Before and After the 536 CE Dust Veil

18 March, 2013

This week On the Reading Rest I have a report on the excavation of six Iron Age settlements. Chronologically they cover the great Roman Iron Age (RIA) settlement expansion and the harsh Pre Carolinian Iron Age (PCIA) conditions partly caused by the dust from the Ilopango eruption 535-536 CE. This eruption created the Tierra Blanca Joven (TBJ) tephra and veiled even the northern hemisphere in years ‘without summer’ (1). In the present, Mälar-Valley, case the settlement area under discussion is slightly peripheral in relation to central local settlements, i.e. the villages still existing today or known from historical records.

Hennius fig 01

Hennius, Andreas et al. 2012. Andreas Hennius (Main author and editor) Äldre Järnålder i Danmarks socken – Sex boplatser vid Säby (Early Iron Age in the Denmark Parish – Six settlements at Säby). Med bidrag av—with contributions from—Anneli Blom, Torbjörn Brorsson, Ylva Bäckström, Erik Danielsson, Dan Fagerlund, Per Frölund, Stefan Gustafsson, Malin Lucas & Örjan Matsson. Upplandsmuseets rapport 2012:15.

There is a summary in English (2) and the report discusses seven settlements inasmuch as one was excavated already six years ago (3).

Hennius fig 02

Light blue shoreline 500 BCE, dark blue shoreline 500 CE

Sä-by means ‘the village by the water’ and the water in question, the (i.e. ‘sea’), has generated several place names. The report will probably be known as the Säby Report for quite a while because it addresses a significant settlement period and a model case – a well-defined settlement area – in a clear-sighted way. In a comparison of three major settlement areas around Uppsala Andreas Hennius, moreover, discusses some of the prominent differences and similarities between the long-term settlement development and the socially and environmentally generated characteristics of the Roman Iron Age (RIA) settlement expansion, its stagnation and eventual demise.

Hennius fig 03In this case, the name Säby points to a settlement situation in the beginning of the Common Era on a narrow peninsula less than 300m from the water. To establish a permanent settlement at this site in a period that eventually changed the area from archipelago to inland, must have stood out as significant in the beginning of the Common Era, but rather pointless c. 500 CE. Present-day Säby with its cemeteries just north of the village, was in all probability the center and beginning of the permanent prehistoric settlement. The cemetery is a typical Early Iron Age (EIA) border cemetery fencing off Säby towards Gnista and other settlements such as Slavsta the odd kilometre north of present-day Säby. As an economic area the central Säby settlement benefits from the land rise, which created new meadows and eventually fields to the west of the village.

Clarity is one of the great qualities of this report because it allows the reader to use the sites in new discussions – testing hypotheses and bringing the report into more general archaeological discussions. I have chosen three topics to illustrate this. (1) Formal rules in the IA building tradition. (2) The layout of the homesteads. (3) The character of the settlements, given that they were not the most central ones in the greater Säby area.

Hennius fig 04

(1) Formal rules in building tradition. If we superimpose the plans of the post holes at D180 upon the older neighbouring houses at D168, the over lapping and systematic dislocation is easy to see. This indicates that the four houses were built with same measuring rod. A similar conclusion can be drawn from the parallelism between the post pairs when they intersect the length axis of the house at oblique angles.

Hennius fig 05

We may thus figure out the length of the principal unit, and a simple analysis shows that this unit was a foot of c. 31.25cm or an ell, i.e. a double foot. Similarly to houses, e.g. from Jutland, the long measures are also the more precise ones (4). This trait can be explained if we suggest that long measures and relatively large units are the more important ones because indirectly they define the general character, size and costs of the house as a construction. Room size and partition walls, on the other hand, are a matter of convenience rather than system. There are indications of the foot being divided into thirds and fourths, which probably means that a foot consisted of 12 inches. The length of some houses can be divided by 1, 2, 3, 4 as well as 6, 9 or 12 feet, i.e. total lengths amounting to 72, 84, 96 or 108 feet. When builders aiming at having 62 feet between the outer post pairs, preferred the series 14.25 – 17.5 – 17— 13.25 = 62f over the technically equivalent: 15.5 – 15.5 – 15.5 – 15.5 = 62f, it stands to reason that 14.25 – 17.5 – 17— 13.25 = 62f, demonstrates room proportion and variation as a value in itself. Perhaps measuring a distance down to ¼ of a foot (c. 7.8 cm) is aethestical overkill and difficult to prove, but the general attitude to constructing architecture with large simple and short complex measures — creating a standardized exterior and differentiated interior — is nevertheless a widespread Roman Iron Age (RIA) Scandinavian phenomenon.

(2) The layout of the homesteads. If we want to look at the settlements and their distribution during the 400 year boom from the 2nd to the 6th century CE, we must start by excluding D170 and D190 because they are too old. And if we want to catch the beginning of the boom rather than its aftermath, we must also for the moment exclude D168 and D180, i.e. two phases of a mid-millennium solitary farms. This leaves us, on the one hand, with D162 and on the other with D169 and D193. The two latter are rather scattered settlements with a mixture of main houses and smaller or lager outhouses. D162 is the eastern part of a much more dense settlement with main houses as well as large and small outhouses. At D162 the settlement area is restricted and the buildings are orientated either E-W or N-S. D 162 is the western part of a small village.

At D169 none of the 9 houses overlap and among the 11 houses at D193 there is only one small overlap between the SE and NW corner respectively of two houses. At D162, on the other hand, 8 of the 15 houses from the 2nd to 6th c. overlap. This means that there is a 10% overlap at D169 &193 and a c. 50% overlap on D162, which in its turn indicates significant differences between the two kinds of settlement.

On D162 there are 7 main houses, 4 large outhouses and 4 small ones. The main houses are more or less parallel to each other, the E-W orientated houses are quite far apart and the N-S houses make up an E-W row of main buildings. The small outhouses are situated among the main houses, but of the four large ones, one is orientated N-S and three E-W. The latter from an E-W row of buildings.

The clue to understanding the village D162 is the way it changes over time. Wisely, the excavators devoted two 14C tests to every house foundation and because of the generally speaking wet conditions on the Säby sites it was possible often to date fragments of the posts found in the post holes. In principle these dates tend to be older than the houses even though they may have been introduced in connection with repair during the lifetime of the house. Moreover, old timber could have been used: side ridges may thus have been redressed to fulfill a secondhand function as posts, and interior posts may have been shortened to become corner posts. Since a post gives the house a terminus post quem (tpq) date, other pieces of charcoal in the holes may be viewed in relation to this date telling us whether the charcoal fell into the hole as a contamination  from earlier activities or not. Typically in a settlement such as D162, age difference between the oldest post and the youngest date – be it a post or a piece of charcoal – grows when we compare the latest end dates to the earliest.

Hennius fig 06The table shows that the settlement comes to an end c. 1550 bp – allowing the outhouses to stand c. 25 years after the last construction phase – sometimes in the first half of the 6th c. CE. The dates, moreover, help us divide the activities of the builders into stages. The first is bp 1775-1740, H 11, 10 & 07. The second is short, bp 1695-80, H15 &14 (in effect two phases of the same house). The third is an even shorter period bp 1655-50, H09 & 08. Lastly there is the phase bp 1600-1575, H19 & 13. Taking the houses with just one date into consideration makes the last construction phase begin in 1605 bp and adds houses to the other phases.

When mapped, these construction periods structure the settlement in a reasonable way. In the phase bp 1775-1740 there are two parallel main houses with each their small outhouse or barn. They are two small farms typical of the ERIA. The pair H11+18 may well have been the very first farm on the site later accompanied by the farm consisting of H10+7.

Hennius fig 07

Hennius fig 08

The short phase, bp 1695-1680, introduces a new orientation of the settlement and larger houses. Instead of a N-S line between two farms, the axis is shifted c. 24° to the East. This suggests that the undated House 12 is partly contemporary with the introduction of this new axis. The new houses, constructed in this and the next phase, are orientated E-W, e.g. House 14, at right angles to the new divide, i.e. 90+24=117° to the East. This indicates a regulation of the plots and makes them the two westernmost estates in a village stretching towards the East. There seems to be no small outhouses belonging to this construction period, but compared with the first phase, the main houses are longer. Measured in roofed square metres the farms have in other words become bigger. In the fourth phase, bp 1655-50, a new farm with a small barn is built on the western plot.

In these two latter phases, i.e. the settlement period combining the two construction phases and covering the period bp 1695-1650, there are probably two farms on the western plot and one on the eastern suggesting that there was now two households on the western plot. This partitioning into South Plot and North Plot indicates that the lands of the original estate were parceled out to two beneficiaries. In the last building phase, bp 1605-1565, after a hiatus or the ebbing out of the main house period, bp 1775-1650 and onwards, four outhouses were constructed on the site. They are obviously plot-related — although perhaps not conpemtorary — and should be seen as someone’s claim to a specific plot and a part in the village. This means that although the plots were without a proper homestead, the estates nevertheless existed.

The development demonstrates that in the first part of the 6th c. at least three small farms disappears from the village D162, without radically affecting the ownership of the land. It is easy to imagine a connection between this abandonment of farms and the dust veil tpq 536 CE.

The way the building phases mirror the boom in the LRIA and its decline during the 5th century CE is typical, but the outhouse period, the very beginning of the PCIA, is equally important because it speaks of the will to claim the land.

Hennius fig 09That D162 is a village is also indicated by the fact that it has its own small cemetery D69, 200m to the West. The small cemetery and the fact that village was no more than twice as big as the excavated area, indicate that D162 is part of the RIA expansion – a new village situated 750 m East of present-day Säby, 750 m North of present-day Villinge, and 750 m South of present-day Kumla – that is, a new village fitted into an existing older structure of settlements with each their cemeteries.

Let us turn to the contemporary settlements: D193 and D169.

Hennius fig 10On D193 there is one small main house, House 33, two large outhouses, H23 (phase 1) & 29, and no less than six small barns or sheds, H24-26, H28 and H30-32. House 27 is a pit house and a dwelling. Houses 33 & 29 are contemporary and may well have been a small 4th century farm and so may the rebuilt House 23 (phase 2) & the outhouse 28. These two farms represent the boom at a site originally characterized by outhouses. The small barn H32 is the last on the site, which it claims in a way similar to the first ones and the last barns at D162. House 32 represents decline and abandonment after the boom. In the late 6th century someone lived in the pit house H27. By that time, there were no other houses on the site.

It is characteristic of this settlement that it doesn’t contain more than perhaps one farm with agriculture, husbandry and cows in the byre. Nevertheless activities are more abundant in the 3rd and 4th century, mirroring the boom character in a modest way. If there had been no aftermath in the form of a pit house the site would have been abandoned already in the beginning of the 5th century.

The site D169 is a sparsely settled area with a series of mostly small main houses, outhouses and sheds. If we follow the 14C dates the site was settled from the 2nd to the 4th century CE and around 300 CE there were probably two households on the site rather than one. Although there were no doubt farms with a multi-functional main house and a barn or shed, there must also have been times when households didn’t have cow sheds. The settlement seems to have been abandoned in the beginning of the 5th century.

D169 and D193 are unregulated peripheral settlements without cemeteries – settlements in which the ruins of earlier houses, rather than plots, guided those who were about to build themselves a new house. Since their farms and abodes were temporary, those who lived at these sites were not the landowning farmers of an autonomous village.


Hennius fig 11The sites D168 and D180, lastly, represent each their short period of farm life in an unregulated settlement. D168 is an old site with sporadic EIA presence, but no farm until the late 5th and early 6th century — i.e. a farm with no part in the RIA expansion. D180 is a younger site and the farm belongs to the central and later part of the 6th century. The youngest, at D168, is dated bp 1540, at D180 the date is bp 1490. Since a house lifetime, given maintenance, of c. 60 years is not unlikely D180 in all probability replaces D168 c. 100 meter to the north. Judging from the 14C dates, the buildings and their layout, the two farms replace each other, one being the first the other the second phase of one and the same estate. The farm is moved during or after the dust veil. D168/D180 is in other words situated before and after this event. The layout of the phases is very similar; it tells us that a farm could consist of a main house, 80 foot or more, and a 40 foot outhouse, which may also have housed a secondary household. The outhouse was sometimes rebuilt and extended. Because the farm consists of a main house and a rather large outhouse it is among the largest ones in the area. Its layout is standard and its farmyard situated between the two parallel farm houses in a relatively orderly way. When all other settlement sites have already been closed down or abandoned in the first part of the 6th century or indeed settled by the odd pit house dweller, D168/180 stands out as an investment in a new farm on a site that has not hitherto been used as a farmstead. D168/180 is going against the stream during at least three or four generations before it disappears.

 (3) The character of the settlements. D162 is the only village in the area. It developed a formal structure with plots. It was reduced and perhaps abandoned in the first part of the 6th century. It didn’t survive into historical times and may well be part of the RIA expansion that came to an end in the Pre Carolingian Iron Age (PCIA). A169 and A193 are marginal settlements abandoned in the 4th-5th century except for the solitary 6-7th c. pit house on D193. During the boom there are one or two small farms on D169 and D193 respectively. It would seem reasonable to suggest a connection between the village D162 and the peripheral settlement at D193. D169 may have linked in with a settlement further North, such as present-day Kumla.

The farm D168/D180 is probably not linked to the village D162 and although technically speaking peripheral, e.g. to present-day Säby, the farm was a full farm with a large outhouse and a multi-functional main house, but similar to homesteads and farms at D169 and D193, D168/180 occupied one of the temporary places in the larger Säby area, albeit for several generation. It seems not to be associated with any graves.

In a chapter on the osteological material Ylva Bäckström discusses the species and not least the balance between bones related to butchering and bones related to cooking – in percent this butchering/cooking relation is expected to be 60:40. In a settlement with more bones than expected related to butchering not all food was prepared at the settlement. Similarly, if bones related to cooking dominate, part of the food was imported to the settlement. In a sufficiently large economic area the balance between the two bone materials will nevertheless be 60:40. Turning to the village D162 there is a small surplus of offal, which means that some meat (mutton and pork) leaves the settlement. This surplus is matched by a small deficit at D193. Adding D162 and D193 to eachother the expected balance 60:40 meets the eye and we may suggest that the production at D162 matches that of D193. The bone material, as well as the layout of the settlements therefore suggest that D193 is a dependent, somewhat specialized, settlement engaged in husbandry linking in with the more central village D162 and its more complete subsistence system. Animals from D193 may thus have been brought to D162 to be butchered or indeed driven off to other parts of the economic system. In return the village D162 would to a certain degree have supported the peripheral settlement D193 – the point being that those living at D193 invested part of their work time in the economy of D162.

Initially, the well-planned farm D168/D180 suggests a subsistence economy or an economy similar to that of the village D162. None the less, the balance between offal and cut of meat is grossly distorted – 80:20 rather than 60:40. The farm D168/180 is in other words involved in exploiting the natural resource of its environment, allowing it to export half of its production of pork, beef and mutton from its immediate surroundings. Since we cannot expect a free market to have existed in the 7th c. CE, we must conclude that D168/180 was strongly tied to a not too distant settlement that needed a surplus, either because this settlement itself had a central position or because it was involved in a network around a more central place with a wish or duty to support people. Since manorial or central farms are characteristic of the PCIA (there are several candidates in the area) it is only reasonable to suppose that cattle farms such as D168/180 existed. Its economic raison d’être can be seen as a consequence of the fact that the other settlements between Säby, Villinge and Kumla were phased out and abandoned.


To sum up: When the isostatic uplift (the rise of land masses once depressed by the weight of the ice sheet) allowed people to settle permanently and farm the greater Säby area, the older part of the landscape with a few temporary Late Bronze Age (LBA) and EIA settlements, could be permanently settled. This happened in the 2nd c. CE when the area was settled with at new village and some marginal settlements. This expansion continued until the 5th century when the marginal settlements disappeared and the village was certainly reduced and probably abandoned. It stands to reason that this reduction could have been caused by overgrazing. Although the marginal settlements were abandoned the plots in the village D162, and the secondary site D193, continued for a while to be built on with an outhouse thus indicating that the estates and the D193 site were laid claim to, although there were no longer any farmsteads. In the first part of the 6th century even the outhouses disappeared and it is difficult not to see a connection between this abandonment and the general shortage of food and fodder created by the dust veil. The dust veil came in the end of a century in which a large number of farmsteads were abandoned. In the Säby area no LRIA estate disappeared, but the outhouses at D162 and D193 came to an end suggesting that nobody rebuilt them when necessary. It seems unlikely that everyone once living on the farms that were abandoned went on to live in the farms that were not abandoned. In the Säby area at least five households disappeared, none continued and only one new farm was created, probably after a hiatus.

The settlement situation at Säby is not unique, but in this particular area the settlement expansion is late and so is its decline. Compared to other well-dates cases e.g. in the Oslo fjord area where most farms are solitary, it become apparent that if the permanent farms of the expansion start early – often after centuries of sporadic presence – they will also come to an early end. If they begin with Common Era the end will come in the 4th century. Similarly, a late beginning results in a late end in the 5th c. CE. This is the prime, albeit indirect reason why abandonment should be seen in connection with the malfunction of the RIA subsistence system – its ecological weakness being its dependency on husbandry in combination with a temptation to graze too many heads and allow too many households to be established.

Since modern exploitation isn’t likely to target present-day farms we seldom excavate the settlements that survived 536 CE and continued to be settled, but it does happen – Old Uppsala being case in point.

In Säby a new farm was established in the 5th century as indeed the only one in the abandoned area. This farm survives the dust veil for a couple of generations and judging from the osteological material its economy was based on livestock that would have survived grazing the area, which 200 years earlier was occupied by five or perhaps ten farms. This can of course be done as long as the population pressure is low and as long as ownership is not disputed. Since the 6-7th c. settlement at D193 was a pit house and not a farm such as D180, it stands to reason that land could not be occupied by anyone in need of a farm. Despite the demographic problems, the landless continued to be landless, some even began to live in pit houses. The marginal areas were not resettled in order to create self-subsiding households, but rather households that fitted into a larger and more hierarchic economic structure and settlement pattern. The pit house at D193 and the farm D168/180 are the typical outcome of the agrarian crisis of the 4th and 5th century and its aftermath the lost summers and the decrease in population during the late 530s and early 540s.

There is a fair chance that Säby is typical of the settlement change from the mid-second to the early seventh century CE. In that case the crisis, enhanced by crop failure and famine in the sixth century, didn’t result in an agricultural reform, it just meant that for a few more generations a reduced population could continue overgrazing and exploitation. The nadir of the agricultural crisis therefore seems to be the seventh century CE.

The economic crisis in mid-millennium Scandinavia shares some structural similarities with the crisis in the centuries around 1400 CE (cf. OtRR 4 March, 2113). It was an agrarian crisis topped by a demographic catastrophe accentuating social stratification and followed by a period of political warfare. Society was marked by its inability to change its mode of subsistence in order to accommodate a larger population and also by its inability to balance the size of its population against its system of subsistence. The decrease in population enhanced social stratification and brought about political power struggle rather than a new social order. The Scandinavian society even resisted urbanization except in its southwestern corner where Ribe was the exception to prove the rule. When migrating Scandinavians become visible in the annals of the 8th century they are the result of a population surplus in the unreformed society. Leaving the old society, they favoured a new one with an economy built on interacting rural, urban and mercantile sectors – they are the aggressive modernists of their day and age.


(1) Information on this is found at

(2) See
and goto page 273

(3) This seventh settlement was reported by Göthberg 2007: Arkeologisk undersökning. Kumla- bosättning och djurhållning under äldre järnålder. Fornlämning 169, Danmarks socken, Uppland. Hans Göthberg. Rapport 2007:15 (3347 kb). Cf.

(4) See The Early Iroin Age in South Scandinavia pages 245ff.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: