The origin of Maltese cart-ruts: cut

by wheels or tools?

Derek Mottershead

Introduction Upper Coralline limestone, Lower Coralline

limestone and Globigerina limestone
The origin and means of formation of the cart- formations, embracing rocks of differing
ruts of Malta have been matters of debate for character which illustrate some of the variety
almost a century. The principal contenders for found within this broad rock type. At the core
rut formation have been wheeled vehicles, of the current issue is the nature of the surface
sleds, slide cars and cutting with hand tools. micromorphology of these varied rocks and
Most recent authors have discounted both sleds its accurate identification, whether created by
and slide cars. The former, to be of sufficient natural subaerial erosion processes or whether
magnitude, would create unmanageable human-induced. It is also germane to consider
amounts of friction, whilst there is neither the effects of the exposure of any such newly
archaeological nor historical evidence for formed surfaces to subsequent centuries of
the latter. The most recent publications on natural weathering and erosion. In this context
this topic have advocated wheeled vehicles 1 great care (and a geomorphological eye) is
or hand cutting. 2 The former combine field required in interpreting such minor relief
observation of rut form with geotechnical forms, particularly as artefacts of human
information on rock strength, and conclude activity.
that the passage of wheels of, for instance, a There is a substantial lexicon of small
two-wheeled cart, would create more than scale forms, of a few millimetres in relief, on
sufficient stress on the rock beneath to cause limestone surfaces exposed to weathering. This
erosion of the local rock. The latter makes includes features of greater or lesser georr:.etric
the assettion that 'there is clear evidence regularity, well documented in standard texts
of ancient tool marks' ,3 and concludes that such as Ford & Williams4 and Trudgill. 5 In
cutting by hand played a significant part in addition to such well codified forms, ::here
rut formation. This conclusion is illustrated also exist forms of rather less regularity. On
by two photographs of field sites showing near-vertical surfaces such as rut walls, these
small-scale rock surface morphologies which include contiguous quasi-circular hollows
are interpreted as ancient tooling marks. The separated by cuspate ridges, and internal
current paper questions such an interpretation, solutional voids within the rock now exposed
and whether the conclusions derived from it by erosion. Weathering of closely jointed rock
can be sustained by the evidence provided. may preferentially expose the planar faces of
The Maltese cart-ruts are formed in bedding or transverse joints. At more random
limestone, a water-soluble rock. Limestone level, internal variations in rock fabric, degree
varies greatly in texture and in its content of cementation or biotic components may
of biotic and non-calcareous components. be expressed on an exposed surface as relief
Rutted terrains in Malta are found on the forms created by differential weathering.

Malta Archaeological Review • Issue 8 2006/2007 43

 limestone is a rock of high porosity 32-41% 7
and consequent very low shearing strength of
2.20-3.85 MPa8 and it is here directly exposed
on a moderately steep slope. Directly exposed
to solar radiation, it consequently experiences
large and rapid temperature changes (far
higher than the diurnal air temperature range),
and also the deposition of damaging marine
salts at high concentration in a near-coastal
environment. Surface rock at this site is also
directly exposed to the impact of falling
raindrops, which directly apply a shearing
force that increases rapidly on slopes of 45° and
above. This hostile weathering environment
promotes dissolution of the soluble limestone,
particularly at grain boundaries, and creates
mechanical stresses at these weak points by
Plate I. An ancient quarry surface at Dwejra, Gozo, showing small angular planar slaking (wetting/drying cycles), expansion/
facets interpreted as masons' marks by Magro Conti and Saliba (2007, Fig. 59).
contraction cycles, thermal shock and salt
crystallisation. Disaggregation of the rock
The Field Evidence will ensue as raindrop impact forcibly detaches
loosely adherent grains and particles of rock. A
The evidence presented by Magro Conti soluble rock oflow mechanical strength in such
and Saliba consists of two medium range a weathering environment is therefore highly
photographs showing the morphology of rock vulnerable to weathering (breakdown) and
surfaces, one an ancient (sic) quarry near erosion (removal of debris). These processes
Dwejra, Gozo, cut in Globigerina limestone, would be most effective within and adjacent to
the other the sidewall of a cart-rut incised into existing planes of weakness within the rock,
Upper Coralline limestone at Misrah Ghar il- such as joints. Under these circumstances rock
Kbir, Malta. breakdown would initially be concentrated
The Dwejra site (Plate 1) appears to show locally around the joints, which would in turn
a bedrock slope of 40-70°, which in a high become exposed as planar facets as weathered
contrast image reveals a number of planar material is removed, to create the type of
micromorphological elements, in the form surface micromorphologies apparent in the
of angular facets approximately vertical illustration.
in orientation. 6 There also appears to be a Rates of weathering and erosion on
significant number of quasi horizontal slope Globigerina limestone can be inferred from
elements, with markedly angular junctions exposed rock surfaces of known age, such
between the various planar facets. No scale as historic structures. Cavities formed by
is given, but they appear to be of the order weathering and erosion on bastions in Valletta,
of a small number of centimetres in vertical for example, commonly show 20 mm of
extent, with lateral dimensions up to maybe erosion over a period of 450 years. Applying
15 cm in several cases. It is implied that this such a rate to an 'ancient' quarry which, if
morphology represents marks created by a related to the cart-ruts, may well be over
chisel or the blade of masons' picks. 2000 years old, suggests that up to 100 mm of
There is, however, a plausible alternative surface rock may well have been eroded over
explanation. These planar forms, with sharp such a timescale from the site illustrated. If
angular boundaries, are also suggestive of this is the case, then it is extremely unlikely
joint planes, lines of fracture within the rock than any original masons' marks remain on
which have been preferentially exposed by this particular rock over such a timescale; if
surface weathering and erosion through the any had existed they would by now have been
action of natural processes. Globigerina obliterated by weathering and erosion. In

44 Malta Archaeological Review • Issue 8 2006/2007

 Plate 2. Obliq-!e view ofa rut sidewall
showing cur:ed parallel grooves,
convex upw:1rds, interpreted as
masons' tooling marks by Magro
Conti and Scliba (2007, Fig. 47).
The field notevook rests on the lichen
covered rock surface into which the
rut is verticaliy incised.

contrast, however, if they are natural features with the axis of rotation within the existing
the relatively fresh forms now apparent would rut. The arc produced by the cutting action,
be under constant recreation, refreshed by however, would tend to undercut the heaC:wall
continuing erosion. to create an overhang at the rut head. This
The evidence presented from Misrah Ghar would then need to be removed, the simplest
il-Kbir presents a different kind of problem. 9 way to achieve this being a vertical blow to the
It comprises a sub-parallel set of curved rock surface above the unsupported overhang.
flutes, of apparently very low relief, formed There is, however, no trace of marks suggesting
on the surface of a cart-rut wall in Upper the latter action.
Coralline limestone. The flutes appear to be A more likely style of attack with a tool
approximately 10 mm in width and ~100 mm would be from the surface downwards (Fig.
in length (Plate 2). lC), which would create a vertical impact of
Their longitudinally curved form is convex the tool head, and tool marks which des~ribe
upwards and steepening downwards. They a circular arc initially orthogonal to the rock
are described categorically as 'hand cut tool surface, then curving downwards towards the
marks'. 10 It is not stated what tool might have user. This procedure would also be the ob~ious
been employed or how such a tool may have way also to extend the headwall, bringing the
been applied. The most likely tool at this scale blade downwards vertically on the surface
would appear to have been a chisel-headed in order to maximise the lateral tensile stress
pick, rotating across the rock surface and against the unsupported open headwall. This
etching a curved groove laterally into the rut approach would make maximum use of the
wall. The interpretational problem created by mass of the pick, and impose far less stress on
these forms is that the axis of rotation of any the wrist in wielding it. To hold the tool low,
tool would have been substantially below the lifting it above wrist height before rotating it is
original rock surface (Fig. lA); in other words, just not a natural or energy-efficient movenent.
the rut could not have been cut downwards It would appear that any energy efficient node
from the rock surface because the orientation of tooling attack requires a vertical blow to the
of the curvature requires any tool to have been rock surface, which would create grooves of
held below the surface itself. An alternative significantly different form to those described
possibility is that the pick was used to excavate by Magro Conti and Saliba. It is concbded
the rut by longitudinal extension, cutting back a that the morphology of the grooves in Plate 2 is
terminal headwall (Fig. lB). This case permits therefore most unlikely to have been formed by
the pick to be held below the ground surface masons' tooling.

Malta Archaeological Review • Issue 8 2006/2007 45

 dissolution has taken place since rut formation.
A
Such dissolution is likely to have diminished
'
' rather than preserved any original tool marks.

f~ Discussion

It is argued above that the evidence presented

Fig. 1. Arcs of
rotation caused by
differing modes of
attack in hand cutting
a rock surface.
A) the attack required
B

7TTnt\ for tooling marks in relation to both ancient

quarrying and cart-rut formation creates
significant difficulties of interpretation.
Furthermore, there are difficulties in both
cases as to whether original forms supposedly
to create the groove
created over 2000 years ago could have
pattern shown in survived subsequent weathering and erosion.
Plate 2. Resolution of this contention requires
B) the arc formed by
an independent test of the nature of masons'
attacking a headcut
wall with the axis marks, and of their capacity for survival over a

~~
of rotation required long period of historic time. Such a test would
to form the groove consist of an examination of locations where
pattern of Plate 2. c
C) the groove
.. the following conditions are satisfied:-
pattern created by a
normal attack on rock
..
''
that stonemasons would have been active
there would have been no competing
forming the ground '
surface.
activities capable of creating cut surfaces, and
cut rock surfaces have been exposed since
their formation.
A further issue is that the supposed tooling The opportunity for such a test exists locally
marks appear to have a length of at least 100 in the form of ancient quarries such as those
mm. It would appear that in order to generate at Misrah Ghar il-Kbir where grooves cut
sufficient momentum to form such marks, into the solid rock have been manufactured in
which would evidently represent a single cutting out regular ashlar blocks, and Imtahleb,
sweep of the tool, a metallic head of significant where cut quarry faces over one metre high
mass would be required. No archaeological remain exposed. If masons' marks are to exist
evidence of such artefacts has been presented anywhere, then they would surely be present at
which, given the abundance of cart-ruts across such quarry sites, where processes other than
the islands, is perhaps surprising. deliberate cutting are most unlikely to have
Furthermore, although the image is been operative. A careful search of both sites,
described as showing an absence of dissolution in glancing sunlight, of surfaces self-evidently
featu::.-es, a number of fossil fragments appear cut by human artifice revealed no marks of the
to stand out from the rut wall surface, and kind interpreted by Magro Conti and Saliba
an a::-ea of has-relief in the wall appears to as tooling marks. If similar marks to those
reveal a sediment-filled cavity. These areas described by Magro Conti and Saliba cannot
of differential relief are, in fact, indicative be found in such locations, where they are most
of cifferential dissolution and show that likely to exist, then it can be concluded either
weathering has indeed taken place since that they were not created by masons or that any
the formation of the rut. Although there masons' tooling marks which may have been
is no obvious alternative interpretation of created have been destroyed by subsequent
these fluted forms other than some material weathering and erosion. In either case it is
',;,:··· varia:ion of unknown source within the rock implied that marks described by Magro Conti
fabric, the arguments presented above suggest and Saliba have some other cause.
strongly both that they were not formed by Recent evidence of the relationship
humm tooling, and that significant subsequent between rut patterns and limestone rockhead

46 Malta Archaeological Review • Issue 8 2006/2007

relief (the form of the bedrock surface beneath
an overburden cover) has led to a new model
of rut formation. 11 It is reasoned that ruts were
initially formed on the surface of soil material
overlying a buried rock surface. Erosion of
the soil by passing traffic gradually exposed
the underlying bedrock, onto which the cart
tracks became superimposed, incising ruts into
the newly exposed rock. In this way, the ruts
would be formed by linear abrasion along the
lines of the wheel tracks. Patterns indicative
of longitudinal abrasion along the length of
the rut are shown, for example, by the ruts
at Imtahleb (Plate 3), where longitudinally
persistent edges and shelves in the rut walls
form consistent cross sections along the length
of the ruts.

Conclusion

The evidence presented by Magro Conti &

Saliba in support of hand tooling, to this
author at least, appears less than conclusive.
In the light of careful consideration of small
scale forms on limestone surfaces, and of Plate 3. Cart-rut at Imtahleb showing multiple lel"els in the rut wails
formed by longitudinal abrasion.
natural weathering and erosion processes, it is
contended that rather stronger evidence would
need to be presented before an unequivocal remain. The rut planform sketches pro"ided
interpretation of human action can safely be make it clear that those that remain are merely
made. This would include consideration of fragments of much more extensive patterns,
the clarity, detail, frequency and distribution whose original length is clearly a substantial
of the forms supposedly created by human multiple of the remaining network fragments.
handiwork. The case would be strengthened by It would appear inherently unlikely that any
independent evidence of their origin, and also a significant fraction of the original whole was
consideration of the likely style of tool usage. It cut by hand, given the resources of human
is particularly the nature of geomorphological labour and tools that that would demand.
features, where similar forms may represent
the end point of more than one process, that
there is danger in inferring process from form Acknowledgements
without independent supporting evidence. I am indebted to Reuben Grima for a
A broader perspective on rut formation stimulating discussion which greatly clarified
can be taken by considering the rut network my thinking on this topic, to Alastair Pearson
as a whole. Summing up the currently extant for thoughtfully reviewing an earlier draft of
ruts presented in the gazetteer of Magro Conti this paper, and to Bill Johnson for drafting
and Saliba12 suggests that c. 35 km of ruts Figure 3.

Malta Archaeological Review • Issue 82006/2007 47

 Notes of the Maltese Islands', inS. Siegesmund, T. Weiss
and A.Vollbrecht (eds.), Natural Stone, Weathering
D. Mottershead, A. Pearson and M. Schaefer, 'A Phenomena, Conservation Studies and Case Studies,
gcotechnical approach to the formation of the 'cart (Geological Society of London Special Publication
ru::s' of Malta', Stone 2 (2007), 5-6. 205, 2002), 33-49.
D. Mottershead, P. Farres and A. Pearson, 'The 8 J. Saliba, The shear strength of Globigerina
influence of environmental change in the formation Limestone. B.E.&A.(Hons.) unpublished dissertation,
of the 'cart ruts' at Naxxar', Stone 2 (2007), 7-8. University ofMalta, (1990).
D. Mottershead, A. Pearson and M. Schaefer, 'The 9 Magro Conti and Saliba, fig. 47.
cart-ruts of Malta: an applied geomorphology 10 Ibid., 57.
approach', Antiquity 82 (2008), 1065-1079. 11 D. Mottershead et al. (2007), and D. Mottershead,
2 W3gro Conti, J. & Saliba, P.C. (eds.), The Significance P. Farres and A. Pearson, 'The changing Maltese
of Cart-ruts in Ancient Landscapes, (Malta. Midsea soil environment: evidence from the ancient cart-
B::ooks, 2007). tracks at San Pawl tat-Targa, Naxxar', in B.J.
3 n:~d., 69. Smith, M. Gomez-Heras, H.A. Viles and J. Cassar
4 D.C. Ford and P. Williams, Karst Geomorphology (eds.), Limestone in the Built Environment: Present
al#d Hydrology, (London. Unwin Hyman, 1989). Day Challenges for the Preservation of the Past.
5 S.T. Trudgill, Limestone Geomorphology, (London. (Geological Society of London Special Publication
LctUgman, 1985) 331, 2010).
6 M!tgro Conti and Saliba, fig. 59. 12 Magro Conti and Saliba. 193-287 in Appendix I.
7 J. Cassar, 'Deterioration of the Globigerina Limestone Gazetteer of cart-ruts in the Maltese Islands.

48