Introduction

State of research and aims of the analysis

Most archaeobotanical work focuses on the importance of cultivated plants. However, it is becoming increasingly clear that the use of wild plants made an indispensable contribution to subsistence, especially in the early phases of the sedentary way of life. In recent years, wild plant resources and their role within the diet of Neolithic communities have been the subject of many publications (Jacomet 2009; Colledge and Conolly 2014; Antolín et al. 2016, 2021). In the lakeshore settlements of the Alpine Foreland, excellent conditions for the preservation of organic material prevail, resulting in the recovery of large quantities of uncharred plant macro remains. The abundance and variety of edible wild plant remains soon led to research regarding their contribution to the diet. In wetland sites, waterlogged plant remains are considered to be more or less reliable indicators of consumed plant resources (Antolín et al. 2020). For that reason and early on in the study of consumption patterns in the Alpine Foreland, attempts to model the caloric importance of individual food sources in the Neolithic diet were made (Jacomet and Schibler 1985; Jacomet et al. 1989; Gross et al. 1990; Ebersbach 2002; Antolín et al. 2016). These models of calculation are repeatedly updated as the dataset is growing. Several Neolithic lakeshore settlements in Switzerland have been sampled extensively, covering most of the inhabited areas. Results from their archaeobotanical analyses are thus fairly representative for the whole settlement [e.g. Arbon Bleiche 3 (Hosch and Jacomet 2004) and Zürich Parkhaus Opéra (Antolín et al. 2020)]; the estimate of the caloric importance of their wild plant resources suggest that they were a much more important part of the daily diet than has been thought.

The present case study aims to evaluate the importance of wild plant resources of a Late Neolithic settlement located on Lake Neuchâtel in Switzerland. Analyses of seeds and fruits are combined with analyses of both macro and micro remains of dung pellets. The integration of both data sets should allow a better understanding of the exploitation of the landscape surrounding the settlement. In order to reconstruct the plant diet of the ancient inhabitants of Grandson during the four settlement phases, the calorific value of a selection of food plants found in the archaeological layers is considered. The aim is to investigate the relative contribution of the collected wild plants versus the cultivated plants to the diet. With the analysis of dung an insight into livestock management, and complementary information on the appearance of the habitats walked by the animals and humans in the vicinity of the settlement, will be provided.

Site Grandson-Corcelettes, Les Pins

The Late Neolithic lakeshore settlement Grandson-Corcelettes, Les Pins (canton of Vaud, Switzerland; UNESCO World Heritage Site; Fig. 1) is located on Lake Neuchâtel (430 m a.s.l.). In 2017 a rescue excavation uncovered archaeological layers preserved at a depth of more than two metres (Burri-Wyser et al. 2019). Due to their location beneath the groundwater table, the preservation of these layers was excellent. Moreover, archaeological material, especially organic, is exceptionally well preserved and plentiful. The rescue excavation comprised a square 10 m2 in area as well as a narrow trench approximately 76 m long and 1.5 m wide, which was dug in ten sections. The site is dendrochronologically dated from 3009 to 2602 bc (excluding the Hallstatt occupation around 552−548 bc). The different stratigraphic assemblages are attributed to the Horgen, Lüscherz ancien, Lüscherz recent and Auvernier-Cordé cultures. The complexity of the sedimentation, the large quantity of wooden piles, the richness of the artefacts as well as the harsh weather conditions did not make excavation and sampling easy.

Fig. 1
figure 1

Grandson-Corcelettes, Les Pins. Geographical location

Full size image

Materials and methods

Sample analysis and evaluation

The present study deals with the results of the archaeobotanical analyses of 21 sediment samples that stem from four Neolithic occupation layers (Table 1). The sediment was processed using the wash-over technique (Hosch and Zibulski 2003; Jacomet 2013; Steiner et al. 2015). Sieves of mesh sizes 4, 2 and 0.35 mm were used. In total 51.1 L of soil were processed; individual sample size mostly exceeded 2 L, which allows for a good representation of large remains (Jacomet 2013).

Table 1 Overview of the archaeobotanically analysed samples
Full size table

Analysis of the organic fractions was done using a stereo microscope (magnification 7.5×–60×). Seeds, fruits and other biological remains (including dung pellets) were sorted and recorded using the guidelines suggested by Antolín et al. (2017a). The larger fractions (4 and 2 mm) were sorted entirely, from the smallest fraction (0.35 mm) only a subsample was analysed due to its richness in plant macro remains. Identification of plant remains was done using a modern seed reference collection. Plant species nomenclature is based on the 2017 checklist of the National Center for Swiss Flora Data and Information (Juillerat et al. 2017). Results were entered into the ArboDat database program (Kreuz and Schäfer 2002). In order to assess the importance of wild plant resources within the diet, the calorific intake of a selection of food plants (both wild and cultivated) was considered. We defined this selection of food plants as follows: it includes all cultivated plants (except for Camelina sativa, as its status as a cultivated plant is not 100% certain), all large seeded edible wild fruits as well as two smaller seeded wild berries. All of our selected plant taxa have been the subject of earlier calorie intake calculations in the Swiss Neolithic lakeshore settlements. It is clear that not all wild plant resources are included in this analysis. Plant species that could be used as vegetables, salads, spices or for medicinal purposes are usually harvested before the seeds mature. Moreover, those parts of the plants used for consumption (leaves and roots) are generally not preserved in archaeological deposits. To calculate calorie intake, we used several quantitative approaches. First, recorded units of quantification were converted into fruit units (Table 2). This conversion is necessary due to differences in seed production between species [e.g. a rosehip contains 25 seeds, while a strawberry contains 87 seeds (Brombacher 2019)]. Then, weight and calorific intake data were calculated for each species using existing formulae (see Jacomet and Schibler 1985; Jacomet et al. 1989; Hosch and Jacomet 2004; Antolín et al. 2016, 2020). Starting values for weight and calorific intake calculations are based on waterlogged food plant concentrations only.

Table 2 General results of the main food plant taxa recorded in the four different occupation phases in Grandson in average density of the minimum number of individuals (MNI), with additional information on seed/fruit weight, calories per gram, as well as the resulting overall calories per taxon
Full size table

Dung pellet analysis

Sheep/goat dung pellets and potential manure fragments were selected from three of the studied sediment samples and examined for plant macro remains (diaspores, epiderms) and micro remains (pollen, spores, non-pollen palynomorphs). In total 20 dung pellets and four manure fragments were studied (the small scope of the analysis was due to budget restrictions). Samples from two occupation phases (Horgen and Lüscherz recent) were considered. Plant macro remains and epiderms were extracted from the dung pellets. The former were determined using a stereo microscope (magnification 6.3×-40×), the latter with the help of a microscope (magnification 100×-400×). Diaspores were recorded fully quantitatively, epiderms semi-quantitatively. Plant taxa are grouped following the same procedure as stated above.

After the extraction of the plant macro remains, the remainder of the samples were prepared for pollen analysis following standard palynological methods (Stockmarr 1971; Moore et al. 1991). For the identification of pollen grains, the keys of Beug (2004) and other illustrations (Reille 1992) were used. The pollen diagram shows percentage values of pollen and spores.

A detailed description of the methods used for dung pellet analysis can be found in various publications (e.g. Akeret and Jacomet 1997; Kühn and Hadorn 2004; Kühn and Wick 2010, 2021).

Results

General results of the archaeobotanical analysis

In total 31,838 seeds and fruits were sorted and identified; this corresponds to an extrapolated value of 214,875 remains, 93.4% of these being preserved in waterlogged form, 6.6% in charred form. The average density of waterlogged remains in the individual samples varied between 1,317.5 and 7,317.7 items per litre (Table 1). The high concentration of seeds and fruits, as well as the presence of fragile remains such as cereal chaff remains, indicate the good preservation of plant macro remains (Brombacher and Jacomet 1997). This is also proven by the presence of whole leaves, waterlogged and charred whole fruits, fragments of charred cereal ears, fungi and coprolites. The analysis of seeds and fruits yielded a total of 134 different taxa. In the following we will briefly mention the cultivated plants, and then put the emphasis on the wild plant spectrum.

Cultivated plant spectrum

Cultivated plants represent 34.5% of the identified plant macro remains. Mainly cereals and oil/fibre plants are represented. Cereals include Hordeum vulgare var. nudum, Triticum monococcum, T. dicoccum and T. aestivum/durum/turgidum as well as “new-glume wheat” or T. timopheevi and possibly T. cf. spelta. Cereals are present as charred grain and waterlogged and charred chaff. H. vulgare var. nudum and T. dicoccum are the dominant cereals. Oil and fibre plants are the most numerous among the cultivated plants. Three species were identified: Linum usitatissimum, Papaver somniferum and Camelina sativa. Charred and waterlogged seeds were recorded for each species; from L. usitatissimum both waterlogged and charred capsule fragments were also found. Pulses are rare, mainly charred seeds but also a few waterlogged pod fragments from Pisum sativum are recorded. Two potentially cultivated plants are identified: Anethum graveolens and Apium graveolens. Both are not indigenous in Switzerland, but have been recorded at several Neolithic lakeshore sites (Jacomet 1988, 2006, 2007).

Wild plant spectrum

More than 65% of the identified plant remains are from wild plants, most of which could have grown in the near vicinity of the settlement. Aquatic plants, representing a small part of the total number of wild plant remains (7.6%), are found throughout the whole stratigraphy but especially in the lowest and uppermost layers. The species detected are characteristic of fresh, clear, stagnant or very slow-moving water (e.g. Najas marina and Chara sp.) and are possibly the result of periodic flooding. The riparian vegetation is not much represented (2.1% of the wild plant spectrum). The weed spectrum, dominated by weeds of summer cereals/annual ruderals, constitutes 27.4% of the total number of wild plant remains. The ratio of weeds of summer crops to weeds of winter cereal is about 9:1. The majority of the weeds identified are species characteristic of nutrient-rich soils (e.g. Chenopodium album and Stellaria media). In some areas the soil may have been slightly acidic, which could explain the presence of a few hundred fruits of Aphanes arvensis. The ruderal perennials could have originated from the village itself, however this seems unlikely given the wet soil conditions. It seems more plausible that they came from the outskirts of the village or from fields that had not been cultivated for several years. Meadow and pasture plants account for an average of 15% of the wild plant spectrum. This is remarkable, as it is generally accepted that grassland does not develop until the late Bronze Age (e.g. Kühn and Heitz-Weniger 2015). Some species possibly stem from fallow fields, but others could have flourished in the open forests of the Jura slopes, where this type of vegetation still occurs today. Plant macro remains of woodland vegetation represent 47.8% of the wild plants (28 species). The majority of these originate from drier areas in the hinterland of the settlement, only a few species originate from the riparian forests along the lake (e.g. Alnus). The greater part of the remains (over 65%) stems from foraging plants, among them fruits, berries, hazelnuts and acorns. The most regularly found species with high nutritional value include Corylus avellana, Malus sylvestris, Fragaria vesca, Prunus spinosa, Rubus fruticosus and Quercus. All of those yielded not only uncharred but also charred remains; noteworthy were the findings of charred M. sylvestris halves and the abundance of charred Quercus cotyledons halves. Also Abies alba needles are present in all samples, they constitute a large proportion of the woodland vegetation and are attested in waterlogged, charred and semi-charred form. Another interesting finding concerns the large number of Picea abies needles in some of the samples. P. abies naturally only grows at higher altitudes. Modern occurrences of the species at lower altitudes are due to plantations. P. abies originally colonised the Jura heights from Savoy around 4300 bc (Burga and Perret 1998; Burga and Hussendörfer 2001).

Reconstruction of the diet according to the calorific contribution of food plants

In order to verify the significance of wild plant resources within the diet, the calorific intake of both the crops and the edible wild plants (as defined above) was calculated and presented in Table 2. These calculations are based on uncharred plant material only. It should be emphasized that the plant macro remains, which form the basis of the calculations, represent minimal values (Hosch and Jacomet 2004), as many processes influence the preservation of plant macro remains in archaeological layers (Jacomet and Kreuz 1999).

In the present study, the small number of samples from the Horgen and Auvernier-Cordé phases must be considered, and their representativeness with respect to human consumption must be treated with caution. What’s more, only a small part of the inhabited area was excavated. The results are therefore only valid locally and are probably not representative for the whole settlement. Yet, we have a good and large sample set in which large-seeded wild fruits are recorded in a representative way.

Considering the calorific contribution of food plants in the different phases of occupation of the site, we can observe the following trends (Fig. 2). During the Horgen period (3009−2976 bc), wild resources account for about 62.3% of the total calorific intake, while domestic resources account for 37.7%. The latter are mainly composed of T. dicoccum and T. monococcum, while Malus sylvestris and Prunus spinosa provide the majority of calories for the wild resources. In the Lüscherz ancient (2937−2844 bc), wild resources account for about 90% of the total calorific intake (mainly Quercus and Malus), while domestic resources account for 10%. In the Lüscherz récent (2841−2714 bc), we have a similar pattern: wild resources account for about 93% of the total calorific intake, while domestic resources account for 7%. More than half of the calorific intake is provided by Quercus, while Malus, Corylus and P. spinosa provide between 10 and 16%. During the Auvernier-Cordé (2707−2681 bc), a change in calorific intake is observed with wild resources decreasing to about 80% of the total intake and domestic resources increasing to 20%. As in the Horgen period, cereals and L. usitatissimum gain in importance; the wild resources comprise mainly Quercus as well as Corylus and Malus.

Fig. 2
figure 2

Percentage of calories obtained from the main food plants in the four different occupation phases studied. The x-axis represents the settlement phase, the y-axis represents the cumulated percentages of calorie intake per plant species

Full size image

Results of the analyses of dung pellets: macro remains and pollen

The most reliable results on the management of domestic animals are obtained by examining plant remains in sheep/goat dung (Kühn et al. 2013; Marinova et al. 2013) and can give complementary information on the appearance of habitats exploited by the animals and humans. The analysis of the dung pellets and ‘manure’ fragments yielded a diverse spectrum of plant species (Fig. 3, ESM 1 and 2).

Fig. 3
figure 3

Percentage composition of feed based on pollen. ★ supplemented by macro remains

Full size image

Small pollen percentages and few finds of epiderms of forest trees, but higher amounts of pollen, diaspores and/or spikes of heliophilous trees and shrubs (such as Betula, Alnus, Corylus, Rubus, Rosa, Lonicera and Cornus sanguinea) point to an open, strongly exploited landscape. This hypothesis is supported by the wide spectrum of non-arboreal pollen (dominated by Poaceae) as well as the evidence of e.g. Daucus carota, Ranunculus, Origanum vulgare fruits and Pteridophyta sporangiae.

Indications suggesting browsing on fallow fields are obtained from a set of dung pellets rich in weed and cereal pollen as well as seeds/fruits of weedy and cultivated plants. The latter might also have been used as feed. Grazing near the lakeshore in wet habitats is shown by the presence of Filipendula, Succisa, Gentiana, Alisma, Chara and other wetland plants. Our results suggest that sheep and goats were mainly kept in the surroundings of the settlement. It is clear that the animals were inside the settlement during certain periods of the year, as they left their dung behind. The feeding of dried leaves of deciduous trees (Quercus, Tilia, Fagus) and fresh twigs of evergreens (Abies) by humans is probable, but not clearly documented. Whether or not the animals were fed by humans in addition to their stay outdoors cannot be conclusively determined. Wheat and linseed may have been fed to them as well as eaten during their stay in fallow fields. In about one third of the samples, eggs of the intestinal parasites Trichuris and Dicrocoelium dendriticum were recorded.

As the dung pellets and potential manure fragments don’t show striking differences in plant content, it can be assumed that the manure fragments are flattened dung and not a mixture of dung and fodder/litter. From our sample set, there are no clear chronological differences between the Horgen and Lüscherz recent phases.

Discussion and conclusions

Wild food plant resources

Alongside cultivated food plants, wild gathered plants formed a large part of the plant diet in the Neolithic period. Their important role has already been described several times (Jacomet et al. 1989; Maier 2001; Hosch and Jacomet 2004). So far, attempts to model the human diet based on the calorific intake of food plants are rare and restricted to lakeshore settlements in the Alpine Foreland. This is mainly due to the conditions of preservation which are better than average in these settlements, and to the extensive sampling strategies of their well-preserved occupation layers. As indicated by Antolín et al. (2016), an accurate estimate of the contribution of wild plant resources to the diet can only be obtained when the total proportion of calories represented by all plant remains in a given layer is known. Neolithic sites which meet the above criteria and for which calculations exist, include Arbon Bleiche 3, layer 13 of Zürich-Parkhaus Opera and Zug-Riedmatt. Hosch and Jacomet (2004) calculated that wild plants could have accounted for up to 40% of the plant diet at the site Arbon Bleiche 3 (3384−3370 bc). In this reconstruction, only lipid rich species, namely Quercus and Corylus, are considered. Antolín et al. (2020) estimated that wild plant resources—included are M. sylvestris, Quercus, Corylus, P. spinosa, Rosa, Fragaria and Rubus idaeus/fruticosus—account for about 57% of the total plant calorific input at the site of Zürich-Parkhaus Opera (Layer 13) (3176−3153 bc). Steiner et al. (2022) equally registered high caloric input of wild plant resources (included are Malus/Pyrus, Quercus, Corylus, P. spinosa, Rosa, Fragaria, Fagus sylvatica, Physalis alkekengi, R. idaeus and R. fruticosus), but observed differences between the chronological units of Zug-Riedmatt (3200−3100 cal bc); this possibly represents a change in the role of gathered food plants throughout the settlement phases. In Grandson, the values obtained for the calorific intake through wild plant resources vary between 62.3% and 93% of the total calorific intake, which is even higher than in the previously mentioned sites. However there are a few issues to consider. In comparison to Arbon Bleiche 3, layer 13 of Zürich-Parkhaus Opera and Zug-Riedmatt, the calculations of Grandson are based on a much smaller number of samples; the excavation covered only a very restricted part of the settlement and a preservation bias towards certain food plants must be kept in mind. We are aware that only a small selection of edible wild plants were included in our analysis, and that many other plants that were collected but did not leave traces, such as leafy greens and roots/tubers, certainly also played an important role in the daily diet. The preservation bias affects the cereal remains as well. Uncharred cereal grain is not preserved; of H. vulgare var. nudum almost exclusively charred remains are preserved and thus not considered here; of T. dicoccum/monococcum chaff is mainly preserved as waterlogged remains and thus maybe overestimated in relation to H. vulgare. Similarly, the role of Pisum sativum is probably not correctly estimated either, as waterlogged pod fragments have not been systematically recorded.

The purpose of our calculations was ultimately to investigate the role of wild-collected plants in the plant diet and to do so in a more progressive way than just comparing the number of remains. We however come to the same conclusion as previous research that wild plants are much more important within the diet than we thought so far and their role in the diet of early agricultural communities is often underestimated, with too much emphasis on cultivated plants (Antolín et al. 2020). Further investigation into the role of wild plants should always be part of future archaeobotanical studies.

The large quantities of wild plant remains not only highlight their importance in the diet, but also give us information on their preparation, consumption and preservation. Some wild plants were gathered for immediate consumption (fleshy berries etc.), while others were collected for longer term storage. In Grandson, the presence of charred apple halves may indicate preparation for preservation. It is thought that wild apples were cut in half and dried near a fire, to make them palatable and durable (Jacomet et al. 1989). Still other wild plants had to be processed to make them palatable for human consumption, for example acorns. To get rid of the bitter component (tannin), acorns were boiled or roasted. Many Quercus remains in Grandson are charred halves of cotyledons in a shelled state. Notwithstanding the laborious process prior to human consumption, it is unlikely that the acorn findings were intended solely for domestic animal feed. Acorn flour has been used as a staple food since Mesolithic times, and the nutritional value of oak fruit is similar to that of grain; because of its higher fat content, acorns are even higher in calories than cereals (Schneider 1990). Like hazelnuts and wild apples, acorns can be harvested in large quantities with little effort. They were probably also used in times of food shortage (e.g. Ayerdi et al. 2016).

Exploitation of the natural environment

The remains of aquatic plants in archaeological deposits are primarily deposited naturally, whereas terrestrial plants are assumed to be the consequence of human activities (e.g. Antolín et al. 2017b). Thus, the archaeobotanical assemblage reflects the way in which the inhabitants exploited the plant resources around the lakeshore village, their range of action. The topographical context with the different exploitable zones is crucial to understand this range of action (Fig. 4). From the archaeobotanical and dung pellet analysis, we have indications that the lakeside plateau was exploited for various activities such as crop cultivation and wild fruit gathering. In the immediate hinterland of the site, the soils developed on glacial moraines are very favourable for agriculture (Source: https://map.geo.admin.ch). Despite the fact that the site is located on a lake shore, there appears to have been almost no exploitation of wetland plants by the inhabitants. Domestic animals foraged in different habitats from the riparian zone to the slopes of the Jura mountains, where more food gathering could have taken place (Fig. 4). The presence of Picea needles indicates that the territory of exploitation extended up to the heights of the Jura mountains. The reason why the inhabitants visited the higher ridges of the Jura mountains is still unknown. Was it for hunting or to exploit the grazing areas with livestock, as observed around Lake Constance (Kühn and Hadorn 2004)? The vertical mobility of the herds has certainly several advantages. First, the pastures near the lake/settlement are preserved for the winter; in the Neolithic period, the cattle grazed mainly in the forest as shown by various studies (Akeret and Jacomet 1997; Akeret et al. 1999; Kühn and Hadorn 2004). Second, the use of remote pastures kept livestock away from fields where they can cause great damage. Yet, it is hardly possible that the Picea needles reached the site as animal dung as their preservation is very good and no signs of chewing or digestion are observed. Picea branches could have been collected for various reasons (fodder, bedding), but the effort seems to be considerable when there are suitable sources closer to the lake, especially during the cold season. The transport of the needles in the fur of animals would be another possibility.

Fig. 4
figure 4

Reconstruction of the immediate surroundings of the site with human and animal exploitation zones

Full size image

From our findings, we can conclude that the territory of exploitation of the Grandson settlers extended from the riparian zone up to the heights of the Jura mountains; the landscape must have been relatively open and the animals were mostly kept close to the settlement area.