An analysis of some constructional features of a Bronze Age

burial mound at Clogher Lower, Co. Roscommon, Ireland.

D.W. Shimnwell, M. F. Robinson¹S C. Cribbin²

(¹School of Geography, University of Manchester and ²Tooraree, Ballyhaunis)

Abstract

A cross-section of a Bronze Age burial mound is described in terms of its constructional materials and chemical analyses of its various horizons. X-Ray fluorescence analyses of the chemical composition of topsoils and a calsilitic horizon above a boulder core reveal characteristic CaO/SiO₂ratios and concentrations of the diagnostic trace elements titanium A and zirconium. These features suggest the incorporation of fired and slaked lake marl mixed with esker sand as a mortar—like sealant layer during the process of mound construction.

Keywords

BRONZE AGE; BURIAL MOUND CONSTRUCTION; SEALANT LAYER

Introduction

During the course of construction of a farm access road in the townland of Clogher Lower, Co. Roscommon, Ireland in May 1993. a section was cut through a burial mound to reveal a central cist, empty apart from a few fragments of cremated human bone. Local information suggested that the cist had been entered in the I 920s and the contents removed. The find was reported (by GC) to the Office of Public Works which commissioned a preliminary survey of the burial mound, prior to the notification of the site and its inclusion in the Sites and Monuments Register for Co. Roscommon (Gibbons et al. 1992). The cross-section (Figure 1 and photograph 1) provided an ideal opportunity for a scientific study of construction materials and methods of a Bronze Age burial mound, a facet of archaeological study which is recommended by Waddell (1990). The site was surveyed by the authors on May 21, 1993 and samples of soil from various parts of the mound were taken for analysis with the specific aims of:

(i) comparing selected properties of the topsoils of the mound. the surrounding fields and the buried soil beneath the mound;

(ii) investigating the nature of what appeared to be either a calcrete. marl or

calsilitic sealant layer between the topsoil and the central stone core of the mound:

Nature of Construction Materials

The burial mound is located at circa 331ft (10I m) OD, National Grid Reference 15612707, on low-lying ground of the upper Suck catchment, north of the eastern end of the 450ft (138m) uplands of the Slieve Dart ridge to the south and the Carrickaphuca-Moanbane hill to the north. The surrounding land is reclaimed pasture developed on gley soils overlying a thin skin of glacial drift. The cross-section of the mound reveals a structure of the dimensions of 12m in diameter and 2.3m in height. A paved cist of fat sandstones, with internal dimensions of width 55cm, depth 70cm and height 42cm is located I.5m below the surface of the mound. It is surrounded by a core of rounded and sub-angular sandstones, limestones and felsites which are dry and lack soil in their interstices. One sandstone block bearing what appeared to be primitive carvings was recovered from the central core and lodged with the Office of Public Works.

The cist sandstones suggests quarrying from the Old Red Sandstones of Devonian age which outcrop along the Slieve Dart ridge, some I5OOm to the south-west, while the nature of the core stones indicates a glacial origin of scattered surface debris from the eskers and other drift deposits in the immediate vicinity. Beneath the stone core, the section reveals two low buried soil mounds with some charcoal fragments which suggest the presence of a circular structure as a primary cremation site, prior to the later construction of the cist—burial mound.

The stone core of the mound is sealed by a grey-white layer of material resembling calcrete, marl or calsilitic daub, varying between 5 cm and 10cm in thickness and water percolating through the layer has resulted in the secondary deposition of calcium carbonate as a thin flowstone scale on some of the stones in the upper part of the core. Above the calcrete/marl/calsilitic daub layer, the mound has been sculpted with topsoil to a depth of some 60cm close to the summit and to a greater depth at the outer edges by natural slump, downwash and poaching by cattle. The upper horizon and differentiated sealant layer have been broken by an intrusion some seventy years ago and the infill is clearly visible.

Characteristics of Soils

Samples from the topsoil overlying the mound, from the adjacent fields and from the buried horizons beneath the tumulus were analysed for colour (using standard Munsell Soil Charts) and pH and subjected to an aqua regia digestion according to Allen (1993) in order to determine the total concentrations of calcium, copper, lead and zinc by means of atomic absorption spectrophotometry. The results are presented in Table I below.

Table I. Characteristics of soils from Clogher Lower tumulus

*Sample*	*Colour*	pH	*Ca²⁺*	*Zn²⁺*	*Pb²⁺*	*Cu ²⁺*
		(%)	-------	(µg^g-1)	-------
Mound Topsoil 1(20cm)	lOYR 3/6	8.1	1.62	63	40	9.25
Mound Topsoil 11(40cm)	lOYR 3/6	7.9	1.59	60	44	9.50
Field Topsoil I (20cm)	1OYR 3/6	7.8	0.78	68	41	8.25
Field Topsoil 11(40cm)	lOYR 3/6	7.8	0.95	60	38	7.75
Buried soil I (north)	lOYR 3/3	7.7	0.65	38	29	5.50
Buried soil II (south)	IOYR 3/3	7.7	0.70	34	30	6.50

The colours of the soil samples proved to be relatively uniform, becoming lighter with presumed age, and the pH ranged from 7.8 to 8.1 in the samples of topsoil compared to a lower value of 7.7 in the buried soils beneath the tumulus. The calcium concentrations in the topsoils overlying the tumulus were greater by a factor of two than those of both horizons in the field topsoil and of both buried soil samples. The concentrations of zinc, lead and copper were also uniformly higher in the mound and field topsoils than in the buried soils.

Nature of the Sealant horizon

Samples of the topsoil, supposed sealant and flowstone scale deposits from the tumulus were taken for X-ray fluorescence analysis using the recommended methodology of Potts (1993) and micromorphological analysis following the procedures described by FitzPatrick (1984). In addition, samples of lake marls from marshlands in the neighbouring townland of Glenmore and from a core at Island Lake, Ballyhaunis, County Mayo and esker sands from Glenmore were analysed for comparison. A presentation of the major diagnostic features of the analysis of ten oxides and seventeen trace elements are presented in Table II from which it may be seen that the Clogher deposits have a uniform CaO/SiO2 ratio of 38/54% while the flowstone scales have a high CaO concentration of 94%. The lake and marshland marls from the upper horizons had CaO/SiO2 ratios in the general range of 53/38% and 56/40%; those from a greater depth are almost pure deposits of calcareous marl with a CaO concentration of 97%.The percentages of alumina in the Clogher samples are also significantly higher than those of other samples. The results of the analyses of esker sands are more variable in terms of both the silica/alumina and calcium oxide/silica ratios and reflect the differences in depositional environments. Of the seventeen trace elements analysed, titanium A and zirconium appear diagnostic in the differentiation of the sealant horizon of the cairn, lake marls and esker sands. The ranges of 1516-2073 ppm TiA and 168-227 ppm Zr in the former are notably different from the respective ranges recorded for the marl samples of 78-90 and 64-80 ppm, but the result for esker sand from Pollanalty shows some similarity to the Clogher sealant deposit.

In an attempt to recreate the sealant layer under laboratory conditions, the combination of 10g of lake marl and 20g of esker sand from Glenmore was undertaken in three processes: (i) combination of the natural materials with an appropriate volume of water; (ii) treatment of the marl in an oven at 800°C for one hour, followed by slaking and then mixing with sand; (iii) procedure (ii) after firing for two hours. After 48 hours, the results of procedures (i) and (iii) had failed to produce a hardened material, but procedure (ii) resulted in an indurated crust of a similar physical consistency to the sealant layer. Analyses of the crust yielded results of a remarkably similar pattern to the sealant layer for the presumed diagnostic characteristics.

Table II. X-ray fluorescent analysis of cairn, materials, lake marl and esker sand deposits

*Sample Location*		*CaO*	*MgO*	*SiO₂*	*Al₂O₃*	Ti	Zr
				(%)		(ppm)
Clogher topsoil	30cm	10.92	00.76	84.14	02.67	68	61
	50cm	11.45	00.70	82.62	03.88	73	66
Clogher sealant deposit	1	38.86	00.16	54.66	03.36	2073	222
	2	39.26	00.32	55.15	03.19	1516	168
	3	38.43	00.36	53.93	03.66	1841	227
Clogher flowstone scale	1	94.47	01.24	04.30	00.30	257	42
	2	94.09	01.23	04.28	00.30	238	44
Island lake marl	(0.5m depth)	52.75	00.57	38.00	00.70	90	64
	(4m depth)	97.33	01.53	00.45	00.06	87	70
	(6m depth)	97.80	01.54	00.45	00.06	93	78
Glenmore marl	(0.5m depth)	54.86	00.41	40.44	00.65	84	80
	(1m depth)	96.24	00.56	00.26	00.04	78	73
Glenmore esker sand	1	14.58	00.84	82.18	01.37	609	91
	2	00.04	00.23	95.03	02.75	732	54
Pollanalty esker sand		02.94	00.45	85.84	06.14	2217	261

Laboratory reconstruction		35.10	02.51	58.97	02.97	1199	222

Discussion

i. Calcium concentrations in soils

The results for the analysis of calcium concentrations in the mound topsoil are contrary to the normal expected situation in soils developed naturally through the processes of pedogenesis in regions of high precipitation and leaching (Chorley 1969, Curtis et al 1976, Trudgill 1977, inter alia). The normal, expected situation may be seen in the results from the analysis of the soil profile from the adjacent field, wherein deposition of calcium carbonate at depth due to a combination of water percolation through leaching and biological and chemical weathering, may be inferred. The concentration of calcium in these horizons is in the same approximate range as those from the buried soils beneath the tumulus. If leaching in a region of high precipitation is recognised as the dominant soil process, then the higher calcium concentrations cannot be explained by such a process, considering that the main effects of leaching would have been due to precipitation falling directly on to the mound rather than to a combination of direct rainfall and lateral water percolation as in the field situation.

The higher values for calcium in the mound topsoil suggest either an anomalous and extremely localised calcification process, typical of mid-latitude grassland soils subject to seasonal drought (Butzer 1964), or an extraneous source of calcium carbonate. Soil calcification would require a dry season to maintain a more or less permanently dry subsoil so that leached carbonates were soon precipitated as the soil water evaporated, with some upward migration of lime-charged soil moisture by capillary action. Explanation of the higher calcium concentrations in the mound topsoil would thus require the invocation of a protracted period of relatively dry climate with seasonal droughts in the centuries following the construction of the tumulus from circa 3000 yrs bp onwards. This situation would seem to be inconceivable for any environment located close to the Atlantic seaboard of western Europe.

ii . The sealant layer as a natural calcrete

The presence of a calcium oxide rich horizon with the basic characteristics of a calcrete is also quite anomalous in this particular topograhical situation. The nature and stages of development of calcretes have been the subject of considerable recent research (Goudie 1983, Tucker & Wright 1990, Wright & Tucker 1991). Such features usually develop in either alluvial/lacustrine or shallow water carbonate environments and are characterised by a series of phases of development, ranging from thin, discontinuous coatings of calcium carbonate on soil surfaces to continuous, often indurated matrices with calcium carbonate contents in the range 20-60%. There are seemingly no reports of the development of calcretes in mounded or hillslope environments. Wright (1990) recognises two main categories of calcrete, alpha and beta, the latter being more common in terrestrial situations and characterised by the presence of several biogenic features such as calcified microbial tubules and needle-fibre calcites. A micromorphological examination of the Clogher samples failed to reveal such features. The peculiar topographical location and the lack of biogenic features would thus suggest that the sealant layer is not a natural calcrete formed in situ.

A comparison of the CaO/SiO₂ratios of the sealant layer and the marsh and lake marls indicates quite distinct ratios for the latter deposits. Clearly, the sealant layer, with a 38/54% ratio is quite distinct from the virtually pure (97/2%) calcareous lake marls from deeper horizons and considerably different from the upper lake and marsh deposits with ratios of 53/38% and 56/40% respectively. The presence of the calcareous flowstone scale on the core stones suggests leaching of calcium carbonate from the sealant layer and deposition in the core. On the assumption that the original sealant deposits originated from the marshlands of Glenmore townland some 500m to the north-east, or from Coolcam turlough. 1000m in the same direction where such deposits have been reported by Coxon (1987a, b and pers. comm.), it may be possible that the leaching process over the past 2000 years has resulted in a reduction of the original marsh marl CaO percentage by 18%.

iii. The sealant layer as an artificial constructional feature

The characteristic CaO/SiO₂ratios and concentrations of titanium A and zirconium clearly distinguish the sealant layer from the mound topsoils and the lake manls. Titanium and zirconium are recognised as two characteristically stable trace elements and are most probably derived from the heavy mineral oxides rutile and zircon respectively, the most resistant minerals to weathering with the highest weathering index of 13 (Hampel 1968, Retallack 1990). Analyses of the assemblages of heavy minerals contained within sands have been traditionally used to determine the sources and transport pathways of sediments (van Andel 1959, Komar et al 1989). In a like manner, the high concentrations of titanium A and zirconium in the sealant layer would seem to indicate the incorporation of esker sand with marl from extraneous sources. The concentrations of the two trace elements in the mound topsoils are relatively low and it is difficult to conceive that they have accumulated towards the base of the profile by natural soil processes in the past 3000 years. It is, however, conceivable that the elements have been concentrated in the sealant layer by the firing and slaking of the marl, a process which would remove many of the minerals with a lower weathering index and leave a greater proportion of the more resistant heavy minerals. The inevitable conclusion is thus that the calsilitic sealant layer with a relatively high content of calcium carbonate is an artificial layer and an integral feature of the construction process.
Limbrey (1975) discusses the various aspects of the presence of calcium carbonate in archaeological excavations, noting the practice of marling with calcareous clays in the reclamation of lowland heaths and moorland, the deposition of calcium carbonate from mortar or plaster in occupation deposits and the presence of calcium carbonate layers in ditch fills. In the latter situation, it is suggested that recrystallisation of calcium carbonate in the lower part may be the counterpart of decalcification in the upper part, the boundary between the noncalcareous and calcareous soil often being very sharp and distinct, Limbrey also remarks that the formation of a soil over a structure such as a cairn built of stones, with no soil in their interstices, is a puzzling phenomenon but does not relate this feature to the presence of a sealant layer of relatively impervious material. Waddell (1990) concurs with this observation and implies that, with the exception of a mound at Ballybrew, Co. Wicklow, where impressions of Carex or Juncus stems were preserved in an encrustation of carbonate of lime on some of the paving slabs of the cist, there is a general lack of records of calcium carbonate deposition in such structures. Further, there are few examples of cross-sectional details of mounds which suggest a sealant layer of any description. There is the report of continous layer of sandy soil at a depth of 90 cm directly above an unprotected inhumation in an earth and stone matrix at Farta, Co. Galway and a mound at Ballyeeskeen, Co. Sligo had a dark brown band overlying a predominantly stone core (Waddell 1990).These may perhaps be tentatively interpreted as sealant layers. In contrast, there is a report of a distinct sealant clay layer over a core of granite rocks in the Neolithic burial mound of La Hogue Bie. Jersey, by Patton (1993). It may thus be that many archaeological excavations have simply overlooked the presence of a distinct sealant layer, a feature which may prove to be of importance in distinguishing funerary types and traditions.

There appear to have been four phases in the construction of the mound: (i) cist emplacement; (ii) cairn construction; (iii) sealant application; (iv) mound sculpting with topsoil. It is possible that all four phases occurred in close temporal sequence. This is suggested by the convex profile of the cairn and the maintenance of mean slope angles of 43° on the southern flank and 38°on the northern flank. Naturally occurring boulder-constituted slopes (screes and clitter) are concave to straight in profile and display modal slope angles of approximately 35° (Tinkler 1966). In the case of a loose cairn structure, lacking the external debris supply of a scree but still incohesive and subjected to transport processes, slope concavity might be expected to develop over time in the absence of any stabilising influence. With respect to the Clogher cairn, this stabilisation may thus have been achieved by the calsilitic sealant acting as a mortar. It is possible, therefore, that the cairn remained uncovered by topsoil for a considerable length of time. Alternatively, or additionally, the rapid completion of the structure with an earth covering would provide a higher degree of slope stability, especially as the sealant layer prevented the inwash of soil from the cairn surfaces to the interstices of the underlying uncompacted stones. The physical demands of construction suggest the possibility of the simultaneous emplacement of cairn stones. calsilitic seal and topsoil layer in an ascending process of building. In this case, the convex profile of the cairn could be protected and the risk of collapse avoided during the construction process. The application of the calsilitic seal could also be achieved from a working surface which allowed gradual closure towards an apical point.

The results of the analyses of soil characteristics and a review of the soil processes which might lead to the formation of the calcium carbonate rich horizon in the burial mound at Glogher would seem to suggest that horizon could not have been formed naturally in such a topographical situation in the past 3000 years. The particular CaO/SiO₂ratio and the high concentrations of the trace elements titanium A and zirconium indicate that the horizon was formed by the combination of lake marl and esker sand and applied as a sealant layer during the process of construction of the burial mound. The laboratory reconstitution of a compound of similar chemical characteristics suggests that the firing of marl to 800⁰C, followed by slaking and mixing with esker sand. The demonstration of the presence of a sealant layer recommends the careful excavation of burial mounds with a view to the definition of a typology of construction styles at different time periods during the Neolithic and Bronze Ages.

Thanks are due to Pat Donnegan for access to the site; to Michael Clarke and Paul Lvthgoe for laboratory analyses: and Graham Bowden for cartography.

Allen. S.E. (Ed.) (1989) Chemical Analysis of Ecological Materials. Second Edition. Oxford, Blackwell Scientific Publications.
Butzer. K.W. (1964) Environment and Archaeology An Introduction to Pleistocene Geography. London, Methuen.
Chorley, RJ. (1969) The role of water in rock disintegration. In: Chorley, RJ. (Ed.) Water, Earth and Man. London, Methuen.
Coxon. C.E. (1987a) The spatial distribution of turloughs, Irish Geography 20: 11-23.
Coxon. C.E. (1987b) An examination of the characteristics of Turloughs, using multivariatestatistical techniques. Irish Geography 20: 24-42.
Curtis. L.F.. Courtney, F.M. & Trudgill. ST. (1976) Soils in the British Isles. London, Longman.
FitzPatrick, EA. (1984) Micromorphology of Soils. London, Chapman & Hall.
Gibbons, M.. Alcock, 0.. Condit, T. & Tunney, M. (1992) Sites and Monument Record. County Roscommon. Sites & Monuments Record Office, Dublin.
Goudie, AS. (1983) Calcrete. In: Pye, K. (Ed.) Chemical Sediments and Geomorphology. London, Academic Press.
Hampel, C.A. (1968) The Encyclopedia of Chemical Elements. New York, Reinhold Book Corporation.
Komar. PD., Clemens, KE., Li. Z. & Shih. S-M. (1989) The effects of selective sorting on factor analyses of heavy-mineral assemblages. Journal of Sedimentary Petrology 59: 590-596.
Limbrey, S. (1975) Soil Science and Archaeology London, Academic Press.
Patton, M. (1993) Statements in Stone. Routledge, London.
Potts, P.J. (1993) Laboratory Methods of Analysis. In Riddle. C. (Ed.) Analysis of Geological Materials. New York, Marcel Dekker Inc.
Retallack. G.J. (1990) Soils of the Past Boston, Unwin Hyman.
Tinkler. K.J. (1966) Slope profiles and scree in the Eglwyseg Valley, North Wales. Geographical Journal 132: 379-385.
Trudgill, ST. (1977) Soil and Vegetation Systems. Oxford, Clarendon Press.
Tucker. ME. & Wright. V.P. (1990) carbonate Sedimeniology. Oxford, Blackwell Scientific Publications.
Van Andel. T.H. (1959) Reflections on the interpretation of heavy mineral analyses. Journal of Sedimentary Petrology 29: 153-163.
Waddell, J. (1990) The Bronze Age Burials of Ireland Galway, Galway University Press.