Alan Smith and Bruce Yardley



To be published in the

Proceedings of the Cumberland Geological Society Volume 7 part 3

October 2008







Alan Smith and Bruce Yardley



‘Lithophones' (or tuned rocks) are essentially percussion instruments that create melodies and harmonies by striking slabs of rock laid out on some type of horizontal rack. Rock xylophones are perhaps a simpler definition. They are known in a wide variety of forms from around the world and many date from quite ancient times.


The so called ‘Musical Stones of Skiddaw', currently housed in the Keswick Museum and Art Gallery, are an outstanding British example and represent one of the largest and most complex types of these musical instruments. Built by the Richardson family in the early Nineteenth Century the instrument achieved fame as it was played nationwide in a touring musical extravaganza, culminating in 1848 with performances before Queen Victoria . (Figure1). Several other similar, but smaller sets of these stones exist, all built from slabs of hornfels derived from the Skiddaw Group in the inner aureole of the Skiddaw Granite.


For the last hundred years or so the stones have been largely forgotten, residing as museum curiosities receiving little attention. The last three or four years, however, have seen a revival of interest. Largely through the efforts of the staff at the Keswick Museum and Art Gallery the stones have gone on tour again with a series of concert performances in both Britain and Europe . They have also featured on a BBC Radio 4 programme recorded at the Museum entitled ‘The Worlds First Rock Band'. The Museum has also become involved in the interdisciplinary project with the University of Leeds under the title of ‘Yorkshire Quarry Arts' aimed at exploring the links between rock, music and the arts.


A considerable amount is known about the history of these ‘musical stones' - who made them, how they were played and the range of music that was played on them. On the other hand the geology of the stones has received little attention. Two particular aspects have never been explained; where was the actual source of the stones and what are the geological processes that have caused these materials to ‘ring' when struck.


The few scattered references to the source of the stones are confused and vague. Inaccurate locations have frequently been transposed from one account to another. Two fundamental points are clear however. All the stones of the main instruments are hornfels, material than can only be found in Lakeland within the inner aureole of the Skiddaw Granite. Second, suitable slabs were not easy to locate. It was clearly difficult to find pieces that had both a musical ring and were of the right proportions and thickness to shape and tune up into musical stones. In the Richardson instrument the largest stone is 1.03m long and only 8cm wide, the smallest 18.5cm long and 6cm wide. All of the 35 stones in the main scale, as well as the 25 on the top row, are close to 2.5 – 3 cm thick. It has been shown by Miller et al ( 2006 ) that there is a strong relationship between the length of the stones and their pitch and vibration. The very long, narrow stones must have been particularly difficult to find and more difficult to tune. Material around about 2.5 - 3cm thick was the optimum to locate.


The earliest reference and earliest Lakeland example of musical stones goes back to the late Eighteenth Century when Peter Crosthwaite, a somewhat eccentric and keen inventor, built a set for his own museum of curiosities in Keswick. His journals record in detail his collecting activities throughout the Cumbria region. An entry dated 10 th June 1785 records ‘ soon this morning found six musical stones at the tip end or north end of Long Tongue on the sand beds of the River Greta'. He later records that these first six stones were in perfect tune, but the remaining ten stones took him six months to find and had to be carefully chipped away, working 12 hours a day, until the desired note rang true (Barnes 2006 and Musical Stones files in Keswick Museum and Art Gallery). This early set of 16 stones is also on show in the Keswick Museum , but they are much smaller, shorter and relatively thick stubby pieces compared with the larger Richardson instrument. Crosthwaite's description of the source is very puzzling. The material is clearly hornfels, but Long Tongue cannot be located. Nowhere does the River Greta come anywhere near the inner aureole of the Skiddaw Granite. A possible explanation is that the original six loose pieces were brought down the Glenderaterra valley south from the area of the aureole, either in the glacial till, or by fluvial action and dumped into the Greta valley somewhere west of Threlkeld where the two streams join.


Clifton Ward (1876) in the first Geological Survey Memoir for the district clearly refers to the source of the stones as within the intermediate zone of the aureole in the belt of what he termed the spotted (or Andalusite) schist. He does not however pinpoint the precise location but all his analyses refer to the outcrop of the granite in Sinen Gill (Figure 2). Later sources, notably Postlethwaite (1913 ) repeat Ward's data on the aureole and specifically locate the source to the ‘spotted schists' of the aureole around Sinen Gill. Postlethwaite, describing the material as ‘sonorous', specifically links not only the Richardson instrument to this source, but also two other similar sets. These are the Daniel Till instrument constructed around 1875 and now in the Metropolitan Museum of Art in New York and one built by G.P. Abraham, the famous Keswick mountaineer (the present location of which is unknown). A description of the Till instrument is instructive of what was required to produce a set – ‘The rocks are Gneiss and Hornblende Schist, very hard and with a regular grain structure and were found, surrounding the granite, in large boulders which were chipped down, to oblong shape, to sizes varying from six inches to four feet in length. It was found that by chipping off their ends the tone was raised ,and by taking off a larger or smaller piece from the middle it was deepened, but unlike most instruments it never gets out of tune' (Moore 2005).


The focus of all the popular accounts and references to the source of the musical stones has therefore consistently centred on the Sinen Gill locality. Not only do all the tourist guides and information material refer to this area, but it has become entrenched in the literary heritage too as in the writings of the classic Lakeland author Hugh Walpole in his novel The Fortress, part of the Herries Chronicle (Walpole 1932). In 1966 E.H. Shackleton in his popular guide to ‘ Lakeland Geology: Where to go, What to see' , goes further and locates the source as the streamside of Roughton Gill (NY 298276), a short way south of Sinen Gill. This same information has been repeated elsewhere, notably by Shipp (1982 and 1992) and by Armstrong (2006) in a description of another set of musical stones housed at Cliffe Castle Museum , in Keighley, W. Yorkshire .


The metamorphic rocks within the aureole of the Skiddaw Granite have not been systematically mapped or examined in any detail apart from the early work of Ward (1875), Rastall (1910) and the survey work of Eastwood et al (1968) on the Cockermouth Sheet 23 which covered only part of the outcrop. Recent BGS survey work on the Keswick Sheet 29 (Woodhall et al 2000) and on the Skiddaw Group (Cooper et al 2004) has not focussed heavily on the aureole rocks. The observations contained in this paper are the result of a series of field transects across the area by the authors and some laboratory examination of specimens collected. The aureole covers about 56 kms ² of the Skiddaw fells, the inner part almost 20 kms² .Most of this ground has a cover of vegetation, peat and glacial till, hence outcrops of the bed rock are very limited. The granite is exposed in three places. The two northernmost outcrops lie along the bed of the River Caldew, at the Grainsgill Beck junction (NY325326) ( A , Figure 2) and in the Blackhazel Beck area (NY312312) ( B , Figure 2), but neither afford good exposures of the aureole rocks. The smaller outcrop in Sinen Gill however provides more extensive ground where the aureole rocks can be examined ( C , Figure 2).


Away from the inner aureole, Skiddaw Slate is a rather homogeneous fine grained slate, and although evidence of bedding and folding is sometimes apparent, the appearance of the rock is dominated by the intense slatey cleavage. Because bedding can be hard to identify, the intense folding that accompanied the cleavage development is not always apparent in outcrop (Simpson, 1968). Remarkably however, in the innermost part of the aureole at Sinen Gill, hornfels outcrops are well bedded on a scale of a few centimetres and are seen to be pervasively folded (Figure 3). Cleavage has been largely obliterated at the grain scale by the contact metamorphism, although some cleavage related features may remain (Figure 3b) but the regional cleavage is approximately axial planar to the observed folds. Pieces of rock similar to those used for the musical stones are seen to correspond to slabs of the more resistant beds within the hornfels.


At first sight it seems completely incongruous that rocks in which traces of bedding were largely destroyed during low grade metamorphism and regional deformation should subsequently redevelop an appearance of bedding only in the regions subject to the most intense contact metamorphism. The reason for this is almost certainly related to small differences in the clay mineral composition of the original sediments. In the innermost aureole, the distinct beds, readily picked out by weathering, are distinguished because of their varying proportions of andalusite and cordierite.


Andalusite forms prismatic crystals, typically 5-10 mm in length and 0.5 to 1mm across, although coarser grained beds are sometimes present. They have a distinctive white appearance on weathered surfaces but may appear pink when fresh. In contrast, cordierite grains occur as dark ovoids, 2-4mm across that are much less conspicuous in hand specimen, although they sometimes weather out as dark greenish black depressions. Under the microscope, it is apparent that the cordierite is extensively altered to a fine grained, yellowish-tinged pinite alteration. In the innermost aureole at Sinen Gill, the more resistant beds appear to be those with a significant content of andalusite in addition to cordierite, while those that have weathered more are rich in cordierite but with little andalusite.


Both cordierite and andalusite are minerals that are very rich in aluminium, and are therefore characteristic of metamorphosed clay-rich sediments. However cordierite fractionates magnesium relative to other ferromagnesian minerals found in pelitic metasediments, and so is more abundant in relatively magnesian compositions. Thus according to the proportion of Fe to Mg in the original sediment, pelitic hornfelses may contain andalusite + biotite + muscovite, cordierite + andalusite + biotite + muscovite, or cordierite + biotite + muscovite. It seems likely that the beds originally differed rather subtly in the proportions of illite, chlorite and other clay minerals. In terms of the formation of cleavage, these minerals properties are so similar that cleavage passes through the weakly metamorphosed beds in a way that does not distinguish between them. Only when the highest grades of contact metamorphism were achieved in the inner aureole was the cleavage destroyed by the recrystallisation of aligned micas and the formation of andalusite and cordierite. At this metamorphic grade however, the coarse grain size makes the mineralogical differences between beds more readily apparent, while the very different weathering characteristics of cordierite and andalusite mean that andalusite-rich beds are relatively resistant compared to those rich in cordierite but with little andalusite. Where the original cleavage is still apparent in the inner aureole, as in Figure 3b, it is because it was originally developed by a combination of phyllosilicate alignment and segregation of quartz-rich and phylosilicate domains. Contact metamorphism has destroyed the alignment, but cannot reverse the segregation.


A further point worthy of note is the variation in size of andalusite within the aureole. In the outer part the matrix is still dominated by fine grained phyllosilicates, but andalusite crystals may reach lengths of several cm. In the inner aureole, matrix minerals have recrystallised throughout and are distinctly coarser, but the andalusites in particular are typically smaller and often stubbier. This pattern of variation has been noted in studies of a number of aureoles in recent years, and has been ascribed by Waters & Lovegrove (2002) to the variation in heating rate away from the intrusion contact. In the inner aureole, heating is fast, reactions are overstepped and so andalusite forms multiple small crystals, as in the chilled margin of a pluton, whereas further out slower heating permits smaller numbers of crystals to grow to larger sizes.


Such textural considerations probably explain the localisation of outcrops in which the rock weathers and splits along beds. The Sinen Gill ringing slabs lie within the inner aureole of the Skiddaw Granite described by Rastall (1910) and Eastwood et al. (1968), and discussed in Mason (1978). Mason (1978) has pointed out that the mineralogy of the hornfelses is very much the same in the inner hornfelses and the surrounding slaty hornfelses of the intermediate zone.. Hence it is not the presence of cordierite and andalusite per se that makes the inner hornfelses so distinctive, but the fact that the abundant micas have completely recrystallized to an interlocking hornfels texture and no pervasive cleavage remains.


In a transects extending north across the inner aureole from Sinen Gill it was observed that ringing slabs could be found quite widely. Isolated boulders and weathered surface slabs scattered over the whole area could be shown to ring when struck. There is some evidence that andalusite-rich hornfelses ring better than cordierite-rich ones, presumably because cordierite weathers intensively.


Close examination of the streamside of Roughten Gill, specifically mentioned by Shackleton as the source, certainly reveals much evidence of quarrying of the bedrock. A 4-5m wide shelf has been cut in the hillside slope and extends for about 150m upstream from the small slab bridge along the north side of the stream. (Figure 4). The fabric in the andalusite-rich hornfelses here dips at around 70 ° to the SE. Slabs have clearly been pulled away from the small quarry faces. However the material does not have a good ring and the slabs come away in much thicker and more irregular pieces than the 2.5 – 3 cm thick pieces in the musical stones. It would seem more likely that this site was used as a source of building slabs for the nearby mine buildings and workings associated with the Brundholme Mine (Glenderaterra Mine) which is downslope from this area in the floor of the Glenderaterra valley (see sketch of mine installations in Tyler 2006).


The upper part of the Sinen Gill valley on the other hand looks a more likely source of material that could have been shaped into musical stones. (Figure 4). The granite is only clearly exposed along a section of the stream bed of Sinen Gill for about 170m. On the spur top approximately 100m south of the Gill (NY 30092 28055) there are some huge granite slabs exposed at the surface (over 5m in length), which may be in situ. The much more extensive outcrop of the granite as shown on the current BGS mapping of the area ( 1:50 , 000 Sheet 29 Keswick) is not immediately evident from the available surface exposures. Nevertheless the hornfelses in the upper area of Sinen Gill must be within perhaps 5-10m vertically above the roof of the granite. The hornfels is well exposed in this area with prominent crags on both sides of the valley and large areas of loose slabby scree ( lightly shaded area on Figure 4 ). The material has a very strong bedded fabric and importantly is seen to break relatively easily into flat slabs close to the optimum thickness of 2.5 - 3 cm seen in the musical stones (Figure 3a). On the south side of the valley the bedding dips consistently at between 65 ° and 80 ° to the SE (ie into the hillside slope and hence produces upstanding craggy outcrops). North of the stream with the dip closer to the angle of the hillside slope there are less continuous outcrops, but some bold rock faces reveal the same hornfels material. On this side there are some tight folds, (Figure 3a) with axial planes dipping steeply to the SE, but the bedding is very clear and uniform slabs of material could easily have been levered away from the surface outcrops or collected from the loose debris. Thin slabs throughout this area ring when struck with a geological hammer. Further, this area is relatively accessible. Accounts in the Richardson family material refer to family outings to the fells with a horse and cart to search for possible stones (Barnes 2006). Possible pieces were then conveyed back to Keswick where Joseph Richardson, (a stonemason by trade), spent long hours shaping and tuning the slabs. The well graded, wide, mine track up to this area from Threlkeld would have provided easy access for horses and carts. Beyond Sinen Gill the area of the inner aureole is poorly provided with tracks. Aerial photographs still show traces of narrower tracks up to the higher parts of Sinen Gill from the main mine trackway (Figure 4 ).


Acknowledgements: The authors would like to thank Joe Cann and Murray Mitchell for assistance with the field work and for helpful comments.


References :

Armstrong, A.C., 2006, A Lake District Rock Xylophone at Cliffe Castle . Bulletin Yorkshire Naturalist Union , 45, 1-2.


Barnes, J., 2006, Joseph Richardson (1790-1855) & Sons and the Famous Musical Stones of Skiddaw, Keswick Characters (Volume 1), Keswick Historical Society & the Friends of Keswick Museum & Art Gallery, 32-40.


Clifton Ward, J.C., 1875, The Geology of the Northern Part of the English Lake District, Memoir of the Geological Survey of England and Wales , 9-12.


Cooper, A.H., et al., 2004, The Skiddaw Group of the English Lake District . British Geological Survey Memoir for parts of Sheets 22,23,24,28,29,30,31 and 48. British Geological Survey, Keyworth, Nottingham .


Eastwood, T., Hollingworth, S.E., Rose, W.C.C. and Trotter, F.M., 1968. Geology of the country around Cockermouth and Caldbeck. (Explanation of one-inch Geological Sheet 23, New Series). Institute of Geological Sciences, London .


Postlethwaite, J. 1913 Mines and Mining in the English Lake District , W.H. Moss, Whitehaven (3 rd Edition). (Reprinted 1975 by M. Moon, Beckermet , Cumbria . Pages 17-18 ).


Mason, R., 1978. Petrology of the Metamorphic Rocks . George Allen & Unwin, London .


Millar, B, Odling, N, Barker, D.W., Ng, K., and Stell, J., 2006, Music Rocks. Bulletin Yorkshire Naturalist Union , 45 , 3-5.


Moore , J.K., 2005, Correspondence between The Metropolitan Museum of Art , New York and the Keswick Museum & Art Gallery. Musical Stones file, Keswick Museum & Art Gallery.


Rastall, R.H., 1910. The Skiddaw granite and its metamorphism. Quarterly Journal of the Geological Societ y, 66 , 116-141.


Shackleton, E.H. Lakeland Geology: Where to go: What to see. Dalesman Publishing Co. Clapham. Chapter 7.


Shipp, T., The Lake District . Cumberland Geological Society. Unwin Paperbacks, Rocks & Fossils Geological Guides. Chapter 5.


Shipp, T., 1982, Lakeland Rocks and Landscape: A Field Guide. Ellenbank Press, Maryport. Chapter 14.


Simpson, A., 1968. The Caledonian history of the north-eastern Irish Sea region and its relation to surrounding areas. Scottish Journal of Geology , 4 , 135-163.


Tyler, I., 2006, The Lakes and Cumbria Mines Guide, Blue Rock Publications, Keswick.


Walpole, H. 1932 The Fortress. MacMillan, London . (Climax to a Long Sequence, 111 In a dark house ).

Waters, D.J. and Lovegrove, D.P., 2002 . Assessing the extent of disequilibrium and overstepping of prograde metamorphic reactions in metapelites from the Bushveld aureole. Journal of Metamorphic Geology , 20 , 135-149.


Woodhall, D.G. et al, 2000, Geology of the Keswick District. (Sheet description of British Geological Survey 1: 50,000 Series Sheet 29. British Geological Survey, Keyworth, Nottingham .




Dr. R. A. Smith,

Rigg Side, Grange Park , Keswick , Cumbria . CA12 4AY



Professor Bruce Yardley,

School of Earth and Environment, University of Leeds , Leeds LS2 9JT






Figure 1. The Richardson set of Musical Stones in the Keswick Museum and Art Gallery built by Joseph Richardson in 1840. It is 3.65m long and 1.2m high. There are 35 stones on the main scale, and 25 on the upper row. Above are two rows of metal plates and a series of bells. (Photographs courtesy of Adam Naylor-Whalley).







Figure 2. The Aureole of the Skiddaw Granite within the Skiddaw Group. ( Based on Eastwood et al 1968).






Figure 3a


Figure 3b


Figure 3. Outcrop of hornfelses upstream of the granite exposures on the northern side of Sinen Gill (NY 303282) but probably within 5m vertically of the roof of the granite: a ) Hornfels with bedding of a scale of a few cms picked out by weathering and defining a tight fold. Above the hammer, bedding slabs have broken off the side of the outcrop. b ) Nearby outcrop of near-vertical hornfels with the remnant SE-dipping cleavage preserved as resistant ridges of segregated quartzofelspathic material. Note the rock nevertheless breaks along bedding or jointing rather than the former cleavage .





Figure 4. The Sinen Gill Area.