En skvatmølle i Ljørring


  • Axel Steensberg




Ljørring, Horizontal Mill, skvatmølle, middle ages, middelalder, Løvenå


A Horizontal Mill at Ljørring, Jutland.

Horizontal water-mills have been in use in Jutland since the beginning of the Christian era 2). But the one here described shows so close a connection with the timber building technique of middle-age Denmark (the log-house or "bulhus") that it can most probably be dated to the Middle Ages. Unlike the horizontal mills of the Scandinavian peninsula and of the Faroe Islands this mill was designed for a quite insignificant fall of water, and in that, as in several other respects, it recalls the water-mills in Ireland described by A. T. Lucas 3).

The approach channel of the mill cut off one of the bends of the river Løvenå, an arrangement which gave greater force of water (Figs. 1-2). The flow of water to the channel could probably be regulated by means of an "upper floodgate", which must have lain further to the west, though no trace of it has been found. The investigation covered the approach channel from the point where it was cut away by a modem drainage ditch, and as far as the site where the actual mill must have stood. The latter had unfortunately disappeared, removed in all probability actually in the Middle Ages. Closest to the drainage ditch some of the oak posts which were used to hold in position the sides of the mill-channel were found still in position. On the plan (Fig. 3) a layer of stones is to be seen at the bottom of the channel; this layer became closer packed as it approached the surviving portion of the wooden channel. The stones in the bottom of the channel were sharp edged and shattered, some clearly by the humic acid in the soil, while some small pieces of flint were white and crackled, probably as the result of action by fire, though this action did not necessarily take place when the stones were in their present place. The stones at the bottom of the approach channel rested throughout on a layer of branches laid parallel to the course of the channel, and the branches were held in position by loose wodden cross-pieces (Fig. 4). The branches prevented the stones sinking into the soft peat soil, while the stones prevented the water digging a deeper channel and undermining the mill.

Where the layer of stones ended, a compound oak channel continued the same course (Fig. 5). It consists of two sideplanks, 4-5 cms. thick, 282-288 cms. long, with a rightangled section, one side of the angle forming part of the floor of the channel, and the other angle its wall. The highest portion of wall still preserved has an interior height of a little over 10 cms., and along its lower edge runs a groove, in which rested the third component of the channel, a plank originally about 285 cms. long, 20 cms. wide and 3 cms. thick. This construction was common in Danish wooden houses of middle-age date. The three parts of the channel were held together by means of rectangular holes cut through the side planks through which the anchoring posts mentioned above were driven. In the southern side plank there were two such holes, about 9 cms. wide and 18 cms. long (Fig. 5), while in the northern plank there was a corresponding hole in the western end in which still rested a long pointed anchoring post. Almost opposite the eastern post-hole in the southern plank, the northern plank showed two nail-holes pierced through the wood, revealing that here an anchoring post had stood on the outer side of the channel, and that the channel plank had been nailed to this post. A similar exterior anchoring post can be seen in Fig. 5 in the approximate centre of the south side of the channel. The western end of the channel rested on a cross-plank, which can be seen on the left in Fig. 5, and in the right background in Fig. 6. Fig. 6 shows in addition, bottom left, two cross-planks on which the eastern end of the channel had rested.

In the interval to be seen in Fig. 3 between the wooden channel described above and a hollowed trough to the east of the excavation there probably lay originally a similar compound channel of oak planks. Like the first it must have rested on cross-planks, through holes in which two anchoring posts stood on either side of the channel's side-pieces, to which they were nailed. One of the cross-planks was found in position, as can be seen in Fig. 3. This channel, too, rested on a layer of stones which in turn lay on a mat of branches laid in the direction of the channel. The end of this channel, now missing, doubtless rested upon the cross-block which the landowner first struck, when he was engaged in removing some of the anchoring posts because they hindered the work of ploughing. This cross-block can be seen on Figs. 7 and 8, but when the picture was taken both it and the southern anchoring post had been taken up and replaced. It proved impossible to replace the anchoring post at exactly its original depth (the inset in the post lies about 10 cms. too high), while the cross-block had been split in taking up. On Fig. 7 the position of the cross-block in relation to the mill-trough can be seen. Unfortunately the closed western end of this hollowed-out trough was destroyed in the course of the owner's first attempt to dig up the timbers. This end-piece had split in the direction of the grain, but was originally of the same height as the sides of the trough, and was therefore level with the upper edge of the cross-block.

This cross-block at the western end of the mill-trough served, however, not only as a support for the now lost central portion of the mill-channel. It probably also bore a sluice-gate to regulate the flow of water to the actual mill-trough, and thereby served to start and stop the mill. Comparison with horizontal mills in Norway and Sweden - disregarding the fact that the approach channel there often meets the mill-wheel at an angle of 90° - shows a distinct measure of agreement with the theory that there should exist a sluice-gate at the point where the approach channel meets the actual mill-trough. Such a sluice-gate is often furnished with a counterbalanced lever such as those used on wells, which would allow the sluice-gate to be easily raised and lowered. Fig. 9 shows just such an apparatus from Värmland in Sweden, from a drawing by N. Keyland. A similar Norwegian horizontal mill, without a seesaw lever, is illustrated in Curwen and Hatt's "Plough and Pasture" 6). The water flowing to the mill through the approach channel runs over the sides of the channel when the sluice-gate is shut and the mill not working, and it must find its own way around the mill on both sides. On Fig. 8 two circular holes can be seen in the cross-block (one in the edge of the crack, the other in the corner of the shadow of the post). Sticks placed in these holes must have served to guide the sluice-gate when it was being raised, and to prevent the water breaking it up when it was lowered 7).

On both sides of the mill-trough lay an embankment built up of white waterlaid sand. In the embankment stood a few posts which probably originally held in position faggots of branches and twigs, but as the process of oxidisation has had more favourable conditions in the sand than in the peat around it, only very little remains of the woodwork were here preserved. Two layers of sand were found in the embankment, separated by a layer of peat in which the roots of trees or bushes which had grown on the embankment were partially preserved. The lower layer of sand sloped downwards towards the east to approximately the level of the bottom of the mill-trough, and rested on fairly deep peat. Several pieces of oak planks and beams were found in the sand. The layers of sand were possibly formed by the water, which, when the mill-wheel lay idle and the sluice-gate was down, would spread out on either side and run away over the embankment. Erosion thereby caused would not be very great, particularly since the embankment was doubtless periodically strengthened with faggots of soft-wood twigs, which are now no longer preserved. The embankment can be dimly seen on Fig. 10 as a light-coloured mound on either side of the mill-trough.

The purpose of the mill-trough, as in the case of the Irish horizontal mill in Fig. 12, is to increase the speed of the water immediately before it meets the mill-wheel. In the case of Morett Mill here illustrated the approach trough has been made funnel-shaped to add speed to the water, but this was perhaps unnecessary at the horizontal mill at Ljørring, with its fall of about 1:28? At the lower end of the trough there must have been a regulator, to ensure that the water only struck one side of the radially projecting scoops of the mill-wheel. In the case of a Renaissance water-mill at Bolle in Jutland 9), a large boulder had been placed in one side of the channel, and there, as at Killogrone Mill in Ireland 10), the water was forced into the left side of the channel, so that the wheel, if viewed from above, revolved in a clock­wise direction. As will he seen in Fig. 13, at Killogrone Mill a stone slab had been placed in position, with a depression at the left side of the upper edge, and the jet from this depression could just reach the hollowed wheel-scoops and turn the wheel. At Ljørring Mill the trough is 216 cms. long, and its greatest height is 31 cms. and its greatest width 48 cms. The thickness of the wood at the bottom is 8-9 cms., and at the sides 5-6 cms., so that the width of the channel would be about 35 cms. At Morett Mill, on the other hand, the inside width of the trough was 50 cms., and its length was no less than 420 cms., while the mill­channel of the Renaissance mill at Bolle was about 90 cms. wide immediately before the narrowing, and the width of the actual narrowing was 35-40 cms. On the other hand, the channel at the mill which existed at Bolle at the beginning· of the Christian era (MD I) was only 35 cms. wide 11).

The mill-trough at Ljørring was provided with nail holes for securing. These holes were found at two points on each side a little above the bottom of the inside channel, and they were positioned in pairs opposite one another. When it was taken up the trough was still nailed fast to the two oak posts which can be seen towards the front in Fig. 7. The interval between these posts and the sides of the trough was filled with two pieces of wood through which the nails were driven. Under the bottom of the trough lay a cross-block of oak, 98 cms. Iong, 25 cms. wide and 9 cms. thick. In both ends of this block were cut rectangular holes, which held these anchoring posts (Fig. 14). Two anchoring posts, which must have stood opposite the nail-holes nearest to the sluice-gate, must have been pulled up at some time before the excavation took place. The upper edge of the mill-trough lay only 3-13 cms. below, the upper edge of the embankment already described.

As mentioned in the introduction, no trace was found of the actual mill. It was probably a wooden building, built of horizontal planks held between vertical posts resting in a hollow footing. The materials could therefore be used elsewhere without difficulty and the building was presumably moved some time during the Middle Ages. Sand washed over the site has levelled all traces in the soft earth. Nor were any potsherds found, which could have dated the site, and the pollen samples which were taken in the peat proved to contain so few well­preserved pollen grains that it was not tempting to base a date on this evidence. On the other hand, it should be possible in the course of time to date the site by means of C-14 analysis. However, all the pieces of timber which have been preserved and brought to the National Museum have now been conserved, so that it would be necessary for such an analysis to dig up anchoring posts from the site. At the time when the discovery was made, in 1952, it was considered that the dating of objects from the Middle Ages by means of C-14 involved too large a degree of uncertainty. But if the margin of error can be brought down to ± 100 years it would be of great value in discoveries of this nature to keep a portion of the woodwork unconserved for future dating.

Axel Steensberg





Steensberg, A. (1959). En skvatmølle i Ljørring. Kuml, 9(9), 130–145. https://doi.org/10.7146/kuml.v9i9.103046