Nursing Under the Old Poor Law in Midland and Eastern England –

This article uses data drawn from the overseers’ accounts and supporting documentation in thirty-six parishes spread over four English counties, to answer three basic questions. First, what was the character, extent, structure, range of activities, and remuneration of the nursing labor force under the Old Poor Law between the late eighteenth century and the implementation of the New Poor Law in the s? Second, were there regional and intra-regional differences in the scale and nature of spending on nursing care for the sick poor? Third, how might one explain such differences? The article suggests that nursing became an increasingly important category of spending for the poor law from the later eighteenth century, but that there were significant variations within and (particularly) between English counties in parochial attitudes toward the provision of nursing for the sick poor. These variations can be explained by applying a matrix of explanatory variables ranging from the minor (differences in how parishes defined “nursing”) through to the major (long-standing cultural attitudes toward the responsibility of parishioners to their sick compatriots and the ingrained expectations of the sick poor). The article also throws new light on the hidden aspects of female labor force participation, pointing to the development of professional nursing networks long before the later nineteenth century.

western England than was the case in the south and east.  For the purposes of this article, however, three observations about the nature of the Old Poor Law are important. First, because the government set no entitlement criteria or notional norms for payment, ecclesiastical parishes and their officials were autonomous bodies. They could choose to raise more or less local tax to fund welfare of all sorts, and so there was no inevitable link between the relative wealth of a parish and the amount it paid out in welfare. Indeed, as we will see below, the relationship was often inverse. Second, while parishes and their officials were autonomous, they were not immune from the accretion of customary practice, the intervention of powerful advocates for the poor or rising medical standards which would in turn feed through to the expectations of the poor. Finally, from the early eighteenth-century until its abolition in , the Old Poor Law was by far the biggest non-familial provider/funder of nursing care in England and Wales.  Understanding nursing in this period is thus indivisible from understanding the practice and sentiment of the poor law.
Some work in this area has pointed the way forward. Jeremy Boulton's study of the parish of St Martins-in-the-Fields from the late seventeenth-century to the s, for instance, suggests that between  and  percent of total parish spending, as well as almost one-third of all "extraordinary" allowances, was absorbed by a small cohort of nurses most of whom provided care in de facto nursing homes.  Such arrangements were short-lived-the foundation of the parish workhouse made nursing less lucrative and more transient  -but the study provides a glimpse of the potential dynamism of nursing under the early Old Poor Law. Against the backdrop of the spiraling cost of poor relief from the s (the so-called "crisis of the Old Poor Law"); however, historians have generally had both little and little positive to say about poor law nursing. Samantha Williams's study of two small Bedfordshire parishes (Campton and Shefford) identifies  "carers" who were retained for broad nursing duties. Almost all were employed on an irregular basis and poorly paid, with the majority themselves becoming dependent upon poor relief at some point in their life cycle.  Anne Borsay suggests that eighteenth and early nineteenth-century community nursing (including but not confined to poor law nursing) "was less abject than conventional wisdom alleges" but can draw on little empirical material relating to the Old Poor Law and ultimately concludes that it was institutional nursing that drove the professionalization process.  Unsurprisingly, the image of the poorly paid, untrained, ineffectual, and amateur (often pauper) nurse has become a leitmotif in relation to the medical care offered by the poor law. This characterization can be tested through the accumulation of individual microstudies, but when these are set against marked differences in intra-and inter-regional support for the sick poor more generally,  the need for a wider frame of reference is clear. The article thus analyses data (overseers' accounts, pauper letters, vouchers, vestry minutes, and overseer correspondence) drawn from thirty-six parishes in Berkshire, Norfolk, Northamptonshire, and Wiltshire between the s and s, constituting by far the largest study of its kind.  I will suggest that spending on nursing care varied between parishes within counties, but even more keenly between those counties. These differences reflected variations in the propensity of parishes to engage nurses, the nature and duration of nursing care provided, and the pay and professional status of the nurses engaged. These proxy variables in turn reflect a matrix of causative relationships (explored fully in the conclusion) which range from the minor (parishes defined and thus recorded "nursing" in different ways), through the significant (variable experiences of epidemic disease) and to the systemic (accumulated local cultures of nursing). Using brief characterizations of comparator parishes, I will suggest that the causative matrix has little place for either variations in wealth or socio-economic structure. In turn, analyzing the data for this substantial set of parishes helps to draw back the chronological focus of recent revisionist work on nursing history and offer a more refined understanding of issues such as the role of the Old Poor Law as a care-giver, parochial responsiveness to demographic shocks, women's labor market position and remuneration, and communal sentiment toward the poor.

D O I N G A N D D E F I N I N G
Collectively, the four counties providing data for this study sit within the broad arc of southern, midland, and eastern England where poor relief measured in money terms was at its most generous from the later eighteenth century. They provide coverage of the three major agrarian production regimes (arable, pastoral, and mixed farming), and all lack the sorts of rapidly growing urban areas, at least until the end of the period covered here, that tend to skew understandings of the role and cost of welfare in other areas. At the same time, all four counties can boast a complex socio-economic patchwork running from the very smallest villages, through areas experiencing long-term industrial decline and to communities at the forefront of rapidly expanding workshop-based industries such as shoemaking and ribbon weaving.
At the level of the individual county, selected parishes are broadly comparable in terms of the presence and longevity of workhouses. On the other hand, I have sought to include material from communities across the size distribution-the range for Wiltshire was between  and , while that for Berkshire was between  and , for instance  -topographical range, migration experience, and wealth spectrum appropriate to each county. The county sample as a whole is underweight in large urban areas, but the data provide an opportunity to explore the cost and character of nursing across a typological spectrum from the smallest . In , the collective parish population sizes were  in Berkshire,  in Northamptonshire,  in Norfolk, and  in Wiltshire. hamlet to small market and industrial towns. To test the assumption in much poor law historiography that the level, form, and distribution of welfare resources were intrinsically related to the wealth or socio-economic composition of individual communities, I have conducted five separate analytical sweeps through different subsets of the intra-and inter-county data: the whole sample; comparing only communities with similar socio-economic and wealth complexion; topographical groupings; population bands; and communities segmented according to industrial or agrarian production regimes. Systematic differences of practice and experience emerging from these different configurations of the data are reported in the analysis that follows, though as a general observation that the failure of socio-economic typology or wealth levels to offer predictive value in terms of the scale, duration, and complexion of nursing provided is striking. By way of example, the five wealthiest and largest parishes (measured by the absolute size of the local tax base) in the sample were Thetford (Norfolk), Caversham (Berkshire), Trowbridge (Wiltshire), Thatcham (Berkshire), and Wellingborough (Northamptonshire). All were broadly urban but none feature in a hierarchy of the top ten largest expenditures on nursing either in absolute or percent of medical relief terms. Moreover, the differences in spending between wealthier urban communities in Norfolk and exactly comparable communities in Berkshire were profound.  From the records relating to these parish samples, it is possible to trace the proportion of all welfare spending that was dedicated to relief of the sick poor, and then more narrowly to focus on a comparative analysis of relative and absolute expenditure on nursing, the character of the nursing labor force (pay, careers, etc.), and the attitude of parochial authorities to nursing within and between counties. While the data are not, and not meant to be, as rich as that deployed in individual microstudies, our analysis provides a wider frame of reference in which to inscribe future studies of this sort.
Initially, however, and in common with other authors, it is necessary to explore several key definitional and practical problems. Failure . A detailed list of parishes and associated archives is available in the Leicester Research Archive.
Journal of the History of Medicine : Vol. , October  to engage with eighteenth and early nineteenth-century poor law nursing is at least partly explicable in terms of ambiguities over how to define what a nurse was.  Can "nursing" be distinguished from the medical attention provided by practitioners, both men and women, regular and irregular? Samantha Williams, following the early work of Margaret Versluysen, certainly thinks so, finding little curative as opposed to palliative care by nurses in Bedfordshire.  Yet, the functional borderland between nurses and (regular or irregular) practitioner was by no means clear even in the census era.  Eighteenth and early nineteenth-century doctors, midwives, wise people, and apothecaries frequently "nursed" their patients in terms of sitting and watching, administering medicines, staying in the house of the sick person, or even taking the sick into their own homes.  In turn, it is sometimes difficult to make a distinction between provision of nursing services and membership of an occupational group of "nurses."  How often did a person have to provide nursing services to be regarded as a nurse, when we know that nursing in this period was provided across a spectrum of engagement, ranging from the full-time professional who might in effect run a nursing home to those providing ad hoc nursing? If a parish turned regularly to a defined group of people, or if they maintained a national register of people who might be called upon at need, we might justifiably call such people a reserve pool of nursing labor, to all intents and purposes "nurses," even if they were not active at any given point. If, however, a parish turned . Margaret Pelling, The Common Lot: Sickness, Medical Occupations and the Urban Poor in Early Modern England (London: Longman, ), -; Sara Mendelson and Patricia Crawford, Women in Early Modern England (Oxford: Oxford University Press, ), - and -. Also Borsay and Hunter, "Nursing and Midwifery," .
. Not least because, unlike other European countries, religious organizations played relatively little part in providing nursing care for the poor in eighteenth-and early nineteenthcentury England. See Colin Jones, The Charitable Imperative: Hospitals and Nursing in Ancien Regime and Revolutionary France (London: Routledge, ); Susan Dinan, Women and Poor Relief in Seventeenth-Century France: The Early History of the Daughters of Charity (Aldershot: Ashgate, ). Wildman, "Change," , suggests that there were  religious groups offering nursing care by , though there was a distinct London focus to their activities. to a nurse because they were physically closest to the sick person, because the sick person had asked for them, or because poor law officials required that nursing be provided as a condition of, supplement to, or alternative for poor relief, the label nurse might be inaccurate at best and actively misleading at worst.
A second and related problem is how to reconstruct what nurses, however defined, actually did? Margaret Pelling argues that sixteenthcentury epidemics redefined the nature and content of nursing, necessitating more precise terminology to describe what was actually involved, such that fine distinctions between "keeping," "helping," "watching," and "nursing" were created.  On the other hand, overseers' accounts for the eighteenth and early nineteenth centuries are replete with entries where the detailed composition of a bill for "nursing widow x" might comprise carrying, washing, mending, childcare, nursing, attending, "doing for" someone, and laying-out the dead body all in one consolidated block. This suggests that officials maintained a very expansive definition of the duties associated with the label "nurse."  Accepting this loose definition raises a third set of issues about the visibility (and hence "countability") of different sorts of nurses and nursing services. The tasks of washing, mending, and generally "doing for" were most likely to be required and would be most extensive and expensive, for those least inscribed into kinship and neighborhood networks. Communities with transient populations would thus have the highest bills for nursing of this sort, while more representative (and less inclusive) nursing packages for those cemented more firmly into their communities would by definition be less visible.  Similarly, nurses whose activities straddled parish boundaries would be less visible, while those providing nursing for chronic cases would tend to be more so. Equally, medical charities founded to work in conjunction with, or as a substitute for, poor law support were often small-scale and transient. Their records thus tend to be patchy, but the effect of these organizations on the nature of local . Pelling, Common Lot, -, , and -. . For Williams, "Caring," , generalized entries encapsulated by the phrase "doing for" comprise the "vast majority" of care work. In the sample used here this is not the case, with "nursing" as a phrase and expenditure item appearing much more commonly and deliberately.
. For a discussion of this point, see Fissell, Patients, Power, .
poor law provision and the sorts of duties undertaken by nurses might nonetheless be profound.  This article treats the services of regular and irregular medical providers (even midwives unless they were paid extra for nursing) as separate from "nursing care." At the same time, it makes a distinction in quantitative analysis between regular "nurses" and the providers of ad hoc "nursing services." I follow overseers themselves in adopting a wide and reflexive definition of nursing in calculating expenditure levels, including services such as washing and laying-out as well as generalized bills related to "doing for" or the boarding of adults and children. There were in each county specialist institutions, such as voluntary hospitals, to which parishes subscribed and in which nursing would have been the greatest element of medical care. Subscriptions have not been ascribed as "nursing" expenditure because of inconsistent information on which, if any, paupers benefited.

QUA N T ITAT IV E OV ERV I EW
The thirty-six parishes (collective population in , ,) spent . percent of their total welfare resources directly on "medical relief," rising to . percent where all cash payments to people on longer-term relief who were intermittently sick are counted.  Irrespective of socio-economic typology or relative wealth levels, this proportion increased over the period from the s to s. Figure  traces the absolute amounts spent on nursing in the counties, offering important initial observations to frame this article.
Together the parishes spent very considerably less on nursing annually than the single London parish of St Martin's, with its roughly similar population, earlier in the eighteenth-century. The distinctive . Anne Summers, "The Cost and Benefits of Caring: Nursing Charities c. -," in Medicine and Charity Before the Welfare State, eds. Jonathan Barry and Colin Jones (London: Routledge, ).
. The category "medical relief" is contingent. Williams, "Caring," -, includes food, allowances, direct medical relief, and subscriptions to hospitals. Eric Thomas, "The Treatment of Poverty in Berkshire, Essex and Oxfordshire -," (Unpublished PhD, University of London, ), , calls for a more expansive definition. In this sample, almost two-thirds of all spending would become "medical relief" if we were to follow the emerging historiography of the aged poor, which has come to link their relief to progressive disability and inability to labor. For the current study, medical relief is defined to include nursing, doctoring costs, drugs, cash allowances, food and fuel, rent, funerals, and short-term allowances to the relatives of deceased paupers. The costs of building dedicated medical facilities in specific parishes are excluded.
King : Nursing Under the Old Poor Law  role usually ascribed to nineteenth-century London in the professionalization of nursing by historians clearly has much deeper roots than is often allowed. There were, however, distinct differences between counties in absolute amounts spent. For much of the period, Norfolk parishes were spending less than one-quarter of their Berkshire counterparts on nursing, while in the later part of the period both Northamptonshire and Wiltshire parishes were spending at least double that of Norfolk parishes.  There was, as suggested above, no systematic tendency for wealthier communities to spend more on medical relief and nursing or for areas with particular socio-economic complexions to spend more than others. Comparing similar agrarian parishes in Norfolk (Salhouse, Shottesham, and Bessingham) and Berkshire (Brimpton, Drayton, and Shellingford) confirms the spending disparities between the two counties reported in figure  and points to fundamental differences of practice and policy rather than simply The need for continuous time-series and the chronological cutoff points for this study impact inclusion of larger urban communities in particular. In Berkshire, for example, I have excluded small parishes such as Hinton Waldrist but also much more important places such as Hungerford, Reading, Englefield, and Tilehurst. differences in the socio-economic complexion of the county samples or wealth levels.
The volatility in absolute spending levels, particularly outside Norfolk, is striking. Relatively, few parishes paid for nursing services in every single year of the period, but the importance of nursing as a spending category could increase exponentially in some periods. In part, this reflects the specific impact of epidemics  and of the susceptibility of spending to the particular and often short-term needs of problematic individuals or families. In the small rural parish of Welton (Northamptonshire), for instance, periodic visitations of smallpox, typhus, and measles could increase spending by  percent compared with an average of the prior three years. This sort of microexperience is reflected in figure , which traces the peaks of spending on nursing at an individual parish level. In Berkshire, the town of Wallingford spent . percent of its medical relief on nursing in  (when the town was visited by smallpox and measles, with more than  cases of severe illness), while in Northamptonshire the small rural parish of Whiston spent . percent in . For Norfolk, the highest parish figure recorded in this sample was  percent for Scottow in , while in Wiltshire, Sopworth spent  percent of its medical relief resources on nursing in .  In most places and at most times the scale of nursing expenditure was not linked systematically to that on doctors, drugs, cash allowances, etc. Nursing was, in other words, a discrete service, a point to which I return below.
An analysis of the percentage of all relief and all medical relief spent on nursing, facilitates wider comparison. Both James Stephen Taylor (for Berkshire) and Samantha Williams (for Bedfordshire) have traced a limited place for nursing in the total poor law budget, and a relative decline in spending on nursing over time. Taylor suggests that the proportion of poor law resources devoted to nursing in seven Berkshire parishes fell from a slim . percent in the s to a tiny . percent by the early s. Williams see a similar linear decline for Shefford (from a more healthy . percent in the s . Although Eric Thomas, "The Old Poor Law and Medicine," Med. Hist., , , -; , argues that the poor were generally inoculated en mass and hence better protected from smallpox at least than the general population.
. Figure  provides tantalising evidence that absolute spending levels on nursing might have been rather higher earlier in the eighteenth century, but the records are not of adequate quality to systematically explore this matter on the scale needed.
to  percent by the s) and a slightly more varied trajectory for Campton (where the figure fell from  percent in the s to . percent in the s before rising and then falling to just . percent in the s).  In the thirty-six parishes analyzed here, the picture is rather more complex. Only in the Berkshire urban parishes of Caversham (. percent) and Thatcham (. percent), the rural Wiltshire parish of Hardenhuish (. percent), and the small rural Northamptonshire parish of Thorpe Achurch (. percent) did nursing absorb more than  percent of the total poor law budget for the period as a whole. At county-level, Berkshire devoted the highest percentage of its total welfare spending to nursing over the period (. percent), while the equally rural and equally wealthy Norfolk devoted the least (. percent). A systematic comparison of subsets of the parishes within and between counties reveals no tendency for wealthy or very poor parishes to devote higher or lower proportions of total spend to nursing. Equally there is no evidence of a systematic patterning to spending between communities of different socio-economic types. The small Northamptonshire village of Crick, for instance, devoted  percent of total spending to nursing while the much richer proximate village of Stoke Albany spent less than . percent of its funds in this manner. The use of total spending as a numerator is, however, problematic. Not all spending recorded in the year-end totals for "poor relief" by overseers was on the poor. Account books frequently indicate spending heads such as contributions to the county-rate, legal disputes, and interest on borrowed money. Moreover, it is imperative in interpreting these figures to allow for differential inflation. In most of the parishes analyzed here, the costs of the different elements of medical relief broadly defined (drugs, midwives, average cash allowances for sickness, etc.) saw modest inflation over the period, whereas the costs of rents, cash allowances for the underemployed, coals, and other generic support increased significantly faster. In other words, differential inflation inevitably ate away at the share of medical relief (and, within this category, potentially nursing) in overall relief for the parishes and counties analyzed.
Arguably, therefore, focusing on the proportion of parochial medical relief devoted to nursing affords a better understanding of underlying attitudes. The results of such an exercise (figures -) are striking. An un-weighted whole-period average for Berkshire of . percent contrasts markedly with that for Norfolk of . percent (and to a lesser degree Northamptonshire and Wiltshire at . and . percent, respectively), suggesting that the differences between the two counties in terms of absolute spending reflect very real underlying differences in attitudes and experiences. These county differentials are maintained or even amplified when we investigate parochial subsets King : Nursing Under the Old Poor Law  devised to control for wealth and socio-economic typology. Thus, the two urban and relatively wealthy parishes of Thetford in Norfolk spent less than  percent of parochial medical relief on nursing whereas the directly comparable Berkshire parishes of Caversham and Thatcham spent  percent.
There were, however, strong intra-county variations in the importance of nursing as a spending category. A detailed consideration of trends for all thirty-six parishes suggests three models of parochial nursing culture. In Norfolk, for instance, we encounter very rural parishes like Scottow ( population ) or Bressingham ( population ) the former with its sometimes substantial but always at least low-level spending on nursing and the latter with nursing the most common form of relief next to cash. For such communities  Journal of the History of Medicine : Vol. , October   nursing was clearly written into the very fabric of poor relief throughout the period. Other counties had their equivalent communities. In Wiltshire, for instance, both Sopworth and Hardenhuish spent more than  percent of their medical relief on nursing across the period. The small Northamptonshire parish of Thorpe Achurch consistently spent more than  percent. While most parishes falling into this first typology were rural or rural-industrial in character, towns too could see nursing written into the fabric of poor law expectations. The urban parish of Trowbridge, for instance, spent less ( just over  percent per year), but it did so consistently across the period. A second model is exemplified by the Norfolk parish of Thetford St Cuthbert ( population ) or the Berkshire parish of Wallingford where spending on nursing was more substantial than elsewhere when it did occur but also more intermittent. To these typologies, we might add a third, represented by the parishes of Salhouse in Norfolk ( population ) or the town of Wellingborough in Northamptonshire ( population ). In the former, nursing as a category of relief disappears altogether from , while in the latter such spending was at best patchy.
In short parochial commitment to nursing as a form of medical relief ranged on a spectrum between fragile and invulnerable in the last decades of the Old Poor Law. Higher absolute spending levels in Berkshire and lower levels in Norfolk in part reflected the distribution of the parochial sample across this spectrum. Crudely, Berkshire had King : Nursing Under the Old Poor Law  more parishes with a consistent and consistently high level of spending across the period than did Norfolk. This remains the case even where we compare like sub-groups of communities, an observation which points to the need to explore a wider range of potential explanatory variables (the composition, scale and remuneration of the nursing labor force within and between counties; the nature of ingrained official attitudes toward nursing; and perhaps to differential expectations on the part of the poor) than existing historiographical focus on the importance of wealth and socio-economic differentials in explaining poor law spending allows.
The process of dissecting poor law accounts  in order to trace the discrete number of nurses employed over the period considered here is hampered by frequent changes of marital status/name/alias and by the tendency for some overseers to intersperse their otherwise detailed record keeping with entries like "nursing for a stranger sick at Widow Hampton's." Such entries can be found periodically across the sample, often relating to nursing that was of short duration. Table  consequently provides a minimum quantitative overview of the scale of the nursing labor force in the four counties. These numbers begin to contextualize the county-level and intra-county differences in resourcing for nursing highlighted above.
Norfolk parishes employed only around half the number of nurses than their Berkshire counterparts, something that is unlikely to be explained by differential sickness rates and which is consistent across all of the typological subsets of the parochial data. At . nurses employed per parish per year, the Berkshire figures are somewhat higher than Samantha Williams found for Campton (.) or Shefford (.) though the close fit of Norfolk to these Bedfordshire figures begins to suggest a wider "eastern England" model of poor law nursing.
As we might expect from the three nursing typologies outlined above, each of the counties demonstrated a range of parochial policies on a spectrum from the employment of many nursing staff to places . Local officials kept a set of accounts detailing payments made to individual welfare recipients and to suppliers of services, including nurses. The accounts were considered annually by a larger group of ratepayers to ensure that financial probity had been maintained.
where employment was much more irregular and short-term. This patterning is not easily explained in terms of simple urban-rural or industrial-rural divisions within county boundaries. For Berkshire, it is certainly true that the town of Caversham employed the most nurses while the rural parish of Shellingford employed the least. Other relatively wealthy urban communities like Thatcham, however, employed relatively few nurses, while rural parishes like Brimpton were a mere step away from the figures exhibited by Caversham. In Northamptonshire, the relationship was even less distinct. The relatively wealthy rural parish of Stoke Albany employed the fewest nurses (followed by the industrial town Wellingborough), while the relatively poor proto-industrial parish of Woodford Halse provided the most. Other predictive variables may have more purchase. It is likely, for instance, that parishes occupied very different positions in the local and regional medical marketplace, with Stoke Albany employing few nurses because it was constrained by lack of supply in its very rural situation, something that might also apply to Thatcham. Differences in the scale of nursing between and within counties might also reflect strong variations in the range of nursing duties that communities would support, in turn a potential function of differences in underlying attitudes toward the sick, a subject revisited below.
More detailed consideration of the evidence suggests that the bald figures of table  and figures - are underpinned by marked spatial differences in the composition of the nursing labor force. Some Note: Sample includes only those for whom a definite identity could be established and thus excludes cases of generalized "nursing". Absence of family reconstitutions as a control on nominal linkage within overseers' records might lead to a small overstatement of the number of discrete nurses, though in turn most family reconstitutions yield only a relatively small core of definitively linked families and individuals. parishes were particularly likely to engage (relatively expensive) male nurses, though again the relative wealth of parishes appears not to have been a major factor in the patterning of such engagement. Burghfield, a relatively wealthy Berkshire parish expending  percent of medical relief on nursing employed John Hudson, Ambrose Allen, Thomas Benham and William Garrard for nursing services between the s and s. Garrard in particular was a regular and well-paid nurse who took patients into his home and was paid accordingly.  In Welton, a relatively poor Northamptonshire parish, the vestry appointed eleven male nurses between  and . By the early s, Henry Ball was the main provider of nursing services in the parish and was paid significant sums for his work, including extras and gratuities in recognition of the standard of his care.  In both places, employment of male nurses generated larger nursing bills than in comparable parishes within and between counties. Across the sample as a whole, however, employment of male nurses appears to have been more uncommon (< percent of all nurses and carers) than Samantha Williams has suggested for Bedfordshire, where  percent of all nurses were men. Moreover, it is important to recognize that male nurses were disproportionately involved in the singular task of caring for the insane poor. In Caversham for instance, three of the five male nurses employed over the period were engaged to deal with the insane John Bearfield.  For some parishes, then, variations in the incidence of insanity and the propensity to employ male labor as a consequence begin to explain differences in the scale of spending on nursing. Moreover, the fact that insanity was essentially a random event at parochial level explains at least partly why we see so little association between levels of wealth and the scale of nursing expenditure.
This said, the vast majority of nurses across the sample and period were female and the vast majority of nursing tasks had little to do with insanity.  Record linkage within the corpus of material used for this article allows us to say more about this female labor force . Berkshire Record Office (hereafter BRO) DP//-, Overseers' accounts. . Northamptonshire Record Office (hereafter NRO) p/-, Overseers' accounts.
. BRO DP//, Overseers' accounts. . Williams, "Religion, Respectability," . Davies, "Female Healers,"  also implicitly assumes that all nurses were female. and about some of the other potential explanations for county-level and intra-county differences in spending. Figure  traces the proportion of nursing events that can be ascribed to certain kinds of nurse. The categories employed are more numerous than those devised by Samantha Williams for Bedfordshire,  reflecting the more complex nursing structures and experiences that emerge when we consider multiple parishes across county boundaries. The numbers have several weaknesses, including the fact that the nursing typologies were not always discrete (see note to figure ) and that record linkage is (even with family reconstitution data) overwhelmingly probabilistic and hence subject to error. At best, therefore, figure  provides a broad overview. This notwithstanding, an initial reading of the data is that in attempting to understand differences in parochial spending patterns within and between counties we must distinguish between professional or semi-professional nurses within parishes, and those who made a living by providing nursing services across a wider ambit.
As Samantha Williams has observed, few parishes were without a person taking up multiple nursing roles within any given year or time period. Figure  suggests, however, that "professional" parish nurses in our sample were more common than in Bedfordshire, particularly for Berkshire and Northamptonshire.  Such, for instance, was Mary White of Ardington, in Berkshire, who nursed three to eight patients a year through s and early s, garnering almost two-fifths of all paid nursing assignments from Ardington poor law officials in this period. Moreover, just as Berkshire appears to have had a larger cohort of professional nurses than other counties, so the range of their activities was also wider. Lucy Goodwin of Ardington provided regular midwifery services, medicines, boarding for sick parishioners and strangers, and general nursing services to a substantial core of the sick poor in the parish. While she was never called a surgeoness, her activities suggest that . Williams, "Caring,"  talks of three types of carer. Also Fissell, Patients, Power, , , and .
. Williams, "Caring," . Focusing on training and status rather than simply on the intensity of their work, Borsay and Hunter, "Nursing and Midwifery," , suggest that "professionalization" was a product of the later nineteenth-century. Because "professional" is a perceived as well as an ascribed and claimed status, I use the term here to reflect parochial employment policies which might indicate trust, skills, or other aspects of worth.
she fulfilled the role.  In the same parish, Dame Williams provided lodging for the sick, children in particular, and general nursing and caring services for the chronically sick and for those lodged by the parish with other families. Women such as these were occupied as nurses rather than simply providing nursing Hence, maternity nurses could also be professional inter-parish nurses, for instance. We have tried to control for such instances by assigning such women to the professional category as they were professional nurses first and maternity nurses second. Other overlaps have been similarly addressed. Contingent nursing was that provided as part of a package by doctors or in designated institutions. Ad hoc nursing is that offered by carers who were named at least twice. Journal of the History of Medicine : Vol. , October  labor. While the relationship was not always consistent at the parish level, judged in the round counties and communities, like Ardington or Caversham for instance, that witnessed greater reliance on professional nurses were also those devoting larger proportions of medical relief to nursing.
In turn, we must distinguish the "professional" group from a smaller but nonetheless very important subset of nurses who traveled between parishes on nursing assignments. While this sort of activity has often been most closely associated with nursing of epidemic disease cases (especially smallpox, for which specialist nurses were paid a risk premium  ), it is clear that officials were also willing to engage such nurses for other aspects of sickness too. In Berkshire, for instance, Mrs Mary Hatt was engaged by at least four of our parishes in the late s, while other Hatt women were to be found as midwives and nurses in almost all of the Berkshire parishes analyzed, a de facto nursing dynasty. Overseers generally made a clear distinction between the role of these extra-parochial nurses and the more proximate nursing labor force. In  at Burghfield (Berkshire) Elizabeth Money, a woman who had provided ad hoc nursing services in the past, was paid s "for attending on Sarah Turner after the [professional] nurse was gone."  For all of the counties analyzed here, there is clear evidence of the existence of a professional extra-parochial nursing network. Even in Norfolk, rural parishes such as Forncett St Peter routinely brought in nursing labor from outside.  There was in turn a positive relationship at the parish level between a tendency to employ extraparochial nursing and the absolute and proportionate spending on nursing, helping to explain some of the intra-regional differences highlighted thus far and suggesting that ingrained commitment to nursing could develop in some parishes and not others.
Professional nursing also overlaps with the "contingent nursing" categorization of figure . In Northamptonshire, Wiltshire, and Berkshire, bills issued by doctors provide a clear indication that they could engage their own nurses and charge the parish for out-of-pocket expenses in this regard. We can be more precise about another aspect . Nursing in general but smallpox nursing in particular was very dangerous for the men and women involved. Fissell, Patients, Power, .
of contingent nursing: that in formal or informal institutional settings. Broadly, the more parishes turned to nursing outside the homes of the poor, the higher the bills they paid for nursing services. On September , , for instance, the vestry of Peterborough St John took out a fire insurance policy (£) for a "newly erected [and clearly substantial given the insurance amount] building used as a nursery and on a new building adjoining used as a place of confinement" at Westgate in the town.  Other parishes converted existing buildings or carved nursing areas out of workhouse accommodation. The workhouse book of Bray (Berkshire), for instance, recorded on the April ,  that "the upper right hand front Garret be reserved as a sick room, into which all persons deemed ill by the Apothecary shall be removed, and a proper steady person appointed as a nurse."  The fact that parishes in Northamptonshire and Berkshire, irrespective of wealth of socio-economic typology, were more likely to provide nursing in institutional settings explains at least part of the variation between the county datasets in the percentage of medial relief devoted to nursing services. Moreover, contingent nursing was also provided by way of parishes buying service and space in the houses and establishments of parishioners. At its lowest level, this involved the boarding-out of individual sick paupers, but multiple boarding was surprisingly common. In Caversham, Mary Clisby was paid in the s and s for providing board and lodging for the sick in her own home, attending and taking in sick travelers and strangers, and for hiring out nurses who were clearly in her employ so that they could attend other cases on behalf of the parish. We can find similar examples in all of the counties analyzed here, though the "nursing homes" were much more common in Berkshire than they were in either Northamptonshire and Wiltshire, and especially Norfolk, where such provision was rare. These places were not on the same scale as those found by Boulton in London, but communities that used or hosted such nursing homes-urban Caversham; rural Woodford Halse and Hardenhuish-tended to devote more of their overall and medical relief resources to nursing than parishes that did not.  Other forms of nursing exhibited less spatial diversity in terms of the regularity of parish spending. Ad hoc nursing care in response to sudden need was a commonplace. In all counties, it was surprisingly common to find such care recorded in the overseers' accounts alongside that provided by more regular nurses for the same person. Hence, in Bray in January , Mrs Hearn was paid s for nursing widow Haines, but Widow Brown was given an additional s for "assisting the nurse."  In terms of specialist nursing, most parishes paid for nursing related to maternity and postnatal care, albeit Berkshire parishes seem to have been more generous payers than those in other counties. Meanwhile, all parishes attached importance to nurses employed to deal with epidemic disease, especially smallpox, typhus, and the diseases of childhood. Smallpox nurses were paid at a considerably higher rate (-s per week plus board and lodging even for nurses living within the parish appears normal) than those employed for regular nursing, and even more than inter-parish nurses of the sort that encountered above. It thus follows that parochial susceptibility to smallpox and other infectious disease was a driver of the regularity and scale of expenditure on nursing and the share of nursing in overall and medical relief.  Clearly, then, some of the intra-and inter-county variations highlighted in figures  and - relate directly to the nature and scale of the nursing labor force. Parishes employing significant numbers of external nurses, or where nursing-type homes were more common, as for instance in Berkshire, spent more (absolutely and proportionately) on nursing care than those which did not. There was also a relationship, though a complex one, between the size of the nursing labor force and the importance and scale of spending on nursing. It does not inevitably follow that those parishes employing most nursing labor would spend more (proportionately or absolutely) on nursing. Much depended on the duration and intensity of nursing seen as normal by officials and on unit pay costs (both issues examined in the next section). Much, however, also depended upon the constellation of nursing rather than simply its volume. In those parishes and counties (such as Berkshire or Northamptonshire) where multiple simultaneous and sequential nursing in a combination between specialist-ad hoc-professional-single-event was deployed for bouts of sickness, nursing expenditure tended to be higher absolutely and proportionately. This, we might imagine, is at least in part related to traditions of ursing under the Old Poor Law, with the slow but definite development of a sense, in Berkshire and Northamptonshire communities in particular, that parishes had basic obligations to provide a comprehensive nursing service for the sick. Where, as for instance in Caversham or Thatcham, the administration of the poor law passed relatively early into the hands of officials who remained in place for a long time, the formation of this sort of parish sentiment was accentuated. At the opposite extreme, parishes where active committees of local ratepayers ("vestries") took power out of the hands of local welfare officials, as for instance in the Northamptonshire town of Wellingborough or the Norfolk town of Thetford, so traditions of nursing provision dissipated. 

N U R S I N G ID E N T I T Y, PAY, A N D P R AC T I C E
The identity, careers, and pay of nurses are not always easy to reconstruct. Some nurses in the sample, especially those who provided ad hoc nursing services, were certainly paupers in receipt of an allowance. Indeed, this was one way in which officials could defray the unit costs of nursing, paying paupers less than others (via an addition to their allowance) or enforcing an expectation that nursing was a condition of relief. Jane Smith of Bray, a s per week pensioner, was paid s per week for nursing smallpox victims between May and July . The sum of s per week was no doubt a considerable addition to her allowance, but it pales into insignificance when compared with what a specialist nurse would have expected.  A few parishes (and generally not the poorest) came to rely heavily on this sort of labor in our period, dampening absolute and relative spending on nursing care. Yet, only a small proportion ( percent) of nurses employed on any basis were, at the point they started nursing, paupers.  Indeed, the relationship between pauper status and nursing is complex. It is a striking feature of the sample, for instance, that many nurses subsequently became ill themselves. The act of nursing made those employed more vulnerable to dependence upon the poor law. Thus, in Ardington in , William Church and his family were given cash payments totaling  s d and Mary White (s d) and (widow?) Wood (s) were paid for nursing them. Immediately after this episode, the overseers recorded that "Mary White very ill" and paid Mary Purton (s) and "Nurse Ballard" (s) for nursing her. "Dame" Ballard then fell sick in her turn and was nursed by a recovered Mary White (along with three others) in .  Unsurprisingly, then, while relatively few women were paupers at the time they nursed, a larger proportion ( percent) would subsequently become so.
Other nursing duties were probably undertaken by neighbors, lodgers, co-residents, and kin, drawing upon a much older custom of nursing for all social groups.  Without family reconstitution data, it is difficult to specify the relationship between the sick and their nurses, even if they shared the same name. Moreover, it requires a close reading of the overseers' accounts (and potentially their linking with rate exemption and paying lists) to understand who was living with/ lodged with whom. That kin and neighbors/co-residents were paid for nursing in all of the parishes is not in doubt; accounting entries such as "Elizabeth Harden for nursing her mother" make a clear connection. For some parishes, a combination of vestry minutes and detailed accounts allows a more systematic perspective. In Welton, for instance, some  percent of all nursing engagements involved work by relatives, neighbors, or co-residents. The employment of such labor was not, however, an avenue for cost-cutting. Such nurses were well-paid, and their services were often employed alongside other forms of medical relief and even alongside the services of other nurses. Here at least the attitude of officials toward nursing feels much more like a tradition than a reflexive balancing of immediate economy with the need to reduce long-term bills by ensuring that the sick returned to work and independence.  . BRO DP//, Overseers' accounts of Ardington. . Pelling, Common Lot, ; Fissell, Patients, Power, . Borsay and Hunter, "Nursing and Midwifery," , suggest that this form of nursing care was "by far the most common." . NRO p/-, Vestry minutes and accounts.
Saying more about the identity of nursing labor in terms of marital status, age distribution, and supplementary occupations is hampered by lack of any wide-ranging census-type data. That widows were over-represented as nurses in most of the sample is clear. Working only from the standard notations in overseers' accounts ("Widow," "Dame," etc.) suggests that more than one-quarter of all nurses were widows, significantly greater than their likely presence in the background population in this period.  For those parishes where vestry minutes and bastardy examinations survive, it is also clear, as Samantha Williams has found for Bedfordshire and Pelling for Norwich, that the mothers of illegitimate children were well represented in the ad hoc nursing labor force. Internal evidence from the overseers' accounts also suggests this. At Hungerford, Berkshire, in May , for instance, Bett Palmer was given s "more than her pay [s] for nursing Widow Rosier," at the same time as she looked after her illegitimate child.  It is impossible to establish whether other generalizations in the literature, particularly that nurses tended to be older women, also hold true for this sample, though pauper letters, overseer correspondence, and settlement examinations do provide incidental indications that this might be so.  Evidence of the duration of employment and the nature of pay for nursing is rather richer and in any case is likely to explain more of the intra-and inter-county variations in the scale and importance of spending on nursing than the age or marital status of the labor force. Of course, reconstructing pay rates for nursing services (let alone income for individual nurses) is fraught with difficulty. Within and between parishes nurses were paid both by the week/month and by episode, and sometimes their pay was supplemented by board and . Williams, "Caring," , suggests that  of her  female carers in Campton and Shefford combined could be assigned a marital status either by title or record linkage. Of these,  percent () were widows while  percent () were married and the remainder were spinsters, the mothers of illegitimate children or changed status.
. BRO DP//, Overseers' accounts. . For instance, BRO DP//, Pauper letters. Richard Dredge, overseer of Reading, wrote to his counterpart in Pangbourne, Berkshire, on September ,  to say that Ruth Warner needed relief. He noted, "The Pauper it appears has supported herself for upwards of ten years by going out nursing. She is now upwards of seventy three years of age and I think a very proper object for to receive parochial relief." See also Margaret Pelling, "Thoroughly Resented? Older Women and the Medical Role in Early Modern England," in Lynette Hunter and Sarah Hutton, eds., Women, Science and Medicine - (Stroud: Sutton, ), -. lodging or other perquisites and sometimes not. Where such additions were usual, as for instance in Forncett St Peter in Norfolk, they were not consistently paid to the nurse as opposed to the sick person or the person lodging the nurse, so that the net income of nurses is difficult to discern in the overseers' accounts. This is made more complicated by the fact that nurses of whatever hue were not always paid in cash, but either requested or were assigned goods in kind, as for instance in Marlborough and Upavon in Wiltshire.
Naturally, rates of pay varied according to the duration of the nursing or the riskiness of the attendance. Rates varied too according to whether nurses were expected to look after individuals or families, and whether they were providing other aid than just nursing. In Caversham, for instance, one nurse was paid extra for giving up her bed to a sick patient. Payments were also unstable over the course of a year. In all of the counties and parishes analyzed here, overseers made a distinction between ordinary nursing and that conducted during the harvest, for which extra pay was given. In August  at Warham (Norfolk), Hannah Basham was given "extra in harvest time doing for the old people as usual s," while Widow Green was paid s d for providing nursing for her family and "Loss of time made no Gleaning."  Similarly, in Ardington, Berkshire, because John Emblin's wife was "ill in the harvest" in  the overseer had to pay an overthe-odds s d for nursing by Martha Alder. In , the tables were reversed and John Lay's wife was ill in the harvest and required nursing, for which the overseers were required to pay Margaret Emblin an eye-watering s d.  Another subtle influence on the level of pay for nursing was whether an illness was expected to lead to a death, for which nurses were paid more than when a patient was expected to recover, even where additional labor was employed for laying-out the body or sitting up with the deceased during their final hours.
Such observations are cautionary when trying to assess unit costs or cost ranges for different types of nursing. To them, and especially in light of the need to make intra-and inter-county comparisons, we might add the sometimes thin detail on how long a period payment to nurses actually covered, and the intensity of that nursing. It is certainly true, for instance, that overseers in the best-documented parishes factored payments according to the intensity of nursing. Thus, in Ardington, Berkshire in  Nurse Ballard was given s d per week for two weeks for "looking after" John Clark but Mary Brown was given s for providing nursing "Eleven weeks at several [times]" to the same person, clear recognition of a differing intensity of engagement on the part of the two nurses. Similarly, in  Elizabeth [Nurse] Ballard and Lucy Goody shared a lump sum payment of s d for what must have been periodic low-level nursing of Taylor's wife, while in  Goody was paid s for two weeks of what must have been intensive nursing for Taylor in the run up to her death. Also in Ardington, the mother of Rachel Miles was paid  weeks at s per week for looking after her, and then  weeks at s, implying a different intensity of nursing recognized by the overseer, and perhaps even a de facto scale of nursing charges.  Nor, perhaps, should we forget that some professional and semi-professional nurses were actually paid even when they were not providing services. In Bray Mrs Lock was retained at s per week.  Against this backdrop, there are striking differences in both the duration of nursing and the rates of pay for those undertaking the task. Berkshire parishes (on average and across all types of nursing) paid their nurses more than in the other counties. This disparity remains even where we group the parochial dataset across all four counties on the basis of wealth, broad topography, or socio-economic typology. In particular, Berkshire parishes paid considerably more than their counterparts in Norfolk, the county that demonstrates the lowest absolute and proportionate expenditure on nursing, and even where we control for the size of the rate base in individual parishes.  Thus, postnatal nursing across a range of Berkshire parishes was paid at double the rate that can be traced for Norfolk or Wiltshire communities, something that also applies to general nursing and attendance on the insane poor. Moreover, there were also important inter-county differences in the duration of nursing. For Berkshire, the average duration across most nursing forms was higher than for the other counties. In turn, the juxtaposition of longer terms of general and . BRO DP//-, Overseers' accounts. . BRO DP //, Workhouse book -. . Williams, "Caring,"  suggests that in  Norfolk parishes were on the higher end of the spectrum. These findings also contrast with Taylor, "Aspects" who sees rather more parsimony in his sample of Berkshire parishes. chronic nursing with simultaneously longer periods of attendance, waiting and other more casual forms of nursing reflect the fact that the Berkshire parishes were particularly likely to engage more than one nurse to provide care for an individual. The Berkshire parishes also demonstrated the longest periods of chronic nursing, while in Norfolk the juxtaposition of long-duration nursing care by relatives with the lowest figures for "looking after" and chronic nursing probably points to a rather heavier reliance on cheaper family labor for long-term care than in the other counties. A sense that differences in spending on nursing within and (particularly) between counties might reflect ingrained cultural and policy attitudes toward the nursing of the sick poor, rather than simply the relative wealth of a community, thus continues to emerge.
Moving from these observations to understand the sorts of livings made by nurses is hampered by occasional unspecified payments for "nursing" and the fractured earning histories that we detect for inter-parish nursing. Nonetheless, for each of the parishes we can construct exemplary nursing careers and earning profiles. Such work demonstrates clearly that some nurses earned sums amounting to a significant proportion of the average male wage in the county concerned once we correct for the earning-effects of male underemployment. In the Berkshire parish of Brimpton, for instance, Nurse Coxhead was paid £ s for nursing Thomas Hockley with the smallpox in August , more than one-quarter of the corrected average male wage in the area. In the same parish, Widow Fry was paid s d for nursing John Lee, plus s d for lending him her bed in , and she then made s d for nursing Luke Pembroke the next year for only weeks work. In , Mrs Pool was paid £ s to board, lodge and nurse John Lawrence for six months, fully two thirds of the corrected annual average male wage for the area at this date.  In Caversham, Mary Wilder nursed two to three people at a time, receiving between d and s d per week per person in the later eighteenth and early nineteenth-centuries.  These are not equivalent to the sorts of sums traced by Boulton in London, but they are nonetheless significant.
Figures for parishes in other counties may well have looked even more remarkable to ratepayers. In Welton, for instance, Widow Heath was paid an average of £ per year for various nursing services between  and , fully three-fifths of the corrected average male wage for the area. Even in the evidently more parsimonious Norfolk parishes Mary Dixon earned an average of £ d per year between  and  for nursing provided to the parishioners of Forncett St Peter. These incomes are within or exceed the range of £-£ per annum that Perry Williams sees as normal for nurses (not just poor law nurses) in this period.  In common with the work of Samantha Williams, only a small minority of all nurses ( percent across the sample as a whole) earned sums that would have made a fundamental contribution to the household economy. This said, there was a significant difference between the experiences of nurses in Berkshire and to a lesser extent Northamptonshire, and those in Norfolk and Wiltshire. In the former counties, a significant minority of all nurses ( percent) gained consistent employment and income from their roles, reflecting a policy of attaching nursing to most cases of sickness and maternity and a local labor force in which nursing was often a discrete occupation rather than an addendum to a range of other work. The opposite was true of Norfolk and to a lesser extent Wiltshire, where nursing tended toward the cheaper end of provision and duration. Across the whole sample some  percent of nurses made between £ and  per year, providing services on a sufficiently regular basis for them to have regarded such sums as part of their family economy and for us to regard them as "nursing labor." A deeply inscribed model of poor law nursing, which concentrates on casual provision, often by other paupers, has a limited basis in reality, particularly for counties like Berkshire. In turn, it is clear that nursing provided a formidable opportunity for female labor force participation, a fact that has been consistently overlooked in recent debate on women, work, and wages, where the supposed casual nature of the work has led nursing to the very margins of our understandings of female labor markets.
Journal of the History of Medicine : Vol. , October   CON C L U SI ON The empirical basis on which to understand nursing under the Old Poor Law has, until now, been relatively slim. This despite the fact that parochial authorities spent many hundreds of thousands of pounds annually on nursing care widely defined, with individual parishes conceptualizing nursing as an integral part of sickness relief, something to be provided as a discrete service separate from consideration of the need for doctors, drugs, and cash allowances. Historiographical views of the nature and quality of this nursing-essentially characterizing it as casual, residual, poor quality, and often undertaken by paupers, in contrast to the professionalized service we see emerging from the s-are deeply entrenched. The material analyzed here questions this view; in counties like Berkshire parishes had access to a substantial core of "professional" nurses drawn from a local and intraregional pool of nursing labor. These men and women were often well paid, undertook nursing intensively, and sometimes founded their own nursing homes.  Such labor was neither casual nor residual. Nor was it cheap. Moving from individual microstudies to a wider analytical canvas allows us both to grasp the scale of nursing and, once done, to understand its importance if we are to gain a proper understanding of issues such as the scale and nature of female labor force participation and the relationship between the intensity of work done and the variability or otherwise of female wages. In the counties analyzed here, nursing must have made a considerable impression on family economies.
Yet there were also notable differences between parishes in each county and, even more markedly, between counties in the level of absolute spending on nursing, the percentage of medical relief devoted to this task and in the intensity and nature of nursing undertaken. Only by assembling considerable datasets and analyzing them on a comparative basis do such variations come to light. These differences were remarkably stable over time and, contrary to many of the implicit assumptions written into the historiography of the poor law, showed little systematic relationship to either the relative wealth of parishes/counties or their socio-economic composition. We have . As an adjunct to this analysis, it is clear that London formed its own nursing enclave in the eighteenth century.
King : Nursing Under the Old Poor Law  suggested a three strand model for characterizing the experiences of individual communities, running on a spectrum from nursing that was deeply ingrained into the basic policy-making of parishes through to places where commitment to nursing as part of relief and medical relief was fragile and inconsistent. This model can help to locate further microstudies of nursing practice at parochial level for the Old Poor Law period. It also highlights some very important spatial= patterning, notably a disparity between a broadly defined "eastern" experience (relatively low expenditure and low intensity nursing) represented by Norfolk and Bedfordshire and a central/southern policy (relatively high expenditure with a strong core of professional nurses) seen most keenly in Berkshire but also Wiltshire and Northamptonshire. This sort of subtle spatial differentiation provides a middle ground for a wider understanding of variation in poor law character between those who see infinite variety in practice across a poor law populated by welfare republic and others, including myself, who have argued for much wider regional constellations of practice that might be understood as "welfare regimes." Constructing an explanatory model for inter-and intra-county variations in attitudes toward nursing is problematic. Different confluences of circumstances could lead to the same outcomes in terms of spending on nursing. For instance, parishes firmly cemented into the regional medical market and those isolated from it might end up spending relatively little on nursing, the one because officials had plenty of choice and could keep down costs, and the other because there was simply not the wide range of nursing choices that might otherwise underpin more substantial costs. Nonetheless, the article has suggested a range of variables that help to contextualize variation in practice and spending within and between counties. Albeit crudely, such influences can be ranked in terms of importance and longevity, as table  suggests. Thus, a heavy susceptibility to epidemic disease could drive higher nursing expenditure in some communities compared with others; the formation of nursing homes was positively related to spending; and parishes that tended to draw on a core of professional and/or inter-parochial nursing labor invariably spent more on nursing than where this was not the case.
Yet as table  suggests, some influences have a totemic status if we are to understand the complexion and cost of nursing provision for the sick poor. Foremost among them is the steady accretion in some parishes and counties of a sense-perhaps a custom-that it was the obligation of local ratepayers to provide both medical and nursing care for the sick poor. This was certainly the case in most Berkshire parishes in a way that is simply was not in Norfolk. In the latter county nursing tended to be provided at the less intensive, less professional, and less expensive end of the policy spectrum. Norfolk communities did not lack the money or the nursing networks to adopt a different approach; they lacked the history of action and the will to change. A collective memory of obligations could be disrupted if a determined overseer took control or if a vestry of local ratepayers took control of decision-making, but surveying all of the poor law material available for the parishes considered here points to longevity of customary practice over the long-term. Parochial officials-and particularly those who remained in post for some time-were remarkably powerful guardians of the collective memory and normative standards. It is these normative standards that lie at the heart of intra, but more especially inter-county variations in the complexion of nursing provision.
In turn, we see these different county attitudes reflected in and shaped by the demands for nursing made by paupers in the letters they wrote back to their parishes of settlement when they became sick and were "out of their place." Paupers brought up or living in


Berkshire, Wiltshire and Northamptonshire inscribed an expectation that they would receive support for nursing into the basic rhetoric of their narratives. They asked, often unsubtly, how much they were to receive toward nursing care or who was to be nursing them. Nursing care, in other words, was to be understood as part of the necessary infrastructure (alongside food and money) in the appropriate parochial response to illness among the deserving poor.  Equivalent writers from Norfolk asked whether they would be given help with nursing. The rhetorical and strategic differences embodied in the two types of letter are slim but nonetheless crucial. They point to a shared set of expectations in both counties, expectations which we in turn see played out in the money spent on nursing and the propensity to turn to professional nursing networks. 