Conserving the tomb chamber complex in the Midas Mound at Gordion in Turkey

Full text

1 INTRODUCTION
1.1 History
The area around the ancient Phrygian capital of
Gordion in central Turkey (modern Yassıhöyük) is
marked by over 100 tumuli, or earthen burial
mounds. By far the largest of these is Tumulus MM,
which still stands 53 m tall with a diameter of circa
300 m at its base (Figure 1). Traditionally consid-
ered the tomb of Midas, the most famous of Phrygi-
an kings, it has now been shown to have been that of
his father but built by Midas after his father’s death.
Tumulus MM was excavated in 1957 by Rodney
S. Young of the University of Pennsylvania (USA),
who found in the tomb chamber proper the body of a
60-65 year-old male lying on the remains of a cedar
coffin, surrounded by grave goods that had been
used in a funerary feast (Young, 1981).
Dating to circa 740 BC, the tomb is the oldest
standing wooden building in the world. It was pre-
served for over 2,700 years because after the burial,
the tumulus sealed the environment around the tomb
chamber, and the soft-rot fungus that would normal-
ly destroy the wood was stopped short by the re-
duced temperature and moisture levels. The archeo-
logical excavation changed that environment, and
thoughts turned quickly to preserving this unique
monument. Much of this preservation has been doc-
umented, but the present paper discusses the conser-
vation efforts from a structural engineering perspec-
tive.
The timber structure miraculously has survived
despite being buried under approximately 53m of a
man-made mound with moisture infiltration a con-
stant threat. The ancient Phrygian culture created a
fascinating protection system for the tomb that is
still not fully recognized.
Figure 1. Tumulus MM (c. 740 BC) viewed from west.
Since the tomb’s discovery by archeologists, all
activities related to the Midas tomb can be catego-
rized as preventative conservation. Most activities
occur each summer from mid-June through mid-
August.
1.2 Tomb construction
The construction of the tomb exhibits the high
skills of the builders of the time, and it reflects more
than 100 years of tumulus building by the Phrygians
(Liebhart, 2013).
The builders began with a large pit some 16 m by
18 m and around 2 m deep (Figures 2-3). This pit
was lined with soft limestone blocks creating a pe-
rimeter wall about 0.8-1.0 m thick. The pit itself was
filled with rough stone cobbles, with larger boulders
Conserving the tomb chamber complex in the Midas Mound at Gordion
in Turkey
D. Biggs
Biggs Consulting Engineering, Saratoga Springs, NY, USA
R. Liebhart
PhD, retired archaeologist, Chapel Hill, NC, USA
S. Gönen
PhD candidate, Bogazici University, Department of Civil Engineering, Istanbul, Turkey
ABSTRACT: The tomb chamber (c.740 BC) in the Midas Mound at the ancient Phrygian capital of Gordion
in central Turkey (modern day Yassıhöyük) is the earliest known standing wooden building in the world. Be-
lieved to be the tomb of the father of the famous and historical King Midas, the structure was constructed and
then buried following a great funerary feast. The tomb is covered by a man-made mound (Tumulus MM) ap-
proximately 53 m tall and 300 m in diameter at its base. The chamber was constructed with cedar floor beams
and pine beams for the walls and double-pitched roof. The tomb chamber proper was reinforced by an exterior
wall of juniper logs. The interior and exterior walls are separated by rubble stone. The structure was sur-
rounded by rubble stone and covered with soil that protected it from moisture for over 2,700 years.
This paper provides information on the tomb’s wooden structure and describes conservation efforts to pro-
tect the tomb and prevent further structural damage.

Figure 2. Stone pit and timber floor. (see Figure 5 for material
designations).
Figure 3. Partially completed tomb chamber (see Figure 5 for
material designations).
creating a kind of inner foundation or stabilizing
ring roughly along the lines of what would be the
tomb chamber. The stone lined pit also prevented
water from accumulating around the timbers. Bed-
ded in the rubble were pairs of juniper logs at the
shorter north and south sides of the eventual cham-
ber. The Phrygians frequently used juniper in similar
ways because of the wood’s high density and re-
sistance to fungal attack. These logs were trimmed
flat to provide a bearing surface for the cedar floor
beams.
On the floor were set the pine wall beams, with
the shorter north and south beams fitting into simple
notches in the east and west wall beams (Figure 4)
in order to withstand the inward pressure of the tu-
mulus. This pressure also helped to hold the tomb
chamber walls in position. The tomb chamber creat-
ed was 5.15 m by 6.20 m, almost twice the size of
the second largest excavated wooden tomb chamber
at Gordion.
As each wall course was installed, juniper logs of
an outer casing were set parallel to the pine walls
and about 35 cm away. At the same time, the perim-
eter limestone wall was raised, with more stone
cobbles filling the spaces between this stone wall
and the juniper outer casting, and between the outer
casing and the pine walls of the tomb chamber prop-
er.
While the timber operation was going on, the tu-
mulus was also being built up against the outer face
of the perimeter stone wall, which was apparently
designed to provide a solid barrier between the
earthen tumulus fill and the rough stone cobbles of
the tomb chamber complex. The perimeter wall was
not designed as a retaining wall for either the tumu-
lus or for the loose stones inside; rather, the entire
construction served to envelop and hold in place the
walls of the tomb chamber proper.
Figure 4. Tomb chamber plan
When the tomb chamber walls reached ca. 2.65
m high, two superimposed pairs of pine beams were
set on the top, and these new beams extended out to
rest also on notches cut into the juniper logs of the
outer casing at east and west. The pine walls were
built up to the tops of these cross beams. Next the
pine beams of the angled supports for the double-
pitched roof were installed, again set in pairs. At this
point, the tomb chamber looked like a wood-lined
pit, surrounded by the tops of the outer casing, cob-
bles, and perimeter stone wall, all encased by the
tumulus, which was rising at the same schedule. It
cannot be determined whether the tumulus at this
stage of construction was built to its final diameter,
but circumstantial evidence suggests that it was.
The tomb was now ready for the funeral of the
dead king (Figure 3), after which the tomb chamber
was sealed by setting two layers of pine beams over
the angled roof supports. A secondary roof of juni-
per logs was built over the tomb chamber, and the
entire tomb chamber complex was covered with a
mound of yet more rough stone cobbles (Figure 5).
Figure 5- Completed tomb.

The tumulus was then finished to a height even
greater than its current eroded height of 53 m. The
tomb chamber roof and walls sagged, bowed, dis-
torted, and cracked under the weight of the tumulus,
but the tomb chamber roof never collapsed: the
Phrygian builders had gotten their “engineering” ex-
actly right. Eventually, the environment in the tomb
chamber reached a state of equilibrium, with a con-
stant temperature of ca. 15º C and a low relative
humidity.
Upon its completion, Tumulus MM contained
over 1,250,000 m3 of earth, hundreds of stone
blocks, millions of rough cobbles, approximately
200 wooden timbers, a vast array of grave goods,
and the body of one man (Liebhart & Johnson,
2005).
During the early years, a slow-acting soft rot fun-
gus was able to damage the wood until it finally died
off. The timber was able to maintain most of its
strength and still feels solid to the touch. In some
sections, it is possible to detect wood that is soft and
powdery.
2 STRUCTURAL CONSERVATION
2.1 1957 Excavation and stabilization
With a tumulus the size of MM, Young knew that
digging blindly through the earthen mound could
prove futile, as well as overly expensive. Because
most of the smaller tumuli already excavated had
tomb chambers off center in the mound, Young was
determined to locate the rock pack that he correctly
assumed to be piled over the tomb chamber during
its construction.
After experimenting on two smaller tumuli with a
water-cooled drill and measuring the depth of each
hole, he started drilling in Tumulus MM in the fall
of 1955, finishing the 96 borings in 1956 (Figure 6).
Figure 6. Partial plan of drillings over the tomb.
The drilling operation was successful in deter-
mining the limits of the rock pile (and thus, the loca-
tion of the tomb chamber), but the water that was
supposed to recycle back out of each hole failed to
do so, and after 2,700 years of being left alone, the
tomb chamber was soaked with water (even partially
filling some of the bronze bowls found on the tomb
chamber floor). This unfortunate spike in moisture
apparently did relatively little immediate damage to
the structural integrity of the tomb chamber, but
there is now no adequate way to assess the effects.
Young began the actual excavation in the spring
of 1957 with miners brought in from the coal dis-
tricts of Turkey. They excavated first via an open
trench cut primarily through the erosion slump of
the mound, then by tunneling toward the location of
the tomb chamber complex (Figure 7). When they
reached the perimeter limestone wall, they removed
some of the blocks, which released a flood of the
loosely packed stone cobbles. Temporary shoring
was used to create a passage to the juniper logs of
the outer casing. A doorway was cut through this in
what was the north half of the west wall; this was
followed by removal of more cobbles. Finally, a
door was cut into the tomb chamber proper, and
Young found himself literally at the foot of the dead
king.
Figure 7. Section through tumulus after excavation (from
north).
Of immediate concern in the tomb chamber was
that the lower members of the central cross beams
had bent downward and cracked from overload by
the pressure of the mound above, though they had
never hit the floor. The damage included splitting
along the wood grain of the lower center cross
beam. Figure 8 shows the inside of the tomb cham-
ber, the damaged center cross beam and the access
opening created to entire the tomb. Actually, the
beam splitting was precipitated by the orientation of
the grain that sloped downward from the east end to
the bottom of the timber girder near its mid-span.
This flaw in construction is peculiar given the high
level of craftsmanship attributed to the Phrygian cul-
ture and the known quality of the workmanship in
the timber building itself.

Figure 8. Interior of tomb.
There has been no attempt to repair the girder.
The goal has always been to conserve the timber
structure and prevent further structural damage. The
archeologists initially installed shoring that included
three wooden posts to prop these beams up—the
first preventative measure to conserve and protect
the structure of the tomb chamber.
Subsequently, steel framing was added to prop up
the girder and it has remained that way since (Figure
8). Eventually, the tomb conservation plan will be
extended to include the steel framing and renewed
corrosion protection will be applied to the framing.
No timber repair or strengthening is planned, only
continual conservation.
2.2 Earthen dome
During the excavation, the removal of the rubble
packing inside the perimeter wall exposed the un-
derside of a natural earthen dome (top of Figure 5),
which had been formed by the tumulus compressing
the fill material onto the rubble piled over the tomb
chamber complex. This dome held together during
the documentation and clearing of the objects from
the tomb chamber, although there were occasional
falls of cobbles that had adhered to the underside of
the dome. The archaeologists had no way to assess
whether this dome would hold, but they proceeded
nonetheless.
During the subsequent operation to protect the
tomb chamber in the following four years, the rest of
the rubble was removed, with wooden and steel
supports installed to hold the juniper logs of the out-
er casing. There were several instances of partial
collapses of the dome, although this did not stop the
protection efforts.
The removal of the stone above the tomb re-
moved the overburden pressure of the mound. The
stone removal around the sides removed the lateral
pressure from the walls. But, how did archeologists
know the earthen dome would be stable with nearly
50 m of earth above it?
From the drilling records, the archeologists en-
countered sand, clay, limestone and rock in their 96
drill holes (Figure 6). Records for drill holes 1 to 57
were lost in 1988. So, we have limited remaining
information from drill hole 58 that was nearly over
the tomb. In the immediate vicinity of the tomb,
they had approximately 22 holes they struck rock at
relatively shallow depths (shaded holes in Figure 6)
and were considered “duds”. They did help to iden-
tify the southwest and northwest perimeter of the
stone pile surrounding the tomb. In the same area,
there were approximately 29 holes that went deeper
(unshaded holes in Figure 6). Notes for hole 67 in-
dicate that they were not sure if the tomb was just
rock.
The drilling data is insufficient to clearly identify
the earthen dome size and its thickness. The arche-
ologists noted that the diameter of the rock pile cov-
ering the tomb was presumed to be about 30m di-
ameter from SW to N and greater from NW to SE
(Notebook 63, 1957). From the data from several
drill holes (58, 59, 67), the clay layer might be 3 to 4
meters thick.
Clearly, the geotechnical aspects of the tumulus
mound represent the most uncertainty attributed to
the construction of the tomb. The drillings were
made to locate the tomb, not to obtain geotechnical
information.
For the conservation of the mound, there should
be a structural evaluation of the earthen dome per-
formed. Therefore, exact information on the earthen
dome is required to verify the characteristics of the
clay as well as the thickness and profile of the dome.
2.2 Concrete protection structure
As part of the continuing conservation efforts, the
Turkish government completed a protective concrete
structure in 1961 to shield the tomb from the earth
pressure above (Figure 9). The structure was con-
structed with a gap between the earthen dome and
the structure. A thin concrete floor slab was con-
structed between the remaining stone wall and the
juniper logs. The slab was intended to provide lat-
eral bracing at the floor level. By 1963, the tunnel
created to access the tomb was also lined with con-
crete and stone for safe entry.
In 2015, the original calculations and sketches for
the concrete structure were found. Figure 10 shows
the original loadings for the concrete structure. The
structure is a concrete moment frame that was ana-
lyzed by hand calculations using the moment distri-
bution method. While the information obtained does
not accurately represent the as-built conditions, it
does provide evidence of the design intent which
was to fully support the earth above the tomb.

Figure 9. Cross section of tomb with protective structure (see
Figure 5 for material designations).
There were two unfortunate side effects of the
concrete construction. One was to introduce exces-
sive moisture into the tomb from the curing con-
crete. This added to the water from the drilling oper-
ation. The second problem created was the new
wood used in the formwork and shoring brought to
the tomb a brown rot fungus. Conservation efforts
have had to deal with these issues for the preserva-
tion of the original construction (Liebhart & John-
son (2005)).
Figure 10 – Sample of original design drawings.
As previously mentioned, the protective concrete
structure to shield the tomb from the earth pressure
above has been in place since 1961. However, it has
not been structurally utilized so far because the
earthen dome is separated from the structure by a
gap. Except for isolated areas where the underside
of the earthen dome has dried out and fallen onto the
roof of the concrete structure, the structure currently
only supports its own weight.
The recently discovered original drawings and
calculations reveal that although the as-built con-
struction differs slightly from the design, assump-
tions used in the calculations can be speculated as
adequate. The as-built construction has larger mem-
ber cross sections and lateral braces for the columns
thereby buttressing the perimeter walls. The design
reinforcement can be considered as well-detailed,
however the precise details of the actual reinforce-
ment remains unknown. It is reasonable to predict
that there would be some discrepancies between the
design and the construction when the working con-
ditions at that time are imagined.
Considering there are many visible defects in the
construction, future conservation efforts intend to
determine the material properties of the structure,
determine the reinforcement distribution, and carry
out a performance assessment of the structure. Per-
formance assessment is going to be made under the
loading of earth pressure above along with the cor-
responding earthquake effects. A reliability assess-
ment can be done if found necessary.
After judging the outcomes of the assessments
and the current conditions, a decision will be made
whether the concrete protection system requires any
strengthening. Complete utilization of what already
exists is preferred over providing new invasive solu-
tions.
2.3 Steel bracing
The removal of the lateral earth pressure on the
tomb required specific structural stabilization. The
timber tomb was constructed with corner joinery
that relies on compression created by lateral earth
pressure to maintain stability. Without the lateral
earth pressure, the concern was the juniper logs
might shift.
In the four years after the excavation of the tomb,
the remainder of the stone rubble was removed, and
timber bracing was installed to prevent outward
movement of the outer logs. This bracing remained
in place for nearly 30 years.
In 1993, a conference was held to discuss the an-
cient Gordion wood. Following the conference, a
significant conservation step was taken. The deci-
sion was to create a new system of bracing that re-
sulted in steel framing being designed and installed
in 2002 to replace the earlier timber system. The
bracing was to prevent movement; not to strengthen
the timbers.
Vertical posts were anchored to the new concrete
foundations and braced back to the concrete struc-
ture above (Figure 11). Each post has its own foot-
ing that buttresses against the column foundations
for the concrete structure. Adjustable rods with
flexible steel heads support each juniper log as seen

in Figures 11and 12. Each head has a neoprene pad
to accommodate seasonal movement of the juniper
logs. An inert membrane of “Marvelseal”, often
used with archival materials, was placed between
the neoprene and ancient wood to prevent moisture
transmission and staining. Tension rod bracing was
added to prevent lateral movement on the upper cor-
ners of the posts during an earthquake.
Figure 11. Steel framing supporting logs laterally.
3 MONITORING
3.1 Structural
There have been several monitoring campaigns that
continue today. To evaluate movement of the juni-
per outer walls and the inner pine walls, two types
of monitoring have been used. The first type of
monitoring includes plumb lines on the interior
which are used to measure offsets to the north and
south wall timbers. Additional offsets are taken
from the vertical posts of the steel framing to the
outer logs. The second type of monitoring uses tell-
tales at all four exterior corners to determine in-
plane movement (Figure 12).
Figure 12. Movement monitors at corners.
For two years, the monitors were only measured
during the summer work season. Those recordings
did not indicate any changes to the tell-tales. A
chance off-season winter visit in 1996 noted a 1 cm
change in the northwest corner. By summer, the
readings had returned to the initial readings. This
indicated there are seasonal changes occurring due
to temperature and humidity effects. Thus, monthly
readings are now being taken year round by the staff
of the Gordion Museum. Tell-tale 5 of Figure 12
was added after the seasonal movement was discov-
ered.
3.2 Environmental
The wood fungal infection and the changes in the
tell-tales both pointed to temperature and humidity
concerns within the tomb. Liebhart & Johnson,
2005 document the on-going efforts to monitor the
environmental conditions regarding temperature and
humidity using electronic data loggers (arrow, Fig-
ure 12).
From a structural conservation perspective, there
are several environmental issues that must be con-
sidered including 1) what is the appropriate humidi-
ty level that must be maintained in the tomb to con-
serve the timber and prevent structural
deterioration? 2) what is the optimum moisture level
in the earthen dome to prevent desiccation of the
underside of the dome? Continued loss of material
from the underside of the dome will cause it to
weaken.
Recent readings indicate the temperature fluctu-
ates seasonally but are relatively constant at 15.5º C
in summer to 12.5º C in the winter. The absolute
humidity is the preferred measurement for humidity.
It represents the water content of air at a given tem-
perature. In spring, the readings inside the tomb are

0.004 kg/m3 and increase to 0.011 kg/m3. These
humidity readings generally mirror the exterior envi-
ronment outside of the mound.
4 FUTURE WORK FOR STRUCTURALLY
CONSERVING THE MIDAS TOMB AND
MOUND
4.1 Earthen Dome
In 2015, soil samples were taken from the underside
of the earthen dome. Proposed testing includes: par-
ticle size analysis, liquid and plastic limits test, hy-
drometer test, X-ray diffraction (XRD) with identifi-
cation of soil type, and scanning electron
microscope (SEM).
From these results, the characteristics of the
earthen dome material will be identified including
the soil type, optimum moisture content, and struc-
tural properties. A long-term project is proposed to
determine more exactly the size and thickness of the
earthen dome so that a finite element analysis might
be performed.
Meanwhile, efforts are on-going to determine
whether a material could be injected into the gap be-
tween the earthen dome and the concrete structure.
The material would transfer the soil pressure to the
concrete structure to prevent settlement of the
mound.
4.2 Concrete protection system
In 2015, Schmidt hammer tests were performed to
determine the concrete strength at various locations.
Results ranged from 19 Mpa to 28 Mpa whereas the
original design was listed as 14 Mpa. Future testing
will include more Schmidt tests along with taking
and testing core samples to better correlate the
Schmidt tests for greater confidence.
Ferro scans are contemplated to identify the loca-
tion, size and cover for the reinforcement. Some
cover removal will be required where existing rein-
forcement is not visible.
Ultimately, an analysis of the concrete structure
will be performed to verify the load carrying capaci-
ty of the structure and provide information for any
future remedial work.
4.3 Steel bracing
Few efforts will be needed for the conservation of
the steel bracing. Interior tomb bracing needs re-
painting for corrosion protection. The outer bracing
must be maintained similarly.
4.4 Environmental
Monitoring efforts will continue indefinitely. Con-
trolling the temperature and humidity in the tomb
may prove to be impossible and is not yet an option
being considered. A program for identifying new
fungal development is contemplated along with a
protocol for conserving the timber surfaces.
5 CONCLUSIONS
The Midas Tomb chamber complex is one of the
world’s most unique timber structures. Conserva-
tion has been on-going since it was excavated in
1957. Maintaining this monument requires the con-
tinual conservation efforts by individuals with ar-
cheological, structural engineering, geotechnical en-
gineering, material science, and environmental
expertise.
6 REFERENCES
Liebhart, R. 2013. Phrygian Tomb Architecture: Some Obser-
vations on the 50th Anniversary of the Excavations of Tu-
mulus MM. In C. B. Rose, ed., The Archaeology of Phryg-
ian Gordion, Royal City of Midas. Philadelphia: University
of Pennsylvania Press.
Liebhart, R. & Johnson, J. 2005. Support and Conserve: Con-
servation and Environmental Monitoring of the Tomb
Chamber of Tumulus MM. In The Archaeology of Midas
and the Phrygians: Recent Work at Gordion, ed. L. Kealho-
fer, pp. 191-203. Philadelphia, PA: University of Pennsyl-
vania Museum of Archaeology and Anthropology.
Young, R.S. 1957. Unpublished excavation records, MM Ex-
cavation 1957, p.18, Gordion Archive at the University of
Pennsylvania Museum.
Young, R. S. 1981. Three Great Early Tumuli. The Gordion
Excavations Final Reports, Vol. I, ed. by E. L. Kohler.
University of Pennsylvania, Philadelphia: University Mu-
seum.