Nutritive value of winter-collected annual twigs of main European woody species,mistletoe and ivy and its possible consequences for winter foddering of livestock in prehistory

Abstract

In Central Europe, forest grazing was probably the main way of providing winter feeding for livestock in prehistory; the main components of this diet most likely included annual twigs of woody species (Acer platanoides, Betula pendula, Carpinus betulus, Corylus avellana, Fagus sylvatica, Fraxinus excelsior, Picea abies, Pinus sylvestris, Populus tremula, Quercus robur, Salix caprea, Tilia cordata and Ulmus glabra), Hedera helix, Viscum album and senescent grassland biomass; however, their nutritive value has never been compared. Biomass samples were collected in the Czech Republic in February 2013 and analysed for concentration of nitrogen, phosphorus, potassium, calcium, magnesium, fibre and lignin. With the help of a recent analogy from the Altai Mountains (Russia), possible consequences of winter grazing on the development of forest vegetation and cattle breeds in prehistory were assessed. Hedera and Viscum were the best forage available in winter, and their nutritive value, according to the highest N and lowest fibre and lignin concentrations, was higher than the nutritive value of all winter-collected annual twigs of woody species. Nutritive value of annual twigs of all woody species was very low, especially compared with the quality of meadow hay, leaf-fodder or senescent steppe grassland biomass. Therefore, Hedera and Viscum might have been preferentially grazed by livestock and collected by ancient farmers for the supplementary feeding of privileged animals. According to recent analogy, annual twigs of woody species were browsed by livestock. Their insufficient quality could be one of the reasons for the low body size of cattle recorded in prehistory. The most realistic model of livestock breeding in prehistory seems to be year-round grazing, opening forests and causing deforestation in the surroundings of prehistoric settlements.

Keywords

agricultural history forage quality livestock feeding prehistory Viscum album year-round forest grazing

Introduction

It is generally accepted that the livestock of prehistoric farmers grazed on pastures, fallows, stubbles and in forests during the vegetation season, but it is still not clear how livestock was fed during the winter time. In Central and Northern Europe, winter time is generally the bottleneck for livestock breeding as the carrying capacity of the landscape is substantially lower in winter than during the vegetation season. This problem was at least partly solved by transhumance in some mountain regions – seasonal movement of people with their livestock between summer and winter pastures (Moe et al., 2007). According to many authors, the shortage of fodder during the winter time was solved by feeding livestock with leaf-fodder from the main broad-leaved woody species, harvested and conserved by drying during the vegetation season (Dreslerová, 2012; Rasmussen, 1989). Livestock was probably also partly fed by winter-collected twigs (Haas et al., 1998). Leaf-foddering should thus have played the main role in winter feeding of livestock in Europe from the Neolithic and at least up to the La Tène period when the first scythes enabled cutting of grasslands and hay making (Hejcman et al., 2013). Therefore, hay making step by step replaced leaf-fodder harvesting, although leaf-fodder was used as a supplement to winter hay feeding in some regions up to the present (Austad and Hauge, 2006; Halstead et al., 1998; Slotte, 2001).

Although generally accepted, the theory of leaf-foddering in the Neolithic is based on relatively weak archaeological evidences of several woody species recorded in coprolites of goats, sheep and cattle discovered in Switzerland, Germany (Akeret and Jacomet, 1997; Akeret and Rentzel, 2001; Akeret et al., 1999; Rasmussen, 1993) and in France (Delhon et al., 2008). However, sheep and particularly goats browse frequently on woody species even if they have non-limited access to high-quality pasture forage (Kühn and Hadorn, 2004; Papachristou and Platis, 2011; Pokorná et al., 2013). Leaves of woody species in coprolites of sheep and goats are thus not direct evidence for harvesting of leaf-fodder, and analysis of coprolites must be supplemented by analysis of charcoals and twigs on archaeological localities (Regnell, 2002). The second argument for large-scale leaf-foddering was the decline of elm (Ulmus sp.) in different periods of agricultural prehistory, detected according to a marked fall of its pollen in pollen diagrams in many countries (Andersen and Rasmussen, 1993; Garbett, 1981; Heybroek, 1963; Peglar and Birks, 1993; Troels-Smith, 1960). Elm is a species with a high forage value of leaves in comparison with other woody species (Hejcmanová et al., 2013), and its decline was interpreted by some authors as resulting from Dutch disease transferred by the beetle Scolytus scolytus, although it could also be triggered by large-scale leaf-foddering together with climate and soil changes (Girling and Greig, 1985; Moe and Rackham, 1992; Parker et al., 2002; Peglar, 1993; Troels-Smith, 1960).

The importance of large-scale leaf-foddering for winter feeding of livestock since the Neolithic can thus be deduced particularly from (sub) recent analogies. The question which remains open is thus whether cattle, sheep and, since the Bronze ages, also horses were mainly fed by leaf-fodder during the winter time or whether they were able to obtain enough fodder by winter grazing. Unique information about a possible solution, in a European context, can be found in the memoirs of the farmer FJ Vavák, who managed his small peasant farm in the Elbe lowland (190–210 m a.s.l.) 50 km east of Prague in the last third of the 18th century (Klír, 2008). In moderate winters, livestock (cattle, sheep, goats and pigs) were grazed year-round on communal pastures and in forests without housing. Storage of conserved forage for winter use was insufficient, and areas of pasture substantially exceeded areas of hay meadow. Year-round livestock grazing without any or with a very limited amount of supplementary feeding was thus most probably the dominant means of livestock breeding in Central Europe from the Neolithic up to the 18th century. Year-round cattle grazing was also documented by medieval written resources from the Czech Republic (Szabó and Hédl, 2013). According to Zimmermann (1999), replacement of year-round livestock grazing on pastures and in forests with livestock housing and winter feeding was the result of an increase in human population density and the necessity of using all natural resources with maximum effectiveness. This is also the reason why the first cow houses were used after the Bronze Age in Western Europe (Netherlands and NW Germany, in particular), where human density was substantially higher than in the Czech Republic where livestock was not definitively closed into barns until the 19th century (Petrášek, 1972).

This leads us to question which plant species could be eaten by livestock during winter forest grazing and ask how this winter forage could fulfil the nutritive requirements of livestock. In the lowlands of Central Europe where the climate borders the oceanic and continental, winter grass biomass is generally senescent (yellow). There are, however, some winter green understory species such as ivy (Hedera helix) and hemiparasitic species such as mistletoe (Viscum album), which were, according to several authors (Akeret and Rentzel, 2001; Bottema, 2001; Deforce et al., 2013; Kühn and Hadorn, 2004; Kühn et al., 2013; Troels-Smith, 1960), also used for winter feeding of livestock. Thus the main role in winter feeding of livestock was probably played by annual twigs of woody species since the Neolithic up to the 18th century, but their nutritive value (syn. forage quality) has never been studied. Nutritive value of forage can be evaluated according to the concentration of macro-nutrients, N, P, K, Ca and Mg, particularly their ratios, Ca:P and K:(Ca + Mg) and by fibre content (cellulose, hemi-cellulose and lignin), which predetermine organic matter digestibility (Hejcman et al., 2006, 2010; Pavlů et al., 2006).

The aim of this study was thus (1) to determine the nutritive value of winter-collected annual twigs of the main woody species (Acer platanoides, Betula pendula, Carpinus betulus, Corylus avellana, Fagus sylvatica, Fraxinus excelsior, Picea abies, Pinus sylvestris, Populus tremula, Quercus robur, Salix caprea, Tilia cordata and Ulmus glabra), leaves of mistletoe (V. album) and ivy (H. helix), and to compare them with senescent winter steppe grassland biomass and meadow hay, and (2) to assess the possible consequences of the nutritive value of twigs for winter feeding of livestock, development of forest vegetation and cattle breeds in Central Europe since the Neolithic.

Materials and methods

Biomass sampling

For the analysis, we selected 15 woody species common in Central Europe at least since the Neolithic (5600 bc). We collected the annual twigs with buds of selected woody species and leaf biomass of ivy and mistletoe from at least three individuals of each species on each of four selected localities during February 2013. Selected localities were broad-leaved forests and their margins in the Bohemia, western part of the Czech Republic: (1) Kozly (50°15′N, 14°32′E, 167 m a.s.l.), (2) Měšice (50°11′N, 14°31′E, 203 m a.s.l.), (3) Kletečná (50°34′N, 13°59′E, 472 m a.s.l.) and (4) Mšec (50°12′N, 13°52′E, 435 m a.s.l.). Mean annual precipitations and temperatures ranged in sampling sites from 614 to 723 mm and from 8–9°C, respectively. We collected in total 60 (15 species × four site replicates) annual twigs and winter leaf biomass samples, which were then oven-dried at 60°C for 48 h and ground to powder.

Annual twigs and winter leaf chemical properties

In twig and winter leaf samples, the concentration of macro-elements (N, P, K, Ca and Mg) and the content of residual ash (ash − (P + K + Ca + Mg)) neutral detergent fibre (NDF) and acid detergent fibre (ADF), and acid detergent lignin (ADL) were determined. NDF represents cellulose, hemi-cellulose and lignin together, and ADF represents cellulose and lignin. The N concentration in the plant samples was determined using an automated analyser TruSpec (LECO Corporation, USA) by combustion with oxygen in an oven at 950°C. Combustion products were mixed with oxygen, and the mixture passed through an infrared detector of CO₂ and by circuit for the aliquot ratio where carbon is measured as CO₂. Gases in the aliquot circuit were transferred into helium as a carrying gas, conducted through hot copper and converted to N.

Biomass samples were burnt in a microwave oven at a temperature of 550°C and weighted in order to determine ash content. Biomass samples were mineralized using aqua regia, and P, K, Ca and Mg concentrations were then determined in the solution using ICP–OES (Varian VistaPro, Mulgrave, VIC, Australia). Residual ash containing mostly Si was calculated as the ash content minus the sum of P, K, Ca and Mg concentrations. NDF, ADF and ADL contents were determined by standard methods of Association of Official Analytical Chemists (AOAC, 1984).

All analyses were performed in an accredited national laboratory, Ekolab Žamberk (http://www.ekolab.zamberk.cz). N:P, Ca:P and K:(Ca + Mg) ratios were calculated from determined concentrations.

Data analyses

Data tested by the Kolmogorov–Smirnov test of normality met assumptions for the use of parametric tests. One-way ANOVA followed by post hoc comparison using the Tukey’s multiple range tests in Statistica 9.0 program (StatSoft, Tulsa, USA) were used to identify significant differences in concentrations of nutrients and NDF, ADF and ADL contents among species.

Unconstrained principal component analysis (PCA) in Canoco for Windows 4.5 program (ter Braak and Šmilauer, 2002) was used to analyse the relationships among annual twig and winter leaf chemical properties and the similarity of the 60 samples. Data were log-transformed before the analysis. The results of the PCA analysis were visualized in the form of an ordination diagram constructed by the CanoDraw program (ter Braak and Šmilauer, 2002).

Results

Calculated by one-way ANOVA, the effect of species was significant on all determined chemical properties of collected winter biomass. Concentration of N ranged from 12.4 g/kg in Fraxinus to 21.1 g/kg in Viscum; concentration of P ranged from 1.1 g/kg in Picea to 2.4 g/kg in Viscum; concentration of K ranged from 4.8 g/kg in Picea to 13.8 g/kg in Viscum; concentration of Ca ranged from 7.4 g/kg in Picea to 21 g/kg in Populus; and finally, concentration of Mg ranged from 0.9 g/kg in Picea to 2.2 g/kg in Hedera (Table 1). Values of N:P ratio ranged from 7.8 in Salix to 12 in Carpinus; values of Ca:P ratio ranged from 3.8 in Viscum to 14.7 in Populus; and finally, values of K:(Ca + Mg) ranged from 0.14 in Tilia to 0.37 in Fraxinus (Table 1). Content of NDF ranged from 431 g/kg in Viscum to 596 g/kg in Tilia; content of ADF ranged from 318 g/kg in Hedera to 533 g/kg in Carpinus; and finally, content of ADL ranged from 112 g/kg in Hedera to 266 g/kg in Betula. Content of residual ash ranged from 15.6 g/kg in Pinus to 51.2 g/kg in Corylus (Table 2).

Table 1.

Concentration (means ± standard error of mean) of N, P, K, Ca, Mg and N:P, Ca:P and K:(Mg + Ca) ratios in leaf-fodder of studied species. Calculated by one-way ANOVA, differences among species for all chemical properties were significant (p < 0.01). Using Tukey’s post hoc comparison test, species with the same letter were not significantly different (p < 0.05). Chemical properties of good-quality meadow hay follow Hejcman et al. (2010, 2012), Hrevušová et al. (2009) and Tallowin and Jefferson (1999), and optimum range for cattle follows Kudrna (1998) and Whitehead (1995).

Species	N (g/kg)	P (g/kg)	K(g/kg)	Ca (g/kg)	Mg (g/kg)	N:P ratio	Ca:P ratio	K:(Mg + Ca)
Acer platanoides	15.6 ± 1.4^ab	1.8 ± 0.2^ab	8.0 ± 1.1^bc	14.7 ± 1.3^ab	1.5 ± 0.2^ab	8.7 ± 0.6^ab	8.5 ± 1.5^abc	0.25 ± 0.05^a
Betula pendula	13.9 ± 0.4^a	1.6 ± 0.1^ab	5.1 ± 0.6^ab	12.3 ± 1.8^ab	1.3 ± 0.3^ab	8.8 ± 0.5^ab	7.7 ± 1.0^ab	0.19 ± 0.02^a
Carpinus betulus	14.7 ± 0.8^a	1.3 ± 0.1^a	3.9 ± 0.4^a	11.1 ± 1.3^a	1.2 ± 0.1^a	12.0 ± 1.2^b	8.9 ± 1.0^abc	0.16 ± 0.01^a
Corylus avellana	17.9 ± 2.1^ab	1.8 ± 0.2^ab	8.5 ± 0.5^bc	10.9 ± 1.3^a	1.8 ± 0.2^ab	10.0 ± 0.9^ab	6.3 ± 1.0^ab	0.32 ± 0.02^a
Fagus sylvatica	14.4 ± 1.1^a	1.3 ± 0.1^a	4.6 ± 0.7^ab	12.2 ± 2.1^ab	1.0 ± 0.1^a	11.7 ± 1.2^ab	9.6 ± 1.1^abc	0.17 ± 0.01^a
Fraxinus excelsior	12.4 ± 0.7^a	1.5 ± 0.1^ab	8.5 ± 0.7^bc	9.9 ± 0.6^a	1.2 ± 0.1^a	8.5 ± 0.6^ab	6.8 ± 0.7^ab	0.37 ± 0.04^a
Hedera helix	16.9 ± 1.0^ab	1.4 ± 0.05^a	9.5 ± 1.1^c	14.7 ± 2.7^ab	2.2 ± 0.3^b	11.9 ± 1.0^b	10.3 ± 0.9^bc	0.27 ± 0.03^a
Picea abies	13.5 ± 1.1^a	1.1 ± 0.1^a	4.8 ± 0.3^ab	7.4 ± 0.7^a	0.9 ± 0.1^a	12.0 ± 0.6^b	6.5 ± 0.3^ab	0.28 ± 0.04^a
Pinus sylvestris	15.7 ± 1.4^ab	1.5 ± 0.1^ab	5.2 ± 0.6^ab	7.8 ± 1.0^a	1.1 ± 0.04^a	10.3 ± 0.6^ab	5.3 ± 1.0^ab	0.29 ± 0.06^a
Populus tremula	12.5 ± 0.7^a	1.5 ± 0.2^ab	6.6 ± 0.8^abc	21.0 ± 2.5^b	1.5 ± 0.2^ab	8.7 ± 0.8^ab	14.7 ± 2.1^d	0.15 ± 0.02^a
Quercus robur	14.0 ± 0.3^a	1.2 ± 0.1^a	5.1 ± 0.7^ab	9.7 ± 0.8^a	1.3 ± 0.2^ab	11.8 ± 0.7^ab	8.1 ± 0.3^abc	0.22 ± 0.02^a
Salix caprea	15.9 ± 0.6^ab	2.1 ± 0.2^ab	5.5 ± 0.5^abc	16.2 ± 1.7^ab	1.4 ± 0.2^ab	7.8 ± 0.5^a	8.2 ± 1.7^abc	0.16 ± 0.02^a
Tilia cordata	15.2 ± 1.2^ab	1.9 ± 0.2^ab	6.1 ± 0.8^abc	20.7 ± 3.0^b	1.5 ± 0.3^ab	8.3 ± 0.7^ab	11.4 ± 1.8^bc	0.14 ± 0.03^a
Ulmus glabra	14.6 ± 0.6^a	1.5 ± 0.05^ab	5.9 ± 0.3^abc	14.7 ± 0.3^ab	1.4 ± 0.1^ab	9.7 ± 0.1^ab	9.8 ± 0.6^abc	0.18 ± 0.01^a
Viscum album	21.1 ± 2.3^b	2.4 ± 0.4^b	13.8 ± 1.2^d	8.5 ± 1.2^a	1.7 ± 0.2^ab	8.2 ± 1.0^ab	3.8 ± 0.4^a	0.68 ± 0.11^b
Meadow hay	20.0–28.7	2.7–3.7	24.1–34.0	6.0–8.4	1.5–4	5–10	1.7–2.5	–
Steppe grass winter biomass	17.9	2	3.7	6.6	1.2	9	3.3	0.22
Optimum range	19.2–25.6	2.3–3.7	5–10	2.9–5.8	1.5–3.5	5–10	1–2	1–2.2

Table 2.

Concentration (means ± standard error of mean) of neutral detergent fibre (NDF), acid detergent fibre (ADF), acid detergent lignin (ADL) and residual ash in leaf-fodder of studied species. Calculated by one-way ANOVA, differences among species for all chemical properties were significant (p < 0.01). Using Tukey’s post hoc comparison test, species with the same letter were not significantly different (p < 0.05). Chemical properties of good-quality meadow hay follow Worrell et al. (1986) and Isselstein et al. (2007) and optimum range for cattle follows Kudrna (1998) and Whitehead (1995).

Species	NDF (g/kg)	ADF (g/kg)	ADL (g/kg)	Res. ash(g/kg)
Acer platanoides	577 ± 14^cd	514 ± 9^cd	225 ± 14^def	25.6 ± 2.6^a
Betula pendula	600 ± 24^cd	525 ± 9^d	266 ± 24^f	23.3 ± 2.9^a
Carpinus betulus	642 ± 27^d	533 ± 12^d	206 ± 10^bcdef	25.3 ± 2.9^ab
Corylus avellana	551 ± 58^bcd	510 ± 45^bcd	255 ± 20^f	51.2 ± 14.8^b
Fagus sylvatica	632 ± 31^d	547 ± 20^d	261 ± 17^f	34.5 ± 5.1^ab
Fraxinus excelsior	587 ± 31^cd	455 ± 28^bcd	144 ± 15^ab	29.0 ± 7.6^ab
Hedera helix	392 ± 9^a	318 ± 10^a	112 ± 5^a	35.6 ± 2.8^ab
Picea abies	498 ± 35^abcd	413 ± 19^abc	166 ± 5^abcde	20.2 ± 1.6^a
Pinus sylvestris	512 ± 12^abcd	416 ± 11^ab	163 ± 7^abcd	15.6 ± 1.3^a
Populus tremula	577 ± 38^cd	491 ± 27^bcd	226 ± 20^def	38.4 ± 5.7^ab
Quercus robur	582 ± 29^cd	495 ± 17^bcd	238 ± 8^ef	20.8 ± 0.1^a
Salix caprea	488 ± 10^abc	455 ± 14^bcd	207 ± 18^bcdef	29.6 ± 2.8^ab
Tilia cordata	596 ± 23^cd	495 ± 25^bcd	215 ± 9^cdef	33.8 ± 5.4^ab
Ulmus glabra	456 ± 6^abc	408 ± 6^ab	213 ± 6^bcdef	37.3 ± 0.3^ab
Viscum album	431 ± 13^ab	327 ± 12^a	150 ± 7^abc	27.0 ± 2.8^ab
Meadow hay	500–680	340	40
Steppe grass winter biomass	724	433	89	76
Optimum range	330–450	190–300	Max. 80

Results of the PCA analysis are presented in the form of an ordination diagram (Figure 1). The first ordination axis explained 39%, the first two axes together 61% and the first four axes together 83% variability of biomass chemical data. The first axis of the diagram can be attributed particularly to concentrations of N, P, K, Mg, NDL, ADF and lignin. The second axis can be attributed particularly to concentration of Ca in the biomass. Concentrations of N, P, K and Mg were positively correlated to each other, as vectors for all these elements were directed into the same area of the diagram and were negatively related to NDL, ADF and lignin concentrations. The highest concentrations of N, P, K and Mg were recorded in Viscum, followed by Hedera and the lowest in Carpinus and Fagus in which, on the contrary, the highest concentrations of NDL, ADF and lignin were recorded. The highest concentrations of Ca and Ca:P ratio were recorded in Populus, followed by Tilia, Ulmus and Salix. The lowest concentrations of Ca were recorded in Fraxinus, followed by Pinus and Picea.

Figure 1.

Ordination diagram showing results of principal component analysis (PCA) of relationships among chemical properties of winter biomass of studied species; N, P, K, Ca and Mg concentrations; N:P, Ca:P and K:(Ca + Mg) ratios; NDF (neutral detergent fibre); ADF (acid detergent fibre); and lignin (ADL).

Discussion

Nutritive value of annual twigs and winter leaves

The main message of our study is that species with winter green leaves such as Hedera and Viscum belong to the best forage, which might have been available in forests during the winter. Their nutritive value, according to the highest N and lowest ADF and lignin concentrations, was higher than the nutritive value of all winter-collected annual twigs of woody species. Similar high nutritive value can be recorded in winter leaves of Rubus fruticosus (Verheyden-Tixier et al., 2008), which was also among plant species recorded in Neolithic coprolites of sheep/goat in the Grande Rivoire rock shelter in France (Martin, 2011). The ancient farmers thus probably collected Hedera and Viscum during the winter time intentionally and used them for the feeding of animals (Haas, 2004; Nicod et al., 2008), because they are the richest source of N and P in winter. Viscum was collected, although the amount of its biomass was relatively small in forests in comparison with the biomass of woody species, and its collection was laborious.

Therefore, we suggest that Viscum was probably used as a supplement for feeding of privileged animals such as lactating and pregnant cows or goats with the highest N and P requirements. According to our personal experience, Viscum is excellently eaten by goats without any detrimental effects on their health (Figure 2). Macro-remains of Viscum discovered at archaeological localities (see Akeret and Rentzel, 2001; Deforce et al., 2013; Kühn and Hadorn, 2004; Kühn et al., 2013; Pokorný et al., 2006) could thus represent remains of winter-collected Viscum used as ‘an extra forage’ for selected categories of livestock. Nowadays, the high nutritive value of Viscum is also well known to hunters in Austria who still use the biomass of Viscum to attract deer to particular places during the winter (G Glatzer, personal communication). In addition, stems and leaves of Viscum are fragile, and they frequently fall down from trees after strong winds. Therefore, Viscum could be eaten by livestock even without any intentional collection. Similar to Viscum, Hedera has higher nutritive value than annual twigs of woody species, and this is why it was probably selectively grazed by livestock in winter. In the experiment by Van Uytvanck and Hoffman (2009), for example, Hedera completely disappeared from forest managed for several years by year-round cattle grazing. Hedera was thus probably scarce or completely missing in prehistoric and medieval forests managed by winter livestock grazing, similarly as other winter green understory species such as Vinca minor or Rubus sp. In grazed forests, Hedera could survive only as liana on trees, but not as ground-covering species. In addition, it is highly probable that old flowering shoots of Hedera were intentionally collected and used as green winter fodder for livestock by prehistoric farmers. A decrease in the pollen production of Hedera in different periods (see Bottema, 2001; Iversen, 1944; Troels-Smith, 1960) could thus be considered as an indicator of human activities in forests connected with livestock breeding.

Figure 2.

(a) Pinus sylvestris is sometimes used by Czech farmers as supplementary winter fodder to improve health status of goats. (b) According to our feeding experiments, Viscum album is very well eaten by goats during the winter. (c) Ruminating young cattle in the evening on their way home from winter pastures in the Altai Mountains. All photographs taken by Pavla and Michal Hejcman in early March 2013.

Nutritive value of all winter-collected annual twigs of woody species was substantially lower than the nutritive value of meadow hay (Table 1) and also spring-collected leaf-fodder of broad-leaved woody species (see Hejcmanová et al. (2013); for a comparison with other plant species, see also Verheyden-Tixier et al. (2008)). Low nutritive value of twigs was given by insufficient concentrations of N and P and also partly by insufficient concentration of Mg for livestock nutrition. On the other hand, twigs were characterized by too high concentrations of Ca, ADF and lignin, and also by too high Ca:P ratio and too low K:(Ca + Mg) (tetanic) ratio. The next characteristic aspect of the chemical composition of twigs was relatively small differences in the concentration of P among individual species in comparison with their leaves in which differences were substantially higher (Hejcmanová et al., 2013). Twigs of Fagus and Carpinus had absolutely the worst nutritive value because of their highest lignin content, which is in accordance with their leaves (Hejcmanová et al., 2013). Leaf-fodder of Ulmus and Tilia were the best of all woody species in Central Europe, and this is also consistent with the nutritive value of their twigs, which was also comparable with Salix. Senescent grassland biomass collected in winter on steppe grassland was of better nutritive value than twigs of woody species because of higher concentrations of N and P, lower concentrations of Ca and lignin and also a lower Ca:P ratio in senescent grassland biomass than in twigs. This indicates that livestock first of all grazed senescent grassland biomass and then, if no other alternatives were available, started to browse trees. A problem of senescent grassland biomass can be high fungus infection and therefore contamination by mycotoxins, particularly in moderate (‘warm’) winters (Skládanka et al., 2011). In some years, senescent grassland biomass was thus probably less consumed by livestock as its consumption could cause health problems. In addition, in summer-grazed forests and grasslands, the amount of senescent grass biomass was low in winter and probably insufficient for livestock feeding.

Despite the relatively low nutritive value of annual twigs in comparison with meadow hay and leaf-fodder, twigs are regularly browsed by different deer species, hares and European bison, and serve as winter forage if no other better alternative is available (Ammer, 1996; Kowalczyk et al., 2011). In addition, winter-collected annual twigs of woody species are frequently used as a supplementary feed to improve the health status of livestock (Figure 2).

Winter feeding of livestock and its consequences for cattle breeds

Despite low nutritive value, annual twigs of many woody species probably played a crucial role in the winter feeding of livestock in many regions since the Neolithic, and this resulted in the development of wood-pasture habitats in Central Europe (Bergmeier et al., 2010). We suggest that the majority of livestock was probably year-round grazed without any or with only a limited amount of supplementary feeding during the winter and that they were thus subjected to insufficient nutrition conditions. Indirect evidence for very bad livestock feeding during the winter time in the past can be a reduction in body size during the process of domestication which is, for instance, up to half in cattle in comparison with aurochs, their ancestor (Petrášek, 1972; Zeder, 2006). One of the reasons consists in the fact that cattle in the surroundings of villages had much more limited access to winter forage than free-ranging animals. According to our personal experience, heifers with insufficient access to high-quality forage over the winter do not reach their full body size. Although evidence for winter feeding of livestock by leaf-fodder has been recorded in archaeological layers, the amount of harvested leaf-fodder was probably not sufficient to cover the nutritional requirements of all livestock (Rasmussen, 1990; Thiébault, 2005). This idea is also supported by written records from the 18th century describing insufficient feeding of livestock during the winter and the importance of forests for livestock winter feeding and survival (Klír, 2008; Petrášek, 1972). In the Czech Republic, for example, a herd of wild cattle survived in the forests of the Doupov Mountains military area in the 1990s without any supplementary feeding and regularly reproduced each year. Their main winter forage was grassland senescent biomass, annual twigs of woody species and bark. The herd was finally hunted for veterinary reasons (farmers’ fear of the spread of diseases) and because of their negative effects on forest regeneration (Z Macháček, personal communication). In Kraansvlak in the Netherlands, European bison, Highland cattle and Konik horses are bred in deciduous forests without any supplementary winter feeding (Kemp and Cromsigt, 2012). Further recent analogies for the system of no winter feeding can be found in the Altai Mountains where sheep, cattle and horses are still bred with no or only limited supplementary winter feeding. As we recorded, livestock, forced by hunger, freely walked out of the village in the mornings to obtain forage in their surroundings. According to our observations and those of local farmers, sheep and cattle were able to explore an area up to approximately 7 km out of the village and horses approximately two times farther. Simultaneous breeding of sheep, cattle and horses thus enables the best spatial use and partitioning of forage resources in the surroundings of the village. As we recorded, the most explored were the steppe south exposed grasslands where senescent grassland biomass was grazed and alluvial forests where annual twigs of Salix sp. and Populus sp. were intensively browsed. Woody species formed bonsai-like shrubs and alluvial forests were very open (Figure 3). In the evenings, livestock slowly returned home (Figure 3), where they were controlled by farmers and sometimes also partly fed by small amounts of hay or other supplementary fodder. The system of no supplementary winter feeding was very efficient in landscapes with low human density as it required almost no labour and the ancient farmer was able to keep relatively high numbers of livestock. The disadvantages were no or rather negative live weight gain of livestock over the winter, the small size of adult animals, delayed maturity of young animals and the very high pasture area required. A well-known adaptation of adult cattle on this system is the fluctuating of live weight between the summer, when they accumulate fat, and winter, when they metabolize energy from fat reserves gained in summer and fall, similarly as deer in the northern environments (DelGiudice et al., 1992). This adaptation is well known from old (primitive) breeds of cattle such as Highland, Galloway, Czech Reddish or Salers. As we learned in the Altai Mountains, the system of no winter feeding can also be realized with modern European cattle breeds such as Hereford, Simental or Aberdeen-Angus (Figures 2 and 3). Doubts of recent authors (see Kreuz, 2008) about the ability of Neolithic cattle to browse twigs of woody species and their ability to survive winter in Central Europe without any supplementary feeding are therefore unjustified. Winter grazing of livestock in forests could thus play the crucial role in their opening in prehistory.

Figure 3.

Winter livestock grazing is still practised in the Altai Mountains (South Siberia, Russia): (a) Cattle grazed senescent biomass on alluvial grasslands or browsed annual twigs of Salix sp. shrubs in the foreground, (b) Professor Vilém Pavlů in bonsai-like alluvial shrub land (Salix sp. and Populus sp.) maintained by winter browsing of annual twigs by cattle and horses and (c) Alluvial Salix sp. forest heavily affected by the winter browsing of cattle and horses.

We suppose that the system of no or limited winter supplementary feeding of livestock was also used in lowland regions of Central Europe from the Neolithic at least up to the Iron Age when the human population density increased, but that to some extent, this system survived up to the 18th century. As the nutritive value of annual twigs of woody species is insufficient for cattle nutrition, browsing of twigs helped livestock to survive the winter but did not enable an increase in live weight or the stimulation of intensive milk production. We suggest therefore that the main selection criteria for prehistoric cattle and other livestock were thus the ability to survive the winter with no or limited supplementary feeding on forage of low nutritional value.

Year-round livestock grazing and its consequences for development of forest vegetation

In Central Europe, the maximum extent of forest occurred during the Atlantic Period (c. 8000–4500 cal. bc in Central Europe (Chytrý, 2012; Ložek, 2007) and 7350–3900 bc in Northern Germany (Dörfler et al., 2012) with higher temperature and precipitations in comparison with the present. Subsequent vegetation changes were ascribed, in addition to climatic changes, to human activities related especially to deforestation and agricultural activities. Marked vegetation changes consisted, for instance, of Ulmus decline (Parker et al., 2002), spread of Carpinus and Fagus (Chytrý, 2012; Herbig and Sirocko, 2013; Pokorný, 2005), and the emergence of various pasture indicators such as Trifolium sp., Plantago lanceolata and grasses whose pollen was recorded in different sediments (Brun, 2011; Poschlod and Baumann, 2010). In two basic models of deforestation because of early agriculture, Iversen (1941) suggested clearance of forests by fire with the aim of producing pastures in woodland as the source of food for cattle, whereas Troels-Smith (1953) formulated a model which places livestock breeding to an earlier time, based on intensive leaf-foddering of animals in byres, particularly by Ulmus. In his model, Troels-Smith assumed the absence of pastures, animal confinement in settlements and limited area for the free ranging of animals according to the lack of pasture indicator plant species in pollen diagrams. At the same time, both authors debated the decline of Hedera pollen in sediments at around 5000 BP. Iversen (1944) suggested it as a climate change indicator, whereas Troels-Smith (1960) proposed that Hedera declined as a consequence of foddering to livestock.

Livestock keeping could have begun, however, even earlier than indicated by both authors, especially by leaving animals to free range in the forested surroundings of human settlements. Wood pastures were common in the past (see, for instance, Jamrichová et al. (2013) and Rösch (2012)), because Central European broad-leaved and mixed forests brim over with forage resources for domestic animals during the vegetation season. Based on our results, we can conclude that even during the winter time, available annual twigs of woody species and especially evergreen Hedera and Viscum provided sufficient nutrient supply for livestock to survive the winter. This idea is supported also by analyses of prehistoric cattle, goat and sheep dung from wetland settlements in the borderland between Germany and Switzerland in which macro remains of woody species were frequently recorded (Kühn et al., 2013). Based on our experience of winter grazing and the ability of livestock to browse annual twigs of woody species, the most realistic model of livestock breeding in prehistory seems to be year-round livestock grazing. Based on this model, forests exposed to intensive winter grazing and browsing by livestock were open with many bonsai-like shrubs and dead debarked trees in surrounding of settlements. Our conclusions are partly in agreement with the theory of Vera (2000) that the species composition and succession of vegetation was governed by large herbivores, and the Central and Western lowlands were covered by a park-like landscape in areas with adequately high density of megaherbivores. There were probably no strict borders between forests and grasslands in vicinity of settlements, and forests were full of grassland species, although this can hardly be detected by pollen analysis as intensive grazing would have prevented the flowering and pollen production of many grassland species. During Neolithic times, the effects of livestock grazing on forest were limited to the vicinity of settlements, therefore settlement density is a very important factor to evaluate the impact of livestock grazing on forests. Additionally, forests grazed by livestock do not necessarily develop a classical meadow park-like landscape. According to our personal experience from the Altai Mountains, open areas are mainly found in wet environments around water courses where Carex sp., Salix sp. and other wetland species are dominating. Livestock grazing had probably important effect on forests, but numbers of livestock in the landscape had been still low during the Neolithic compared with later times, and the grazing effect was thus not strong enough to enable development of large-scale grasslands. This may explain why large-scale grasslands are not visible in hundreds of pollen diagrams from the Neolithic (see Kreuz, 2008). Park-like landscape can develop if the density of livestock in forests is adequately high, and well-detectable, large-scale effects of livestock grazing on vegetation is mainly evidenced from the Bronze or Iron Ages onwards (Pokorný, 2005). In the discussion of livestock forest grazing, the time period, the settlement density, the cultural developments and density of browsing animal must be taken into account. Based on our experience, we believe that substantial part of deforestation in the surrounding of prehistoric settlements could be directly ascribed to year-round livestock grazing.

Conclusion

Based on the higher nutritive value of winter-collected Hedera and Viscum than winter-collected annual twigs of woody species, we suppose that both species could have been collected by prehistoric farmers and used for supplementary feeding of privileged animals. In addition, Hedera was probably preferentially grazed by livestock during the winter up to its total disappearance in forests. Prehistoric farmers could also feed livestock directly in forests by driving herds to intentionally cut trees or their branches with Hedera or Viscum. The presence of Hedera pollen in sediments could thus indicate forests which were not used for winter feeding of livestock, but validation of this conclusion requires further research. Winter-collected annual twigs of all woody species were characterized by very low nutritive value, much lower than the quality of meadow hay, leaf-fodder or senescent steppe grassland biomass. After grazing of senescent grassland biomass, annual twigs of woody species were probably browsed by livestock. Insufficient winter nutrition could thus explain the low body size of cattle recorded since the Neolithic up to the 18th century. Year-round livestock grazing practised by ancient farmers seems to be the key driver for the formation of open forest with many bonsai-like shrubs.

Footnotes

Funding

The study was funded by the Czech University of Life Sciences, Prague (projects CIGA 20114205 and IGA 20134268), and by the Czech Science Foundation (project GACR P505/12/1390).

References

Akeret

Jacomet

(1997) Analysis of plant macrofossils in goat/sheep faeces from the Neolithic lake shore settlement of Horgen Scheller – An indication of prehistoric transhumance? Vegetation History and Archaeobotany 6: 235–239.

Akeret

Rentzel

(2001) Micromorphology and plant macrofossil analysis of cattle dung from the Neolithic lake shore settlement of Arbon Bleiche 3. Geoarchaeology 16: 687–700.

Akeret

Haas

Leuzinger

, et al. (1999) Plant macrofossils and pollen in goat/sheep faeces from the Neolithic lake-shore settlement Arbon Bleiche 3, Switzerland. The Holocene 9: 175–182.

Ammer

(1996) Impact of ungulates on structure and dynamics of natural regeneration of mixed mountain forests in the Bavarian Alps. Forest Ecology and Management 88: 43–53.

Andersen

Rasmussen

(1993) Radiocarbon wiggle-dating of elm declines in northwest Denmark and their significance. Vegetation History and Archaeobotany 2: 125–135.

Association of Official Analytical Chemists (AOAC) (1984) Official Methods of Analysis. 14th edition. Washington, DC: Association of Official Analytical Chemists.

Austad

Hauge

(2006) Pollarding in western Norway. In: 1er Colloque Europeen sur les trognes, Vendome, 26–28 October.

Bergmeier

Petermann

Schröder

(2010) Geobotanical survey of wood-pasture habitats in Europe: Diversity, threats and conservation. Biodiversity and Conservation 19: 2995–3014.

Bottema

(2001) A note on the pollen representation of ivy (Hedera helix L.). Review of Palaeobotany and Palynology 117: 159–166.

10.

Brun

(2011) Anthropogenic indicators in pollen diagrams in eastern France: A critical review. Vegetation History and Archaeobotany 20: 135–142.

11.

Chytrý

(2012) Vegetation of the Czech Republic: Diversity, ecology, history and dynamics. Preslia 84: 427–504.

12.

Deforce

Bastiaens

Van Neer

, et al. (2013) Wood charcoal and seeds as indicators for animal husbandry in a wetland site during the late Mesolithic–early Neolithic transition period (Swifterbant culture, c. 4600–4000 bc) in NW Belgium. Vegetation History and Archaeobotany 22: 51–60.

13.

DelGiudice

Mech

Kunkel

, et al. (1992) Seasonal patterns of weight, hematology, and serum characteristics of free-ranging female white-tailed deer in Minnesota. Canadian Journal of Zoology 70: 974–983.

14.

Delhon

Martin

Argant

, et al. (2008) Shepherds and plants in the Alps: Multi-proxy archaeobotanical analysis of Neolithic dung from ‘La Grande Rivoire’ (Isère, France). Journal of Archaeological Science 35: 2937–2952.

15.

Dörfler

Feeser

van den Bogaard

, et al. (2012) A high-quality annually laminated sequence from Lake Belau, Northern Germany: Revised chronology and its implications for palynological and tephrochronological studies. The Holocene 22: 1413–1426.

16.

Dreslerová

(2012) Forest in the prehistoric landscape II. Archeologické Rozhledy 64: 199–236.

17.

Garbett

(1981) The elm decline: The depletion of a resource. New Phytologist 88: 573–585.

18.

Girling

Greig

JRA

(1985) A first fossil record for Scolytus scolytus (F.) (elm bark beetle): Its occurrence in elm decline deposits from London and the implications for Neolithic elm disease. Journal of Archaeological Science 12: 347–351.

19.

Haas

(2004) Mikroskopische Analyse von Schaaf- und Ziegenkoprolithen. In: Jacomet

Leuzinger

Schibler

(eds) Die jungsteinzeitliche Seeufersiedlung Arbon Bleiche 3: Umwelt und Wirtschaft. Archäologie im Thurgau, vol. 12. Frauenfeld: Kanton Thurgau, pp. 351–357.

20.

Haas

Karg

Rasmussen

(1998) Beech leaves and twigs used as a winter fodder: Examples from historic and prehistoric times. Environmental Archaeology 1: 81–86.

21.

Halstead

Tierney

Butler

, et al. (1998) Leafy hay: An etnoarchaeological study in NW Greece. Environmental Archaeology 1: 71–80.

22.

Hejcman

Hejcmanová

Pavlů

, et al. (2013) Origin and history of grasslands in Central Europe – A review. Grass and Forage Science 68: 345–363.

23.

Hejcman

Pavlů

Hejcmanová

, et al. (2006) Farmer decision making and its effect on the subalpine grassland succession in the Giant Mountains, Czech Republic. Acta Societatis Botanicorum Poloniae 75: 165–174.

24.

Hejcman

Strnad

Hejcmanová

, et al. (2012) Response of plant species composition, biomass production and biomass chemical properties to high N, P and K application rates in Dactylis glomerata- and Festuca arundinacea-dominated grassland. Grass and Forage Science 67: 488–506.

25.

Hejcman

Szaková

Schellberg

, et al. (2010) The Rengen Grassland Experiment: Relationship between soil and biomass chemical properties, amount of elements applied, and their uptake. Plant and Soil 333: 163–179.

26.

Hejcmanová

Stejskalová

Hejcman

(2013) Forage quality of leaf-fodder from main broad-leaved woody species and its possible consequences for the Holocene development of forest vegetation in Central Europe. Vegetation History and Archaeobotany. Epub ahead of print 5 September. DOI: 10.1007/s00334-013-0414-2.

27.

Herbig

Sirocko

(2013) Palaeobotanical evidence for agricultural activities in the Eifel region during the Holocene: Plant macro-remain and pollen analyses from sediments of three maar lakes in the Quaternary Westeifel Volcanic Field (Germany, Rheinland-Pfalz). Vegetation History and Archaeobotany 22: 447–462.

28.

Heybroek

(1963) Diseases and looping for fodder as possible causes of a prehistoric decline of Ulmus. Acta Botanica Neerlandica 12: 1–11.

29.

Hrevušová

Hejcman

Pavlů

, et al. (2009) Long-term dynamics of biomass production, soil chemical properties and plant species composition of alluvial grassland after the cessation of fertilizer application in the Czech Republic. Agriculture, Ecosystems & Environment 130: 123–130.

30.

Isselstein

Griffith

Pradel

, et al. (2007) Effects of livestock breed and grazing intensity on biodiversity and production in grazing systems I: Nutritive value of herbage and livestock performance. Grass and Forage Science 62: 145–158.

31.

Iversen

(1941) Landnam in Danmarks Stenalder: En pollenanalytisk undersøgelse over det første landbrugs indvirking paa vegetationsudviklingen. Danmarks Geologiske Undersogelse II, Række 66. Copenhagen: Reitzel.

32.

Iversen

(1944) Viscum, Hedera and Ilex as climate indicators: A contribution to the study of the post-glacial temperature climate. Geologiska Föreningen i Stockholm Förhandlingar 66: 463–483.

33.

Jamrichová

Szabó

Hédl

, et al. (2013) Continuity and change in the vegetation of a Central European oakwood. The Holocene 23: 46–56.

34.

Kemp

Cromsigt

(2012) Vijf jaar Wisenten in het Kraansvlak 2007–2012. Velserbroek: PWN Waterleidingbedrijf Noord-Holland, p. 40.

35.

Klír

(2008) Osídlení zemědělsky marginálních půd v mladším středověku a raném novověku. In: Klápště

Měřínský

(eds) Dissertationes Archaeologicae Brunenses/Pragensesque, vol. 8. Praha: Filozofická fakulta Univerzity Karlovy v Praze, pp. 1–222.

36.

Kowalczyk

Taberlet

Coissac

, et al. (2011) Influence of management practices on large herbivore diet – Case study of European bison in Białowieża Primeval Forest (Poland). Forest Ecology and Management 261: 821–828.

37.

Kreuz

(2008) Closed forest or open woodland as natural vegetation in the surroundings of Linearbandkeramik settlements? Vegetation History and Archaeobotany 17: 51–64.

38.

Kudrna

(ed.) (1998) Produkce krmiv a výživa skotu. Prague: Agrospoj.

39.

Kühn

Hadorn

(2004) Pflanzliche Makro- und Mikroreste aus Dung von Wiederkäuern. In: Jacomet

Leuzinger

Schibler

(eds) Die jungsteinzeitliche Seeufersiedlung Arbon Bleiche 3: Umwelt und Wirtschaft. Archäologie im Thurgau, vol. 12. Frauenfeld: Huber, pp. 327–348.

40.

Kühn

Maier

Herbig

, et al. (2013) Methods for the examination of cattle, sheep and goat dung in prehistoric wetland settlements with examples of the sites Alleshausen-Täschenwiesen and Alleshausen-Grundwiesen (around cal 2900 bc) at Lake Federsee, south-west Germany. Environmental Archaeology 18: 43–57.

41.

Ložek

(2007) Zrcadlo minulosti – česká a slovenská krajina v kvartéru. Prague: Dokořán.

42.

Martin

(2011) Une bergerie néolithique dans le Vercors: analyse des macro-restes végétaux des ‘fumiers’ fossiles de la Grande Rivoire (Isère, France). In: Wiethold

(ed.) Carpologia. Articles réunis à la mémoire de Karen Lundström-Baudais. Actes des rencontres d’Archéobotanique, Glux-en-Glenne, Bibracte 20, Centre archéologique européen, pp. 27–38.

43.

Moe

Rackham

(1992) Pollarding and a possible explanation of the Neolithic elmfall. Vegetation History and Archaeobotany 1: 63–68.

44.

Moe

Fedele

Maude

, et al. (2007) Vegetational changes and human presence in the low-alpine and subalpine zone in Val Febbraro, upper Valle di Spluga (Italian central Alps), from the Neolithic to the Roman period. Vegetation History and Archaeobotany 16: 431–451.

45.

Nicod

P-Y

Picavet

Argant

, et al. (2008) La bergerie néolithique de la Grande Rivoire. In: Jospin

J-P

Favrie

(eds) Premiers bergers des Alpes: de la préhistoire à l’Antiquité. Gollion: Infolio, pp. 74–79.

46.

Papachristou

Platis

(2011) Intake by goats browsing kermes oak alone or choices of different browse combinations: Implications for Mediterranean grazing systems. The Rangeland Journal 33: 221–227.

47.

Parker

Goudie

Anderson

, et al. (2002) A review of the mid-Holocene elm decline in the British Isles. Progress in Physical Geography 26: 1–45.

48.

Pavlů

Hejcman

Pavlů

, et al. (2006) Effect of continuous grazing on forage quality, quantity and animal performance. Agriculture Ecosystems and Environment 113: 349–355.

49.

Peglar

(1993) The mid-Holocene Ulmus decline at Diss Mere, Norfolk, UK: A year-by-year pollen stratigraphy from annual laminations. The Holocene 3: 1–13.

50.

Peglar

Birks

HJB

(1993) The mid-Holocene Ulmus fall at Diss Mere, South-East England – Disease and human impact? Vegetation History and Archaeobotany 2: 61–68.

51.

Petrášek

(1972) Geneze chovu skotu v českých zemích. Prague: Academia.

52.

Pokorná

Hejcmanová

Hejcman

, et al. (2013) Activity time budget patterns of sheep and goats co-grazing on semi-natural species-rich dry grassland. Czech Journal of Animal Science 58: 208–216.

53.

Pokorný

(2005) Role of man in the development of Holocene vegetation in Central Bohemia. Preslia 77: 113–128.

54.

Pokorný

Boenke

Chytráček

, et al. (2006) Insight into the environment of a pre-Roman Iron Age hillfort at Vladař, Czech Republic, using a multi-proxy approach. Vegetation History and Archaeobotany 15: 419–433.

55.

Poschlod

Baumann

(2010) The historical dynamics of calcareous grasslands in the central and southern Franconian Jurassic mountains: A comparative pedoantracological and pollen analytical study. The Holocene 20: 13–23.

56.

Rasmussen

(1989) Leaf foddering of livestock in the Neolithic: Archaeobotanical evidence from Weier, Switzerland. Danish Journal of Archaeology 8: 51–71.

57.

Rasmussen

(1990) Leaf foddering in the earliest Neolithic agriculture: Evidence from Switzerland and Denmark. Acta Archaeologica 60: 71–86.

58.

Rasmussen

(1993) Analysis of goat/sheep faeces from Egolzwil 3, Switzerland: Evidence for branch and twig foddering of livestock in the Neolithic. Journal of Archaeological Science 20: 479–502.

59.

Regnell

(2002) Charcoals from Uppakra as indicators of leaf fodder. Acta Archaeologica Lundensia 40: 105–115.

60.

Rösch

(2012) Forest, wood, and ancient man. Interdisciplinaria Archaeologica Natural Sciences in Archaeology 2: 247–255.

61.

Skládanka

Nedělník

Adam

, et al. (2011) Forage as a primary source of mycotoxins in animal diets. International Journal of Environmental Research and Public Health 8: 37–50.

62.

Slotte

(2001) Harvesting the leaf-hay shaped the Swedish landscape. Landscape Ecology 16: 691–702.

63.

Szabó

Hédl

(2013) Socio-economic demands, ecological conditions and the power of tradition: Past woodland management decisions in a Central European landscape. Landscape Research 38: 243–261.

64.

Tallowin

JRB

Jefferson

(1999) Hay production from lowland semi-natural grasslands: A review of implications for ruminant livestock systems. Grass and Forage Science 54: 99–115.

65.

ter Braak

CJF

Šmilauer

(2002) CANOCO Reference Manual and CanoDraw for Windows User’s Guide: Software for Canonical Community Ordination (Version 4.5). Ithaca, NY: Microcomputer Power. Available at: http://www.canoco.com.

66.

Thiébault

(2005) L’apport du fourrage d’arbre dans l’élevage depuis le Néolithique. Anthropozoologica 40: 95–108.

67.

Troels-Smith

(1953) Ertebollekultur-Bondekultur. Resultater af de sidste 10 Aars Undersogelser I Aamosen. In: Aarboger for Nordisk Oldkyndighed og Historie, pp. 5–63.

68.

Troels-Smith

(1960) Ivy, mistletoe and elm: Climate indicators – Fodder plants. Danmarks Geologiske Undersogelse 4: 1–31.

69.

Van Uytvanck

Hoffman

(2009) Impact of grazing management with large herbivores on forest ground flora and bramble understorey. Acta Oecologica 35: 523–532.

70.

Vera

FWM

(2000) Grazing Ecology and Forest History. Wallingford: CABI Publishing.

71.

Verheyden-Tixier

Renaud

Morellet

, et al. (2008) Selection for nutrients by red deer hinds feeding on a mixed forest edge. Oecologia 156: 715–726.

72.

Whitehead

(1995) Grassland Nitrogen. Wallingford: CAB International, 397 pp.

73.

Worrell

Clanton

Stroup

, et al. (1986) Effect of harvest date on meadow hay quality I: Nutritional attributes, voluntary intake and rate of passage in growing cattle. Journal of Animal Science 63: 1527–1537.

74.

Zeder

(2006) Central questions in the domestication of plants and animals. Evolutionary Anthropology 15: 105–117.

75.

Zimmermann

(1999) Why was cattle-stalling introduced in prehistory? The significance of byre and stable and of outwintering’ in settlement and landscape. In: Proceedings of a Conference (ed Fabech

Ringtved

), Århus, Denmark, 4–7 May 1998, pp. 301–318.