by Desconhecido
Hence, documentary sources give us no direct evidence on vegetation or environmental change in connection with the Black Death and the crisis. But indirectly they point to large-scale and profound changes. In countries with richer documentary sources, changes in rents, taxes and prices indicate changes in the agricultural production that must have been connected to vegetation change. In Sweden, the many deserted farms and the significant population drop (reflected in Peter’s Pence and several other sources) in particular indicate a social crisis that must have had major environmental and ecological consequences. In this way, we may use historical evidence of the societal crisis to discuss its possible environmental consequences, but much of the reasoning would be hypothetical.
A different approach is to use pollen analysis or other palaeoecological techniques to study vegetation changes more directly. In the study presented in this chapter pollen data from southern Sweden are used to interpret vegetation and agricultural change in the wake of the Black Death. Pollen analysis is a widely used method for studying vegetation change, but its potential is still rather unexplored in the context of the late-medieval crisis. In general it is used to study long-term vegetation changes of a distant past, in particular of prehistoric periods from which we have no written sources. However, also from later periods, like the Middle Ages and later, pollen analysis and other palaeoecological techniques have the potential to make important contributions.
Fig. 4. The basic steps of pollen analysis, from peat coring on a bog, via subsampling to microscoping. The pollen photo is taken at magnification ×400 (photos: Per Lagerås and Henrik Pihl)
Some earlier studies, both in Sweden and other countries, have shown that reforestation and other vegetation change in connection to the Black Death may be identified in pollen diagrams.2 However, studies are few and usually based on only one or a few sites, which have made it difficult to draw general conclusions. In this study a large set of pollen data based on numerous individual pollen studies is used. The aim is to look beyond local variation to identify possible regional changes in vegetation and land use that may be associated with a large-scale societal crisis.
The method of pollen analysis will be presented below but one principally important difference from the use of written sources should be mentioned here. Pollen data first of all reflect vegetation. But because vegetation in a cultural landscape is much influenced by agricultural land use, pollen data also reflect agriculture. Furthermore, because agriculture and, in particular, crop growing indicates nearby settlement, pollen data may be used to interpret not only vegetation and agriculture, but also settlement dynamics. And when large sets of pollen data are compiled, like in this study, the observed settlement dynamics may even be discussed in terms of population dynamics. Thus, the line of reasoning is from pollen to vegetation, from vegetation to land use, from land use to settlement dynamics, and finally from land-use and settlement dynamics to population dynamics. This is opposite to the line of reasoning offered by written documents. They primarily deal with social and societal matters, which indirectly may provide hints on vegetation change.
Hence, a palaeoecological approach offers new perspectives by taking changes in vegetation and agricultural landscape as starting points for discussion. In so doing it highlights the environmental and ecological dimensions of the social crisis.
Data and methods for this study
Pollen analysis is the most widely used method for studying past vegetation. The method was introduced already a century ago,3 and even though there has been considerable development in data processing, modelling and in dating techniques, the basic principle remains the same (Fig. 4).4 Pollen is released during flowering and in particular wind-pollinated plants disperse enormous amounts of pollen, of which the majority does not fulfil its mission of pollination but just falls to the ground. The pollen grains are very robust to decay and if they get embedded in wet, anaerobic environments, like peat or lake mud, they may be preserved for thousands of years. One cubic centimetre of such sediment may contain as many as hundreds of thousands of pollen grains. Since pollen carries information on past vegetation and landscapes, peat bogs and lake-bottom sediments with their microscopic content of pollen are nature’s own historical archives.
From cores retrieved from such archives several small samples are taken out at various depths representing different times of deposition. The samples are treated with chemicals to extract and concentrate pollen and thereafter analysed using a high-magnification microscope. The identification and counting of pollen grains is a time-consuming task and takes skill and patience. Usually between 500 and 1000 pollen grains are identified in each sample to get a statistically significant result. Some characteristic pollen grains can be identified down to species level but most of them only to so-called pollen types, which may include several species, or to plant genera or family. After identification, the raw pollen counts from each sample are transferred to percentage values and usually presented in a diagram with individual graphs for each pollen taxon. The diagram can be used to interpret past vegetation around the site, and the changes in pollen percentages observed throughout the stratigraphic sequence reflect vegetation change through time. Peat or other plant remains from the analysed core are radiocarbon dated to obtain an absolute chronology.
After its introduction, pollen analysis was for long used only to study what was regarded as natural vegetation development, in particular changes in forest composition, and how it may have reflected climate. This was partly due to the limited knowledge on pollen identification at the time, which was much restricted to tree pollen, but partly also to the general opinion during the early twentieth century that past human impact on the vegetation was small and not measurable. However, from the 1940s onwards, with the first identification of pollen from cereals and from plants thriving in pastures, pollen analysis gradually became a method also to study past agriculture.5 After an influential paper by Björn E. Berglund,6 pollen diagrams in the 1970s and 1980s were frequently used to identify periods of agricultural expansion and decline, which were discussed in relation to Ester Boserup’s model of a stepwise societal development and to ideas on carrying capacity in relation to different agricultural techniques.7
Until the 1980s pollen analysis was mostly used to study long-term trends over several millennia and the pollen diagrams that were produced had a rather poor time resolution. They also had a poor spatial resolution since they used cores from large lakes and bogs with a large pollen source area, and aimed at describing the regional development rather than the local. This changed in the 1990s. From then onwards there has been an increasing interest in Sweden and in many other countries for more detailed pollen studies of local vegetation development with a shorter time perspective. Some of these local pollen diagrams have been produced within the context of forest ecology, with the aim to answer questions on tree successions and woodland dynamics at stand scale.8 Others have been produced in connection with archaeological excavations in order to interpret local vegetation and land use that may be linked to the investigated remains.9 In both cases a detailed picture of local conditions with a high temporal resolution is preferred before a more general picture of long-term trends over a large region.
Hand in hand with this development, there has been an increasing awareness of problems connected to radiocarbon dating. Chronologies of most early pollen diagrams were based on radiocarbon dating of bulk lake sediments, which due to the so-called reservoir effect may provide erroneously old dates (the error is usually 100–300 years but occasionally much larger).10 In modern studies radiocarbon dating of bulk samples from lake sediments is usually avoided. Instead bog moss and other macro-remains (fruits, leafs, etc.) from terrestrial plants are used which has resulted in more accurate chronologies.11 This advance has been facilitated by the AMS technique (Accelerator Mass Spectrometry), by which very small samples can be dated.
Sweden has a long tradition of pollen analysis and a wealth of pollen data, an
d in particular during the last two decades several high-resolution pollen records have been published. These detailed, local records are the foundation for the present study. Because the aim is to reveal possible vegetation changes connected to the late-medieval crisis, which in a palaeoecological perspective is a rather sudden and short-lasting event, care has been taken to use pollen records that have a reasonably high resolution (i.e. many pollen-analysed levels) for the last 1000 years and have good and independent chronologies. Pollen records with chronologies based on radiocarbon dating of lake sediments or on pollen-analytical correlation with other sites have been excluded.
Fig. 5a. Map showing pollen sites in southern Sweden used in the present study. Symbol colours distinguish between lowland sites (< 100 m a.s.l.) (red dots), and upland sites (> 100 m a.s.l.) (blue dots). Sites: 1 Skeakärret, 2 Torup, 3 Häggenäs, 4 Skärsgölarna, 5 Östra Ringarp, 6 Grisavad, 7 Bocksten, 8 Yttra Berg, 9 Trälhultet, 10 Exhult, 11 Köphult, 12 Rosts täppa, 13 Bjärabygget, 14 Baggabygget, 15 Råshult, 16 Siggaboda, 17 Flahult, 18 Lindhultsgöl, 19 Öggestorpsdalen, 20 Åbodasjön, 21 Store mosse, 22 Storasjö, 23 Skärpingegöl, 24 Bråtamossen, 25 Mattarp
Fig. 5b. Map showing the pollen sites used in the present study including three sites in middle and northern Sweden (yellow dots). Sites: 26 Fjäturen, 27 Kalven, 28 Kassjön. For site descriptions and references to original publications see Appendix 1
Based on the above criteria a set of pollen data from 25 sites from southern Sweden (Fig. 5a) has been established. The vast majority of them (21 sites) are situated in the South-Swedish Uplands (Sw: Sydsvenska höglandet), and therefore much of our conclusions and discussions will focus on these uplands. One reason why the uplands have attracted so many pollen-analytical studies is that they are rich in well-preserved peatlands. In contrast, the agricultural plains of the lowlands have very few preserved peatlands, due to cultivation, drainage and peat cutting, and hence there are only a few useful pollen records (four sites) from the lowlands in the data set. In addition to records from the uplands and lowlands of southern Sweden, there are three sites from further north – two of them from middle Sweden and one from the northern part of the country (Fig. 5b) (for details about the pollen sites see Appendix 1).
The South-Swedish Uplands range c. 100–377 m above sea level and today they are, to a large degree, covered by forest, in particular planted spruce and pine forest. The high degree of forestation is due to the relatively poor natural conditions for agriculture. In comparison to the lowlands of Scania in the south, the mean annual temperature of the uplands is c. 1–2° lower and the growing season is 10–20 days shorter.12 The Quaternary deposits are dominated by sandy till, rich in stones and boulder, reflecting the hard gneissic bedrock, whereas the agricultural plains of the lowlands are characterised by more fertile, calcareous clay and silt.
The strong bias in our pollen data to marginal uplands makes it difficult to draw general conclusions on a national scale or to make comparisons between lowlands and uplands. However, because settlements of marginal areas were not isolated, ups and downs in marginal areas in one way or the other reflect changes in central areas too. Furthermore, marginal areas appear to be most suitable for studies of environmental responses to societal change. In comparison to central areas they show a more discontinuous land-use and settlement history, which makes it relatively easy to distinguish periods of agricultural expansion from periods of stagnation and decline. In particular semi-open landscapes characterised by a mosaic of woodlands and agricultural land are suitable for a palaeoecological approach, because in such landscapes changes in human impact are likely to result in vegetation changes that are distinct enough to be detected by pollen analysis. Therefore studies on marginal areas may be used to take the temperature on society as a whole.
In addition to pollen records, conclusions presented in this chapter are drawn from dendrochronological dating. The basic principle of this technique is the counting and measuring of tree rings for comparison with reference tree-ring series of known age.13 Trees, timber or any wooden objects that contain tree rings can be dated with very high precision, even down to a single year if the outer tree ring is preserved. In Sweden the construction of reference tree-ring series and systematic dendrochronological dating started in the 1970s.14 Since then a lot of dating has been performed, in particular of timber from churches and other standing buildings but also of wooden objects from archaeological excavations.
In middle and northern Sweden well-preserved extant log buildings (mostly small barns and storage buildings) have attracted a lot of dating activity, and many of them have turned out to be very old.15 The oldest ones are from the late thirteenth century and quite a few are from the first half of the fourteenth century. What is most interesting for the scope of this book is that the dates show a gap of about 100 years between the 1360s and the 1460s. The gap obviously reflects a pause in building activities due to stagnation and population decline in connection with the late-medieval crisis.16 To investigate if building activities also show a similar break in southern Sweden, a large number of dendrochronological dates have been compiled and analysed. Our data set includes all dates from the provinces of Småland and Östergötland performed by the Swedish National Laboratory for Wood Anatomy and Dendrochronology at Lund University (all together 1882 dates). Since southern Sweden has very few standing profane buildings from the Middle Ages the data represent first of all planks, posts and other wooden objects documented during archaeological excavations, but also standing churches.
Dendrochronological data from Småland will be used in this chapter as a complement to pollen data from the same region, first of all to discuss the dating of the late-medieval decline, but also in relation to woodland development.17
Abandonment of fields and farms
Among the different aspects of the late-medieval crisis, abandoned farms are probably the most acknowledged. In fact the mentioning of abandoned or deserted farms in written records from the centuries after the Black Death is one of the strongest indications of the societal crisis and of the associated drop in population numbers. Also the physical remains of them are sometimes preserved. The archaeological study of deserted farms and villages is a well-established field of research in several countries, particularly in England but also for instance in Germany, where they are referred to as Wüstungen. Even though farms may be abandoned for different reasons, and have so been from time to time throughout history, the Late Middle Ages appear to have been a period of unusually extensive abandonment.
In Sweden relatively few have been archaeologically investigated and conclusions on late-medieval farm abandonment are based mainly on documentary evidence.18 In particular cadastral registers have been extensively studied.19 However, the earliest of these registers are from the late fifteenth century and most of them are from the sixteenth or seventeenth centuries, i.e. several centuries after the Black Death of 1350. Therefore they do not catch the actual process of abandonment and they do not show the original number of deserted farms. What the cadastral registers do show is that many farms were still deserted one or several centuries after the Black Death.
The reason why abandoned farms are mentioned in these records in the first place is that they were still someone’s property. They were uninhabited but owned by other farms in the vicinity or by big landowners and they were regarded as a resource (or a potential resource) with an economical value. When they were registered they were uninhabited and probably had no buildings left, but their land was owned and managed by other farms, usually for pasture or mowing. The term “abandoned farm” (Sw: ödegård) in the cadastral registers thus means a piece of land that once belonged to a farm that is now uninhabited. Because the registers are not from the nadir of the societal crisis, but rather from afterwards when there was strong expansion of agriculture and settlement, many once abandoned farms were probably already re-inhabited and therefore no longer deserted. In the records they will appear as ordinary tax paying farms, und
istinguishable from farms that were never abandoned. Furthermore, there were probably many abandoned farms in particular in distant areas that were never reclaimed or registered. They were simply deserted, overgrown and forgotten and therefore missing in the historical record.
For these reasons it is not possible to say, based on the historical documents, how many farms that were originally abandoned in the wake of the Black Death and this is why different scholars have arrived at different conclusions. According to the cautious estimation by The Scandinavian Research Project on Deserted Farms and Villages, only about 15% or less of the farms in several parts of Sweden were abandoned.20 However, for one study area, situated in the uplands of southern Sweden, they suggested a more extensive desertion, about 25–40%.21 Janken Myrdal presented a quite different conclusion. According to him, in many areas of southern Sweden 30–60% of the farms was abandoned and in marginal uplands even more.22
From this background we will now turn to pollen data and look specifically for indications of farm abandonment. Because crop growing is closely associated with settlement, cereal pollen may be used as an indicator of settlement. In Figure 6 the cereal pollen graph for each individual site in the data set is presented. Except for the three sites from middle and northern Sweden, presented at the top the diagram, the sites are presented in order of altitude, starting with the lowermost site at the bottom. The graphs are plotted on a time scale and most of them cover the last 2000 years. Because cereals do not survive in the wild, cereal pollen is the best pollen indicator of crop growing and indirectly of agricultural settlement. With the exception of rye, all cereals are poor pollen producers, and even though the pollen grains are dispersed by wind, they are relatively large and heavy and do not travel long distances. Consequently, cereal pollen in most pollen diagrams reach only low percentages of the pollen sum (i.e. of the total number of identified pollen grains in the sample), but even low percentages may be used as indications of arable fields in the vicinity.