---
abstract: |
  Debate about the Cape’s 1713 smallpox epidemic has centred on its disputed death toll. This article reconstructs its severity and social consequences from a systematic search of the colony’s daily journal for 1700–1720, tax returns, and Orphan Chamber estate papers. The journal traces disease from March into November and connects it to disrupted harvesting and renewed marriage. Estate documentation increased fivefold in 1713. Between 1712 and 1714, the recorded settler population fell by almost a quarter, and privately enslaved children declined by a similar proportion while adult counts remained stable. Survivors reorganised their lives: relatives arranged children’s inheritances, spouses formed new unions, and Khoesan communities chose new leaders from the kin of their dead captains. Recovery ran through unequal institutions, as the protection of some children’s property entailed the transfer of enslaved people. Population totals alone understate the social reach of the epidemic.
author:
- Johan Fourie[^1]
bibliography: references.bib
reference-section-title: References
title: "A disease never seen here: Smallpox and family reconstruction at the Cape, 1713–1714[^2]"
---

> Figures and typeset tables are omitted from this Markdown version.
> The complete paper, with all figures, is in JF_ADiseaseNever_v1.pdf.


**Keywords:** smallpox; Cape Colony; family history; slavery; historical demography

**JEL codes:** N37; I15; J12

# Introduction

In November 1713, farmers at the Cape of Good Hope were preparing to harvest after months of smallpox. The Company’s journal reported that most of the Khoesan people on whose labour they relied had died. Some farmers were turning to scythes, an unfamiliar practice locally. Three weeks later, the clerk recorded the 32nd marriage since July and connected these new unions to marriages broken by the epidemic. Smallpox had changed both the work needed to sustain families and the families themselves. Its severity lay in these changes as well as in the number of people who died.

This article reconstructs the epidemic’s demographic and social consequences by bringing three bodies of evidence together. A systematic search of the full supplied journal, covering 7,666 usable daily texts across 1700–1720, establishes the recorded sequence of illness and responses. Tax returns reveal changes in adults, children, and the composition of administrative households in Cape District and Stellenbosch–Drakenstein. An audit of all 90 estate documents dated 1713–1714 connects the administrative surge to bereavement, inheritance, and the arrangements survivors made. Read together, these sources show a prolonged mortality crisis followed by family and community reconstruction under markedly unequal conditions.

Three findings follow. First, the crisis was severe across several recorded populations and lasted well beyond the town’s initial outbreak. The journal reports disease already widespread among Company-owned enslaved people by late March, rural deaths in August and continuing illness in Drakenstein in November. Estate documentation rose fivefold in 1713. The two district returns recorded almost a quarter fewer settlers in 1714 than in 1712, and privately enslaved children declined by about a quarter while adult counts remained almost unchanged. These changes in recorded stocks are not mortality rates. Their significance emerges from their convergence with direct reports of death and disruption in the other sources.

Second, population totals conceal substantial reorganisation. The Company’s own account explicitly states that arrivals and births offset part of its losses. Estate papers follow children’s property into relatives’ and other recipients’ custody, record preparations for remarriage, and trace one child’s belongings to Batavia. The journal connects renewed marriage to epidemic bereavement. Survivors from four Khoesan communities sought successor captains after reporting devastating losses, and each successor was a brother, nephew, or son of the dead captain. Recovery involved the formation of new relationships and responsibilities, not simply the replenishment of a population.

Third, the institutions that supported this reconstruction distributed protection unequally. The Orphan Chamber secured recognised heirs’ material claims while administering the transfer of enslaved people. Their relationships appear in the same papers, but receive far less protection and description. The wider journal also records enslaved men providing assistance and cleaning in the Company hospital before the epidemic. Care depended partly on coerced labour, while losses among Khoesan workers disrupted settler production. These connections place the epidemic within the organisation of colonial society rather than treating disease as an external shock with uniform consequences.

The contribution builds on a substantial historical literature. Ross (1977) used the tax returns to question received accounts of the epidemic, while later population reconstructions made explicit demographic assumptions where direct enumeration was unavailable (Fourie and Green 2015; La Croix 2018). Historians of the Cape have situated disease within dispossession, labour coercion and changing political relations (Marks 1972; Worden 1985; Guelke and Shell 1992). The present analysis extends that work by connecting systematic corpus retrieval, disaggregated population counts and named family histories. This combination establishes consequences that neither aggregate demography nor selected journal passages reveal on their own.

For family history, the central implication is that the effects of mortality continued through the organisation of survival. Recent work on the direct costs of epidemics draws attention to care, disrupted production and the institutions through which households absorb loss (Curtis et al. 2025). The Cape evidence brings those processes together in a colonial setting. Family reconstruction could preserve an inheritance, create a new marriage or renew leadership; it could also transfer control over people whose own relationships the administration scarcely recorded. A history confined to the death toll misses this distribution of consequences.

The article first sets out the setting and sources, then reconstructs the epidemic through the full journal corpus. It examines estate activity and population changes before following families and communities through the aftermath. The final discussion brings the evidence on care into comparison with other smallpox histories. The supplementary appendices provide the retrieval protocol, source concordances, coding audit, and demographic sensitivity checks.

# The Cape and the epidemic

The Dutch East India Company, or VOC, established its Cape settlement in 1652 to provision ships travelling between Europe and Asia. By the early eighteenth century, the settlement included Cape Town and an expanding agricultural hinterland. Cape District encompassed inland farms as well as the town. Stellenbosch–Drakenstein covered another substantial farming area; a comparison between these districts is therefore not simply a comparison between a port and a frontier. Settler production depended on family members, hired or contracted workers, enslaved people, and Khoesan labour under changing and often coercive conditions. Company employees and Company-owned enslaved people formed administrative populations distinct from the freeburgher tax units (Worden 1985; Fourie and Green 2018).

These distinctions matter for exposure and for the survival of evidence. An enslaved person listed with a farmer’s assets was not thereby a member of that farmer’s kin group, even when both lived or worked on the same property. Conversely, enslaved people had family relationships that a fiscal return might not record. Company ownership also does not establish that everyone in a muster resided in the slave lodge. The institutional label identifies the owner and the counting procedure; it does not identify a single residential population.

Smallpox spread through close contact with infectious people and could also be transmitted through contaminated material. Illness could incapacitate several people within a household, making food, water, and nursing harder to obtain. Recovery usually conferred lasting immunity, so prior exposure mattered. But a population described as encountering smallpox for the first time could contain immigrants who had already survived it elsewhere. The Cape’s mixed origins make a uniform assumption of susceptibility particularly inappropriate. Neither the tax returns nor the estate documents establish individual infection histories (Fenner et al. 1988; Schneider and Davenport 2026).

Prevention and treatment must also be distinguished. Variolation deliberately exposed a person to smallpox material before natural infection; it had histories in Asia and Africa before its prominent adoption in eighteenth-century Europe. Vaccination came later. Neither intervention should be confused with a remedy given to somebody already ill (Fenner et al. 1988). The absence of a specific effective cure in 1713 does not imply the absence of medicines, attempted treatments, nursing, or beliefs about healing. Nor can modern medical knowledge establish what a Cape official believed available. These distinctions are essential when interpreting the journal’s remarks about Khoesan knowledge of the disease.

By early May the journal described accumulating deaths in the town. Its entry of 6 May reported 13 white bodies awaiting burial, all of people born locally or previously unexposed to smallpox, and then reported Khoesan deaths. On the following day the government arranged the burial of nine Khoesan bodies. The clerk described sorrow in the streets. In June the journal reported that illness was easing locally while continuing in the countryside. These are observations by officials within a crisis, not a surveillance system with uniform coverage. The entry of 13 June gave a cumulative count of 160 men, women, and children since the beginning of April, “at this place”. It did not state a usable population denominator or precisely delimit the social groups included.[^3]

Accounts of the epidemic have often centred on Khoesan depopulation. That emphasis reflects the epidemic’s place in the history of colonial expansion, not merely a dispute over arithmetic. Theal’s (1897) account and subsequent syntheses conveyed extreme losses; Ross and later demographic historians questioned how those descriptions could be converted into population estimates. Elphick (1984) discussed depopulation over a longer period, which should not be silently equated with deaths during a single outbreak. A reported fraction of a particular group, a conjecture about a wider population, and a reconstruction over a decade answer different questions. The analysis below preserves testimony about devastation while avoiding a numerical precision that these sources cannot sustain.

# Sources and measurement

## Tax returns and estate papers

The *opgaafrolle* record fiscal declarations organised around named heads and associated persons and assets. The analysis uses the Cape of Good Hope Panel transcriptions and their documented household-project release, with a documented mapping ledger. The transcription work, including that of Chris de Wit and Hans Heese, is described by Fourie et al. (2024). It is distinct from the TANAP transcription of Council of Policy resolutions. The focal workbook indexes identify the Hague archival references for 1712 and 1714.[^4]

The unit is an administrative tax unit, not a verified residential household. Counts distinguish settler men, women, sons, and daughters, and privately enslaved men, women, boys, and girls. The category translated as sons or daughters need not correspond to a fixed biological age band. The analysis therefore uses “adult categories” and “child categories” when precision matters. It does not add named heads to numeric columns or assign enslaved children to the head as offspring. *Knechts*, recorded separately, are not folded into the four settler categories. The totals consequently do not represent every person living within either district.

Blank cells remain missing in the analytical data. Reported totals sum the numeric counts actually entered; they do not establish that every blank means no person. This is a consequential limitation because the focal sheets contain many blanks and no explicit numeric zeros in these population columns. Two uncertain entries in the 1714 Stellenbosch–Drakenstein sons column remain unknown. A separate problem arose in harmonisation: the historical spelling *Meijsies* was not mapped to enslaved girls in the shared release. The analysis recovers that column from the local transcription, records every override and retains its blanks. Otherwise the apparent absence of girls before 1714 would be a processing error rather than a demographic result. Appendix A gives the definitions and sensitivity checks.

The MOOC8 series consists of estate inventories and associated papers of the Orphan Chamber. The Dutch transcriptions used here were produced by the TEPC Transcription Project of 2004–2008 at the Western Cape Archives and Records Service; GLOBALISE now provides a public rendering of the XML. These records have been used extensively to study Cape wealth (Fourie 2013). They also record inheritance, dependants, property custody, and transactions involving enslaved people. Their selection reflects property and administrative circumstances, including provision for minor heirs. They are not a representative register of all deaths.

The extraction counts dated document headings consistently across the series and preserves each document’s text and locator. The focal audit covers all 90 headings dated 1713 or 1714. It separates the heading date from dates within the document, flags repeated estates, and identifies passages used for the family cases. One archival number, MOOC8/3.54, occurs with two different headings; both documents are retained under distinct date-qualified identifiers. Names in headings are not automatically deceased persons: some declarations name the living spouse or declarant. The cases below are selected for explicit relationships and recorded arrangements, not to estimate the prevalence of orphanhood or a representative family response.

## The journal and computational assistance

The *dagregister* is the daily journal, not the Council of Policy’s resolution series. It contains weather, shipping, and local affairs, sometimes reporting meetings or incorporating correspondence. A reference to a Council meeting does not make the entire entry a resolution. For the decisive passages, the analysis uses the journal manuscripts held by the Nationaal Archief, VOC inventories 10730–10733, and compares them with the supplied Dutch text. Leibbrandt’s published extracts provide corroboration but are abridged and sometimes attach material to a different date. Appendix C records the concordances and worked examples.

The dagregister was transcribed and supplied by the Tracing History Trust and shared by Helena Liebenberg. The full supplied workbook contains 7,670 distinct daily identifiers across the 21 calendar years 1700–1720. Four entries lack usable text, leaving 7,666 texts and approximately 858,000 words. Every usable entry enters the analysis. This defines the population of texts being measured; completeness relative to every page of the surviving archival series has not been independently established.

The analysis separates reproducible lexical measurement from interpretation. Overlapping searches retrieve smallpox, illness, death, hospital, and medical vocabulary, accommodating historical spelling variants. Annual counts describe entries containing these terms. Reading every smallpox candidate in context identifies 17 relevant entries and rejects one false match. Medical passages are screened across the entire period, with extended readings of the institutional and treatment examples discussed below. Entry identifiers, search expressions, matched passages, and decisions are retained, allowing a reader to move directly from a result to its source. Co-occurrence of a group name and illness somewhere in an entry is not treated as proof that the group experienced that illness.

Smallpox references are rare, so sampling cannot establish how completely they are recovered. A hand-coded random sample of 79 entries contains none (Appendix D). The smallpox series therefore inspects all 18 retrieved candidates. Its scope is explicit references recovered from the supplied text, not an estimate of every implicit discussion of disease.

# The epidemic in the full journal

## Routine recording and an exceptional outbreak

The full corpus places the outbreak within the Company’s routine observation of sickness. Illness vocabulary appears in 644 entries and hospital vocabulary in 250. Of the illness matches, 557 also contain maritime vocabulary. These overlapping lexical counts describe a journal in which shipping and health were often recorded together; they do not assign every illness to a sailor. Figure 1 shows that generic illness mentions did not reach their highest frequency in 1713. By contrast, all 17 inspected smallpox references fall in 1713 or its immediate aftermath: 16 in 1713 and one in February 1714. Counting routine sickness and identifying a particular epidemic thus produce different historical pictures.

*[Figure not reproduced here — see JF_ADiseaseNever_v1.pdf]*
Health vocabulary and smallpox references in the full supplied journal, 1700–1720. Panel A counts entries matching the specified illness or hospital vocabulary per 100 usable entries. Panel B counts smallpox references after contextual inspection of all 18 candidates. Shading marks 1713–1714. Source: Tracing History Trust transcription; author’s retrieval and contextual review.

Spelling variants matter for retrieval. The 3 May entry uses *gepokt*, referring to previous experience of smallpox, rather than the more familiar *kinderpokjes*. Searching for this variant recovers an explicit comparison: illness seemed to be abating somewhat among Company-owned enslaved people while increasing sharply among the *Affricanen*, the Cape-born, who had never had smallpox. In the journal’s usage the term denotes people born at the Cape, not Africans in general; on 6 May the clerk described the 13 white dead in the same terms, as locally born or previously unexposed. Local birth, and hence the absence of earlier infection, was the marker of vulnerability he recorded. Conversely, *capok* in a criminal proceeding produces a false lexical match. The candidate ledger records both decisions. These examples show why a reproducible search benefits from reading matches in their historical context.

The dated sequence establishes a prolonged and geographically uneven crisis. On 31 March the clerk described smallpox, which had broken out among Company-owned enslaved people some time before, as still increasing daily. By 9 April about 150 were bedridden with illnesses including smallpox, and on 18 April the clerk reported 11 deaths in the preceding day. In May the journal described spread into the countryside. Thanksgiving for 15 recoveries on 11 June accompanied a report that the disease was easing locally while continuing outside the town. Further rural deaths were reported in August; people were still ill in Drakenstein on 28 November. The epidemic therefore extended from before the end of March into late November, with recovery in one place overlapping with illness elsewhere.[^5]

## Labour, flight, and the resumption of marriage

Repeated reports connect disease to the colony’s dependence on labour. On 10 May officials feared that illness in the countryside would interrupt cultivation. By late November, approaching the harvest, the clerk reported that most of the Khoesan workers whom farmers had customarily employed had been carried off, and that some farmers were using scythes in response. The manuscript explicitly connects the change in work to those losses. This is direct evidence of an adjustment during the crisis, although it does not measure harvest losses or establish lasting technological change.[^6]

Khoesan people also appear as actors responding to infection. On 19 May the journal relayed a report that some of the remaining Cape Khoesan, fleeing the disease, had crossed the mountains towards another group, which had killed all but one of them for fear that they would spread smallpox. The clerk called this a severe policy. The report is mediated and its events are not independently corroborated, but its content matters: flight and an attempt to exclude infection sit alongside the clerk’s better-known claim that Khoesan people were unfamiliar with the disease and its healing. Reading the passages together resists an account in which Indigenous people simply failed to respond.[^7]

The journal then makes an unusually direct connection between mortality and family formation. On 17 December it recorded the 32nd couple married since 21 July, describing renewed marriage in relation to unions broken by smallpox and as necessary to the colony’s welfare. This does not identify all 32 couples as widowed survivors or establish an excess marriage rate. It does show that a contemporary observer understood marriage as part of recovery. The estate papers below reveal some of the property and kin arrangements through which that reconstruction occurred.[^8]

The full corpus therefore does more than supply quotations. It distinguishes routine health reporting from named smallpox references, recovers a sequence across the agricultural year, and connects illness to work and family formation. It also provides the wider evidence on medicine and care considered in Section 8. The frequency of a term is one result; what the retrieved passage establishes is another. Keeping the two connected makes the journal a systematic source for the epidemic’s social history.

# Estate activity during and after the epidemic

Estate documentation increased sharply in 1713 and remained elevated in 1714 (Figure 2). The annual count rose to 54 documents, compared with a mean of 10.8 in the five preceding years, and then fell to 36. The short baseline makes the familiar fivefold comparison explicit. Over 1700–1712 the mean was lower, producing a larger ratio; the result therefore depends on the comparison window. Pooling 1713 and 1714 yields 90 documents, against 21.6 over two years at the 1708–1712 mean. These calculations describe administrative activity. The difference is not a count of excess epidemic deaths.

*[Figure not reproduced here — see JF_ADiseaseNever_v1.pdf]*
Orphan Chamber documents by heading year, 1700–1720. Counts include repeated estates and associated declarations. The line marks the 1708–1712 mean; 1713 and 1714 are highlighted. Source: author’s extraction from TEPC MOOC8 XML.

Population growth could raise estate activity even without a mortality crisis. The pooled number of recorded tax units increased from 629 in 1708 to 656 in 1712, a much smaller proportional change than the document surge. This comparison gives a sense of scale, but it is not a growth-adjusted death rate: taxable declarations and estates subject to the Chamber are different populations. Changes in wealth, family structure, probate practice, or the survival of records could also change the likelihood that a death generated a document. No compatible annual series of people at risk of entering this probate collection has been recovered.

Repeated documents are a more immediate problem. The estates of Willem Basson and Helena Clement generated separate inventories for Hout Bay and the town within two days in June 1713. Catharina Cruse’s estate produced an inventory, a record of goods retained, and a record of their distribution. The papers concerning Wessel Pretorius and Geertruij Elberts span July and August 1713 and January 1714. Counting each of these as another death would confuse stages of administration with new bereavements. The focal register links nine explicit sets of related documents, while retaining the original document counts for consistency with the pre-epidemic series.[^9]

The 1714 count contains demonstrable continuation of earlier estate work and delays in filing. The January division of the Pretorius–Elberts estate followed its inventories in 1713. A declaration of his own goods by Pieter Christiaanse de Jager, a living man, is dated 8 December 1713 in its body but was submitted to the Chamber on 10 March 1714. It names no deceased person, but it shows that a document could enter the series in a later year than it was written. Jan Oberholster’s January 1714 inventory was made before his proposed marriage to Judith du Plessi; both were widowed, but the document does not date the earlier deaths.[^10]

It follows that the elevated 1714 count cannot be read solely as continued epidemic mortality. Nor can all of it be reassigned to deaths in 1713. The series records the workload generated by deaths and by subsequent arrangements among survivors. Read in that way, persistence into 1714 is itself informative: the administrative consequences of bereavement continued after the most visible crisis. The distinction also removes the basis for treating probate incidence as a lower bound on colony-wide mortality.

# Population counts and household composition

## Changes in recorded populations

Both districts recorded fewer people in the four settler categories in 1714 than in 1712 (Table 1). Recorded privately enslaved children also declined in both districts, whereas the combined adult counts changed little. The comparison brackets the most disrupted epidemic year. It does not assume that the two surrounding returns were complete, identically timed or equally accurate. Figure 3 therefore displays the available years on either side, including 1713; the underlying sex-specific levels appear in Appendix A.

|  | Cape District |  |  | Stellenbosch–Drakenstein |  |  |
|:---|---:|---:|---:|---:|---:|---:|
| 2-4(lr)5-7 Category | 1712 | 1714 | Change (%) | 1712 | 1714 | Change (%) |
| Tax units | 290 | 271 | -6.6 | 366 | 271 | -26.0 |
| Settler men | 273 | 256 | -6.2 | 354 | 266 | -24.9 |
| Settler women | 183 | 148 | -19.1 | 185 | 132 | -28.6 |
| Settler sons | 220 | 133 | -39.5 | 262 | 205 | -21.8 |
| Settler daughters | 194 | 141 | -27.3 | 296 | 207 | -30.1 |
| Settler total | 870 | 678 | -22.1 | 1,097 | 810 | -26.2 |
| Enslaved men | 927 | 931 | 0.4 | 556 | 554 | -0.4 |
| Enslaved women | 171 | 164 | -4.1 | 72 | 77 | 6.9 |
| Enslaved boys | 85 | 61 | -28.2 | 47 | 37 | -21.3 |
| Enslaved girls | 58 | 47 | -19.0 | 42 | 31 | -26.2 |
| Enslaved people, total | 1,241 | 1,203 | -3.1 | 717 | 699 | -2.5 |

Recorded people and tax units, 1712 and 1714

The contraction was substantial. Across the two districts, the recorded settler population fell from 1,967 to 1,488, a decline of 479 people. Children account for most of this reduction. In Cape District, their number fell by about a third, compared with about a ninth among adults; in Stellenbosch–Drakenstein both categories declined by approximately a quarter. Privately enslaved children also declined by about a quarter in each district, while the pooled adult count was unchanged. The crisis left a pronounced difference between the recorded trajectories of adults and children, especially among privately enslaved people.

*[Figure not reproduced here — see JF_ADiseaseNever_v1.pdf]*
Recorded adult and child categories, indexed to 1712 = 100. Cape District includes inland farms. Points show available returns; connecting lines guide the eye across missing years. The shaded year is 1713, retained with its recording limitations. Source: data and definitions as in Table 1.

The surrounding years reveal why the meaning of this contraction requires care. Cape District’s 1713 return contains only 189 tax units, compared with 290 in 1712 and 271 in 1714; recorded adults rebound sharply in the latter year. Stellenbosch–Drakenstein had experienced earlier declines in children and fluctuations in declarations. These series combine demographic events with movement and changes in coverage. They establish the scale and composition of recorded change during an independently documented mortality crisis, rather than a stable counterfactual from which epidemic deaths can be subtracted.

## What the changes can identify

For a consistently defined population, a change in recorded numbers combines demographic flows and recording:
$$
\begin{equation}
\Delta R=B-D+I-O+\Delta E,\label{eq:accounting}
\end{equation}
$$
where $B$ and $D$ denote births and deaths, $I$ and $O$ entries and exits, and $E$ the difference between recorded and actual numbers. For child categories, entries and exits also include transitions between age or status categories. Private enslaved populations could change through purchases, sales, transfers, and manumission as well as movement. The equation is an accounting framework, not an estimated model.

The identity clarifies two distinct findings. A declining count need not measure all the people lost: births and arrivals can offset deaths, as the Company muster explicitly demonstrates in Section 7. Conversely, departures, category changes, and omitted declarations can reduce a count without an equivalent number of deaths. Pooling the districts removes movements between them only where both ends are consistently recorded. Stable adult totals therefore do not establish protection from disease, and the child decline cannot by itself establish an age-specific mortality rate.

Even within these limits, the compositional result is informative. Applying Cape District’s settler adult decline proportionally to its initial child count leaves about 93 more children than were recorded in 1714. The difference could reflect mortality, movements of dependants, ageing, or differential omission. The family documents show why such movements are historically relevant: bereavement prompted new arrangements that could take children beyond the original tax unit or beyond the Cape. Appendix A records the alternative accounting scenarios. The geographic claim remains at the district level; these returns do not measure local contact with Khoesan communities.

## Recorded children within tax units

Changes in the number of declarations account for part of the population pattern. Cape District’s tax-unit count fell modestly between 1712 and 1714, while Stellenbosch–Drakenstein’s fell by about a quarter. Recorded settler children per tax unit fell in Cape District but were almost unchanged in Stellenbosch–Drakenstein (Table 2). This is a more precise description of the geographic difference than treating every missing child as a death.

|  |  |  |  |  |
|:---|---:|---:|---:|---:|
|  | Cape District |  | Stellenbosch–Drakenstein |  |
| Measure | 1712 | 1714 | 1712 | 1714 |
| Recorded settler children per tax unit | 1.43 | 1.01 | 1.52 | 1.52 |
| Recorded enslaved children per tax unit | 0.49 | 0.40 | 0.24 | 0.25 |
| Settler child/adult count ratio | 0.91 | 0.68 | 1.04 | 1.04 |
| Enslaved child/adult count ratio | 0.13 | 0.10 | 0.14 | 0.11 |
| Units with positive settler child count (%) | 45.2 | 37.3 | 40.4 | 43.9 |
| Units with positive enslaved child count (%) | 17.2 | 12.9 | 10.9 | 10.3 |

Child counts within recorded tax units

The enslaved measures reveal a similar distinction. Recorded children per tax unit fell in Cape District but rose slightly in Stellenbosch–Drakenstein, where fewer declarations accounted for a similar average number. Thus contraction could appear through fewer units, fewer children within units, or both. The returns describe this reorganisation at an administrative level. The next section identifies the relationships and arrangements behind particular family histories.

# Bereavement, kinship, and the organisation of survival

## Children and inheritance

The estate papers make the loss of parents visible. Jan Nijs and Lijsbet Jansze de Roode were both dead when their inventory named three children as beneficiaries. Dirk Dirksze van Schalkwijk and Maritie Olivier left four explicitly described minor children. The Basson–Clement documents name Arnoldus and Matthijs as their two minor children. These families experienced the loss of both parents by the time of their documentation in the epidemic year. Individual causes and dates of death are not consistently recorded, so the cases establish bereavement and its consequences rather than a count of smallpox orphans.[^11]

The papers concerning Catharina Cruse allow a closer view of what followed. Cruse, already the widow of Willem Corssenaer, left five minor children. The July inventory lists Aletta aged 18, Barbara 16, Johanna 13, Willem 10, and Johannes six. In August, relatives requested that some goods remain unsold. A subsequent document records belongings distributed among the children and delivered to close relations or others with whom they were then living. It names Joan Casper Rigter, Johannes Swellengrebel and Jacobus Cruse as recipients on their behalf. This is evidence that care and property administration extended beyond the deceased parent’s home. The receipts identify responsibilities for particular belongings; they do not, by themselves, locate every child’s residence or establish who provided daily nursing.[^12]

The later annotation records items sent with Willem to Batavia in April 1714. His family history crossed both a calendar year and the boundaries of the Cape settlement. Although the document does not explain the journey, it demonstrates a form of mobility that unmatched local returns could register as disappearance. Following the estate through time illustrates how a survivor’s departure could appear as a loss in unmatched returns.

Other documents show how kin and local institutions acted together. In the estate of Johannes Viellion and Chatrina Snijman, Hercules des Pres, described as a relation, and the *krankbesoeker* Harmanus Bosman helped arrange an inventory to protect the surviving child’s interests. The goods were secured pending the Chamber’s further instructions. Oberholster’s inventory before remarriage established the property position of an earlier marriage before a new union. Set beside the journal’s December account of renewed marriage, these papers show the legal work involved in family reconstruction.[^13]

Maintaining an estate also required labour. When the *heemraad* Arnoldus Kruijsman inventoried Gerrit Elbertsz’s farm, he left it under Jeronimus Stevensz’s supervision because the *knecht* was sick and unable to act, promising a daily payment. The illness is not identified as smallpox, but the arrangement records a direct cost of incapacity: somebody had to be found and paid to continue the work. This estate-level response complements the journal’s account of changes in harvesting. Both show how illness affected the organisation of production.[^14]

## Enslaved people and unequal protection

The Chamber’s records did not protect everyone’s relationships in the same way. Cruse’s inventory lists Tamer with a child among the enslaved people in the estate. The Basson–Clement town inventory records Pasqual and two children. Their appearance makes it incorrect to say that estate papers contain nothing about enslaved families. Yet the information is much thinner than the names, ages, and entitlements recorded for the recognised heirs. The entries do not provide a complete genealogy, and they do not tell us whether these people remained together after the estate was settled.[^15]

Transfers of enslaved people could accompany provision for an owner’s family. In January 1714, the division of the Pretorius–Elberts estate allocated livestock and two enslaved men, Piet Spangiert and David, to Johannes Wessels through his marriage to Geertruij Pretorius. The other half was to be sold for the benefit of the minor son Hendrik. The transaction is evidence of inheritance sustained through the transfer of people treated as property. It does not record their wishes, relationships, or eventual residence. An institution securing one set of children’s material claims could therefore participate in the redistribution of other people.[^16]

Nor was probate confined absolutely to people classified as European settlers. A document concerning Abraham de Veij explicitly identifies him as a free Chinese man and records assets allotted to his widow Maria Jacobs. It also includes an enslaved person among those assets. The example does not make the collection socially representative. It shows why its boundaries must be described through legal and administrative access, rather than a claim that it contains only European deaths and no marginalised people.[^17]

The Company’s own population appears through a different record. On 14 May 1713, the journal noted that the Company had lost about 95 enslaved people to the disease in two months. On 26 August, it reported a muster of 441 Company-owned enslaved people, 127 fewer than in the previous year. It attributed the shortfall chiefly to smallpox. It then noted recruitment from Ceylon and local births, stating that, after allowing for these additions, the loss during the year approached 200. This is contemporary evidence that additions concealed part of the Company’s gross loss. It is not evidence that private owners acquired replacements at the same rate, or that everyone counted lived in the lodge.[^18]

The scale was considerable: losses approaching 200 were equivalent to about 35 per cent of the previous muster of 568. This compares reported losses with an initial stock rather than estimating the mortality risk of a closed cohort, since births and arrivals also entered the population. More importantly, the Company’s calculation demonstrates a mechanism the private returns leave unobserved: additions could conceal a large part of gross loss. A population’s numerical persistence could coexist with profound disruption among the people who composed it.

## Khoesan testimony and succession

The surviving testimony of Khoesan communities belongs within this history of family and institutional disruption. On 13 February 1714, people from four *kraals*, described in the journal as gatherings or communities, appeared at the Castle. They reported that smallpox had carried off the rest, including their captains, and that one in ten of their company remained. They sought the appointment of successors. The passage concerns four communities, not a single band. It records a catastrophic account given by survivors, transmitted through the clerk’s language and the occasion of a political request.[^19]

The follow-up describes succession through kinship as well as colonial recognition. On 15 February the journal recorded four successors, each identified as heir to a dead captain: Scipio Africanus, brother of Asdrubal; Hanibal, brother of Jazon; Hercules, Hartloop’s brother’s son; and Kolinga, son of the elder Koeinga. Staffs bearing the names of the officeholders marked the Company’s recognition. The episode therefore records more than a fraction of survivors: it shows communities reconstituting leadership through brothers, a nephew, and a son after the deaths of incumbents. The Company’s role expressed a claim to authority, but the survivors’ initiative and the recorded hereditary relationships also matter. These observations do not establish the communities’ internal deliberations or imply that colonial recognition exhausted their political life.[^20]

No usable Khoesan population denominator for this period is supplied by the two tax series analysed here. Their focal forms lack a separately usable Khoesan population count. This is a period- and source-specific statement: later opgaaf forms do contain categories for Khoesan people. The estate papers likewise disclose selected individuals and relationships without constituting a Khoesan death register. The fraction reported by the four communities can consequently be retained as testimony about those communities, but cannot be expanded into a measured mortality rate for the Khoesan as a whole.

This limit does not make uncounted suffering less important. Colonial dispossession and coercive labour relations had already altered access to land, resources, and security (Marks 1972; Guelke and Shell 1992). The journal’s report of burial and its account of survivors seeking renewed leadership show consequences that a tax count misses. Its description of people as unfamiliar with the disease and its healing is evidence of the clerk’s interpretation. It cannot establish that Khoesan people attempted no treatment or lacked practices of care. The distinction is especially important because colonial explanations of Indigenous mortality could themselves diminish Indigenous knowledge and agency.

# Care and comparison

## Medicines and the organisation of care

The wider corpus shows that the Company recorded medical activity well before 1713. In November 1704, correspondence copied into the journal described exhausted medicine stocks aboard a ship in Saldanha Bay and the dispatch of medicines, water, vegetables, and men. A further dispatch specified bread and medical supplies for two vessels. In 1708 the journal recorded remuneration and medicine charges for hospital treatment of Company servants with venereal disease. These entries establish that medicine procurement, attempted treatment and its costs were part of the administration’s recorded business.[^21]

The records also identify some of the labour sustaining care. A resolution copied into the journal on 13 April 1711 retained 16 enslaved men in the hospital to assist sick people and keep the institution clean. The same entry required prescriptions to be recorded in a book, renewing instructions issued the previous year. Enslaved people were therefore part of the hospital’s provision of assistance, not only potential recipients of it. These arrangements show an organised capacity for care embedded in coercive relations; they do not establish its adequacy or the distribution of access during smallpox.[^22]

This context permits a more focused assessment of the claim, repeated in Ramaphosa’s (2020) public letter, that medicines were imported from Batavia in 1713 and used for colonists while withheld from Khoesan people. The full-corpus search located no corroborating passage for that specific transaction or allocation. The result carries more weight than an isolated failed keyword search because the wider record supplies positive examples of medicine requests and dispatches. It nevertheless concerns one administrative source: the prescription-book instructions themselves demonstrate that treatment was also documented elsewhere. The specific import-and-withholding account remains unsubstantiated by this investigation. Unequal access to care is a broader question, for which the coercive organisation of labour supplies relevant evidence but no patient-level comparison.

## Severity, care, and comparative history

The Cape evidence establishes severe disruption across several sources, each observing a different part of the crisis. Table 3 brings those quantities together. Their different denominators preclude a common mortality ranking, but their historical implications can be compared: losses could overwhelm ordinary arrangements, and surviving institutions could reorganise dependency in unequal ways.

| Evidence | Observed quantity | Historical interpretation |
|:---|:---|:---|
| MOOC8, 1713–1714 | and 36 dated documents | Exceptional estate activity after bereavement; documents can repeat an estate and follow death with a lag. |
| Two district returns | Recorded stocks by category | Contraction and changing composition, combining mortality with other population flows and recording changes. |
| Journal, 13 June 1713 | reported deaths since 1 April | Heavy local loss during a defined interval; full population coverage is uncertain. |
| Company muster report | present; 127 net decline; loss near 200 after additions | Arrivals and births masked part of gross loss in this approximate institutional account. |
| Four Khoesan communities | Report that one in ten remained | Survivors’ testimony of catastrophic local loss, followed by the appointment of successors from the dead captains’ kin. |

What each Cape measure establishes

Comparative research helps explain why the organisation of care belongs in an account of severity. Schneider and Davenport (2026) find much lower case fatality in endemic Sweden than in epidemic Iceland in 1707–1709. Simultaneous illness among adults and children could remove the people who normally supplied water, food, and nursing. This is a plausible mechanism for the Cape, where the journal describes illness across age and social groups. The present sources identify some institutions and workers providing assistance, although they do not measure household nursing or its effect on survival. Their contribution is to make care historically concrete: it required labour, supplies, and arrangements for continuity.

Colonial comparison broadens the explanation. Kelton’s (2015) study of Cherokee responses to smallpox places Indigenous practices and disruption through colonial violence at its centre. It demonstrates the value of examining responses and conditions alongside prior exposure. At the Cape, reported flight from disease, the loss of workers, and the initiative of survivors in renewing leadership through kin bring Indigenous action into the same history as settler family reconstruction. The comparison concerns these processes rather than a transferred mortality percentage. Appendix E explains the different documentary bases and the limits of numerical comparison with European cases.

The estate histories show that severe loss could coexist with continuing action by relatives and officials. Making provision for children, maintaining a farm and settling property before remarriage required people to assume new responsibilities. These were arrangements after bereavement rather than observations of nursing during infection, but they reveal a crucial part of the epidemic’s consequences. Survival involved reorganising dependency, and the institutions directing that reorganisation privileged some relationships while treating other people as property. This is where the Cape evidence extends a history of mortality into a history of families and inequality.

# Conclusion

The 1713 smallpox epidemic was a prolonged crisis whose consequences reached through families, work, and political authority. The full journal traces illness from March into November and connects it to disrupted cultivation, changes in harvesting and renewed marriage. Estate documentation rose fivefold, the recorded settler population contracted sharply, and child counts declined among both settlers and privately enslaved people. The Company reported heavy losses despite additions, while survivors from four Khoesan communities described local catastrophe. Taken together, these sources establish severity without requiring their different counts to be forced into a single mortality rate.

What followed was an uneven process of reconstruction. Relatives and officials protected children’s inheritances; surviving spouses formed new unions; Khoesan communities installed the brothers, nephew, and son of their dead captains as successors. Yet these arrangements operated within institutions that recognised some people’s claims while treating others as transferable property. Enslaved people sustained care and production, appeared with children in estate lists, and passed between owners through the same legal processes that secured heirs’ futures. Recovery preserved responsibilities and relationships for some while exposing others to decisions over which they had little control.

The paper shows how loss altered the organisation of survival. Systematic corpus analysis supplies the sequence and institutional context; disaggregated returns reveal changes hidden by population totals; estate histories identify the people and arrangements behind those aggregates. The sources do not identify a colony-wide death toll or every family’s experience. They do establish that the epidemic’s social consequences extended beyond its numerical losses, and that family reconstruction was itself part of the history of colonial inequality.

# Data availability

Scripts to reproduce the numerical tables and figures from deposited aggregates, source locators, input hashes, mapping corrections, derived census and probate tables, the full-corpus retrieval protocol, candidate and care registers, and smallpox decisions are available at <https://github.com/johanfourieza/research/tree/main/2026/1713smallpox>. The detailed estate-document audit and complete journal-entry ledger are retained by the author. Access to these files and the underlying transcriptions is subject to the custodians’ terms; the package does not assert unrestricted redistribution rights over supplied workbooks. The journal images and public TEPC rendering are identified separately from the dagregister transcription supplied by the Tracing History Trust.

# References

Curtis, Daniel R., Bram van Besouw, and Bram Hilkens. 2025. “Household, Family, and Community Responses to the Direct Costs of Epidemics.” *The History of the Family* 30 (3): 329–39. <https://doi.org/10.1080/1081602X.2025.2543382>.

Elphick, Richard. 1984. “The Khoisan to c. 1770.” In *The Shaping of South African Society, 1652–1820*, edited by Richard Elphick and Hermann Giliomee. Longman.

Fenner, Frank, D. A. Henderson, I. Arita, Z. Ježek, and I. D. Ladnyi. 1988. *Smallpox and Its Eradication*. World Health Organization. <https://iris.who.int/handle/10665/39485>.

Fourie, Johan. 2013. “The Remarkable Wealth of the Dutch Cape Colony: Measurements from Eighteenth-Century Probate Inventories.” *Economic History Review* 66 (2): 419–48. <https://doi.org/10.1111/j.1468-0289.2012.00662.x>.

Fourie, Johan, and Erik Green. 2015. “The Missing People: Accounting for the Productivity of Indigenous Populations in Cape Colonial History.” *Journal of African History* 56 (2): 195–215. <https://doi.org/10.1017/S002185371500002X>.

Fourie, Johan, and Erik Green. 2018. “Building the Cape of Good Hope Panel.” *The History of the Family* 23 (3): 493–502.

Fourie, Johan, Erik Green, Christiaan Burger, et al. 2024. “‘Waar Is Beter Dorp in Zuid Africa Dan Stellenbos?’: What the Stellenbosch–Drakenstein Tax Censuses Reveal.” *South African Historical Journal* 76 (4): 420–46. <https://doi.org/10.1080/02582473.2025.2500410>.

Guelke, Leonard, and Robert Shell. 1992. “Landscape of Conquest: Frontier Water Alienation and Khoikhoi Strategies of Survival, 1652–1780.” *Journal of Southern African Studies* 18 (4): 803–24.

Kelton, Paul. 2015. *Cherokee Medicine, Colonial Germs: An Indigenous Nation’s Fight Against Smallpox, 1518–1824*. University of Oklahoma Press.

La Croix, S. 2018. “The Khoikhoi Population, 1652–1780: A Review of the Evidence and Two New Estimates.” *Studies in Economics and Econometrics* 42 (2): 15–34. <https://doi.org/10.1080/10800379.2018.12097332>.

Leibbrandt, H. C. V. 1896. *Précis of the Archives of the Cape of Good Hope: Journal, 1699–1732*. W. A. Richards & Sons, Government Printers. <https://archive.org/details/precisofarchives00cape_7>.

Marks, Shula. 1972. “Khoisan Resistance to the Dutch in the Seventeenth and Eighteenth Centuries.” *Journal of African History* 13 (1): 55–80. <https://doi.org/10.1017/S0021853700000268>.

Ramaphosa, C. 2020. *From the Desk of the President, 24 February 2020*. The Presidency, Republic of South Africa. <https://www.thepresidency.gov.za/node/4133>.

Ross, Robert. 1977. “Smallpox at the Cape of Good Hope in the Eighteenth Century.” In *African Historical Demography*, edited by C. Fyfe and D. McMaster. Centre of African Studies, University of Edinburgh.

Schneider, Eric B., and Romola Jane Davenport. 2026. “What Is the Case Fatality Rate of Smallpox?” *Population Studies*, ahead of print. <https://doi.org/10.1080/00324728.2026.2620692>.

Theal, George McCall. 1897. *History of South Africa Under the Administration of the Dutch East India Company, 1652 to 1795*. Swan Sonnenschein.

Worden, Nigel. 1985. *Slavery in Dutch South Africa*. Cambridge University Press.

[^1]: Department of Economics, Stellenbosch University. Email: <johanf@sun.ac.za>.

[^2]: I thank Helena Liebenberg for sharing the dagregister transcription produced by the Tracing History Trust; Karli de Kock and Martie van Wyk for research assistance; Tim Riswick and Fran Marco Gracia for detailed comments; the LEAP research group at Stellenbosch University; and Robert Ross and Eric Schneider for earlier conversations. The analysis depends on the work of the Cape of Good Hope Panel’s transcribers and data custodians, the TEPC Transcription Project, the Western Cape Archives and Records Service, the Nationaal Archief, and the GLOBALISE team maintaining access to the TEPC materials. Their work is distinct from the computational assistance described below. This work was supported by Riksbankens Jubileumsfond under grant M20-0041. The author reports no competing interests. This paper was created with the help of several models of Anthropic’s Claude and OpenAI’s Codex. Cite this paper as: Fourie, Johan. 2026. “A disease never seen here: Smallpox and family reconstruction at the Cape, 1713–1714.” Working Paper, Department of Economics, Stellenbosch University.

[^3]: Nationaal Archief (NA), VOC archive 1.04.02, inv. 10732, journal entries of 6–7 May and 13 June 1713, digital scans 114–116 and 135–136. The June count is under 13 June in the manuscript, whereas the abridgement by Leibbrandt (1896, 256) associates it with 11 June. The original date is used here.

[^4]: The transcription indexes refer to NA, VOC 4068, pp. 244–263 (Cape, 1712) and pp. 220–243 (Stellenbosch–Drakenstein, 1712); VOC 4073, pp. 25–33 (Cape, 1714) and pp. 14–24 (Stellenbosch–Drakenstein, 1714). These are the transcribers’ archival locators. Source versions and hashes are recorded in the replication files.

[^5]: Tracing History Trust dagregister transcription, entries vc-1713-90, -99, -108, -123, -130, -162, -214, and -332. For the November passage, NA, VOC 1.04.02, inv. 10732, scan 260. Appendix C and the replication package supply the full 17-entry chronology. The 31 March entry states that the disease had already been spreading for some time (*sedert een geruijmen tijd*); the first retrieved reference is therefore not the epidemic’s first infection.

[^6]: Dagregister, 10 May and 28 November 1713, entries vc-1713-130 and -332; the latter checked against NA, inv. 10732, scan 260. The clerk wrote of *het meerendeel der Hottentots*, the majority of the Khoesan; this proportion is retained as a report, not converted into a mortality estimate.

[^7]: Dagregister, 19 May 1713, entry vc-1713-139; NA, inv. 10732, scan 121. The report uses conditional wording (*zouden hebben*); the attribution and uncertainty are preserved.

[^8]: Dagregister, 17 December 1713, entry vc-1713-351; NA, inv. 10732, scan 275. The image contains both the count and the connection the clerk drew between smallpox, marriage, and the colony’s welfare.

[^9]: Western Cape Archives and Records Service, MOOC8/2.76–77; MOOC8/3.37–39; MOOC8/3.56–59. TEPC transcriptions. The associated document register preserves both identifiers and dates.

[^10]: MOOC8/3.58, 11 January 1714; MOOC8/2.105, declaration dated 8 December 1713 and submission dated 10 March 1714; MOOC8/2.101, 15 January 1714.

[^11]: MOOC8/2.72, 22 May 1713; MOOC8/2.75, 11–12 May 1713; MOOC8/2.76–77, 28 and 30 June 1713.

[^12]: MOOC8/3.37, inventory of 5–8 July 1713; MOOC8/3.38, 9 August 1713; MOOC8/3.39, 11 August 1713, with an added note of 16 April 1714. The text describes recipients as relatives or others with whom the children lived; the distinction between custody of property and custody of a child is retained.

[^13]: MOOC8/3.49, 4 July 1713; MOOC8/2.101, 15 January 1714. A krankbesoeker was a visitor or comforter of the sick; the title is not treated here as proof of medical qualification or of treatment in this particular case. The Oberholster inventory does not date his first wife’s death.

[^14]: MOOC8/3.48, inventories undertaken in July and registered on 31 July 1713.

[^15]: MOOC8/3.37; MOOC8/2.77. The Dutch lists group Tamer with a child and Pasqual with two children. The analysis does not infer additional parentage, ages, or subsequent separations.

[^16]: MOOC8/3.58, 11 January 1714.

[^17]: MOOC8/2.67, 28 April 1713.

[^18]: Dagregister, 14 May 1713, entry vc-1713-134 (*die ’er in de tijd van twee maande wel 95 door verloren heeft*); NA, VOC 1.04.02, inv. 10732, 26 August 1713, scans 176–177. The distinction between the Company population and people enumerated in private returns is maintained throughout.

[^19]: NA, VOC 1.04.02, inv. 10733, 13 February 1714, scans 51–52. Leibbrandt (1896, 258) abbreviates this episode under 15 February. The journal separates the appearance on 13 February from the appointments on 15 February.

[^20]: Supplied transcription, entry vc-1714-46 (*broeder en erfgenaam*; *broeders zoon en erfgenaam*; *zoon en erfgenaam*), which refers back to the entry of 13 February; Leibbrandt (1896, 258) gives the same four names. NA, VOC 1.04.02, inv. 10733, scan 56, shows the Hercules and Kolinga appointments; the portion naming the two brothers was not located in the image sequence used here. Spellings of names vary between manuscript and transcription.

[^21]: Tracing History Trust dagregister transcription, 1 and 15 November 1704 (vc-1704-306 and -320), and 8 November 1708 (vc-1708-313). The latter entry incorporates a Council resolution. Complete entries and relevant passages are retained in the care-context register. These institutional examples are based on the supplied transcription, rather than a new image concordance.

[^22]: NA, VOC 1.04.02, inv. 10730, 17 February 1710, scans 58 and 61–63 (prescription instructions on scan 62); inv. 10731, 13 April 1711, scans 162–167 (hospital labour on scan 165; prescriptions on scan 166). Supplied transcription: vc-1710-48 and vc-1711-103. The latter describes the 16 men as already assigned to the hospital and orders their retention for assistance and cleaning. The prescription requirement concerns a separate record, not the inclusion of each prescription in the journal.