Lars Aasbjerg Jensen, Poul Klenz Larsen and Tim Padfield


Maglehøj is a Neolithic passage grave in Denmark. It was built about 5200 BP. The chamber walls consist of large irregular stones with infill of small horizontal sandstone slabs. The horizontal joints were sealed with rolls of birch bark. When the chamber was opened in 1823, the bark was well preserved, but today is brittle and fragile.

Measurements of the ventilated grave showed episodes of condensation in spring and early summer as warm air entered the still cool grave chamber. A second year of measurement, with the grave entrance sealed, showed a constantly high RH with intermittent condensation in late summer caused by the temperature gradient between floor and ceiling. A third year with the floor made impervious to vapour showed a RH over 90% but no evidence of condensation on the stones.

The measurements show that the mantra that ventilation is always good, does not apply to a moderately ventilated enclosure with high thermal inertia and low moisture buffer capacity. It will experience a more variable RH than ambient, often reaching condensation, even though the annual average RH will be lower than in buried, unventilated spaces.


Maglehøj (figure 1) is a Neolithic passage grave near Hårlev, about 30 km south of Copenhagen [1]. It has a well preserved chamber (figure 2) within the grass covered earth mound. The chamber walls consist of large irregular stones with infill of small horizontal sandstone slabs. The horizontal joints between the slabs were sealed with rolls of birch bark (figure 4), which has been carbon 14 dated to 5200 BP. When the chamber was opened in 1823 the bark was well preserved, according to written accounts. Today the bark is brittle and fragile, crumbling in many places or already lost in some areas. This apparently rapid deterioration may be caused by the microclimate resulting from the free ventilation of the grave through the open entrance. A recent excavation (figure 3) shows the care taken to drain water away from the chamber through a gravel layer outside the massive wall stones.


Figure 1: The neolithic passage grave Maglehøj, 30 km south of Copenhagen


Figure 2: The interior of the grave. Surface temperature sensors (type K thermocouples) are sprung against the stones with about 20 mm of the thin wires set close to the stone to minimise heat conduction from the chamber air.


Figure 3: The excavation trench reveals the gravel drainage surrounding the chamber structure.


Figure 4: The birch bark rolls inserted between the sandstone slabs which fill the spaces between the large irregular stones. The birch bark gave the opportunity to date the construction of the tomb to about 5200 BP.

The measurement campaign

We have measured the microclimate within the tomb and its surroundings for three years. The first year with free ventilation through the entrance, the second year with the entrance sealed and insulated, the third year with a water vapour barrier laid on the floor.

Measurement techniques

Relative humidity (RH) and temperature sensors were placed outside and within the grave chamber. Resistive wetness sensors were attached to several of the large stones. Surface temperatures were measured with type K thermocouples sprung against the rock surface. Rainfall was measured beside the earth mound. Gypsum soil moisture sensors were placed at several depths beside and on top of the mound and below the chamber floor. The sensor layout is summarised in figure 10.

All these sensor signals were collected by a Campbell Scientific data logger which transmitted the data sets every hour by radio telemetry.

Measurement difficulties

The high humidity within the grave chamber degraded the sensor performance. The datalogger was enclosed in a container with silica gel drying agent but the signal processing for the capacitive humidity sensors was necessarily integrated in the exposed sensor head. The chamber RH sensor failed during the second year. Since the RH reading of the chamber air space is essential to interpretation of the other sensor signals, this was a serious setback, which we should have anticipated by installing duplicate RH sensors. The wetness sensors had a tendency to delaminate from the rock surface. The gypsum sensors dissolve over time, depending on the water movement around them. Only the thermocouples were reliable, suggesting that we should have supplemented the capacitative RH sensor in the chamber with a thermocouple based psychrometer to confirm the very high RH values, which are notoriously difficult to measure accurately but which turned out to be important to interpreting the microclimate data.

The measurements

The first year of measurement was made with the chamber freely ventilating through the metal lattice door. Figure 5 shows the contrast between inside and outside temperature and RH. The shaded areas indicate moments when the dew point of the outside air exceeded the inside temperature, so condensation was likely. Indeed condensation was observed on some of the ceiling stones (figure 6). Generally, the inside RH was more variable than that outside, being generally lower than outside in winter, when the tomb was relatively warm.

first year graph

Figure 5: The microclimate in the grave chamber when it had a lattice door permitting ventilation. The pale traces show the outside RH and temperature. During the winter the chamber temperature is generally above ambient, so the RH is below ambient, going down to 50% on several days. During the spring, warm outside air enters the still cold chamber causing condensation on the stone surfaces. The periods of condensation risk, when the outside air dewpoint is above the inside temperature, are marked at the bottom of the graph.


Figure 6: Condensation on a ceiling stone. The birch bark can be seen in the crevice just below the condensation.

After one year of measurement, which clearly showed the variability of the RH, and that the daily ambient temperature cycle influenced the interior, the entrance was sealed with a solid insulated door to prevent movement of both air and heat. However airtightness was not achieved because of the porosity of the earth mound, with numerous burrows providing tortuous air passages towards the grave chamber. The microclimate during this second year is shown in figure 7.

second year graph

Figure 7: The microclimate in the grave chamber after the entrance was sealed and insulated. The faint traces show the outside RH and temperature. The temperature is much more stable and the RH is now always over 90%, with no evidence of influence from the weather. Condensation was still observed on the underside of the ceiling stones. In the autumn, the ceiling stones cool down faster than the floor, so moisture in the floor will evaporate and condense on the ceiling. The RH is measured in the centre of the chamber, so it always shows a value below 100%. The condensation periods marked at the bottom of the graph are times when the RH calculated for the ceiling surface exceeds 100%.

The inside temperature became more stable and the RH became very high, but stable. There was still condensation on the ceiling. This was due to the temperature difference between the floor and the ceiling, so that there is condensation when the floor is warmer, even though only by a few degrees.

The vertical movement of water vapour is not symmetrical because the floor is porous sand while the ceiling is impermeable igneous and metamorphic rock. Water condensed on the floor will disappear by capillary movement. So for the third year we installed a polyethylene membrane over the floor, with perforations at the low points to allow condensate to flow downwards, but so few perforations that evaporation upwards is inhibited. The microclimate during this period, to date, is shown in figure 8.

third year graph

Figure 8: The microclimate in the grave chamber after the entrance was sealed and insulated and after a polyethylene membrane was laid over the floor. The pale traces show the outside RH and temperature. The temperature follows the seasonal trend with no daily variation. The RH is now around 90%. The disturbances in the RH trace are due to our interventions to check calibration, to change the sensors and to make perforations in the floor membrane to allow condensed water to flow down into the porous sand below. No condensation has been observed but cannot be ruled out because the wetness sensors have proved unreliable.

Soil moisture

The entire period of soil moisture measurement is shown in figure 9. The soil of the mound is wet during the winter and dry from July to the beginning of winter. At the chamber floor and below, the picture is confusing. The floor can be described as dry to the touch except for late summer in the ventilated chamber (2013), a period when the soil above the chamber has become too dry for plant growth. The polyethylene cover applied in 2015 caused the underfloor moisture content to increase, and thus implies a significant decrease in the upward flux of water vapour into the chamber.

soil moisture graph

Figure 9: The suction pressure measured by the gypsum blocks throughout the measuring period. Dry conditions are at the bottom of the graph. The soil sensor was too dry to give a measurable signal from July to winter every year. The chamber floor would be described as 'dry' except for a small period in July 2013 in the ventilated chamber. After installation of the polyethylene floor cover in 2015 the moisture content of the floor increases at every depth.


The climate instability caused by unforced and moderate ventilation of a space surrounded by rock and earth is not surprising, but emphasises the general point that ventilation of spaces with high thermal inertia but low moisture buffer capacity risks condensation, particularly in early summer.

The distillation of water to and fro between ceiling and floor was an interesting and unpredicted observation. The inhibition of the upward movement of water vapour by the polyethylene membrane was impressive, but needs confirmation with better instrumentation. We need accurate hygrometry at high RH and thermopiles to provide accurate temperature differences between surfaces, to supplement the temperature measurements which can only be relied on to one degree. There is no evidence of water penetrating from above, except for a definite leak near the entrance. The gravel drainage layer seems still to be functioning well.


The comprehensive field measurements have proved both difficult to keep accurate and consequently difficult to interpret securely. The high and varying RH caused drift and eventual destruction of the sensors and their electronics. Gypsum soil moisture sensors gradually dissolve and the surface wetness sensors had very short lifetimes.

However, the measurements clearly show that the long-held mantra that ventilation is always good, whatever the hygrothermal characteristics of the construction being protected, is wrong. In particular, an environment with high thermal inertia and low moisture buffer capacity will always experience a more variable RH than ambient, often reaching condensation, even though the annual average RH will be lower than in buried, unventilated spaces.

It also seems that the source of water vapour within the chamber is the floor, which argues against covering the mound with a watertight membrane.

This investigation was prompted by the presumed rapid deterioration of the birch bark sealing between the small sandstone slabs which filled the gaps between the large stones. Our investigation gives no direct information about the cause of deterioration but it does show the variability of the climate in the ventilated chamber, contrasted with the temperature stability and constantly high RH which we must presume prevailed over most of the time since the grave was built.

Technical information

Datalogger: Campbell Scientific CR1000WP with AM16/32B 16/32 multiplexer mounted in ENC 16/18 box.
Temperature and RH inside and out: HC2S3 Rotronic Hygroclip2 temperature and RH probe mounted in radiation shield MET21.
Stone surface wetness: Campbell Scientific 237F Wetness sensing grid resistors
Raingauge: Texas Electronics TR-525M serial No. 37474-1105
Soil moisture: Delmhorst gypsum block
Temperature measurements: Thermocouple type K
Data transmission: Wavecom Fastrack Supreme 10 (Open AT) WM20392
Power supply: two 65AH lead-acid batteries, good for 3 months operation.

sensor layout

Figure 10: The sensor layout


We thank archaeologists Jørgen Westphal and Torben Dehn for the information about the history of archaeological studies and for supporting the work reported here. This investigation was supported by the Danish Agency for Culture and Palaces.


[1] Archaeological description of Maglehøj (in Danish):

This is a reformatted version of the article submitted to the International RILEM Conference on Materials, Systems and Structures in Civil Engineering, Conference segment on Moisture in Materials and Structures. 22-24 August 2016, at the Technical University of Denmark, Lyngby, Denmark.


Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.