Rivets and roves

During the time between the Hjortspring Boat (circa 350 B.C.), on which the planks were sewn together with lime bast cord, and the Nydam Boat (circa 350 A.D.), iron working evolved enough to allow blacksmiths to make the rivits efficiently and at a relatively low cost. This evolution, however, does not mean every boat after Nydam was built with iron rivets, the sewing technique lived on alongside rivets for many hundred years.

It requires a large portion of iron to produce rivets enough for even a small boat like Gislinge, which in its finished state will contain between 450 and 500. Every rivet weighs around 10 grams so we require at least 4.5 kg of raw iron which in turn takes about 45 kg of bog iron to produce.

The Nydam Boat represents the evolutionary step just “below” Viking longboats, carrying up to 45 warriors 36 of which rowed, but was still without a sail. It would take another apx. 350 years before the Scandinavians mastered that technology. The earliest Scandinavian ship found carrying mast and rig is the Oseberg Ship from 820 A.D.    

To fit the boards together, a hole is drilled through both keel and plank using a spoon shaped drill. The hole is preferably just a little smaller in diameter than the rivet. The rivet is then fitted through the hole and the tip is cut of which leaves a blunt end that can be hammered (riveted) over/around the rowe to prevent the rivet from being pulled back out. This procedure is followed every time another one of the four planks is fitted.

As part of the Gislinge Boat Project, a subproject concentrating on extracting iron from bog iron and turning it into quality rivets and rowes for the boat will take place at the museum in Roskilde between July 27th and August 2nd.  Through five separate, controlled experiments archaeologist and ironworker Mads Jylov and blacksmith Jonas Bigler will carefully follow and document the process. One of the important elements will be to investigate just how many times you need to fold a piece of iron to get the needed strength for a rivet.  

Bog ore experiment

Below, you can read Mads Jylov's text about the project's bog ore experiment:


The aim of the experiment is to conduct four smelts, producing four iron bloom which match the type of bloom known from Danish archaeological material. The intention is to produce bloom which has a similar chemical composition and density as the archaeological material. The bloom’s density is an indicator its porosity and the ratio of iron and slag. This ratio is of vital importance in terms of further research with the material generated by the experiment. The bloom will be cleaned and forged to rivets, which will be used in the Gislinge boat.


Viking Age and Medieval smelting furnaces are pretty much unknown within Danish archaeological material. The finds are limited to a handful of sites and are often only identified as areas with concentrations of slag. Remains of the actual furnace are essentially non-existent. From the surviving material, the following can be concluded:

1) Furnaces were placed over a small hollow, ca. 40-50cm in diameter and 20cm deep.

2) The hollow opens out into a pit on one side.

3) The furnace walls are ca. 15-20cm thick at ground level and are made of sand-grogged clay.

4) The types of slag found indicates that tap-furnaces were in use.

Experiment design

For the experiment, a scaled-down version of slag-tapping furnace will be used. The furnace is 90cm high and has an inner diameter of 25cm. The air intake is 20mm in diameter and is positioned 60cm from the top of the furnace, i.e. the centre of the air intake is ca. 25cm above the furnace bottom. The furnace is placed over a circular hollow 25cm wide and ca. 30cm deep. The air intake is located within a 2cm thick plate of fire-proof material. This plate can be removed when the furnace is repaired.

During the experiment, a Frederiksen electronic bellows, calibrated to blow with a speed of 50km/hr and an air volume of ca. 459L/min, will be used.

The tuyere (the pipe leading from the bellows) has a diameter of 19mm and is positioned 4cm from the furnace’s air intake, angled 20 degrees downwards towards the base of the furnace.

Mixed charcoal of deciduous wood with a max. size of 5 x 5cm us ised.

All charges are composed of 1kg charcoal and 1kg roasted bog ore.

The bog ore is from the Guldager area.


The furnace base is packed with a 20cm thick layer of ash and small charcoal.

The furnace is heated up using wood with a max diameter of 15cm. The furnace is heated up over ca. an hour using its own natural updraught.

The furnace is then filled to the top with charcoal and the bellows come into use.

The furnace is charged as space becomes available.

From here, the furnace is charged ca. every ten minutes.

When the slag reaches up to the base of the air intake, it is tapped by making a hole under the bellows plate. When the air from the bellows begins to come out through this hole, it is then sealed and the slag begins to gather again.

After all charged have been filled, the furnace is allowed to burn down to ca. 10cm above the air intake.

The remaining slag is tapped and the area under the bellows plate is dug out.

The bloom sits fast under the air intake and extends ca. 20 cm into the furnace.

The bloom is hammered together on a wooden block, directly from the furnace.

The experiment is finished and material collection and registration begins.