In the ninth century the Dutch coastal areas in the present Netherlands were defended by Danish rulers against invading pirates on behalf of the Carolingian kings. The Franks themselves were hardly able to protect these regions 'because they were inaccessible for an army on account of countless courses of different streams and impenetrable swamps' (1). That is why they seeked cooperation with Danish exiles, who earlier showed their skill to fight in areas abounding in water, using longships of the Nordic type. However one can ask whether Frisian ships were more suitable in stead of their own longships during their long stay in Frisia. The latter, having a deep keel, were rather inconvenient to penetrate the Frisian mud-flats and coastal lagoons. Moreover it is questionable if they had sufficient ships at their disposal all the time. That is why the Danish rulers had to make use of the locally built waddenschip, the Frisian cog, a ship that was completely adapted to the local circumstances.

In the High Middle Ages the cog was a popular cargo vessel in the Frisian coastal waters. But finds of this ship, that can be dated before the twelfth century, are lacking. When we look for its Early Medieval roots we have to take other clues into consideration, as can be found in iconographical and etymological sources. Therefore we follow the trail of the Frisian merchants with their cogs to Scandinavia. Then we return with Danish exiles back to the Frisian coasts where they created a defence organization based on the local cogs.

The origin of the cog
First we have to ask ourselves if the cog, well-known in de Hanseatic area since the thirteenth century, did derive from a ship that was denoted as such in the ninth and tenth century. Although we do not have any finds of Early Medieval forerunners at our disposal, we have to presume that there must have been common characteristics with the Hanseatic cog, merely based on the name. A fast development, caused by the expanding trade in bulk goods, was the reason that little remained of the original nature of the small Frisian waddenschip, but the name cog sustained. The Frisians must have adapted a ship, that was originally of continental origin, to the specific circumstances of the shallow Frisian tidal waters. They must have used flat-bottomed ships which could navigate up the creeks and sink to the bottom at low tide for loading the ship. With such a ship merchants could sail to Denmark, well accessible through the tidal basin behind the barrier islands, called Wadden Sea.
Nordic ships were less suitable to navigate in the tidal waters of Western Jylland. For this purpose the Danes could have built ships of the Frisian type. For exactly wrecks of the earliest known cogs, from the twelfth century, were discovered at Kollerup and Kolding on Jylland.
In the western part of the Limfjord a closely related type still can be found nowadays, but disappeared from the Dutch waters in the 19th century. It is the so called kaag or kåg, with a carvel built bottom and slanting sides. In the Nordic shipbuilding tradition ships with carvel built bottoms are completely lacking (2). Cog-like ships without a keel are certainly not a Nordic development and must rather be considered to be a Frisian equivalent of the knarr, the well-known Nordic cargo-ship.
Early cogs rather seem to have developed from a Celto-Roman pram-tradition than that they were inspired by Nordic influence (3). Roman ships of the Zwammerdam-type, found in the Netherlands, had sides with overlapping strakes, secured with iron nails. The nails were bent twice, just like we can find at later cogs (4). Nails of Nordic vessels were on the contrary riveted, that is clenched by hammering out the point. The earliest known cogs outside Denmark were found in the Zuyder Zee polders in the Netherlands and are dated in the first half of the thirteenth century.
On coins from the Frankish emporia Dorestad and Quentovic, minted during the reign of Charlemagne and Louis the Pious, crescent-shaped ships are depicted. According to Ellmers we can recognize hulk-like vessels (5). Apparently the hulk was regarded as representative for those trading towns, this cargo ship must have been the most important 'money-maker'. None of the depicted ships on the coins from Dorestad and Quentovic displays any similarity with cog-like vessels. For them we must not search in Frisia or any place in the Frankish empire, but strange enough in Birka in Sweden. Among many burial goods a couple of dozen bracteats were found in graves from the ninth century (6). Apart from pure imitations, some specimen were inspired by the mentioned coins from Dorestad, but with completely different ships. Next to typical Nordic ships, with curved stems and a row of shields alongside, we can find another type with a flat bottom and high, steep stems. Mainly based on these characteristics Ellmers argued that we are dealing with Frisian cogs (7). These coines can be dated in the first decades of the ninth century (8). According to Ellmers (9) the depicted (proto-) cogs make it clear that the coining could not have taken place without Frisian influence and thus must be located in Haithabu (Hedeby) rather than Birka, because Haithabu was the main emporium in the border zone between Danes, Slaves, Frisians and Saxons. Here Frisian merchants had there own 'trading-quarter' ever since the eighth century (10). Therefore it is well possible that these coins were struck by Frisians. It is not likely that Danes should have depicted a Frisian ship on their coins, considering the special cultural meaning of Nordic shipbuilding. Everywhere in the Frisian and Lower Saxon coastal areas the toponym Kaag, Koog or compositions of these names frequently occur, originally meaning the coastal plane flooded by the tides. The relation with the cog (in Dutch: kogge), specially built to function in these tidal areas, is evident and indicates the Frisian origin of the cog. They must have been very suitable for the (coastal) route to Denmark and could very well function in the tidal harbours of Ribe and Hollingstedt. Through Haithabu the cog must have spread into the Baltic Sea and developed into the well-known Hanseatic ship. In the process many Nordic details, like the T-keel, were adapted (11). Because of its greater cargo capacity and its relatively simple construction it repressed the Nordic ship type.
Along the Blekinge coast, in the south of Sweden, various toponyms occur that refer to cogs (12). It is not known if they already existed before the Hanseatic period. However this is certainly the case with the Kugghamn (13), a bay near the trading-place of Birka. This bay must have been named after the cogs of the Frisian merchants, who frequently aimed at this harbour.

Naval cog organization
There is a hiatus of several centuries between the moment that the Haithabu coins were minted and the age of the oldest excavated cogs. However from the period in between there are some historical sources on the subject. For we have an indirect clue of the occurrence of the cog in the ninth century. We can deduce this from the way in which the general military duty (expeditio) locally was organized in the High Middle Ages. This was arranged in a different way in the county of Holland (the coastal territory in the western part of the Netherlands) than in other places. Holland was divided into naval districts, which were called koggen (cogs). In each district the inhabitants had the obligation to provide one heerkogge (war-cog), sufficiently manned and equipped. However this duty was not introduced by the counts of Holland, because references to these regulations can be found before their appearance. Already in a list of properties of the Church of Utrecht, drawn up in the late ninth or early tenth century, a tribute was mentioned which is called cogsculd (14). That was also registered in two charters dated in the tenth century (15). We may assume that we are dealing in these cases with a tax related to the cog duty.
The phenomenon of naval districts is also well-known outside Frisia. For there is a remarkable resemblance between the cog-districts in Holland and similar areas in Scandinavia.
According to the Heimskringla, the chronicle of the kings of Norway, King Håkon Haraldsson ordered to divide the land in skipreiður (ship-districts), reaching inland from the sea 'as far as the salmon swims upon the rivers' (16). All over Scandinavia we can find those districts, they were divided into smaller units, named after rowers. Each unit had to deliver one rower. And this reference is just what we also find in Holland. Because here the cog-districts were divided into some 20 to 30 smaller units, which were called riemen (oars), each responsible for the contribution of one able-bodied man, or armed rower.
It is well conceivable that under constant pressure of piracy some sort of organized naval coast guard already was created in the migration period, with a fleet of free Germanic peasants patrolling in the summer season. This organization can not have shown much variation along the Frisian and Danish coasts, if one considers the tight historical ties between the Frisians and the Danes. It is well possible that an organization was established, that later was employed by different rulers for their own purposes. Rulers, while they centralized their power, took over the defensive organization of the peasants (militia) and amplified it to a general military duty (expeditio). That happened in Scandinavia where powerful kings spread their power over a large area (17). We may assume that the leidang (leiðangr) was introduced in this way by local rulers and later taken over by the kings who were supported by these local rulers as an expeditio.

Danish rulers in Frisia
Finally Danish rulers and not the Carolingian kings forced the Frisians to use their cogs, by shaping the existing militia for their purposes. Apparently they must have had sufficient authority to impose this obligation to the local population. In Frisia one Danish ruler, Hrœrekr (Roric), was particularly active for about 25 years, except for some interruptions. He most likely was responsible for moulding the existing coastal defence into an organization to suit his own aims.
Initially Hrœrekr must have entered the Frisian waters with Nordic longships. But these ships were not very suitable for these waters. Although this kind of ships had only a shallow draught, there was a considerable chance that they would run aground in shallow water with their prominent keel. Indeed the recorded targets of Viking raids in the Low Countries were all accessible through the big rivers.
As defender of the Frisian coastal area, but especially as ruler over the Frisian population, Hrœrekr also had to penetrate those places that were only accessible through shallow lagoons and tidal channels. For that purpose he had to look for a type of ship that was better adapted to the Frisian waters than his longships. He anyway needed new ships, for it will not have lasted long before a part of his retinue of exiles and outlaws stood out to sea seeking glory and plunder. And while doing this they must have taken the required ships. Therefore it is presumable that Hrœrekr employed local Frisian shipwrights. They could build a type of ship in their own tradition, the cog, that was fully adapted to the circumstances of the shallow Frisian tidal waters (18).
After a the majority of his men left for England, Hrœrekr possibly kept some longships by means of strategic backup. He could not rely on the cog-organization any more. Surely after the inhabitants who had to serve cog-duty - they were called Cokingi (19) - revolted against this widely interpreted defence duty.
In spite of the fall of the Danish lordship and the reduction of the central Frankish authority the cog alliances survived and stayed a defensive organization for a long time after. Only with the introduction of the counts of Holland the ship duty was annexed and transformed into a general military duty (expeditio).
The presence of Norsemen did not left a trace behind of naval districts outside Frisia, also not in regions like the English Danelaw and Normandy. In England before the Norman invasion the maritime duty was admittedly organized in districts. But differently from Frisia those districts were based on unities, the hides (20), with a more common administrative character. A general military duty, with a milition of free farmers, did occur much earlier than the time of the Danish colonization. It appears that the maritime duty was fitted in the already existing model. Moreover all duties based on hides only occurred outside the Danelaw (21).
In Normandy several laws have come down to us, but a naval duty based on a Nordic model is unknown. The magnates - the tenants and the clergymen of William the Conqueror - provided the necessary ships and armed warriors for the Battle of Hastings in a feudal manner. Not in England or Normandy, without a maritime background worth mentioning in the Early Middle Ages, but in Frisia where the population, like in Scandinavia, for a long time was familiar with navigation, a ship duty could be fitted in successfully. In an area abounding in water like Frisia a maritime defence was the most logic. Danish rulers could employ the Frisian cog to their benefit. It is no coincident that the organization was named after this ship.

Notes
(1) Regino, see Rau, 1960, 268-270
(2) Crumlin-Pedersen, 1965
(3) Bill, 2003, 11
(4) Bill, 2003, 12
(5) Ellmers, 1972, 59
(6) Malmer, 1966, 60
(7) Ellmers, 1972, 64ff.
(8) Malmer, 1966, 60
(9) Ellmers, 1972, 72
(10) The so-called Südsiedlung, Lebecq, 86-88
(11) Ellmers, 1972, 72-73
(12) It concerns the toponyms Koggaskärsbåde, Kuggaskär, Kuggaviken, Kuggeboda, Kuggen and Kuggskär
(13) Arbman, 1961, 40
(14) in Almere regalis decima census qui vocatur cogsculd, in the list of properties of the church of Saint Martins in Utrecht, see Muller & Bouman, 1920, no. 49
(15) Muller & Bouman, 1920, no. 111, from 948 and no. 113, probably from 949
(16) Hákonar sögu góða (the Saga of Håkon the Good), see Jóhnsson, 1911, 82-83
(17) Bjørkvik, 1970
(18) The finding of a different type of rivets according to the local tradition in Viking-settlements in Russia also makes the use of locally constructed ships possible (Bill, 1997, 200)
(19) AB 867, see Rau, 1958, 166
(20) A hide is a certain amount of arable land.
(21) Cam, 1944, 94