Braathen Dendrokronologiska Undersökningar
Wooden buildings on Gotland

Eke and Hemse, two stave-churches.

At Statens Historiska Museum, Stockholm, timber from two stave-churches Eke and Hemse is preserved. The timber has been re-used as part of wooden floors in the present medieval stone-churches. For taking bore-cores, among many pieces of oak-wood, such pieces were chosen which might either give long and well shaped sequences of annual rings or have part of a surface which might be the border between sap-wood and heart-wood. The reason for a selection following the later choice is that both for ships and buildings sap-wood in oak-wood is removed. In fresh wood it is difficult to perceive where the layer of sap-wood starts, and part of it frequently remains in the formed timber. Sap-wood is easily eroded, mainly because of insect attacks, and a curved surface remains.
For ordinary grown oaks in Sweden the layers of sap-wood contain minimum 9 annual rings (in very young trees) and maximum 20 annual rings (in old and slow growing trees) with an arithmetic medium of 14. The standard deviation is +3, -2 years. If the border between sap-wood and heart-wood can be confirmed, the felling year thus can be fixed with an error of +3, -2 years.

A table concerning dating of the bore-cores from Eke and Hemse stave-churches follows:

Identification No of annual rings Sapwood Year of the outermost measured annual ring Year of the innermost measured annual ring
Eke 3 85 - 1075 991
Eke 1 191 - 1074 884
Eke 2 181 - 1072 892
Hemse 4 113 - 1086 974
Hemse 1 115 - 1068 954
Hemse 2 79 - 1068 988
Hemse 3 215 - 1067 853
Hemse 5 127 - 1061 935

'Eke 3' and 'Hemse 4' have been chosen at a spot with a curved surface ( = a small part of the surface of a cylinder). If that surface (it lacks remains of sap-wood) is the boarder between sap-wood and heart-wood, the timber for the stave-church of Eke was felled in the period 1087 - 1092 and for the stave-church of Hemse in the period 1098 - 1103.

It can not be taken for granted that the judgement concerning the boarder between sap-wood and heart-wood is correct, and a more restricted judgement is that the stave-church of Eke was erected after 1084 and the stave-church of Hemse after 1095 where the nearest 3 years after those dates have a very small degree of probability and that there is no upper limit.

On Gotland there are traces from early wooden churches in at least the churches of Lojsta and Silte. There is no evidence that most of the early stone-churches in Scandinavia replaced stave-churches.

In the beginning of the 11th century Christianity was common in at least the more dense populated parts of Scandinavia. According to Adam of Bremen, in 1072 there existed in the archbishop's diocese of Bremen about 500 churches, most of them wooden. The main part of them were placed in northern Germany and Denmark. This tells that at that time there were not many churches left for the rest of Scandinavia.

Th. Bartholin, Lund, has dated remains of three stave-churches from Lund in the province of Scania (Skåne, Schonen). One of them, S:t Stephan, was built in 1050, S:a Maria Minor was built after 1050 and Drotten II was built before 1100.

I have dated remains of three stave-churches from the western part of Sweden. The stave-church of Vänga was built in the period 1063 - 1065 and extended or rebuilt after 1236. The stave-church of Kinnaromma was erected after 1120 and Hällestad after 1133. The stave-church of Hedared was erected in the period 1498 - 1503.

In the diocese of Skara in western Sweden there were about 550 parishes in the medieval ages. In that region there are traces of 11 wooden buildings which according to historians of Art might remain from stave-churches.

Literature: Emil Ekhoff, Svenska Stavkyrkor, Stlm 1914 - 1916.

Sproge.
A building from Sproge in frame work construction.

When relaying a floor in the medieval stone-church of Sproge many planks of oak-wood and other pieces of wood were found. Some of these were shaped as floor planks and others had once been wall-planks in a framework construction (drops of tar show that those planks have been pointed horizontally and have rested on side-edge. Border between sap-wood and heart-wood is missing. In order to get an indication of when the building might have been erected, a number of samples were taken as follows below:

Identification No of annual rings Sapwood Year of the outermost measured annual ring Year of the innermost measured annual ring
Sproge 5 165 - 1131 937
Sproge 15 193 - 1128 936
Sproge 9 214 - 1127 914
Sproge 6 217 - 1115 899
Sproge 8 134 - 1105 963
Sproge 1 209 - 1103 895
Sproge 11 114 - 1102 989
Sproge 14 117? - 1100 985
Sproge 3 180 - 1098 919
Sproge 4 160 - 1096 937
Sproge 13 149 - 1089 941
Sproge 10 156 - 1088 933
Sproge 12 172 - 1087 916
Sproge 7 255 - 1063 809
Sproge 2 156 - 1057 901

Those readers who are familiar with calculus of probability know that there exists a model of calculation by means of which one can get an expected value for the youngest number of year in the whole collection, but the basic conditions are lacking here, among others, because the samples are not chosen by random. We should however restrict us to the fact that the building has been erected after 1131 + 9 = 1140 and that the nearest 3 years after 1440 have a low degree of probability.

What conclusion would we have made if we would have dated one of the pieces only and if that piece were 'Sproge 2', which gave the oldest number of year?

One answer should be that if we want to date a structure containing much timber without waney edge or sap-wood or the curvature of the trunk, and we for some reason are restricted to take one sample only, we should desist from trying to date the object.