A comparison of early Neolithic crop and weed assemblages from the Linearbandkeramik and the Bulgarian Neolithic cultures: differences and similarities
- First Online:
- Cite this article as:
- Kreuz, A., Marinova, E., Schäfer, E. et al. Veget Hist Archaeobot (2005) 14: 237. doi:10.1007/s00334-005-0080-0
- 377 Views
The spread of early agriculture from the Mediterranean to central Europe is still poorly understood. The new subsistence reached western central Europe during the second half of the 6th millennium cal b.c. This paper presents a comparison of crop and weed species from 33 Bandkeramik sites from Austria and Germany and six Bulgarian Neolithic sites. The aim is to investigate whether the early cultivation system brought in from the eastern Mediterranean was adapted to European conditions in Bulgaria or further West. Some characteristics of the potential weeds are interpreted with respect to the cultivation systems and the origin of the species.
KeywordsEarly NeolithicBulgariaGermanyAustriaCultivation systemsWeeds
From recent excavations in Bulgaria new archaeobotanical evidence is available (Marinova 2000, 2001; Marinova et al. 2002; Popova 1995a, b; Thanheiser 1997). It allows a comparison of agricultural data from the Bulgarian Neolithic with that of the early Neolithic in Germany and Austria. This comparison offers the opportunity to investigate whether the early cultivation system brought in from Turkey and Greece was adapted to European conditions in Bulgaria or further West. The spread of the Neolithic to the western Mediterranean is not discussed in this paper.
The Karanovo Culture, named after the famous tell site at the border of the Thracian plain, became a synonym for the Neolithic and Chalcolithic of Bulgaria (Fig. 1). Today, different regional groups are differentiated. They all have in common a similar settlement structure and architecture as well as a red-slipped pottery with white, later dark red, ornamentation (Georgiev 1981; Lichardus-Itten et al. 2002; Nikolov 2000, 2002, 2004; Todorova 1981). In addition some early groups apparently produced monochromatic pottery (for a critical review see Lichardus-Itten and Lichardus 2003). In the following we will use the terms Early, Middle and Late Bulgarian Neolithic to avoid these group names.
The first farmers of Bulgaria settled in the foothills around the Thracian plain and in those of south-western Bulgaria. Possibly the Struma valley played an important role during the introduction of Neolithic subsistence from Thessaly to Bulgaria (Perles 2001; Nikolov 2004). It is still a matter of dispute whether the new subsistence arrived from Greece exclusively or from Anatolia via Turkish Thrace or from both regions (for example Parzinger 1993, p 84). The state of research does not allow a final statement concerning this question. Recent excavations, for example at the Early Neolithic site of Yabalkovo at the southern border of the Bulgarian part of the Thracian plain, will provide new insights into this subject (K. Leshtakov personal communication).
Finds of the earliest Bandkeramik have been made in a huge area of western central Europe. In the second part of the Bandkeramik from phase II to phase V (chronological phases based on the pottery after Meier-Arendt 1966) the distribution area became even larger (Lüning 2000). Finds and settlements are spread between the Paris Basin and the Black Sea (Fig. 1).
Archaeological and botanical evidence points to western Hungary as the centre of Bandkeramik origin. This Hungarian Bandkeramik culture was possibly strongly influenced by the Neolithic Starčevo-Köros-Čris complex of eastern Hungary, Serbia and Romania (Fig. 1; Bánffy 2001; Kalicz 2001; Lichardus-Itten and Lichardus 2003; Lüning 1991, 2000; Lüning et al. 1989).
Similarities and differences of both the Bandkeramik and the Bulgarian Neolithic cultures are reflected for example by the structures of settlements and buildings. In both early Neolithic cultures the houses were built with a timber framework. This is an important difference to the mud-brick houses of Greece and Turkey (Parzinger 1993, pp 294ff.; Perles 2001, pp 172ff.). The Bandkeramik sites are open flat settlements each comprising just a few houses. The houses were about 30 m long and 6 m wide. There are postholes and wall-ditches as well as some pits (see papers in Eckert et al. 2003), but the ground surface is eroded, so that the house floors are not preserved. In contrast to the Bandkeramik, the Bulgarian Neolithic sites are either multilayer flat settlements or tell sites (Georgiev 1961, 1981; Hiller 1993; Lichardus-Itten et al. 2002; Todorova 1981, Todorova and Vaissov 1993). It is still open to discussion why some Neolithic settlements in Bulgaria stopped gaining height before becoming a real tell.
In both areas botanical material can be recovered from rubbish pits. In addition, in Bulgaria, culture layers of levelled houses as well as ground floors of buildings and their surroundings are preserved in situ as they have been covered by layers of settlement waste. The latter are the places where the storage finds or other massive concentrations of plant remains can be found, if the house had burnt down (Dennell 1978; Dotcheva 1990; Marinova 2001; Thanheiser 1997; Tschakalova and Božilova 2002; Tschakalova and Sârbinska 1986).
In contrast to Bandkeramik settlements characterized by single standing long-houses, the Bulgarian Neolithic villages consist of rows of houses (Georgiev 1961, 1981; Hiller 1993; Lichardus-Itten et al. 2002; Todorova 1981, Todorova and Vaissov 1993). The Bulgarian Neolithic rectangular houses were about one third the size of the Bandkeramik long-houses. All these differences have important social implications (Parzinger 1993, p 295). They could imply for example different family or group sizes and structures.
The first farmers of both cultures - the Bulgarian Neolithic and the Bandkeramik - settled in landscapes very well suited for agriculture. In Bulgaria these are mostly regions with brown soils (Cambisols) and a sub-Mediterranean to sub-continental climate. Present-day average mean temperature covers the range between 10 and 14°C, average precipitation—with two maxima, the main in May/June and a secondary one in November/December—between 500 and 700 mm (Egger 1997; Horvat et al. 1974; N. Ninov 2002; Velev 2002; Kopralev 2002).
In Germany and Austria the landscapes settled first were mostly characterized by Chernozem soils, developed from loess or fluvial sediments, and by today’s warm and dry climate. Present-day average mean temperature lies between 7 and 9°C, average precipitation - with a maximum during (June/)July/August—between 550 and 650 mm (Bakels 1978; Kreuz 1990, p 7ff.; Lüning 2000; Sielmann 1971). High lake levels could be interpreted as signs of a wetter climate in central Europe at that time. This might have been induced by precipitation, possibly in form of heavy rainfalls (Beug et al. 1999; Bouzek 2001; Haas et al. 1998; Hormes et al. 2001; Kalis et al. 2003; Kreuz 1990, p 8; Magny 1998; Maise 1998; Schmidt and Gruhle 2003; Spurk et al. 2002).
The reconstruction of the temperatures around 6000 b.c. suggests that in south-eastern Europe the mean winter temperatures might have been almost at today’s level and the summer temperatures were slightly lower than today. Cooler and wetter conditions than today are discussed (Davis et al. 2003).
Landscapes of both areas have been reconstructed by palynologists and anthracologists as more or less densely wooded. As in central Europe, early Neolithic human impact is difficult to trace by pollen analysis in Bulgaria, as there are only weak signals indicating cultivation. In both regions this might be due to the fact that the deposits analyzed are not adjacent to the settlements and fields, and the former woodland canopy had acted as a pollen filter.
In Germany and Austria, the forest cover was also formed by deciduous woodland of comparable species to those in Bulgaria like Quercus petraea and Q. robur, Fraxinus excelsior, Ulmus sp., Tilia sp. Acer sp. and others (Bakels 1978; Beug 1986, 1992; Kreuz 1990 pp 17ff., 1995, in press a; Liese-Kleiber 1997; Litt 1990; Schäfer 1996; Schweizer 2001; Van Zeist 1967; Van Zeist and Van der Spoel-Walvius 1980). There are different opinions concerning for example the percentages of Quercus (oak) and Tilia (lime) trees and other woodland species or the kind of woodland cover of the flood plains, but this subject is not of interest here. In our investigation areas, the forest cover on Chernozems was formed by deciduous woodland. The results of charcoal analysis can be interpreted as managed hedges serving as supply of firewood (Groenman-van Waateringe 1970; Kreuz 1988, 1992).
Archaeozoological investigations from settlements of both cultures revealed that the spectra of domestic and wild animal species are comparable (Arbogast et al. 2001; Benecke and Ninov 2002; Kovachev and Georgiev 2002; Ninov 1992, 1999). It is difficult to estimate what was the main domestic animal at the different sites. This is due to the fact that preservation of bones differs not only between sites but also within one single excavation area due to changing soil conditions. It has to be stressed that there is no dominance of sheep or goat detectable in either region.
Figures 2a and b show the location of the 33 Bandkeramik settlement sites investigated archaeobotanically. Determinations have been carried out by Nicole Boenke (Götzis, Austria), Angela Kreuz, Elena Marinova, Ursula Thanheiser (Wien, Austria) and Julian Wiethold. The data from Hesse, northern and southern Germany and south-western Austria are methodologically comparable to each other. Therefore all data have been included in the following evaluation which has been carried out with the help of our archaeobotanical database programme ArboDat (Kreuz and Schäfer 2002; for the explanation of terms see also there). As it is often difficult to differentiate the Bandkeramik phases III to V based on fragmentary pottery finds, we combined the archaeobotanical results into one later Bandkeramik group LBK III–V. Features which could not be dated archaeologically more exactly than phases ‘LBK IIff.’ or ‘LBK II/III’ are not included in the calculations.
Due to the soil conditions in settlements of both cultures plant remains are preserved by charring or mineralization only. The contexts sampled at the Bulgarian Neolithic sites are often not “real” archaeological features but parts of bigger stratigraphic units such as layers or floors. There the squares excavated were taken as context units or “features”.
It has to be mentioned that post-holes and ditches contain very few or mostly no plant remains at Bandkeramik sites. Their density of plant remains is not comparable with that of pits, which—in their last function—were used for deposition of settlement waste. Therefore concentration values are based on samples from pits only (Table 3) and are always calculated without storage finds (66 Bulgarian Neolithic, 458 Bandkeramik features). The other calculations are based on samples from 93 Bulgarian Neolithic and 494 Bandkeramik features (see Tables 2 and 3).
Surprisingly, the seed concentrations of the two areas resemble each other. Extremely low seed concentration values occur only at sites of Bandkeramik phase I (Fig. 3a). Due to unknown reasons, fewer charred seeds were deposited in pits at that time.
The samples are rarely sufficiently rich in crops and weeds for statistical analysis. In addition they almost never derive from a single crop species or crop processing stage. So we are dealing with mixtures of crop remains and by-product material, waste and residues caused by accidents involving fire which were found in open contexts. For all these reasons and due to the different state of research and the different datasets of our Bandkeramik and Bulgarian Neolithic sites, the following comparison has to be more of a qualitative character.
Archaeobotanical records of cultivated plants from 33 Bandkeramik sites in Germany and Austria. LBK I Bandkeramik phase I, LBK II Bandkeramik phase II, LBK III-V Bandkeramik phases III-V; phases according to the chronology of Meier-Arendt (1966). Types of plant remains: gr grains, gb glume bases, ra rachis fragments, no culm fragments, s seeds
There are some hints given by storage finds, that einkorn and emmer were sometimes grown as maslins (mixed crops) by Bandkeramik as well as Bulgarian Neolithic farmers, partly even together with pea (Kreuz in press a; Marinova 2001, p 98; for the general discussion of maslins see Jones and Halstead 1995). In addition, the Bandkeramik farmers maintained lentil and flax fields. Papaver somniferum (opium poppy) is not recorded before Bandkeramik phase II. It may point to direct or indirect contacts with the western Mediterranean (Bakels 1982; Kreuz 1993). Opium poppy is absent from all Balkan Neolithic sites.
The Neolithic crop assemblage of Bulgaria comprises two additional cereals: Hordeum sp. (barley) and Triticum aestivum s.l./durum/turgidum (naked wheat), as well as three additional pulses: Cicer arietinum (chickpea), Lathyrus sativus vel cicera (grass pea) and Vicia ervilia (bitter vetch) (Table 4, Fig. 13; Marinova 2001). All in all, these five more crop species grown by the farmers of the Bulgarian Neolithic imply a different agricultural system.
If we look at the countries adjacent to Bulgaria and outside the Linearbandkeramik area we can see that the two additional cereals, barley and naked wheat, were grown in the sphere of influence of all cultures belonging to the Starčevo-Körös-Čris complex including eastern Hungary (Table 6 and Fig. 1). The pulses do not all reach eastern Hungary, but it is difficult to decide whether this is due to the state of research or to other reasons (data from Greece: Evi Margaritis unpublished data from the site Dispilio/Kastoria; Hubbard and Housley 2000; Valamoti 2004, Valamoti and Jones in press; Hungary: Amy Bogaard and Ferenc Gyulai unpublished data (see also Bogaard et al. in press); former Yugoslavia: F. Bittmann and D. Kučan unpublished data from Okoliste/Visoko in Bosnia; Ksenija Borojevič, Dragana Filipovič from the Vinča site (Borojevič 1998; Borojevič and Filipovič in press); Van Zeist 2003; Romania: Cârciumaru 1995; Cârciumaru and Monah 1987; F. and D. Monah 1996; Turkish Thrace: Reinder Neef personal communication for the site Aşaği Pinar; general overviews: Hopf 1991; Kroll 1991; Wasylikowa et al. 1991).
The spectrum of Bandkeramik crop species is limited. Only some of the species cultivated in Neolithic Bulgaria finally reached the area of the Bandkeramik in Austria and Germany in early Neolithic times. This phenomenon has to be discussed further (see below).
The relationship of einkorn and emmer
The results suggest that einkorn was the dominant Bandkeramik cereal. This is quite surprising, as einkorn would seem to be the worse choice. The yield of einkorn is almost half of that of emmer (Körber-Grohne 1988; Van der Veen 1997; organic farmers personal communication). In addition the lower tillering rate of einkorn allows more weeds to grow in the fields in relation to emmer (organic farmers personal communication). So why should anyone prefer einkorn? Einkorn is considered to be more winter hardy than emmer (Körber-Grohne 1988, pp 322ff.). But another characteristic seems even more important. Einkorn is the only cultivated cereal which, due to the characteristics of its straw, keeps standing after heavy rainfall (Fig. 10 in Kreuz in press a). Emmer on the other hand, as all other cereals, tends to lodge (personal communication of organic farmers and own observation). Lodging of cereal plants may reduce the yield seriously. In the event of frequently occurring heavy rainfall during the Atlantic period (see above) einkorn would have been the better choice (Kreuz in press a). Due to that possible climatic interpretation of einkorn dominance it seems worthwhile to explore this question further.
Potential weeds as indicators for agricultural practices
Storage finds of crops with associated weeds are almost absent from Bandkeramik sites. In the data from both regions no relationship is evident between certain species and the different crops. How can we know if a taxon found in ‘normal’ samples was a weed or not? To answer this question we need chorological and ecological data for each potential weed species (Kreuz 1990, pp 143ff.).
The central European flora can be divided into two groups: there are plants which came into a region without anthropogenic influence and help. These are the indigenous species called idiochores (Idiochoren; for the terms see Schroeder 1969, 1974). Some of them are able to establish themselves as ruderals or weeds in the fields. Such potential weeds are called apophytes (Apophyten).
Secondly there are species which could only come to a region and persist by direct or indirect human influence. These species, which do not grow in natural stands, are called anthropochores according to their method of dispersal (Anthropochoren). We can expect that all anthropochores of the Neolithic were brought in with seeds or by other means during the colonization of the landscapes. That is why they might represent weeds. In addition in the Neolithic they almost certainly indicate a movement of people (see below).
Bandkeramik potential weed species and their characteristics (following Oberdorfer 1990, 2001 and Kästner et al. 2001) and numbers of plant remains. APO apophyte, ANT anthropochore; height: low 0–40 cm, medium 50–80 cm, high >80 cm; life-form: s summer-annual, w winter-annual/biennial, per perennial, / indifferent; area: euras eurasiatic, including no-euras, euras-smed, euras(...); eurassubozean eurasiatic-suboceanic, including subatl subatlantic; kont kontinental; med et al. mediterranean, including med-smed, smed-med, med(...); smed submediterranean; omed eastern mediterranean, including eastern submediterranean
Among the apophytes there are no real woodland species. This is probably due to the fact that woodland species are not adapted to grow as weeds under the “steppe-like” (Van Zeist 1987) and regularly disturbed conditions of a field (shrubs and trees are excluded from the data set presented here). On the contrary, almost all Bandkeramik apophytes today normally grow in the floodplains of river valleys. Possibly they were brought to the fields with the dung of cattle which grazed the floodplains as well as the harvested fields.
Possibly we have recorded here a change in the harvesting technique (Kreuz in press a). Following Hillman (1981) and Reynolds (1985, 1993, p 189), many more and also lower- as well as medium-growing weed plants and their seeds are collected by sickle- than by ear-harvesting. There is archaeological evidence that such a change in harvest technique took place. Flint working techniques and the quantities of lithic artefacts changed markedly between Bandkeramik phase I and the following phases: in earliest Bandkeramik settlements the percentage of sickle blades is lower and less standardized (de Grooth 2003, p 402; Gronenborn 1997, p 102 and pers. comm.; Kind 1997, p 140). Possibly, sickles were less in use for harvest during Bandkeramik phase I than in the later Bandkeramik phases.
In contrast many fewer low growing species were found at the contemporary late Bulgarian Neolithic sites (Fig. 7). That is why we cannot exclude the possibility that ear harvesting remained important to the Bulgarian Neolithic farmers. Ear-plucking is an efficient method for harvesting the hulled wheats when fully ripe, as it prevents more ears from falling to the ground. It is for example written in the Bible that ear-plucking was practised by the disciples of Jesus (the gospel according to St. Mark, mk.02, 23–28).
On the other hand, experimental harvesting has revealed that harvesting with flint sickles is three times faster than ear-plucking (L. Peña-Chocarro, Como (I), and L. Zapata-Peña, Vittoria-Gasteiz (E) personal communication, see also Ibánez et al. 2001). The point in question is whether the use of this technological innovation was of different importance to the Bandkeramik compared with the Bulgarian Neolithic farmers.
There exist different opinions concerning the intensity of field management and the time of sowing during the Neolithic; their discussion goes beyond the scope of this paper (Bogaard 2004, chapter 7; Bakels and Rousselle 1985, p 55; Behre and Jacomet 1991, p 86; Willerding 1980; Lüning 2000). Important information derives again from the characteristics of the potential weeds.
In both cultures, perennials form an essential part of the weed assemblage. Therefore we should ask whether at least parts of the fields were not cultivated intensively. Characteristics of vegetative propagation and dispersal of the species concerned indicate whether they are able to reproduce, for example from their rhizomes or runners. Such species could possibly survive under intensive soil treatment by hoeing or similar activities. Actually, some of the Bandkeramik perennial species as for example Agrostis capillaris/stolonifera, Carex muricata, Galium verum, Lotus uliginosus or Plantago lanceolata are able to propagate vegetatively, but others are not. To decide whether their occurrence in the samples really is a sign of the intensity of field management (Bogaard 2004; Jones et al. 1999), further research is needed.
Geographical origins of the potential weeds
Bulgarian Neolithic potential weed species and their characteristics and numbers of plant remains (for explanations see Table 7)
We can state that most Bandkeramik potential weeds and many of the Bulgarian Neolithic weeds were introduced with seeds into the fields and settlement sites from elsewhere. There is again no visible chronological development concerning the species of the early and late Bulgarian Neolithic sites. In contrast there is a strong increase of Mediterranean species during the Bandkeramik phases III–V (Fig. 12). More introduced species per site—especially Mediterranean ones—occur from Bandkeramik phase II onwards (without figure).
In this context Papaver somniferum, the western Mediterranean opium poppy, (see above) and possibly Vicia faba, the Mediterranean Celtic bean (Buxó 2004), which have been found at later Bandkeramik sites (Fig. 13), have to be remembered. These archaeobotanical phenomena still have to be connected with an archaeological counterpart to explain these new external influences. New waves of immigrants in the second half of the Bandkeramik period could be a possible explanation (Kreuz in press a).
The agricultural system of the Bulgarian Neolithic culture is based on 10 crop species (Fig. 13). Naked wheat and barley are higher yielding than emmer and einkorn, which are the only Bandkeramik cereals. However, they need more nitrogen than the hulled wheats, and naked wheat needs more careful weeding (Körber-Grohne 1988; Kreuz in press b). The characteristics of these additional cereals therefore have important consequences for a cultivation system. The additional pulses of the Bulgarian Neolithic, Cicer arietinum, Lathyrus sativus/cicera and Vicia ervilia also have special demands of cultivation.
Different dietary customs can be derived from the different crops: barley and naked wheat have a different taste and other requirements of processing and preparation than emmer and einkorn. The number of pulse species in the Bulgarian Neolithic might indicate that they were a certain substitute for meat in the diet.
The earliest Bandkeramik cultivation system with only five species was different from the neighbouring Starčevo, Körös, Čris cultures and the Bulgarian Neolithic. We could speculate whether manpower was a problem for Bandkeramik groups, which later even led to a time-saving change in lithic technology and harvest technique (see above). On the other hand there might also have been other priorities concerning parts of the agricultural system.
The weed assemblages let us assume that ear-plucking remained an important harvesting method throughout the Bulgarian Neolithic. At least some of the cereals were grown as winter crops there. In contrast the sowing time of the Bandkeramik farmers—as indicated by the weeds—points to (at least mainly, more likely exclusively) summer crop cultivation. In this case the fields could be grazed after harvest until the next spring. This would have been useful for the farmers if they put more emphasis on stock breeding.
Comparing the two cultivation systems, the Bulgarian Neolithic system seems more time-consuming than that of the Bandkeramik. This has important social implications. Social differences between the Bulgarian Neolithic and Bandkeramik are also indicated by different settlement structures and house types (see above).
To summarize, we can state that an important agricultural and social change of early Neolithic subsistence occurred somewhere in the transition area from eastern to western Hungary. It took several centuries until cultivation of barley and naked wheat crossed the rivers Tisza and Danube. In western central Europe this was the beginning of the middle Neolithic with the rising of the Großgartach and the Rössen Cultures, including—among others—settlement structures, technologies and ritual practices different from the Bandkeramik ones (Eisenhauer 1999; Lüning 2000, p 16ff.). The background of this delayed adaptation of two important cereals—Hordeum sp. (barley) and Triticum aestivum s.l./durum/turgidum (naked wheat)—in western central Europe is not yet understood. Future archaeobiological and archaeological work—especially in the key area of Hungary - is therefore urgently needed.
Our warmest thanks to F. Bittmann, A. Bogaard, K. Borojevič, D. Filipovič, F. Gyulai, M. Kohler-Schneider, D. Kučan, E. Margaritis, R. Neef, T. Popova and S.M. Valamoti for their unpublished data from Austria, Bulgaria, Greece, Hungary, former Yugoslavia and Turkish Thrace. We thank the Deutsche Forschungsgemeinschaft for funding the project “Archaeobotanical investigations of Early, Middle and Late Neolithic landscape development and agricultural systems in Hesse, Germany, and adjacent areas” (KR1569/2-4), the referees S. Jacomet and G. Jones for helpful remarks, M. Elibal, A. Posluschny and H. von Schlieben for the help with the scans, maps and photographs and James Greig for improving our English