Post-Disaster Archaeological Investigations at the Kathmandu Valley UNESCO

World Heritage Property (Nepal)

INTERIM REPORT OF THE NOVEMBER-DECEMBER 2016 SEASON OF FIELD ACTIVITIES

Robin Coningham, Kosh Prasad Acharya, Christopher Davis, Ram Bahadur Kunwar, Ian Simpson, Anie
Joshi and Kai Weise

February 2017
Contents

1. Introduction ........................................................................................................................................ 3
2. Summary of Field Activities at the Kasthamandap in October-November 2015 ................................ 3
3. November-December 2016 Field Activities and Rationale ................................................................. 4
4. Preliminary Results of the 2016 Field Activities.................................................................................. 5
5. Research Conclusions.......................................................................................................................... 9
6. Observations and Recommendations ............................................................................................... 11
7. Acknowledgements........................................................................................................................... 14
8. References ........................................................................................................................................ 14
9. Figures ............................................................................................................................................... 16

2
1. Introduction

1.1 The two major earthquakes that struck Nepal on the 25th April and the 12th May 2015 caused a
human catastrophe, devastating large areas of the county and neighbouring regions, leading to
substantial loss of life, livelihoods as well as post-disaster physical and mental trauma. This natural
disaster, and its associated aftershocks, also generated a cultural catastrophe, damaging and
destroying parts of Nepal’s unique cultural heritage, including monuments within UNESCO’s
Kathmandu Valley World Heritage Site of Universal Outstanding Value. Not only beautifully ornate
temples of wood, brick and tile, the monuments of the Kathmandu Valley represent a major source
of income and economic growth through international and national tourism and are a key
component of Nepal’s fragile economy as one of the country’s major sources of foreign currency.
Furthermore, these sites play a central role in the daily lives of thousands of Nepalis, representing
portals where the heavens touch the earth and where it is possible for ordinary people to reach out
and commune with their guiding goddesses and gods – they are of intangible value. For these
reasons, the damaged heritage cultural sites of Nepal are subject to a major programme of
consolidation, conservation and reconstruction. The Kasthamandap, traditionally thought to be one
of Kathmandu’s oldest monuments, collapsed during the 7.8 Magnitude Gorkha Earthquake on the
25th April 2015 and has become an icon and focus of the reconstruction of Kathmandu and Nepal’s
lost cultural heritage.

2. Summary of Field Activities at the Kasthamandap in October-November 2015

2.1 The collapse of the Kasthamandap led to many fatalities and, in the aftermath, the Government
of Nepal, Kathmandu Metropolitan City and local communities pledged to rebuild quickly. However,
despite the depth and breadth of architectural research on its timber superstructure, nothing was
known about its foundations. In response, Durham University and the University of Stirling, in
collaboration with the Department of Archaeology, Government of Nepal, were funded by UNESCO
in October-November 2015 to undertake post-disaster archaeological assessments and evaluations
of sites and monuments within the Kathmandu Valley UNESCO World Heritage Property that had
collapsed during the 2015 Gorkha earthquake. In addition to Ground Penetrating Radar Survey
around the Kasthamandap and within Hanuman Dhoka’s Durbar Square, the major focus of our
investigations was in the ruins making where the Kasthamandap had stood. Utilising a multi-
disciplinary approach of contextual systematic excavations, geoarchaeological investigations,
scientific dating methods and architectural analysis our pilot investigations studied the origins and
development of the monument. To do this, the team excavated a single evaluation trench into the
rubble of the south-west corner (Figure 1). This trench quickly demonstrated that the foundations of
the monument had not been severely damaged by the 2015 earthquake or by prior seismic activity.
Moreover, it demonstrated that much of the damage to the foundations could be directly attributed
to post-disaster interventions, including the removal of material in the emergency phase of post-
disaster management (Figure 2). Our excavation revealed that the brick foundation walls reached
depths of two metres and had been set within mud mortar, on a surface prepared prior to
construction and with an organised sediment fill material, which we suggested gave the foundations
of the monument resilience and flexibility during seismic shock (Figures 3 and 4). The organised
foundation sediments contrasted sharply with the river deposited material of the immediate
surrounding landscape. We were also able to date the construction of the foundation wall to the
seventh century CE, illustrating that the monument was constructed earlier than traditionally

3
ascribed in textual sources and we were also able to identify pre-monument occupation in the
vicinity of the monument dating to the second century BCE (Coningham et al. 2016).

2.2 We were also able to demonstrate that the design of the foundations in the monument's south-
west corner had included cross-walls for stability, bracing the massive one metre thick and two
metre deep outer walls against a central free-standing brick pier. This had given added strength to
the central superstructure as the pier in turn had supported a saddlestone into which one of the
monument’s massive six metre high timber pillars had locked (Figure 5). In our conclusions, we
hypothesised that that the pattern cross-walls ran across the entire interior of the main foundation
wall of the monument, forming a 9-celled mandala configuration beneath the twentieth century
floor levels. Clearing debris and rubble from the areas of floor surface which had not been removed
by JCB in the aftermath of the earthquake, we were able to identify the location of the saddlestones
of the south-west, south-east and north-west of the monument. However, we were unable to find
the north-east saddlestone as its projected location was simply tiled. Moreover, on inspection of the
four major central timber pillars demonstrated that whilst three still possessed wooden tenons, the
base of the fourth had rotted and had no tenon. This discovery led us to conclude that conservators
had simply cemented tiles around and under the rotten pillar base rather than replacing it during the
monument’s major 1960s renovation. This work, we believe, had entirely disengaged the north-east
timber pillar from the secure framework of saddlestones, brick piers and braced foundations below,
creating a key weakness within the structure (Coningham et al. 2016). We were also concerned that
the north-east brick pier and saddlestone may have been missing due to sustaining damage from
earlier seismic events. In such a scenario, we anticipated that the pier might have collapsed or the
saddlestone split, thus necessitating a significant engineering intervention into the monument’s
foundations.

3. November-December 2016 Field Activities and Rationale

3.1 As our 2015 mission was limited in time and involved work at other monuments across the city,
we were unable to evaluate the stability of all the Kasthamandap's foundations or confirm the
location of the missing north-east saddlestone. However, we recommended to the Government of
Nepal and UNESCO in our report in December 2016 that further rescue excavations be undertaken
so that "more of the monument is exposed, which would allow its construction to be fully
understood and would also remove the debris from the JCBs...The larger excavation trench would
not only provide sequences but also architectural plans." (Coningham et al. 2015). No further
funding was available from UNESCO or the Government of Nepal and work at the site ceased with
the Municipality erecting a chainlink fence on metal poles around the ruins. Almost a year later, a
new stream of funding through the Arts and Humanities Research Council (UK)’s Global Challenges
Research Fund (AH/P006256/1) to Professor Robin Coningham and Professor Ian Simpson and a
Conservation Trust award from the National Geographic Society (#C333-16) to Mr Kosh Prasad
Acharya, facilitated a season of multidisciplinary field activities between the 9th November and 20th
December 2016. The fieldwork at the Kasthamandap was undertaken with the following objectives:
 To evaluate the nature and condition of the monument’s foundations;
 To evaluate the subsurface archaeology adjacent to the temple;
 To evaluate the cultural sequence of the monument, including subsurface deposits complex;
 To scientifically date and characterise the cultural sequences of the site;
 To assess the condition of the monument’s foundation and the phasing of its construction to
assist the preparation of plans for its reconstruction;

4
  To offer training in the post-disaster analysis and evaluation of monuments.

3.2 To achieve these objectives, a collaborative team from the Department of Archaeology,
Government of Nepal, Durham University and University of Stirling undertook inter-disciplinary
rescue excavations at the Kasthamandap. The investigations also offered a key opportunity to
provide training and capacity building for officers from the Department of Archaeology, Government
of Nepal; Kathmandu Metropolitan City; the Central Cultural Fund, Government of Sri Lanka; the
Department of Archaeology and National Museum, Government of Myanmar funded by the Oriental
Cultural Heritage Sites Protection Alliance; and postgraduate students at the Lumbini Buddhist
University in rescue archaeology. We hope that this valuable training will contribute to the
mitigation of the risk of unchecked reconstruction and enhance opportunities for recording and
protecting heritage during post-disaster interventions. Such action is required now due to the focus
on rapid reconstruction in the post-earthquake political and social climate and the potential of
damage posed by the 'Build Back Better' approach due to the lack of awareness of subsurface
heritage and raise awareness of the valuable structural information that can be gathered from these
remains of the Kathmandu Valley.

4. Preliminary Results of the 2016 Field Activities

4.1 Excavations began at the Kasthamandap on the 12th November following a public briefing
meeting for policy makers, practitioners and residents in Nasal Chowk within Hanuman Dhoka’s
Palace complex (Figure 6). The first step in understanding the Kasthamandap’s morphology and
recording what survived of its foundations involved the removal of the loose rubble above the site.
Indeed, the entire footprint of the Kasthamandap had been reconstituted with rubble after the
emergency response phase and then levelled to create a flat surface across the monument (Figures 7
and 8). The monument’s western, southern and eastern edges were marked by dry brick wall in
order to retain the rubble behind. Following the pilot post-disaster heritage protocols currently
being developed by Durham University’s UNESCO Chair with the Department of Archaeology,
Government of Nepal, the surviving footprint of the Kasthamandap was gridded into 10 by 10 metre
squares. The rationale behind this grid is that any rubble and artefacts recovered or removed can be
provenanced to a particular locality within the site during excavation. This was particularly
important as the rubble comprised a mixture of modern and historic materials, the result of the
destruction of in-situ archaeological deposits and their mixing with material from more modern
phases of the structure. Indeed, the rubble was found to contain a number of structural elements,
such as saddlestones, which would have originally formed an integral part of the link between the
foundations and the superstructure (Figure 9).

4.2 Once the rubble was removed, we were able to document the substantial damage to both the
monument’s floor and it’s structural foundations from the deployment of bulldozers and JCBs the
during humanitarian efforts to recover the injured and dead from the brick, clay, tile and timber
debris forming the collapse above the Kasthamandap (Coningham et al. 2016). We were even able
to identify the individual scrape marks of the buckets of this machinery across the surviving tiled
surface of the monument (Figure 10). The bulldozers and JCBs had also clearly destroyed large
portions of the south-east and south-west of both the inner foundation wall as well as the
monument’s outer wall (Figures 11, 12, 13 and 14). This damage included the ripping out of a

5
number of the large single and double saddlestones placed along the top of the outer wall and inner
foundation wall, which had served to link these elements to the superstructure above.

4.3 While the majority of the foundations exhibited no evidence of earthquake damage, we were
able to identify some evidence of minor earthquake damage (Pers. Com. Dr Paolo Forlin). This
included small vertical cracks in bricks immediately below two of the double saddlestones in the
western stretch of north-south inner foundation wall (Figure 15). Limited to the upper courses, this
type of earthquake archaeological effect, particularly below a structural element, provides evidence
of stress within a building during a seismic event as the pressure of ‘up and down’ movement causes
the wall below to sustain damage from the weight of the saddlestone with cracks forming in the
brickwork. However, whilst it is not possible to date these cracks to a particular seismic event, the
extent of this damage is extremely limited. To reiterate our earlier comment, the exposed
foundations exhibited no major structural damage attributable to the 2015 earthquakes or previous
seismic events and appear exceptionally resilient. It is highly likely that the use of mud mortar to
bond the walls has saved the building from further stress, limiting earthquake damage to only a few
locations. As noted by the engineers from the Tokyo National Research Institute for cultural
Properties, more devastating collapses have been prevented on other buildings by ensuring that
joint failure precedes brick failure (TNRICP 2016: 58). We did, however, record other indicators of
earthquake damage, such as impact damage on areas of the tiled surface. Indicated by concave
surfaces within the tiling, they were most likely caused by falling elements of the superstructure
during the 25th April 2015 earthquake (Figure 16).

4.4 Once the remaining tiled surface of the Kasthamandap was exposed, it was again possible to
confirm the visible presence of the three large saddlestones around the central Gorakhnath Shrine.
Indeed, we were able to confirm that each had had tile placed alongside the edges of the sockets in
these large stone architectural elements. Clearly observable in the south-west, south-east and
north-west, the socket for the fourth, which should have been present in the north-east, was not
visible (Figure 17). Therefore, we opened a half-section over the location where we postulated it
should have been located. Once the floor tiles were removed from the eastern side of this half-
section, we were able confirm that the saddlestone was indeed present, sealed below the tiled
surface (Figures 18, 19 and 20). We can therefore state that this key structural element had been
sealed below the most recent floor surface during the twentieth century renovations within the
Kasthamandap. This confirmed our earlier hypothesis that the major structural timber pillar for the
superstructure in the north-east was not locked into its resilient foundation below. A contributing
factor in the collapse of the Kasthamandap, this element was free-standing, potentially moving at a
different rate to the rest of the structure, perhaps leading it to slip and causing the monument’s
complete collapse. We also noted that some of the other, smaller saddlestones had also been tiled
over, presumably due to rotten wooden tenon joint and some of their mortice sockets filled – again,
contributing to additional weaknesses within the building between the superstructure and
foundations (Figures 21 and 22).

4.5 This discovery allowed us to refute our earlier concerns that the north-east saddlestone might
have split or sustained damage in earlier seismic events but also allowed us to provide additional
information on traditional construction and architectural practices. Indeed, once the north-east
saddlestone was cleaned, we identified the residue of a copper plate above its socket as already

6
recorded on the three large saddlestones identified in 2015. As importantly as the presence of this
possible damp proofing between the saddlestones and foundations and timber tenon, we also
recovered, the corroded remains of a copper shoe around a fragment of degraded timber from
within the saddlestone’s mortice joint (Figures 23 and 24). Still preserved with nail holes through
the copper sheeting, it illustrates traditional techniques, now lost. The presence of the copper shoe,
with wood tenon preserved inside, suggests that the pillar’s tenon had already severely deteriorated
and that individual tiles had been pushed under it during its 1960s restoration rather than
dismantling the entire monument. In-field XRF analysis was used to determine the elemental
composition of the copper plate and confirmed that copper (Cu) comprised approximately 84.5% of
the material and with an associated set of trace elements.

4.6 Once we had removed the copper shoe in the mortise socket of the north-east saddlestone, we
were able to record and then remove two circular gold foil mandalas (Figures 25, 26 and 27). Both
foil discs were incised with floral designs. The presence of the mandalas at the base of the
saddlestone’s socket was not surprising as we had already recorded the presence of a single circular
gold foil mandala in each of the sockets of the three other major saddlestones excavated in 2015
(Coningham et al. 2015, Coningham et al. 2016). We reiterate our earlier suggestion that their
presence was part of elaborate consecration and construction rituals for the monument, potentially
linked to the creation of an axis mundi (Coningham, et al. 2015; Slaczka 2007).

4.7 The surviving tiled surface within the monument was then removed, allowing us to identify a
series of settings and relatively modern interventions as floor surfaces across the Kasthamandap.
The tile floor itself was set in cement, which had been laid over a rough surkhi surface (Figure 28).
When we peeled this thin surkhi layer, we exposed a brick setting or paving, which included stamped
Rana period bricks (Figure 29). This had been laid above another brick paving level, which
incorporated a mixture of old and modern bricks (Figure 30). This phase also contained an area of
stone paving in the north-west of the monument, east of the north-west Ganesh Shrine and north of
the Gorakhnath Shrine (Figure 31). This paving included a number of reused architectural
fragments, including a short decorated pillar (Figure 32). We postulate that this delineation of stone
paving, may define one of the partitions present within the Kasthamandap when it was occupied
residentially and by shops until their subdivisions and partition walls were removed during the 1966
renovations.

4.8 This surface of reused stone and its contemporary surface of mixed modern and old bricks was
then removed onto an underlying sterile silt. This silt was then removed to expose the full layout of
the cross-walls hypothesised by the team in 2015 (Coningham et al. 2016), that of a nine-celled
mandala across the monument (Figure 33). We were therefore able to confirm that after the initial
construction of the foundation wall in the seventh century CE, the monument was reconfigured in
the ninth century CE with the creation of the nine-celled mandala by the use of single-brick thick
cross-walls. Most likely imbued with cosmological and symbolic value, it is also highly likely that
these cross-walls acted as bracing walls between the central brick piers supporting the four major
timber pillars and the inner foundation wall. With the entirety of this pattern now confirmed, we
then attempted to answer the secondary question of whether the north-east saddlestone and brick
pier had sustained any seismic damage.

7
4.9 Each of the nine-cells contained soil fills and, as we had investigated two cells in the outer cells of
the mandala pattern to the south and west in 2015, we decided to investigate the central cell. This
we did by excavating a deep trench between the north-east and north-west saddlestones and their
brick pillars, defined on the northern edge by the east-west cross-wall and on the southern side by
the Gorakhnath Shrine itself. The first fill encountered was very different from those excavated in
the south-west corner of the monument in 2015. Rather than comprising a silty material, the fill was
a prepared clay, possibly forming an upper surface within the central cell of the nine-celled mandala
pattern. Below this, we encountered two more fills which were siltier and contained large clay
lumps as part of prepared fills. From an analysis of the soil composition, it became clear that these
deposits were not random accumulations but had been carefully prepared and placed within the
cells. Kubiena tin samples from this stratigraphy have been taken for micromophorphological
geoarchaeological analysis and the stratigraphy was also subjected to XRF analysis to determine
their composition. Preliminary analyses of these sediments has given clear indication of a ‘recipe’ of
contrasting materials from different sources, deliberate organisation of materials and marked
contrast with the naturally deposited sediments prior to monument construction.

4.10 As we excavated further, it became clear that neither the north-west nor the north-east free-
standing brick piers and their saddlestones had been damaged by seismic activity (Figure 34).
Already demonstrated by our excavation of the south-west pier and saddlestone in 2015, we believe
that the use of brick in mud mortar enhances the piers’ resilient to seismic shock. Similarly, the east-
west cross-wall between the northeast and northwest brick piers was also undamaged from seismic
events. On reaching the depth of 26 brick courses below the top of the cross-wall and brick piers,
we identified a slight change in the composition of the soil fill. This was accompanied by the
exposure of additional cross-walls between the taller cross-walls and running under the Gorakhnath
Shrine in our north-facing and east-facing trench sections (Figure 35). We recorded three of these
cross-walls in our deep trench, two running north to south and one running east to west (Figures 36
and 37). A copper alloy coin was recovered next to the eastern-most of these lower north-south
cross-walls and will be analysed later (Figure 38). It may represent a foundation deposit within
foundations of the monument. It is significant to note that these lower level walls are not present in
the outer cells of the mandala as they were not identified in our deep excavations in the west and
south-west of the monument in 2015.

4.11 The lower cross-walls were found to extend to a depth of eight courses. Abutting the larger
cross-walls that formed the nine-celled mandala across the entire structure, we noted that they
were similar in form and also constructed as a single line, one brick thick. Similarly, they were not
bonded to the large cross-walls but abutted them. Therefore, we conclude that they were laid later
than the main cross-walls but, before OSL dating samples are analysed (Figure 39), we are unable to
determine whether this was during the same construction phase or later. The walls do confirm,
however, that a second smaller mandala was constructed at a lower depth, within the central cell of
the monument, possibly below the central sanctum, within the larger nine-celled mandala formed
between the brick piers and the foundation walls. The larger mandala, the mandala within the
central mandala and the gold foil mandalas found within the sockets of the four large central
saddlestones allude to the cosmological and symbolic significance associated with the construction,
reconstruction and renovation of the Kasthamandap.

8
4.12 Following the confirmation of the major 1960s restoration errors, we also examined the
surviving wooden elements of the Kasthmandap, which had been salvaged and stored within the
Hanuman Dhokha after the earthquake. This study was conducted in order to understand more
about the impact of the earthquake on the structural system of the superstructure and its relation to
the sub-surface foundations. The analysis started with a condition assessment of the major four
timber pillars, their connecting beams and the brackets. Three of the pillars had a base tenon to tie
the pillar into its saddlestones and a second tenon to lock the upper part of the pillar into the
brackets and beams above. All the latter were broken during the earthquake but one showed signs
of having been renewed at some stage in the past. The base of the fourth pillar has clear evidence
of 12 cm deep cut, indicating the replacement of the tenon piece at some time in the past. The 7cm
deep base tenons of two of the pillars are in good condition, while one of them is slightly rotten.
The fourth tenon is completely absent, indicating a major structural weakness within the
superstructure. Among the 16 brackets used in its central structure, six are in intact, while the other
10 brackets show evidence of the impact of the movement of the pillar, hence resulting in different
break patterns. Similarly, there is evidence of major movement of the pillars in the damage pattern
seen in the beams as well. Only one of the beams has one end intact, which must have been holding
the north-east column with the replaced tenon, which failed easily during the earthquake. One of
the major pillars was sawed into two pieces by a chain saw during the emergency rescue phase
immediately after the earthquake. Most of the outer pillars on the ground floor had decayed bases
and broken tenons at the top, while the base tenons of the pillars of the upper floors are in good
condition with the broken tenons due to earthquake impact on the overall structure. The north-
west pavilion of the Kastamandap is still preserved complete with wooden planking details below
the tiled flooring in cement mortar, added during the 1966 restoration (40, 41, 42 and 43).

5. Research Conclusions

5.1 Our two seasons of rescue excavations at the Kasthamandap, one of Kathmandu’s most
significant monuments, have provided invaluable new data for archaeologist and historians as well
as information for engineers and architects tasked with its reconstruction. Furthermore, our
findings have provided new historic and archaeological evidence which indicates that the
Kasthamandap’s brick foundations were originally laid in the seventh century CE but that it had been
subject to a major campaign of remodelling with the introduction of bracing cross-walls within 200
years. This much earlier date indicates that the Kasthamandap’s foundations may be attributed to a
late Licchavi date, even if one favours the later range of the OSL dates (Coningham et al. 2016).
Reflecting on Slusser’s earlier comment that “most of the principal national shrines, the temples and
stupas, can be traced to Licchavi foundations, for the most part the superstructures represent
restorations…the many scattered architectural fragments above ground attest to the splendour of
Licchavi architecture” (1982: 39), we can now attest to the fact that the ground plan of the
Kasthamandap had been firmly established by the beginning of the Transitional Period in 879 CE.
Furthermore, deep excavations in 2016 identified a further nine-celled mandala within central cell,
below the Gorakhnath Shrine (Figure 44, 45, 46 and 47).

5.2 The Kasthamandap’s bracing cross-walls not only offered a structural strength but the resultant
plan of nine units or cells, three north to south and three east to west, formed a mandala. The nine-

9
celled mandala was further elaborated within its central portion. During the deep excavation to
investigate the presence or absence of seismic distortion or damage to the north-east brick pier
which supported the formerly ‘missing’ north-east saddlestone, we encountered additional cross-
walls of eight courses of single brick. Similar in size to those of the cross-walls previously identified,
we postulate that these deeper cross-walls formed a further nine-celled mandala at the centre of
the monument. This further demonstrates the symbolic mandala that the monument’s foundations
formed and the microcosm created within the Kasthamandap’s foundations. We await the OSL
dating of these newly identified cross-walls within the central larger mandala but, at present, we
believe they may be of a similar phase. We can also state that the larger nine-celled mandala within
the original monument’s foundations and brick piers were initiated and then refilled with sand and
clay in the ninth century CE.

5.3 Reference should also be made to the similar, though smaller, foundation (S21) excavated by the
Italian mission at Harigaon Satya Narayana between 1984 and 1988 (Verardi 1988). Measuring 2.2
metres square and surviving to a depth of 0.66 metres, S21’s square foundation walls and cross-
walls formed nine square pits (ibid.: 65). Verardi dated the framework of walls to the Transitional
Period (ibid.: 68) and clearly recognised them as forming a navakunda, “a mandala subdivided into
nine padas, which is one of the models envisaged in the traditional treatises on Newar architecture”
(ibid.: 65). As at the Kasthamandap, Verardi found the nine kundas filled with sterile sand and
recorded contemporary rituals within the Valley, which involved the “custom of constructing
foundation walls with nine pits in a sacred building…After the prescribed ritual, the pits are filled
with sand or earth. The ritual documented…prescribes that in each pit nine different kinds of grain
are thrown…According to another recorded ritual, it is the powder of the pancarangis, or ‘five
minerals’ (gold, silver, copper, brass and iron)… which is thrown in the kundas. The foundation of
the sacred building, conceived and laid in the above way, is then sealed with a paved floor after
having been consecrated” (ibid.). Initially interpreted as the foundations for a small temple, he later
suggested that they had formed the base for a stupa, which was later entirely destroyed in the
eighth century CE (Verardi 1992: 78). This reinterpretation was later critiqued by Tiwari, who stated
that it formed the sacred foundation for a square temple (Tiwari 2002: 111), and this bears a
remarkable similarity the morphology exposed within the foundations of the Kasthamandap during
our recent excavations. It is anticipated that future results of XRF analysis, particle size distribution
analyses and Kubiena tin micromorphology will assist us in ascertaining whether the fills within the
different cells exposed within the Kasthamandap in deep stratigraphy during our 2015 and 2016 field
season’s show any evidence of differing soil composition.

5.4 The symbolic mandala pattern is reaffirmed through the discovery of gold foil mandalas within
the four large central saddlestones, forming the major elements of the monument’s superstructure.
Such objects are relatively rare and, in her review of consecration rituals in Asia, Slaczka noted 200
examples from the eighth to the fourteenth century CE (2007: 2). Her evidence suggests that
elaborate rituals were conducted across the region and treatises, like the twelfth century AD
Kasyapasilpa, indicate that deposits, or garbhanyasas, provided “life-breath” to a building (ibid.: 84)
and prosperity and welfare to those performing the ceremony (ibid.: 201). Other texts describe the
placing of gold under main pillars (ibid.: 212), perhaps relating to rites of cosmological significance,
such as creating an axis mundi (ibid.: 212-213).

10
5.5 The discoveries of our 2016 season have not only furthered our knowledge of foundation
deposits within saddlestones but also provided more information on traditional construction and
architectural methods. Once the surface of the north-east saddlestone was cleaned, we were able
to confirm the presence of the residue from a copper plate as already recorded on the three large
saddlestones exposed in 2015. Identified as a damp proof course between the brick foundations
and the timber superstructure, the presence of the remains of a copper shoe originally nailed
around the tenon of the large wooden pillar is also striking. Still with nail holes present, it further
illustrates now forgotten traditional techniques of ensuring the longevity of the timber tenon – a key
element in ensuring the strength, resilience and integrity of the entire monument. The presence of
the copper shoe and wood tenon inside, suggests that the base of the pillar had already severely
deteriorated but that conservators in the 1960s merely pushed tiles under the pillar base rather than
dismantle the entire structure. The improper restoration of one of the major structural pillars of the
Kasthamandap became the weakest point during the seismic event.

6. Observations and Recommendations

6.1 A major contributing factor to the collapse of the Kasthamandap was the tiling over of the north-
east saddlestone during the monument’s conservation in the twentieth century conservation, which
detached a key element of the superstructure from with the resilient foundations below. Therefore,
we recommend that the inner foundation wall and outer wall of the Kasthamandap should be
retained as far as possible and should only be strengthened or replaced if there is valid
justification. In this instance, we advise that valid justification includes where there are missing
elements from damage caused by JCBs during the post-disaster emergency response, as well as the
resetting of saddlestones ripped out from the structure, or in areas where there is slight
earthquake damage in the upper courses below in-situ saddlestones.

6.2 Some debris and rubble remains within the fenced compound and needs to be carefully sorted
and removed prior to consolidation and reconstruction activities begin. This rubble is particularly
thick on the southern half of the monument and contains both modern and archaeological material.
It is recommended that the remaining debris and rubble within the fenced compound is removed.
As it contains both historic and modern material, we recommend that a grid of 3 by 3 metres is laid
out and squares cleared systematically to ensure spatial provenance. It is also recommended that
this material is carefully sieved as it is highly likely that it will include modern personal effects,
which may assist with identification of those killed in the monument, and antiquities. The material
is also likely to contain reusable bricks, which should be separated and stored. These activities
should be undertaken within an Archaeological Watching Brief.

6.3 Having exposed the remaining in-situ elements of the Kasthamandap, we can confirm that the
majority of damage to the monument’s foundations directly related to the activities associated with
the post-disaster emergency recovery phase. This included severe damage to the southern half of
the monument, where bulldozers and JCBs has cut through the upper courses of both the internal
foundation and exterior walls. Displaced or loose bricks should be removed from the damaged
portions of the walls and reused bricks of the same size and strength be used to reconstruct the
walls to their original heights and thicknesses. The mud mortar originally used in the monument’s

11
foundations should be analysed in a laboratory for clay type and particle size distribution and
replicated for reuse within the reconstruction. It is as significant to model the effect of moisture
levels on joint strength. These activities should be undertaken within an Archaeological Watching
Brief.

6.4 The deployment of machinery during the post-disaster emergency recovery phase also displaced
a number of the double saddlestones that supported the Kasthamandap’s timber superstructure as
well as almost all of the single saddlestones which had rested above the outer wall. A number of
these saddlestones are reusable and lie both within and above the rubble at the site. Undamaged
double and single saddlestones should be reclaimed from the rubble at the site and should be
reinstalled in their original locations and rebonded. If damaged beyond use, duplicates should be
prepared from similar stone.

6.5 We recorded some evidence of minor earthquake damage within small portions of the upper
courses of the inner foundation wall. These comprised vertical cracks from earthquake damage
below two of the double saddlestones on the western stretch of north-south inner foundation wall.
This type of earthquake archaeological effect, particularly below a structural element, provides
evidence of stress within a building during a seismic event as the pressure of ‘up and down’
movement of a saddlestone during an earthquake causes the wall below to sustain damage with
cracks forming in the brickwork (Pers. Com. Dr Paolo Forlin). Whilst it is not possible at present to
date when these cracks formed, the evidence of this damage is limited, and the foundations
exhibited little major structural damage due to the 2015 earthquakes or previous events and appear
exceptionally resilient. The cracked bricks should be removed and reused bricks of the same size
and strength be used to reconstruct the walls to their original heights and thicknesses. The mud
mortar originally used in the monument’s foundations should be analysed in a laboratory for clay
type and particle size distribution and replicated for reuse within the reconstruction. These
activities should be undertaken within an Archaeological Watching Brief.

6.6 As stated above, we recommend that the core of the monument should be rebuilt utilising as
much of the original foundations as possible and we also recommend that its superstructure be
reconstructed with as many original materials as possible using traditional techniques. However,
due to the damaged condition of many of the timbers, we understand that many elements will have
to be freshly cut, seasoned and prepared. We recommend that this process is accompanied by the
scientific analysis, including dating, of timber from selected unrecyclable elements in order to allow
the development of a clear understanding of past cycles of construction and maintenance to guide
the cyclical renewal of the monument in the future. We therefore recommend the extraction of
samples from major timber elements from the Kasthamandap to construct a timeline of past
construction and maintenance. These activities should be undertaken within an Archaeological
Watching Brief. While reusing salvaged elements, proper repair details should be designed
considering its structural role and each of the techniques used should be documented and noted
for the long term maintenance plan.

6.7 We recognise that our excavations, and those of the JCBs, have created a number of deep voids
within the Kasthamandap’s foundations, now refilled with a mixed soil. We strongly advocate that
this mixed soil is replaced by clean silts and clays, with similar composition and compaction to that
excavated during our investigations.

12
6.8 We recommend that the mandala wall patterns are protected and preserved in-situ below any
new floor surface within the reconstructed monument. We also recommend that this important
symbolic design is incorporated into whatever floor surface is selected. Indeed, the subsurface
alignments of these cross-walls could be marked in different brick patterns on the modern upper
surface, as seen at Jarrow Monastery in the UK (Figure 48), serving to inform visitors to the
mandala’s location.

6.9 Our observations at the Kasthamandap highlight the strength of traditional foundations within
this active seismic region with a lack of earthquake distortion present throughout the deep sections
of foundations and cross-walls exposed. Indeed, historical sources, some corroborated by analysis
of soil stratigraphies which have identified some non-documented seismic events, suggest major
earthquakes in c. 1100 CE, 1224 CE, 1255 CE, 1260 CE, 1344 CE, 1408 CE, 1430 CE, 1681 CE, 1767 CE,
1803 CE, 1810 CE, 1823 CE, 1833 CE, 1869 CE and also more recently in 1916 and 1934 CE (Gautam
et al. 2015 1-3, Mugnier et al. 2013, Sapkota et al. 2013, Schwanghart et al. 2015). Thus, the
foundations of the Kasthamandap have survived undamaged through many of these catastrophic
events. We suggest that the collapse of the superstructure of the Kasthamandap may therefore be
linked to superstructure maintenance issues, particularly the breaking of the tenon and mortace
joint on the north-east saddlestone. We recommend that structural engineers create a 3D analysis
model using line, surface and joint elements in order to consider performance against earthquake
vibrations and model the effect of the structural weaknesses and conservation errors identified.

6.10 From our investigations at Kasthamandap, it is also clear that there is a need to undertake inter-
disciplinary research whereby archaeologists, geo-scientists, engineers and architects can work in a
collaborative environment to ensure a sympathetic, authentic and sustainable rehabilitation
reconstruction and renovation of a monument. In this environment, each discipline can utilise its
expertise to inform and aid the development of plans for conservation and reconstruction. There is
a pressing need to initiate, design and deliver short courses on the subsurface archaeological
heritage of the Kathmandu Valley for non-archaeological specialists. In association with an
introduction to modern archaeological evaluations and the necessity for rescue and urban
archaeology and the practicalities of urban planning and watching briefs, this will raise awareness
of the need to consider the evaluation and assessment of a monument’s original foundations, to
understand whether these contributed to collapse, or are a vital part of its resilience, the latter as
found at the Kasthamandap.

6.11 There are a number of UNESCO World Heritage sites at risk from direct seismic events within
the Himalayan belt in Afghanistan, Bangladesh, India, Myanmar, Nepal and Pakistan, as evidenced
from the 2015 Nepal and 2016 Myanmar earthquakes. In addition, it is also noteworthy mentioning
the potential impact of related tsunami damage in India and Sri Lanka from primary events
elsewhere through short-term environmental shocks. Whilst national and international
communities have been able to mobilise emergency medical, rescue and engineering teams, there is
currently little capacity for emergency heritage teams within the region. Indeed, expertise within
urban and rescue archaeology within the South Asian region is extremely limited with most
professional archaeologists having gained their experience from ‘green field’ archaeological sites and
excavations. The mission’s involvement in the training of archaeologists within the Kathmandu
Valley, with the development of links with colleagues from Sri Lanka, India and Myanmar, indicates
the potential for developing a broader pan-Asian team who would be able to accompany

13
multidisciplinary emergency teams and thus reduce the immediate adverse impact on cultural
heritage of emergency and recovery activities. We recommend the urgent developing of South
Asian-wide capacity in rescue archaeology to respond to cultural disasters and emergencies at
UNESCO World Heritage Sites and to develop capacity in rescue excavation and watching briefs in
urban situations.

7. Acknowledgements

We would like to acknowledge the support of the following individuals for their help and expertise in
the field and during the archaeological activities in 2016: Hon. Minister Mr. Jiwan Bahadur Shahi MP,
Hon. Ashta Laxmi Shakya MP, Hon. Rajya Laxmi Shrestha MP, Mr Bhesh Dahal, Mr Bharat Subedi, Mr
Kai Weise, Dr Roland Lin, Mr Christian Manhart, Mrs Nabha Basnyat-Thapa, Mrs Nipuna Shresta, Mr
Damodar Gautam, Mrs Manju Singh Bhandary, Mr Bhaskar Gaywali, Mrs Saubhagya Pradhananga,
Mrs Aruna Nakarmi, Mr Raj Kumar Banjara, Mr Jagat Bahadur Katuwal, Miss Shanti Sherma, Miss
Sunita Bhadel, Mr Ranjan Kumar Dulal, Mr Kulnath K.C., Mrs Maiya Kaiti, Mrs Bindhaya Karki, Mr U
Ko Ko Aung, Mr U Ye Lwin, Mr Priyantha Kumara, Mr Thusitha Herath, Mr Pushpa Kumara, Mrs
A.M.D Kumar, Ms Pratima Sapkota, Mr Govinda Neupane, Mr Hari Prasad Bhusaal, Mr Surendra
Bhandari, Mr Niranjan Adhikari, Mr Susil Kuma Gautam, Mr Dinesh Chandra Regmi, Mr Kiran Man
Chitrakar, Professor S.R. Tiwari, Professor P.N. Maskey, Professor Prishanta Gunawardhana, Mr
Gaurab Shrestha, Mr Purushottam Awale, Mrs Patricia Voke, Ms Brooke Tully, Mr David McCaughie,
Mr Adem Kilic, Mrs Zahra Hussein, Mr Urban Coningham, Mr Augustus Coningham, Ms Anouk
Lafortune-Bernard, Professor K. Krishnan, Mr Vrushab Mahesh, Dr Paolo Forlin, Dr Nina Mirnig and
Dr Keir Strickland. Particular thanks are recorded to Ms Anie Joshi for the preparation of the
architectural plan of the layout of the Kasthamandap from data produced during the archaeological
investigations. We are also grateful to Mr Ramesh Ratna Tamrakar and Mrs Durga Devi Tamrakar as
well as the Kasthamandap Café for allowing us to take photographs of the Kasthamandap from their
properties during the field season.

We would like to thank the Arts and Humanities Research Council (UK), for a Global Challenges
Research Fund Grant (AH/P006256/1) and the National Geographic Society for a Conservation Award
((#C333-16) for their generous financial support. We are also grateful for the institutional assistance
and support provided the Department of Archaeology, Government of Nepal, Durham University,
the University of Stirling and Lumbini Buddhist University. We also thank the Oriental Cultural
Heritage Sites Protection Alliance for providing funds for Officers from the Department of
Archaeology, Government of Myanmar to join the project, and also to the Central Cultural Fund,
Government of Sri Lanka for providing funds for their officers to participate in the field programme.
Finally, we would like to thank the Kathmandu Metropolitan City and communities of Hanuman
Dhoka for their support and interest in our mission.

8. References

Coningham, R.A.E., Acharya, K.P. Davis, C.E. and Kunwar, R.B. 2015. Post-disaster urban
archaeological investigation, evaluation and interpretation in the Kathmandu Valley World Heritage
property. Report and Recommendations of a mission conducted between 5/10/2015 and
22/11/2015. Report for UNESCO Kathmandu Office and Department of Archaeology, Government of
Nepal.

14
Coningham, R.A.E., Acharya, K.P., Davis, C.E., Kunwar, R.B., Simpson, I.A., Schmidt, A. and Tremblay,
J.C. in press. Preliminary Results of Post-Disaster Archaeological Investigations at the Kasthamandap
and within Hanuman Dhoka, Kathmandu Valley UNESCO World Heritage Property (Nepal). Ancient
Nepal 191: 28-51.

Gautam, D., Pradhananga, S., Kunwar, R.B. and Sharma, M.K. 2015. Preliminary report of Monuments
Affected by Earthquake April 25 2015. Kathmandu: Department of Archaeology.

Korn, W. 2007 (second edition). The Traditional Architecture of the Kathmandu Valley. Kathmandu:
Ratna Pustak Bhandar

Mugnier, J-L., Gajurel, A., Huyghe, P., Jayangondaperumal, R., Jouanne, F. and Upreti, B. 2013.
Structural interpretation of the great earthquakes of the last millennium in the central Himalaya.
Earth-Science Reviews 127: 30-47

Slaczka, A.A. 2007. Temple Consecration Rituals in Ancient India: Text and Archaeology. Leiden: Brill.

Slusser, M.S. 1982. Nepal Mandala: A Cultural Study of the Kathmandu Valley. Princeton: Princeton
University Press.

Sapkota, S.N., Bollinger, L., Klinger, Y., Tapponnier, P., Gaudemer, Y and Tiwari, D. 2013. Primary
surface ruptures of the great Himalayan earthquakes in 1934 and 1255. Nature Geoscience 6: 71-76.

Schwanghart, W., Bernhardt, A., Stolle, A., Hoelzmann, P., Adhikari, B.R., Andermann, C., Tofelde, S.,
Merchel, S., Rugel, G., Fort, M. and Korup, O. 2015. Repeated catastrophic valley infill following
medieval earthquakes in the Nepal Himalaya. Science 351(6269): 147-150.

Tiwari, S.R. 2002. The Brick and Bull: An Account of Handigaun, the Ancient Capital of Nepal. Lalitpur:
Himal Books.

Tokyo National Research Institute for Cultural Properties (TNRICP). 2016. Project for Investigation of
Damage Situation of Cultural Heritage in Nepal: Structural Survey of Historic Buildings. Tokyo: Tokyo
National Research Institute for Cultural Properties.

Verardi, G. 1988. Harigaon Satya Narayana, Kathmandu: A Report on the Excavations Carried out in
1984-1988. Rome: IsMEO.

Verardi, G. 1992. Excavations at Harigaon, Kathmandu. Rome: IsMEO.

15
9. Figures

Figure 1: View to the northeast of excavations at the Kasthamandap in November 2015.

Figure 2: JCBs clearing rubble around the Kasthamandap on the 26th April 2015 (image courtesy of
Kai Weise).

16
Figure 3: East facing section of excavations at the Kasthamandap in 2015

Figure 4: 2015 Schematic elevation of the Kasthamandap’s superstructure superimposed on the

17
Figure 5: Saddlestone with cross-walls connecting to the brick pier uncovered in southwest corner of
the Kasthamandap in 2015 excavations.

Figure 6: Briefing meeting held at Hanuman Dhoka Palace in advance of the archaeological
investigations at the Kasthamandap in 2016.

18
Figure 7: The Kasthamandap, looking west, prior to excavations, November 2016.

Figure 8: North-facing section, illustrating building-up of rubble to recreate the footprint of the
Kasthamandap.

19
Figure 9: Saddlestone identified within the reconstituted rubble of the Kasthamandap.

Figure 10: Clear marks from the bucket of machinery cut across the tiled surface, looking south.

20
Figure 11: Post-disaster damage and diagnosis plan.

21
Figure 12: Cut from the JCBs through the southeast corner of the Kasthamandap’s foundations,
looking west.

22
Figure 13: Damage to outer wall of the Kasthamandap, with overlying rubble from 2015 visible in the
west and east facing sections.

23
Figure 14: Surviving tiled surface across the Kasthamandap, illustrating significant damage to the
southeast and southwest corners of the monument.

Figure 15: Vertical cracks visible below a double saddlestone in the east facing elevation of the
western north-south foundation alignment.

24
Figure 16: Impact damage on the tiled surface of the Kasthamandap, most-likely from falling
superstructure during the 25th April 2015 earthquake.

Figure 17: Photograph of excavations in 2015, illustrating location of missing northeast saddlestone.

25
Figure 18: Half section exposing the ‘missing’ northeast saddlestone.

26
Figure 19: Half section of northeast saddlestone, showing its alignment with the southeast
saddlestone to the south

27
Figure 20: Presence of the four saddlestones around the Gorakhnath shrine.

Figure 21: Double saddlestone set on foundation wall with one socket cemented over.

28
Figure 22: Tiled over double saddlestone on northern east-west stretch of foundation wall, looking
west.

Figure 23: Copper shoe within socket of northeast saddlestone and evidence of copper plate residue
on the surface of the saddlestone around this socket.

29
Figure 24: Nail holes visible in copper shoe within the socket of the northeast saddlestone.

Figure 25: Gold foil mandalas present at the base of the socket of the northeast saddlestone.

30
Figure 26: Detail of gold foil mandalas present at the base of the socket of the northeast saddlestone

Figure 27: Post-excavation photograph of the gold foil mandalas from the northeast saddlestone.

31
Figure 28: Surkhi surface across the Kasthamandap, again showing the JCB cuts to the southeast and
southwest.

Figure 29: Twentieth century Rana period brick paving or setting across the Kasthamandap, again
showing the JCB cuts to the southeast and southwest.

32
Figure 30: Brick paving, utilising a mixture of modern and old brick across the Kasthamandap, again
showing the JCB cuts to the southeast and southwest. A stone paved surface is present in the
northwest.

Figure 31: Detail of stone paving in the northwest of the Kasthamandap.

33
Figure 32: Architectural stone fragment within the stone surface to the northwest of the
Kasthamandap.

Figure 33: Exposed nine-celled mandala formed by cross-walls between four central brick piers and
saddlestones and within the foundation wall.

34
Figure 34: Brick pier below the northeast saddlestone, indicating no evidence of seismic damage.

35
Figure 35: Lower cross-walls, looking northwest, identified running north-south within the central
cell of the nine-celled mandala.

36
Figure 36: Lower cross-walls, looking east, identified running north-south within the central cell of
the nine-celled mandala.

37
Figure 37: Lower cross-walls, looking south, identified running north-south and east-west within the
central cell of the nine-celled mandala.

Figure 38: Coin identified in lowest fill, associated with the lower cross-walls within the central cell of
the nine-celled mandala.

38
Figure 39: OSL dating sample being taken from lower phase of cross-walls forming mandala within
the central cell of the large nine-celled mandala.

39
Figure 40 Detail showing linkage between the Kasthamandap’s foundations and superstructure.

40
Figure 41 Four main timber pillars from the Kasthamandap.

41
Figure 42 Beams and brackets from the Kasthamandap.

42
Figure 43: Planking and joists in the north-west pavilion of the Kasthamandap.

43
Figure 44: Architectural plan of the phasing of the Kasthamandap, illustrating locations of outer wall,
foundation wall, cross-walls and surviving in-situ saddlestones as well as pre-earthquake exterior
footprint of the monument

44
Figure 45: Schematic section through the Kasthamndap based on 2016 feidlwork.

45
Figure 46: 2016 Schematic elevation of the Kasthamandap’s superstructure superimposed on the
east-facing archaeological section of the foundations (Architectural elevation after Korn 2007).

46
Figure 47: Provisional architectural phasing of the Kasthamandap.

47
Figure 48: Different phases of subsurface wall alignments represented through differing brick
patterns at the site of Jarrow, United Kingdom.

48
Figure 49: Visit to excavations of the Kasthamandap by Minister Shai, Minister of Culture, Tourism
and Civil Aviation (Government of Nepal)

Figure 50: Puja held outside the Kasthamandap on the 19/12/16, with the cross-walls and
foundations of the monument visible.

49
Figure 51: Team photo, December 2016.

50