Textiles and environment in the showcase containing Saint Canute the Holy († AD 1086): Radiocarbon dating and chemical interactions

The cathedral in Odense, Denmark, has for nine centuries held the relics of the Danish King St Canute the Holy and his brother Benedikt. They were both murdered in the predecessor church at the site in AD 1086, and Canute was sanctified in already in AD 1100. The history of the relics has been that of turmoil at times, varying from initial worship of the Catholic believers, to being walled up and hidden away after the protestant reformation in AD 1536, and since the 19th Century on display as important heritage objects of national importance. In the present work we have characterised some of the textiles and analysed the air inside the glass showcases exhibiting the 11th Century wooden coffins holding the remains of St King Canute the Holy and his brother together with some precious textiles. Contrary to previous belief, we now prove that all the textiles analysed have the same age, which is consistent with the time of the enshrinement of the King and his brother in AD 1100. It is also shown that some of the textiles were treated with paraffin wax, most likely during attempts at conservation at the National Museum in the nineteenth century. The results of the air chemistry analyses show the problematic side of simultaneously storing of slowly decaying wood, fine textiles, and human bones in rather airtight environments. The wood continuously releases organic acids, the soaring concentrations of which are potentially harmful to the 11th Century textiles and probably also to the bones.


Introduction
A puzzle in Danish history concerns two very tangible objects: two coffin-shaped 1.5 m long reliquaries situated in the crypt of St Canute's Church in Odense [1][2][3][4]. One of the wooden coffins is now missing a lid, while the other has preserved the lower part of a hipped lid (Figs. 1 and 2). The two reliquaries were made of oak and were once covered by gilded copperplates of which only fragments have survived. Several Romanesque altar-frontals, retables, and crucifixes manufactured in the same technique have survived elsewhere in Danish churches and an example of these can give an impression of the original appearance of the shrines (Fig. 3) [5,6].
Each reliquary contains the bones of a male individual. One is identified as King Canute the Holy (Kong Knud 2. den Hellige, in Danish, [7]), who was born c. 1040-45 and ruled 1080-1086; the other most likely his younger half-brother Benedikt. Several well-preserved textiles of silk and linen are present in the reliquaries. The dates of the textiles and their historical context are one of the two main research questions of this investigation, which shed new light on the complicated story of the shrines. Were the textiles specifically chosen for the enshrinement of the saint and were they contemporaneous with the enshrinement? Or do they represent a more random selection of old and contemporary textiles as has previously been suggested [8,9]? The other research objective encompasses understanding the chemistry of the environment inside the showcases holding the reliquaries. Are the bones and the textiles well preserved and well-kept for posterity in the present configuration?

The martyrdom of St Canute and his brother Benedikt
Canute the Holy was a son of King Svend Estridsen, who reigned c. 1047-1074. Among the numerous children, which Svend had with various women, five sons became kings of Denmark one after the other. Canute was the second of them. His ideal was that of a strong royal power and his attempts to realize it in cooperation with the Church made him many enemies. In July 1086 he was killed together with his brother Benedikt and 17 housecarls by an army of unsatisfied magnates. Since the king had taken sanctuary in St Alban's Church in Odense, the killing took place on holy ground, and this was the first step towards a politically very promising sanctification instigated by other members of the royal family still ruling the country. Already in 1095 Canute's body was transferred to a stone coffin under a new large travertine church which was being erected to his honour southwest of St Alban's Church [10]. And thanks to the fourth king in the line of brothers, Erik I Ejegod, in 1100 the pope promised to canonize Canute. On April 19th in the "sixth year" of Erik's rule, i.e. either 1100 or 1101, Canute's bones were transferred from the stone coffin to a reliquary in the new St Canute's Church. Details of the killing were passed on in two brief accounts which are now only known from later copies: "Tabula Othiniensis" originating from the first transfer of the bones in 1095 and an "Epitaphium St Canuti" which was placed among the saintly bones during the transfer to the reliquary in   [11,12]. Less than 40 years after the killing a detailed account of the life and death of King Canute was written by Aelnoth of Canterbury, who was a priest at the Benedictine monastery attached to St Canute's Church which had recently been founded by King Erik I Ejegod with inspiration from England [13]. In his "passio gloriossimi Canuti regis et martyris" Aelnoth described among other things how the royal bones were wrapped in silk and placed in a marvellous shrine forged of yellow shining metal and decorated by precious blue and yellow stones. He states that Edel, Canute's widow, who had married Duke Roger I of Apulia and Calabria in Southern Italy, contributed to the furnishing of the shrine with rich gifts. The reliquary was placed "on the life-giving altar table", either the high altar or more likely an altar in the crypt which is more consistent with Aelnoth's additional information that Canute the Holy in this way was "displayed to all who sought his protection" [14]. The entrance to the choir with the most holy high altar was reserved for clerics only. Both a Russian prayer from c. 1135 [15] and an anonymous legend about Canute from 1220-1250 [12] mention that not only Canute but also his brother Benedikt were venerated as saints, and the Icelandic Chronicle "Knytlinga-saga" from c. 1250 tells that both brothers were enshrined [16]. But for the next centuries only the veneration of Canute is mentioned in the historical sources, especially in connection with his day of martyrdom in July, when his shrine as well as a separate reliquary for his skull and other reliquaries of the church were carried in a procession around the town of Odense. Since the end of the 13th Century, the travertine church was gradually replaced by a Gothic brick cathedral, which as an unusual feature for its time got a crypt similar to the one in the old Romanesque church, undoubtedly due to the cult around the reliquaries [10].

The fate of the reliquaries after the reformation
In connection with the Danish civil war 1534-1536 the leader of the rebellion Count Christopher of Oldenburg in October 1534 sent a letter to the prior of St Canute's monastery in Odense, asking him to deliver all gold, silver, and precious objects present in the cathedral, except for four chalices. The order may not have been obeyed. But in 1536 a proper silver tax was imposed by King Christian III who was about to win the war and needed payment for his mercenaries. This time the monastery in Odense was obliged to pay 4000 "lod" silver, c. 62 kg. The prior paid up and mentioned in a letter that with royal permission he had taken part of the silver from a shrine in the church. Among the silver was also a gilded copper picture [17]. The shrine in question is likely to have been a reliquary. But an indisputable mention of a reliquary shrine in St Canute's Church in Odense does not actually occur until 1582.
In 1582 the church was being rebuilt, encompassing the demolition of both the crypt and a chapel to the north side of the church. During these works a well-preserved large reliquary was uncovered, made of oak with metal mounting that according to some sources were gilded and had a decoration with rock crystals [1]. All sources agree that this was the reliquary of St Canute, and in their descriptions, they also quote the now lost inscriptions which were made for the enshrinements of Canute, "Tabula" and "Epitaphium". One author who claimed to be an eye-witness to the find, the clerk of the now secularized St Canute monastery, Joachim Konninck, stated that the text "Epitaphium" was written on a piece of white parchment placed among the bones. The text was not discovered immediately since the bones were covered by textiles. According to Konninck, the reliquary was lined with several brownish silks that were still very beautiful and very rare, and it further contained a double yellow silk with yellow silk embroidery and a cotton lining as well as a golden piece of textile enclosing a linen made of smooth cotton wrapped around the bones [12]. The find aroused an interest among scholars all over Denmark, and several foreign Catholics expressed their wish to see the bones and eventually acquire some of them. Thus in Several decades then elapsed without any mention of reliquaries. In 1694 some minor repair works were being done in the choir, and on that occasion a small hole was noticed in the east wall behind the altar, allegedly filled up with feathers and down. The hole turned out to be opening into a niche containing two reliquary shrines which were stripped of their metal fittings except for a few fragments. They were placed vertically so that both bones and textiles lay in a heap at the bottom of each shrine. It is not known why and when the two shrines were walled up nor when they were stripped of their metal fittings. The reliquary of Canute must have lost its fittings sometime after 1582. The whereabouts of the other shrine are quite unknown, though it is a tempting suggestion that it might be the abovementioned unspecified shrine which was stripped of its silver and copper mounting around 1536. An investigation of the reliquaries was carried out in 1696 by the Odense antiquarian Thomas Bircherod. He mentions that the shrine with the hipped lid contained several pieces of silk and some pillows of wool and silk filled with down [1,12,18]. The textiles in the shrine with the columns were described as linen or cotton with a resemblance to altar-cloths. After the investigation both the reliquaries were once more placed in the narrow niche in a vertical position. But in 1833 they were permanently removed from their unpleasant hiding place. On this occasion, the textiles in each shrine were once more listed [19]. The shrine with the hipped roof contained a fairly large silk blanket with depictions of eagles, a couple of smaller pillows and a long, narrow, round pillow with a yellow cover which is described as cotton (it was not realized at the time that it is in reality silk), as well as a piece of red and blue "changeant"-coloured, thin silk.
The shrine with the columns only contained a pillow and a fringed linen cloth [1].

The situation of the reliquaries and the storing of the textiles since 1833
After 1833 the shrines were furnished with new glass lids and displayed behind an iron fence at the east end of the northern side aisle. The most important of the textiles were taken out of the shrine with the hipped lid and send to the Museum of Nordic Antiquities (Oldnordisk Museum), the original name of the National Museum in Copenhagen, while smaller fragments remained in Odense.
In 1875 the reliquaries with their bones and textiles underwent a thorough investigation and restauration. The reliquaries were placed on new, brickwork pedestals in the reconstructed crypt and both bones and textiles were placed inside them. The shrine with the columns got a hipped wooden lid as an imitation of the lid on the other shrine. Both shrines were fitted with a glass cover on the top. But the arrangement did not live up to the intentions expressed in the statement by the scientific commission, dated 12.12.1874, which stressed that the textiles were so fragile that they could only be saved for posterity by being hermetically sealed [20]. Most of the textiles which had been removed from the shrine with the hipped lid in 1833 were now placed in the shrine with the columns because their high quality seemed more fitting for the shrine of a king. The report from 1886 lists the following textiles: 1. A piece of red and blue, thin shot-silk cloth, which seemed originally to have served as lining inside the shrine (Table 1, sample K2/KLR-7186). The reliquaries were protected from the audience by panes of glass mounted in brass racks. Problems with humidity in the crypt caused mouldy growth on the coffins and their content which had to undergo repair and cleaning several times. In 1936 the reliquaries were transferred to new glass showcases designed by the architect Knud Lehn Petersen (Fig. 6). In 1956 the "eagle carpet", which had been stretched out as a cover over a part of the shrine with the columns, was removed and placed in its own, new showcase, which in 1990 was replaced with a new one [1]. The samples investigated here were taken by a conservator in 2008 and they are listed in Table 1.
The present investigation was aimed at two different aspects of the relics. First, the authenticity and the dating of the textiles was investigated, including a chronological comparison with previously reported radiocarbon dates of the bones and dendrochronological dates of the timber in the shrine with the pillars. In this connection some of the textiles were analysed to screen for contaminants in connection with the radiocarbon dates. Secondly, we performed a chemical assessment of the environment inside the glass showcases, which has bearings on the state of conservation and possible contaminants relevant for a proper assessment of the reliability of the radiocarbon dates. In addition, the possible presence of conservation materials or contaminants on the textiles was evaluated, and one  piece of coloured textile was analysed to characterize the dyeing material used in its production.

High-performance liquid chromatography with a diode-array detector (HPLC-DAD) for dyes analysis
The HPLC analyses were performed using a HPLC system consisting of a PU-2089 Quaternary Pump with degasser equipped with an autosampler AS-950 coupled to a spectrophotometric diode array detector MD-2010 (Jasco, Japan). The data were processed with the Chrom-Nav software. The spectra acquisition was in the range of 200-650 nm every 0.8 s with a resolution of 4 nm.
The chromatographic separation was performed at 30 °C on an analytical reverse phase column Poroshell 120 EC-C18 column (3.0 mm × 75 mm, 2.7 μm particle size) with a Zorbax Eclipse plus C-18 guard column (4.6 mm × 12.5 mm, 5 μm particle size), both Agilent Technologies, USA. The eluents were A: formic acid (FA 0.1% v/v) in HPLC grade water and B: formic acid (FA 0.1% v/v) in HPLC grade acetonitrile. The injection volume was 20 µL, the flow rate was 0.4 mL min −1 and the elution program was 15% B for 2.6 min, then to 50% B in 13.0 min, to 70% B in 5.2 min, to 100% B in 0.5 min and then hold for 6.7 min; re-equilibration took 11 min. 0.462 mg of sample KLR-7199 were subjected to a mild extraction with EDTA/DMF solution. The treatment consisted in the following steps: addition of 200 μL of 0.1% Na 2 EDTA in H 2 O/DMF (1:1, v/v) solution, extraction in ultrasonic bath at 60 °C for 1 h, filtration with PTFE filters and injection of 20 μL of the extract in the HPLC-DAD system. Peak assignment was based on comparison with analytical standards and relevant reference materials.

Pyrolysis gas chromatography mass spectrometry (Py-GC-MS)
The analyses were performed using a multi-shot pyrolyzer EGA/PY-3030D (Frontier Lab, Japan) coupled with a 6890 N gas chromatography system with a split/ splitless injection port and combined with a 5973 mass selective single quadrupole mass spectrometer (Agilent Technologies).
The samples (80-120 mg) were placed in stainlesssteel cups and directly analyzed without further sample pre-treatment [21][22][23]. The pyrolysis conditions were optimized as follows: pyrolysis chamber temperature 550 °C, interface 280 °C. The GC injector temperature was 280 °C. The GC injection was operating in split mode and the best analytical results were obtained with a split ratio of 1:10 [24]. The chromatographic separation of pyrolysis products was performed on a fused silica capillary column HP-5MS (5% diphenyl-95% dimethyl-polysiloxane, 30 m × 0.25 mm inner diameter, 0.25 μm film thickness, J&W Scientific, Agilent Technologies), preceded by 2 m of deactivated fused silica pre-column with internal diameter of 0.32 mm. The chromatographic conditions for the analysis were: 36 °C for 10 min, 10 °C min −1 to 280 °C, 300 °C for 2 min, 15 °C min −1 to 300 °C. The carrier gas was helium (purity 99.9995%) with a gas flow set in constant flow mode at 1.2 mL min −1 . The MS parameters were as follows: electron impact ionization (EI, 70 eV) in positive mode; ion source temperature 230 °C; scan range 50-700 m/z; interface temperature 280 °C. Perfluorotributylamine (PFTBA) was used for mass spectrometer tuning. MSD ChemStation (Agilent Technologies) software was used for data analysis and peak assignment was based on a comparison with literature mass spectra, standard compounds previously analysed in the same conditions, and libraries of mass spectra (NIST 8.0).

Carbon isotopes and radiocarbon dating
Radiocarbon dating was performed on nine textile samples at the Groningen Center for Isotope Research in the Netherlands. The dating was performed by AMS (Accelerator Mass Spectrometry).
The samples were first decontaminated by Soxhlet extraction in hexane after which they were subjected to the standard ABA (Acid-Base-Acid) pre-treatment procedure [25]. The samples were combusted to CO 2 by an elemental analyser, coupled to a stable isotope ratio mass spectrometer (EA/IRMS) [26]. Part of the CO 2 is cryogenically trapped and subsequently transferred into graphite. The graphite was pressed into targets for the ion source of the AMS, which measured the 14 C content of the graphite. The AMS was based on a 2.5 MV tandetron accelerator, manufactured by High Voltage Engineering Europa. The machine is described in detail in van der Plicht et al. [27]. The 14 C content was measured as the 14 C/ 12 C ratio, expressed as the 14 C activity 14 a [28]. From this the 14 C age was calculated, expressed in BP (Before Present, defined unit for 14 C-time). Reporting is by convention relative to the Oxalic Acid standard, using a halflife of 5568 years and includes correction for isotopic fractionation to δ 13 C = − 25‰ [28].
The IRMS provides a measurement of the stable isotope ratio δ 13 C. This is expressed in a delta value, which is a measure of the difference of the 13 C/ 12 C isotope ratio of a sample and that of a standard (or reference) material: It is expressed in permille. For 13 C, the reference is an international standard limestone known as Pee Dee Belemnite (PDB) which has an accurately measured absolute 13 C concentration [29].

Climate monitoring
The temperature and relative humidity of the air was monitored inside the showcase, in the crypt, and outside the church, by use of electronic climate sensors integrated in battery-driven dataloggers. For indoor measurements Tinytag Plus 2 loggers were used (Gemini Dataloggers, Chichester, United Kingdom), while the outdoor conditions were monitored using a weather station grade datalogger mounted in a sun and rain screen and placed on a light pole at 2.5 m height (HOBO Pro v2, Onset Corp., Bourne MA, USA). The accuracy of such sensors is typically within ± 0.3 °C and ± 3% RH. The climate parameters were recorded at one-hour intervals for a full year.

Showcase leakiness
The air exchange rate of the glass showcase for St Canute was established by using CO 2 as a ventilation tracer gas [30]. The analysis was performed using two Bacharach CO 2 2810 Analyzers (range 0-10,000 ppm); one located inside the showcase and one located in the open air of the crypt (Bacharach Inc., New Kensington PA, USA). CO 2 was released from a compressed gas cylinder and into the showcase up to a concentration of c. 3,000 ppmv, after which the showcase was quickly closed. The CO 2 concentration was recorded each minute inside and outside the showcase for half a day (the battery lifetime of a fully charged instrument). Although a longer measurement period could be desired, the 12-h monitoring was found sufficient, as the indoor environment was very stable. The air exchange rate between the showcase and the room air was calculated based on the decay rate of the CO 2 concentration inside the case using the formula: Air Exchange Rate = ln(C1/C2)/(T2−T1), where C1 and C2 are the concentration surplus above ambient (C showcase -C room ) and T1 and T2 the two times.

Air pollution Passive sampling
Several sampling methods were used for detecting traces of compounds emitted from indoor materials, from human activity, from outdoor pollution, and de-gassing from the reliquary objects. The methods were all characterized by using passive sampling, where the compound was collected onto the sampling media over long time (several days or weeks), as the gas diffused through a sampling tube or membrane at a rate proportional to its concentration. Although slow and less sensitive than active sampling methods (by air pump), passive sampling was preferred because it caused less disturbances of microenvironments in the showcase, and allowed that average concentrations to be determined rather than a short-term incident measurements [31,32]. For all passive samplers, commercial analytical services were used. Samplers were mounted in an aluminium clip rack and placed at three sampling locations: inside the showcase, on a shelf in the crypt, and under the outdoor sun-and rain-screen used for the climate sensors. All compounds were measured by duplicate samplers at each location, except Volatile Organic Compounds (VOCs), which was not monitored outdoors. The reported values are the mean concentration of two samplers. The uncertainty of passive sampling methods was typically about ± 25% relative accuracy.

VOC screening
For a broad screening of VOCs Tenax TA tubes were used as samplers and analysed by Automated Thermal Desorption Gas Chromatography Mass Spectrometry (ATD-GC-MS). This was performed twice. The first time it comprised of 18.8 days of exposure. The Tenax tubes were analysed by the commercial company Eurofins A/S (Galten, Denmark). The second time, a year later, the analysis was performed with 8 days of exposure time and analysed by Norwegian Institute for Air Research-NILU (Kjeller, Norway). At the same time as the first sampling a short-term qualitative screening of the showcase air was performed by exposing (for 10 min) a solid-phase micro-extraction (SPME) glass fibre needle with a highly absorbing coating (85 µm Carboxen ® /Polydimethylsiloxane) to the inside of the showcase while the lid was slightly open. The fibre uptake was analysed by Solid Phase Micro Extraction Gas Chromatography Mass Spectrometry (SPME-GC-MS) at the Conservation Laboratory at the National Museum of Denmark (Kongens Lyngby, Denmark).

Specific compounds
Passive VOC screening by Tenax TA constitutes a broad but semi-quantitative method. The main organic component found in this way, acetic acid, was subsequently re-sampled twice at all locations using a passive sampler specific for organic acids (sampler and analysis by IVL Swedish Environmental Research Institute, Gothenburg, Sweden). The exposure time was 28 days. At these two incidents the ambient pollutants nitrogen dioxide and ozone was measured as well by passive sampling tubes provided with analysis by Gradko Environmental Ltd. (Winchester, United Kingdom).

Corrosion test
As a dosimeter for determining the microenvironment's corrosivity a 20 × 50 mm coupon of pure lead was exposed inside the showcase for 1 year (Merck lead foil 0.25 mm, no. 1.07365.0500 pro analysis). Before exposure, the coupon was polished with P-800 grinding paper and rinsed and degreased in acetone. The amount of corrosion was quantified by weighing the coupon on a microbalance before and after the exposure, and normalising to the weight-gain per square meter of metal surface [33].

Analyses of organic molecules in the textiles
The HPLC-DAD chromatogram of the blue thread of the sample B4/KLR-7199, extracted at 600 nm, is reported in Fig. 7. The chromatogram revealed the presence of only one peak, whose retention time and UV-Vis spectra match with those of indigo. This indigoid compound is the molecular marker for the identification of Indigofera o Isatis vegetal species [34,35].
The Py-GC-MS chromatogram obtained for the sample B4/KLR-7199 is shown in Fig. 8. The pyrolytic profile was characterized by the presence of pyrrole and phenol derivatives, together with diketopiperazines (m/z 154, 168, 194 [36]) characteristic of the pyrolysis of the proteinaceous portion of the material. This specific pyrolysis profile agreed with that obtained for a reference silk fibre. The analysis also allowed to identify the presence of several aliphatic hydrocarbons compatible with the presence of a paraffine wax, a petroleum derivative introduced in 1830′s. This material is commonly applied during restoration attempts.
The Py-GC-MS chromatogram obtained for the sample K11/KLR-1195 is shown in Fig. 9. The Py-GC-MS chromatogram was characterized by the presence of laevoglucose and furan derivatives. This profile suggested the presence of a cellulose-based fibre. The comparison with the reference material database allowed to hypothesize the use of jute. The analysis also showed traces of several aliphatic hydrocarbons compatible with the presence of a paraffin wax.
The Py-GC-MS chromatograms obtained for the samples K10/KLR-7194 and B1/KLR-7196 are shown in Fig. 10. Both the profiles feature the pyrolysis markers of a cellulose-based material. The comparison with the reference materials database allowed to hypothesize the use of jute fibres.

Radiocarbon
The results of the radiocarbon dating are listed in Table 2.

Calibration
The 14 C dates are calibrated into calendar ages using the latest calibration curve, IntCal20 [37]. This curve is annual for the time range under investigation. Hence the dates are not rounded. Both 1-and 2-sigma calibrated age ranges are shown in Tables 2 and 3. The code used for     paraffine wax, probably used in the restoration of the textiles. This material is completely soluble in hexane and indeed its removal was ensured by the Soxhlet extraction procedure of the samples.

Climate conditions
The indoor temperature was uniform in the crypt and in the showcase, with an annual average of 18 °C (Fig. 12). Daily temperature variations were tiny, while the temperature seasonally varied between 16 °C in the winter and 21 °C during summer. The moisture content of air (absolute humidity) inside the showcase was constantly 9 g m −3 , resulting in an almost constant 60% relative humidity (RH) year-round. In the church room, the annual average humidity was c. 9 g m −3 , but with larger seasonally fluctuations causing the relative humidity to vary between c. 50% and 70% RH for the main part of the year (with short-term spikes between 35-77% RH) (Fig. 12). During summer, the indoor air had the same moisture content as the outdoor air. However, during winter the indoor air contained a surplus of water vapour of about 4 g m −3 .

Air exchange rate
The CO 2 injected inside the showcase decreased by 500 ppmv in concentration over 10 h, which corresponded to an air exchange rate between the showcase and the church room of 0.8 showcase volumes per day.

Air chemistry
While the air of the crypt contained a mixture of trace gases originating from outdoor and indoor sources, the showcase only contained internally generated compounds (Fig. 13).

Indoor air pollutants
All sampling methods determined acetic acid to be the main organic component inside the showcase. In the IVL sampler, acetic acid was above quantification, i.e. in excess of 2500 µg m −3 , followed by formic acid quantified at 270 µg m −3 . Also detected, by passive sampling on Tenax TA (NILU and Eurofins methods), were furfural, acetone, acetonitrile, pinene, propanoic acid, tridecane and tetradecane, and a large amount of unidentified hydrocarbons (VOC sum c. 6000 µg m −3 , toluene equivalents). Acetic acid were by Tenax collected at a lower (less complete) amount; about  Table 3 below 1400 µg m −3 , than by the dedicated organic acid sampler by IVL, which could indicate an even higher total VOC concentration inside the showcase. By the qualitative sampling on SPME were, besides acetic acid, also detected monoterpenes and phenoxy derivatives.
The air of the crypt contained 15 µg m −3 organic acids, and c. 60 µg m −3 other VOCs including acetone. The outdoor air contained traces of acetic acid, which, however, was below the detection limit (< 1 µg m −3 ).

Ambient pollutants
The church is situated in an urban environment with much car traffic. Nitrogen dioxide, which is abundant in car exhaust gasses, was present inside the church room at about the same level as outdoors, 30 µg m −3 , while the much more reactive pollutant ozone was diminished indoors to 5 µg m −3 (6% of the ambient level). Inside the showcase both nitrogen dioxide and ozone were below the detection levels, i.e. < 0.5 µg m −3 .

Corrosion test
The lead coupon was, after 1 year's exposure inside the showcase, completely covered by grey-white corrosion, and had gained weight corresponding to 8.4 g per m 2 surface (Fig. 14).

Discussion
The average date of the nine textiles dates is 940 ± 15, and the calibrated age interval is 1045-1155 (1σ) and 1040-1158 (2σ), which is a large interval due to the nature of the calibration curve in the eleventh and twelfth centuries. However, within this uncertainty the nine textiles are all in accordance with a manufacture   13 The distribution of compounds in air between the showcase, the crypt of the church, and outdoors. VOC was not measured outdoor, < DL = measurement below detection limit date close to the enshrinement of the brothers AD 1100. All the tests conducted, individually and averaged, are in accordance with the expected time of AD 1100, and all the objects investigated are therefore as such authenticated.
Samples of the tibia of St Canute and the femur of his half-brother Benedikt were radiocarbon dated in 1999 [7]. According to historical sources mentioned above, Benedikt was murdered at the same time and on the same occasion as St Canute. The radiocarbon age and present-day calibrations are listed in Table 3. Even if the dating uncertainties are high, the dates agree with a death of both men in AD 1086 within the 1σ calibrated interval, and the dates are also in accordance with the textile dates of the present work. However, there are more complications to the history of the shrines and the textiles than that.
The wood of the coffin with the columns now holding the skeleton of St Canute has been dendrochronologically dated to after c. 1074 [39]. No bark or sapwood were present, so it is likely that the year of harvesting the tree was up to c. 25 after the year of the last year ring, 1074. This makes it probable that the timber was harvested c. 1100, the year of the enshrinement of St Canute.

Earlier investigations and a possible displacement of the textiles
The restoration of the reliquaries in 1875 included relocations of some smaller bones as well as most of the textiles. It was based on investigations by an expert commission which had been set up the previous year and which in 1886 on the 800th anniversary of the martyrdom of St Canute published a report about its results [20]. The members of the committee did not, however, all agree in the interpretations, and the discord has stimulated several later investigations of the reliquaries and their content of bones and textiles.
Dendrochronological investigations of the shrines and anthropological analyses of the skeletons have been carried out leading up to the 900th anniversary of the martyrdom in 1986 [40,41] and as a preparation for the publication of "Odense Sankt Knuds Kirke" (St Canute's Church in Odense) in the thorough inventory "Danmarks Kirker" in 1995 [1,7,39]. This has answered two of the most discussed questions which emanate from the chaotic find history of the reliquaries.
When the two shrines were found in 1694 the main informant, professor Thomas Bircherod (1661-1731), thought that the shrine with the hipped lid contained the remains of St Canute since only this shrine showed traces of mountings with rock crystals/precious stones as described when the shrine of the king was uncovered in 1582 [42]. Since Medieval sources state that King Canute himself had acquired relics of the two well-known English saints St Alban and St Oswald, Bircherod concluded that the other shrine, which had no lid and had been decorated with columns, must contain the remains of one of these. The identification of the bones in this shrine was already challenged by the son of Thomas Bircherod, the antiquarian Jacob Bircherod (1693-1737), who instead suggested Benedikt. Yet the idea of St Alban as the other saint remained viable even up to the new investigations in the 1990′s. As late as 1992 a new suggestion was put forward: namely that the other buried person might be the nephew of St Canute, King Erik 3. Lam, who became king of Denmark in 1137 but abdicated in 1146 and retired to become a monk in St Canute's monastery in Odense where he died [43].
After the investigations in the 1980′s and 1990′s it can be concluded that the so-called shrine with the columns was made c. 1100 [39], that it has had a lining with textiles as mentioned in the description of St Canute's shrine in 1582, and that it contains the bones of the elder of the two skeletons in question, a man, who met a violent death somewhere between 980 and 1200 (1σ) at an age of 35-45 years (Fig. 15). This reliquary can safely be identified as the shrine of St Canute (Fig. 1). A trace element analysis revealed a remarkably high content of Au in the skull which could be the result of the skull being kept in a separate, gilded reliquary, precisely like the one which Queen Christine donated around year 1500 for the head of King Canute as recorded in her account book [7]. Another explanation may be that the skull had rested for centuries on pillows with golden metal threads since golden textiles were mentioned in the description of St Canute's shrine when it was recovered in 1582. The shrine with the hipped lid (Fig. 2) could not be dendrochronologically dated but has usually from a stylistic point-of-view been judged to have a slightly younger date than the shrine with the columns. There are no traces of , who as far as is known died quite peacefully as a monk. It thus seems certain that the two deceased men can be identified with St Canute and his half-brother Benedikt [7].
Further investigations of the skeletons have not fundamentally changed this interpretation but have confirmed the violent death of the two men in the shrines and stressed that the structure of the bones and the skulls are so similar that it points to a family relation. Since the two men did not have the same mother it is not possible to prove the relationship with mitochondrial DNA analyses [44].
Several questions have been raised concerning the textiles in the shrines, especially their date and place of origin as well as the iconography of the two silks which are adorned with pictures: The Eagle Silk and the pillow with the bird motif [9,45,46]. It is the general and most likely assumption that the preserved silks must have been part of the costly gifts that were sent from South Italy to St Canute's shrine by his widow, Edel. As duchess of Calabria and Apulia Edel had access to high-quality silks from the Byzantine Empire. Perhaps the textiles were brought to Denmark by St Canute's half-brother, King Erik 1. Ejegod, who visited Italy in order to promote the sanctification of Canute and who is known to have visited Bari in Apulia on this occasion.
The textiles are the largest and best-preserved high-status textiles from the High Middle Ages in Denmark. At the time of Canute's canonization and enshrinement, silk weaving in Europe was not yet established outside the boundaries of the Byzantine Empire and silk was both a precious and a much-coveted import article. The colours, weaving, and motifs of the fabric determined its value even in a figurative sense since textiles played an important representative role in religious ceremonies as well as in the homage to rulers and princes. The preciousness of silk meant that it was considered a suitable fabric for wrapping up bones of saints, whether in tiny lead boxes hidden in the altar tables or in larger reliquaries like the ones in Odense.
The most recent technical analyses of the textiles have been published by Krag [46]. The yellow silk of the 148 × 35 cm monochrome pillow/mattress is linenwoven with a thin, firm warp thread and a woof of carded, loosely spun silk of a poorer quality than the one used for the warp. The appearance of the material has given it a certain resemblance to cotton, which is also the term used in older descriptions of the cloth. The yellow threads of these monochrome pillows were analysed to be dyed with Persian berries of the Rhamnus family [46]. The linen-woven white cloth with fringes, earlier thought to be an altar-cloth, measures 305 × 75 cm and has a selvage along one side. The other sides have been torn off. A number of holes placed rather at random along the edge may suggest that this cloth was once nailed inside one of the shrines as lining. The Eagle Silk (today c. 110 cm high and 133 cm wide, but originally at least 195 cm, perhaps even 230 cm wide, Fig. 5) is woven as samite, a fine weaving technique developed in Sasanid Persia (AD 226-661), in which the weft predominates so that the horizontal threads cover most of the weave without interruption creating a very smooth surface. The back of the textile is equally smooth, yet with the opposite colours. The weave of the Eagle Silk is close, and the warp is relatively coarse, z-spun, alternating in reddish-brown and undyed silk, while the weft is of unspun red and bluish-black silk in various thickness. The red threads were dyed with madder and sappanwood, while the dark blue colours were primarily achieved with woad and indigo, in agreement with the identification provided by HPLC-DAD in this study and by the data reported in Krag [46]. The pillow with birds juxtaposed in pairs consists of several pieces sewn together, of which one piece, 30 × 40 cm, covers the  (Fig. 4). They have been cut off without regard to the pattern. The material is very loose in structure. The weave is of samite type and has a closer density than the Eagle Silk. The warp consists of fine, almost unspun, undyed silk threads and the woof of similarly unspun, slightly golden-and light-blue silk threads varying greatly in thickness. It probably had three colours. The ground colour is today yellow, and the pattern appears in two faint blue shades. The yellow threads were dyed with redwood, Persian berries, indigo or woad, the blue ones with indigo or woad [46]. The organic analyses of the present work focused on samples not previously analysed (see Table 4).
The eagles on the Eagle Silk and the juxtaposed birds on the pillow, probably peacocks flanking a stylized tree or a cross, can be seen as symbols of Imperial Power and Christianity respectively, although the pattern with the peacocks originated in a non-Christian context. The production site of the Eagle Silk has been suggested as being Sicily/Southern Italy [8], Spain [47] or even the imperial workshops in Constantinople itself [48,49]. At the base under the foot of each eagle an inscription-like pattern is to be seen. Attempts have been made to find either Greek, Islamic, Armenian, or Georgian letters in this pattern, but none of the suggestions seem trustworthy. It is more likely that it is simply an ornamental imitation of such bands with inscriptions that could be depicted on imperial Byzantine silks. The quality of the Eagle Silk does not quite live up to the high standards of the imperial workshops. But it was surely produced in an area under Byzantine influence, while the pattern on the pillow with the birds has been compared to a group of Sasanid-Persian silks. It may have been made in a town situated on the Silk Road from China to the Mediterranean, perhaps Tashkent in Uzbekistan [9]. The silk has usually been dated to c. 900, meaning that it would be about 200 years old when it was re-used for the pillow in Odense. However, such dates have not been encountered in the present work. The new radiocarbon dates reveal that all the nine dated textiles from Odense can be dated inside the same chronological framework which fits very well with the enshrinement in 1100 or 1101, and it seems clear that a fine selection of contemporary textiles was sent to Denmark to be used for the enshrinement. There is no indication of a chronological difference between the textiles. The pillow with the birds was thus not a 200 years old fabric, and the somewhat coarse way that the silk for the pillow with the birds was cut up in order to suffice should not be interpreted as a sign of reuse. Rather it indicates that various whole pieces of silk of different size were send to Denmark where they were cut out and sown into bedcover, pillow-cases etc. for the reliquaries. The sewers had to make do with the available pieces of silk and in the case of the pillow with the birds they succeed in creating a nice front while it obviously did not matter so much how the back of the pillow looked. The remaining silk was simply cut out into small strips with no regard to the pattern and motifs.
It may be questioned if the present distribution of the silks in the two shrines is valid. The textiles comprise pillows, cushion, sheets, and blankets reflecting the bedclothes of a live person yet with an exclusive character [50]. As previously mentioned, all the best silks were found in the shrine with the hipped lid as well in 1694-96 as in 1833, and it was the expert commission of 1875 that Table 4 Comparison of previous work (Krag [46]) and this work decided to place most of them in the shrine with the columns-the shrine of St Canute. The argument was that the canonized royal saint must have deserved the best possible textile equipment. In that way the textiles of the two shrines were fundamentally interchanged so that the more humble, fringed linen textile which lay in the shrine with the columns ended up in the shrine with the hipped lid instead. Yet the experts were well aware that this new distribution of the textiles rested on a feeble ground, as they declared in December 1874 in a letter to the Ministry of Church and Education: "We are not able to judge in which of the shrines the found weaved fabric belong" [20]. The only archaeologically valid argument for the new placing of the textiles concerns the so-called red and blue "changeant"-coloured, thin silk because tiny pieces of this cloth is to be found under the nails on the inside of the shrine with the columns, thus indicating that it is indeed a remain of the original lining of that shrine.
In 1582 the lining of the shrine with the pillars was described as brown but that is not as strange as it sounds since the colours of this silk is not easily recognized in dim daylight inside the church and the silk even today at first glance gives the impression of being brownish. So, this textile may have been found inside the shrine with the columns and later removed from its original shrine to the shrine with the hipped lid sometime after 1582 but before the recovery of both shrines in 1694-96. The damages to the fabric seem to indicate that it was rather ruthlessly torn off. Since it was later found in the other shrine, we may deduce that the shrine with the hipped lid was accessible during or shortly after the despoiling of the shrine with the columns. However, the 1875 argument for placing all the best silks in the shrine with the columns is not convincing. The precondition for the argument that the best silks should belong in the king's shrine is that all original textiles are preserved. That is clearly not the case. The description of the king's shrine from 1582 mentions a double yellow silk with yellow silk embroidery as well as a golden piece of textile. These silks seem to have been lost. On the other hand, the description does not mention such characteristic textiles as the Eagle Silk or the pillow with the birds. The most reasonable conclusion is that most of the preserved textiles right from the beginning lay in Benedikt's shrine where they were also found in 1694-96 and in 1833. The silks are indeed fine and suitable for a saint. But as the somewhat enforced use of the silk for the pillow with the birds indicate, Benedikt's shrine could not necessarily claim the best silks or the best handicraft. The silks in St Canute's shrine were probably of even better quality. That might explain why most of the textiles which were found in the shrine in 1582 have disappeared, probably stolen like all the metal fittings on the shrine. The shrine was clearly plundered for all valuables. It is a likely-but unverifiable suggestion-that the shrine with the hipped lid was the shrine which is recorded to have been stripped of its silver and copper fitting in connection with the large confiscations of precious metal in 1536. This would not have involved the textiles which remained relatively safely inside the shrine with the hipped lid for the next 300 years. The Py-GC-MS analyses of the white textiles in the shrine with the hipped lid highlighted mainly the use of a cellulose-based material which would correspond with the theory of a slightly humbler equipment. Since the present distribution of these textiles does not necessarily reflect the original state of things it is not, however, possible to draw any safe conclusions.

Conservation and chemistry within the showcase
The showcases-almost museum pieces in themselvesseem at first glance to be a both logical and excellent way of exhibiting the bones, the textiles, and the shrines together as a collected and coherent relic and national treasure. But is this in fact a good solution? We have investigated the protection capacity of both the church in the urban environment and of the showcase within the church.
The air exchange rate of the showcase was low compared to the surrounding room: 0.8 day −1 . This is not low by today's standard for modern museum showcases but signifies that the showcase offers a considerable protection from the outside air, although the objects inside the showcase are still exposed to the ambient environment to some minor degree. On the other hand, it can be assumed that the exchange of air between the church and the outdoors is about one order of magnitude higher, so the weather will influence the indoor climate of the church room, especially with regard to moisture content. Although we did not measure the air exchange of the church room, we know from other measurements in similar Danish church buildings that the natural ventilation typically provides an air change of about 5 building volumes per day [51].
Climatically this provides a "box in a box" situation: The reliquary is housed in a closed glass box, which is placed inside a church building, which again is in a temperate climate urban environment. This is reflected especially in the hygrothermal behaviour, becoming increasingly more stable inside each "boxed" micro-climate (Fig. 12).
The indoor temperature during winter is controlled by central heating (radiators), while no temperature control is turned on during summer. The high thermal inertia of the heavy construction of the church ensures a very steady temperature, with no appreciable variations between the church room and the showcase.
The synchronous trace of how the indoor relative humidity follows variations in the outdoor temperature ( Fig. 12) suggests that the amount of water vapor inside the church is somewhat controlled by the outdoor air, which is also what is expected from the assumed ventilation rate of ca 5 day −1 . On the other hand, the relative humidity inside the showcase is almost unaffected by room-scale or outdoor humidity fluctuations; it is maintained at the same annual average level. It is wellknown that hygroscopic materials act as a humidity buffer when enclosed, and the stable humidity can be attributed to the large amounts of textiles and wood in the coffin.
Despite the winter heating, the indoor relative humidity was rarely below 50% RH, even though the outdoor air was drier (4 g m −3 less). This indicates an additional source of water vapor indoors, which very likely could be moisture evaporating from the walls or floor of the lowlying crypt.
The analyses of the relatively well-protected air space inside the showcase show the presence of several organic compounds, but mainly acetic acid, a natural deterioration product from organic materials. It is a viable interpretation that the VOCs are derived mainly from the wood pedestal or most likely the coffin; furfural and organic acids from degassing of degrading wood -especially oak, the monoterpenes from wood resin, and the phenoxy derivatives from the deterioration of lignin. On room-scale the level of organic compounds, including organic acids, was much lower than inside the showcase, due to the higher rate of dilution by the outdoor air. The detection of acetone may be attributed to human bio-effluents.
While the showcase by its slow air exchange maintained internally-generated compounds in high concentrations, it retarded at the same time external pollutants. Both ozone and nitrogen dioxide, which were present in the church room, were diminished in concentration to below the limit of detection inside the showcase.
The corrosion of the lead coupon exposed inside the showcase was not surprising, taking the presence of organic acids into account. Corrosion of lead artifacts is a common phenomenon in wooden showcases [52], and although we did not analyse the corrosion product it is very likely basic lead carbonate; caused by acetic acid in combination with atmospheric moisture and CO 2 . The corrosion rate of 8.4 g m −2 in 1 year suggest a highly corrosive environment [33].
Containment inside a low-ventilated showcase is a 'double-edged sword' for the preservation of heritage artefacts. The glass case clearly protects the objects from ambient air pollutants and retards the fluctuations in humidity observed in the surrounding air. As long as the temperature is kept steady at the current level the showcase ensures an overall acceptable and very stable hygrothermal environment. Likewise, it must be supposed that the showcase in the same way acts as a protecting barrier for particles; at least coarse dust, although this was not investigated.
On the other hand, acid air compounds in high concentrations could pose a potential hazard to the reliquary objects. Acetic and formic acid could deteriorate the metals (fortunately, no lead objects are present), the textiles, and the bones in the coffin. Although lower than the so-called 'acceptable damage concentration' suggested for acetic acid and bone by one study [53] we still consider the observed level inside the showcase a risk, as it have remained in an undisturbed environment for a very long time (more than a century). Probably the textiles are at the highest risk. As the source of the organic compounds is in the exhibit itself, there is no way to avoid the build-up in concentration of the organic acids other than either by an increased ventilation of the showcase, or by adding pollution scavenging materials somewhere inside the showcase. Ventilation conflicts with the protection from ambient climate factors, including the risk of soiling by dust, which could only be safely administered with a design with filtered air inlets. The use of adsorbents, e.g. activated charcoal, may be an effective mitigation strategy, however; it requires frequent control and replacement while exhausted, and it may pose a design challenge to locate the sorption media in correct positions inside the case.
Although ozone and nitrogen dioxide does not enter the showcase in any detectable amounts, the presence of either compound inside the church room still poses a risk to other, openly exposed heritage materials in the church room, such as dyes and pigments in paintings (see summary of literature by Tétreault 2003 [52]). The presence of indoor nitrogen dioxide at a level comparable to that of outside, warrants an explanation. In urban environments, nitrogen dioxide originated mainly from car exhaust, or is a reaction product from nitrogen monoxide (also from car exhaust) and ozone, which will be formed at times or places of little sunlight (at night, or indoors). The phenomenon of high indoor NO 2 concentrations was previously observed in a large art exhibition hall, not unlike the design of a church room [54]. In our case, the source of nitrogen monoxide is probably the busy roads surrounding the church, and ozone is naturally present outdoors. The indoor level of ozone is low, which, besides the loss from surface deposition may be attributed to consumption via air chemistry producing extra nitrogen dioxide indoors.