religions

Article
Faith in Numbers—Re-quantifying the English
Quaker Population during the Long
Eighteenth Century
Andrew Fincham
Department of Theology and Religion, University of Birmingham, 1046 Bristol Road, Birmingham B30 6LJ, UK;
AJF396@bham.ac.uk




Received: 21 November 2018; Accepted: 24 January 2019; Published: 30 January 2019


Abstract: The Religious Society of Friends was traditionally disinclined to define “membership” in
pursuit of the Quaker mission to create a worldwide church. As a result, for almost two centuries
before the first census of 1847, there are no records of the numbers within the Society. By the time of the
census, Victorian Friends were seeking an explanation of what they generally perceived as a decline.
The first, and still most widely accepted attempt to address this question, was John Stephenson
Rowntree’s Quakerism, Past and Present, who grounded his 1857 essay on estimates extrapolated from
summaries of Quaker records of marriages, births and deaths. This paper re-examines for the first
time the assumptions made by Rowntree, and deploys both more recent demographic estimates and
findings from the detailed Births, Marriages and Deaths records for over 33,000 Quakers in London
from 1650 to 1809 to create an alternative population model charting stock and flows in the Quaker
population of England by decade from 1650 to 1809. The paper seeks to reconcile such population
estimates with accepted trends in English Quakerism across the period of the long eighteenth century.

Keywords: Quaker; Religious Society of Friends; long eighteenth century; membership; statistics;
population; demographics; fertility; life expectancy; birth; marriage; death

1. Introduction
It is axiomatic within the curious world of Quaker academics that part of the attraction for
researchers lies in the abundance of archive material. From the 1650s to the present, the Society has
found no contradiction in silent worship and copious notetaking, as if a sect testifying to the necessity
of simplicity sought to leave a monument built from volumes of minutes. The seeds of this centricity
of the written may be found in the “pamphlet wars” of the mid-seventeenth century, when following
the abolition of the Star Chamber Court all parties exploited an unfettered press to publish an
explosion of opinion;1 the contemporary collection of London stationer George Thomason amounted
to 26,000 tracts, broadsides and pamphlets collected between 1640 and 1661.2 In this environment,
early Friends promoted publications and counter-publications, while their written records documented
increasingly organised internal deliberations conducted by a hierarchy of administrational committees
and sub-committees. These oversaw areas of importance to the Society, from scriptural justifications for
beliefs and practices, to shared finances, international travel, the licensing of members, marriages and

1 The original power to regulate printed material was enacted in June 1643: “An Ordinance for the Regulating of Printing”
(Firth and Rait 1911).
2 (Greenberg 1988; Bortescue 1908).

Religions 2019, 10, 83; doi:10.3390/rel10020083 www.mdpi.com/journal/religions

Religions 2019, 10, 83 2 of 15

ministers—down to less obvious aspects of life and death, such as advice on curtain-fringes and
serving cheese at burials.3
Given the abundance of such records in the archives, the absence of any reliable census of
membership from 1650 until 1840 requires an explanation. The traditional (and practical) response
ascribes this to a disinclination of the Society of Friends to define “membership”, which was finally
resolved by the definition of marriage regulations in 1737. Elizabeth Isichei added a spiritual rationale,
suggesting Quakers had “misgivings about King David’s punishment for numbering the people
of Israel”.4 Notwithstanding, for almost two centuries early Friends did not consider it necessary,
or advisable, to compile lists of membership of the Society. Instead, the relationship of individuals
with the organisation spanned a spectrum which can be thought of as four concentric circles: from the
inner circle of “weighty Friends”, through those “in unity” on the benches, to the disorderly walkers
who were “dealt with”, out to the “disowned”—a pseudo-excommunicated status, which for much of
the long eighteenth century (and even later) could be redeemed through the expression of suitable
contrition, usually required in writing.
Not quantifying numbers was an inevitable result of this rather fluid approach to membership,
which was itself a product of the contemporary Quaker belief system which acknowledged “God in
Everyone” in pursuit of their mission to create a worldwide church. Yet as this vision faded in the
eighteenth century, first the Society and later historians began to seek explanations for what was
generally perceived as a decline. The first—and most widely accepted—attempt to address this
question was contained in John Stephenson Rowntree’s Quakerism, Past and Present, subtitled “Being an
enquiry into the causes of its decline . . . ” This was published as the winning entry in an open essay
contest run in 1858 on behalf of “A Gentleman . . . who laments that the Society of Friends is less in
number than at the beginning of the century . . . and who . . . is anxious to obtain light respecting
the causes of this change”.5 Significantly, Rowntree grounded his enquiry into the decline upon a
foundation of numbers extrapolated from summaries of Quaker records of marriages, births and
deaths, applying modified contemporary ratios to arrive at a total membership population. At the
heart of his estimates are the marriage records: Rowntree first formulates a marriage ratio of 1 per
150 of the Quaker population for the year 1800. This is based on Quarterly Meeting (QM) marriage
records and his estimate of the number of Quakers at the close of the century, which he back-projects
from the Society’s 1847 census adjusted for those dying, leaving or joining and estimated from samples
of “representative” QM records. This ratio he then modifies to 1:140 for the late seventeenth century,
an increase intended to account for missing records. He next selects the high point of QM marriage
registers (2820 in the decade beginning in 1670), and multiplies this by 140 to produce an initial
estimate of almost 40,000 Quakers in England and Wales in 1680.6 To this, Rowntree goes on to add or
subtract extrapolated figures for various factors and geographies in pursuit of a maximum Quaker
population, the decline of which his narrative addresses to account for the census membership figure
of 15,345 in 1847.7
Questions of methodology notwithstanding, the figure of 40,000 has proved remarkably enduring.
Fifty years later, the formative Quaker historian William Braithwaite, while not acknowledging
Rowntree, came up with the same figure: his approach was to estimate the number of male Quakers
from those imprisoned in the Fifth Monarchist uprising of 1661, and use a family size of approximately 6

3 For a review of advices contained in the early Books for Discipline, see Fincham in (Healey 2019).
4 (Isichei 1970).
5 (Rowntree 1859).
6 (Rowntree 1859, p. 71). The QM marriage records used by Rowntree include “Hereford, Worcester and Wales”,
numbering 666 from 1650 to 1779, or 3% of the Society total (81); the proportion from Wales is not given, thus while
Wales is nominally included in the model, Quaker dynamics in the Principality are too small to be addressed separately in
this paper.
7 (Rowntree 1859, p. 87).
Religions 2019, 10, 83 3 of 15
Religions 2019, 10, x FOR PEER REVIEW 3 of 15

to factor 8 Fifty years after Braithwaite, Isichei’s study of Victorian Quakers

study ofinVictorian
the women and children.
Quakers characterizes Rowntree’s work as careful “calculations based on all
characterizes Rowntree’s
available statistics”, andwork
acceptsas careful “calculations
his estimates based on
accordingly all available
without furtherstatistics”,
question.and9 Inaccepts
the samehis
estimates accordingly 9
decade, Alan Gilbert without further question.
uses Rowntree as his source In the
forsame decade,
Quaker Alan Gilbert
numbers during uses Rowntree
the long as his
eighteenth
source
century,for
10 Quaker
following numbers
Richardduring
Vann the long
in his eighteenth
1969 work The Social 10
century, following Richard
Development of English Vann in his 1969
Quakerism. Yet
work The Social Development of English Quakerism. Yet Vann introduces a note
Vann introduces a note of caution as to the quality of the records, noting that even contemporaryof caution as to the quality
of the records,
Friends noting thatateven
were concerned the contemporary Friends Based
laxity of registration. were concerned at the laxity
upon a comparison ofoflocal
registration.
meeting
Based
minutesuponanda comparison
registers of of local he
births, meeting
mademinutes and registers
an assessment that atof some
births,periods
he madeand an assessment
in some places that
at some periods
registrations and be
might in some
off byplaces
as much registrations
as 20%, and might be off byrather
concluded as much as 20%, and
unhelpfully concluded
that when tryingrather
to
unhelpfully that when
determine whether trying
there were to fewer
determine
friendswhether
in the there were fewer
eighteenth centuryfriends
than in
in the
theeighteenth
seventeenth, century
“one
than 11 By
man’singuess
the seventeenth,
is almost as good“oneas man’s guess 11is By
another’s”. almost as good
the time Adrian as Davies
another’s”.
addressed time in
the issue Adrian
2000,
Davies
he could addressed
conclude the thereissue
wasin“a2000,
declinehe in
could conclude
the sect’s there was
numbers” “a decline
without in thetosect’s
committing numbers”
any numerical
without
estimatecommitting
at either endtoofany thenumerical
process.12 estimate at eitherhave
Other historians end of process.12 Other
theencouraged
been historians
to follow have
suit, and in
been encouraged to follow suit, and in the face of such consensus, any re-evaluation
the face of such consensus, any re-evaluation of Rowntree might appear unnecessary. However, as of Rowntree might
appear unnecessary.
the history However,
of the Society as theattracts
of Friends historynew of the Society ofsoFriends
historians, attracts new repays
the historiography historians, so the
revisiting:
historiography repays revisiting:
the Quaker experience merits more the Quaker
textureexperience meritstransition
than a simple more texture fromthan a simple transition
seventeenth century
from seventeenth
“enthusiasm” century
through “enthusiasm”
eighteenth through
century eighteenth
“quietism” centuryactivism”
to “social “quietism”intothe “social activism”
nineteenth, in
and
the nineteenth,
Rowntree’s and Rowntree’s
approach approach
merits another’s merits another’s guess.
guess.

2.
2. Constancy
Constancy as
as aa Route
Route to
to Accuracy
Accuracy
The
The first
first questions
questions concern
concern the
the lack
lack of
of methodological
methodological consistency.
consistency. Rowntree
Rowntree was
was primarily
primarily
concerned with an explanation for the decline in nineteenth century Quakerism, yet the application
concerned with an explanation for the decline in nineteenth century Quakerism, yet the application of
his “marriage ratio” to the records of the Society across the whole period suggests a rather different
of his “marriage ratio” to the records of the Society across the whole period suggests a rather different
picture
picture (Figure
(Figure 1).
1).

Rowntree "Marriage*140"

45,000
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
-

13
Figure 1. Estimates of
Figure 1. of English
English Quaker
Quaker population
population using
using JS
JS Rowntree’s
Rowntree’s “Marriage
“MarriageRatio”
Ratio”(140).
(140)13 .

Applying the ratio of marriages against the numbers of marriages recorded in the QM registers
shows that the fall in Quaker population to the mid-nineteenth century level of approximately 15,000
8 (Braithwaite 1912). His total assumes that the 4200 recorded as jailed represented approximately 50% of adult male Quakers,
occurs in the 1730s, after which it remains relatively static (aside from a small anomaly in the 1760s,
and estimates between 5 and 7 for families.
9 (Isichei 1970, p. 112).
10 (Gilbert 1976); unfortunately, his source is Rowntree.
9
11 (Isichei
(Vann 1970, p. 112).
1969).
10
12 (Gilbert2000;
(Davies 1976);
Coleunfortunately,
1957). his source is Rowntree.
13 Rowntree does not provide figures for marriages for the last three decades. Values for these have been extrapolated from
11 (Vann 1969).
Birth per Marriage ratios, and cross-validated with ratios to London and Middlesex Quarterly Meeting marriage records,
12 (Davies 2000; Cole 1957).
as discussed below.
13 Rowntree does not provide figures for marriages for the last three decades. Values for these have been
extrapolated from Birth per Marriage ratios, and cross-validated with ratios to London and Middlesex
Quarterly Meeting marriage records, as discussed below.
Religions 2019, 10, 83 4 of 15

Applying the ratio of marriages against the numbers of marriages recorded in the QM registers
shows that the fall in Quaker population to the mid-nineteenth century level of approximately 15,000
occurs in the 1730s, after which it remains relatively static (aside from a small anomaly in the 1760s,
probably arising from the recognition of Quaker marriage in the Act of 1755). This would suggest that
any explanation for the decline of Quakerism would need to be grounded in the period after 1670,
when historians accept the movement was at its most vigorous. Clearly, as Vann suggests, the marriage
registers cannot be wholly accurate, but there is also a question as to the applicability of this
“marriage ratio”. The evidence of the birth and burial registers used by Rowntree present additional
problems: deaths outnumber births by 19,442 over the period, which leaves a 50% differential by the
end of the century, whereas the size of the population estimated is very low given the 113,897 births
recorded, which requires an average births per marriage (BpM) rate above seven—well in excess of
both Quaker and general population norms.

3. Examining Marriage
These issues can be remedied by accepting and attempting to quantify the under-recording
of marriages. The “missing marriages” can be back-projected through estimates of BpM using the
QM birth registers, which Rowntree provides for all decades of the period.14 Demographers have
mapped trends based upon birth records over the period for general populations, with fertility
tracked as the gross reproduction rate (GRR) representing the number of female births per 1000 of
population, and Richard Smith provides a review of fluctuations over the period 1541–1871 from
which relevant data can be extracted.15 Allowing differences in the male–female birth balance to even
out over time (while accepting any consequent errors), the GRR becomes a useful benchmark for
fertility during the period. There exist also indicators of trends in BpM. However, two significant
issues arise with such indicators: first, these are constructed by aggregating different data sets across
diverse smaller communities over long periods. Variables include changing patterns of nuptiality
(resulting from changes in wages or occupation), differing patterns of age at marriage (or not marrying),
falling decennial mean birth intervals (increasing marital fertility) and not least the significance
(or otherwise) of general changes in demography.16 While the interactions of such indices provide a
fruitful source of debate for professional demographers, they do not materially assist the historian
seeking to verify behaviours of discrete populations over specific periods. Secondly, the Quakers
themselves are traditionally regarded by their biographers as “different”: in the words of Professor
Robynne Rogers Healy, perhaps the leading authority on the sect during the long eighteenth century,
they exhibited “a community cohesiveness that spanned geographic boundaries while it continued
to keep Quakers apart as a ’peculiar’ people”.17 Unfortunately, sect-specific studies seem to identify
almost as many differences between groups of Friends as between Friends and non-Quakers.18
The largest (and most recent) demographic study, Vann’s and Eversley’s Friends in Life and Death:
The British and Irish Quakers in the Demographic Transition, 1650–1900, demonstrates the difficulties
which arise from using the demographic technique of family reconstruction, which demands creative
assumptions which do not align easily with statistical analysis. Noting the “immense differences in the
demographic data” arising from the authors’ estimates in England and Ireland,19 historian J. William
Frost concludes:

14 (Rowntree 1859, p. 79).

15 (Smith 1981).
16 See, respectively: (Galloway 1988; Schellekens 1997; Dixon 1971; Wrigley 1998; Goldstone 1986).
17 (Healy 2006).
18 (Wells and Zuckerman 1972). Helpfully, “by the late eighteenth century marriage patterns in the colonies started to move
into line with those found in Europe”; p. 428.
19 (Vann and Eversley 1992).
Religions 2019, 10, 83 5 of 15

The reasons for using Quaker records for demography are clear; the birth, death,
and marriage records provide more reliable information about a numerous population
over a large area for a long period than any other English records. While demographers
then have the difficulty of comparing Quakers with other English-men and women because
Friends were by no means typical, that is not an issue for the social history of London
Yearly Meeting.20

This argument is convincing, and leaves no better alternative to the statistician than to make a
count from the registers: both the number of marriages per decade and the number of births, and from
a single Quaker population. The registers chosen for this somewhat laborious study are those from the
London and Middlesex Quarterly Meeting (L&M QM), both because they represent the largest single
QM in terms of records from 1650 to 1809, and because they have been transcribed electronically thanks
to significant effort by the Quaker Family History Society. This simultaneously increases the accessible
data set and reduces the effort, while methodologically introducing one or two challenges in that the
transcriptions retain a sizeable number of duplicates arising from a decision to preserve alternative
versions of the recorded data held in Monthly and Quarterly Meeting registers. The QM registers were
essentially transcriptions of the records of subordinate Monthly Meetings, and as such introduce errors
through misspellings, transposing names, months and days, general omissions, and (occasionally in
the birth records), changes of given name between the dates of entry. For this pursuit of the numbers,
the primary sorting criteria used date and surname, names and monthly meeting—thus enabling a
common-sense approach to de-duplication.21 This approach also has the inestimable advantage of
providing a basic validation of the data given in Rowntree for the totals of marriages for L&MQM by
decade: clearly, had the figures used in 1856 proved wildly inaccurate, any re-evaluation of this work
would need to take a different course.
Happily, the numbers correlate well (Table 1), and permit acceptance of other raw data used by
Rowntree with some degree of confidence.

Table 1. Comparison of Rowntree’s marriages by decade with the London and Middlesex Quarterly
Meet (L&M QM) marriage registers.22

Decade From N(M)JSR Accuracy N(M) L&M QM

1650 0 3
1660 171 101.2% 173
1670 485 104.8% 509
1680 555 103.2% 573
1690 485 103.4% 502
1700 479 101.0% 484
1710 395 101.4% 401
1720 315 101.1% 319
1730 196 100.8% 198
1740 187 101.9% 191
1750 169 100.0% 169
1760 179 102.0% 183
1770 154 102.6% 158
1780 152 100.0% 152
1790 131 100.0% 131
1800 151 100.0% 151
4203 4291

20 (Frost 1993).
21 An error would arise should two couples with identical names, from the same monthly meetings, marry on the same
date—since one of these records would be discounted. In fact, the records show very few dates with more than one marriage,
and none where couples are from the same monthly meeting.
22 (Rowntree 1859, p. 81), Note Six, Table, line Item “London and Middlesex”.
Religions 2019, 10, 83 6 of 15

4. Evidence of Births
Having analysed the marriages, the L&M QM Birth register was subjected to a similar process
Religions 2019,
Religions 2019, 10,
10, xx FOR
FOR PEER
PEER REVIEW
REVIEW 66 of
of 15
15
of de-duplication, then totalled by decade to enable a Quaker-specific BpM rate to be calculated for
L&M
L&M QM. QM. This
This waswas then
then compared
compared to
to the
the general
general Quaker
Quaker BpM
BpM rate
rate derived
derived from
from Rowntree’s
Rowntree’s work
work
L&M QM. This was then compared to the general Quaker BpM rate derived from Rowntree’s work
(Figure 2).
(Figure 2).2).
(Figure

9.00
9.00
8.00
8.00
7.00
7.00
6.00
6.00
5.00
5.00
4.00
4.00
3.00
3.00
2.00
2.00
1.00
1.00
--
1660 1670
1660 1670 1680
1680 1690
1690 1700
1700 1710
1710 1720
1720 1730
1730 1740
1740 1750
1750 1760
1760 1770
1770 1780
1780 1790
1790 1800
1800

Corrected BpM
Corrected BpM (L&MQM,
(L&MQM, 30%)
30%) BpM (JSR)
BpM (JSR) Gross Reproduction
Gross Reproduction Rate
Rate ** 2
2

Figure 2.
Figure 2. Comparison
Comparison of
of births
births per
per marriage
marriage (BpM)
(BpM) for
for L&M
L&M QM
QM and
and Rowntree,
Rowntree, gross
gross reproduction
reproduction
rate (GRR) (doubled) for comparison 23
23 .
rate (GRR) (doubled) for comparison.
comparison .23

Two clear
Two clear observations
observations can
observations canbe
can bemade:
be made:Quaker
made: Quakerfertility
Quaker fertility
fertility followed
followed
followed that
that
that of
thethe
of the
of general
general
general population
population
population (as
(as
(as
per per
thethe
GRR)GRR)
overover
thethe period;
period; and
and (assuming
(assuming that
that Quakers
Quakers in
in general
general had
had families
families
per the GRR) over the period; and (assuming that Quakers in general had families which similarly which
which similarly
ranged
ranged inin size
size to
to those
to those of
those of L&M
of L&M QM),
L&M QM), Rowntree’s
QM), Rowntree’s marriages
Rowntree’s marriages must
marriages must be
must be understated.
be understated. Fortunately,
understated. Fortunately,
Fortunately,
Rowntree himself believed
Rowntree himself believed this
this was
was the case—hence his
the case—hence
case—hence modification of
his modification
modification of the
of original marriage
the original
original marriage ratio.
marriage ratio.
ratio.
Using Rowntree’s
Rowntree’s data
data for
forbirths
birthsbybydecade,
decade, wewe can
can use
use the L&M
the L&M QMQM BpM
BpM
Using Rowntree’s data for births by decade, we can use the L&M QM BpM to estimate the numberto estimate
to the
estimate number
the number of
marriages
of marriages
of these
marriages these births
these birthsrepresent,
births represent,and
represent, andsubtracting
and subtracting the
subtracting the number
the number from
number from his record,
from his arrive
his record, at
record, arrive an
arrive at estimate
at an
an estimatefor
estimate
the
for number
for the
the number
numberof “missing”
of “missing”
of “missing”marriages (Figure
marriages
marriages 3). 3).
(Figure
(Figure 3).

500
500
450
450
400
400
350
350
300
300
250
250
200
200
150
150
100
100
50
50
0
0
1650 1660
1650 1660 1670
1670 1680
1680 1690
1690 1700
1700 1710
1710 1720
1720 1730
1730 1740
1740 1750
1750 1760
1760 1770
1770 1780
1780 1790
1790 1800
1800

JSR Marriages
JSR Marriages Marriages Needed
Marriages Needed for
for Births
Births Annual "Missing"
Annual "Missing" Marriages
Marriages

Figure 3.
Figure 3. “Missing
“Missing Marriages”:
“MissingMarriages”: comparison
Marriages”:comparison of annual
of
comparison annual marriages
marriages
of annual (from
(from
marriages BpMBpM
BpM
(from for L&M
for L&M QM) QM)
QM)
for L&M and
and
Rowntree.
Rowntree.
and Rowntree.

Evidently, more
Evidently, more marriages
marriages are are needed
needed to to account
account forfor the
the births
births than
than are
are in
in the
the registers.
registers. Vann
Vann
notwithstanding,
23 correction is seeking
required to an
later explanation
data arising from for this
“marrying under-recording
out” (under-registrationsolely
notwithstanding, seeking an explanation for this under-recording solely in the slackness of the
A of in the
marriages) slackness
and of the
over-registration of
births for “non-members”.
contemporary clerking This is
clerking procedures discussed
procedures runs below.
runs against
against the
the wider
wider view
view that
that Quaker
Quaker records
records were
were generally
generally
contemporary
well kept. A much stronger explanation reflects the growing nature
well kept. A much stronger explanation reflects the growing nature of the movement in the of the movement in the early
early

23
23 A correction
A correction is
is required
required to
to later
later data
data arising
arising from
from “marrying
“marrying out”
out” (under-registration
(under-registration of
of marriages)
marriages) and
and over-
over-
Religions 2019, 10, 83 7 of 15

Evidently, more marriages are needed to account for the births than are in the registers.
Vann notwithstanding, seeking an explanation for this under-recording solely in the slackness of
the contemporary clerking procedures runs against the wider view that Quaker records were generally
well kept. A much stronger explanation reflects the growing nature of the movement in the early
decades: the recorded births being the natural result of marriages contracted before the parents became
Quakers, thus QM marriage registers would not reflect subsequent births.24

5. Considering Burials
Having begun to reconcile QM birth and marriage registers, the third pillar of burial records must
now be considered before the three can be incorporated in a mathematical model. Burial numbers
for the Society by decade are provided by Rowntree and, intriguingly, from 1750 these also include
an estimate of those included who were considered to be “non-members”.25 This topic remains a
challenge to historians and demographers alike, since most scholars of Quakerism accept that the
status of being “in membership” remained subject to interpretation over much of the period, with the
concept of “being in unity” or “disowned” at least as widely used as that of member or non-member.
Thus, these non-members might be either former Friends or, equally possibly, the (close?) relations of
Friends. Considering the total numbers of burials shows that from 1670 to 1730 inclusive, more than
10,000 burial entries occur for each decade, while over the whole period from 1650 to 1809, the total
number of burials recorded is in excess of 133,000. As with births, demographers have estimated a
general crude death rate (CDR) over the period, which when applied to the recorded Quaker deaths
by decade would produce an estimate of population. In practice, the CDR suffers the same problems
of relevancy as the birth and marriage rates. Happily, the death registers for L&M QM have also been
transcribed, and provide useful detail on burials throughout the period. Sorting deaths per decade by
age group provides a data set of 33,231 deaths. The age groups chosen were 0–14 years, 15–34 years,
35–49 years, 50–65 years and “Aged”. This allows for the calculation not only of life expectancy at
birth (l0 ), but also after the age of 14 (l15 ), necessary since until 1680, and in some locations after,
deaths in this group accounted for more than 50% of burials.26 The data permits the calculation of
an age at death weighted by numbers per decade for London and Middlesex, giving a L&M l0 rising
from 21.6 years in the decade beginning in 1660, to 38 years by the end of the century, with a peak of
39 years in the 1750s. Although the L&M QM statistical analysis shows a more rapid increase to l0 32 by
1700, the trend indicated aligns with the estimates Vann and Eversley produced using the alternative
“family reconstruction” methodology, which gave London l0 of 26 years at 1700, rising to 45 years
by 1849; outside London their findings indicate expectancy might be ten years greater. Both sets of
figures and trends are also in line with general demographic estimates created by the founders of
the field, E.A. Wrigley and R. S. Schofield. Modifying expectancy to eradicate the effects of plague,
the wealth of individuals, the health and even the prosperity of particular parishes within London
permit a wide range of modified estimates to account for local population results, which is ultimately
unhelpful: that the L&M QM registers record the deaths of around half a dozen centenarians per
decade is sufficient pause for thought. However, it is safe to conclude that there was an increase in the
age of the population over the period (Figure 4).

24 The solution as proposed by Vann. I am grateful to Professor Pink Dandelion for originally drawing my attention to
this possibility.
25 (Rowntree 1859, p. 82), Note 7, Table.
26 A key paper from the founders of the field is: (Wrigley and Schofield 1983). Modifying the assumptions detailed in this
aggregated study of 13 unconnected parishes in England over the period to reflect Quaker differentials presents a challenge
which is best answered by Professor Frost’s observation, quoted above.
Religions 2019, 10, 83 8 of 15
Religions 2019, 10, x FOR PEER REVIEW 8 of 15

Figure 4. Age
Figure 4. Age structure
structure of
of L&M
L&M QM
QM burial
burial registers.
registers.
6. Calculating Population from Marriages
6. Calculating Population from Marriages
Having established some data points from the L&M QM registers, we can begin to assemble
Having established some data points from the L&M QM registers, we can begin to assemble the
the rules for a model. Analysis of the L&M QM burial records shows the proportion of the Quaker
rules for a model. Analysis of the L&M QM burial records shows the proportion of the Quaker
population that made up each age group, which in turn provides an estimate of the size of the segment
population that made up each age group, which in turn provides an estimate of the size of the
regarded as the “first marriage” cadre: 15–34 years. The crude first marriage rate (CFMR) has been
segment regarded as the “first marriage” cadre: 15–34 years. The crude first marriage rate (CFMR)
mapped by demographers for the period and indicates the number of (first) marriages per 1000 of
has been mapped by demographers for the period and indicates the number of (first) marriages per
the 15–34 population, a factor which needs “adjustment” to address known Quaker propensities
1000 of the 15–34 population, a factor which needs “adjustment” to address known Quaker
for initially marrying later, and later not marrying at all. While the extent of this can be debated,
propensities for initially marrying later, and later not marrying at all. While the extent of this can be
and following indications from the study of Quaker marriages by Wells and Zuckerman, and that of
debated, and following indications from the study of Quaker marriages by Wells and Zuckerman,
Vann and Eversley, this effect is here represented by a compression of the general CFMR by around 5%
and that of Vann and Eversley, this effect is here represented by a compression of the general CFMR
at the start of the period, to 10% by the end, both of which require a larger population per marriage.
by around 5% at the start of the period, to 10% by the end, both of which require a larger population
The CFMR, the revised estimate of marriages (derived from the BpM from the L&M QM Registers)
per marriage. The CFMR, the revised estimate of marriages (derived from the BpM from the L&M
and the size of the 15–34 cadre relative to that of the adult population of Quakers can be used together
QM Registers) and the size of the 15–34 cadre relative to that of the adult population of Quakers can
to generate an estimate of the Quaker population (QPOP) in England over the period (Table 2).27 As a
be used together to generate an estimate of the Quaker population (QPOP) in England over the period
reference point, Rowntree’s ratio of 1 marriage per 140 of the Quaker population can be seen to sit
(Table 2).27 As a reference point, Rowntree’s ratio of 1 marriage per 140 of the Quaker population can
in the centre of the range of values generated across the period, derived from dividing QPOP by the
be seen to sit in the centre of the range of values generated across the period, derived from dividing
predicted number of marriages generated from the BpM ratio (Table 2, Col. 7).
QPOP by the predicted number of marriages generated from the BpM ratio (Table 2, Col. 7).
The above QPOP derives from a steady BpM trend which requires up to one-third of the births
registered to have occurred in families whose marriage was either contracted before they became
Quakers, or was conducted outside the Society. Helpfully, Rowntree includes in the 1800 birth total
of 6910 the note that only 4863 were births of members, giving a reference point of 30% for that
decade. The estimated ratio of “convinced married” and “married out”, and some of the implications,
are examined below.

27 Data from (Smith 1981, p. 601) Figure 3, “Estimated crude first marriage rates (marriages per 1000 persons aged 15–34)
compared with gross reproduction rates (both 25-year moving averages), England, 1541–1871.” Save an anomaly in 1760
(16%), the trend within the 15–34 years cadre shows a consistent trend across the period, falling steadily from 20% in 1660 to
settle between 11–13%. This effectively offsets any impact caused by deriving marriage cadre size from that of deaths across
adjacent decades.

27 Data from (Smith 1981, p. 601) Figure 3, “Estimated crude first marriage rates (marriages per 1000 persons
aged 15–34) compared with gross reproduction rates (both 25-year moving averages), England, 1541–1871.”
Save an anomaly in 1760 (16%), the trend within the 15–34 years cadre shows a consistent trend across the
period, falling steadily from 20% in 1660 to settle between 11–13%. This effectively offsets any impact caused
by deriving marriage cadre size from that of deaths across adjacent decades.
Religions 2019, 10, 83 9 of 15
Religions 2019, 10, x FOR PEER REVIEW 9 of 15

Table 2.
Table 2. Estimate
Estimate of
of the
the English
English Quaker
Quaker population
population (QPOP).
(QPOP).

15-34 Crude Estimated Marriages Predicted Predicted QPOP

Cadre First Size of (from Marriages QPOP per Estimated
Decade Marriage #15-34 Rowntree) (by Births) Marriage from CFMR
Rate Cadre Ratio
(CFMR)
1650 42.1% 18.0 556 10 135 10 1,320
1660 42.1% 17.0 10,588 180 303 83 25,150
1670 36.4% 18.0 15,667 282 399 108 43,040
1680 32.2% 18.5 14,043 260 447 98 43,613
1690 24.9% 19.5 11,231 219 312 144 45,103
1700 23.4% 20.5 10,829 222 274 169 46,279
1710 22.9% 21.0 9,190 193 225 178 40,133
1720 22.9% 22.0 7,727 170 232 146 33,744
1730 22.9% 22.0 5,682 125 182 136 24,811
1740 21.8% 21.0 5,238 110 191 126 24,028
1750 20.0% 22.5 4,800 108 160 150 24,000
1760 23.9% 22.0 5,773 127 185 130 24,105
1770 20.0% 22.0 4,818 106 155 156 24,091
1780 19.5% 21.5 4,605 99 132 179 23,614
1790 19.1% 21.0 4,238 89 124 179 22,141
1800 19.6% 21.0 4,286 90 128 172 21,914
Marriage numbers are annualized.
Marriage numbers are annualized.

The abovefor
7. Accounting QPOP derives from a steady BpM trend which requires up to one-third of the births
Deaths
registered to have occurred in families whose marriage was either contracted before they became
Having constructed a mechanism to align births and marriages, the model requires some
Quakers, or was conducted outside the Society. Helpfully, Rowntree includes in the 1800 birth total
accounting for deaths. As noted above, a simple subtraction of registered burials from births using
of 6910 the note that only 4863 were births of members, giving a reference point of 30% for that
QM records by decade given by Rowntree gives a balancing item, which can be used to adjust the
decade. The estimated ratio of “convinced married” and “married out”, and some of the implications,
total for the following decade. However, it is preferable first to validate if the estimated size of QPOP
are examined below.
for each decade could reasonably be expected to generate the numbers of Quakers recorded for that
period in the burial
7. Accounting registers. This requires reviving the life expectancy trends calculated earlier from
for Deaths
the L&M QM burial registers, which needs to be (following Wrigley and Schofield) “adjusted upwards
Having
to remove theconstructed
depressing ainfluence
mechanism to align 28births
of London”. L&Mand QMmarriages, the model
registers account for requires some
approximately
accounting for deaths. As noted above, a simple subtraction of registered burials from births
17–20% of total Society marriage records, so a similar ratio might be used to approximate the uplift in l0 using
QM records
needed by decade
to reflect enhancedgiven by Rowntree
longevity outsidegives a balancing
the capital, ascendingitem, through
which canthebe used to
period. adjust the
Frustratingly,
totalapproach
this for the following
does notdecade.
deliver However,
the “right”itnumbers:
is preferable first it
instead tosuggests
validate that
if thea estimated
much largersizeQPOP
of QPOP
was
for each decade could reasonably be expected to generate the numbers of Quakers
required in the second half of the period, since fewer deaths arise from a longer-living population.recorded for that
period in the burial registers. This requires reviving the life expectancy trends calculated earlier from
To sustain the QPOP estimates from the birth and marriage registers actually requires a fall in average
the L&M QM burial registers, which needs to be (following Wrigley and Schofield) “adjusted
l0 during the long eighteenth century in order to generate the deaths recorded.
upwards to remove the depressing influence of London”. 28 L&M QM registers account for
Despite being apparently counter-intuitive, such a trend is likely to have been the case. This arises
approximately 17–20% of total Society marriage records, so a similar ratio might be used to
because the original population of convinced Quakers was not made up of a general cross-section
approximate the uplift in l0 needed to reflect enhanced longevity outside the capital, ascending
of the late seventeenth century population, who would exhibit death rates similar to the general
through the period. Frustratingly, this approach does not deliver the “right” numbers: instead it
population. Instead, these were largely younger adults: the median age of converts, according to
suggests that a much larger QPOP was required in the second half of the period, since fewer deaths
Vann’s restructuring, was thirty-three.29 The implications of this are clear: a population consisting of
arise from a longer-living population. To sustain the QPOP estimates from the birth and marriage
registers actually requires a fall in average l0 during the long eighteenth century in order to generate
the deaths recorded.
28 (Wrigley and Schofield 1983, p. 160)
29 Vann (1969, p. 48) His chapter on the “Theory and Practice of Conversion” uses data from Buckinghamshire and Norfolk
from 1654 to 1740.

28 (Wrigley and Schofield 1983, p. 160)

Religions 2019, 10, x FOR PEER REVIEW 10 of 15

Despite being apparently counter-intuitive, such a trend is likely to have been the case. This
arises because the original population of convinced Quakers was not made up of a general cross-
section2019,
Religions of the
10, 83late seventeenth century population, who would exhibit death rates similar to the
10 of 15
general population. Instead, these were largely younger adults: the median age of converts, according
to Vann’s restructuring, was thirty-three.29 The implications of this are clear: a population consisting
those whowho
of those had survived Death’s
had survived predations
Death’s on youth
predations onwould
youthhave a significantly
would higher life higher
have a significantly expectancy.
life
Calculating l from
expectancy. Calculating
15 the L&M QM burial registers suggests figures between 10 and 15 years
l15 from the L&M QM burial registers suggests figures between 10 and greater
15
over
yearsthe period
greater (Figure
over 5).
the period (Figure 5).

60.00

50.00

40.00

30.00

20.00

10.00

0.00
1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800

Life Expectancy at Birth Life Expectancy at 15 Years

Figure 5. Life expectancy calculated from the L&MQM burial registers.

Figure

AA Quaker
Quakerpopulation
populationwhich whichtransitioned
transitionedfrom fromone oneconsisting
consistingentirely
entirelyofof“convinced”
“convinced” converts
convertsto
one sustained
to one sustained substantially
substantially by births within
by births the Society
within wouldwould
the Society thereby expect expect
thereby to experience a general
to experience a
decline
generalin averageinlife
decline expectancy.
average The effectThe
life expectancy. would
effectbewould
equallybefelt in regions
equally felt inoutside
regionsLondon, but from
outside London,
the
butL&M
fromdata
the L&M (Figure 5) it(Figure
data is clear5)that even
it is bythat
clear the end
evenofbythe end ofl0the
century,
the is considerably
century, l0 is lower than l15
considerably
for most
lower of the
than period.
l15 for most Assuming
of the period. thatAssuming
Quakers inthat general had lifestyles
Quakers in generalthathadaligned more
lifestyles thatclosely
alignedto
other
more Quakers
closely tothan othertoQuakers
the general thanpopulation,
to the general it can be shownitusing
population, can betheshown
L&M using
QM datathe that
L&MwhenQM
the overall
data l0 values
that when experienced
the overall byexperienced
l0 values the general by QPOP “shadow”
the general QPOPthose calculated
“shadow” from
those the L&Mfrom
calculated QM
registers
the L&MinQM such a manner
registers so that
in such in the first
a manner so decades
that in thethe first
QPOP l0 is higher
decades (the “convert”
the QPOP l0 is highereffect),
(the
but subsequently
“convert” effect), falls
but over time to approach
subsequently falls over parity,
time then the estimated
to approach parity,QPOP
then generates the number
the estimated QPOP
of burials recorded
generates the number in the registers
of burials (Table 3).
recorded in The
the modifier needed3).
registers (Table toThe
create appropriate
modifier needed shadowing
to create
appropriate
of L&M QM shadowing of L&M QMby
l0 can be approximated l0 can be approximated
ensuring that the CDR by(deaths
ensuring perthat
1000)theused
CDRin(deaths
the QPOPper
1000) used infirst
calculations thelags
QPOP andcalculations
then alignsfirst withlags
thatand thenby
noted aligns with that noted
demographers for thebygeneral
demographers for
population
the general
over population
the period (Figure over 30
6). the period (Figure 6). 30

The CDR used above for the Quaker population is generated by modifying the L&M QM l0 to
ensure it remains lower than that of the general population at the beginning of the period, an effect
which is removed in order for the Quakers to draw in line with overall demographics as the Quaker
population ceases to be made up of adult converts, and the impact of early-year mortality takes effect.

30 (Wrigley 1985); Figure 1.

29 Vann (1969, p. 48) His chapter on the “Theory and Practice of Conversion” uses data from Buckinghamshire
and Norfolk from 1654 to 1740.
30 (Wrigley 1985); Figure 1.
 Figure 6. Modified crude death rate (CDR) used for QPOP vs. general CDR (Wrigley).

The CDR used above for the Quaker population is generated by modifying the L&M QM l0 to
ensure it remains lower than that of the general population at the beginning of the period, an effect
which is removed in order for the Quakers to draw in line with overall demographics as the Quaker 11 of 15
Religions 2019, 10, 83
population ceases to be made up of adult converts, and the impact of early-year mortality takes effect.

Table
Table3.3.Reconciliation
Reconciliationof of
QPOP, crude
QPOP, death
crude raterate
death and and
QM burial registers
QM burial by decade
registers (1650–1809).
by decade (1650–1809).

Total Non- Estimated

Burials Non- QPOP
NM's Members Under- Burials L0
Projected Member Crude L0
Decade QPOP as % of buried recording of from L&M QM
from QPOP Burials Death Modifier
Total (not all Burials for Rowntree Burials
Recorded Rate
identified) Non-Members
1650 1,320 628 11% 80 - 80 - 709 47.6 21.0 100%
1660 25,150 6,406 2% 120 - 120 - 6,599 25.5 21.6 55%
1670 43,040 10,167 - - - - - 10,142 23.6 21.2 50%
1680 43,613 11,196 - - - - - 11,245 25.7 25.3 65%
1690 45,103 10,659 - - - - - 10,657 23.6 30.0 71%
1700 46,279 11,208 - - - - - 11,274 24.2 32.2 78%
1710 40,133 10,963 - - - - - 10,876 27.3 32.9 90%
1720 33,744 10,130 8% 900 - 900 - 11,016 30.0 33.3 100%
1730 24,811 7,591 14% 1,200 - 1,200 - 8,769 30.6 32.7 100%
1740 24,028 6,481 19% 1,500 - 1,500 - 7,925 27.0 37.1 100%
1750 24,000 6,114 11% 720 70 650 - 6,834 25.5 39.3 100%
1760 24,105 6,487 14% 1,046 196 850 - 7,514 26.9 37.2 100%
1770 24,091 6,645 14% 1,122 872 250 - 7,771 27.6 36.3 100%
1780 23,614 6,373 21% 1,701 1,701 - 8,161 27.0 35.2 95%
1790 22,141 5,693 23% 1,669 1,669 - 7,344 25.7 36.9 95%
1800 21,914 4,903 25% 1,628 1,628 - 6,503 22.4 38.0 85%
Total 121,645 plus 6,136 5,550 - 133,339
Religions 2019, 10, x FOR PEER REVIEW 11 of 15

Table
40.0 3 shows the application of this “modifying factor” to the L&M QM l0 from which is derived
a revised
35.0
crude death rate, which is applied to the QPOP estimate to generate a projected number of
burial records which approximate to the burials shown in Rowntree. As with the birth records,
30.0
Rowntree gives some figures for non-members buried for some decades: while such instances are
25.0
recorded from 1750, it seems likely that such a practice began earlier, and the separate recordkeeping
20.0 to take effect. Historians have not established the extent of the practice, but if Rowntree is
took time
correct15.0
in accounting for one-quarter of all burials as non-members in 1800–1809, it seems that the
10.0
practice was widespread. A more likely scenario is presented in the table, with the practice starting
5.0 around 1720, and increasing over the next decades until suppressed (temporarily) by the
in earnest
-
1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800

Crude Death Rate (QPOP) CDR

Figure 6. Modified crude death rate (CDR) used for QPOP vs. general CDR (Wrigley).

Table
The CDR 3 shows
usedthe application
above for the of this “modifying
Quaker populationfactor” to the L&M
is generated QM l0 from
by modifying which
the L&MisQM derived
l0 to
a revised
ensure crude death
it remains lowerrate,
thanwhich
that ofis the
applied
generalto the QPOP estimate
population to generate
at the beginning a projected
of the period, an number
effect
of burial
which records which
is removed in orderapproximate
for the Quakers to thetoburials
draw inshown in Rowntree.
line with As with the as
overall demographics birth
the records,
Quaker
Rowntree
populationgives
ceasessome
to befigures
made up forofnon-members
adult converts, buried
and theforimpact
some decades: while
of early-year such instances
mortality are
takes effect.
recorded from 1750, it seems likely that such a practice began earlier, and the separate recordkeeping
Table
took time to3.take
Reconciliation of QPOP,have
effect. Historians crudenotdeath rate and QM
established theburial
extentregisters by decadebut
of the practice, (1650–1809).
if Rowntree is
correct in accounting for one-quarter of all burials as non-members in 1800–1809, it seems that the
Total Non- Estimated
practice was widespread.Burials A more likely scenarioNon- is presented in the table, with the practice starting
QPOP
NM's Members Under- Burials L0
in earnest around Projected
1720, and increasing over the Member
next decades until suppressed Crude
(temporarily) by theL0
Decade QPOP as % of buried recording of from L&M QM
from QPOP Burials Death
revival of Quaker discipline around Total 1750. This speculation Burials
(not all allowsforthe trends burials toModifier
in non-memberBurials
Rowntree
Recorded Rate
identified)
make sense, without materially affecting the model. Non-Members
1650 1,320 628 11% 80 - 80 - 709 47.6 21.0 100%
1660 25,150 6,406
8. Swings and Roundabouts 2% 120 - 120 - 6,599 25.5 21.6 55%
1670 43,040 10,167 - - - - - 10,142 23.6 21.2 50%
It remains
1680 to be validated
43,613 11,196 if- the decadal
- fluctuations
- in the QPOP
- - estimates
11,245 can be explained
25.7 25.3 65%
by the 10,659 from
45,103data derived
1690modified - the QM- birth, marriage
- -
and death - registers.
10,657 In 23.6
order to 30.0
do so, 71%
1700 46,279 11,208 - - - - - 11,274 24.2 32.2 78%
1710 40,133 10,963 - - - - - 10,876 27.3 32.9 90%
1720 33,744 10,130 8% 900 - 900 - 11,016 30.0 33.3 100%
1730 24,811 7,591 14% 1,200 - 1,200 - 8,769 30.6 32.7 100%
1740 24,028 6,481 19% 1,500 - 1,500 - 7,925 27.0 37.1 100%
1750 24,000 6,114 11% 720 70 650 - 6,834 25.5 39.3 100%
1760 24,105 6,487 14% 1,046 196 850 - 7,514 26.9 37.2 100%
Religions 2019, 10, 83 12 of 15

three established but unquantified elements in the history of the Religious Society of Friends need to be
accounted for. The first is that of “convincements”—those who flocked to join the Society in the heady
millenarian days when the movement believed it was founding a new world church. The second
concerns those who left, either voluntarily as migrants to seek new lives elsewhere (many to Ireland
and others across the Atlantic), both before and after the founding of Pennsylvania in 1684. Finally,
there are the “apostates”, those who left once it was clear that the second coming of Christ was
increasingly delayed, and those who preferred to avoid the “sufferings” that the Quakers attracted
through refusing to conform with the prejudicial laws against sectarians. A subset of this group,
known as the “disowned”, represents those who were rebuked by the Society for not being in unity;
during this period the most common reason for being in such a state was “marrying out”, either with a
non-Quaker partner, or simply by the established church.
Considering the first of these, the model suggests that there must at least be some who came to the
Society married, and subsequently entered their children on the register. There also need to be many
who swelled the ranks in order for the CFMR to generate the remaining marriages. Accepted Quaker
historiography would see this as an increasing swell of converts during the last decades of the
seventeenth century, falling rapidly by the turn of the eighteenth century, yet retaining still some
appeal to attract new membership over the period. The “emigrants” class has been well-charted by
Quaker historians, although (perhaps again, unsurprisingly) the extent of migration remains disputed.
Traditionally, many thousands were believed to have left for the New World: Richard Vann cites
Joseph Illick’s estimate of 8000 by 1685, while noting that other historians have been “less venturesome
in committing themselves to figures”. Vann argues for a lower earlier total based on the number
of surviving Quaker removal “certificates”, giving about 1000 before 1685, and a further thousand
in the 1690s. The most recent full-length study of transatlantic Quakerism, curiously, if tellingly,
makes no attempt to address the subject.31 Rowntree has no doubts that it was a considerable number,
chalking up five hundred per annum between 1676 and 1700, and a “very considerable number”
afterward—unfortunately his source for this was an appendix to an amateur collection of statistics,
created in the mid-nineteenth century, and lacking provenance.32 Most scholars would probably
accept a number between eight and ten thousand over the period, tailing off sharply after 1700,
as uncontroversial. The third group are highly likely to have begun to leave the nascent Society almost
as soon as it began, towards the end of the 1660s after the chastening effect of the Great Fire of London
and the plague had failed to deliver a New Jerusalem. Many of those who had become Quakers
in search of personal betterment (of which there may have been many) are likely to have made the
transatlantic journey even without Quaker certification, or simply abandoned sectarian religion in the
spirit of an increasingly secular eighteenth century. Extant scholarship on the Quaker apostates tends
to focus on the trajectories of prominent individuals, which unfortunately fails to help statistically,
leaving this area very much in the dark. The final group—that of the “disowned”—is similarly lacking
in quantifiable data. The solution adopted here is to make an assumption that the missing marriages
until the formal marriage regulations of 1737 represent married couples joining the Society before
contributing their subsequent children to the registers, while those after this point increasingly include
couples who “married out”, and were disowned, before contributing theirs. While the accuracy of
such an approach can only be guessed at, it has the advantage of modelling quanta while fitting an
accepted interpretation of historical events which accepts significant disownments as a feature of the
period of Quaker Discipline which began around 1755.
Taken together, these flows produce a model which attempts to reconcile data derived from
known QM records with factors from wider demographic work, while accounting for changes in the
English Quaker population for each decade over a period of one hundred and fifty years.

31 (Landes 2015).
32 (Rowntree 1859, p. 74), Appendix to [John?] Thurnam’s “Statistics” page 12.
 Religions 2019, 10, x FOR PEER REVIEW 13 of 15
accepts significant disownments as a feature of the period of Quaker Discipline which began around
1755.
accepts significant disownments as a feature of the period of Quaker Discipline which began around
1755.Taken together, these flows produce a model which attempts to reconcile data derived from
known Taken
Religions 2019, together,
QM 10,
records
83 thesefactors
with flows from
produce a model
wider which attempts
demographic work, whileto reconcile data derived
accounting from13
for changes inofthe
15
known QM records with factors from wider demographic work, while
English Quaker population for each decade over a period of one hundred and fifty years. accounting for changes in the
English
TheQuaker population
projected for each
populations ofdecade over aare
the model period of one
shown in hundred
Figure 7,and fifty years.
against estimates from the
The projected populations of the model are shown in Figure
Rowntree approach. One of the first observations might be that the figure for thefrom
The projected populations of the model are shown in Figure 7, 7, against
against estimates
estimates from theof the
the Rowntree
beginning
Rowntree
approach. approach.
One of theOne of
first the first
observations observations
might be might
that be
the that
figurethe figure
for the
nineteenth century approximates Rowntree’s estimation of 19,800—not perfectly well, but rather for the
beginning beginning
of the of the
nineteenth
nineteenth
century
better than century
approximates approximates
Rowntree’s
using Rowntree’s Rowntree’s
own estimation estimation
projections from of
of 19,800—not 19,800—not
marriage.perfectly perfectly
well,in
Looking well,
but but
rather
more rather
better
detail atthan
the
better
using than using
Rowntree’s Rowntree’s
own projectionsown projections
from marriage. from marriage.
Looking in Looking
more detail in
at more
the detail at
consequences theof the
consequences of the assumptions above, some further points of comparison occur (Table 4).
consequencesabove,
assumptions of the assumptions
some furtherabove,pointssome further points
of comparison occurof comparison
(Table 4). occur (Table 4).

50,000
50,000
45,000
45,000
40,000
40,000
35,000
35,000
30,000
30,000
25,000
25,000
20,000
20,000
15,000
15,000
10,000
10,000
5,000
5,000
--
1650 1660
1650 1660 1670
1670 1680
16801690
16901700
17001710 1720
1710 1730
1720 1740
1730 17501750
1740 17601760
17701770
1780 1780
1790 1790
1800 1800

QPOP Rowntree 'Marriage*140'

QPOP Rowntree 'Marriage*140'

Figure 7. The projected QPOP model compared with Rowntree.

Figure
Figure 7. The
The projected
projected QPOP
QPOP model compared with Rowntree.
Table
Table 4.
4.English
EnglishQPOP
QPOPstock: convinced,
stock: migrant,
convinced, apostate
migrant, andand
apostate disowned flows.
disowned flows.
Table 4. English QPOP stock: convinced, migrant, apostate and disowned flows.
Decade QPOP Births Converts Converts Total Convinced QM Migrants Disowned Apostates
Decade QPOP Births Converts Not
less Already ConvertsConverts
Total <10Convinced
years% +/- QM All Migrants Disowned
Marriage Apostates
Remainder
Deaths Already Married
less Married Not All
Converts <10 years%P.A. +/-Regions not onMarriage Remainder
the Balancing
Deaths Married Married P.A. Regions not onItem
register the Balancing
register Item
1650 1,320 2,395 2,499 16,522 19,021 75.6% 73 0 - 0
1660
1650 25,150
1,320 663
2,395 2,452
2,499 16,851
16,522 19,302
19,021 44.8%
75.6%64 73 0
- 0 -
-2,767 0
1670
1660 43,040
25,150 - 389
663 2,340
2,452 6,456
16,851 8,796
19,302 20.2%
44.8% 5 64 -1,000
- 0 -
-6,456 -2,767
1680
1670 43,613
43,040 -- 2,034
389 3,749
2,340 5,670
6,456 9,4198,796 20.9%
20.2%21 5 -3,500
-1,000 -
- -436 -6,456
1690
1680 45,103
43,613 -- 1,527
2,034 1,865
3,749 5,187
5,670 7,0529,419 15.2%
20.9%17 21 -2,500
-3,500 -
- -226 -436
1700
1690 46,279
45,103 -- 2,200
1,527 1,048
1,865 1,111
5,187 2,1597,052 5.4%
15.2% -7 17 -1,000
-2,500 -
- -2,869 -226
1710 40,133 - 2,518 638 1,204 1,842 5.5% -5 -500 -2,609-
1700 46,279 - 2,200 1,048 1,111 2,159 5.4% -7 -1,000
- -2,869
1720 33,744 - 3,662 797 405 1,201 4.8% -6 -500 -441
-1,822
1710 40,133 - 2,518 638 1,204 1,842 5.5% -5 - -500 -2,609
1730 24,811 - 2,276 963 2,977 3,940 16.4% 14 0 -182
-124
1720 33,744 - 3,662 797 405 1,201 4.8% -6 -500
-441 -1,822
1740 24,028 - 2,381 1,431 3,124 4,555 19.0% 17 0 -191
0
1730
1750
24,811
24,000 -
- 2,276
1,256 688
963 2,977
2,160
3,940
2,848
16.4%
11.8% 10
14 -182
0
0
-352
0
-124
1740
1760 24,028
24,105 -- 2,381
1,504 1,431
490 3,124 3,383
2,893 4,555 19.0%10
14.0% 17 -191
0 0 -665
0 0
1750
1770 24,000
24,091 -- 1,256
1,185 292688 2,160
2,168 2,848
2,460 11.8% 7
10.4% 10 -352
0 0 -680
0 0
1760
1780 24,105
23,614 -- 1,504
1,344 - 490 2,893
1,889 3,383
1,889 14.0%
8.5% 5 10 -665
0 0 -661
0 0
1770
1790 24,091
22,141 -- 1,185
631 - 292 2,168 1,771
1,771 2,460 10.4% 4
8.1% 7 -680
0 0 -694
0 0
1780
1800 23,614
21,914 - 1,344
407 -- 1,889
- -1,889 8.5% -3
0.0% 5 -661
0 0 -756
0 0
1810
1790 22,141
21,600 - 631 19,252- 70,386
1,771 89,639
1,771 8.1% 4 -9,000 0 -4,622
-694-17,308 0
1800 21,914 407 - - - 0.0% 0
-3 -756 0
1810
First, the21,600
First, the number 19,252emigrants
numberofofprojected
projected 70,386is about
emigrants is89,639
about
halfhalf
that that of estimated
of estimated -9,000
apostates,
apostates, -4,622 a -17,308
suggesting
suggesting
aratio
ratioofofone
oneininthree
threeleaving
leavingto to
promote
promote Quakerism,
Quakerism, while the the
while remainder dropped
remainder awayaway
dropped for other
for other
First,
reasons—in the number
fact, the of projected
balance could emigrants
shift easilyis about
either half
way. that of
Second, estimated
all majorapostates,
movements
reasons—in fact, the balance could shift easily either way. Second, all major movements from emigrants suggesting
from a
ratio of
emigrantsone in
and three leaving
apostates to
togetherpromote
are Quakerism,
completed by while
the very the remainder
early eighteenthdropped
century,
and apostates together are completed by the very early eighteenth century, when the Society has the away
when for
theother
reasons—in
Society
least “new fact,
hasblood”:
the the
least balance
“new
over 95% of could
blood”:Friendsshift
over 95% ofeasily
would haveeither
Friends wouldway.
been haveSecond,
beenfor
convinced all major
convinced
more than movements
for more
ten than ten
years. from
This fits
emigrants and apostates together are completed by the very early eighteenth century,
the profile of the Society in Quietist mode, minimizing engagement with outsiders and with the world. when the
Society has the least
Third, following “new
a low blood”: over 95%
in convincements at of Friends
about would
the time thehave
last been convinced
apostates for morethere
were leaving, than is
tena
steady if not spectacular recovery, until around ten or twelve percent of the Society enters membership
every decade. This aligns somewhat with the religious “awakening” movements of the period, not least
Methodism and associated spiritual ventures. Finally, an estimate has been made of the actual annual
Religions 2019, 10, 83 14 of 15

impact these movements would have had on those in the Society’s Quarterly Meetings (see figures
in grey, centre of the table, Column QM +/− P.A.). With approximately twenty-five QMs across the
country, the aggregate experience would have been at first to see dozens of new Friends joining each
year, tailing off until a perhaps (net) half dozen or so would leave or migrate, before that too tailed off,
and by the middle of the century perhaps a dozen or half dozen might join. At the level of monthly
meetings, or even lower preparatory meetings, the effect would be highly individual (and sometimes
invisible) at all dates from 1700, with at most one or two joining or leaving each local meeting in
any year.

9. Conclusions
In thus examining Quaker numbers during the long eighteenth century, the original motivation
was to see if they could be made to “stack up”—that is, given sufficient manipulation, they could
be incorporated into a model which accounted for known data. This would have enabled some
conclusions to be drawn regarding the accuracy (or otherwise) of the registers over time, and allowed
some comments to be made regarding Rowntree. Having brought the model this far, it seems fair
to suggest that not only do the registers “stack up” but that estimates for the numbers of Quakers
in the Society can be projected with some accuracy, in populations which not only align to general
demographic indicators, but also reflect known dynamics within Quaker history.
Yet at no point do these projections become facts. Professor Peter H. Lindert’s warning that
the “uncritical consumption of detailed estimates can be critical to intellectual health” is ever valid,
and his desire to “replace old, rough, tentative guesses with new, rough, tentative guesses” remains an
aspiration.33 That said, there would appear to be more than sufficient correlations of a positive nature
to justify using these figures, while awaiting further work to augment the fundamental assumptions
within the model.
With respect to such further work, the model prompts new questions in several areas. The first
concerns the nature and extent of eighteenth-century discipline, specifically: what was the real extent
of those disowned by the Society? Previous explanations have required large numbers of Quakers
to be expelled in order to reduce numbers to known “Rowntree” levels, yet the model would tend
to suggest that such expulsions were minimal, and thus any resurgence of discipline was either
exaggerated, or perhaps less drastic in execution and consequence. A similar question regarding
numbers arises in the case of how many early Friends emigrated—as has been noted, there exists a
level of uncertainty regarding this traffic where clarity would be particularly useful. Finally, the model
assumes that the Society of Friends remained able to attract a number of converts throughout the
period—albeit a relatively small number when compared with the awakenings that would go on to
establish Methodism. Much useful work can be done on the nature of the beliefs of this membership
compared with those convinced earlier, as well as perhaps on any differences between such new
members and birthright Friends which might help account for later developments within the Society.
Finally, we might do well to remind ourselves that the two Biblical attempts to number the
Children of God from “Beersheba even to Dan” differ by a significant margin: while we may hope to do
better, this should serve to emphasize the enduring difficulties associated with historical reconstruction
of population data.34

Funding: This research received no external funding.

Conflicts of Interest: The author declares no conflict of interest.

33 (Lindert 1980, p. 704).

34 cf. II Samuel 24 and I Chronicles 21. 1,300,000 versus 1,570,000, or approximately 20%.
Religions 2019, 10, 83 15 of 15

References
Bortescue, G. K., ed. 1908. Catalogue of the Pamphlets ... Collected by George Thomason, facs. ed. London, Nedeln:
Online Library of Liberty, vol. 2, pp. iii–xxv.
Braithwaite, William. 1912. Beginnings of Quakerism. London: Macmillan, p. 512.
Cole, W. Alan. 1957. The Social Origins of the Early Friends. Journal of the Friends’ Historical Society 48: 99–118.
Davies, Adrian. 2000. The Quakers in English Society, 1655–1725. Oxford: Clarendon Press, p. 46.
Dixon, Ruth B. 1971. Explaining Cross-Cultural Variations in Age at Marriage and Proportions Never Marrying.
Population Studies 25: 215–233. [CrossRef] [PubMed]
Firth, Charles Harding, and Robert Sangster Rait, eds. 1911. Acts and Ordinances of the Interregnum, 1642–1660.
London: His Majesty’s Stationery Office, pp. 184–218.
Frost, J. William. 1993. Review Friends in Life and Death: The British and Irish Quakers in the Demographic
Transition, 1650–1900. Quaker History 82: 45–47. [CrossRef]
Galloway, Patrick R. 1988. Basic Patterns in Annual Variations in Fertility, Nuptiality, Mortality, and Prices in
Pre-industrial Europe. Population Studies 42: 275–302. [CrossRef] [PubMed]
Gilbert, Alan. 1976. Religion and Society in Industrial England: Church, Chapel and Social Change, 1740–1914. London:
Longman, p. 36.
Goldstone, Jack A. 1986. The Demographic Revolution in England: A Re-examination. Population Studies 40: 5–33.
[CrossRef] [PubMed]
Greenberg, Stephen J. 1988. Dating Civil War Pamphlets, 1641–1644. Albion: A Quarterly Journal Concerned with
British Studies 20: 387–401. [CrossRef]
Healey, Robynne, ed. 2019. Quakerism in the Atlantic World during the Long Eighteenth Century. University Park:
Penn State.
Healy, Robynne Rogers. 2006. From Quaker to Upper Canadian: Faith and Community among Yonge Street Friends,
1801–1850. Montreal: McGill-Queen’s Press—MQUP, p. 32.
Isichei, Elizabeth. 1970. Victorian Quakers. Oxford: OUP, pp. 112–113.
Landes, Jordan. 2015. London Quakers in the Trans-Atlantic World, the Creation of an Early Modern Community.
London: Palgrave Macmillan.
Lindert, Peter H. 1980. English Occupations, 1670–1811. The Journal of Economic History 40: 685–712. [CrossRef]
Rowntree, John Stephenson. 1859. Quakerism, Past and Present. London: Smith, Elder & Company,
Preface. Available online: https://archive.org/details/quakerismpastan02rowngoog/page/n4 (accessed on
20 January 2018).
Schellekens, Jona. 1997. Nuptiality during the First Industrial Revolution in England: Explanations. The Journal of
Interdisciplinary History 27: 637–654. [CrossRef]
Smith, Richard M. 1981. Fertility, Economy, and Household Formation in England over Three Centuries.
Population and Development Review 7: 595–622. [CrossRef]
Vann, Richard T. 1969. Social Development of English Quakerism. Harvard: HUP, pp. 159–162.
Vann, Richard T., and David Eversley. 1992. Friends in Life and Death: The British and Irish Quakers in the Demographic
Transition, 1650–1900. Cambridge: CUP.
Wells, Robert V., and Michael Zuckerman. 1972. Quaker Marriage Patterns in a Colonial Perspective. The William
and Mary Quarterly 29: 415–442. [CrossRef]
Wrigley, Edward Anthony. 1998. Explaining the Rise in Marital Fertility in England in the ‘Long’ Eighteenth
Century. The Economic History Review 51: 435–464. [CrossRef]
Wrigley, E. A. 1985. Population Growth: England, 1680–1820. Recent Findings in Research in Social and Economic
History, 1–5, (ReFresh Issue 1 Autumn).
Wrigley, Edward Anthony, and Roger S. Schofield. 1983. English Population History from Family Reconstitution:
Summary Results 1600–1799. Population Studies 37: 157–184. [PubMed]

© 2019 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).