THE ROOT OF THE SCATTER: NINETEENTH CENTURY ARTIFACT … · for root crops to remain unhoused. This prac-tice is borne out in Rempel's study (1967:46), which illustrated floor plans

THE ROOT OF THE SCATTER: NINETEENTHCENTURY ARTIFACT AND SETTLEMENT

PATTERNS IN RURAL ONTARIO

Eva M. MacDonald

In southern Ontario, the most commonhistoric period site type encountered duringcultural resource management (CRM) work isthe small rural farmstead. The fact that thesesites are located most often within the plough-zone, however, can make their analysis chal-lenging as little in the way of structural fea-tures has survived. This paper will discuss thedocumentary evidence, settlement patterndata, and the frequency variations withinartifact groups on 15 sites excavated by theconsulting firm Archaeological Services Inc.(ASI) between 1986 and 1995. Particular atten-tion will be paid to the site formation pro-cesses, in addition to the contextual evidence ofnineteenth century land use, in order to explainthe variation observed within artifact groupsand settlement patterns in this study. Inparticular, it will be suggested that the pres-ence or absence of a root cellar, or other largesubsurface features, would appear to haveimportant implications for the interpretation ofsite formation processes that occurred after thefarmstead was abandoned. Also import ant arethe methods we choose as archaeologists toexcavate these sites.

INTRODUCTION

The idea to research site formation pro-cesses on nineteenth century sites presenteditself following mitigative excavations con-ducted on the Log Cabin (A1Gu-107) site inRichmond Hill, and the James Brown (BbGw-22) site in Barrie. On both sites, over 90 percentof the total artifact assemblage came from afeature interpreted as a root cellar, eventhough nine or more features were documentedduring each excavation (MacDonald and Austin1995; Welsh and MacDonald 1994). Thequestion to be asked in each case was, "Whatprocesses of deposition transformed the rootcellar from a functional feature into anarchaeological feature (Schiffer 1987), and whydid most of the artifacts end up there?" This led

to a review of the types of features identified onsites in general, as well as a search for pat-terns in the types and frequencies of artifactsdeposited in the features. The results of thisresearch are presented below.

ASPECTS OF THE SITEFORMATION PROCESS

Each archaeological site is the product of acomplex set of human actions and environ-mental processes that can result in differentsettlement patterns and artifact frequencies.Michael Schiffer (1987:3) has analyzed this interms of the Site Formation Process, wherebyartifacts are transformed from a systemiccontext into an archaeological context. Thefollowing processes that contribute to theformation of an archaeological site will beoutlined below: the establishment of the farm,the construction of farm buildings and theirdiffering uses, the length of occupation, refusedisposal practices, farm abandonment, andchanging land use. It should be noted thatother variables concerning the site occupants,such as family size, economic status, andethnic affiliation, also contribute to aspects ofthe site formation process, although this datais not presented in this paper.The Nineteenth Century Farmscape

To characterize the archaeological sites interms of what could be learned through archi-val research, the sample represents ruralfarmsteads with one storey log or framehomes, the majority of which were first occu-pied during the 1830s and 1840s (Table 1).

While not the earliest pioneer period inOntario, the 1830s was a time of tremendousgrowth in population as a significant number ofpeople emigrated from England, Scotland andIreland to southern Ontario (Glazebrook 1968:24). Some of the most interesting accounts ofpioneer life in Ontario were written by English

56 ONTARIO ARCHAEOLOGY No. 64, 1997

daily life in Upper Canadain the second quarter ofthe nineteenth century. Ofthe landscape aroundThornhill, she wrote:

You have now a fieldor two quite clearedand almost level; nowseven or eight moreor less dotted withstumps from four tofive feet high; now afield or strip of landthickly set with hightapering poles. Thefences are universallyzig zag walls whichare generally untidy,and at all timesperhaps more pictur-esque than neat. Thecott ag es are moreor less distant fromthe road. They aremostly plank, withbarns at hand [Miller1968: 21].

The site formationprocess begins, there-fore,with the settler whochooses a location andconstructs a house, fen-ces and ancillary build-ings necessary to themen and women whoemigrated during thistime. In 1828, MaryGapper O'Brien camefromEngland to stay with hertwo brothers who werehomesteading inThornhill, north ofToronto.Her diaries provided afirst-hand account of

MACDONALD THE ROOT OF' THE SCATTER:... 57

progressed at the same rate, nor were all partsof a township settled at the same time, produc-ing a "complex mosaic of farmscapes" in agiven area at a given time (Kelly 1975:71). Theancillary buildings came in all shapes andsizes, and these structures were often placedwithout any regard to order or symmetry,reflecting a pattern characteristic of farmersfrom the British Isles (Rempel 1967:70).

The construction material used in the farmhouse in particular is thought to reflect thestage of a farm's development (Wightman1974). The log house was associated with theearliest settlers in Ontario because it reflectedthe use of a material which was a by-productof the forest clearing process (Rempel 1967:34).It was distinguished from the log shanty, whichcould be thrown up in a day if necessary, bythe presence of glazed windows and a chim-ney. The dimensions of the log house weregenerally 16 feet by 20 feet, which conformedto the by-laws passed in many parts of south-ern Ontario to regulate settlement duties in theearly nineteenth century (Rempel 1967:55). Thefloor was made from cleft planks smoothedwith an adze and pinned to logs laid directlyon the ground. The chimney was built of stoneor local brick and the fireplace was betweenfour and six feet wide with a brick or stonehearth (Rempel 1967:66). Earlier chimneys alsoincorporated wattle and daub in the absenceof brick or stone. Untempered, fired clay ad-hering to a stone chimney platform has beendocumented archaeologically on the Gould(BaGs-25) site, excavated by ArchaeologicalServices Inc. in 1997. As described by JohnMcGregor (in Rempel 1967:59-60), who pub-lished his observations of Canada in the 1830s,"a wooden framework placed on a stonefoundation is raised a few feet from theground, and, leading through the roof with itssides closed up with clay and straw kneadedtogether, forms a chimney."

Sub-floor pits, as opposed to full basements,were associated with early house forms. Sam-uel Strickland (1971:1:165) advised settlers toexcavate a small cellar near the fireplace,"commodious enough to hold twenty or thirtybushels of potatoes, [and] a barrel of pork ortwo," as the winter in Canada was too severefor root crops to remain unhoused. This prac-tice is borne out in Rempel's study (1967:46),which illustrated floor plans from log housesbuilt in the 1820s and 1830s with a cellar trap

door indicated in the kitchen or living roomarea. The sub-floor pit could vary in depth andshape and function, however, as evidenced bythe variety of sub-surface pits reviewed in theliterature by archaeologist Richard Kimmel(1993). Kimmel (1993:110) has called for ataxonomy of pits based on archaeological andhistorical documentation, as it is currentlydebated in the United States as to whether ornot hearth-front storage pits can be interpretedas evidence of Afro-American slave occupation.Certainly researchers in Ontario could benefitfrom such a taxonomy, as Robert Mayer(1995:10) has observed that most large semi-subterranean features on historic archae-ological sites are labelled root cellars, yet othertypes of features incorporating pits did exist,as discussed below.

As soon as 15 or 20 acres had been cleared,the settler was advised to erect a frame or logbarn (Strickland 1971:1:170). The earliest barnswere based on European models and theyevolved as they were applied to the NorthAmerican experience (Ennals 1972). The typeof agriculture that was practised determinedthe form of the barn. In southern Ontario priorto 1850, wheat farming was predominant, andfew implements other than a plough and a fewhand tools were required. The harvested graincould be stored in a relatively small amount ofspace and was threshed as it was needed(Ennals 1972:265). Consequently, the first barnswere one-storey log or frame structures, set onthe ground or on a course of stones, accommo-dating two storage bays and one centralthreshing floor (Rempel 1967:191). Specialpurpose pits were also excavated under barnstructures to keep dairy products cool andstore root crops grown for animal fodder(Glazebrook 1968:39).

Farmers generally left their cattle exposed tothe elements year-round to fend for them-selves. Horses and working oxen, however,would be sheltered during the winter in theirown stable, perhaps within the two-bay barn(Ennals 1972:265). Two-level barn structureswith an earthen ramp and stables did notbegin to appear in the farmscape on a regularbasis until after 1870, when wheat farming wasgradually superseded by a mixed agriculturethat integrated livestock with cultivation (En-nals 1972:267).

The three household-related buildingsessential to every farmstead in earlier times


MACDONALD THE ROOT OF THE SCATTER:.. . 59

were the ice house, woodshed, and privy (Noble1984:85). Almost any kind of small framebuilding could serve as a woodshed, used tokeep the wood dry for use in the fireplace orstove (Noble 1984:86). The ice house requireda pit to be excavated in addition to construc-tion of a building. Mary Gapper O'Brien notedthat while her brother had hired a workman todig the ice house pit and finish the walls inNovember, the ice house itself was not builtuntil the following January as it was importantfor the ground beneath the structure to befrozen before it was filled with ice (Miller 1968:25-26). The pit itself may have been lined withmortar or stone, and the structure probablyhad double plank walls filled with an insulatingmaterial such as wood shavings, straw, leavesor bark (Noble 1984:85).

With regards to the privy, J.H. Hammondwrote:

There is no building which is so gen-erally located in the wrong place, asthat diminutive house to which aname is applied that expresses abso-lute importance of such a retreat. It isstrange that a house which everyoneis ashamed to be seen to enter,should be in one of the most conspic-uous positions so that from all backwindows of the dwelling house, it isthe most apparent object in view[Hammond 1858:150].

It was an improvement when the privy wasplaced next to a shed, so that "the unfortunateperson who was obligated to retire to it mightskulk around the shed, and allow it to be con-jectured that he might have gone on some lessignoble errand" (Hammond 1858:150). On ruralsites, the privy was usually a frame buildingthat could be easily removed to permit periodiccleaning of the vault excavated underneath it(Noble 1984:87). Alternatively, in lieu of exca-vating a vault, a portable receptacle such as abucket could be used in the privy and thewaste could be "recycled" on adjacent fields(Geismer 1993:59).

Water-related structures were also importanton the farm (Noble 1984:81). Where springswere not present, the family depended uponthe well from which to draw water. These werelocated close to the house for convenience.Often a mechanical pump was employed to

make it easier to lift the water, or if the watertable was close to the surface, a "draw-well"could be employed. According to CatharineParr Traill, the plan was simple:

A long pole, supported by a post, actsas a lever to raise the bucket, and thewater can be raised by a child withvery trifling exertion. This method isby many persons preferred to eitherrope or chain, and from its simplicitycan be constructed by any person atthe mere trouble of fixing the poles[Traill 1929:71-72].

One household-related structure not foundon every farm was the outdoor bake oven.Many regional and ethnic variations have beendocumented in North America (Noble 1984:92-97), and Catharine Parr Traill (1929:71) foundthe beehive shape of the clay-built bake ovento be a familiar sight as she travelled in east-ern Ontario during the summer of 1832.

Additional outbuildings that may have beenpresent included crop storage and processingfacilities such as root houses and smokehouses. The root house was different from theroot cellar in that it was not an internal, house-hold feature. Mary Gapper O 'Brien describedthe root house excavated by her brother inOctober of 1829:

Southby has finished wheat sowingand is now making a root house, thatis, a house to preserve roots from thefrost. This is a pit of six or eight feetdeep and roofed with logs which areagain covered with earth [Miller 1968:71].

In November of 1828, her brothers wereconstructing "a smoke house for drying thebacon and hams on three pigs for the comingyear" (Miller 1968:26). Thomas Kenyon (1983)documented a number of smoke houses builtfrom a variety of materials in Haldimand andNorfolk counties, all of which were square torectangular structures erected over a centralfire pit. In general, they were located close tothe farmhouse. Butchering was done in the latefall and during winter to reduce livestock andconserve precious grain. Smoking the meatpreserved it as well as improved the flavour.

While not exhaustive, the above descriptions


illustrate the variety of buildings that weretypical of the rural farmscape in nineteenthcentury Ontario. Finally, in addition to distin-guishing the types of structures, it is also im-portant to note the length of site occupation asrevealed by the land use history. The sample inthis paper ranges from sites occupied brieflyby one family, to multi-component sites occu-pied by more than one family (Table 1). Thiscould help to explain the increased number offeatures and artifacts on sites occupied bysuccessive generations or different familiesover a long time period. Overlapping occupa-tions may also result in the obliteration of earlyfeatures.

Refuse Disposal

Another important factor in the site formationprocess is the pattern of refuse disposal on thefarm. Stanley South's "Brunswick Pattern," forexample, defined the different waste streamson eighteenth century Anglo-American sites inorder to predict the location of points of entryand exit to dwellings (South 1977:47-82). Thiswas based on the assumption that refusewould be deposited adjacent to the home,primarily at the back, but also at the frontdoorway, and in nearby depressions (South1977:47). South also predicted that refuse witha high food bone content would be discardedaway from the home where the smell would notbe so offensive. As a surface deposit, themidden was subject to displacement throughlandscaping activities around the house, andthrough animals such as pigs, rats and dogsscavenging for food (South 1977:48).

Refuse could also be placed in non-function-ing features, such as abandoned cellar pits,which would create stratified subsurfacedeposits while the farm was still occupied.Holes were also dug to bury garbage, asevidenced by the number of nondescript"refuse pits" containing artifacts that aredocumented on archaeological sites (Table 2).The disposal of refuse in a "back forty" dump,located at a greater distance from the house,was more common during the late nineteenthcentury. This phenomenon was documented atthe Pickard (AhGx-24) site in Ancaster, wherethe surface scatter originating from a houseoccupied between 1822 and 1839 was mixedwith refuse dumped on the site in the earlytwentieth century by the occupants of a differ-

ent house located approximately 500 metresaway (Welsh and MacDonald 1996:20).

Finally, privy pits were also used for refusedisposal, both after they had been abandonedand while they were still in use. For example, astudy of beetle fauna collected from a stratifiedseventeenth century privy in Boston, Massa-chusetts, concluded that the beetles originatedfrom yard and kitchen wastes, as well as housesweepings periodically deposited in the vault(Bain 1998:45).

House Abandonment and Changes in Land

Use

Eventually the occupation of the farm houseended. In southern Ontario, the broader pro-cess of rural house abandonment took placeduring two periods in the nineteenth and earlytwentieth centuries (Kenyon 1995). The firstperiod is characterized by the replacement ofthe original log house with an improved frame,brick or stone structure. While not all loghouses disappeared immediately, as theycould be used as ancillary buildings when thenew structure was built (Kenyon 1995:8), thechange in land use signalled an end to anearlier domestic pattern, and could potentiallyobliterate early features on a multi-componentarchaeological site.

The final change in land use transformed thefarm house and yard into an agricultural field,and several choices could be made at thistime. The architectural hardware could besalvaged and taken away from the site, alongwith the lumber, or the buildings could berazed and burned to remove them quickly. Thebuilding foundation, if one existed, chimneyand hearth stones may also be removed so asnot to impede the plough. The former houselocation could also be used as a place to dumprefuse, a pattern noted above on the Pickardsite in Ancaster.

The fact that many farm buildings incorpo-rated an excavated pit is also significant to thesite formation process. To protect livestock andfarm machinery from falling into the holes, itwould be expected that the abandoned wells,cellars, privy pits and root houses would befilled in opportunistically after they were nolonger needed, incorporating the surfacemiddens as secondary fill, and possibly incor-porating structural material to help fill in the

hole. This razing process has beendocumented archaeologically, andcan be detected based on certainstratigraphic traits related to theactivity of filling in an open holeonce a building has been pulleddown (White and Kardulias1983:70).

Potentially, the balance of themidden remained in what wouldbecome the ploughzone as theland use of the former homelotchanged. Hence, when an archae-ological site is stripped of its top-soil in order to expose the settle-ment pattern, it should be notedthat the use of a Gradall also con-tributes in a major way to the siteformation process, as part of thesite has been removed before thearchaeologist begins to record thefeatures.

ANALYSIS OFSETTLEMENT PAT ERN

AND ARTIFACTFREQUENCY DATA

Eight artifact group categorieshave been used to organize thedata in this study: Kitchen Group,including ceramics, domestic glassand utensils; Faunal Re-mains;Architectural Group, includingnails, window glass and hardware;Tobacco Pipe Group; ClothingGroup, including buttons, sewingitems, and shoe parts; PersonalGroup, including slate pencils,grooming items and coins; ActivityGroup, including livery, farmtools, arms and toys; and finally,Unidentified Artifacts. Thesegroups are loosely based on South(1977:95-96) and are utilized inthis paper as a means tocommunicate information in func-tional groupings.

Each of the 15 sites was discov-ered during Stage 2 archaeologicalassessments conducted insouthcentral Ontario (Figure 1).Eleven of the sites were discov-

MACDONALD THE ROOT OF THE SCATTER:... 61

ered as scatters of material in ploughed fields.The remainder were discovered through testpitting, although these too had been subject tocultivation to some degree prior to their discov-ery. The topsoil from each site was removed byGradall during a Stage 4 salvage excavationbefore the features in the subsoil were re-corded and hand-excavated. In general, thetopsoil was not sampled in order to concen-trate efforts on the collection of settlementpattern data and artifacts with a feature pro-venience, hence the data are derived fromartifacts with a feature provenience.

While it is not standard practice amongCRM firms in Ontario to conduct anythingbeyond a controlled surface collection duringStages 2-3 investigations on historic sites, itshould be noted that the Archaeological As-sessment Technical Guidelines (Stages 1-3 &Reporting Format) for Ontario stipulate that,"for most sites it will be necessary to determinethe nature of ploughzone and subsoil artifactdeposits and/or the potential for the presenceof cultural features, stratigraphy or buriedmidden deposits during Stage 3 investigations"(MCTR 1993:6). The Guidelines make nodistinction between prehistoric and historicsites. Elsewhere in the Northeast, the plough-zone is sampled routinely to varying degrees,pointing to a philosophical difference in howvaluable the data in the ploughzone is per-ceived. In the state of Maryland, for example,the use of a Gradall is not permitted on anarchaeological site until the ploughzone hasbeen adequately sampled and subsurfacefeatures have been exposed (Shaffer and Cole1994:29). Different strategies have been devel-oped and are used by CRM archaeologists inthe Northeast to extract useful data from theploughzone, and it would be worth adoptingsome of the strategies here in Ontario. Theseinclude taking samples for soil chemical signa-tures that are used to interpret activity areaswithin a site (cf. Pogue 1988), as well as sam-pling the ploughzone and analysing the pat-terns of artifact distribution contained therein.

Settlement Pattern Data

The number of features to be found on a siteis variable (Table l), ranging from as few astwo, as on the Robert Smith (AkGw-55) site inBrampton, to 69 at the Devil's Pulpit (AlGx-9)site in Caledon. Assuming that the site occu-

pants probably utilized an area larger thanthat stripped by the Gradall, the variability inthe number of features may be related to theamount of site surface exposed by the Gradall,but there does not appear to be a direct ratiobetween the number of features and the totalarea of topsoil stripped. The total number ofartifacts per site is also variable, ranging from171 artifacts to 9,099, excluding the ploughzone(Table l).

Of the 258 features in the sample, only 21percent (n=55) could be classified according toa functional type other than "refuse filleddepression," "pit," or "unknown" (Table 2).These included 13 "structural" features, nineroot cellars, six animal burials, six post pits,four wells, four midden deposits, three drains,three privies, three chimney platforms or fire-places, two landscape features, one bake ovenand one cistern.

The "structural" features are interpreted asthe location of a former structure based on thepresence of materials such as flagstones,foundation remnants containing mortar, brickor stone, remnants of wooden boards, orattached drainage features, in addition to theirmorphology. For example, Feature 2 at theHighbush (AkGs-16) site (Figure 2) in Pickering,and Feature 1 at the Pickard site (Figure 3),were interpreted as the locations of formerstructures based on the combination of severalof the above attributes. In general, this type offeature is relatively shallow, ranging from 16 to32 cm deep, and the soil fills are often lensedor mottled with ash and charcoal, as for exam-ple Feature 1 at the Pickard site (Figure 4). Itwas postulated that these features were form-ed by the removal of an existing structure, andthat the building footprint subsequently be-came filled with refuse (MacDonald and Austin1994:8; Welsh and MacDonald 1996:7).

Indeed, if one compares the Artifact Groupfrequencies in six features identified as thelocation of a former structure (Table 3), theKitchen Group frequency is highest on theLoring (AIGw-43) and Wilkinson (AjGw-140)sites, suggesting the formation of a midden inthese features. Upon re-examination of thedata, it is also possible, however, that Feature 2at the Highbush site represents the location ofa sub-floor pit due to the relatively higherfrequency of artifacts in the Activity, Personal,Clothing and Tobacco Pipe Groups. HeatherHenderson (personal communication 1997) has



Figure 1. Location of Sites used in the Studysuggested that rectangular, relatively shallowfeatures such as that at Highbush may repre-sent the remains of sub-floor pits, due to thehigher frequency of these particular artifactgroups. The pattern is consistent with the lossof smaller, personal items indoors. The highfrequency of Architectural Group artifacts,principally nails, in Feature 5 of the FletchersCreek (AjGw-84) site in Brampton, however, iswhat one would expect of a wooden structurethat was left intact after the site was aban-doned (Table 3).

Only two intact house foundations weredocumented among the 13 "structural" fea-tures in the study sample. At the Flynn (AjGw-69) site in Mississauga, the mortared field-stone foundation measured six metres by eightmetres, with internal walls and floor joistsintact in several locations (Figure 5). An explor-atory trench was hand-excavated within thestructure, in front of the fireplace, but fewartifacts were found. The house including theroot cellar had been filled solidly with demoli-tion rubble. The bulk of the artifacts came frommidden deposits to the north and east of thehouse, conforming to South 's "BrunswickPattern" of refuse disposal (MacDonald and

Williamson 1992:10). Intact fireplaces such asthat at Flynn were not common in this study,however, a rectangular concentration of largestones and ash within a darker soil matrix inFeature 2 at the Pickard site (Figure 6) proba-bly represents a chimney platform shorn of thechimney itself (Welsh and MacDonald 1996:7).

A more common settlement pattern in thisstudy is one where no contiguous foundationsexisted to guide the interpretation of the home-lot pattern. However, some features may repre-sent portions of a foundation that can be usedto interpret the location of the house, such asthat documented at the Lampman (AhGx-96)site in Ancaster (Garner and Austin 1996:6).Feature 4 at Lampman (Figure 7), for example,a rectangular basin-shaped feature with brickfill, also contained two coins dated to between1790 and 1830 that may have been placed inthe foundation of the original structure for goodluck (Garner and Austin 1996:7). At the Devil'sPulpit site (Figure 8), some of the pit featuresmay represent holes dug to support woodenuprights for simple frame structures, as well asfence posts, or planting holes, as they con-tained relatively few artifacts (Austin andWilliamson 1990:8), however, the pattern is not

Figure 2. Settlement Pattern of the Highbush Sit

straightforward. On other sites such as JamesBrown (Figure 9), the root cellar feature wasinterpreted as the former location of the house,based on nineteenth century descriptions citedabove (Welsh and MacDonald 1994:6).

Root cellars were identified on nine of the 15sites in the sample (Table 2). As was noted in astudy of Illinois cellar excavations (Phillippeand Walters 1987:41), the interpretation ofthese features as root cellars is based primar-ily on their morphology, and physical relation-ship to other features, as opposed to theirartifact content, as they generally do not con-

tain de facto refuse. In Ontario, Mayer (1995:10)has also observed that most root cellars con-tain secondary deposits of artifacts.

In this study sample, the majority of rootcellars were square to rectangular pits with flat

basin profiles, plain dirt walls and hard-pack-ed floors. In some instances, a wooden liningwas evident, as at the Highbush site (Feature 5,Figure 2). The depth below grade (after theploughzone had been removed) of the cellarfeatures ranged from between 39 cm at theHighbush site, where the subsoil was predomi-nantly clay, to 115 cm at the Devil's Pulpit site,where the subsoil was sandy. The absence of aramped entrance indicates that these featuresare sub-floor pits within the house asdescribed by Samuel Strickland above. Onlyone "Keyhole Cellar" (Phillipe and Walters1987:41) with exterior entrance was docu-mented — at the Flynn site (Figure 5). Little orno stratigraphy was noticeable in many of theroot cellar features, and the fill often includedlarge, flat field stones placed at the bottom of


the pit that may have originated from thefireplace or house foundation. At the LogCabin site, the fill had been placed on top of athin, dark soil layer (Figure 10) interpreted hereas organic material that had been worked intothe floor when the cellar was in use.

A more complex, stratified deposit wasfound in the root cellar at the James Brown site(Figure 11), including fire-reddened soil andcharred building debris, which suggest thatpart of a superstructure was burned in the pitafter the house was abandoned (Welsh andMacDonald 1994:6). An underlying soil layercontained bone that had been gnawed by

rodents and dog-sized carnivores (Thomas1994:Table 5), indicating that the bone hadbeen exposed in a surface deposit before itwas placed in the cellar hole as a secondarydeposit. Gnawed bone was also common inthe Log Cabin and Pickard site assemblages(MacDonald and Austin 1995:17; Welsh andMacDonald 1996:18).

On sites where no sub-floor pit or root cellarwas identified, as at Lampman, the circular,basin-shaped pits located close to the proba-ble house foundation may have been dug forstorage, and then filled with refuse when theywere no longer used for their original purpose.


Figure 4. Feature 1, The Pickard Site

The dearth of privy features in the studysample is notable (Table 2). Privy pits with drylaid stone walls are a common feature type onurban archaeological sites. As the dense,organic waste material expected in these

features is not always present, and their sizeand shape may vary, they are often identifiedby the presence of floral indicators, such asraspberry and strawberry seeds (Geismar1993). Urban privies, including the one at


Inge-Va in Perth, Ontario, which containedover 15,000 artifacts (Doroszenko 1993), arenoted for their massive deposits of ceramicsand glassware (Beaudry 1994).

In this sample of rural farmsteads, however,privy features were identified on only two of the

15 sites, Langstaff (AlGu-95) in Richmond Hill,and Pickard (Table 2). Interestingly, at theLangstaff site, over half of the site assemblagewas concentrated in this feature, including agreat amount of architectural hardware, and afine collection of smoking pipes (Austin et al.


Figure 6. Feature 2, The Pickard Site

1992:11). It is possible that more privies werenot located because the variant that employedthe removable bucket was preferred, or be-cause they were capped with sterile soil thatdid not contribute to the surface scatter ofartifacts disturbed by the plough, and hencelay beyond the estimated site area that wasstripped of topsoil during the Stage 4 salvageexcavation.

The two privies (Features 5 and 6) at thePickard site, for example, were located ap-proximately 20 metres to the north of the maincluster of features (Figure 3), and may haveonly been uncovered because an area greaterthan the surface scatter was stripped by the

Gradall due to the presence of a prehistoriccomponent. The privy features were rectangu-lar, stratified, basin-shaped pits with basaldeposits of darker soil (Figure 12). A similarfeature was identified as a privy at the multi-component Southdale Site (AfHh-35) insouthwest Ontario (Timmins 1990:7). It wasdocumented 20 metres away from the maincluster of historic features and had cut into thewall of a prehistoric Iroquoian longhouse. Theprivy appeared to lie beyond the historicalartifact surface scatter and may only havebeen uncovered because of the longhouseinvestigation (Timmins 1990:Figures 2, 3).Alternatively, privies have not been identified


Figure 7. Settlement Pattern of the Lampman Site

on the majority of sites because the featureslacked the classic indicators such as seeds ordark "night soil." Perhaps there is a privy

variant on rural sitesthat has yet to be identi-fied with confidence.

The dearth of barns,stables, granaries andother farm-related out-buildings in the sampleis also interesting, al-though their absence inthe archaeological re-cord has been notedelsewhere in the North-east (Garrison 1996).Again, it is possible thateither they have notbeen recognized withinthe pattern of posts andfeatures recorded for agiven archaeological site,that their constructionand use has left nothingto the archaeologicalrecord, or not enough ofthe total site area hasbeen investigated. Moreattention should be paidto interpreting the entirefarmscape settlementpat-tern, not just thehouse and immediateyard. At a minimum, it issuggested that at least a20 metre radius beyondthe feature(s) interpretedas the former houselocation should beinvestigated, in or-der todocument a morecomplete homelot pat-tern.

Artifact GroupFrequencies

A comparison of theartifact group frequen-

cies in the 15 site as-semblages also pointsto the potential differ-

ences in the site forma-tion processes, as the

frequencies arewide-ranging in the

Kitchen, Faunal andArchi-

tectural Groups (Table 4). If one searches for


patterns in the site formation processes, how-ever, similarities between sites become appar-ent. Two of the three sites where remnantmidden deposits were hand-excavated in one

metre squares, Flynn and Devil's Pulpit, alsocontained a large percentage (72 percent and65 percent respectively) of Kitchen Groupartifacts. The frequency of Faunal Remains


Figure 9. Settlement Pattern of The James Brown Site

was relatively low, between 10 percent and 12percent, perhaps reflecting the poor rate ofpreservation in an open-air context, and theArchitectural Group ranged between 13 per-cent and 18 percent. These sites were alsodistinguished by the fact that the assemblageswere concentrated in these middens; 77 per-cent of the artifacts from Devil's Pulpit, forexample, were found in the slope midden,while the root cellar contained only 20 percentof the assemblage.

In contrast, although a sheet midden wasidentified at the Lampman site, Kitchen Groupartifacts made up only 44 percent of the as-semblage (Table 4). This may be due to thefact that 50 percent of the excavated artifact

assemblage was concentrated in Feature 28,a well, rather than in the midden (Figure 13).Forty-eight percent of the well assemblageitself was comprised of Kitchen Group arti-facts, however, the Architectural Group madeup a further 32 percent of the well assemblage,reflecting a deposit of structural debris.

The effect that the site formation process hason a site assemblage can also be illustrated byexamining the James Brown and Devil 's Pulpitsite assemblages by comparing the frequencyof the artifact groups within each featurecontext. Where the midden was found to berelatively intact, and hand-excavated duringStage 4, as at the Devil's Pulpit site, the fre-quency of artifacts in the Kitchen Group is high


Figure 10. Plan and Profile of Typical Root Cellar Feature, The Log Cabin Site

(Figure 14). Where no midden is identified, andover 90 percent of the assemblage originatedin the root cellar, as at the James Brown site,

the frequency of Kitchen Group artifacts isgreatly reduced, relative to other classes thatdominate the assemblage (Figure 15). For


Figure 11. Plan and Profile of Root Cellar Feature, The James Brown Site

example, close to 50 percent of the JamesBrown assemblage consisted of architecturalhardware, confirming the deposition of struc-tural debris within the root cellar feature afterthe site had been abandoned and the housedismantled.

Trends can also be seen in a comparison of

the artifact group frequencies in seven rootcellar assemblages with more than 100 arti-facts. Where more than 80 percent of the totalsite assemblage originated as secondarydeposit in a root cellar, and the cellar is theonly major feature identified, as at the LogCabin, James Brown and Robert Smith sites,


the frequency of artifacts in the Kitchen Groupis low (Table 5). Concomitantly, the frequencyof Faunal Remains is relatively higher, indicat-ing the better rate of bone preservation in asealed context.

On the Highbush, Loring and Flicka (AkGw-18) sites, however, the root cellar representedonly a quarter to half the site assemblage,indicating that the artifacts were depositedamong more subsurface features on thesesites. In these cases, the root cellar proveni-ence contained a greater frequency of KitchenGroup artifacts, and a lesser amount of animalbone (Table 5). These frequencies fall closer tothe ranges of those on sites where a middendeposit has remained relatively intact. Itshould be noted that the relative length ofestimated site occupation does not appear tobe a factor. Although one could argue thatthere were fewer Kitchen Group artifacts to befound on the Log Cabin, James Brown andRobert Smith sites because there was less timefor the artifacts to accumulate on these sites,the Highbush site was also thought to havebeen briefly occupied (Table l).

SUMMARY ANDCONCLUSIONS

This paper has provided a review of thedocumentary evidence, settlement patterndata, and frequency variations within artifactgroups on 15 rural farm sites occupied duringthe nineteenth century in southern Ontario. Allof these archaeological sites lacked intact,above-ground structures. In order to interpretthe apparent differences among the sites, withrespect to the artifact group frequencies withinthe various features, processes of site forma-tion have been examined.

To briefly summarize the trends identifiedwithin the sample, where a midden deposit hasbeen identified and excavated as a feature, 65percent or more of the total assemblage iscomprised of Kitchen Group artifacts. Thisgroup is reduced to under 30 percent of thetotal site assemblage on sites where artifactsare concentrated as a secondary fill within onelarge feature, such as a root cellar. This trendin the relative reduction of Kitchen Groupartifacts was also noticed on sites with deepfeatures such as wells and privies, where atleast half of the site assemblage was concen-trated in those features. Architectural Group


artifacts, by contrast,are better representedin these features,indicating the reuse ofthese features as aplace where structuraldebris could be aban-doned. These pat-terns should be con-sidered preliminaryand should be testedwith an expandedsample to confirmtheir validity.

The observed pat-terns can be ex-plained partially bythe context of thesesites, located as theyare in agriculturalfields. The middendeposits that hadformed around thestructures have beenused opportunisticallyto fill the cellar holes,privies and well shaftsonce the structuresare abandoned,resulting in a sortingprocess apparentlyb i a s e d aga ins tK i tchen Groupartifacts. Thisphenomenon wasdocumented by Kath-leen Wheeler (1995),who re-evaluated thesocial status of thewidow Rider in Ports-mouth, New Hamp-shire. Wheeler com-pared the assem-blages from an open-air kitchen middenand a privy in thewidow 's yard, andconcluded that theartifacts in the privyfeature in no way re-flected the status ofthe widow as deter-mined by archival


Figure 12. Plan and Profile of Privy Feature, The Pickard Site

Figure 13. Bar Chart Showing Artifact Distribution by Class and Feature Context for The Lampman Site

sources. Instead, the midden was more indica-tive of the day to day activities in the Riderhousehold, as the privy had been filled selec-tively, as well as cleaned periodically of some ofits artifacts.

More importantly, the observed patterns inthe data have implications for what the exca-vation techniques and analysis of the features

on these sites can tell us about the formeroccupants. For example, how is the MeanCeramic Date of the site or an economic scalinganalysis of the ceramic assemblage affectedwhen it would appear that on some sites, partof the ceramic assemblage is missing? ASI iscurrently collecting ploughzone data duringStage 3 assessments on historic


Figure 14. Bar Chart Showing Artifact Distribution by Class and Feature Context for a Site with a RelativelyIntact Midden (Devil's Pulpit)

Figure 15. Bar Chart Showing Artifact Distribution by Class and Feature Context for a Site with a DominantRoot Cellar Feature (James Brown)

sites, and it is hoped that further research willdetermine what percentage of the site area inthe ploughzone should be retained before thetopsoil has been removed to record the settle-ment pattern. Also, as demonstrated above onthe Pickard and Southdale sites, it can beexpected that additional features would bediscovered if an area of the ploughzone great-er than that of the original scatter of artifactswas removed by the Gradall during Stage 4investigations. The questions raised in thispaper can only be answered if more attention

is paid to the site formation processes, includ-

ing how we choose to excavate historic sites.

Acknowledgements. This paper began itslife as a presentation at the 29th annual meet-ing of the Council for Northeast HistoricalArchaeology (CNEHA), held in Louisbourg,Nova Scotia, in 1995. I would like to extend mysincere thanks to my colleagues at Archaeo-logical Services Inc., particularly Ron William-son and David Robertson; members of CNEHA,especially Mary Beaudry, Dena Doroszenko


and Heather Henderson; and Neal Ferris andone anonymous reviewer, for their criticalevaluation of the ideas presented in earlierdrafts of this paper. David Robertson draftedthe figures. Finally, for any archaeologistinterested in the historic period in Ontario, thewritings of Ian Kenyon will always remain aninspiration and cannot go without acknowl-edgement. This paper is respectfully dedicatedto Ian's memory.

REFERENCES CITED

Austin, S., E. MacDonald and A. Allan

1992 Archaeological Services Inc.: Archae-

ological Salvage Excavation of the

Langstaff Site (AIGu-95), Part Lot 37,Concession 1 EYS, Town of RichmondHill, Ontario. Ms. on file, Ontario Min-istry of Citizenship, Culture and Rec-reation, Toronto.

Austin, S., and R. Williamson1990 Archaeological Services Inc.: Archae-


Devils Pulpit Site (AIGx-9), Lot 9, Con-

cession 1 EHS, Town of Caledon,Ontario. Ms. on file, Ontario Ministryof Citizenship, Culture and Recre-ation, Toronto.

Bain, A.1998 A Seventeenth-Century Beetle Fauna

from Colonial Boston. Historical Ar-

chaeology 32(31:38-48.

Beaudry, M.

1994 "Above Vulgar Economy": Material

Culture and Social Positioning amongNewburyport's Merchant Elite. Paperpresented at the 27th AnnualConference of the Society forHistorical Archaeology, Vancouver,B.C. Ms. on file, Department of Ar-chaeology, Boston University, Boston,Massachusetts.

Doroszenko, D.

1993 One Party Too Many? Container andTable Glass from the Inge-Va Privy.Paper presented at the 26th AnnualConference of the Canadian Archae-ological Association, Montreal. Ms. onfile, Ontario Heritage Foundation,Toronto.


Ennals, P.1972 Nineteenth-Century Barns in South-

ern Ontario. Canadian Geographer16(3):256-270.

Garner, B. and S. Austin

1996 Archaeological Services Inc: Archae-ological Salvage Excavation of theLam pm an Site (AhGx-96, City ofHamilton, Ontario. Ms. on file, On-tario Ministry of Citizenship, Cultureand Recreation, Toronto.

Garrison, J. R.1996 Remaking the Barnyard. The Archae-

ology of Farm Outbuildings in theConnecticut River Valley of Massa-

chusetts, 1770-1870. In Historical Ar-chaeology and the Study of American

Culture, edited by L. DeCunzo and B.L. Herman, pp.359-394. The HenryFrancis DuPont Winterthur Museum,Inc., Winterthur, Delaware.

Geismar, J.1993 Where is the Night Soil? Thoughts on

an Urban Privy. Historical Archaeol-ogy 27(2):57-70.

Glazebrook, G. P. de T.

1968 Life in Ontario: A Social History. Uni-versity of Toronto Press, Toronto.Hammond, J.

1858 The Farmers and Mechanics PracticalArchitect. John P. Jewett, Boston. Kelly,K.

1975 The Impact of Nineteenth CenturyAgricultural Settlement on the Land.

In Perspectives on Landscape andSettlement in Nineteenth CenturyOntario, edited by J. Wood, pp. 64-77.McClelland and Stewart Limited,Toronto.

Kenyon, I.1995 "Weeds Upspring Where the Hearth

Should Be": Rural House Abandon-

ment in Southern Ontario. Kewa1995(6):2-16. (editor's note: now see

Kenyon, this volume).Kenyon, T.

1983 Nineteenth Century Notes: Smoke-

house. Kewa 1983(8).Kimmel, R.

1993 Notes on the Cultural Origins and

Functions of Sub-Floor Pits. HistoricalArchaeology 27(3):102-113.

MacDonald, E., and S. Austin

1994 Archaeological Services Inc.: Archae-ological Salvage Excavation of the

Highbush Site (AkGs-16), Part Lot 31,Concession 1, Town of Pickering,Ontario. Ms. on file, Ontario Ministryof Citizenship, Culture and Recre-ation, Toronto.


ological Salvage Excavation of the LogCabin Site (AIGu-107), Part Lot 55,

Concession 1 WYS, Town of Rich-mond Hill, Ontario. Ms. on file, On-tario Ministry of Citizenship, Cultureand Recreation, Toronto.

MacDonald, R., and R. Williamson1992 Archaeological Services Inc.: Archae-

ological Excavation of the Flynn Site

(AjGw-69). Ms. on file, Ontario Minis-try of Citizenship, Culture and Recre-ation, Toronto.

Mayer, R.1995 Some Research Notes on Root Cellars,

Ice Houses and Coolers. Arch-Notes95(6):10-17.

Miller, A.

1968 The Journals of Mary O'Brien (1828-

1838). MacMillan, Toronto.Ministry of Culture, Tourism & Recreation

(MCTR)1993 Archaeological Assessment Technical

Guidelines (Stages 1-3 & ReportingFormat). Cultural Programs Branch,Archaeology & Heritage Planning,Toronto.

Noble, A.

1984 Wood, Brick &Stone: The North Amer-

ican Settlement Landscape. Volume 2:Barns and Farm Structures. Univer-sity of Massachusetts Press, Amherst.

Phillippe, J., and W. Walters Jr.1987 Rats, Damp and Foul Miasma: Some

Thoughts on the Literature and Ar-chaeology of Pioneer Cellars. Wiscon-

sin Archeologist 67(l):37-46.Pogue, D. J.

1988 Anthrosols and the Analysis of Ar-chaeological Sites in a Plowed Con-text: The King's Reach Site. NortheastHistorical Archaeology 17:l-36.

Rempel, J.

1967 Building With Wood. University ofToronto Press, Toronto.


Schiffer, M.1987 Formation Processes of the Archaeo-

logical Record. University of NewMexico Press, Albuquerque.

Shaffer, G., and E. Cole

1994 Standards and Guidelines for Ar-chaeological Investigations in Mary-land. Maryland Historical Trust Tech-nical Report Number 2. Office of Ar-chaeology and Office of PreservationServices, Maryland Historical Trust,Crownsville, Maryland.

South, S.

1977 Method and Theory in Historical Ar-chaeology. Academic Press, New York.

Strickland, S.

1971 Twenty-seven Years in Canada West,or the Experience of an Early Settler.2 vols. Reprinted. M.G. Hurtig Ltd.,Edmonton. Originally published 1853.

Thomas, S. C.1994 Inventory of the James Brown Faunal

Assemblage. In Archaeological Ser-vices Inc.: Archaeological SalvageExcavation of the James Brown Site

(BbGw-22), Part Lot 3, Concession 12

(Former Township of Innisfil), City ofBarrie, Ontario, Appendix A. Ms. onfile, Ontario Ministry of Citizenship,Culture and Recreation, Toronto.

Timmins, P.1990 Southdale: A Multi-Component His-

toric and Prehistoric Site in London,Ontario. Museum of Indian Archaeol-ogy Research Report 20. University ofWestern Ontario, London, Ontario.

Traill, C. Parr

1929 The Backwoods of Canada. Re-printed. McClelland & Stewart Lim-ited, Toronto. Originally published1855.

Welsh, B., and E. MacDonald1994 Archaeological Services Inc.: Archae-

ological Salvage Excavation of theJames Brown Site (BbGw-22), Part Lot3, Concession 12 (Former Township of

Innisfil), City of Barrie, Ontario. Ms. onfile, Ontario Ministry of Citizenship,Culture and Recreation, Toronto.



Pickard Site (AhGx-24), Historic Com-ponent, Part Lot 51, Concession 3,

Town of Ancaster, Ontario. Ms. onfile, Ontario Ministry of Citizenship,Culture and Recreation, Toronto.

Wheeler, K.

1995 Mary Rider as "Active" Widow. Paperpresented at the 28th Annual Confer-ence of the Society for Historical Ar-chaeology, Washington, D.C. Ms. onfile, Independent ArchaeologicalConsulting, Portsmouth, New Hamp-shire.

White, J., and P. Kardulias1983 The Dynamics of Razing: Lessons

from the Barnhisel House. HistoricalArchaeology 19(l): 65-75.

Wightman, W.R.1974 Construction Materials in Colonial

Ontario 1831-61. In Aspects of Nine-teenth-Century Ontario: Essays Pre-sented to James J. Tolman, edited byF. Armstrong, H. Stevenson, and J.Wilson, pp. 114-134. University of To-ronto Press, Toronto.

Eva M. MacDonald

Archaeological Services Inc.528 Bathurst Street, Toronto, Ontario M5S 2P9


Documents

THE ROOT OF THE SCATTER: NINETEENTH CENTURY ARTIFACT … · for root crops to remain unhoused. This prac-tice is borne out in Rempel's study (1967:46), which illustrated floor plans