Progress in Old World Palaeoethnobotany, Van Zeist. Wasylikowa & Behre (eds) C »99 r Balkema, Rotterdam ISBN 90 6191 8812

Western Continental Europe 13

CorrieC. BakelsInstitute of Prehistory Leiden University Netherlands

ABSTRACT: Subject of this contribution are the published seed and fruit remains from theNetherlands, Belgium, Luxemburg and northern France. The material is reviewed in a chronologi-cal order, beginning with the Palaeolithic and ending with the Middle Ages and Later HistoricalPeriods. Where possible, the interpretation of the botanical data by the various authors has beensummarized. Most of these interpretations are of an economic nature.

1 INTRODUCTION

Western continental Europe comprises theNetherlands, Belgium, Luxemburg and thenorth and west of France. The southernmostFrench départements included in this researchare Charente-Maritime, Charente, Vienne,Indre-et-Loire, Loir-et-Cher, Loiret, Yonne,Haute-Marne, Haute-Saône and Haut-Rhin.The greater part of the area consists of rollingcountry dissected by rivers, the only mountainsbeing the Vosges. Two features which pliiy arather important rOlc in the following are thecombined delta of the rivers Rhine, Meuse andScheldt, and the tidal marshes in the north.Both are referred to as the "wetlands".

The research done before the sixties was notvery systematic in the sense that the botanistswho investigated macroremains did not formpart of the archaeological teams responsible forthe excavations. Summaries of the early worktan be found in van Zeist (1970), De Ceunynckand Verbruggen (1985) and Marinval (1988).Since then systematic sampling, supervised byspecialists, has become more and more therule, especially in the Netherlands. The largerDutch archaeological institutes have their ownbotanists. These institutes are the Biologisch-Archaeologisch Instituut at Groningen, theAlbert Egges van Giffen Instituut voor Prae- enProtohistorie at Amsterdam, the Instituut voorPrehistorie at Leiden and the Rijksdienst voorhet Oudheidkundig Bodemonderzoek at Amers-foort. Unfortunately, the situation is less ad-vanced in Belgium, Luxemburg and France.

The aim of the following will be to give asurvey of what is known on the subject of fruitsand seeds in the studied area. Unfortunately,for lack of room, other macroremains cannotbe dealt with here and even the remains ofwild plants can hardly be discussed in the waythey deserve. The material will be reviewed ina chronological order. Only fully publishedreports will be discussed. However, in a fewcases the results of an incompletely publishedor internal report have been included as theyfill an otherwise serious gap.

2 THE PALAEOLITHIC,EPIPALAEOLITHIC AND MESOLITH1C( -6400 BP)

This paragraph covers the longest period ofman-plant relationships, but presents the short-est list of data. No seed and fruit remainswhatsoever have yet been ascribed to thePalaeolithic proper. This is not due to inad-equate research. At the intensively sampledMiddle Palaeolithic site of Maastricht-Belvé-dère, for instance, even charcoal had crumbledto powder. Remains of seeds, if at all present,must have disintegrated beyond recognition(Roebroeks 1985: 111). Many of the Palaeo-lithic sites may simply not have favoured thepreservation of seeds over such a long stretchof time.

The gathering economy of the Epipalaeolithicand Mesolithic did leave some traces, but well-preserved finds are scarce. Charred fragments

279

Table 1. Sites with Epipalaeolithic and Mesolithic hazelnut shells.

site BP author(s)

Abri de BellefondCoincyMontbaniChambre des Fées, CoincyManiefelsen, OberlargWeelde-PaardsdrankNeerharen-De Kip

8190±1907280±350

90306990 ±1359170±HX)

Patte (1971)Parent (1973)Parent & Planchais (1972)Hinout (1964)Thevenin & Sainty (1974)Huyge & Vermeersch (1982)Lauwers & Vermeersch (1982)

of Corylus avellana are mentioned most often(Table 1). Other species have been recordedfor the Late Mesolithic (Atlantic Period).Newell (1973) mentions Trapa natans, Quercussp. and Prunus avium as remains of the DutchLeien-Wartena complex, but the full data havenever been published. The Prunus avium stonesfound in the often-cited Belgian site of Obglab-beek-Ruiterskuil (Vermeersch et al. 1974) may,on second thought, be subrecent. These uncar-bonized remains came from a podzol and noclear connection has been established betweenthe remains and the Mesolithic artefacts; theydo not look very old either (Vermeersch, pers.comm.; Bakels, unpubl.). The famous Frenchsite of Téviec (Morbihan) is reported to haveyielded carbonized pears of the species Pyruscordata (Boone & Renault-Miskovsky 1976),but these may require reexamination.

3 THE EARLY NEOLITHIC(6400-5900 BP)

Food production was introduced fromelsewhere. In the largest part of the studiedarea this way of life was made known bypeople of the Bandkeramik Culture. As thisculture originated in central Europe, to beprecise in the Carpathian Basin, agriculturemust have reached our region from the east viaGermany.

Bandkeramik settlements have been exca-vated in the southeastern part of the Nether-hinds, in the loess region of Belgium, and inthe northern part of France, including the ParisBasin. Synopses of the carbonized finds foundat these sites have been published by Bakels(1979, 1984), Heim (1983) and Bakels andRousselle (1985). The main crop was Triticumdicoccum, in most cases with an admixture ofTriticum monococcum. There is no evidence tosuggest that einkorn was cultivated as a crop inits own right. The Triticum aestivum reportedto have been found at Oudoumont (Belgium)

was incorrectly identified. Other crop plants arePisum sativum, Lens culinaris and Linum usita-tissimum. Hordeum vulgäre var. nudum hasonly been identified at two Belgian sites at thevery edge of the loess belt and in the FrenchAisne valley. Another uncommon crop plant isPapaver somniferum var. setigerum, which hasso far only been recovered from Dutch sites.Presumably domesticated in the West Mediter-ranean Basin, it is the only southwestern Euro-pean element in a collection of crop plantswhich came from central Europe and, ultimate-ly, from the Near East (Bakels 1982).

The crop plants are accompanied by a setgroup of weeds, at least in the Netherlands andBelgium (Table 2). Insufficient French siteshave been investigated to allow us to draw upa list of the most common species of France.The list of the Netherlands and Belgium closelyresembles that of the German Rhineland. Theweeds, first and foremost Lapsana communis,indicate the cultivation of small fields whichwere shaded part of the day, presumably be-cause they were surrounded by remnants of theprimeval forest. It is difficult to concludewhether the Bandkeramik farmers sowed theircereals in autumn or spring, because the weedspecies identified permit either conclusion

Table 2. Frequency of weeds in seven Dutchand Belgian Bandkeramik sites.

Bromus secalinus 6Chenopodium album 6Polygonum (Bilderdykia) convolvulusLapsana communis 5Phleum sp. 4Polygonum persicaria 3Bromus sterilis + tectorum 3Galium aparine 3Vicia hirsuta + tetrasperma 2Echninochloa crus-galli 2Rumex sp. 2Galium cruciata 2

280

(Bakels & Roussette 1985). Further data on«•'rop processing will be published by Bakels (inpress c). The carbonized cereal remains can besplit into two categories. The first consists ofnot yet dehusked kernels which could representunsuccessful grain parching. The second is theby-product of dehusking: the chaff. A thirdcategory of waste is unripe seeds of Chenopo-dium album, which are interpreted as the re-mains of the cleaning of vegetables. Charredfragments of wild nuts and fruits, such as Cory-lus avellana, Prunus spinosa and Malus sylves-tris, show that these were still being gathered,but the relative importance of this activitycannot be assessed.

The above-mentioned finds all come from pit-fil ls in open-air settlements, the only exceptionin this region being the cave of La Baume deOonvillars (Haute-Saône), a site which deservesspecial attention. Layer XI of the cave sedi-ments revealed three hearths, six vessels, threequerns, some tools, etc. and a cubic metre ofcarbonized cereals. One of the hearths yieldeda "C date of 6250±300 BP. All artifacts hadbeen deposited in an orderly and well-plannedmanner and the cave is therefore thought tohave been suddenly abandoned (Pétrequin1970). M. Villaret identified the grain andfound that just over half of the assemblageconsisted of remains of Triticum dicoccum, aquarter was Hordeum vulgäre var. nudum andless than a quarter Triticum aestivum. Thegrain was clean consumption grain, free ofweeds and almost free of chaff. Two of thechaff fragments recovered looked as thoughthey may have belonged to Triticum spelta, but't is not certain whether this wheat was indeedknown here (Villaret, in Pétrequin 1970). Un-fortunately, the report does not state whetherthe cereals had been stored as a mixture orwhether they were found separately. The greatbotanical importance of the find is the pres-ence of large quantities of naked barley andnaked wheat, which are strange to the Band-keramik world, but did occur in the south ofFrance. The pottery is not of the classicalBandkeramik style either. Gonvillars may repre-sent a point of contact between the centralEuropean (Bandkeramik) tradition and theMediterranean Neolithic.

Whereas the Bandkeramik culture has beenfairly extensively studied, very little research hasbeen done on the second Early Neolithic cul-ture in our region, that of western France. ThisNeolithic culture is known best for its cairns(funeral monuments). Before the cairn of Dis-signac near St. Nazaire (Loire-Atlantique) wasconstructed the site seems to have been occu-

pied by Neolithic settlers who left carbonizedremains identified by J. Heim as Triticum aesti-vum, Hordeum vulgäre var. nudum and Corylusavellana (Heim, pers. comm.). Charcoal fromthe ancient surface yielded dates ranging from6250±150 BP to 5780±150 BP but a date of494()±140 BP was also obtained (PHelgouach1977, 1981).

4 THE MIDDLE NEOLITHIC,LATE NEOLITHIC ANDCHALCOLITHIC (5900-3700 BP)

This time range covers the Rossen, Cerny,Chasséen, Funnel Beaker, Michelsberg, Seine-Oise-Marne and Beaker Cultures, to mention afew well-known names.

The direct heirs of the Bandkeramik Cultureare Rossen and Cerny, but it is now clear thatill least as far as the Rossen people are con-cerned, their agricultural tradition was no directcontinuation of Bandkeramik practices. Al-though the field weeds indicate that their fieldswere laid out according to the same pattern,they grew a greater variety of cereal crops.Triticum aestivum and Hordeum vulgäre var.nudum were also cultivated as main crops,alongside the emmer and einkorn wheat ofearlier times. The new cereals are thought torepresent a fresh influence from, ultimately,southern France (Bakels, in press a,b). In thislight it is a pity that so little is known aboutthe French Early Middle Neolithic proper. Theonly relevant seed assemblages come from theVilleneuve-St. Germain group. Two pit fillsfrom the type site have been published (Bakels1984) and they did not contain naked wheat ornaked barley. So far, no research has beendone on the other follower of the Bandkeramiktradition, the Cerny Culture.

The still very scarce data available on westernand northern France, Belgium and Luxemburgmake it almost impossible to draw a picture ofthe Neolithic and Chalcolithic in those regions.The available information is presented in Table3.

Contrary to what one would expect in acountry with a tradition of palaeoethnobotanicalresearch, the information obtained on the per-iod concerned in the drier parts of the Nether-lands is no better. The people of the MiddleNeolithic Michelsberg Culture, the FunnelBeaker Culture and a few more local groupsleft virtually no material remains from whichsuitable samples could be obtained. SomeFunnel Beaker sherds show imprints of seedswhich provide at least some information on

281

crop plants. Triticum dicoccum, Triticum mono-coccum, Hordeum vulgäre var. nudum andLinum usitatissimum have been identified (vanZeist 1970; Bakels 1976).

Little more is known about the Late Neo-lithic. The most detailed information comesfrom Eeserveld, where van Zeist (1970) foundseparately stored supplies of emmer, nakedbarley and acorns. The Bornwird samples camefrom an old plot of cultivated land whichshowed ard marks.

The lack of data from the upland sites is tosome extent offset by that from the Dutchwetland sites. In these wetter regions traces of

human occupation usually are buried beneaththick layers of sediments, but some lie closeenough to the surface to be excavated. Theseare situated either in the higher parts of thelandscape or in areas where erosion or humanaction have removed large parts of the overly-ing sediments. All kinds of plant remains havebeen preserved in the wet environment and asa rule the excavation of sites includes intensivesampling programmes.

The oldest traces of Neolithic occupationwere discovered at Bergschenhoek near Rotter-dam. Around 5400 BP people camped there inan environment of fresh-water marshes andlakes. They used a floating island consisting of

282

peat, which they repeatedly visited during thelate autumn, witness the bones of migratingbirds. They left a hearth, some pottery sherdsand parts of their hunting and fishing gear,such as fish traps, but no remains of cultivatedplants. Plant remains include Corylus avellanashell fragments, Prunus spinosa stones andcarbonized apples of Malus sylvestris (vanZeist, pers. comm.).

Another Middle Neolithic seasonal settlementhas been excavated at Swifterbant. It is datedto around 5300 BP and is thought to have beenoccupied in summer. People lived on the leveesof a system of fresh-water creeks. The land-scape was one of willow and reed marshes.Swifterbant did yield cereal grains, all of whichwere carbonized. The majority are grains ofHordeum vulgäre var. nudum, the rest aregrains of Triticum dicoccum and a single grainof what is thought to be Triticum aestivum(Casparie et al. 1977; van Zeist & Palfenier-Vegter 1983). Rachis and glume fragments ofnaked barley were also found at this site andthis is taken as an indication that the crop wasgrown locally, in spite of the unfavourableedaphic conditions. "For, if the plants had beengrown elsewhere, the threshed grains, which are

considerably less bulky than the ears, wouldhave been transported" (Casparie et al. 1977:51). The quality of the crop, as expressed bythe size of the grains, is not poor at all, butthis tells us nothing about grain yields or fre-quency of crop failures. According to van Zeistand Palfenier-Vegter, it is unlikely that thelocally grown crops constituted the main itemof the diet of the inhabitants of Swifterbant.Wild species were collected too, as is clearfrom the presence of seeds of Corylus avellana,Malus sylvestris, Crataegus monogyna, Rosa sp.and Rubus fruticosus. As for some of theplants found, such as Chenopodiaceae, Polygo-num species and Urtica dioica, it is not certainwhether they were eaten or whether they sim-ply represent the local vegetation of a dis-turbed habitat (van Zeist & Palfenier-Vegter1983).

A rather important series of layers containingNeolithic occupational waste, separated byclean sediments, was excavated on the slopes ofa now almost buried river dune near Mole-naarsgraaf: the Hazendonk. The dune con-stitutes a dry spot in an environment of lakes,reed marshes and alder carrs. The sequencebegins with the "Hazendonk 1" phase which is

Fig. 1. Location of the sites mentioned in the text. Dotted line: boundary of the area underreview. Shaded: ares with wetland sites. For the sake of clarity, Dutch settlements (nos 1-27)belonging to the Roman period and historical times are presented on an inset map.

1 Groningen2 Leeuwarden3 Gasseite4 Odoorn5 Wijster6 Pesse7 Noordbarge8 Kampen9 Dalfsen

H) Kootwijk11 Ede-Veldhuizen12 De Horden13 Haarlem14 Amsterdam15 Valkenburg16 Leiden7 Alphen-aan-den-Rijn

18 Woerden19 Naaldwijk20 Ouddorp21 Aardenburc220ss

3 's-Hertogenbosch24 Dommelen25 Sittard

26 Voerendaal27 Maastricht

28 Bornwird29Elp30 Eeserveld31 Angelsloo +

Emmerhout32 Noordbarge33 Gees34 Ommen35 Swifterbant36 Ermelo37 Colmschate38 Opperdoes39 Twisk40 Bovenkarspel41 Aartswoud42 Zandwerven43 Assendelver

Polders44 Zijderveld45 Hazendonk

46 Bergschenhoek47 Vlaardingen48 Hekelingen49 Oss50 Son & Breugel51 Dommelen52 Maastricht53 Broekom54 Neerharen55 Opglabbeek56 Weelde57 Evergem58 Ittre59 Givry60 Titelberg61 Düdlingen62 Peppingen63 Weiler zum Turm64 Suippes65 Chfllons-sur-Marne66 Menneville67 Cuiry-lès-

Chaudardes68 Chassé my69 Montbani

70 Coincy71 Villeneuve-

St. Martin72 St. Pierre-en-

Chastre73 Champlieu74 Béthisy-St. Martin75 Compiègne76 Catenoy77 Baillet-en-France78 Villiers-le-Sec79 Le Fort Harrouard80 Orléans81 Téviec82 Dissignac83 Abri de Bellefond84 Chazelles85 Cronenbourg86 Geispolsheim87 Rosheim88 Entzheim89 Etival-

Clairefontaine90 Gonvillars91 Manlefelsen

283

Table 3. Carbonized seeds from the Middle Neolithic, Late Neolithic and Chalcolithic periods.Gr. = Grossgartach, Röss. = Rossen, Epiröss. = Epirossen, VSG = Villeneuve-St. Germaingroup, Mich. = Michelsberg, Vla. = Vlaardingen, PFB = Protruding Foot Beaker, SOM = Seine-Oise-Marne, c = cf. Dates in years BP.

upland sites:Rosheim, Gr. . . + Erroux (1976)Maastricht, Ross. 5800 + + + + + . + . Bakels (in press a,b)Entzheim, Epiröss. 5400 + + . + . c Hopf (1975)Givry, Epiröss. 5300 . + + + . . . Heim (1979)Ittre, Epiröss.-Mich. + . + + + . . . Heim (1987)Villeneuve-St. Germain, VSG ++ + . . . Bakels (1984)Cuiry-lès-Chaudardes, Mich. . + . . c. . + . + . . . Bakels (1984)Geispolsheim, Mich. + + + + Hopt'(l987)Cuiry-lès-Chaudardes, SOM + . . . Bakels (1984)Eeserveld, PFB 3955 ±50 . + . + . . . . + . . . . van Zeist (1970)Born wird, PFB before 3930 ±50 . . . + + . . . van Zeist (1970)

wetland sites:Swifterbant 5300 . + c + + + . + van Zeist &

Palfenier-Vegter (1983)Hazendonk 5300 . + . + + + . + Bakels (1981)Hazendonk, Mich. 5000-4800 . + . + . . . . + + + + + Bakels (1981)Hazendonk, Vla. 4600-4400 . + . + . . . . + + . + + Bakels (1981)Hekelingen, Vla.4 300-4000 . + . + . . + . + + + +. Bakels (1988)Vlaardingen, Vla. 4300 . + + . + van Zeist (1970)Zandwerven, Vla. . + . + + . . . van Zeist (1970)Aartswoud, PFB 4000-3900 + + + +. . + . + + +. . Pais (1984)Hazendonk, BB 3600 + + . . . Bakels (1981)

contemporary with Swifterbant, i.e. datable to5300 BP. The next two phases are dated to5000 and 4800 BP, respectively, and these areascribable to a local version of the MichelsbergCulture. These were followed by three "Vlaar-dingen Culture" (a Dutch Late Neolithic Cul-ture) phases, dated to 4600, 4400 and 4100 BPand finally a Bell Beaker/Barbed Wire Beakeroccupation phase, dated to 3600 BP. A prelim-inary report on the seeds shows that Hordeumvulgäre var. nudum was known there through-out all of these phases. The other cereal found,Triticum dicoccum, occurred here up to andincluding the Vlaardingen phase 1 only. The

material comprises carbonized kernels andchaff. In spite of the intensive research carriedout, no non-carbonized cereal remains werefound. The cereals were accompanied by weedseeds, the most important of which was Broniussecalinus (chess). In one of the Hazendonk 1assemblages the emmer/barley/chess grain ra t iois 100/400/150, which suggests that the chesswas eaten along with the cultivars (Bakels1981). As at Swifterbant, the question arises asto whether the crops were in fact cultivated onthe small river dune itself. During the occupa-tion of Hazendonk 1, the dry surface area ofthe dune amounted to approximately 1.2 ha; by

284

the lime of the last Vlaardingen occupationphase it had already shrunk to 0.4 ha, due tocontinuous sedimentation and peat growtharound its base and on its slopes. There wasvery little space for fields, because it must notbe forgotten that there must have been roomfor huts and for activities such as fish cleaning.In spite of the abundance of chaff, it seemsmost unlikely that crops were grown locallyand, contrary to the views held by the re-searchers working at Swifterbant, the crops arethought to have been brought to the site in theear. The Hazendonk excavation also yieldedwild plant foods such as Malus sylvestris,Prunus spinosa and Crataegus monogyna, all ofwhich were carbonized. They could have beenpicked locally, but again, there was only littledry surface area and local harvests cannot havebeen very plent i ful . It is, however, impossible toq u a n t i f y the amount of plant food gathered.

The same question, of whether crops wereraised locally or not, arose in connection with athird wetland site: Hekelingen near Rotterdam.Here the narrow levee of a fresh-water tidalcreek revealed traces of intermittent Vlaardin-gen Culture occupation between 4300 and 4000BP. The waste included both kernels and chaffof Triticum dicoccum and Hordeum vulgärevar. nudum, in addition to remains of Linumusitatissimum. Some of the species collectedare Corylus avellana, Prunus spinosa, Malussylvestris, Quercus sp., Trapa natans andRanunculus ficarin (tubers). As the levee wasonly about 20 m wide and was bounded by anactive tidal creek on one side and by a back-swamp with active alder-carr peat formation onthe other, arable land was scarce. Even if thelevee was occupied by only a small group ofpeople, a considerable stretch of land musthave been cleared for cultivation but there isno evidence of this in the pollen diagrams.Therefore, the grain was most probably im-ported and perhaps the acorns too, as it isunl ikely that oak trees grew nearby (Bakels1988).

It the grain found at both Hazendonk andHekelingen was imported, where was it pro-duced? The coastal dunes and the broaderlevees have been suggested as possible sources.Unfortunately, only few samples have beentaken in these areas. One sample, from Vlaar-dingen, shows, surprisingly, not a main crop ofnaked barley and emmer, but of Triticum aes-tivum, with an admixture of Triticum dicoccum.The sample also contained four grains of hulledHordeum vulgäre (van Zeist 1970). As yet, thisis an unique find. More in accordance with ourexpectations are the Vlaardingen Culture sam-

ples from Zandwerven, which contained theremains of Triticum dicoccum and Hordeumvulgäre var. nudum (van Zeist 1970). At Zand-werven ard marks indicated the existence oflocal fields. But this, of course, does not provethat such sites produced the grain for the siteson the smaller parcels of dry land, whetherthese were seasonally occupied or not.

The discussion of Dutch Neolithic wetlandsites will be concluded with the mentioning of aLate Neolithic Protruding Foot Beaker/BellBeaker settlement near Aartswoud. This site isalso situated on the levee of a tidal creek, inthis case a brackish one. The landscape wasopen and, as opposed to the sites described sotar, the levees were not yet covered with trees.The surroundings were vast salt marshes insome stage of desalination. Linum usitatissi-mum, Triticum dicoccum and Hordeum vulgärevar. nudum were the staple crops. Triticumaestivum was a scarce admixture (Pals 1984).The few hulled barley grains published for thissi tf are now interpreted as carbonized milk-ripe naked barley grains. The presence of ro-bust carbonized Gramineae stems was originallythought to indicate local agriculture, but thepossibility that they are the tops of Phragmitesrather than cereals cannot be ruled out (Pals,pers. comm.). However, the carbonized weedremains are dominated by Althaea officinalis,Atriplex sp., Scirpus sp. and Chenopodiumalbum, and the first three support a local prov-enance of the crops because they grow inbrackish environments and presumably camefrom the lower-lying margins of the quite re-cently desalinated fields. Wild plants were stillgathered to a varying extent. Remains of Rubussp., Malus sylvestris, Quercus sp. and Corylusavellana have frequently been found. It is quiteunlikely that the apples, acorns and hazelnutswere gathered near the settlement. As the siteis thought to have been occupied the yearround, there must have been organized "fruitgathering expeditions in autumn to the highersandy soils which were at least 15 km away"(Pals 1984).

To conclude, the research carried out in thewetlands has provided us with a question whichis never raised by the contemporaneous uplandsites, namely are the sites at which edible plantremains and especially crop remains are foundalways producer sites? The answer is plainlyno. However, wetland sites may form a specialcase. A non-agrarian economy, such as fishingand fowling, might have been more rewardingin these areas than a food-producing system.

285

5 THE BRONZE AGE (37(X)-2600 BP)

Unfortunately, most of our knowledge of theBronze Age concerns burials and hoards andvery little is known about Bronze Age settle-ments. However, recent excavations are begin-ning to alter this unbalanced situation and thisis of benefit to palaeoethnobotany.

To begin with western and northern France,the richest collection of seeds published so faris that from Le Fort Harrouard near Dreux(Eure-et-Loir). This hillfort is situated on apromontory between a steep-banked dry valleyand the river Eure and must have controlled alarge area. It certainly occupied a central posi-tion. Part of the site was excavated at thebeginning of this century by Abbé J. Philippe,who not only recorded the architectural re-mains and rich mobilia, but also took a largenumber of seed samples (Bakels 1982-'83,1984). Predominant in the Early and MiddleBronze Age assemblages is Panicum miliaceum,found as lumps of fused grains still covered bythe palea and lemma. These lumps contain noremains of other species, not even of weeds.Other species encountered in the samples areTriticum dicoccum, Hordeum vulgäre, Viciafaba var. minor and Quercus sp. The richestfinds are from the Late Bronze Age. The ubiq-uitous presence of lumps of carbonized grainsuggests that the hillfort, or some quarters of it,must have burned down at least twice, onceduring the Late Bronze Age I and again duringthe Late Bronze Age III. The charred foodsupplies comprise Triticum dicoccum, Hordeumvulgäre, Panicum miliaceum, Vicia faba var.minor, Pisum sativum and Quercus sp. Lesscommon are Triticum aestivum, Malus sylves-tris, Rosa sp., Prunus spinosa and Corylusavellana. The stores were very clean, only afew remains of Avena sp. and Bromus sp. werefound. Some years ago the investigation wasstarted again and the first botanical samplestaken were identical to those described above.

Recent research at another hillfort, Catenoy(Oise), revealed the same species as those ofLe Fort Harrouard: Triticum dicoccum, Hor-deum vulgäre, Panicum miliaceum, Vicia fabavar. minor and Pisum sativum (Bakels, un-publ.).

A 19th century excavation of a third hillfort,St. Pierre-en-Chastre near Vieux-Moulin (Oise),yielded a small decorated Late Bronze Agevessel full of beautifully preserved Hordeumvulgäre with an admixture of some Triticumdicoccum, Panicum miliaceum and a few weeds.The grain must have been carbonized in thepot, because the apical hairs and rachillae were

undamaged. The function of the vessel is stillnot clear. It is far too small to have been usedfor storage purposes, even for the grain of onemeal or for the sowing of one field. It mayhave been an offering (Bakels 1984). The re-mains found at other French Bronze Age sitesare presented in Table 4.

No Bronze Age finds from Belgium are asyet available, but two Late Bronze Age (Ur-nenfelderkultur) sites have been discovered inLuxemburg: Düdlingen-Budersberg and Peppin-gen-Keitzenberg, which yielded rich assemblagesof carbonized seeds. The crop plants identifiedin these assemblages are Triticum dicoccum,Triticum spelta, Hordeum vulgäre, Panicummiliaceum and Lens culinaris. Prunus spinosa,Quercus sp., Sambucus ebulus and Rosa sp.must have been collected as wild plants andfruits. The assemblages are also rich in wildherb species (Bakels, unpuhl.).

The rather scarce Dutch Bronze Age findsrecovered from the higher grounds are pre-sented in Table 4. The seeds from one site,Oss-IJsselstraat, were found in a well which hasbeen dated to 320()±30 BP. This MiddleBronze Age well provided remains of a ratherunexpected species, which is not mentioned inthe table, namely an uncarbonized stone ofPrunus insititia (Bakels 198()a).

Bronze Age wetland occupation is discussed byBuurman (1979 and especially 1988a), who in-vestigated material from the province of NorthHolland. In this region the occupants settled ongully ridges, the remnants of a previously tidalflat landscape. The silted-up tidal gullies be-came sandy ridges during the inversion of reliefthrough differential shrinkage. At the timewhen they were occupied they were the highestand driest parts of the landscape. Here findshave been recovered datable to the MiddleBronze Age, from 3200 to 2850 BP, and theLate Bronze Age, from 28(X) to 2650 BP.

The arable fields were situated on top of theridges where the light, well-drained soil was theeasiest to till. The lower-lying clay basins wereused as pasture-land. There may have beenfresh-water lakes in the lowest parts of thesebasins (Buurman 1988a). The fields themselveswere recognizable by the ditches marking theirboundaries. Other, circular, ditches and circlescomposed of pits, both with diameters of about4 metres, are interpreted as the remains ofdrainage ditches around corn stacks. This inter-pretation is based on the discovery of carbon-ized cereal stems, heavy and fine chaff andkernels in two of them (at Bovenkarspel).Seeds of low- and tall-growing weeds were also

286

Table 4. Carbonized seeds from the Bronze Age.EB = Early Bronze Age, MB = Middle Bronze Age, LB = Late Bronze Age. Dates in years BP.

upland sites:Fort Harrouard, EBComniegne, EBFort Harrouard, MBChazelles, MBSon, MB

Oss-IJsselstraat, MBZijderveld, MB/LBEmmerhout, MB/LBElp, MB/LBFort Harrouard, LBCatenoy, LBVieux-Moulin, LBChazelles, LBComniegne, LBDudlingen, LBCronenoonnPeppingen, LB

wetland sites:Twisk, MBBovenkarspel, MBOpperdoes, MBBovenkarspel, LBOpperdoes, LB

i§u

I•D3c 5S E-S5 U 3 C

Ja u.2 MW».«

E3

1 _ OT H •r* CQi « g yj _ «2 ca t- (/ï

^ U U , ? > ' c > > S 2 ei."? S C^ 3 n * t ' i r - _ f c c < 3 - . s c - £ t o > >'£.IlillilsPSSfJ3 3 3 U U a1" c u C u 3 w SS.u. u .0-0 -o. u n E „, g E-Sg =

l**alalJ8|Ira|H(-H-C3-a.>eujj OU 2 o-

+ . . . + + Bakels (1984)+ Bakels (1984)

+ . . . . + + . . . + . . . Bakels (1984). . . + + + + . Marinval (1983)+ . . . + + . . . . + . . . Bakels &

van der Ham (1980)3200±30 . . , . . + Bakels fl980aj3370±80 + ' . . . + van Zeist (19703320+60 + . + + . . van Zeist (19703200-2750 + . . + + + + . . van Zeist (1970

+ + . . + + + + . . + + ++ Bakels (1984)+ . . . + + + + Bakels (unpubl.)+ . . . + + Bakels (1984)

+ . . . Marinval (1983)+ . . . _ + Bakels (1984)+ . + . + + + . . . . + . . Bakels (unpubl.)

+ . Halt & Zumstem (I960). . + . + + . . + . . . . + Bakels (unpubl.)

3350±35 + . . + Buurman 1988a+ . . + + Buurman 1988a+ . . + + Buurman 1988a. . . . + . . . . + . . . . Buurman 1988a

. + . . + . Buurman 1988a

found in these circular ditches and it is thoughtthat whole sheaves were destroyed by an acci-dental fire. Further processing of the crop tookplace near or in the houses, as is apparentfrom the distribution of fine chaff and weedseeds.

The first crop plants to have been sown areTriticum dicoccum, Hordeum vulgäre var. nu-dum and Hordeum vulgäre. Barley and emmerwere also common. Naked barley, however,occurred only in the colonization phase. It wasrather soon replaced by hulled barley, which isthe only cereal to have been cultivated in theLate Bronze Age, because emmer vanished too.This change is ascribed to the increasing wet-ness of the area; only the highest gully ridges•emained inhabitable. Nevertheless, at least oneother crop plant became common in the Late

Bronze Age: Linum usitatissimum, which showsthat crop cultivation did not end with a mono-culture of hulled barley.

The most striking aspect of the rather hetero-geneous collection of Bronze Age plant remainsdiscussed in this paragraph is the appearanceof new crop plants. Hulled barley had eitheralready taken the place of naked barley by thebeginning of the Bronze Age or gradually didso in the course of this period (in the wetlandsof North Holland). Millet was (suddenly?)present everywhere and the stored supplies ofthe inhabitants of the hillforts of northernFrance included horsebean. Of the variouswheat species, emmer was most common,whereas naked wheat has only been identifiedat Le Ford Harrouard. Einkorn has not been

287

found at all in Bronze Age contexts, but thismay he due to the small number of sites sam-pled. Opium poppy seed has not yet beenrecovered either. The presence of remains ofspelt wheat and lentil in the assemblages of theUrnfield Culture of Luxemburg deserves someattention.

6 THE IRON AGE (26(K)-2(KX) BP)

Most records of Iron Age fruits and seedsconcern the Lute Iron Age. Early Iron Agefinds are relatively scarce. As usual, most mate-rial comes from settlement sites, but there isone exception: a La Tene I burial near Châ-lons-sur-Marne (Marne) is reported to haveyielded a tew carbonized grains of Triticumdicoccum and Hordeum sp. (Hopf 1969). It isnot clear, however, whether these are con-nected with the burial ceremony or ended upin the grave by mere chance.

Four of the settlement sites mentioned in theliterature are oppida: defended central places.These are Bethisy-St. Martin (Oise), Villeneu-ve-St. Germain (Aisne), Etival-Clairefontaine(Ardennes) and Titelberg (Luxemburg). Verylitt le information has been published on them;the available data are presented in Table 5,with the exception of those relating to Etival .Part of a floor of a I.« Tene III house at Eti-val was covered wi th two separate heaps ofcharred grain. These consisted of "wheat, barleyand rye varieties still cultivated today" andTriticum turgidum (Billorel 1968). This largequant i ty of grain may need re-examining.

As for the remaining non-wetland settlementsites mentioned in the literature, all of whichare rural sites, the finds listed were recoveredfrom silos in at least seven cases (Table 5).These usually cylindrical, conical or beehive-shaped pits with flat bottoms were used fromthe times of the Neolithic Bandkeramik Cultureonwards, but they were most spectacular in theIron Age, when large aggregations of pits weredug, for example at Suippes (Marne) and Son& Breugel (the Netherlands). One of the mostpressing problems concerning the charredgrains found in these silos is whether theyformed part of the original contents of the pitor whether they represent dumped waste. Inmost cases, the lack of concentration of car-bonized seeds near the bottom of the pit oradhering to the pit wall suggests that we aredealing with waste material. This is furtherconfirmed by the fact that the silos were clearlyused as dumping places for other kinds ofdomestic waste too. The general composition of

the seed assemblages by no means suggeststhat we are dealing with the remains of astored product. At Suippes the cultivatedplants: wild herbs ratio ranges between 1:1.5and 1:42 (Bakels 1984). The silos of Son con-tained a mixture of all kinds of cultivated andwild plants from different habitats (Bakels &van der Ham 1980).

One pit discovered at Menncville (Aisne),two pits found at Neerharen-Rekem (Belgium)and a number of pits sampled at Colmschate(the Netherlands) could possibly be identifiedas silos containing stores carbonized in situ.One of the Compiègne (Oise) pits and a pitdiscovered at Evergem-Ralingen (Belgium) maypossibly also be added to this list. The floors ofthese pits were covered with a thick layer ofcarbonized seeds. Surprisingly, none of theselayers consisted of the remains of one singlespecies (Table 5) but with the exception of thecontents of a pit excavated at Colmschate, theassemblages were poor in weed seeds.

Some authors are of the opinion that layersof pure grain are indeed to be interpreted asstores in s i tu and hence indicative of the pri-mary function of the pit. They ascribe thecarbonized state of the grain to the burning outof the silos to sterilize them for continued use.But it is diff icult to obtain conclusive evidenceof this. According to the experiments carriedout by Reynolds (1974) such assemblagesshould at least include some germinated grain,but no germinated kernels are reported to havebeen found at the sites in question. The onlyrecord of germinated grain found in a storagepit concerns the site at Weiler zum Turm (Lux-emburg) and this was clearly not burned on thespot (Hopf 1980). Another indication are red-fired pit-floors and walls. These were observedin one of the Neerharen pits, in one of theColmschate structures and at Weiler zumTurm. They form a very small minority in thev.-ist number of silos excavated in this region.Al though these red walls indicate that some-th ing was burned in the pits, they are not nec-essarily proof of cleaning by fire. In spite ofthe large number of silo contents analysed, westi l l do not know whether the layers of carbon-ized grain indicate the cleaning of the silo byfire and whether they are in fact to be seen usan indication of the primary function of the pit.

In the wetlands the first terpen appeared. Ter-pen are a r t i f i c ia l mounds bui l t by man to serveas dry habitation places. They consisted oflayers of clay, plant matter and dung, and wereenlarged whenever necessary. Older buildingswere incorporated in the mounds and the re-

288

Table 5. Carbonized seeds from the Iron Age, upland sites. EIA = Early Iron Age,MIA = Middle Iron Age, LIA = Late Iron Age, c = cf. Dates in years BP.

ED

TT

.>ësi

Ips^U^ell |-ËHlS.^iS+'-ÉsSd.««

IHinl*w1Jiï*p p p p o u 3 u ( g e P ° « ï G _ 5 < / )U U O O - O T 3 . y - = C M C g " 1-'?*3•s »s'c s i- t- E 8 "'3 3 ? ° u u^=w l i doal

T?teîberg, Tène II 2256±55 + + + . . + . . + Rowlett et al. M 9821TitelbergLTène III . . + . . + Rowlett et al . (1982)Bélhisy, Tène III . • + J°UVC O9.™)Villeneuve-St. Germain, Tène III + Bakels (1984)

Weiler zum Turm, Hal ls ta t t D + + + . . + . . . . + Hopf (1980)Colmschate, E IA . + . + + + + . + . C + + + . . Buurman (Menneville, Tène I . + . + . .Suippes Tène I + + . + . + + . + . . . + . + . Bakels (1984)Neerharen, Tène I 2530±50

2435±35 . + ... + .. + . + • + • • • Roymans (198SajEvercem Tène I 249Ô±60 ! + . . . + + . + + . . . + . . Ceunynck et al. (1984)Son,1v1IA 2460-2250 . + . + . + + . . . + + + + + . Bakels &

van der Hain (1980)

Anjdsloo. EIA 2570+55 . + . . + + + . + + van Zeist (1970)Ommen ÈIA • + • • van Zelst (WErmelo, EIA/MIA 2140+60 . + . - . + + • + • . • + • yan Zeist JCompiègne, Tène I . + . + . + + . + Bakels (19S4)Chassémy, Tène la . + + . + c . . ++ . Hopf (1969)Dommelen, MIA . + . . . + + . . . . + . . Roymans (l_985b)Oss-Ussen, MIA + • • Bake's C'Gees, MIA 2420±35 . + . . . + . . + van Zeis (1970)Oss-Ùssen, l .IA • • + Bake's .( M,71Noordbarge, LIA + + + vfn Zeif (1976)Broekom, Tène III . + Vanvmckenroye (1988)

suiting structures could be seen as wetland tells.Terpen occur in the coastal areas of the Neth-erlands and especially in the northern prov-inces. Terp bodies are excellent preservers ofplant material and it were terp bodies thatformed the object ol the first systematicpalaeoethnobotanical research to be carried outin the Netherlands (Beijerinck 1929-1931). Mostol the results obtained in these investigationswere recapitulated by van Zeist (1974).

The crop plants of these sites are Hordeumvulgäre, Linum usitatissimum and Camelinasaliva, and also some Trilicum dicoccum andVicia faba var. minor. Mosl of the lerpen lie insalt marsh areas. The presence of the remainsof crop plants does not necessarily imply tha tthese plants were grown in the immediate sur-

roundings, but the presence of threshing re-mains of barley, gold of pleasure and flax showthat at least these products were not imported.Experiments conducted by van Zeist et al.(1976) and Bottema et al. (1980) have shownthat it is in fact possible to raise crops in iheunprotected salt marshes, provided that thearea is not flooded by sea water too often,especially during the seedling stage. Avenasaliva, Hordeum vulgäre, Linum usitatissimum,Camelina saliva and Vicia faba var. minor gavereasonably reliable yields. Triticum species andPanicum miliaceum were less suitable or evenunfit for cultivation in these areas. The risk offlooding meant that the sowing had to be donein the spring. The field weeds found in thesamples taken from the terpen indeed point to

289

summer crops. Nevertheless, some productswere certainly imported and this does not onlyapply to the Triticum dicoccum found; remainsof Corylus avellana and Prunus spinosa alsohave been found, and these shrubs cannot havethrived in the neighbourhood.

The salt marshes were not the only wet envi-ronments where farming communities lived.The fresh-water tidal areas and peats were alsoinhabited. Intensive research is at present beingcarried out on Iron Age farming in these re-gions. Some reports have already been pub-lished on one of these projects, namely theAssendelver Polders Project (Therkorn et al.1984; Pals, in Brandt et al. 1987).

The Assendelver Polders lie at the edge of avast peat area east of the Older Dunes in theprovince of North Holland. One of the excava-tions carried out as part of the project yieldeda beautifully preserved Early Iron Age farm-stead (site "Q"). Building materials were MCdated at 2465±30 BP and 2520±30 BP. Theinhabitants' subsistence economy was based onstock raising, but the presence of threshingremains of Camelina saliva suggests that thiscrop at least was indeed grown on the localpeat. This w;is possible because in this area thepeat was drained by gullies; in fact it was thisdrainage that had made the land suitable forhabitation in the first place. The samples alsoindicated the presence of cereals, predominant-ly Hordeurn vulgäre and some Triticum dicoc-cum, Panicum miliaceum and Avena saliva, butthese are not considered to have been localproducts. The same applies to Linum usitatis-simum. It is not possible to determine whetherthe inhabitants of this farmstead had their owncereal fields at some distance from the site, ina non-peaty area, or whether the cereals weretraded.

The most remarkable aspect of the dataobtained so far for the Iron Age is the ap-pearance of Camelina saliva, at both uplandand wetland sites, and the common occurrenceof Triticum spelta at upland sites. Vicia (abavar. minor, which has only been encountered atFrench hillforts in Bronze Age contexts, isfound at rural settlements in Iron Age contexts.However, its former absence at such settle-ments it presumably due to the scanty set ofavailable Bronze Age data. One Iron Age siteyielded remains of Secale céréale. But as onlya few grains were found, it is not quite certainwhether rye was already being cultivated as acrop plant at this time, though the possibilityremains (van Zeist 1976).

Because from the Iron Age onwards cropswere contaminated by many weeds, not only of

the taller species but of low-growing ones aswell, it is often assumed that some aspect ofthe farming practices, for example harvestingmethods, must have changed fundamentally(Knorzer 1971; but see also Willerding 1988).In Western continental Europe, insufficientdata have so far been obtained to help thediscussions on this subject any further. Weedseeds are certainly much more abundant inIron Age samples, in absolute numbers as wellas in number of species, than in Neolithic ones,but at present the Bronze Age and the IronAge cannot be adequately compared becausethe sources of data are too diverse. However, ifwe are to pinpoint this change in farming prac-tices it is better to date it around the transitionfrom the Middle to the Late Bronze Age thanaround the transition from the Late BronzeAge to the Early Iron Age. Samples taken atthe Late Bronze Age sites of Diidlingen andPeppingen already contained numerous weedseeds.

7 THE ROMAN PERIOD (2000-1500 BP)

This period derives its name from the fact thatin these times Rome controlled the greaterpart of the studied area. What was at that timethe largest branch of the Rhine marked thenorthern border of the Roman Empire: thelimes. The branch passed Arnhem, Utrecht andLeiden. The area south of the limes cameunder the influence of the Roman marketsystem. North of it, life went on more or lessas before, although some Roman influence isobservable in these regions too. Within theEmpire different classes of settlements startedto arise. There was greater differentiation thanthere had been in the preceding Late Iron Age.Besides rural settlements, there were also mili-tary camps and proper towns. The inhabitantsof the latter two were not food-producersthemselves and this led to the emergence offood-producing centres: commercial farmswhich are called villae rusticni'.

In contrast to the large amount of informa-tion on this period yielded by other sources,the data obtained from palaeoethnobotamcalresearch are still scanty. The only villa rusticaon which some details have been published isthe villa at Voerendaal (the Netherlands). Herea granary (a horreum), and a threshing floorindeed produced the expected remains, i.e.cereal grains and chaff, respectively. The villaspecialized in the production of Triticum spelta,but Triticum dicoccum, Triticum aestivum andHordeum sp. were also harvested (Willems &

290

Kooistra 1987).The cargo of a second century Roman ship

which sank in the Rhine at Woerden (theNetherlands) was at first assumed to have con-sisted also of spelt wheat (Haalebos 1986), butthe final investigation showed that it was em-mer (Pals & Hakbijl, in press).

Very little has been published on the con-sumer side of the agrarian economy too, butseveral large-scale investigations are now beingcarried out. As far as towns and other civilsettlements are concerned, only Champlieu(Oise), Maastricht (the Netherlands), Aarden-burg (the Netherlands) and Ouddorp (theNetherlands) can be mentioned. The Gallo-Roman grain from Champlieu is a very cleanmixture of Triticum aestivum, Hordeum sp.,Secale céréale, Avena sp., Pisum sativum andLens culinaris. What this assemblage was in-tended for is not clear (Bakels 1984). A largecharred grain deposit at Maastricht consistedmainly of Triticum aestivum with an admixtureof Triticum spelta, Secale céréale and onegrain of Hordeum sp. With the exception ofsome Bromus mollis/secalinus and Agrostemmagithago seeds, this grain was free from weedstoo (van Zeist 1970). The waste product ofsuch grain was discarded on the banks of theriver Meuse in Maastricht. The dump, dated tothe end of the second/beginning of the thirdcentury AD, contained a large amount of un-carbonized Triticum spelta chaff, Linum usita-tissimum capsules, and smaller amounts ofPanicum miliaceum, Papaver somniferum, Ma-lus sp., Corylus avellana, Prunus sp., Ficuscarica and the kitchen herbs Coriandrum sati-vum, Anethum graveolens, Satureja hortensisand Apium graveolens. Remains of field weedswere also encountered (Kuijper 1984a). AtAardenburg a small carbonized weed-free sam-ple of Triticum aestivum was found. It datesfrom between AD 170 and 270 (van Zeist1970). The same wheat species dominated afind from Ouddorp, whilst a second assemblagefound at Ouddorp, with a HC date of 2030±35BP, was dominated by Hordeum vulgäre. Bothwere only slightly contaminated by seeds ofother species, among which one Vicia faba var.minor seed (van Zeist 1970).

Information on the military section of societyis provided by the samples from the castella atValkenburg and Alphcn-aan-den-Rijn, both inthe Netherlands. Mil i ta ry granaries in the civil-ian settlement (vicus) next to the Valkenburgcastellum, which was burned down in the firstcentury AD, possibly during the revolt of thelocal inhabitants in AD 69, contained separatelystored cleaned Iriticum aestivum and Hordeum

vulgäre. The grain was in good condition andshowed no signs of infestation by insects (Palset al. 1989).

The remains of a grain supply within thecastellum itself, which was also burned downduring the same revolt, consisted, on the con-trary, of undehusked Triticum dicoccum andTriticum spelta which had begun to germinate -an indication that the storage room was verydamp. Two more deposits were found in thisdestroyed castellum, one consisting mainly ofTriticum aestivum and the other of Hordeumvulgäre, stored in the chaff (van Zeist 1970).All these Valkenburg samples contained onlyfew weed seeds.

Virtually pure chaff was found too. The floorof a well close to the two granaries is the vicus,but dated to the second century, was coveredwith a 20-cm-thick layer of Triticum speltachaff. This chaff was mixed with large weedseeds only, which showed that it was the re-mains of the dehusking of a stock similar tothose described above.

There was, of course, more to eat than thecereals mentioned. Charred plant matter andwaterlogged material revealed the consumptionof Secale céréale, Avena sativa, Vicia faba var.minor, Vicia sativa spp. obovata, Linum usita-tissimum, Papaver somniferum, Beta vulgaris,Corylus avellana, Juglans regia, Coriandrumsativum, Ruta graveolens and Satureja vulgaris,to mention only a few species.

The final stage of the consumption processhas been found too. In both the Valkenburgand Alphen-aan-den-Rijn castella an officers'latrine was excavated. The one found at Val-kenburg yielded remains of Papaver somni-ferum. Prunus cf. avium and Fragaria vesca(van Ledden-Hulsebosch 1955). The recentlyanalysed contents of the Alphen structure in-cluded fragments of cereal bran, among whichbran of Triticum spelta, some chaff remains ofTriticum spelta and Triticum dicoccum, seedsof Vicia faba var. minor, Papaver somniferum,Coriandrum sativum, Anethum graveolens,Apium graveolens, Olea europaea, Ficus carica,Vitis vinifera, Prunus persica, Malus sp. andCrataegus laevigata, and pollen of Pimpinellaanisum, Foeniculum vulgäre, Carum carvi andAnthriscus cerefolium (Kuijper & Turner, inpress).

Besides the meals of the living the meals forthe dead have also been studied. At both Val-kenburg and Naaldwijk remains of funeralpyres revealed charred food remains. Most ofthe remains found at Valkenburg were of Lensculinaris, whilst remains of Triticum sp., Hor-deum sp., Vicia faba var. minor and Olea euro-

291

paea were also encountered (Pals et al. 1989).A grave at Naaldwijk contained five seeds ofVicia faha var. minor (Bult et al. 1988).

The finds mentioned so far represent therather clean crop products belonging to themarket economy. They were transported andstored in large amounts. The sources of pro-duction have been traced with the aid of largeweed seeds such as the seeds of Orlaya grandi-flora and Caucalis platycarpos. Seeds of theseplants were found mixed with the spelt chaffrecovered from Valkenburg, formed part of theWoerden cargo and were also found in thelatr ine at Alphen, but the plants themselvescannot have thrived in the surroundings of theplaces were they were found. They belong tothe more southern weed flora of Belgium andFrance. The villa at Voerendaal lies just withinthe area of this flora district and Orlaya wasindeed found there. The weeds show that atleast part of the grain supply for the Romanarmy stationed along the frontier was importedfrom regions further south. More informationon this can be found in Pals et al. (1989) andGroenman-van Waateringe (1989).

How the native rural society responded to theproducer-consumer system is not yet clear, atleast not from botanical evidence. Again, thepublished information concerns Dutch sitesonly, but research on this subject is now beingcarried out in the French Aisne valley.

The most complete study of plant remainsfound at a site within the Empire is that car-ried out at a native settlement close to thelimes: De Horden near Wijk bij Duurstede,which was inhabited from 50 BC to AD 225.Some 300 soil samples were processed andtheir contents were handled with the aid ofCorrespondence Analysis. The main trend alongthe time axis is an increase in grassland speciesand a decline in the numbers of cereal re-mains. The increase in the grassland compo-nent is interpreted as an indication of a switchin the economy from subsistence farming in thepre-Roman period to the commercial produc-tion of meat and hides in the Roman period(Lange 1988). Hordeum vulgäre and Triticumdicoccum, and to a lesser extent Panicum milia-ceum, Pisum sativum, Vicia faba var. minorand Linum usitatissimum, were invariably thelocal crops, but some influence of the Romanworld is reflected in the very small quantities ofTriticum aestivum, Triticum spelta, Lens culina-ris, Anethum graveolens and Apium graveolens.

An even longer continuous range, from theEarly Bronze Age to the first half of the thirdcentury AD, has been analysed at Oss-Ussen,

but this vast set of data has not been fullypublished yet. The large excavation revealedthree contemporaneous settlements dating fromthe Roman period. Only one of them showssome signs of Romanization and it is at thissame settlement that Coriandrum sativum andBeta vulgaris were found. The preliminaryreports do not mention the distribution of theother food plants mentioned: Triticum spelta,Triticum dicoccum, Hordeum vulgäre, Secalecéréale, Panicum miliaceum, Vicia faba var.minor, Linum usitatissimum, Papaver somni-ferum and Juglans regia (van der Sanden1987a, b). It is therefore not yet clear which ofthese crops were cultivated locally. A contem-poraneous settlement nearby, Oss-IJsselstraat,yielded the remains of Triticum dicoccum,Hordeum vulgäre and Panicum miliaceum only,which could imply that the spelt of Oss-Ussenis attributable to Roman influence, as it is atDe Horden (Bakels 198()a).

Similar series of data have been published forthe area north of the limes. One, relating tothe Assendelver Polders, shows Hordeum vul-gäre as the principal crop and Panicum milia-ceum, Camelina saliva and Linum usitatissimumas additional ones. Remains of Triticum dicoc-cum and Avena saliva are scarce and they areinterpreted as admixtures (Pals, in Brandt et al.1987). The occurrence of Triticum pollen, how-ever, suggesls lhat emmer production wasknown, and Groenman-van Waaleringe (1989)thinks that ihe absence of emmer from Ihelocal record is due lo Ihe facl lhal il wasIraded to the Roman garrisons.

Olher dala were oblained from the terpen inIhe northern coastal area. The inhabilanls ofihe lerpen apparently conlinued their formerway of life of growing Hordeum etc. (van Zeist1974), although some material imports showthat these people were in conlacl wilh theRoman Empire. A rather unusual find of thiskind, which was, however, not recovered from aterp, but came from a "normal" seulement, is asmall Roman bronze flask filled with Raphanussativus seeds and some Apium graveolens,Origanum vulgäre and Malus sylvestris. Thecontents are interpreted as some kind of medi-cine (Buurman 1988b).

Only incidental finds are known from thehigher soils north of the limes. They consist ofHordeum vulgäre (site at Wijster), Hordeumvulgäre with some Triticum dicoccum (Dalfsen),Secale céréale with some Hordeum sp., Avena(sativa) and Panicum miliaceum (Noordbarge),and Secale céréale with a minor admixture ofPanicum miliaceum, Hordeum vulgäre andHordeum vulgäre var. nudum (Ede-Veldhuizen)

292

(van Zeist 1970, 1976). A well at Wijsteryielded stem fragments of braked Linum usi-tatissimum.

From the foregoing it is clear that the nakedwheat and spelt found in the Roman militaryand civil settlements were not produced nearor north of the limes. They must have beenbrought in from other areas. It is a historicallyknown fact that large quantities of these wheatswere grown on the fertile soils in the south ofthe Dutch province of Limburg (and the adja-cent German Rhineland), in Belgium and innorthern France, but, as already stated above,the botanical evidence is still restricted to onlyone villa rustica.

8 THE MIDDLE AGES AND LATERHISTORICAL PERIODS (1500 BP- )

No pnlaeoethnobotanical data are yet availableon the period immediately after the collapse ofthe Roman Empire. The record starts with thesixth century AD. Quite a number of publica-tions have recently appeared on medieval sitesfrom this time onwards. For France the reporton the farms (villae) at Villiers-le-Sec (Seine-et-Oise) and Baillet-en-France (Seine-et-Oise)should be mentioned (Ruas 1988). The Villiers-Ie-Sec series covers the second half of the sixthup to and including the l l t h century. Secalecéréale was the dominant cereal, closely fol-lowed by Triticum aestivum. Avena sativa andHordeum vulgäre came next, whilst remains ofPanicum miliaceum were found only once, inan early (Merovingian) feature. Vicia faba var.minor was as common as oats and barley, atleast in the later (Carolingian) periods. Lesscommon was another legume: Pisum sativum.The list also mentions Linum usitatissimum,Cannabis sativa, Vit is vinitera, Malus sp., Pyruscommunis, Prunus insititia, Juglans regia, Cory-lus avellana, Prunus persica and Humulus lupu-lus. The finds do not appear to change withtime.

Another series of data has been obtained inthe Netherlands. Van 7,eist and co-authors(1986) compared three rural sites on the sandysoils of the province of Drenthe: Pesse, Odoornand Gasselte. They concluded that Secale cere-ale, Hordeum vulgäre, Avena sativa and Avenastrigosa were the cereals cultivated in this area.The scarce finds of Panicum miliaceum cannotbe considered evidence of the intentional cul-tivation of millet. The data suggest that thedominance of rye shifted in the course of timetowards a more balanced cultivation with more

Avena. In addition to cereals, Linum usitatis-simum, Vicia faba var. minor and Pisum sati-vum were cultivated. It is questionable whetherCamelina sativa was still being grown, becauseits few representatives were found in associa-tion with Linum, which suggests that it mayhave been a weed in flax fields. Remains ofkitchen herbs were also found: Anethum gra-veolens, Satureja hortensis, Apium graveolensand Foeniculum vulgäre. The large amount offield weeds indicates that the harvest wasthreshed and cleaned in the settlement. Theweeds show that some of the crops were sownin autumn, but that spring-sowing was alsopractised

In this period the coastal sites, in particularthe terpen, were still, continuously, inhabited.In his 1988 paper, van Zeist analyses the pos-sibility tha t the coastal sites imported cropproducts from the settlements on the higher,sandy soils. He obtained a positive result forthe coastal settlement of Leeuwarden, but theevidence obtained for the majority of settle-ments in the coastal marsh was less convincing.Leeuwarden was a kind of regional marketplace and its botanical record may have beenaffected by the town's trading activities. Theimported product was unthreshed rye (andmosses). For further information on coastalsites the reader is referred to van Zeist's 1974paper. The economy of the inhabitants of thearea under review there seems to have re-mained unchanged for a long stretch of time.Until the first dikes were built, around AD1000, the farming conditions remained more orless the same.

It is not possible to list all the early medievalsites in this paragraph, but two other sitesshould be mentioned: Dommelen (AD 700-1250) and Kootwijk (AD 750-1000), both ruralsites on sandy soils. The Dommelen series hasnot yet been published in full, but preliminaryreports show that Secale céréale dominatedduring both the Merovingian and Carolingianperiods. Hordeum vulgäre and Avena sativawere of minor importance. Linum usitatissimumhas also been found, as seeds and, in wells, asstem fragments (van Vilsteren 1985). The mostremarkable find, however, is a fruit of Fagopy-rum esculentum, which was dated to the middleof the twelfth century on the basis of the pot-tery found in association with it but to920±100 BP on the basis of charred grainfound in the same context. It is the earliest findso far recovered in this area, excluding pollen.The assemblages gave rise to thoughts on themethod of harvesting. During the later phasesof occupation this must have be done using a

293

Table 6. Some species found in cesspits, after Vermeeren in press (1), Kuijper 1984b (2), Marinvalin Petit 1983 (3), Bakels 1980b (4), van den Brink 1988 (5), Pals 1983 (6), van Zeist 1987 (7),Paap 1983 (8).

B C 8Q ~ •£<5 Q 2

I -I 5s u * -

g I2 S 3a-s- c. ?? eg rf e jj

•* M U TZ _^ •

Ü 2 T * ^ O cu. -- ~ C3 u k« CO 05 .« Z J O -5 •-

"Grain":Cerealia branAvena sativaHordeum sp.Oryza sativaPanicum miliaceumSc ca le céréaleTriticum aestivumFagopyrum esculentum

Pulses:Lens culinarisPisum sativumVicia faba

Fruits and nuts:Castanea sativaCucumis sativusCucumis meloCucurbita pepoCocos nuciferaCorylus avellan:iFicus caricaFragaria vescaFragaria sp.Juglans regiaMalus domesticaMespilus germamcaMorus nigraOlea europaeaPhoenix dactyliferaPhysalis alkekengiPrunus aviumPrunus cerasusPrunus domestica

Prunus dulcisPrunus insititiaPunica granatumPyrus communisRibes sp.Rubus caesiusRubus fruticosusRubus idaeusSambucus nigraVaccinium sp.Vitis vinifera

Greens, herbs etc.:Apium graveolensAnethum graveolensAnthriscus cerefoliumBeta vulgarisBrassica sp.Capsicum annuumCarum carviCichorium intybusCoriandrum sativumFoeniculum vulgäreLepidium sativumPetroselinum crispumPortulaca oleraceaSinapis alba

Others:Cannabis sativaCarthamus tinctoriusHumulus lupulusLinum usitatissimumPapaver somniferumReseda luteola

'• '>' .

1:

+

' l

" ,- r

Q

D" '

•3

+

- 'B: .

" .

:U

5

++

- •

0

i- ;

•

/,

1

ogen

bosc

h,

u

.'•

+

+

:. :•

i

!:

::;

+

+

r 1

' ;

- 1'

Mc '

::U

ï

• •

.

i

;

p

K,

ronin

-

+

+

iu••-''

\\•••

3[<++

+

é<. ,i

Q•-.p• Q

ronin

C

+

+

+ + . + + + . + + ++ + + + + + + + + +

scythe, as 50% of the weeds are known to below-growing species. The size of the weedseeds found together with the remains of cropplants suggests the use of grain sieves with amesh width of about 2.5 mm (van Vilsteren1984).

Kootwijk is a clear case of an abandonedvillage. The inhabitants left the site when drierclimatic conditions had caused the pool whichprovided their drinking water to dry out com-pletely. Again the samples indicated Secalecéréale as the most common cereal. Hordeumvulgäre and Avena sativa came next whilstVicia faba var. minor, Linum usitatissimum and

Camelina sativa were less frequent in the sam-ples. The list of crop plants also mentionsReseda luteola. Malus sp., Corylus avellana,Rosa sp., Fragaria vesca, Rubus idaeus andRubus sect. Rubus were gathered. The materialwas too scanty to permit conclusions as to ashif t in the relative importance of any crop inthe course of time. The presence of enclosedfields around the village suggests that the cropswere raised in an infield - outfield system (Pals1987).

The investigation in the town of Leeuwardenmentioned above is an example of the work

294

done on towns and non-rural sites such asmonasteries and castles. More and more sam-ples are being taken for seed analysis duringthe archaeological investigations carried out inconnection with the restoration of old buildingsand the renovation of old town quarters. Thepublished results are too numerous to be dis-cussed in full here. The lists of crop plants,cultivated and gathered fruits, kitchen herbs,industrial crops and the like are usually ratherlong because the remains are often found wellpreserved in waterlogged and even mineralizedcondition. Table 6 gives an idea of the contentsof cess-pits, a rich source of material from thelate Middle Ages to quite recent times.

An example of the research done on cess-pitsis provided by the complexes from the 13th upto and including the 19th century investigatedin Amsterdam. One of the results of thisresearch is that the town has always importedfood, ever since its beginning as a small village.Assemblages dating from AD 1500-1550 con-tained Neslia paniculata, which is in accordancewith the historically testified large-scale importof grain from Prussia and Poland at the time.Oryza saliva was quite common in a 17th-18thcentury Jewish quarter and the presence ofScirpus mucronatus suggests that this rice camefrom the Mediterranean, perhaps Italy (Paap1983). This same Jewish quarter can be dividedinto a richer and a poorer part. A thoroughanalysis of the differences between the twosocial classes has not yet been published, butone of the preliminary reports mentions thatthe main difference found is that the rich peo-ple ate other kinds of plums than their poorerneighbours (Paap 1984). The published datareveal no other striking differences in the occu-pants' diets.

Here the survey of the palaeoethnobotany ofhistorical times must end. However, there is anabundance of literature available on this sub-ject. This deserves a synopsis of its own.

9 CONCLUSION

The article by van Zeist mentioned in the in-troduction stated that 25 years of collectingseeds in the course of systematic excavationsprovided the opportunity to fill up at leastsome of the gaps in our knowledge of cropplants in the Netherlands (van Zeist 1970: 43).This knowledge was mostly limited to whichplants occurred in which periods. The same istrue of the information contained in two othersummaries of the results obtained in previousresearch.

Twenty years later it appears that we havestill been busy fillings up gaps in the species-time record and are continuing to do so. In thisrespect no progress has been made in our wayof working and thinking. Fortunately, progresshas been made in other respects. It is observ-able in the work of those authors who had theopportunity to devote themselves to one site, toone small and geographically restricted area, orto one archaeological culture. The questionsraised are mostly of an economical nature. Oneof these is whether people grew their own foodor not. Others concern the location of thefields, the time of sowing, the harvestingmethods and the processing of harvested crops.Conspicuously, much of the more advancedwork concerns wetland sites. The reason isundoubtedly that more time and effort arespent in investigating these sites because theyare rich in plant remains. It is striking thateven the carbonized remains obtained fromsuch sites can provide answers to some of themain questions. In one case, one of the prob-lems to be solved led to experimental work, i.e.the growing of crops in a salt marsh un-protected by dikes.

As stated above, true progress is dependanton the amount of time available for research.Where researchers are scarce, or have to workwith samples taken by non-specialists, or haveto hurry from one excavation to another, pro-gress is limited to the filling in of gaps in ourtime-table knowledge of plants.

REFERENCES

Bakels, C.C. 1976. Graanindrukken in TRB-aarde-werk. In P.J.R.Modderman, J .A.Bakker & H.A.Heidinga (eds.), Nederzettingssporen uit het Mid-den-Neolithicum (TRB), Late Bronstijd en Middel-eeuwen in het Beekhuizer Zand onder Harderwijk,Prov. Gelderland. Analecta Praehistorica Leidensia9:69.

Bakels, C.C. 1979. Linearbandkeramische Früchteund Samen aus den Niederlanden. In U.Körber-Grohne (Hrsg.), Festschrift Maria Hopf. Archaeo-Physika 8:1-10.

Bakels, C.C. 1980a. De bewoningssporen van deMaaskant I: Plantenresten uit de Bronstijd en Ro-meinse Tijd gevonden te Oss-IJsselstraat, Prov.Noordbrabant. Analecta Praehistorica Leidensia13:115-131.

Bakels, C.C. 1980b. Een Sittardse beerput en mest-vaalt. Archeologie in Limburg 9:2-3.

Bakels, C.C. 1981. Neolithic Plant Remains from theHazendonk, Province of Zuid-Holland, the Nether-lands. Zeitschrift für Archäologie 15:141 -148.

Bakels, C.C. 1982. Der Mohn, die Linearbandkera-mik und das westliche Mittelmeergebiet. Archäolo-gisches Korrespondenzblatt 12:11-13.

Bakels, C.C. 1982-'83. Les graines carbonisées de

295

Fort-Harrouard (Eure-et-Loir). Antiquités Natio-nales !4(5):59-63.

Bakels, C.C. 1984. Carbonized seeds from NorthernFrance. Analecta Praehistorica Leidensia 17:1-27.

Bakels, C.C. 1987. Oss-Ussen en de tuinboon. Bij-drage tot de Studie van het Brabantse Heem 31:137-139.

Bakels, C.C. 1988. Hekelingen, a Neolithic site in theswamps of the Maas estuary. In H.Küster (Hrsg.),Der prähistorische Mensch und seine Umwelt(Festschrift U.Körber-Grohne). Forschungen undBerichte zur Vor- und Frühgeschichte in Baden-Württemberg 31:155-161.

Bakels, C.C., in press a. The Crops of the RossenCulture: signifantly different from their Bandke-ramik predecessors, French influence? Actes ducolloque "Rubané et Cardial" 1988, Etudes et Re-cherches Archéologiques de l'Université de Liège.

Bakels, C.C., in press b. The Crops of the RossenCulture. Acta Interdisciplinaria Archaeologica,Nitra.

Bakels, C.C., in press c. Tracing crop processing inthe Bandkeramik culture. In J.Renfrew (ed.), Newlight on Early Farming: Recent developments inPalaeoethobotany. Edinburgh.

Bakels, C.C. & R.W.J.M.van der Ham 1980. Verkooldafval uit een Midden-Bronstijd en een Midden-IJzertijd nederzetting op de Hooidonksche Akkers,Gem. Son en Breugel, Prov. Noord-Brabant.Analecta Praehistorica Leidensia 13:81-91.

Bakels, C.C. & R.Rousselle 1985. Restes botaniqueset agriculture du Néolithique ancien en Belgique etaux Pays Bas. Helinium 25:37-57.

Beijerinck, W. I929-I93I. De subfossiele plantenres-ten in de terpen van Friesland en Groningen Wa-geningen. 3 Vols.

Billoret, R. 1968. Etival-Clairefontaine. Gallia 26:392-393.

Boone, Y. & J.Renault-Miskovsky 1976. La Cuillette.In H.de Lumley (éd.), La Préhistoire Française I,p.684-687. Paris, Editions du Centre National de laRecherche Scientifique.

Bottema, S., T.C.van Hoorn, H.Woldring & W.H.E.Gremmen 1980. An agricultural experiment in theunprotected salt marsh, part II. Palaeohistoria 22:127-140.

Brandt, R.W., W.Groenman-van Waateringe & S.E.van der Leeuw 1987. Assendelver Polder Papers I.Cingula 10

Brink, W.van den 1988. Zaden en pollen uit een 16eeeuwse beerput uit de Postelstraat. In H.W.Boekwijt& H.L.Janssen (eds.), Kroniek van bouwhistorischen archeologisch onderzoek 's-Hertogenbosch,p.113-124.

Bult, E.J., P.van der Hout & T.Immerzeel 1988. EenRomeins grafveldje aan de Tiendweg bij Naald-wijk. Westerheem 37:118-125.

Buurman, J. 1979. Cereals in Circles - Crop Proces-sing Activities in Bronze Age Bovenkarspel (theNetherlands). In U.Körber-Grohne (Hrsg.), Fest-schrift Maria Hopf. Archaeo-Physika 8:21-37.

Buurman, J. 1986. Graan in IJzertijd-silo's uit Colm-schate. Nederlandse Archeologische Rapporten 3:67-73

Buurman, J. I988a. Economy and environment inBronze Age West-Friesland, Noord-Holland (fromwetland to wetland). British Archaeological Re-ports, British Series 186:267-292.

Buurman, J. I988b. Roman medicine from Uitgeest.

In H.Küster (Hrsg.), Der prähistorische Mensch undseine Umwelt (Festschrift U.Körber-Grohne). For-schungen und Berichte zur Vor- und Frühgeschich-te in Baden-Württemberg 31:341-351.

Casparie, W.A., B.Mook-Kamps, R.M.Palfenier-Vegter, P.C.Struijk & W.van Zeist 1977. The pa-laeobotany of Swifterbant. A preliminary report.Swifterbant contribution 7. Helinium 17:28-55.

Ceunynck, R.de & C.Verbruggen 1985. Over de oor-sprong van de kultuurplanten in Viaanderen.VOBOV-INFO 18/19:7-15.

Ceunynck, R.de, M van Strijdonck & J.Vanmoerker-ke 1984. IJzertijd-kuil te Evergem-Ralingen.VOBOV-INFO 17:14-20.

Erroux, J. 1976. Les débuts de l'agriculture en Fran-ce: les céréales. In J.Guilaine (ed.), La PréhistoireFrançaise II, p. 186-191. Paris, Éditions du CentreNational de la Recherche Scientifique.

Groenman-van Waateringe, W. 1989. Food for Sol-diers, food for thought. In J.C.Barrett, A.P.Fitz-patrick & L.Macinnes (eds.), Barbarians and Ro-mans in North-West Europe. British ArchaeologicalReports, International Series 471:96-107.

Haalebos, J.K. 1986. Ausgrabungen in Woerden(1975-1982). In Studien zu den MilitärgrenzenRoms III. 13. Internationaler Limeskongress Aalen.Forschungen und Berichte zur Vor- und Frühge-schichte in Baden-Württemberg 20:169-174.

Halt, J.J. & H.Zumstein I960. Découvertes d'un fourde potier de l'Age de Bronze final à Cronenbourg.Cahiers Alsaciens Archéo Art. Hist. 1960 (4):l7-26.

Heim, J. 1979. Recherches paléobotaniques au sitenéolithique (Roessen) de la "Bosse de l'Tombe" àGivry. Bulletin de la Société Royale Belge d'An-thropologie et de Préhistoire 90:65-78.

Heim, J. 1983. Apports récents de la paléobotaniqueà la connaissance de l'importance des activités cul-turales (agricoles) des néolithiques anciens entreRhin et Seine. Dissertationes Archaeologicae Gan-denses 21:62-70.

Heim, J. 1987. Etude des restes végétaux carbonisésprovenant des fosses d'âge Epi-Roessen/Michels-berg situées à Ittre (Brabant-Belgique). Helinium27:61-70.

Helgouach, J.P 1977. Informations archéologiquesPays de la Loire. Gallia Préhistoire 20:438-440.

Helgouach, J T I98I. Informations archéologiquesPays de la Loire. Gallia Préhistoire 24:425-437.

Hinout, J. 1964. Gisements tardenoisiens de l'Aisne.Gallia Préhistoire 7:65-106.

Hopf, M. 1969. Carbonized plant remains fromChassémy (Aisne). World Archaeology I: 130-133.

Hopf, M. 1975. Fosses néolithiques d'Entzheim. Re-vue Archéologique de l'Est et du Centre-Est 26:115-117.

Hopf, M. 1980. Sämereien aus der Siedlung Weilerzum Turm. Publications de la Section Historique del'Institut G.P.de Luxembourg 94:125-127.

Hopf, M. 1987. L'habitat néolithique récent (groupede Munzingen) de Geispolsheim (Bas-Rhin). Pflan-zenfunde aus Geispolsheim. Cahier de l'Associationpour la Promotion de la Recherche Archéologiqueen Alsace (APRAA) 3:130-133.

Huyge, D. & P.Vermeersch 1982. Late NeolithicSettlement at Weelde-Paardsdrank. Studia Praehis-torica Belgica 1:117-209.

Jouve, M. 1973. La cabane gauloise du Barillet,Béthisy-St. Martin (Oise). Revue archéologique del'Oise 3:27-37.

296

Knörzer, K.H. 1971. Urgeschichtliche Unkräuter imRheinland. Vegetatio 23:89-111.

Kui jper , W. 1984a. Plantenresten uit Romeins Maas-tricht. Archeologie in Limburg 21:3-8.

Kuijper , W.J. 1984b. Planten- en dierenresten inlaat-middeleeuwse beerputten op het terrein vanhet St. Agnietenklooster in Leiden. Bodemonder-zoek in Leiden, Jaarverslag 1984:131-139.

Kuijper, W.J. & H.Turner, in press. Diet of a Romanofficer at Alphen aan de Rijn, the Netherlands, inthe first century AD. Review of Palynology andPalaeobotany.

Lange, A.G., 1988. Plant Remains from a nativesettlement at the Roman Frontier: De Horden nearWijk bij Duurstede: a numerical approach. Ph.D.Thesis, University of Groningen. [Will appear inNederlandse Oudheden, Amersfoort.]

Lauwers, R. & P.M.Vermeersch 1982. Un site duMésolithique ancien à Neerharen-De Kip. StudiaPraehistorica Belgica 1:3-52.

Ledden-Hulsebosch, J.C.van 1955. Chemisch enmikroskopisch onderzoek van monsters van Valken-burg Z.H. Jaarverslag van de Vereniging voorTerpenonderzoek 33-37:190-191.

Marinval , P. 1983. Études de quelques semencesarchéologiques provenant de niveaux de l'Age duBronze de la Grotte du Queroy, Chazelles (Charen-te). Société Archéologique et Historique de la Cha-rente, Bulletins et Mémoires 4:203-214.

Marinval, P. 1988. L'alimentation végétale en France,du Mésolithique jusqu'à l'Age du Fer. Paris, Édi-tions du Centre National de la Recherche Scienti-fique.

Newell, R.R. 1973. The post-glacial adaptations ofthe indigenous population of the northern europeanplain In S.K.Kozlowski (ed.). The Mesolithic inEurope, p.399-476. Warsaw.

Paap, N. 1983. Economic plants in Amsterdam: Qua-l i ta t ive and Quantitative Analysis. British Archae-ological Reports, International Series 181:315-325.

Paap, N.A. 1984. Palaeobotanical investigations inAmsterdam. In W.van Zeist & W.A.Casparie (eds.).Plants and Ancient Man, p.339-344. Rotterdam,Balkema.

Pals, J.P. 1983. Zaden en vruchten uit een 16e eeuw-se beerput opgegraven in het stadsdeel Bakenes.Haarlems Bodemonderzoek 17:32-37.

Pais, J.P. 1984. Plant remains from Aartswoud, aNeolithic settlement in the coastal area. In W.vanZeist & W.A.Casparie (eds.). Plants and AncientMan, p.313-322. Rotterdam, Balkema.

Pals, J.P. 1987. Reconstruction of landscape andplant husbandry. In W.Groenman-van Waateringe &L.H.van Wijngaarden-Bakker (eds.). Farm Life in aCarolingian Village, p.52-96. Assen/Maastricht,Van Gorcum.

Pals, J.P., D.C.M.Beemster & A.P.Noordam 1989.Plant remains from the Roman castellum Praeto-r ium Agrippinae near Valkenburg (prov. of Zuid-Holland). In U.Körber-Grohne & H.Küster (Hrsg.),Archüobotanik. Dissertationes Botanicae 133:117-134.

Pals, J.P. & T.Hakbijl, in press. A deposit of water-logged grains in a ship at the Roman fort of Lau-rium in Woerden, prov. Zuid-Holland. Review ofPalynology and Palaeobotany.

Parent, R. 1973. Fouille d'un atelier tardenoisien à laSablonniére de Coincy (Aisne). Bullet in de la So-ciété Préhistorique Française 70:337-351.

Parent, R. & N.Planchais 1972. Nouvelles fouilles surle site tardenoisien de Montbani (Aisne). Bulletinde la Société Préhistorique Française 69:508-532.

Patte, E. 1971. Quelques sépultures du Poitou, duMésolithique au Bronze moyen. Gallia Préhistoire14:139-244.

Petit, D. 1983. Orléans, Saint-Pierre-Lentin 1977-1978. Revue Archéologique du Loiret 9:11-34.

Petrequin, P. 1970. La grotte de la Baume de Gonvil-lars. Annales littéraires de l'Université de Besançon,Paris.

Reynolds, P. 1974. Experimental Iron Age StoragePits: an interim report. Proceedings of the Prehis-toric Society 40:118-131.

Roebroeks, W. 1985. Archaeological research at theMaastricht-Belvedere pit: a review. Analecta Prae-historica Leidensia 18:109-118.

Rowlett, R.M., A.L.Price & M.Hopf 1982. Differen-tial grain use on the Titelberg, Luxemburg. Journalof Ethnobiology 2:79-88.

Roymans, N. 1985a. Carbonized grain from two IronAge storage pits at Neerharen-Rekem. ArchaeologiaBelgica 1:97-105.

Roymans, N. 1985b. Nederzettingssporen uit deMidden-IJzertijd op de Kerkakkers bij Dommelen.Bijdragen tot de Studie van het Brabantse Heem27:11-18.

Ruas, M.P. 1988. Alimentation végétale, pratiqueagricole et environnement 7e-IOe siècle. In Unvillage au temps de Charlemagne, p.203-217. Paris.

Sanden, W.A.B.van der I987a. Oss-Ussen: écologie eneconomie. Bijdragen tot de studie van het BrabantseHeem 31:81-88.

Sanden, W.A.B.van der I987b. The Ussen-project:large-scale settlement archaeology of the period 700BC-AD 250, a preliminary report. Analecta Prae-historica Leidensia 20:95-123.

Therkorn, L.L., R.W.Brandt, J.P.Pals & M.Taylor1984. An Early Iron Age farmstead: Site Q of theAssendelver Polders project. Proceedings of thePrehistoric Society 50:367-372.

Thevenin, A. & J.Sainty 1974. Archenheim, Ober-larg. Archeologia 75, octobre:49-61.

Vanvinckenroye, W. 1988. De Romeinse villa op deSassenbroekberg te Broekom. Publicaties van hetGallo-Romeins Museum Tongeren 38.

Vermeeren, C.,in press. Botanisch onderzoek aanmiddeleeuwse beerputten uit Kampen. Publicatiesvan de Stichting Archeologie Ussel/Vechtstreek.

Vermeersch, P.M., A.V.Munaut & E.Paulissen 1974.Fouilles d'un site du Tardenoisien final à Opglab-beek. Quartär 25:85-104.

Vilsteren, V.T.van 1984. The medieval village ofDommelen: a case study for the interpretation ofcharred seeds from postholes. In W.van Zeist &W.A.Casparie (eds.). Plants and Ancient Man, p.227-235. Rotterdam, Balkema.

Vilsteren, V.T.van 1985. Verkoolde zaden en vruch-ten van de Middeleeuwse nederzetting bij Dom-melen. Bijdragen tot de studie van het BrabantseHeem 27:41-43.

Willems, W.J.H. & L.I.Kooistra 1987. De Romeinsevil la te Voerendaal. Archeologie in Limburg 32:29-38.

Willerding, U. 1988. Zur Entwicklung von Acker-unkrautgesellschaften im Zeitraum vom Neoli thi-kum bis in die Neuzeit. In H.Küster (Hrsg.), Derprähistorische Mensch und seine Umwelt (Fest-schrift U.Körber-Grohne). Forschungen und Be-

297

richte zur Vor- und Frühgeschichte in Baden-Württemberg 31:31-41.

Zeist, W.van 1970. Prehistorie and Early HistoricFood Plants in the Netherlands. Palaeohistoria14:43-173.

Zeist, W.van 1974. Palaeobotanical studies of settle-ment sites in the coastal area of the Netherlands.Palaeohistoria 16:223-371.

Zeist, W.van 1976. Two early rye finds from theNetherlands. Acta Botanica Neerlandica 25:71-79.

Zeist, W.van 1987. Plantenresten uit twee beerputtenin Groningen. In A.Carmiggelt, H.van Gangelen,G.Kortekaas & W.van Zeist (eds.). Uitgeput Huis-raad, p.65-74. Groningen, Stichting Monument enMateriaal.

Zeist, W.van 1988. Botanical evidence of relations

between the sand and clay districts of the north ofthe Netherlands in medieval times. In H Küster(Hrsg.), Der prähistorische Mensch und seine Um-welt (Festschrift U.Körber-Grohne). Forschungenund Berichte zur Vor- und Frühgeschichte in Ba-den-Württemberg 31:381-387.

Zeist, W.van, T.C.van Hoorn, S.Bottema & H.Wol-dring 1976. An agricultural experiment in the un-protected salt marsh. Palaeohistoria 18.111-153.

Zeist, W.van & R.M.Palfenier-Vegter 1983. Seedsand fruits from the Swifterbant S3 site. Final re-ports on Swifterbant 4. Palaeohistoria 23:105-168.

Zeist, W.van, G.J.de Roller, R.M.Palfenier-Vegter,O.H.Harsema & H.During 1986. Plant Remainsfrom Medieval sites in Drenthe, the Netherlands.Helinium 26: 326-274.

298