so that those who could afford to leave the city might do so and merchants might consider the effects on their businesses. However Graunt realised, perhaps following discussions with Petty, ¹ that much additional information might be deduced.

Before describing how Graunt made use of the Bills it is first necessary to say something by way of context. Graunt's use of language is of course of its time but his publication contained many of the elements of a modern scientific paper. He described in some detail how he collected his information, how he analysed it and how he interpreted the results to arrive at his conclusions. In doing so he recognised the importance of a critical appraisal of the reliability of his information and he corrected for errors where this was possible. He formulated hypotheses to test his “Conceits [conceptions], Opinions, and Conjectures” and, where possible, he confirmed his findings by using different methods. He usually presented his findings in a way that allowed others to check his results and to make their own judgements. He actually invited criticism: “For herein I have, like a silly [in the earlier sense of ignorant] Schole-boy, coming to say my Lesson to the World (that Peevish, and Tetchie Master) brought a bundle of Rods wherewith to be whipt, for every mistake I have committed.” ¹⁰

His work differs from a modern scientific paper in several respects. For example, deficient information forces him to make many assumptions, the rationale for which is sometimes unexplained. Much of what would now constitute a separate discussion is interwoven with his findings. The discussion is sometimes overtly political or religious or both. Most important of all, Graunt's interpretation of his findings is that of a tradesman, albeit a very clever and well educated tradesman, in mid-seventeenth century England. This was a country that was still recovering from a civil war which had ended in regicide, a country which was then governed by an authoritarian military ruler but which had finally reverted to a still fragile monarchy only two years before Graunt's work was published. The political turmoil was overlaid by intense religious dissent and punctuated by repeated epidemics of bubonic plague for which there was no prospect of any effective prophylaxis or treatment. All of these events permeate Graunt's thinking.

Methodology

Graunt worked mainly from the annual summaries of 1604–60, sometimes also referring to weekly data. He excluded data from 1636–47 “as containing nothing Extraordinary and as not consistent with the Incapacity of the Sheet.” ¹⁰ To provide comparisons from a rural area he included information from Romsey (Hampshire) in the first edition. In later editions he added information from the London Bills of 1665, the year of the Great Plague, and from two more rural areas – Tiverton (Devon) and Cranbrook (Kent).^4,15–17 He thought it important to tell his readers exactly how the London Bills were compiled. Deaths were reported by the parish Sexton to the parish Searchers who were “antient [sic] Matrons, sworn to their office.” By viewing the body and “by other enquiries” the Searchers decided on a cause of death which they communicated to the parish Clerk. He then compiled an “accompt” [account] of all burials in the previous week which he forwarded to the Company of Parish Clerks which published a weekly compendium of results from all the parishes in the city. ¹⁰

Graunt was well aware of the system's imperfections and, like so many of those who have followed in his footsteps, he emphasised the necessity of both accuracy and consistency in collection of data. The Searchers were untrained and often uneducated, even illiterate. Illnesses might be given different assignations in different parishes. He also knew that, “after the mist of a Cup of Ale, and the bribe of a two-groat fee instead of one,” the poorly paid Searchers could be induced to ascribe a death to “ulcers” or “sores” rather than to the “French-Pox” (syphilis). ¹⁰ He recognised that christenings were under-reported, particularly during the republican Commonwealth of 1649–60, because only those christenings performed by a priest of the Church of England were recorded and not those of dissenting non-conformist Christians. He corrected for this by monitoring maternal mortality rates which, unlike stillbirths and abortions, were accurately recorded and which remained constant in proportion to population changes.

Having made a critical assessment of the available data including its accuracy, completeness and sources of error, Graunt took a pragmatic approach to many of its deficiencies. For example, if someone died suddenly, “the matter is not great, whether it be reported in the Bills, Suddenly, Apoplexie, or Planetstrucken, &c.” Likewise, if “a man of seventy five years old died of a Cough … I esteem it little errour (as to many of our purposes) if this person be … reckoned among the Aged and not placed under the title of Coughs.” ¹⁰ Many of the Searchers’ classifications were a matter of common sense which required no special knowledge or training (for example drowning, scalding, suicide, smallpox, dropsy, ague) and in some cases the Searchers were able to report the opinion of the attending doctor. ¹⁰ We may perhaps condone Graunt's expedient attitude on the grounds that, having assessed its deficiencies, he used the only data which was available to him.

How graunt analysed the bills

Graunt wrote that he was familiar with The history naturall and experimentall of life and death in which Francis Bacon (1561–1626) employed the tenets of natural philosophy to construct an observational study on how to live a long and healthy life. ¹⁸ Graunt's own methodology reveals that he had also studied Bacon's Novum Organum because he follows Bacon's advice that we cannot possibly analyse a mass of disparate data “unless we put its forces in due order and array, by means of proper and well arranged … living tables of discovery of these matters which are the subject of investigation, and the mind then apply itself to the ready prepared and digested aid which such tables afford.” ¹⁹ Graunt's data was numerical and he was probably the first person to use numerical tables to analyse data on health and disease. So original was his use of numerical tables that he thought it necessary to include instructions on how to read them.

Using Bacon's guidelines for classification and compilation Graunt reduced several large volumes into “a few perspicuous Tables”. He then applied the principles of a seventeenth century merchant's double book-keeping methods to interrogate his tables. To do this he used what he described as “the Mathematiques of my Shop-Arithmetique”, ¹⁰ the tools of which were: division which he used to compare ratios, subtraction and addition to follow numerical changes over time, time period analyses to search for seasonal and other cyclical events, and averaged annual valuations to smooth out random variations. He used the Merchant‘s or Arithmetic Rule of Three to calculate the fourth unknown factor in a ratio relationship where three of the four numbers were known.^20,21 In presenting his results Graunt used proportions on 60 occasions, ratios on 20 occasions and percentages on six occasions. Using these methods he analysed his data ²² by groupings such as age, sex, parish, season, acute or chronic disease and other variables. Where necessary or appropriate he reclassified his data, often classifying like with like (which he did most commonly to rationalise the different terms used by Searchers) and sometimes to distinguish between what, in the manner of the times, were considered “intrinsick” or inherent values and those which were considered “accidental, or extrinsick”, these being contingent on some other value. This distinction could provide evidence, for example, of why one locality might be more healthy than another or even whether government policies appeared consistent with nature and therefore, by implication, with God's wishes. ²³

In searching for regularities, and in particular, for stable ratios in his findings, Graunt anticipated the suggestion made by Isaac Newton (1643–1727) in his Philosophiæ Naturalis Principia Mathematica of 1687. ²¹ He also used his book-keeping skills to search for irregularities and for errors and inconsistencies in the data extracted from the Bills, such as those arising from the under-reporting of baptisms.

It should be noted that Graunt sometimes used approximations in his calculations and offered no explanation as to why sums used in some calculations were inconsistent or differed from those in his tables.

Graunt‘s findings and conclusions

As it is not possible to discuss all of Graunt's findings and conclusions here I have selected some of those most likely to be of interest to readers of this Journal and have grouped them under the headings of Demographic, Epidemiological, and Political and Religious, these being terms which are familiar to the modern reader. However the omissions and re-ordering do, to some extent, detract from the logical sequence which Graunt used to demonstrate the validity of his methodology.

Demographic

When Graunt heard claims of several million people residing in London he recognised that such a figure was not compatible with the number of burials which never exceeded 15,000 annually. He approached the problem of estimating the city's population in three different ways. All three involved a number of assumptions, some of which were only guesses.

His first method was based on the observation that there were typically about 12,000 baptisms, which he equated with births, annually. On the assumption that women aged 16–40 had a child every other year there must be about 24,000 sexually active women and if they accounted for half of all women then there were 48,000 women in London. On the further assumptions that they represented 48,000 households and that the average household consisted of eight people (husband, wife, three children and three lodgers or servants) the total population would be 384,000. Having established that the ratio of males to females in London at both birth and death was 14:13 (and 15:14 in country districts) Graunt concluded that there were 199,112 males and 184,886 females in London. ¹⁰ Had he had information on age at death he might have realised the need for age-specific death rates in men and women.

His second method was based on the observation that there were about 13,000 deaths per annum and the assumption that in each year three out of eleven households experienced a bereavement. This implied 47,667 families and a total population of 381,338. His third method was based on the large scale map of London published by Fairthorne and Newcourt in 1658 and estimates of the number of houses in a given area and the number of people in each house. By this method he calculated a population of 380,160. To his original calculation of 384,000 he added one fifth to include “Westminster, Stepney, Lambeth and other distant Parishes” and this gave a figure of 460,800. ¹⁰

What Graunt did not know when he first published in 1662 was that a survey of London's population had been made in 1631 for taxation purposes. When he discovered this survey he updated it to account for changes in parishes and used his own estimate of the increase in population in the intervening three decades. This yielded a population of 403,000 in 1661 which he published in the third edition of his book in 1665. He had acknowledged in the first edition that his own calculations had been made “perhaps too much at random” and in the third edition he accepted that “… it will appear, that I computed too many rather than too few, although the most part of men thought otherwise.” ¹⁰ The most important errors in his original assumptions were that he had overestimated both the size of a household, which was probably nearer five than eight, and the fertility rate which he had assumed to be 500 per 1000. ² Nevertheless his estimate was similar to those calculated by modern demographers. ²⁴

Graunt observed how large decreases in population in London during plague years were followed by an increased birth rate in subsequent years. However much of London's long term growth in population of about 6000 p.a. was attributable to immigration from other towns and from the countryside. As the ratio of births to deaths was higher in the countryside than in London, he concluded that the population in the countryside remained stable.

Graunt also estimated the population of England and Wales using three methods. Having stated, without giving his source, that London accounted for one fifteenth of all tax revenues and on the unstated assumption that tax revenues were proportional to population size, he concluded that the population of England and Wales was 6.44 million. He arrived at similar figures based on assumptions about population density and on a separate calculation based on the assumed average populations of each of the 10,000 or so parishes in the two countries. ¹⁰

Once again he may have overestimated the true figure, perhaps by about twenty to thirty per cent.^25,26 Despite these relatively minor inaccuracies his population estimates represented the first rational attempt to calculate a population size as distinct from one derived from a survey or census. They also provided denominators for his epidemiological results.

Epidemiological

The tables made very clear the marked differences in annual death rates between what Graunt termed “Acute” or “Epidemical” diseases and those which he described as “Chronical”. Deaths due to the latter remained relatively constant. For example, during the period under study, annual deaths due to Consumption (tuberculosis) were never less than 1713 and never more than 3610, whereas deaths due to Plague were recorded as 35,417 in 1625 but only 4 in 1627 and none at all in 1629. The distinction between acute and chronic illness had been recognised since at least the time of Hippocrates but Graunt's quantitative analysis yielded new findings. For example he recognised that the total number of deaths during plague years increased by more than the number of deaths attributed to plague itself. From this he inferred that deaths due to plague were under-recorded by about one quarter and that the true number of plague deaths in 1625 should therefore have been about 44,000. Logistical limitations almost certainly resulted in under-recording of plague deaths during major epidemics, although not perhaps to the extent suggested by Daniel Defoe (c. 1660–1731) in his Journal of the Plague Year (1722). ²⁷ Searchers were bribed to ascribe some deaths to causes other than plague so that other members of the household were not locked in their own homes with a red cross daubed on the barred door for all to see. However Graunt's assumption that all excess deaths were due to plague was almost certainly an overestimate. Excess mortality during an epidemic, which is estimated by comparison with deaths during the same period in previous years in a defined geographical area, is a complex mixture of deaths directly attributable to the epidemical disease and those due to indirect or collateral deaths which are multifactorial. ²⁸

That Graunt recognised the distinction between mortality and morbidity is shown by his discussion of lunacy where he distinguished between those “that die by reason of their Madness“ and those that die of other diseases “unto which Lunacie is no Superseadeas [bar].” ¹⁰ Graunt was the first to estimate an infant mortality rate, though he actually expressed it as a ratio rather than a rate. Deaths attributed to Chrysomes, Infants, Overlaid and some other categories could all be attributed to infants but, as the Bills did not include age at death at this stage, he had to make assumptions about the proportions of other deaths which were attributable to very young children. He concluded that “about thirty six per centum of all quick [i.e. live] conceptions died before six years old.” ¹⁰ He referred to conceptions because he included abortions and stillbirths. If these are excluded the figure is thirty two per cent. ²⁹ Despite his unsubstantiated assumptions his figure is similar to modern estimates of infant mortality in 17^th century London.^30,31

Both Pelling and Rusnock have noted that Graunt neither commented on his extremely high estimate of infant mortality nor made any remedial suggestions.^12,17 Rusnock attributed this to a contemporary fatalistic attitude to infant mortality which was to persist for another century. Pelling thought Graunt's omission particularly remarkable because of his connections with the Hartlib Circle. She concluded that members of the predominantly male circle saw very young children as the province of women and not of men. Graunt did not hesitate to make suggestions for reform in other areas but he may have considered that it was not the place of a haberdasher to suggest medical remedies to improve infant mortality, especially as he hoped to use one of his medical friends to promote his work at the newly founded Royal Society which numbered many doctors among its members.

Graunt was the first to describe what has been called the “Urban Penalty”, namely the price paid in poor health and reduced longevity by those living in large conurbations. Having noted that in London there were now more deaths than births but that before 1600 there were, as in the countryside, more births than deaths:

I considered, whether a City, as it becomes more populous, doth not, for that very cause, become more unhealthfull, I inclined to believe, that London now is more unhealthfull, then heretofore, partly for that it is more populous, but chiefly, because I have heard, that 60 years ago few Sea-Coals were burnt in London, which now are universally used. For I have heard, that Newcastle is more unhealthfull than other places, and that many People cannot at all endure the smoak of London, not onely for its unpleasantness, but for the suffocations which it causes. ¹⁰

He had heard about the hazards of smoke due to Sea-Coals (coal brought by sea from Newcastle) from the diarist John Evelyn (1620–1706) who a year earlier had published a treatise on this subject and had urged that the burning of this coal should be banned in London. ³² In fact there had been complaints about the ill effects of sea-coal in the capital for at least 300 years. ³³

Whereas modern demographers have emphasised the importance of diseases due to overcrowding on infant and child mortality, ³⁴ Graunt did not mention the high infant mortality in this context, claiming that those most susceptible to pollution were immigrants from the countryside. He thought that the lower birth rate in London was attributable to a combination of a relative lack of women of reproductive age (the wives of merchants often being left in the countryside), the prohibition on marriage of apprentices, the presence of more unattached soldiers and sailors in the city, and infertility due to adultery and fornication. ¹⁰

Graunt also attempted an estimate of a case fatality rate in an outbreak of malignant fever in Romsey, concluding that “seven might be sick for one who died.” He was thereby enabled to distinguish this illness from plague which carried a much higher fatality rate. ¹⁰

Graunt's analysis of his Tables enabled him to refine nosology. For example, he observed how deaths attributed to “liver-grown” decreased pari passu with the increase attributed to “rickets” and, having also noted that in some years “liver-grown”, “spleen” and “rickets” were aggregated together under the same heading, he concluded that they were in fact the same condition in which “rickets” replaced the older name over several decades. Clarke described how several early writers on rickets had noted the presence of hepatomegaly and suggested that “liver-grown” might reflect the presence of co-existent nutritional diseases such as scurvy or kwasiorkor, both of which can occur in conjunction with rickets and both of which which are causes of hepatomegaly. ³⁵ After allowing for changes in burial numbers, Graunt also noted that deaths attributed to rickets and scurvy increased, those due to gout remained stable and those due to stone and strangury decreased.

Graunt sought to dispel erroneous beliefs and misconceptions. Noting that “many persons live in great fear, and apprehension of some of the more formidable, and notorious diseases” Graunt produced a table showing “the respective numbers [of such causes], being compared with the Total 229,250 [deaths], those persons may the better understand the hazard they are in.” The table indicated the relatively small numbers of deaths from murders, accidents, leprosy, bladder stone, apoplexy, palsy and various other potentially distressing causes. ¹⁰

In spite of Graunt's emphasis on the importance of numerical evidence rather than unproven preconceptions he did make occasional unsubstantiated statements. For example, he held that longevity was improbable in places with a high prevalence of chronic disease, though the single example which he did give (albeit without evidence) of metal workers seldom living to old age has subsequently been proven correct. He was also mistaken in his belief that the dramatic fluctuations in the prevalence of plague could only result from changes in the quality of the air and not from contagion. ³⁶ Although he provides no evidence it is probable that he was correct in claiming that longevity was the best measure of salubrity.

Having commented on the extent of Graunt's assumptions it should also be mentioned that many of his demographic and epidemiological observations required no conjecture: for example that autumn was the unhealthiest season, and that the more sickly the year the fewer births occurred and vice versa.

Political and religious observations

As indicated by his title, Graunt's “Observations” fell into two categories, Natural and Political, for which reason there were two dedications. The dedication relating to Natural Philosophy, examples of which have been discussed, was to Sir Robert Moray, the President of the Royal Society. The dedication relating to the Political Observations, which also included trade, religion and matters relating to morals and society, was to Lord Roberts (more usually spelt Robartes), the Lord Privy Seal, and some of these will now be mentioned.

Graunt was unequivocal in his belief that what we would now refer to as demographic and epidemiological data, such as he had collected, were indispensable if Government and Trade were to “be made certain and Regular” with consequent benefits for peace and prosperity. The term coined in about 1672 by his friend Petty to describe this process was Political Arithmetic. ³⁷ This was defined by Charles Davenant (1656–1714) as “the art of reasoning by figures, upon things relating to government” and was the forerunner of what is now described as economics. ³⁸ Graunt emphasised the importance of good demographic data not only to increase trade but also to avoid wasted effort “so as Trade might not be hoped for where it is impossible.” He was particularly keen to see “England as considerable for Trade as Holland”, believing that the Dutch had “more experience and skill in Trade” than the English. ¹⁰ Graunt's concept of promoting trade and prosperity through enhanced demographic knowledge was entirely consistent with Bacon's principles ¹⁵ but he became rather less Baconian when he moved into the realms of politics, religion and social policy.

Graunt's assertion, without supporting evidence, in his political dedication, “That the troublesome seclusions in the Plague-time is not a remedy to be purchased at vast inconveniencies” foreshadows opinions on the effectiveness of quarantine in preventing cholera in the nineteenth century and more recent restrictions during Covid-19. Then as now the loudest opposition to such measures came from merchants and tradespeople like Graunt.

As an officer in one of the Trained Bands he would have been one of the Parliamentarians defending London against the Royalist army; but following the Restoration he appears to have become an ardent, or perhaps pragmatic, royalist because in the dedications in his book he made some very flattering comments about King Charles II and observed that there was no evidence to support the “seditious” superstition that the coronation of a new king was accompanied by an outbreak of plague. His political priority now appears to have been a stable government which supported both the monarchy and strong moral and religious principles. It is not surprising that a merchant should long for a period of stability after the political turmoils of the previous decades.

Graunt introduces both political and social themes when commenting on his finding that the ratio of males to females is approximately equal. Most people at this time believed that females outnumbered males by three to one. This erroneous belief had led some to propose that polygamy, a common topic of debate at this time, should be made lawful. Two controversial Bills were brought before Parliament, one in 1658 and a second in 1675, the second arguing in favour of polygamy for peopling the nation and preventing the promiscuous use of women. ³⁹ However Graunt believed that his findings confirmed that the “Christian religion, prohibiting polygamy, is more agreeable to the Law of Nature, that is the Law of God, than Mahumetism, and others, that allow it.” He also argued that, where polygamy was allowed, women were viewed as mere servants rather than enjoying what Graunt claimed was parity with their husbands in a monogamous society. Hence “… the irreligious Proposals of some, to multiply People by Polygamy, is withall irrational, and fruitless.” ¹⁰ Pelling and Aldridge have discussed this topic in its wider contemporary aspects.^16,40

The need for denominators in epidemiological calculations was not the only reason for Graunt's interest in population sizes. The size of a country's population, and whether increasing or decreasing, was a major topic of discussion at a time when it was seen as a measure of a ruler's power, status and wealth. ²⁰ It was for this reason that Graunt was interested not only in the total population but also in population density and therefore, by implication, what size of population a country could sustain. ²⁶

Graunt's estimate of the number of men able to bear arms, and therefore to be eligible for military service, was also made for political purposes. Graunt defined this group as those aged 16–56 and then embarked on perhaps the most tortuous of his many sequences of assumptions and extrapolations. His calculations have been analysed in detail by many writers, among them Glass ² , Sutherland ⁴ , Rusnock ¹⁷ and Hald. ²⁹ In essence he used the estimate already described for those under six years old and made assumptions to estimate those surviving above 76 years. He then constructed a table to fill in the intervening seven decades. He does not explain how he did this but others have shown that he overestimated mortality in the earlier decades.^2,29 He intended to multiply the proportion who were alive between the ages of 16 and 56 by his previously calculated figure for the total male population but mistakenly used the proportion who died during that period. His table has sometimes been described as the first Life Table but it is not a Life Table in the modern actuarial sense. He was certainly the first to describe, with many assumptions, the dying out of a cohort. However it was a cohort which he himself had defined without knowing the ages at which individuals died or the size of the whole population and without any allowance for the immigration which he had shown was occurring. The first life table based on actual data of age at death was constructed by the astronomer and polymath Edmond Halley (1656–1742) in a population which is thought to have been stable but in which he had to estimate the total number. ⁴¹ Others then added further refinements. As one actuary has observed, the true significance of Graunt's table is not that it was defective but that he was the first to conceive the idea of presenting mortality in terms of survivorship. ⁴²

It was probably Graunt's observations on trade, population size and other political issues rather than his flattery which so impressed Charles II, at a time when the latter was keen to re-establish the authority of the newly restored monarchy, and which led Charles to personally recommend Graunt's election to the Royal Society. ⁴

Contemporary influences on Graunt's thinking

Apart from Petty and Bacon, who else might have influenced Graunt's thinking while he was working on his Observations? Morabia has suggested the philosopher René Descartes (1596–1650) and the physician Thomas Sydenham (1624–1689). ⁴³ Descartes’ Discours de la Méthode pour bien conduire sa raison, et chercher la vérité dans les sciences was published in 1637 and an English was published in 1649 so there was ample opportunity for Graunt to have studied it. That he followed Descartes’ reductionist approach is evident in his statement that he “reduced several great confused Volumes into a few perspicuous Tables, and abridged such Observations as naturally flowed from them, into a few succinct Paragraphs, without any long Series of multiloquious Deductions.” ¹⁰ The combination of Baconian tabulation and Cartesian methodological reductionism led naturally to the creation of groupings using variables like sex, parish, disease, seasons and years, and it also greatly simplified analysis of the data.

Graunt's rationalisation of the Searchers’ diverse nomenclature for the same or similar diseases is comparable with Sydenham's reductionist approach to nosology and both distinguished between acute and chronic disease. However Sydenham's first known publication was in 1666, four years after Graunt's Observations, and he did not discuss the aetiological differences between acute and chronic disease until his Treatise on Gout in 1683. It is possible that Graunt may have met Sydenham through Petty before 1662. Petty and Sydenham were contemporaries in Oxford where the latter, as an undergraduate at Magdalen Hall, was appointed as a Parliamentary delegate at Wadham (which was Petty's college) to identify those with Royalist views. Although they may have met there is no evidence that Sydenham took any interest in the activities of the members of the Oxford Philosophical Society in which Petty was active. Only when Sydenham moved to London in 1655 did he finally start to practice medicine and even then it was on a part time basis, his main initial income deriving from employment in a government post as Comptroller of the Pipe. He probably began full time medical practice following the Restoration in 1660 which would have ended any remaining political aspirations, but he did not became a licentiate of the Royal College of Physicians until 1663. ⁴⁴ It therefore seems unlikely that he had any influence on Graunt's thinking.

A source of influence on Graunt's political thinking might have been Thomas Hobbes (1588–1679). ⁴⁵ Hobbes, Graunt and Petty had much in common. Hobbes's wider family, though not his father, were clothiers. All admired Bacon and all believed in the importance of strong and stable government. Petty had been Hobbes's amanuensis in Paris where he had drawn the images for Hobbes's work on optics. ⁴⁶ Thereafter he regularly recommended Hobbes's writings to his friends and correspondents. ⁴⁷ Graunt might have met Hobbes through Petty or through Aubrey who was another mutual friend.

Hobbes described his vision of the ideal State in Leviathan (1651). The leviathan, a mythical sea animal of enormous size, serves as a metaphor for the State in which a “commonwealth” of people (the body of the leviathan) forms a “covenant” or social contract with a sovereign authority (the head) which may be an individual or a group of individuals. The covenant is a symmetrical or reciprocal agreement, a balance, in which all the individuals in the body agree to give up certain rights, retaining only those which they are content for others to retain over them. Similarly there must be a balance, symmetry or proportionality and mutual reciprocity between the head and body of the leviathan. In agreeing to ensure peace and stability in exchange for the people's obedience, the sovereign also gives up rights while taking on responsibilities for which he may be held responsible. ⁴⁸ Petty envisaged a similar arrangement but insisted that the covenant “should be shaped and drawn up by the people themselves; otherwise, the monarch will be susceptible to the daily change of affairs and to his temperament.” ⁴⁶

Graunt borrowed the “head and body” image of Hobbes's leviathan and the concept of its balance and reciprocity when he suggested that the rapidly-growing metropolis of London “is perhaps a Head too big for its Body, and possibly too strong.” He also suggested that information of the type which he had collected was necessary for “good, certain and easie Government, and even to balance Parties, and factions both in Church and State.” [my italics] He was more explicit in the third and fourth editions of 1665 when writing about the numbers of workers in different trades and professions: “The proportions of all which I have alwaies thought is necessary to be known, in order to an exact Symmetry of the several members of a Common-wealth.” ¹⁰

Graunt's achievements in historical perspective

The publication of Graunt's magnum opus was meticulously choreographed. At a meeting of the Royal Society on February 5^th 1662 the London physician Daniel Whistler (1619–84) brought in fifty copies of the book to be distributed among the one hundred members of the Society. He then read the flattering dedication to the Society's president and Graunt was proposed as a candidate for election. On February 12^th a committee was appointed to examine the book. The committee's members included three of the “founding twelve” Fellows of the Society: John Wilkins (1614–72) an Anglican clergyman shortly to become Dean of Ripon and later Bishop of Chester, the London physician Jonathan Goddard (1617–75) and Graunt's staunch friend William Petty. The other members were George Ent (1604–89), the president of the Royal College of Physicians and two other London physicians Caspar (sometimes Jaspar) Needham (1622–79) and Whistler who had introduced Graunt's book to the Society and was presumably, like Petty, a good friend. All had links which antedated their involvement with the Royal Society. Whistler, Petty and Goddard held or had held professorships at Gresham College (geometry, music and physic respectively), and reference to entries in the Oxford Dictionary of National Biography shows that Wilkins and Ent had attended meetings held in Goddard's lodgings. Graunt already had his own contacts at Gresham College where, as already noticed, he used his influence to obtain the professorship for Petty. It comes therefore as no surprise that Graunt was elected into the Society on February 26th just two weeks after the committee was appointed. He then subscribed his name as a Fellow on March 5^th. ² Leaving aside the specific individuals chosen to sit on the committee, the fact that five of the six were physicians is an indication that the importance of Graunt's work was initially perceived to lie in its medical and social findings rather than for its novel methodology.

The book was an immediate success. A second edition appeared later in 1662, and third and fourth editions in 1665. A fifth edition in 1676, after Graunt's death, was probably overseen by Petty. ⁴⁹ The work was soon recognised as an innovative way of using numerical analysis in both scientific research and in matters of everyday importance. Moreover, and as Kreager has observed, Graunt made a completely new concept intelligible to a broad audience of natural philosophers, physicians and mathematicians as well as to merchants, churchmen and public officials. ⁵⁰ Samuel Pepys, a civil servant at the Admiralty but not yet a Fellow, bought a copy on March 24^th 1662 and thought Graunt's observations “appear to me upon first sight very pretty.” ⁸ The potential for political arithmetic to provide the information which could enhance stable and effective government attracted other senior civil servants to the work of Graunt and Petty. ⁴⁵

Once their attention had been drawn to the London Bills other European cities started to produce similar documents. ²⁹ In Paris the process began in 1667, a year after a review of Graunt's book by “Sr. G. P.” (probably Petty) had appeared in Le Journal des Sçavans. ⁵¹ This was probably the first scientific journal in Europe when it began publication on January 5^th 1665, two months before the Philosophical Transactions of the Royal Society on March 6^th. ⁵² Ultimately the tables produced by individual cities were superseded by national collections of demographic and epidemiological data. However the collection of accurate data and of nosology remained problematic. In a paper to the Medical Society of London in 1768, nearly a century after Graunt's death, the London physician John Fothergill (1712–80) lamented the continuing use of searchers who “are, for the most part, ignorant poor women”. ⁵³ In Britain the situation only began to improve when the Births and Deaths Act of 1836 came into force in 1837. Thereafter the work of men like Edwin Chadwick (1800–90) and William Farr (1807–83) led to The Annual Reports of the Registrar General becoming a vehicle for administrative and social reform, as had been envisaged by Graunt. Nosology still remains a problem, as evidenced by the continuing updates to the International Classification of Diseases and the Diagnostic and Statistical Manual of Mental Disorders.

In the 1720s the London physicians John Arbuthnot (1665–1735) and James Jurin (1684–1750) presented numerical evidence to show that the risk of dying from smallpox inoculation was considerably less than that from natural smallpox. ¹⁷ Both followed Graunt's example of presenting their data as ratios or proportions and both used information extracted from the Bills of Mortality. Jurin's interest in the subject had been triggered through correspondence with the Halifax physician Thomas Nettleton (1683–1742). Nettleton's use of the phrase “Merchants Logique” to describe Jurin's method of calculation whereby an “Account of Profitt and Loss” will “find on which side the Balance lyes …” parallels Graunt's use of “my Shop-Arithmetique”. ⁵⁴

Others who followed Graunt's methodology or his use of the Bills of Mortality or both included the physicians Thomas Short (1690–1772), William Black (1749–1829), William Heberden the Younger (1767–1845) and the demographers and political economists Gregory King (1648–1712, Charles Davenant (1656–1714) and Johnann Peter Sûssmilch (1707–1767).^55,56 Sûssmilch's work influenced that of Thomas Malthus (1766–1834) whose Essay on the Principle of Population is believed to have shaped the thoughts of Charles Darwin when formulating his theory of natural selection. ⁵⁷ Mathematicians including Nicolaus Bernouilli (1687–1759) and Gottfried Wilhelm Leibnitz (1646–1716) also used Graunt's “mortality table”. ⁵⁸ The re-publication of Graunt's book with other works by the historian and secretary of the Royal Society Thomas Birch (1705–1766) in 1759 helped to establish both medical statistics and political economy as recognised disciplines. ⁵⁹ However, like Graunt and others before them, those working in these subjects in the eighteenth century were often obliged to rely on estimates of population until after the first census of Great Britain in 1801.

Graunt was studying the Bills of Mortality and working on the first four editions of his book at a time which witnessed the dawn of probability theory. In 1654 Blaise Pascal (1623–1662) and Pierre de Fermat (1607–1665) began their correspondence on how the stakes should be divided fairly if a game of chance were interrupted unexpectedly. In 1655 Pascal published Traité du Triangle Arithmétique and in 1657 Huygens published On Reckoning at Games of Chance, widely regarded as the first published treatise on probability theory. ⁶⁰ In 1662, the same year that Graunt published his Observations, Antoine Arnaud (1612–1694) published La Logique on l’art de penser, also known as the Logique Port Royal. Those present at the Port Royal discussions included Pascal, Huygens and Fermat and the work contains the first mention of the word ‘probability’ in the context of numerical calculations. ⁶⁰

When Graunt used ‘probability’ he did so only in the context of conjectural opinion. He never referred to ‘risk’ but, as already mentioned, he wrote of ‘hazard’ where we might now use ‘risk’. As a merchant he would have been familiar with the concept of risk in business, consequent upon such matters as broken contracts, ships and cargo lost at sea (Lloyds of London was founded a few years later in 1686 to insure against such risks) and epidemics of plague. It is significant that Daniel Defoe chose a well-to-do saddler to be his mouthpiece in his Journal of the Plague Year which can be read as a study in risk management. ²⁷ That Graunt had a good grasp of risks and probabilities in business is suggested by his reputation as an arbitrator; but did he also have an understanding of probability in relation to numerical data? Sutherland cites passages in the Observations which suggest that he did and that he was certainly aware of chance variation. ⁴ Kreager sees him as an “early probabilist” but as one in a transitional phase who retained links to the Scholastic tradition and who still used Rhetorical language.^50,61 Pelling also finds an Aristotelian echo in his work. ¹⁶

It is likely that Graunt would have heard of Huygens and his ideas, if not before publishing his Observations then certainly soon after. Huygens had met Moray at meetings in Gresham College while staying in London between March and May 1661 ⁶² and Moray sent Huygens a copy of Graunt's work in March 1662. Huygens admired the clarity of Graunt's observations and the logical way in which they were developed. ⁶³ The two men would certainly have had the opportunity to talk at meetings of the Royal Society between June and September 1663 when Huygens was in England and after his election as a Fellow on 22 June 1663. However Kreager found no evidence of Graunt “participating in abstract mathematical discussions at the Royal Society” and noted that he did not adapt his methodology in the third and fourth editions of his work in 1665 to take account of new thinking on probability in mainland Europe. ⁶¹ Perhaps Graunt saw no need to re-analyse his descriptive statistical data in the context of the newly emerging probability theory. After all, he could claim to have made a convincing case for the conclusions which he had reached and, as Hacking has observed, The Logique Port Royal “has whole sentences of exactly the same form as are found in Graunt”. ⁶⁰

Some conclusions

As the founding father of the study of what became known as human demography, Graunt was the first to make a reasoned estimate of the numbers in a population and the first to document that more boys are born than are girls and in what proportion. Where previously there had been only conjecture, Graunt supplied factual evidence. Because he followed population data over time he was the first to quantify population movements such as immigration from rural to urban areas and he was also able to document the time required for London to repopulate itself after an epidemic.

As the founder of epidemiology Graunt produced the first known tables of aggregated health data including the first numerical statement of the relative frequencies of different causes of death. He devised sample surveys to make comparisons between different urban parishes and between rural and urban areas, the latter providing the first scientific evidence of the ‘urban penalty’. He was the first to quantify the extremely high infant mortality in early modern England and he also attempted an estimate of a case fatality rate.

As the first to make numerical studies of diseases over time Graunt was able to quantify and clarify distinctions between epidemic and endemic diseases and to demonstrate that many of the latter were chronic illnesses. His time studies also allowed him to show that autumn was the least healthy of the seasons and that London had become less healthy as it grew in size. His comparisons of plague years with non-plague years revealed that plague deaths were under-estimated. His time studies demonstrated that mortality from some conditions, mainly chronic illnesses but also accidents and suicides, remained relatively stable whereas others increased over time. He would have recognised the progressive increases in scurvy and rickets in London over time as a measure of the city's escalating un-healthiness but would not have known the reasons why. While noting that the dietary factor in the development of rickets in seventeenth century London cannot be known, Brimblecombe has claimed that the increase in rickets at this time directly paralleled the increase in coal imports and he speculated that some of the increase in rickets was attributable to a reduction in sunlight as a consequence of a soot laden atmoshphere. ⁶⁴

Although Graunt was not a statistician in the modern sense of the word, he was the first to create a template for the numerical analysis of data of a type that we would now characterise as demographic and epidemiologic. As the first person to collect, collate and analyse data on births and deaths and their causes he was the founding father of vital statistics. He was the first to quantify migration and, although he lacked the mathematical skills to develop tests of correlation, he was the first to develop concepts of statistical association, of statistical inference, and of sampling. He was the first to make a rational estimation of population sizes and, as the first to study biological numerical data over time, he was the first to demonstrate the existence of stable ratios such as the male/female ratio of births and the ratio of accidental to total deaths.

Graunt placed much emphasis on those of his findings which were stable over time, and especially on those ratios which remained stable. This may be seen in the context of an Enlightenment desire to find order amid disorder in nature and to link this with order and stability in political, social, moral and religious life. The importance which Graunt attached to political stability has already been emphasised. In like manner, his views on polygamy, which he considered to be supported by his finding of a near equal sex ratio, were an extension of his perception of the natural order into the realms of social and religious life.

Graunt's revolutionary methods of analysing data changed people's concepts of normality and their perceptions of the world in which they lived. They provided the foundation on which the study of demography and epidemiology could advance. However Graunt was an arithmetician and not a mathematician. ⁶⁵ Even if he had fully grasped the potential of combining his methodology with the new ideas on probability emanating from those present at the Port Royal discussions he would not have had the skills to exploit that opportunity. It was not until 1669 that Christiaan Huygens began this process ¹⁵ which, according to Hald, was only successfully accomplished by Thomas Bayes (1702–1761) in a posthumous publication of 1763 and, working independently, by Pierre-Simon Laplace (1749–1827) in 1774. ²⁹ Its first use in the analysis of demographic data was by Daniel Bernouilli (1700–1782) in 1771, almost a century after Graunt's death. ²⁹ Without probability analysis no certain distinction could be made between genuine statistical regularities and artefacts. Neither was it possible to be sure if a sample was truly random and nor was there any check on inherent bias in favour of preconceptions. ⁶⁶ Although Graunt seems to have had an intuitive sense of how much interpretation his data would stand, ⁴ the more extravagant speculations of some political arithmeticians in the eighteenth century contributed to the creation of separate disciplines of statistics and political economy. Probability theory continued to develop, particularly in astronomy, and it was the astronomer and polymath Adolphe Quetelet (1796–1874) who re-united the disciplines by linking advances in probability theory to social, demographic and epidemiological studies. ⁶⁶

A sad end

Graunt was an active member of the Royal Society in the four years after his election and was elected to the Council in December 1664. However he stopped attending after the Council meeting on March 21 1666. ⁶⁷ This was more than five months before the Great Fire of London which destroyed Graunt's home, business and the properties in Lothbury which he had bought with Petty and Martyn. Graunt was left with major financial difficulties but the properties were rebuilt with Petty's financial assistance. The only mention of Graunt's name in the Society's records thereafter was when he proposed Dr Henry Clerke (1621/22?-1687) for the Fellowship in October 1667. ⁶⁷ It is unlikely that Graunt had much contact with Clerke who spent little time in London and took no part in any of the Society's meetings. However Clerke was a close acquaintance of Petty, deputising for him over many years as Tomlins Reader in Anatomy in Oxford. ⁶⁸ By 1673 Graunt was four years in arrears with his subscriptions to the Society ⁶⁹ but he remained in good standing with the Drapers’ Company where he was appointed Renter Warden, a post with major financial responsibilities, ⁶⁵ and a member of the Court of Assistants in 1671. He last attended the Court in March 1673 but continued to hold both positions until his death. ⁷⁰ The Drapers’ Hall as Graunt would have known it before it was destroyed in the Great Fire had been the London mansion of Thomas Cromwell (1485–1540) which the Drapers acquired in 1543 after Cromwell's property had been sequestered by Henry VIII.

Meanwhile Graunt's business had failed to thrive. Petty's correspondence suggests irritation with financial mismanagement by Graunt who, according to Aubrey, became bankrupt.^1,2 His financial problems may have been aggravated by religious preoccupations. Having been brought up a Puritan he later became a Socinian, a form of anti-Trinitarianism allied to Unitarianism, before turning in his later years to Roman Catholicism, “of which religion he dyed a great zealot.” ¹ He was summoned for recusancy but died of jaundice before the adjourned case was heard. ³ Aubrey recorded that “His death is lamented by all good men that had the happinesse to know him; and a great number of ingeniose persons attended him to his grave. Among others, with teares, was that ingeniose and great virtuoso, Sir William Petty, his old friend and acquaintance …” ¹ His widow was left penniless but was supported by Petty who augmented her pension from the Drapers’ Company of £4 annually. ² In a letter sent to Petty eight years after Graunt's death the diplomat Sir Robert Southwell (1635–1702) wrote of “Poore John … in his State of fuddling and frailty.” ⁹ It was a sad end to what had been a public-spirited and distinguished life.