HAL Id: jpa-00215727
https://hal.archives-ouvertes.fr/jpa-00215727
Submitted on 1 Jan 1974
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
MÖSSBAUER STUDY OF FIRING CONDITIONS USED IN THE MANUFACTURE OF THE GREY
AND RED WARE OF TURENG-TEPE
R. Bouchez, J. Coey, R. Coussement, K. Schmidt, M. van Rossum, J.
Aprahamian, J. Deshayes
To cite this version:
R. Bouchez, J. Coey, R. Coussement, K. Schmidt, M. van Rossum, et al.. MÖSSBAUER STUDY OF FIRING CONDITIONS USED IN THE MANUFACTURE OF THE GREY AND RED WARE OF TURENG-TEPE. Journal de Physique Colloques, 1974, 35 (C6), pp.C6-541-C6-546.
�10.1051/jphyscol:19746115�. �jpa-00215727�
JOURNAL DE PHYSIQUE Colloque C6, supplkment au no 12, Tome 35, Decembre 1974, page C6-541
MOSSBAUER STUDY OF FIRING CONDITIONS USED
IN THE MANUFACTURE OF THE GREY AND RED WARE OF' TURENG-TEPE
R. BOUCHEZ
Institut des Sciences NuclCaires, UniversitC de Grenoble, France J. M. D. COEY
Laboratoire des Transitions de Phase, C. N. R. S. Grenoble, France R. COUSSEMENT, K. P. SCHMIDT, M. VAN ROSSUM Instituut voor Kern- en Stralingsfysika, Universiteit te Leuven, Belgium
J. APRAHAMIAN
Institut Dolomieu, Universitt de Grenoble, France J. DESHAYES
Institut d'Art et d'ArchCologie, UniversitC de Paris I, France
RBsumB. - Des tessons de ceramique ancienne, des paleo-sols et des sols recents de la plaine de Gorgan du N-E de 1'Iran sont etudies par spectroscopie Mossbauer, par diffraction de rayons X et par analyse klkmentaire, pktrographique et mineralogique. Les especes minerales argileuses et les mineraux, surtout ferriferes, ont ete identifiks. Des essais de cuisson dans l'air ou dans une atmo- sphere reductrice ont BtC realises de 650 B 1 150 OC.
L'on a trouvk que la separation des doublets quadrupolaires du fer 3+ et du fer 21- est fonction de la temperature de cuisson ; ce qui permet d'estimer la temperature utilisCe par les anciens potiers.
Les tessons de ctramique rouge et de ceramique grise de Tureng Tepe IIA (3 000-2 800 B. C.) examines sont respectivement oxydes et rkduits par rapport aux sols. Leur temperature de cuisson est respectivement (1 000 + 50) OC et (980 f 50) OC ; la ceramique grise semble donc cuite a une temperature un peu plus faible (50 a 100 OC) que la ceramique rouge.
Abstract. - Ancient potsherds, paleo-sols, and modern soil samples from the Gorgan plain of Northeastern Iran have been characterised by Mossbauer spectroscopy, X-ray diffraction, elemental and mineralogical analysis. The clay and other mineral species, and especially the iron-bearing minerals in the soils were identified, Test firings were carried out in the temperature range 650- 1 150 OC in air or in a reducing atmosphere.
The quadrupole splittings of the ferric and ferrous doublets are functions of the firing tempera- ture, and thus the temperature used by the ancient potters, can be estimated.
Sherds of the red and grey ware from Tureng Tepe IIA (3 000-2 800 B. C.) were examined, and respectively found to be oxidized or reduced compared with the soils. Firing temperatures of (1 050 50) OC and (980 f 50) OC were deduced for samples of the two types of pottery thus the grey ware could be fired at a temperature of 50 to 100 OC lower than the red ware.
1. Introduction. - Potsherds are a primary source of archaeological information, and any method of characterising them which can reveal their date, place or conditions' of manufacture, or the uses to which they were put is potentially valuable to archaeologists.
Many physical methods have been employed in recent decades, as recently reviewed in Science in Archaeo- logy [I], notably elemental, mineralogical and petro- graphic analyses by many techniques, including
Mossbauer spectroscopy, and also measurements of the mechanical, magnetic and thermal (especially thermoluminescent) properties. The Mossbauer studies of ancient pottery to date have been reviewed by Kostikas, Simopoulos and Gangas [2]. In practise it is often necessary to use both scientific and archaeo- logical methods in order to solve an archaeological problem [3]. Here we will briefly state such a problem, and then describe the progress made towards resolving
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19746115
C6-542 R. BOUCHEZ, 3. M. D. COEY, R. COUSSEMENT. K. P. SCHMIDT, M. VAN ROSSUM, AND COLL.
it using several analytical techniques, but with the emphasis on the results obtained by Mijssbauer spectroscopy.
2. The historical problem. - Two quite different kinds of pottery are found in the Tureng Tepe culture in the Gorgan plain of Iranian Turkestan [4] at the beginning of the third millenium B. C. One is red, with black decoration painted on the surface, the other is completely grey and polished. The two types of ceramic coexist over five archaeological levels comprising the period IIA (several centuries), and then the red painted ware disappears but the grey ware persists throughout the periods IIB and 111 up until the 17th century B. C. Just after its disappearance in north-eastern Iran a new late grey ceramic, appa- rently derived from the previous one, makes its appea- rance e. g. in north-western Iran North India, in regions which are known to have been populated by Indo-Europeans. Therefore the grey ceramic could be typical of Indo-European populations [4].
It is important to decide whether the grey ware came to north-eastern Iran at the beginning of the third millenium with people migrating from elsewhere (possibly nomads from the central asian steppes) or whether its appearance is connected with a change of manufacturing technique developed in situ. Our knowledge of the occupation of Iran by Indo-Euro- pean peoples depends on which hypothesis is the correct one. Either Indo-European populations origi- nated in the Gorgan plain, or else it was just a stage in their migrations.
A rather complete program of scientific analyses of these bronze age grey ceramics is being undertaken in order to clarify this problem. Before discussing the present Mossbauer results, it is necessary to outline some earlier results and describe the materials studied.
3. Previous results. - Grey and red sherds from Tureng Tepe IIA, dating from the beginning of the bronze age ( x 3 000 B. C.) have been compared from several points of view in earlier work 15-81, The distri- bution of major (Si, Al, Fe, K), minor (Ti, Ba) and some trace elements (Mn, Cr, Rb) were determined by X-ray fluorescence [7] and a semi-quantitative picture of all the trace elements present in concentra- tions greater than a few parts per million was obtained by spark mass spectrometry [5, 61. The major mineral phases were determined by X-ray diffraction [7], and a preliminary petrographic analysis of thin sections was made. Five sherds of each type were analysed and it was shown that there were no significant differences either in their elemental composition on in their quartz content. Signs of rubifaction of tourmaline in the red sherds suggested that they were fired at tempe- ratures in excess of 950 OC in an oxidizing atmosphere, blue-green tourmaline being a minor phase (0.5 %) in all the ceramics.
The most striking difference between the two types
of ceramic appeared in their Mossbauer spectra [5, 81.
The total amount of iron of the ceramics is about 5 %.
The iron in the grey sherds was predominantly ferrous ( x 70 %) whereas in the red sherds it was almost all ferric ( z 90 %). The same relative amount (15 %) of iron was present in both ceramics in a magnetically ordered phase identified as hematite (a-Fe,O,) from its hyperfine field of 513 kOe. This result suggested that the essential difference between the two ceramics was the firing process, e. g. the atmosphere present during the firing cycle. We therefore used Mossbauer spectrometry in parallel with the X-ray crystallogra- phic method and petrographic study in order to try to learn more about the origin and firing conditions of the pottery.
4. Materials. - The first group of materials studied were a set of soil samples taken at Tureng Tepe, and at other sites on the Gorgan plain.
C 1. A red paleosol taken beside the wall of kiln
(( F D, excavated close to Tureng Tepe, dating from period 111 (2 500-1 700 B. C.) and used to manu- facture grey ware [9].
C 2. A red soil from the top of a hill south of Gorgan used by the present-day potter.
C 3. A grey paleosol taken near the kiln (( D D,
also excavated at Tureng Tepe, and dating from period I11 [4, 101.
C 4. Another soil sample used by the present-day Gorgan potter taken from his workshop.
C 5. A grey soil used by a brickworks 61 km west of Gorgan for the manufacture of red bricks in an open furnace with a strong draught. The furnace was fuelled with thick black oil.
These samples are all part of the thick loess deposit on the Gorgan plain. Their colour depends mainly on their degree of humidity. For example, if C 2 is put in water, it resembles C 5.
The second group of materials were the soils fired in ancient times. They include red and grey sherds from Tureng Tepe IIA and grey sherds from Tureng Tepe 111. These are denoted by the letters R and G.
In addition some samples from the kiln F were exa- mined. They were taken from the grey-green inner layer of the wall (F I), the vitrified hearth (F 2) and some grey sherds from Tureng Tepe I11 embedded in the walls.
The third group of materials were samples of C 2 and C 3 clay fired under laboratory conditions using an electrical furnace, different firing temperatures and either an oxidizing or a reducing atmosphere. The oxidizing atmosphere was simply air, whereas the reducing atmosphere was achieved by burning sawdust in the furnace. The firing time was one hour in the first case and three hours in the second. The samples fired in reducing or oxidizing atmospheres resembled in appearance the grey or red sherds.
M~~SSBAUER STUDY OF FIRING CONDITIONS USED IN THE MANUFACTURE (3-543
5. Results. - Mossbauer absorbers were prepared with about 80 mg/cm2 of soil or powdered ceramic, which is equivalent to less than 0.1 mg/cm2 57Fe. In the case of the sherds, the material was taken from the central part, the surface layer(= 2 mm) being discarded.
Sources were either 57Co in CU or Rh, but all isomer shifts are expressed relative to an iron metal absorber.
Most of the data were collected with source and absorber both at room temperature, but some with source and absorber at liquid nitrogen or liquid helium temperature as will be discussed later.
The spectrum of a typical soil sample C 1, is shown in figure 1. There are three main elements in the spectrum : I a ferric quadrupole doublet, I1 a ferrous quadrupole doublet and I11 a ferric magnetic hyperfine pattern. Neither the relative intensities of the three elements nor their hyperfine parameters (except for the ferric quadrupole splitting) vary significantly
I I I
-8 -4 0 4 8
t VELOCITY ( m m l s ) +
1.35
-2 -1 0 2
VELOCITY ( r n m / s > -
FIG. 1. - Room temperature Mossbauer spectra of the soil sample C 1. The least squares fits are shown by the solid lines.
The three elements in the fit, I, I1 and I11 are the ferric doublet, the ferrous doublet and the ferric hypefine pattern referred to in the text. The lower spectrum is the central part of the upper
one, but on an expanded Gcale.
among the five soil samples. The data, summarised in table1 show that the soils are rather homogeneous in composition over a wide area, at least in so far as the iron-bearing minerals are concerned. Information about the identification of these minerals can be obtained from the Mossbauer parameters. Unlike the other patterns, doublet I cannot be un ambiguously identified with a single mineral species, because the lines are broad (0.74 mm/s) and the quadrupole splitting rather variable. Minerals with similar qua- drupole splitting and isomer shifts are the clays illite, montmorillonite [ I l l and kaolinite [12], and also amorphous ferric hydroxide [13] and very fine particles ( 5 200 A) of goethite and hematite [14]. The last three possibilities can be largely eliminated by the evidence of the Mossbauer spectra of the fired clay taken at room temperature and liquid helium tempe- rature as will be discussed below. Doublet I1 is iden- tified with chlorite, for which 6 = 1.13 mm/s and A = 2.61 mm/s [15] and pattern I11 is identified with hematite with a particle size of at least 200 A.
X-ray analyses were subsequently made of the whole sample and the clay fraction (< 2 p). The clay fraction was abundant and contained illite, kaolinite and chlorite in approximately the ratio 3 : 1 : 1. Mont- morillonite was not found and interlayer clays were found only in very small quantities. The coarser frac- tion contained quartz ( z 15 %) and calcite (= 5 %).
Hematite was not detected by X-rays, but this is only to be expected as it constitutes less than 0.5 percent of the total soil. The assignments made from the Mossbauer spectra are thus consistent with the iden- tification of the minerals in the soils by X-ray diffraction.
Some typical results for the red and grey sherds are shown in figure 2. If we assume that the recoilless fractions are identical for the three components, their relative abundance is given by the ratio of their corresponding intensities. The intensity ratio for the patterns I, I1 and I11 is 70 : 13 : 17 for the red sherd,
MiiSSbauer parameters of five unfired clays from the Gorgan plain. The isomer shifts 6 are given relative to iron metal. Fe3+ and Fe2+ are typically six-coordinated : 6, = 0.36 and a,, z 1.1 mm/s.
(The relative precision is respectively 3 % and 2 %for 6 ofthe ferric doublet I and of the ferrous doublet 11, 3 % and 1 %for the quadrupole splitting A of the doublets I and 11, and 1 %for the internal magnetic field He,,.)
- ---
-1 6
-
/-- I 7 A ----A- -Clay A I 6 He,,
% mm/s mm/s % mm/s mm/s % mm/s kOe
C6-544 R. BOUCHEZ, J. M. D. COEY, R. COUSSEMENT, K. P. SCHMIDT, M. VAN ROSSUM, AND COLL.
98
96
T:4'K 94
-10 -5 0 5 10
VELOCITY (mrnls)- VELOCITY (rnrn1s)-
FIG. 2. - Mossbauer spectra measured at room temperature T = 300 K and helium temperature T
-
4 K for a grey and a red sherd from Tureng Tepe IIA. The least squares fits are shownby the solid lines.
and 13 : 72 : 15 for the grey one. The average of the unfired soil samples is 58 : 3 1 : 1 I thus the red sherds are somewhat oxidized and some of the grey ones are greatly reduced, as compared with the soil. There is only a small difference in the intensity ratio of the three elements of the spectra taken at room tempera- ture or at liquid helium temperature, so that super- paramagnetic particles of iron oxides can represent at most 20 % of the total iron in the sherds.
Although the parameters of pattern I11 are similar in the sherds and the soil, there are differences in doublets I and 11. Differences are expected because the firing process disorganises and eventually destroys the clay minerals and promotes their recrystallisation as three-dimensional silicates [16, 20, 211. By X-ray diffraction analysis the only traces of illite were found in a few of the grey sherds, and since the illite peaks disappear between 950 OC and 1 000 OC [19, 221, the firing temperature of the grey ware should be in this range.
The mineralogical transformations induced in clay by firing depend on temperature, time and atmosphere offiring. It is possible to monitor these transformations using the 57Fe Mossbauer effect. MacKenzie [16] has studied the transformations produced by heating kaoli- nite up to 1 280 OC in air, but there is no such Moss- bauer work on the other clay minerals. However the mineralogical transformations produced by heating clay minerals in air (but not under reducing condi- tions) had been extensively studied by X-rays [19, 221.
The valence state of the iron, and the transformations themselves should be sensitive to the firing atmosphere and to the impurities present in the clays. We have taken Mossbauer spectra of the sample C 2 fired in oxidizing conditions at temperatures between 650 OC and 1 150 OC, and of the samples C 2 and C 3 fired at 850 OC and 950 OC in both oxidizing and reducing conditions in an effort to determine the firing condi-
tions and temperature for the ancient pottery of Tureng Tepe. In the oxidizing firings, the intensity of doublet I1 (ferrous iron) falls almost to zero at temperatures in excess of 700 OC. Pattern I11 (hematite) increases from 11 to 25 % of the total iron amount due to the crystallisation of amorphous or micro- crystalline iron oxides or hydroxides which contri- buted to doublet I in the unfired sample. Some hema- tite may also be formed directly from the clay minerals, hence the conclusion above, that the ferric oxides and hydroxides do not give the main contribution to the intensity of the doublet I in the soil. This is in contrast to the behaviour of Attic soils [2].
The variation of the quadrupole splitting of doublet I in the red air-fired samples is shown in figure 3. Similar
-/ y GORGAN CLAY C2 4
A DEFORMATION I
FIG. 3. - The splitting of the ferric doublet I versus the firing temperature for the clay C 2 fired in oxidizing conditions. The values measured of the splitting parameter A of the sherds R 3
and R 4 of Tureng Tepe IIA are also marked.
behaviour has been reported by Janot and Delcroix [I], and had also been indicated in the first Mossbauer effect experiment on ancient pottery by Cousins and Dharmawardena in 1969 [18]. In our measurement the isomer shift is constant at 0.35 mmls. The quadrupole splitting A , of the ferric doublet shows a broad maxi- mum near 750 OC wherc it reaches the value of 1.3 mm/s.
In this region there is also a 50 % increase in the line- width. At higher temperatures it decreases and approaches the value 0.9 mmls. The same behaviour is also found for doublet I1 in the grey samples fired under reducing conditions, though the isomer shift for these samples is different (0.46 mmls). The large increase in A , up to 750 OC as compared with the unfired soil is explained by the dehydration and dehydroxylation of illite and kaolinite. The clay lattice is left very disorganized with a variety of defects and asymmetric coordinations for the iron, resulting in an increase of the electric field gradient and in an increase in the linewidth. The subsequent slow decrease in A is
the result of a gradual recrystallisation of the disor- dered clay minerals as three-dimensional silicates (particularly spinel followed by mullite [21, 221) or even vitrification, which may occur in the same tempe- rature range and which depends strongly on the concen- tration of fluxing impurities such as CaO, Na,O, K,O or FeO in reducing conditions. Typically it is known 1231 that the formation of silicates from ferrous clays occurs at a temperature of about 50-100 OC lower under reducing conditions than under oxidizing conditions.
The contraction of the fired clay in this temperature range is the basis of the thermal expansion method developed by Tite [20] in order to determine the firing temperature of ceramics. The quadrupole split- tings A , of two red sherds (R3 and R,) are marked on figure 3, and we deduce that the firing temperature was not lower than about 950 OC. This value is in agreement with the rubifaction of the tourmaline mine- ral observed in these red sherds. However, this lower limit should be increased to 1000 OC, because there are no illite X-ray peaks observed in the red sherds.
An upper limit ( x 1 100 OC) is estimated from the defor- mation temperature of the fired clay C 2. Complete vitrification appears for this clay at 1 225OC. Therefore the firing temperature of these two red sherds should be in the range 1 050 + 50 OC.
The clays fired in a reducing atmosphere suggest a characteristic temperature dependence for the ferrous quadrupole splitting. Values of A were obtained for the laboratory-fired clays at 850 and 950 OC, and a value at higher temperature ( x 1 1500C) was derived from the vitrified hearth of the kiln F in which the iron was 85 % ferrous. A monotonic decrease is found in the range 850-1 150 0C. From the quadrupole splittings for three grey sherds a firing temperature of x 980 OC can be deduced. The illite X-ray peaks subsisting in some grey sherds suggest an upper temperature limit of 950-1 000 OC for these sherds. The firing temperature of the grey sherds could be thus 980 + 50 OC. Therefore the grey sherds seem to have been fired at a lower temperature than the red ones.
Further studies of the phase transitions of clay minerals fired under reducing conditions seem to be important.
6. Conclusion. - The characterisation of Gorgan soils and ancient ceramics of Tureng Tepe IIA in the Gorgan plain allows us to draw the following conclu- sions.
The soils in the Gorgan region are rather uniform, and probably very similar to the paleosols.
The clay fraction contains illite (50-60 %), chlorite (20-30 %) and kaolinite (15-20 %) but practically no montmorillonite or interstratified clays. The quartz fraction is about 15 % for these soils, and the calcite is about 5 %. The iron is found in the ferric form in the illite and kaolinite (50 % of the total iron) and in the ferrous form in the chlorite (20 % of the total iron).
The rest of the iron is present as hematite and crypto- crystalline ferric oxides and hydroxides.
A principal difference between the red and grey ceramics is the relative amount of Fe3+ to Fe2+. The iron in the red sherds is oxidized relative to the soils, whereas the iron in the grey sherds is reduced.
Laboratory firings of the Gorgan soils in oxidizing or reducing conditions produced red or grey ceramics, although the conditions used cannot have exactly reproduced those of the ancient pottery. The grey and red sherds each contain more ferrous iron than the corresponding grey or red laboratory-fired ceramics.
Nevertheless it is very likely that the grey ware was fired in a strongly reducing atmosphere, and the red ware in a partly oxidizing one.
The ferric and ferrous quadrupole splittings together with the X-ray spectra may be used to set limits upon the firing temperature used in the manufacture of ancient pottery, provided that samples resembling the paleo-sols are available for calibration firings. The ferric quadrupole splitting decreases rapidly from 800 to 9500C, but more slowly between 950 and 1 150 OC. Therefore great accuracy cannot be achieved in the latter range by Mossbauer spectroscopy alone.
Comparing Mossbauer and X-ray data, firing tempe- ratures of (1 050 + 50) OC and (980 + 50) OC were derived for two red sherds and one grey sherd from Tureng Tepe IIA. Using the X-ray data alone, the firing temperature of five samples of the red ware is above 1 000 OC because the illite peaks present in the local clays disappear in all the red sherds. On the contrary, the illite peaks are still present in some samples of the grey ware showing the grey ware could be fired at a temperature 50 to 100 OC lower than the red ware. Many more sherds must be examined before it can be affirmed that this difference is a characteristic feature of the two types of pottery.
This Mossbauer study reveals a characteristic beha- viour when a clay is fired in oxidizing conditions : a disorder state is created in the clay lattice after dehy- dration and dehydroxylation ( x 650 OC) followed at x 900-1 000 OC by ordering and silicate formation.
Further Mossbauer spectroscopy and X-ray diffrac- tion studies are necessary to determine the phase transitions especially in reducing conditions, for diffe- rent firing temperatures, firing times and reducing atmospheres. These studies should be done using local soils and corresponding ancient pottery from different historical periods, and in connection with the historical evolution of pottery kilns.
To solve an archaeological problem, such as the origin, the evolution and the spread of the grey ware of ancient Iran, it emerges from this study that the Mossbauer spectroscopy should be used together with other analytical methods, such as X-ray diffraction, petrographic and mineralogical studies, and also ele- mental analysis. Finally, not only the sherds, but also local clays resembling the sherds and fired in the laboratory should be analysed.
As a late result, on one sherd belonging to the
36
C6-546 R. BOUCHEZ, J. M. D. COEY, R. COUSSEMENT, K. P. SCHMIDT, M. VAN ROSSUM, AND COLL.
period IB from Tureng Tepe (V millenium) has been analysed, but only by Mossbauer spectroscopy. The quadrupole splitting of the ferric doublet is found to be (1.02 $. 0.02) mm/s. Compared with the results obtained on the sherds of the red ware from Tureng Tepe IIA, we can conclude that the firing temperature of this sherd from Tureng Tepe [I is lower by about 150 OC, giving an estimation of the temperature of (900 + 50) O C This result is valid provided that the shed was fired in the region with a local clay, and not
mported.
Acknowledgments. - We are deeply grateful to L. Courtois and P. Bordet for illuminating discussions,
and for performing preliminary petrographic and mineralogical analysis of some sherds of Tureng Tepe IIA.
We would like to thank M. Blanc for his help in firing the soil samples, and for his invaluable comments on the manufacture of the ancient ceramics.
The authors would like to thank also J. De Bruyn, G. Langouche and A. Meykens for their help in the Mossbauer experiment.
The authors are very grateful to Mrs. M. B. S&le for her kind assistance in the sample preparation.
One of us (R. Bouchez) would like to express his appreciation to Prof. P. Mariens for his kind hospita- lity during his stay in Leuven.
References
111 Science in Archaeology, Brothwell D . and Higgs E. S.
(editors) 2nd edition (Thames and Hudson, London) 1969.
[2] KOSTIKAS, A., SIMOPOULOS, A. and GANGAS, N. H., J.
Physique Colloq. 35 (1974) C 1-107.
[3] WILLIAMS, J. L. W., JENKINS, D. A. and LIVENS, R. G., J. Archaeolog. Sci. 1 (1974) 47.
141 DESHAYES, J., Archaeology 22 (1969) 10 ; Ugaritica 6 (1969) 139.
[5] STEFANI, R., BERTHOD, J., PELLET, J., BOURGUILLOT, R., CORNU, A. and BOUCHEZ, R., Proceedings of the Inter- national Conference on the Application of Nuclear Methods in the Field of Works of Art, Rome (1973) (to be published).
[6] STEFANI, R., BOURGUILLOT, R., CORNU, A. and BOUCHEZ, R., Centre de Recherches ArchCologiques du CNRS, Note no 2 (1972).
[7] BOUCHEZ, R., SCHMIDT, K. P., VAN ROSSUM, M., LAN- GOUCHE, G., MEYKENS, A., COUSSEMENT, R., APRAHA-
MIAN, J. and DESHAYES, J., Proceedings of the Internatic- nal Conference on the Application of Nuclear Methods in the Field of Works of Art, Rome 1973 (to be pu- blished).
[8] SCHMIDT, K. P., VAN ROSSUM, M., MEYKENS, A., LAN- GOUCHE, G., COUSSEMENT, R. and BOUCHEZ, R., J. Physique Colloq. 35 (1974) C 1-105.
[9] DESHAYES, J. and BOUCHEZ, R. (to be ptiblished).
[lo] DESHAYES, J., Iranica Antiqua VI (1966) ;
DELCROIX, G. and HUOT, J. L., Syria 49 (1972) 1.
[ l l ] MALATHI, N., PURI, S. P. and SARASWAT, I. P., J. Phys. Soc.
Japan 26 (1969) 680.
[12] MALDEN, P. J. and MEADS, R. E., Nature 215 (1971) 844.
[13] COEY, J. M. D. and READMAN, P. W., Earth Planetary Sci.
Letters 21 (1973) 45.
[14] VAN DER KRAAN, A. M., Thesis, Technical University of Delft (1972).
[15] COEY, J. M. D., SCHINDLER, D. W. and WEBER, F., Can. J.
Earth Science (1974).
[16] DUNCAN, J. F., MACKENZIE, K. J. D. and STEWART, D. J., Trans. Faraday Soc. Symposium (1968) 103 ;
MACKENZIE, K. J. D., Clay Minerals 8 (1969) 151.
[17] JANOT, C. and DELCROIX, P., Centre de Recherches Archko- logiques du CNRS, Note no 4 (1974) ; J. Physique 35 (1974) C 6.
[18] COUSINS, D. R. and DHARMAWARDENA, K. G., Nature 223 (1969) 733.
[19] PERINET, G., Transactions of the VIIth International Cera- mic Congress, London (1960) 371.
[20] TITE, M. S., Archeometry 11 (1969) 131.
[21] The X-ray identification and crystal structures of clay minerals. Brown G. (Editor) (Mineralogical Society London) (1961).
[22] GRIM, R. E. and KULBICKI, G., Bulletin Soc. Fr. Ciramique 36 (1957) 21.
[23] BODIN, V., Technologic des Produits de terre cuite (Gau- thier-Villars, Paris) 1956.
