This week On the Reading Rest I have an article from 1956, no less, because I glanced through a catalogue from a recent exhibition of prehistoric gold objects from Niedersachsen in Germany and saw a picture of the hoard from Ellerbeck exhibited in Niedersächsisches Landesmuseums in Hannover. It may be looked up at: .

There was nothing definitely wrong with the text in the catalogue or on the net, but I thought they were both so cautious that they missed a number of facts in a probable micro history. This led me to the open stacks in the cellars of the nearest university library and a fact-filled article from 1956:

Berghaus, Peter. 1956. Der römischen Goldmünzenfund von Ellerbeck, Lkr. Osnabrück. Die Kunde, N F 7:1/2, pp 30-40.

There are three roots to the micro history of the hoard from Ellerbeck in Osnabrück, the character of the coins, the composition of the hoard and its time and place.

The coins in the Ellerbeck hoard in time and space. The diagram to the left shows the first and often only year in which the coins were minted. The lines go to the different mints and the dotted lines indicate that the imitations were propably minted somwhere in Gaul.

The hoard is made up of 25 gold coins; 22 official Roman solidi and 3 contemporary and probably equally valuable imitations. Two of the imitations are almost flawless, in mint condition, and stamp identical. They are imitations of a type of coins minted for Valentiniaus represented in the hoard by a another pair of stamp identical coins, again in mint condition. These coins, the youngest ones in the hoard, were minted in the mid 360s AD.

In the 4th century AD, Ellerbeck is situated outside the declining West Roman Empire in a region where people had access to Roman gold coins. Still, it is very unlikely that anyone circulating coins in the Osnabrück area could by chance get hold of these two pairs of stamp identical coins. Gold coins are produced at certain occasions to pay for expenses and they come in many series of identical coins. Since these series are relatively short, circulation will rapidly disperse the coins on so many hands that only few have two or more identical ones. In Ellerbeck, therefore, the stamp identities and the mint condition among the coins from the 360s indicate that these coins were never widely circulated. Instead they went more or less directly, i.e. through a few transactions only, from the official and unofficial mints to their last owner.

The oldest coin and five others minted around 350 AD, were pierced with an awl just above the emperor’s head, i.e. pierced from the obverse with a technique that doesn’t deprive the coin of any weight. When you add a loop to a coin it becomes a pendant and a charm, but since piercing, apart from creating a semi loop, is also a form of testing (is the coins solid gold?), pierced coins are not necessarily pendants. Nevertheless piercing indicates some sort of circulation characterizing the coins from the 350s.

The money box and coins from Ellerbeck. Based on a photo from Landesmuseum Hannover.

These 25 coins were found in a cylindrical money box, approximately 5 by 4.3 cm. Its bottom was a hinged lid and just below the top there was a slit . Since the Ellerbeck piggy bank would have contained more than 25 solidi, we may suspect that instead of rattling around in the box, the coins were stored in small a parcel. A black powder that fell out together with the coins when the farmer, who found the box, destroyed the slit to get out the contents, may have been the remains of the wrapper.

Formally speaking, the coins were minted between 346 and 364+ AD (the ‘+’ indicates that the imitations were perhaps, but not necessarily minted a few years later). Because all the coins were produced during very short periods a diagram of the first year of their minting gives a good impression of the chronological composition of the hoard: in the mid-360s a small series of mostly uncirculated coins were added to a large series of circulated coins from the early 350s (346-47; 350-3 AD).

The total weight of the hoard is 108.55 gr or approximately 1/3 of a libra, the Roman pound [1]. In theory, when it comes to Roman pounds, 1/3 of the Canonical libra equals 327.45/3 = 109.15 gr, while 1/3 of the Logarikē libra, i.e. the ‘counting pound’ referred to in matters of economy and coinage equals 322.24/3 = 107.41 gr. In theory, therefore, 72 solidi make up a counting pound, but in the present hoard the average coin weight isn’t up to that: instead of 4.475 gr, in theory what we expect, the average weight is no more than 4.342 gr. For being a hoard of 25 coins with little time depth this average weight is exceptionally low and that explains why the total weight is close to and a little above 1/3 of a Logarikē libra. Generally speaking, in Barbaricum one third of a pound cannot be composed of 24 coins because among barbarians north of Limes, the Roman border, only a few coins weigh 4.475 gr or more and that explains why there are 25 coins in the hoard, instead of the theoretical 24. The Ellerbeck hoard is the result of a wish to hoard a money box with 1/3 of a pound in minted gold, rather than just gold. Take away any one of its 25 coins and the collection will be underweight.

The emperor’s bust on the obverse is a sign of the purity of the metal, and makes solidi prestigious objects. The money box too indicates status, because only outstanding actors on the economic scene need a handy way to box and carry around a fortune. Systematically piercing coins above the emperor’s bust sends a message too signaling barbarian knowhow, suspicion and caution when dealing with solidi.

Given all this, the 25th coin in the Ellerbeck hoard indicates that in order to be sure of having composed 1/3 of a pound of high quality gold the collector added the 25th coin. If he had striven to form 1/3 of a Canonical libra, then the 25th coin wouldn’t have helped. His 108.55 gr would still have been 0.6 gr below the 109.15 gr needed. Since the average weight is very low, and below 4.366 gr that would have matched the Canonical libra although it is still a low average weight in a barbarian hoard, we can be quite sure that the collector wanted to compose a hoard of 1/3 of a Logarikē libra by means of 25 coins.

Since the weight of most of the coins, which the collector had access to, were very close to the correct weight (i.e. 1/72 of Logarikē libra or 4.475 gr), the 25th coin threatened to create overweight. To compensate for this risk he put in the three exceptionally lightweight solidi. In the diagram they stand out as dropouts weighing 3.69, 3.78 and 4.0 gr respectively. Together these coins save the collector 1.929 gr of pure gold presently worth 10.5 $. With 1.93 gr of gold one could also have made 8 small gold foils, fragile, but loaded with ideological contents – at least in upper-class Scandinavia.

Figure foils, guldgubbar is Scandinavian terms, are popular pictures of human beings and some creatures. They are not easy to interpret.

If, for the sake of the argument, we disregard the three lightweight coins and recalculate the weight of the hoard on the basis of the 22 heavier coins, 25 of those would have weighed 110.32 gr. This overweight could perhaps have been tolerated had it been a matter of composing 1/3 of a Canonical libra, but since that is not the case, the hoard would have been 2.7 gr to heavy and we may venture to say that a 2.7 gr difference, at the expense of the collector, is too much. This difference is in other words unacceptably wasteful among the rich in 4th century Lower Saxony. Although we weren’t invited to their cocktail parties, it nevertheless feels good to know that had we been invited, we would now have been able to say something perfectly acceptable concerning ‘overweight in money boxes’ and ‘Roman currency’ (on the ‘perfectly acceptable’, see Gustave Flaubert: Dictionnaire des idées reςues) [2].

One may of course clip a coin to bring down the weight, but putting in 25 coins instead, just to demonstrated that we are fully aware of the character of the Roman coinage and its problems, is also a sign of quality, and the acceptance of a 1 per cent overweight in a sum as ‘small’ as 1/3 of a pound is not alarming given the fact that the gold is coined. A wealthy man, to quote the cocktail party, can always find the lightweight coins he needs to adjust the total weight. Although in reality access to a range of different solidi is pivotal for the composition, the impression it gives will balance skill against inability, perfection against largess or sense against fixed ideas eventually highlighting the collector and his ability to solve a problem.

Twelve of the coins are struck for the short-lived usurper Magnentius and his caesar Decentius (350-353 AD). They had their rapidly shrinking power base in Gaul and spent the whole of their three years on the road to defeat and suicide campaigning. They would have minted as much as they could to cover their expenses, and for a short time in the northwestern part of the Empire, notably Northern Italy and Gaul, their money would have been abundant.

When Constantius defeated Magnentius and Decentius (353) he sent Claudius Silvanus as his magister militum to Cologne to check the Germans. This Silvanus did with bribes funded by the taxes he collected, and taxes, as it happened, had to be payed in minted gold. Spending taxes was a short-run solution and his only option because Trier, where the mint was situated, closed its gates to him. In 355 AD he revolted and was promptly killed by an envoy from Constantius whose agents knew of the revolt in advance. The envoy was the loyal general Ursicinus and he simply succeeded Silvanus as magister militum. Ammianus Marcellinus, who else, has the whole story.

Two thirds of the money in the Ellerbeck hoard link-in with coins from Trier in the early 350s and with Magnentius, Dencentius and Constantius, in that order. It seems plausible that this part of the hoard mirrors payments and bribes to Germans in the 350s. In the process these benefactors seem to have acquired some knowledge concerning the weight of Roman coins as well as the political games of the deteriorating Western empire. In the end, they were able to feed their piggy banks and put aside or part with well-composed round sums of money, attesting to wealth as well as cunning calculation.

Generally speaking, the composition of the hoard mirrored the coins that would have been available in this part of the world when sometime in the latter part of 360s the collection and its peculiarities were put together, wrapped up and boxed. Probably more than 25 coins were looked upon and read, while figuring out their weight and searching for low-weight coins. When the task was finished, the collector would have looked back upon a short period in the history of the declining Roman Empire and his own life. He was probably satisfied with his solution alluding to the 350s and happy to get rid of the usurper’s coinage and of lightweight and pierced coins as well as some modern imitations that were difficult to circulate. Imitations may always be doubted because, unlike the emperor’s picture on his own coinage, the picture on an imitation doesn’t guarantee that the coin is pure gold. Heavy coins are obviously a drawback if we try to make 25 coins equal the official weight of 24, but perhaps also a sign of good measure, needed when you pass on three extremely low-weight coins.  Imitations, not surprisingly, entice with overweight coins and let in the low-weight dropouts.

The Ellerbeck collection is hardly a random sample that makes up someone’s 25-coin fortune. Instead, it seems to be a round sum cunningly composed to become a gift to god or man, a payment, a dowry, a due, a cro etc.  In the chaotic monetary situation of the 360s, the composition highlights both a wish to pay nomore than 1/3 of a pound in minted gold, and a wish to get rid of coins that were difficult to circulate on a nominal basis.  The result was a money box of assorted coins. Let’s hope it satisfied the receiving end.

On the Reading Rest 12 December 2011, a similar preoccupation with solidi was touched upon. In a letter from the 590s Gregory the Great pointed out to his vice deacon Peter, that Peter exerted too many solidi per pound when he collected taxes and dues by weight in solidi. Peter demanded 73 and ½ solidus per pound instead of 72 and Gregory understood that 73 1/2 was far too much to compensate for the slightly underweight coins of the 590s. Peter’s scam consisted in cheating on the peasants when, referring to the underweight of coins, he exerted too many per pound. This allowed him to compose payments similar to the sum in the Ellerbeck money box and keep the change for himself. Gregory wasn’t pleased.

We have no reason to think that the man who composed the Ellerbeck hoard was dishonest, he just knew his business and showed it in the composition of the sum. He might have told someone about his toil or given the box to someone who would understand, but probably he just composed his sum, his complex material narrative, and parted with it. Be this as it may, everything was lost until in 1933 the box was unearthed by Landwirt Schürmann. And little by little, not least owing to Peter Berghaus’ article, the money box started once again to become history.

[1] Although in Romanian, there is a handy dictionary at:
(keyword ‘Libră’).

[2] More on Flaubert’s Dictionnaire des idées reςues—Dictionary of Received Ideas, at: for some translated examples

or download at


This week On the Reading Rest I have an article that I read some years ago when I was about to visit the Mosfell excavations. Now I read it again because I am looking for something that I glanced through then, but came to think of as important only later.

Byock, J. et al. 2005. Jesse Byock, Philip Walker, Jon Erlandson, Per Holck, Davide Zori, Magnús Guðmundsson and Mark Tveskov: A Viking-Age Valley in Iceland: The Mosfell Archaeological Project. Medieval Archaeology Vol XLIX. Pp. 195-218.

I like the article because it describes a project which links written and archaeological sources, historical arhaeology at its best, and because it’s open to any prolific method available. It concerns a long-term research project focusing on an Icelandic valley settled in the Landmám period, i.e. the 60 year era around AD 900 when a wave of Scandinavian settlers colonized Iceland. Like many archaeological research projects it is in essence an exploratory endeavour applying all kinds of methods, eventually finding its own style of research – proving this discovering that. The project is a smorgasbord of interesting things we can do when we decide to investigate an Icelandic valley. Most, but not all the methods are applied systematically, their application nevertheless, provides good examples of their usefulness.

Landnám Mosfell was moved slightly further into the valley and took the name Mosfell with it.

I suspect that The Mosfell Archaeological Project (MAP) must be a wonderful project to work in as a student, because rather than being a somewhat pointless field school, it probably tends to become a miniature archaeological university allowing long-term student engagement to create small informal scientific careers. This university-like framework happens to create itself now and again in long-term archaeological projects working faraway from any university in a place where there is little to talk about but aims, methods and interpretations. The Vallhagar project on Gotland just after the World War II was a case in point inasmuch as it educated a large part of the first post-war generation of Scandinavian archaeologists. But it wasn’t ‘Unique!’, a great number of senior archaeologists will refer to their formative dig after a beer or two.

The homepage is worth visiting

The archaeology of Landmán and Free State Iceland (up and until AD 1262) is breath-taking in itself, but it also informs us on a number of points when it comes to the archaeology of the Scandinavian Iron Age. This has do to do with a point which the article makes in the following way:

From modern socio-historical and anthropological viewpoints, early Iceland is a social laboratorynote, … …. The society that evolved during the Viking Age on this large island … … avoided the establishment of most official hierarchies without going so far as to create egalitarianism. The politically active population consisted mostly of free land-owning farmersnote. These latter kept some slaves, but free workers and cotters were more common. Leadership was in the hands of small-scale Scandinavian chieftains known as goðar. (p. 203)

One can put it in this way without being wrong, the notes support the description and prove it to be widespread, but instead of seeing the Landnám era as the birth of a nation, emphasizing frontier-spirited settlers, we may speak of the conception of a Nation when the old fools of the Late Pre Carolingian Iron Age (AD 550-750), who couldn’t thrive in Carilingian Iron Age (CIA) Scandinavia, meet Virgin Land unable to defend herself against old-fashioned Iron Age courtship enticing with the right to form a household of your own in a house or on a farm of your own, with a say in the assembly – almost a model New Iron Age society based on traditional values. It can in other words be argued that the Landmám society on Iceland was a retake of the Scandinavian Iron Age, however with a strong Pre Carolingian Iron Age emphasis on leadership, the góði being the epitome of the Chieftain [1].

Góði is a tricky word. It derives its meaning from goð which means ‘god’, i.e. one of those who we should worship because we can trust them and seek their help. The word was later used to designate God or Christ, who would also be called an allvaldr – ‘all power’. Obviously the ruler qualities of a góði come closer to those of a fró (originally *frawjaz), the Iron Age ruler who helps people to better their conditions freeing them from their sufferings, than a valdr, who is simply the cause of ‘power’ as a social phenomenon.

This hint at the divine and caretaking leader matches the fact that the word góði may also mean that which is good, derived from the adjective góðr. The noun góði means benevolence too and in practice therefore it can be said that ‘a góði does the góði’ because he is benevolent or góðan. There is in other word a distinct echo of the Iron Age leader and his divine connections in the term, not least while the two words are heteronyms, i.e. spelled and pronounced the same way, but with different meanings. The pun qualities of the words góði/góði must be taken into consideration because we cannot help hearing them.

The characteristics of the Landnám suggest that we can use some of the cultural phenomena found in Iceland as model Iron Age echoes in periods when cultural expressions, e.g. in South Scandinavia, have already diverted from Iron Age ideals.

And this is where Mosfell enters the scene as an example.

During the CIA, halls settings are sometimes characterized by a small side building in a fenced area connected to the main building. It has been suggested that these side buildings are temples and that they represent a return to Early Irons Age ideas of the hall as a ceremonial freestanding building next to the main house of the magnate farm [2].

When this renaissance phenomenon, one of several typical of the CIA, starts to occur we may suspect that part of the cultic functions that could be housed in the large Pre Carolingian Iron Age halls is moved out of these halls in the Carolingian Iron Age.

It seems that Mosfell is indeed a hall farm in the Norwegian style and thus comparable e.g. to Borg in Lofoten albeit smaller. But it has a side house and thus a layout that reminds one of its South Scandinavian counterparts. The remains of the first side building, next to the pagan founders grave at Mosfell, have almost been destroyed. On the plan they are mapped as the remains of an earthen floor only, but the second building is a well-preserved church and churchyard. The situation and the development at Mosfell therefore indicates that it was possible (at least among the higher echelons of society) to redefine the small temple next to CIA halls as indeed a church.

In South Scandinavia, with its dense population and rapid march towards a Christian parish organization, this development is difficult to see either because early the churches take over the magnate farm in order to become the church of a local community (e.g. Jelling, Tamdrup or Lisbjerg) or because they were built to meet a demographic need (e.g. Sebbersund or Veøy). In Iceland and at Mosfell, parish organization is unimportant and acculturation therefore smooth and in harmony with a number of pagan ideological concepts. Consequently comparing Iceland to South Scandinavia suggests that, primarily, South Scandinavian churches are expressions of the Medieval rather than the Christian. In comparison with the pilitical and social transition, the ideological change was smooth and probably not more radical than earlier prehistoric transitions.


[1] Góði and other words mentioned in the following can be looked up in J. Fritzners Ordbok:  

[2] This development is outlined in The Early Iron Age of South Scandinavia pp 230 ff.
On side buildings see also
Jørgensen, Lars. 2003. Market and mannor at Lake Tissø, 6th to 11th century: A survey of Danish productive sites. In: Tim Pestell and Katharina Ulmscheider (eds). Markets in Early Medieval Europe: trading and productive sites, 650?850. Oxford.