THE HISTORICAL ARCHAEOLOGY Ol = THE MEDIEVAL CRISIS IN SCANDINAVIA

Anna Sundberg, Bo Ulfhielm R Stig Welinder In the wake of the Black Death in i g go Europe saw demographic disaster, economic decline, and social and political breakdown. Thousands of farms were deserted. This is the Medieval Agrarian Crisis. The latest decades of outland archaeology, primarily within the frames of rescue archaeology, have made it possible to outline the course of the crisis in the forested parts of middle Scandinavia. The r 4th and r gth centuries were a time of economic change rather than economic decline. However, various areas changed in different ways. When taking outland production into account the medieval crisis has to be conceptualised in another way; it was not solely an agrarian crisis. It was also early industrial expansion and change towards extensive farming.

Ö degå rdsproj ektet), collected large amounts of data and discussed the crisis from a Scandinavian perspective (Gissel et al. zg8z).The crisis was reconstructed as percentages of deserted farms and population de-  cline from written records on ownership, buying and selling, and taxes and rents in relation to arable land.That is what was recorded by the medieval bureaucracy, and accordingly the crisis is often denoted an agrarian crisis in Scandinavia.Little relevant archaeology was avail- able at the time of the research project.
In Norway desertion percentages of more than 4o % were recon- structed in most of the country.Typical percentages in Denmark and Sweden were ~o-go %, and in Finland g -zo %.The differences were due mainly to the use of different kinds of data and different methods.
Recent research hints that all in all about 4o -go % of the Scandina- vian population died in the Black Death and the subsequent pandemic plagues during the r4th and zgth centuries (Myrdal zoog:z66 -z89), in parts of the area possibly more (Antonson zoo').
A main problem in the discussions of the zgyos was whether the plague pandemics were the most important or possibly the only rea- son for the crisis.It was noted that some farms were deserted already before the xg gos, and thus it was discussed whether the Black Death hit a society already in crisis, for instance due to overexploitation and climatic deterioration.However, less effort was invested in the reconstruction of the build-up of the crisis than in the recovery of society after it.The simple reason was the easier access to informative written records from the x 6th and z yth centuries than before z g go.
We will try to discuss the medieval crisis in middle Scandinavia go years after the Scandinavian Research Project on Deserted Farms and Villages.We will use the data from historical archaeology that have come to light during this period, not least within the frames of rescue archaeology.The increase in available data has been huge, and various kinds of sites that were hardly known in the zg6os have been recorded and excavated in large numbers.This is valid especially for areas out- side the cultivated and urban parts of the country.
It is notable that the large-scale Ystad Project on long-term change in the south Scanian cultural landscape found few indications of medieval decline in agrarian activity (Andersson &: Anglert zy8g).Pollen data and theoretical calculations suggested that the landscape was not used above its carrying capacity towards the z4th century.Mats Widgren has suggested the same for all of southern Sweden (Widgren rygga).
The bistorical archaeology of the medieval crisisi n Scandinavia In parts of Trendelag, in areas more marginal from the viewpoint of corn cropping than in southern Scania, the archaeological record on land clearance and farmyards displays a more complex and dynamic process than the written record for the period t zoo -r 6oo (Berglund et  al. zoo').One reason is that the written record has little to tell about outlying-land production.
The most recent study of expansion and stagnation in marginal land is the one on north-western Scania by Per Lagerås (z, ooy).After some centuries of expansion some farms were deserted while others were not during mid-medieval time.Anders Ödman has previously studied the same area (Ödman tqg8, zoor).

OUTLAND ARCHAEOLOGY
The Scandinavian Research Project on Deserted Farms and Villages studied the agrarian landscape, that is, farmyards and the farms' in- land of arable land and meadows.A complete farm, however, also con- sisted of its outland (Andersson tgg8; Hansson et al. zoo':77 -88).
The outlying land was used for forest grazing, iron production, fish- ing and hunting, slash-and-burn cultivation, for timber supplies, tar, birch bark, and for many other purposes.In most of Scandinavia the outland production was of more importance within the economy of the peasant households and within the market economy than the in- land production (Svensson tyg 8).
Today there is a great deal of data especially on charcoal and iron production and on the use of pitfalls.Dozens of iron-smelting furnaces, charcoal pits and pitfalls have been radiocarbon dated.It is possible to discuss whether the medieval crisis had another course when seen in the perspective of outlying-land production than that of the agrarian crisis reconstructed from written data on the cultivated landscape (cf. Holm tyga; Nilsson z, oog).

THE RESEARCH AREA
We have chosen to study an area in the middle part of the Scandina- vian Peninsula.The research area consists of the county administra- tive districts (Nor.fylke; Swe.län) Värmlands län, Hedmark, Gävle- borgs län, Dalarnas län, Jämtlands län and Trendelag, which actually is two fylke (Fig. t).
The part of Scandinavia in focus is a generally forested area with cultivated land along the coasts, fiords and river.valleys, and around Royal, church and feudal power was little exerted in the area until the r4th century or even later.The exceptions are the establishing of the royal and episcopal cities of Hamar and Trondheim in Hedmark and Trendelag, respectively, in the roth and zzth centuries.In these Norwegian fylke there also were large estates (R@skaft z. oog).Urban-  ism and monasticism in Dalarnas län and Gävleborgs län are r gth- century phenomena, while the strongholds are mainly from the z4th century.Many of them disappeared during the early z g th century.This is valid also for the single stronghold in Jämtlands län and for those in Värmlands län.
Market economy was present in all of the research area during the Middle Ages, but the organising power was little visible as monumen- tal sites.The obvious exceptions were the best agricultural districts, for example parts of Hedmark and Trendelag, and the Bergslagen mining district in the south-eastern part of the research area (Fig. r).

THE DATA SET
We have collected all available radiocarbon dates from the time span AD r -x Soo for various kinds of sites in the research area (Table T).In some of the administrative districts in the research area there are local antiquarian institutes with reassuring records of nearly all the radio- carbon dates ever obtained within the districts.In other districts some of the co-authors have spent much time scrutinizing libraries and ar- chives for dates.
Table t.The number of dates/dated sites in the data set organised according to county adminis- trative district and tvvpe of site.
From sites and features with more than one dated sample we have used reasonable mean values.Thus, the figures in Table r and accordingly the graphs in Figures z. -c/ note the number of features and sites, not the number of individual dates.The hundreds of dates from the Gråfjell area in Hedmark are not included in our data set (Amundsen zooy; Rundberget zoo').They are commented upon in the text.
All the dates are calibrated and thus meant to be readable as the common historical time-scale (Ramsay et al. z, oo6).We have not sys- tematically recalibrated dates previously calibrated according to vari- ous methods.All dates are associated with statistical errors of various sizes.In the graphs in Figures z,g we have placed each date in its sta- tistically most probable century.
There are other methods for presenting large amounts of radiocar- bon dates (e.g.Magnusson rg86).Our graphs (Figs.z, -g) represent both an honest attempt to present the actual amount of available data and easily survivable overviews to be used in the interpretative discus- sion.Of course the latter has to be guided by cautiousness and common sense.What are seen in the graphs are long-term trends, not what happened in single centuries.
What is also seen in the graphs is a summary of modern land pres- sure and exploitation resulting in rescue archaeology in addition to various research strategies within field projects with research grants.These factors may differ among the various districts in the research area, thus making comparisons between the districts problematic.
In Hedmark almost all the dates derive from rescue archaeology dur- ing the last go years.This is valid also for Gävleborgs län and Dalar- nas län.The advantage is that a more or less random selection of sites have been excavated within those part of the districts where land has been exploited.Bias based on the idiosyncratic ambitions of individual research projects is not present, or at least only to a limited extent.In Dalarnas län, systematically the oldest clearance at each rescue-archae- ologically excavated site has been radiocarbon dated.This of course influences the graph in Figure g.Otherwise, the graphs for Dalarnas län are based on a selection of relevant data from a list of more than zoo individually dated samples.
In Värmlands län r go dated samples out of a total of :r 86 derive from research projects.The small group of researchers that planned the field and laboratory programs was especially interested in the Late Iron Age and the Middle Ages.Accordingly the late radiocarbon dates are few.Rescue archaeology in Jämtlands län is a rarity.Nevertheless a remarkable number of pitfalls and bloomery furnaces have been ra- diocarbon dated.
Trendelag is an area with high modern land pressure and a univer- sity department responsible for the rescue archaeology in the district.
For our data set it has been surmountable to present only the dates of iron-production furnaces.Bloomery iron production is a special re- search interest of the department of archaeology at the local university (Stenvik zoo').It is, however, obvious that radiocarbon dating of iron production has not only been popular in Trendelag during the latest decades (Figs.g-6).
In addition to radiocarbon dates we will discuss dendrochronologi- cally dated log houses.There are excellent published overviews of tens of houses from the Swedish part of the research area, most of them from Dalarnas län and Jämtlands län.Houses presumed to be older than z6oo have been systematically dated.Houses older than the very late zgth century are not known (Landström R Bartholin zc18y).
Quite another kind of data is pollen diagrams.We have not made a systematic survey of all available informative and reasonably welldated pollen diagrams from the research area.Our selection has been easily found in published literature.The pollen data will underline trends seen in the radiocarbon dates of the archaeological sites.The actual diagrams will not be presented here but only a simple overview of events of interest for our discussion (Table z, ).
It is notable that dates of the actual houses may indicate and date the desertion of farmyards.Pollen diagrams, on the other hand, indicate and date the desertion of the arable land and pasture of the farms.

HOUSES, FARMYARDS AND VILLAGES
In Gävleborgs län and Jämtlands län the number of abandoned farm- yards has been calculated (Fig. z, ).The dates are based on interpreta- tions of series of radiocarbon dates and on the interpretation of the presence and absence of various artifact types.The abandonment of a farmyard may indicate that the farm had been deserted, or that the farmyard had been moved to a new nearby location.It is tempting to suggest that the increase in the number of deserted farmyards in the z4th and x yth centuries (Fig. z) is related to the medieval crisis.
For Värmlands län, Hedmark and Dalarnas län the graphs are based on radiocarbon analyses of post-holes, hearths, wall-ditches, and a multitude of other kinds of remains from general settlement activities (Fig. z).The patterns in the three histograms are very different.At least in part, this reflects ambitions within the local rescue archaeology and, as concerns Värmlands län, local research programs.However, there are no dates from Hedmark and Värmlands län after the z4th and xgth century, respectively.It is reasonable to interpret this as population de- cline and desertion of farms.
In the graph for Dalarnas län in total there is a stop in log-house building for a century after zg6g (Raihle zcfc1o:Fig.8).The pattern is the same in other areas where many log houses are still standing (Land- ström R Bartholin rg8y; Hovanta rygg).For a century there was no need and possibly no labour force to build new log houses.This is not an indication of a total desertion of farms but of the breakdown of the continuous everyday life on the farms.According to the radiocarbon dates, however, life went on in the farms and villages.
The archaeological data on houses, farms and villages correspond to the description of an agrarian crisis by the Scandinavian Research Project on Deserted Farms and Villages.It is indicated that farms were deserted in our research area.Archaeology does not, however, have data to quantify the desertion as was done by the historical research project in the r godos from the written record, except possibly for Jämtlands län.Helge Salvesen, a member of the historical research project, estimated the number of deserted farms in Jämtlands län to about E). ro.Mod-  ern estimates based on, among other things, field registrations suggest 6oo -8oo (Hansson et. al. z, ooy:6g -yo; Antonson zoog).As much as go -6o % of the farms may have been deserted in Jämtlands län.

PASTURE AND ARABLE LAND
The occurrence of clearance of arable land and possibly grazing areas is dated from charcoal underneath clearance cairns and lynchets.The process of clearance has been very different in those parts of the re- search area with sufficiently many radiocarbon analyses for an evalu- ation (Fig. g).
In Dalarnas län and Jämtlands län clearance of land expanded dur- ing the r rth century.In Jämtlands län this activity declined after the r z, th century, while it continued and culminated in Dalarnas län dur- ing the rqth century, that is, the very century of the medieval crisis.In Hedmark, too, clearance activity increased around the ryth century after a long period with no dated clearance cairns and lynchets.Värmlands län has few dates, but the presence of clearance activity in the period r zoor zoo is indicated.The long series of dates from the site Röysfetet in Hedmark (Amund- sen zoo':Figs.rqo -rgr) starts in zoo, culminates 8oo -rooo as con- cerns dates of clearance, and culminates in r z, oo -r yoo as concerns dates of the continuous use without any decline indicating desertion or less intense use of the area.Thus, the chronology of the process of land clearance is not in agree- ment with the general view of farm desertion after r glo.New land was continually cleared.In Dalarnas län there was even a boom-period during the zyth century, while the corresponding period in Hedmark is linked to the recovery towards the end of the crisis.
The pollen diagrams summarised by us (Table z, ) are all from what is today forested areas or the summer-farm landscape.The latter is where the seasonal summer farms (see below) were in use until the late z EJth or early z, oth century.The chronology of the diagrams is based on in- terpolation of radiocarbon dates.It is actually less accurate than what is shown in our overview (Table z, ).Grain has been cultivated and grazing has taken place in all of the for- est and upland areas covered by the pollen diagrams.It is notable how corn cropping or forest grazing started in many areas around rgoo-r4oo.In considerably fewer areas this was a time of decline or a total stop in agricultural activity.
In the research area, the time of the medieval crisis was more of- ten a period of clearance of land and expansion of pasture than stag- nation.This is especially apparent in Dalarnas län.In general it was a period of dynamic change as concerns cultivation and stock-breeding in outlying land.

SUMMER FARMS
The summer farms (Nor.seter; Swe.fäbod) were seasonally used for forest and upland grazing of the cattle and, at least during the period t 6oo -:r ctoo, dairy production for the market.Summer farms, or rather some kind of cattle-posts away from the main farms, have been in use in Scandinavia since the Early Iron Age.
There is a series of radiocarbon analyses with close to a reason- able number of dates from Hedmark (Fig. 4).It shows how the use of summer farms increased after the z4th century.The houses at the site Reysfeltet in the Gråfjell area are dated to the period rgoo -ryoo (Amundsen zoo': Fig. z.z. y).At the site Deset 8stseter they are dated to r goo -z goo (Amundsen zoo': Fig. z6g).These summer farms were established around the time of the crisis.The same, or at least forest grazing, was demonstrated for Gävleborgs län, Dalarnas län and Jämt- lands län with the aid of pollen diagrams sampled near early modern summer farms (Table z. ).

IRON PRODUCTION
The study of pre-industrial iron production has a long tradition in Scandinavia.Long series of radiocarbon dates of iron-smelting fur- naces and slag are available from all the districts in the research area (Fig. g).The production of charcoal is closely linked to iron production, but the excavation of charcoal-production sites started later in Sweden, so there are fewer radiocarbon dates of charcoal pits than from the actual iron-production sites with the exception of Hedmark in Norway (Fig. 6).
There are obvious local boom-periods in the iron production, for example zoo -zoo in Jämtlands län, 8oo -z ooo in Värmlands län, and yoor T. oo in Dalarnas län.The graphs for charcoal production mirror those for iron production during the latter two boom-periods (Figs.g -6).
The kind of bloomery iron production in focus here -not to be confused with water-powered blast furnaces (see below)came to a stop in Värmlands län in the T4th century, and decreased notably in Hedmark at the same time, as did charcoal-production.The same hap- The historical archaeology of the medieval crists in Scandinavia Värmlands län 0 100 200 300 400 500 600 700 800 900 100011001200130014001500160017001800 l. ig. 6.The radiocarbon dating of charcoal produc- tion AD 7 -tsoo (number ot radiocarbon dates of charcoal pits and sucks).pened in Trendelag a century earlier.In Gävleborgs län and Dalarnas län iron production decreased already around the r r th century but did not came to a total stop until the z 6th century.
Thus, in the south-western part of the research area the decline in iron production was related to the medieval crisis during the z4th cen- tury.In the other district boom-periods came and went independent of that crisis during a millennium.
Jämtlands län has a pattern all in itself.The :r4th century saw a re- markable increase in iron production (Fig. g).The second boom-period of iron production in the district started during the medieval crisis.The same is valid for the north-western part of Gävleborgs län adjacent to Jämtlands län.The summary of more than 6oo radiocarbon analyses from south- ern Norway by Jan Henning Larsen (zoo4:Figs.g -6) shows a continu- ously decreasing number of dates per century from the z g th century and onwards for iron production and the z4th century and onwards for charcoal production.It is obvious that this kind of iron production for the market came to a stop as a consequence of the medieval crisis, CvRRENT S5YE1715H AKcHAEQLQGY VQL 17, 009 The historical archaeology of the 777edieval crisisin Scartciinavia Hedmark 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 although the production had decreased already before the z4th century in some areas like Trendelag and Gråfjell (Rundberget zoo6, zoo':g zz).
In the local studies in Setesdal (Bloch-Nakkerud r 078') the r 4th century is pinpointed as the time of a total stop in production.The long series of dates of charcoal pits may even hint at the decades around T g go.
The Middle Ages saw a general change in the bloomery-furnace production, from production for the market to small-scale production for local household use.In some areas this change was related to a breakdown of the production during the z4th century and the medi- eval crisis.In Jämtlands län the opposite took place.
The changes outlined above must, however, also be related to the introduction of large-scale iron production in blast furnaces.That will be done below.

TAR PRODUCTION
There are only few and fairly recent excavations of tar production sites (Hennius et al. zoog; Hennius zoo6).In the research area there is a small series of radiocarbon dates from Hedmark (Fig. y).It suggests that tar production continued uninterrupted through the medieval cri- sis.The series of dates from the Gråfjell area starts in tyoo (Amund- sen zoo': Fig. z8z).

MASS CATCHING
Pitfalls for moose and reindeer have been excavated in fair numbers since the zc76os, but in the research area there are many radiocarbon dates only from Jämtlands län (Fig. 8).There is no obvious pattern in the distribution of the dates.One reason may be that it is difficult to get high-quality samples for analyses of pitfalls from a stratigraphical point of view.
Egil Mikkelsen in his survey of mass catching in southern Norway (I994) has demonstrated that pitfalls were in continuous and equal use until about z zoo.Large-scale constructions requiring a big labour force to maintain and harvest them, on the other hand, stopped being used at the end of the zgth century.The use of hunting cottages and the ac- cumulation of bones outside the cottages in the west Norwegian high mountains also decreased around Tgoo (Mikkelsen 1994:Fig. yg).Egil Mikkelsen's interpretation was that wild reindeer was close to extinction towards the end of the z g th century due to overexploitation.However, it is notable that the mass catching did not start again after a period of recovery.The medieval crisis did not cause the breakdown of mass catch- ing of reindeer in the Norwegian high mountains, but it contributed to its not being restarted after its breakdown for other reasons.
Mass-catching of moose and reindeer in pitfalls and other construc- tions decreased in parts of the research area around the zgth and z4th centuries.The relation to the medieval crisis is not obvious, as dem- onstrated by the local study in southern Norway (Mikkelsen z994).
Jämtlands län again is an exception.The number of pitfalls in use in- creased during the z gth century in the same way as iron production a century earlier.

MINING AND BLAST FURNACES
The earliest archaeologically excavated and radiocarbon-dated blast fur- naces for iron production are situated in the southern part of Bergslagen, slightly to the south of the research area.They were established in the z z, th and zgth centuries (Th.Eriksson zooz, ; Eriksson R Pettersson Jensen z, ooz, ; Anund 5c Skyllberg z, oog).The Falun copper mine and smelteries were, according to the earliest radiocarbon dates, established in the z 3 th century, and the activity increased in the x g th and z 6th centuries (Fig. 9).
The very start may have occurred even earlier (Qvarfort z9yy).
The charcoal production in charcoal stacks is linked to the new un- limited need for charcoal in the early industrial production of iron and copper.The earliest radiocarbon dates are from the r g th and r 6th cen- turies in Gävleborgs län and Dalarnas län.The early modern peaks in Figure 6 are mainly formed by dates from charcoal stacks.
Thus, the decline in bloomery-furnace iron production in Värm- lands län, Gävleborgs län, Dalarnas län and Trendelag around the r zth century (Fig. 4) was due to the introduction of the new industrial methods.Full-scale production according to the old technique contin- ued only in Hedmark, where the medieval crisis meant a stop, and in Jämtland, where the production actually increased (Fig. y).In these two districts the continued bloomery-furnace production was based on new, more efficient types of furnaces, in Jämtlands län from the r rth century (Magnusson rg86:ryo -ryr) and in Hedmark from the r4th century (Holm rygg:gnag).
New technology, both small-scale and industrial scale technology, was introduced in the centuries around the medieval crisis.The total production increased, also during the r4th century, rather than de- creased.However, there were changes in organisation and relations to household and market production associated with the various technol- ogies.This must be discussed on the basis of all the data on economy present in our database (Figs.z -9,

LONG-TERM CHANGE
In spite of the data being of variable quality in the different parts of the research area, one conclusion is justified.There is not one unani- mous pattern of change in all of the area.We can grasp the variation by outlining three different trajectories of long-term change.Common, however, are a pattern of expansion, change generally around the r4th century, and the emergence of a new kind of economy.
The royal and episcopal structure in Trendelag had its economic basis in rich and expanding farmland and outlying-land production, and this is also the case in Jämtlands län.The reoccurring plagues from r glo and onwards meant a demographic disaster with desertion of hundreds of farms, archaeologically demonstrated in Jämtlands län and known from written history in Trendelag.Many of the deserted farms in Jämtlands län were turned into grazing land, hay-making land and summer farms (Hansson et al. zoo':8q -gz, ).Iron production in bloomery furnaces increased, in Jämtlands län during the r4th century and in Trendelag during the r 6th century.
In Dalarnas län the disaster of the Black Death is drastically demon- The historical nrchaeology of the medieval crisisi n Scandi navin strated by the total stop in log-house building in the z g gos.At about the same time farms were deserted in Gävleborgs län.Simultaneously, land clearance and the establishing of summer farms increased in Dalarnas län.The idea is that bread, beer, ox-traction and ox-hide were needed in the new mining and blast-furnace industry denoted Bergslagen in the southern parts of Dalarnas län and Gävleborgs län.The clearance of woods for arable land and grazing in Dalarnas län during the t4th century is especially notable.Continuous forest clearance during the t4th and t gth centuries has been noted also in southern Sweden (Eric-  sson z004).
The outlying-land production in Värmlands län, especially the northern forested part, and Hedmark and other parts of Norway col- lapsed at various points of time during the t g th and :t4th centuries, in some areas somewhat earlier, after centuries of expansion and development of a market economy.The relation to the mid-t4th-century dis- aster is far from obvious although in some areas there may be a close connection.In Hedmark there was land clearance and the establishing of summer farms in the outland.A corresponding expansion in forest grazing is seen also in Dalarnas län, Gävleborgs län and Jämtlands län.The idea is that a general change towards extensive household-oriented farming took place during the r4th and t gth centuries (cf.Antonson 2.004:2.06-2I2.I 2.009).
As far as there are archaeological data available on farming settle- ment, these data generally confirm the conventional reconstruction of a medieval agrarian crisis indicated by deserted farms from the mid- :r4th century onwards.Our data, however, mostly concern outlying- land production, like mass catching and iron production, in forest and low mountain outland.The traj ectories of change outlined above, thus, are not to be applied to central farming areas.Nevertheless, they are of importance in understanding the main course of the medieval crisis.

THE NEW ECONOMY AND THE FLEXIBILITY OF OUTLAND USE
The &4th and r gth centuries saw a demographic catastrophe and the breakdown of the social and economic structure.In some areas there was change in the local economy, among other things the establishing of Bergslagen in the south-eastern part of the research area and the corresponding industrial development in parts of southern Norway, already well before or just a few decades before 3 g go.In other areas the change is closely linked to the crisis and especially the time after the culmination of the crisis, that is, the change was part of the recov- ery from the crisis.
Notable examples of change in local economies are the increase in bloomery-furnace iron production in Jämtlands län and less conspicuously in Trendelag, the clearance of corn-producing land and the in- crease of forest grazing and the establishing of summer farms in Hed- mark and Dalarnas län, and the establishing and increase in mining and blast-furnace iron production in the latter area and Gävleborgs län.
In forested land and low mountain land, that is outland, the economy during the Early Middle Ages in Scandinavia was a broad-spectrum economy with local market specialisations and relations to re- gional markets.The local economies were linked to royal, aristocratic and episcopal elites from an agricultural point of view, and in addition to that to urban centres like Hamar and Trondheim in the Norwegian part of the research area.In the wake of the demographic disaster in the late tqth century the market broke down in parts of the research area, and in some areas the market-oriented production decreased or disappeared.Outlying-land production, however, has an inbuilt flex- ibility in its broad-spectrum economy.If one kind of production had to be abandoned, other possibilities existed.Outland production is char- acterised by continuous flexibility and, thus, a readiness for changeand actual continuous change.
There are signs in our data that the outland production in the re- search area changed according to local demands and boom-periods in the centuries before the z4th century, parallel to a general expansion in population, land use and market economy.During that century, and in most parts of the research area, there was intensification in change.The local economies reacted to the crisis in different ways in different areas: the emergence of a new kind of market economy based on new iron-production technology; the change towards extensive cattle breed- ing and dairy production; production for the local, developing, mining and blast-furnace industry; or a general re-orientation from a regional market economy to a local household and subsistence economy.

THE MEDIEVAL CRISIS
The course of events that started during the z4th century has been de- noted 'the agrarian crisis'.Certainly many farms were deserted and a great deal of arable land was left uncultivated (Myrdal r gnag: t ty -r zg ).
The Black Death was a demographic disaster that triggered downward feedback of social instability and economic stagnation, a classic exam- ple of a crisis that eventually was turned into a new start within a new structure (Myrdal z, oo6).
However, the outlying land and outland production did not see much of a crisis.The many-sidedness in the inland-outland economy offered many different choices when the local economies had to ad- just to regional change.The outland production contained flexibility.It functioned as kind of protective net in times of crisis, and so it did when the labour force and market broke down during the medieval crisis.
Farms and arable land were the backbone of medieval society as concerns ownership, identity and prestige.Accordingly, it was tax-pay- ing and bureaucracy, and accordingly, farms and arable land have been the basis for the study and conceptualisation of the changes in the zyth and z gth centuries as an 'agrarian crisis'.In the outland those centuries meant continuous structural change within the frames set by the series of plagues, the development of early industry, and the increasing influence of the royal power and aristocracy.An interesting problem to discuss from other kinds of data than those recorded by us is whether the areas where outland production actually increased during the me- dieval crisis reflect direction by local and regional authorities in order to compensate for the lost taxes and rents from the deserted farms, or if it was a way for peasants to hide their production in the forests and mountains from the authorities, which turned increasingly repressive during the crisis.
Fig. J.The radiocarbon dat- ing of clearance of arable land AD 1 -1goo (number of radiocarbon dates).
Overview of relevant events In selected pollen diagrams from the research area (from Hafsten I 99z; Emanuelsson zoo I; Karlsson zoo I ; ivlattsson zoo I ;Hansson et al.Eoos:Tab.3 I ; Wallin zoo6).
Fig S The radiocarbon dating of small-scale iron production AD 2 -2 Soo (number of radiocarbon dates, mainly of furnaces).

Fig
Fig. y.The radiocarbon dating of tar production AD r -rgoo (nuIuber of radiocarbon dates of tar pits).