|
Chapter 2
THE STRUCTURES OF THE DIFFERENT LAYERS.
I. The number of strata and their age.
As
far as I know, there is no place in the world where so many clearly
distinguishable layers of buildings and masses of rubble are preserved
than on the hill of Hissarlik. I know many places where two, three or
even more layers of ruins lie on top of each other and together form a
mass of rubble several meters high, but that the remains of old
buildings and their layers of earth reach a height of 15 meters,
and that in these ruins there are nine or even more chronologically
separated strata can be clearly distinguished, so far only occurs in
Hissarlik.
Quite apart from answering the question of whether
the hill of Hissarlik really once supported Homeric Troy, such a
strange site of ruins clearly merits the careful attention of
archaeologists and prehistorians. Here you can make reliable
observations about the relative and sometimes also the absolute age of
many buildings and countless objects. Here are ruins from a very
distant time, from which we do not know any buildings in Europe. And
yet every visitor to the excavation site can still explore the walls
and floors of the various layers and easily collect numerous potsherds
and other objects that belong to a specific layer and thus a specific
millennium. Walking through the field of ruins, one finds oneself in a
large museum, so to speak. From the different layers of the rubble
walls one can remove clay shards, like from the parts of a museum
cupboard, and then determine from the altitude to which millennium or
even which century they belong. In Hissarlik, for example, it is easy
to carry out comparative studies on the culture and art of several
millennia.
But, one will ask, are the layers really so clearly separated that they are easy to recognize and errors are not possible?
In
fact, those who visit the hill of Hissarlik for the first time will
find it difficult to find their way around the ruins at first. In view
of the large quantity of building remains of the same type, he will
initially hardly be able to distinguish between the different strata
and will have difficulty recognizing the buildings belonging together
from the same period as such. The unsystematic way in which the
excavations were initially carried out makes understanding the ruins
even more difficult. Large, irregular gaps have been dug into the
mound, through which the coherence of the individual strata has been
completely destroyed in many places. The stratum present in one part of
the mound cannot, therefore, be traced to another part at once. In
addition, some layers do not go horizontally through the mound, but are
higher in one part and lower in the other. The connection between such
dismembered layers lying at different heights is, of course, often
difficult to recognize, even for a practiced eye, and can sometimes
only be ascertained at a single point. Fortunately, however, on closer
study, the most important layers are unmistakably recognizable by their
construction, their material and their state of preservation and can
easily be found again in the different parts of the castle.
Whoever
studies the ruins and layers of earth in detail for a long time, or is
led around the field of rubble by an expert, soon gains an overview and
precise knowledge of the various layers and their main structures. From
the confusion of stones, earth and remains of fire, three layers are
initially clearly separated and serve as unmistakable signposts as you
continue to explore the debris field.
At the top one recognizes
the Roman layer (IX) by its large foundation walls of regularly cut
ashlars of soft limestone; then, somewhat lower, the Mycenaean layer
(VI) catches the eye by its beautiful and strong embanked walls of
large stones; and thirdly, even deeper, the prehistoric castle complex
(II layer) is unmistakable, because its wall remains are made of small
stones and, what is particularly characteristic, of half-burned mud
bricks. The simpler intermediate layers can then soon be classified
between these main layers. Admittedly, in many cases only small pieces
of them are now visible, because most of them had to fall victim to the
uncovering of the main layers. On the other hand, large pieces have
been uncovered and preserved of the three main layers: the Roman (IX)
mainly in the eastern half of the castle, the Mycenaean or Homeric (VI)
in the southern and eastern parts, the prehistoric castle (II) in the
middle and in the northwestern part. The intermediate strata are, apart
from smaller places, to be recognized almost exclusively on the
vertical walls of the inner cones and on the still untouched edges of
the mound in section. Only a single piece of the lowest layer (I) has
been uncovered in the large north-south ditch a little west of the
middle of the mound.
This distribution of strata throughout the
hill is best illustrated by looking at our Plate III, in which the
ruins of the various strata are distinguished by colour. The three main
layers are denoted by the three colors blue (IX - Roman), red (VI -
Homeric) and black (II - prehistoric). Anyone who wants to find out
more about the distribution of the individual strata must consult the
individual plans of these settlements on our Plates IV - VII.
If
the individual strata can be distinguished on the spot by studying
them, their areas often overlap in such a way that it is not possible
to say with certainty: one stratum ends here and the other begins at.
An exact separation can only be achieved in those places where a
clearly recognizable floor made of clay or limestone or gravel has been
preserved, or where the old floor height on the walls can still be
determined from the fact that the foundations located under the earth
were built less carefully than the once visible upper walls.
But
even where there is such a sharp dividing line between two strata, it
is often not possible to say with certainty which of the two must be
reckoned with an object coming to light immediately below that line.
If, for example, a shard is found under the floor of a room, it can be
doubtful whether it should be counted among the stratum to which the
floor belongs, or among the preceding, older one. Although in most
cases the latter allocation will be more correct, there are also cases
where such a shard belongs to the upper stratum. For example, when
digging the foundations or laying the floor in the ground; the floor
can also be younger than the house to which it belongs.
Of
course, the allocation is even more uncertain in all those places where
no dividing line between two layers can be seen. One then has to make
the attribution to a particular stratum to the best of one's
conscience, taking into account both the circumstances of the find and
the nature of the object. In this state of affairs it is clear and
should never be forgotten that the finds of two consecutive strata can
never be separated with absolute accuracy: they will always merge. One
would deceive oneself and others if one wanted to claim that certain
objects are found exclusively in a single layer. In such a case one
must always keep open the possibility that they also occur in the
immediately preceding and immediately following settlement.
As
far as the exact number of strata found in Hissarlik and their counting
is concerned, I have essentially to repeat what I said about this in
the book Troy 1893 (p. 86).
The nine layers that we distinguish are counted from below and always
denoted by Latin numerals, i.e. I is the lowest and oldest layer, II is
that above and IX is the top and youngest layer. In the 1st layer, 2
different periods are distinguished, in the 2nd layer even 3 periods. 2
periods are now also counted for the seventh stratum. In each of these
three cases, the existing castle walls and inner buildings were
radically rebuilt, and not completely rebuilt. Since this also involved
raising the floor and expanding the castle in some places, one might
think it more correct to count each period as a separate stratum, and
then instead of the 9 strata there would be 13 altogether. However, the
slight elevation of the terrain and the partial use of the older walls
speak against such a change.
With an increase in the number of
layers, the previous numbering of almost all layers would have had to
be changed. This would obviously cause great confusion in the
publications, diaries and find notes. You had to add to each statement
whether the older or younger census was available. Without such an
addition, confusion and error would easily arise. We have therefore
decided to keep the old numbering that Schliemann first carried out in
the book Ilios, despite a few shortcomings, and only as was done first in the book Troja 1893,
instead of the top layer VII, which in Schliemann all historical finds
to distinguish three special layers. The uppermost, Roman layer
therefore bears the number IX, and layers VII and VIII correspond to
the period from the destruction of Troy to the Roman emperors. This
small alteration, in which the six lower strata remain unchanged, was
necessary because between the Mycenaean and the Roman ruins, in those
places where the conditions for observation were particularly
favorable, two strata were actually found which differ in time and
space from them have to be divorced.
The best way to get a good
overview of the 9 layers, which is easy to remember, is to break them
down into 3 groups, each of which includes a main layer and one or more
sub-layers. We then get the following overview, in which the main
layers are highlighted by blocked print :
1st group:
a. lowest prehistoric settlement, layer I.
b. prehistoric castle, level II.
2nd group:
a. three prehistoric settlements, layers III — V.
b. Mycenaean or Homeric castle, layer VI.
3rd group:
a. pre-Greek and ancient Greek settlements, layers VII—VIII.
b. late Greek and Roman acropolis, layer IX.
While
the relative age of all strata is perfectly established, since each of
them is obviously younger than those below, the absolute age cannot be
determined for all strata.
The top layer (IX) certainly belongs
to the Roman period, as numerous inscriptions found in it prove. It
contains several large buildings, the construction of which, as far as
we know, is mostly attributed to the Roman emperors. We assume that
they were all built after the destruction of Ilion by Fimbria (85 BC).
On the other hand, since almost no Byzantine remains have been found at
the castle, we can, especially since this is only an estimate, assume
the existence of the IX layer in the period from the birth of Christ to
around 500 AD.
Level VIII, of which only a few buildings, but
several individual walls and numerous other objects (such as potsherds,
clay reliefs, inscriptions, building elements) have been found,
corresponds to the polis or Komopolis Ilion mentioned by Strabo with
the temple of Athena, which Xerxes and Alexander the Great visited. It is the
small town of Ilion, whose houses Demetrios of Skepsis saw in the 2nd
century BC still without roof tiles, i.e. with clay roofs (Strabo IX,
594). Taking into account the age of layer VII, we can assign this
layer to roughly the time from the birth of Christ to around 700 BC.
At
layer VII we distinguish two periods. At the top is a village, of which
the walls of the houses are still preserved in several places on the
hill, walls that differ from all the others by a special type of
construction, namely the use of high-edged slabs. Among them are houses
from an older period, which are mostly used again by the younger
residents. It is the magazines next to the castle wall that we used to
describe (e.g. in Plan IX of the Athenian Communications
of 1894) of the VI had been assigned to the 7th shift, but had to count
more correctly on the 7th. They have nothing to do with the Homeric
Castle VI, but are certainly only after the destruction of the stately
interior buildings of the VI Layer built over their ruins. The ancient
potsherds found in the houses of both periods allow us to assume that
the VII layer is older than the 7th century. How high it reaches is to
some extent determined by the age of the next lower layer (VI). Since
this must be placed in the second millennium BC, the period from 700 to
1000 BC remains for layer VII, if we again use round numbers.
The
important castle of the VI layer, with its strong curtain wall and its
large inner buildings, is dated by the Mycenaean vases found within its
buildings, together with the native monochrome pottery. Because these
vases are generally attributed to about the second half of the 2nd
millennium BC. If we hereafter the period of the VI. If you limit the
strata by the round numbers 1000 - 1500 BC, we shall probably come
pretty close to the truth.
The dating of the older strata (V —
I) is more difficult because we lack any reliable basis for determining
their age. The various buildings and objects found cannot be dated
absolutely, nor can we determine the duration of the individual layers
with any degree of certainty. Here we must content ourselves with
estimates. In the book Troy 1893
I have assumed the same rounded periods of 500 years each for the
prehistoric villages (layers V - III), for the 3 periods of the
prehistoric castle (layer II) and for the ancient I layer by pointed
out the fact that each of the upper strata also averaged such a period.
As far as I can see, this stipulation has hitherto been generally
accepted and may therefore be retained. But I don't want to omit to
expressly point out again the uncertainty of this dating. Not only by
centuries, but even by thousands of years our approach can deviate from
reality. With this reservation we place the 3 layers of poor
settlements (V - III) in the period 1500 - 2000 BC, the prehistoric
castle of the 2nd layer with its two periods in the time of 2000 - 2500
and the lowest 1st layer with its two periods in the time from 2500 -
3000 BC.
Finally, the various strata and their duration may be summarized in a table for a more convenient overview.
Strata : Settlement : Approximate period :
I ancient settlement perhaps 3000-2500 BC
II prehistoric castle perhaps 2500 - 2000 BC
III - V 3 prehistoric villages perhaps 2000 - 1500 BC
VI Homeric Troy about 1500 - 1000 BC
VII 2 pre-Greek settlements ca.1000 — 700 BC
VIII Greek Ilion ca. 700 — 0 BC
IX Acropolis of Roman Ilion ca. 0 — 500 AD
The
precise location of the various strata and their extent across the
mound will be discussed in more detail when discussing the structures
of the individual strata. Here, for the reader's orientation, a general
picture of their position is to be given, following a schematic section
through the hill.
The accompanying fig.6 shows this section, in
which, for the sake of greater clarity, the dimensions of length are
drawn on a smaller scale than the dimensions of height. On the same
scale, the latter would appear too small. The differences in height are
thus more conspicuous than on the great section through the hill (Plate
VIII), in which the lengths and heights are given on the same scale.
Fig.6: Schematic section through Hissarlik hill showing the nine different strata.
The
natural rock is highlighted by a special hatching and forms the basis
for the 9 layers lying on top of each other. Its course has only been
determined at a few points and therefore not yet secured in all
details. So we do not know at what depth the ground rock lies in the
Simoeis plain drawn on the left edge of the picture.
The
walls of the 1st layer are built on the rock or just a little above it;
marked on average by light hatching and denoted by I. After this
ancient settlement was destroyed, some walls were built before the
construction of the 2nd castle, which we consider to be the 2nd period
of the 1st layer. They are not all indicated in our drawing, but will
be discussed later.
Photo 5: The walls of Stratum I (p.32).
The second layer was built over the rubble
of primitive settlement I as a castle with dwellings inside and a
strong fortress wall outside. Its level is almost 5m above the rock and
on average forms a horizontal line in the middle of the hill. The 3
periods that we distinguish in this layer can only be recognized in the
drawing by the 3 castle walls, which are drawn side by side on the
southern (right) side and correspond to the extensions of the castle.
On the left side, where the northern slope dropped low to the Simoeis
valley, no extension was possible; the curtain wall here seems to have
remained in the same place during all three periods. However, their
exact location is not known and can therefore only be given as a guess.
Anything that survived from the wall was destroyed during Schliemann's
first excavations.
All
buildings of the 2nd level were destroyed by fire, only the foundations
and a few small pieces of the walls survived. "Burnt City" was what
Schliemann called these ruins, which showed the traces of intense fire
everywhere. After their destruction, they formed a large scree hill on
which the settlers of the III class settled and built their huts. In
our drawing the simple walls of the III layer are not drawn; the layer
is indicated only by a line and the number III. The same is true of
layers IV and V; its walls were not shown either, so that the other,
more important layers stand out all the better in the drawing.
Due
to the repeated destruction and new buildings, the originally low hill
had already become considerably higher in prehistoric times. In the
middle the rubble and debris had already reached a height of 100 m
above the rock. The general shape of the hill had also changed as a
result: the castle, which was horizontal at the top, had gradually
become a hill sloping from the middle to all sides.
The builders
of the VI layer found such a place and erected a strong castle rising
in several concentric terraces on it. That was Priam's fortress,
Homer's Troy. In our section, the walls of the VI stratum are drawn
entirely in black, and their floor is also indicated as a thick black
line, in order to be easier to see. On the right we see the southern
ring wall, which reaches down to the rock. To the left is a wide
gallery and then a higher building, the retaining wall of which is
similar in cross-section to a castle wall. It is building VI M of our
plan III, which is drawn in section here. Further to the left was a
second, even higher terrace, of which only part of the retaining wall
has been found. It is no longer possible to determine the height at
which the middle of the VI castle was located, because the entire
middle part, as will be described in more detail later, was removed in
Roman times in order to create a horizontal surface for the district of
Athena. The northern castle wall VI is drawn on the left edge, although
it is no longer preserved in the middle of the hill. I added them
according to the dimensions and the height of the large tower VI g.
preserved on the north-east corner of the hill. The fact that the inner
buildings in the northern part of the castle also rise again in
terraces towards the middle is ensured by the two buildings VI P and VI
Q, which have been preserved to a certain extent and have only been
partially uncovered. What dimensions these terraces had in the middle
of the north side of the hill is not known; the drawing is only
intended to give a schematic picture that is easier to understand than
reality.
After the destruction of the VI castle, simple dwelling
houses and probably also a temple of Athena were built twice in a row
on top of the enlarged mound of rubble, they belong to layers VII and
VIII. The walls of the houses are on the edge of the hill above the
lowest terraces of the destroyed ones Castle VI preserved. In the
middle they were undoubtedly also once there, but they fell victim to
the great reconstruction of the castle by the Romans and could
therefore not be found at all. The Greek temple visited by Alexander
the Great will also have stood in the middle. Remains of it have not
yet been found. In the drawing, the two layers are again only indicated
by a line and the numbers VII and VIII.
The last layer belongs
to the Roman period. The mound was leveled, at least in its eastern
half, by removing the higher center, and a large, almost horizontal,
sacred precinct of the Ilian Athena was built with a temple, columnar
avenues, and a propylaion. A new ring wall also surrounded the entire
hill, which now formed the acropolis of the great Roman city of Ilion.
Due to the height of the mass of rubble, the ring wall does not reach
down to the rock, but, as the pieces that have been preserved show,
stands partly on the rubble covering the slope and partly on older
walls. I have marked the floor of the sacred precinct with a double
line and at the same time the height to which layers VI-VII were
removed at that time. In the west half of the castle, where other
public buildings may have stood, the ground seems to have been slightly
higher. Of course, no more details can be given about this, because
Schliemann destroyed almost all the buildings there when he built his
large ditch.
Insofar as the older strata had been removed in
Roman times, they are only indicated on average by dotted lines. To
what height they reached is not known; the highest point may have been
about 15m above the rock.
The stately buildings of the Romans
fell victim to destruction like all the earlier buildings. From the
Byzantine period onwards, the large stones were used in the surrounding
villages, partly to build houses, partly as grave monuments in the
cemeteries. Hundreds of columns and pieces of entablature still lie
around there. On the Acropolis itself, only the foundations of the
Roman buildings remained and were gradually covered with rubble and
earth so high that fields could be planted again.
2. The building materials. [p.35]
In
order to be able to better adjust the structures of the individual
layers and to be able to distinguish them from one another, it is
useful to take a look at the various building materials used in them.
It is true that in Troy, unlike in other places, completely different
building materials were used in the individual development periods or
strata, which would allow us to deduce the time of its construction
from the type of stone of a building without further ado. Because in
almost all layers, in addition to loamy earth, almost exclusively the
local porous limestone was used. But nevertheless the structures of
most layers show so many peculiarities that it is possible to recognize
them individually and easily distinguish them from those of the other
layers. For one thing, the different strata do not use the same types
of porous limestone; Differences in terms of the degree of hardness and
the structure of the stones are unmistakable. Furthermore, the size of
the built stones is very different. There are also particularly large
differences in the processing of the individual stones and their
assembly. Finally, other types of stone, such as marble, syenite and
sandstone, also occur in some strata.
Three types of building
materials are indisputably primarily used at Troy: porous limestone,
adobe, and wood. They coexist in almost all layers, apparently because
they are the only materials extracted on site. The long ridge, at the
end of which lies Troy's Castle, consists of almost horizontal layers
of tertiary limestone, which is easy to break up and is still used
today as building blocks for houses in the surrounding villages and for
the fortifications of the Dardanelles. The individual layers are of
different degrees of hardness. There are hard stones that are very
weather-resistant and are therefore suitable for all walls, but there
are also soft stones that could only be used for foundations and
similar invisible walls.
In addition, the two river valleys to
the north and west, the plains of the Simoeis and the Skamander, offer
good clay for the production of unfired and fired bricks today, just as
they did in antiquity. The latter were only made in Roman times. In all
other periods of Ilion's long history, the sun- or air-dried mudbricks
were used as widely as they are still in some villages of the
countryside today. Finally, the mountains of the upper Scamander
valley, especially the Ida itself, undoubtedly provided plenty of good
timber in the Alterturae, since even today the timber that came from
the Ida mountains and was rafted down the Scamander is an important
article of trade for the Troad.
The
local limestone was used in two ways in building walls: either
irregular pieces of larger or smaller dimensions were taken and the
so-called quarry stone masonry was made with clay or lime mortar, or
larger, regularly cut blocks were joined together without mortar. The
former type is found chiefly in layers I — V and VII — VIII, the
buildings of II being distinguished in part by the larger dimensions of
the stones; the latter type has been used preferentially for layers VI
and IX. Some walls of VI have the stones worked almost all around,
others show only the bearing surfaces and the outsides worked, while
still others are left completely unworked and rough. A very peculiar
type of stone masonry occurs in the upper part of the castle wall of
VI, the stones being cut to the size and shape of bricks. Apparently
brick masonry is imitated here in more weather-resistant material.
In
stratum IX only the foundations and supporting walls are made of
quadrangular blocks of limestone, while marble was mostly used for the
superstructure of the porticoes and temples. This precious material,
which had to be fetched from great distances, has now disappeared in
Hissarlik except for a few pieces, some of it has migrated to the
cemeteries of the surrounding villages, some to lime kilns. Small
stones of all kinds and bricks with lime mortar were used in the houses
of the IX layer. While the hard types of tertiary limestone are
preferred in the other strata, so soft a material is often used for the
foundations of IX that it unfortunately soon decomposes in the open
air; its good preservation up to now can only be ascribed to the fact
that the walls built with it have always been underground.
We
find unfired bricks mainly in layers II — VI, in the former in such
large quantities that Castle II could justifiably be called a "brick
castle" or "clay castle". In layers III - V, mudbricks only occur in
individual pieces of wall; in general, the walls of these village-like
settlements consist of small rubble stones with mud mortar. In the
sixth layer, the Homeric castle, only a few mud bricks were found in
the rubble covering the ruins during the excavations; they must have
belonged to the destroyed upper walls. It was only afterwards that I
noticed that the superstructure of the north-east tin also consisted of
unfired bricks, some of which are still preserved. In contrast to the
second layer, Troy VI was a "stone castle" because the mud bricks were
only used occasionally for their superstructures.
Two different
types of unfired bricks can be distinguished: one has chopped straw
(chaff) added to the clay from which the bricks are formed, the other
contains only clay without additives. That species occurs almost
exclusively in Troy; I have found only a few bricks in which straw
seems to be entirely wanting. The clay is mostly unstrained, not only
small stones and shells, but even pot sherds and larger stones are found in it. The clay in the bricks of the castle walls is particularly impure.
The
dimensions of the bricks differ greatly not only in the different
layers, but also in one and the same layer. So there are no standard
dimensions prescribed or customary. In general, the bricks are
much larger than the Roman bricks, which are usually only 0.29 m (1
ft.) wide and long and only about 3-5 cm thick. They are more like the
Greek bricks described by Vitruvius and also in ancient walls,
e.g. in the enclosing wall of Eleusis, have actually been found.
The table below gives an overview of the mass of bricks observed in the
buildings of the various strata.
Structure Mass of clay bricks and layer Length Width Height (p.36)
1st building II A 0.66 o,tg 0.45 0.46 0.11-0.13
2nd building II B 0.70-0.72 0.46 0, 48 0.10 0.11
3. The same 0.69-0.71 0.20-0.22 0.10-0.11
4. Castle wall II 0.65 0.32 0.12
5. The same in a different place . 0.45 0.22 • 0.23 0.12
6th house in III 0.52 0.43 0.13
7th building in VI ... . 0.57 0.29 0.10
8. Other building of VI. 0.40 0.41 0.26 0.11
9. Tower of VI. 0.38-0.40 0.26 0.27 0.05-0.06
We
can simplify this colorful series of numbers and make them clearer if
we try to determine the intended formats of the bricks and the ancient
masses on which they are based. As is well known, the bricks in
antiquity mostly had round dimensions of cubits or subdivisions of the
cubit. If we now try to determine the size of the ancient linear
dimensions from the reported brick masses, namely the hand width, the
span, the foot and the cubit, this seems to be because of the
extraordinary variation. At first glance, it is quite impossible for
the numbers in the table to be distinguished. On closer inspection,
however, and considering only average dimensions, we soon find that the
dimensions of most bricks are simply related.
For example, in
the case of the bricks N 1 and 2, the length is evidently equal to the
width and this in turn four times the height. Expressed in numbers,
these bricks have the ratio 6:4: 1. If the height were a hand's breadth
, the breadth would be 4 handbreadths or 1 foot and the length 6
handbreadths or 1 cubit. In reality the height of 0.11 - 0.12m is
evidently greater than a handbreadth, for the handbreadth of 4 fingers
in man is and Correspondingly also with the Greek measures of length
averaging about 0.08'' in size. The height therefore corresponds to
about i '/^ hand breadths or 6 fingers. We thus obtained for the bricks
N" i and 2 a ratio of 9:6:1 '/,, hand widths. The length of these
bricks of 9 hands measures 1 '/., cubit, their width of 6 hands is one
cubit With the bricks N" 5 from the castle wall II, the simple ratio
6:3:1 1/2 hand widths or 1 cubit : 1/2 cubit : 1/2 cubit is evident.
From
this we obtain a value for the size of the old cubit that fluctuates
between 0.45 and 0.48 m. The difference of about 0.03m between the two
numbers could be explained by the fact that the bricks dry out,
because, as is well known, their circumference changes not only when
they are fired, but also when they dry in the air. But such a reduction
in size would evidently have to be essentially the same for all bricks.
Since this is not the case, however, we have no right to
declare one of the values between 0.45 and 0.48 m for the real length
of the Trojan cubit. On the contrary, it is quite conceivable that the
bricks do not have a specific standard measure of the cubit, and that
they are only made according to a natural human measure. The maker of
the bricks could make a wooden form by his own hands, 3 spans long, 2
spans wide, and 1/2 span thick, and did not need to check whether this
span really corresponded to the span or half a cubit in use. Besides,
it is not impossible that in the oldest times there was no fixed
standard at all in Troy.
If you still want to deduce a Trojan
cubit measure from our brick masses, then only the largest value, i.e.
0.48m, may be assumed as the length of the cubit, because the smaller
values can be better explained by the shrinkage of the bricks during
drying than the larger values with a smaller cubit mass. A cubit of
0.48m would then correspond to a foot of 0.32m and a hand of 0.08m. It
may be mentioned that an ancient foot of this size actually exists at
the stadium uncovered at Olympia, that the Graeco-Aeginean foot
(0.328111) is only slightly larger, and finally that a foot of 0.32 m
accepted by some scholars for Asia Minor and Babylon (p39). But the
foot of 0.32 m and the corresponding cubit of 0.48 m for Troy have not
yet been proven with certainty. On the contrary, since the dimensions
of several buildings lead to a completely different measure, namely to
a cubit of about 0.521", we must leave the question of the size of the
Trojan cubit undecided for the time being. The only thing that is clear
is that the mass of the bricks is best fit a cubit of about 0.48 m.
On
the other hand, it is worth examining the formats of the bricks a
little more closely and comparing them with the brick shapes otherwise
known from antiquity. We may proceed from the observation that bricks,
both today and in antiquity, are formed in such a way that their
length, width and height are in a simple relationship to one another in
order to achieve a good bond. Such proportions can now also be
recognized without difficulty in all our bricks. In fact, without
altering the numbers significantly, we find the following simple
proportions:
N 1 = 6: 4: 1
2 = 6 : 4 : 1
3 = 6 : 2 : 1
4 = 6 : 3 : 1
5 = 4 : 2 : 1
6 = 5 : 4 : 1 (?)
7 = 6 : 3 : 1
8 = 3 : 2 : 1
9 = 3 : 2 : 1/2
How do these proportions relate to those used in ancient Greek bricks?
According
to Vitruvius (II, 3, 3) three types of bricks were distinguished in
Greece, namely rectangular ones 6 hands (= 1 cubit) long and 4 hands (=
1 foot) wide, which were called Lydian, and 2 types of square bricks,
of which the larger were 5 hands wide and long, and the smaller 4.
According to Vitruvius, the larger square bricks were used for public
buildings, the smaller ones for private buildings. Corresponding
half-bricks belonged to both types of square bricks, which were
necessary for the production of a regular association.
If we
compare the Trojan brick formats with this, it turns out that the large
bricks of buildings II A and II B (our No. 1 and 2) may be called
Lydian because of their shape. They are larger than the Vitruvian ones,
but they are the same proportions in that they are also 1 1/2 times as
long as they are wide. (p.40) We find the corresponding half-bricks in
the bricks No. 3. Bricks of the VI layer (No. 8 and 9) can also be
counted among the Lydians. All bricks of this format offer the
advantage that they can be put together to form a good bond even
without half bricks if the wall is at least 2 bricks wide.Thus, in
building II A the bond is achieved by placing two bricks of 1 1/2
each in one layer 2 cubits and in the other 3 stones of 1 cubit each
make up the wall thickness. In building II B, on the other hand,
half-bricks are found; the whole stones are all along the length of the
wall, and the necessary change of joints is produced by the fact that
the two whole and the half-bricks forming the thickness of the wall,
the latter alternating with the one and the other outside is arranged.
Square
bricks, as they were mainly common both according to Vitruvius and the
surviving buildings in Greece, do not appear in our list at all, but we
find the ratio 2 : 1 in mottled bricks from different layers and can
recognize half-square bricks in them. In the case of the castle wall of
the second layer I have noted two different sizes of this format,
namely No 4 with 0.65 and 0.32 m in old measurements seems to
correspond to 2 and 1 feet, while No 5 with 0.45 and 0. 23111 can
probably be equated with 1 ell and 1/2 ell. No 7 is also half a square
brick. A good bond was very easy to achieve with the bricks of this
format. That is why our modern bricks also have this ratio between
length and width, but with smaller dimensions (0.25 : 0.13m) and taking
the joint width into account. The same format was also common in the
bricks of ancient Egyptian buildings.
Among the Trojan bricks on
our list is only one different format, namely No 6, in which the length
is related to the width as 5:4 or 6:5. As such a format is quite
unusual and impractical, I suspect that E. Burnouf, who measured this
brick (see Ilios, p. 355) had before him an incomplete brick No 1. I myself have not seen bricks of format No. 6.
As
far as the height of the bricks is concerned, there is great agreement
here. Most bricks are 0.10 to 0.131 m, i.e. on average 0.1151 m high,
so according to ancient measurements they should obviously be 1/2 span
or 1 1/2 hand or measure 6 fingers. Only in the case of the bricks that
are still in the superstructure of Tower VI g and in the case of a few
pieces of brick of unknown origin have I observed half the height (0.06
m = 3 fingers).
In addition to limestone and adobe, another
important Trojan building material is wood. It is not only used for the
roofs and ceilings of the buildings of all layers, but was also used as
longitudinal and transverse beams within the walls, especially in the
second layer. Strong timbers were walled in the walls of brick or
stone, parallel to the facade and perpendicular to it, and sometimes
formed a full scaffolding which gave greater strength to the walls.
Also to protect the free-standing corners of the wall, strong vertical
wooden posts were used as paraposts or ants in the second layer.
Plate 6: T (p.40)
The
covering material for the roofs of the buildings seems to have been
primarily (p.41) earth, spread out in a thick layer almost horizontal
on the surface over strong wooden beams and overlying reeds. In the
case of building II A, this type of coverage is positively secured by
the find circumstances, because here a layer of earth with remains of
burned wooden beams was found inside above the burned floor and below
the ruins of the collapsed burned brick walls, which can hardly be
anything else. than the earth of the horizontal roof that collapsed
when the house fell. Even today the horizontal earthen roof is almost
universal in the villages of the Troas, a sure testimony to its
practicality and its equity.
We do not know whether some
structures were covered with reeds or wooden shingles or similar
materials without a layer of earth. The possibility of such a roof
formation cannot be denied for Troy, especially since it was certainly
in use in other parts of Asia Minor. In this case, of course, the roofs
must not have been horizontal, but on the contrary must have had a very
great slope, even a greater one
than the later Greek tiled roofs.
These latter, with bricks of fired clay or marble, certainly did not
come into use in Troy until Roman times. Because firstly no roof tiles
were found in the older layers, which would be inexplicable if even a
single building had been covered with them, and secondly we know from
Strabo that the village of Ilion was known by skepticism at the time of
Demetrius (i.e. in the II. Century BC) was without tiled roofs.
However, this statement by Demetrius (in Strabo XIII, 594: xcxxx xxxxx
xxxxx xxxx) was often understood to mean that the houses at that time
had no roofs, that is, they were destroyed; but apparently Strabo and
Demetrios meant something else, they rightly considered it a sign of
low culture and great poverty that the houses had no tiled roofs.
The
earthen roofs, which I believe must also be assumed for the large
buildings of the II and VI layers, were not completely horizontal, but
had to have a slight slope from the center to all sides so that the
rainwater could run off. If the roof were not sloping, the water would
not only sink into the layer of earth, but also penetrate into the
interior of the house if it rained for a longer period of time. On the
other hand, the layer of soil must not slope too much, otherwise the
soil would be swept away by the rainwater. A steeply sloping earthen
roof, such as F. v. Reber assumes for the Megaron of Tiryns (see Abhandl. der Bavarian Akad. der Wiss. XXI, p. 506), I
therefore consider inadmissible. Adequate slope exists when the
surface of the roof is made 0.20 to 0.30m higher in the center than at
the edges.
(p.42)
It is well known that the layer of earth must not be laid directly on
the strong beams of the ceiling, but that a layer of reeds or similar
material must be placed between them. For Troy the use of reeds is
proved by the fact that several pieces of burnt earth with the remains
of burned reeds have been found in the II stratum. When the houses
caught fire, the roof beams and the overlying reeds had burned away and
had turned the earth above into terracotta, in which the holes in the
reeds remained visible.
Fired clay and marble roof tiles are
found only in the top layer. There they are found in large quantities
and in various forms. As far as my observations go, there are two types
of clay bricks: first, flat, curved base bricks, the joints of which
were covered with similarly shaped cover bricks; their full masses are
not known because only fragments have been found. I noticed a red coat
of varnish on some bricks of this type, but mostly they are
unvarnished. Secondly, there are flat, square bricks, with the rims
turned at right angles on both sides, and a small round rod at the top.
The next brick reached over the latter, which prevented the water from
receding. These flat tiles include roof tiles of various shapes. In the
case of a flat tile, the original size has been ascertained; it was
0.545 m wide and 0.810 m long. This second type of roof tile is no
longer in use in Greece today. In antiquity it was used in particular
for temples and public buildings and is therefore probably also used in
Roman Ilion for the roofs of temples, columned halls and other public
buildings.
3. The first layer, the oldest settlement. [p.42]
Only a small piece of the lowest layer on the hill of Hissarlik
is revealed in squares D 2 to D 6. The first walls of this ancient
layer were found in 1872 when Schliemann dug the large north-south
ditch across the hill. After removing all the upper strata and cutting
through the great brick buildings of the "burned city," he came upon a
few small parallel walls immediately above the solid primeval soil,
which soon turned out to be the scanty remains of the oldest
settlement. As buildings of very old times, at least the third
millennium BC, they may claim our full interest in spite of their
simplicity. They have been carefully examined several times by
prehistorians.
 In 1879 Schliemann had a larger piece of the
first layer excavated in the presence of R. Virchow to give him the
opportunity to study the utensils, kitchen leftovers and all kinds of
remains from the first settlement. Also in the later excavations of
1882, 1890, 1893 (p.43) and 1894 we always uncovered a small piece of
this ancient layer and thus collected quite rich material for
determining the cultural status of the first inhabitants. During these
excavations, some walls came to light that were younger than layer I,
but not yet belonging to layer II. We must therefore also distinguish
between two periods within the I layer. The finds made in the buildings
will be reported in Chapters IIb and IV; only the buildings will be
described here.
Fig.7: The walls of Layer I exposed in the great north-south trench. (p.43)
The
number and extent of the uncovered walls increased only slightly as a
result of the later excavations; even now it is so small that the
ground plan of the entire complex is not entirely recognizable. In
order to understand the plan of the walls and to determine the purpose
of the rooms, a larger part of the settlement would have to be
uncovered. However, since this is only possible by destroying another
part of the higher layer, the prehistoric castle II, we have not yet
been able to decide on this in view of the great importance of the
latter. Should the ruins of the second layer, which are not very
permanent, be damaged or even destroyed later by the effects of the
weather, the excavations can be resumed and a larger piece of the
lowest layer examined.
What
is known up to now of the first layer is shown in the floor plan
communicated in fig.7 and the main plan of all layers published on
Plate III. The walls are also drawn in the section on Plate VIII and
can be seen directly above the rock. In fig.7 all the walls of the
first stratum are partly emphasized by solid black drawing, partly
cross-hatched. Unfortunately, almost all walls are in such a sad
condition (p.44) that their thickness and sometimes even their
direction can no longer be precisely determined. Even during the
excavation they were preserved only in their lower parts up to 1m high;
since then they have been severely damaged and some have fallen into
disrepair, which was unavoidable given their simple design. Therefore,
when preparing the plan, I was not only allowed to stick to the walls
that are still recognizable, but also had to consult the older plans
and drawings.
Most of the walls have an east-west direction, and
disappear into the earthen walls of the great north-south ditch, the
slopes of which are indicated by simple dashes in the accompanying plan
(fig.7). The extent of the walls to the west was still unknown when the
plan included in the book Ilios
was being drawn up (see our fig.4 on page 11, Chapter 1) and is only
known through the excavations of 1882 and 1890, at least for some walls
(d, e, g and k) have been identified. These are connected at the
western edge of the ditch by a transverse wall f, the northern end of
which is not yet known. We do not know how far the walls extend to the
east. Only one thing we can say with certainty, that they could not
have extended beyond the eastern boundary wall of Layer II. Because by
analogy with the upper layers, we can safely assume that the extent of
the lowest settlement was smaller than that of the II layer. In
addition, nothing of the first layer was actually found during the
excavations in the eastern part of the castle hill. The small
transverse wall h no longer exists; but since it is drawn in the older
plans, I have included it in the plan, although there are doubts as to
its former existence.
The different thicknesses of the
walls, which the observer immediately notices, is certainly due to the
fact that some of the inner walls are thinner and some of the outer
walls are thicker. Walls a, c and m may be regarded as the latter; they
are characterized as outer walls by their slope, which is also
indicated in the ground plan, and as fortress walls by their strength.
If the walls a and c in our plans, in contrast to the earlier drawings,
encircle at their western end, then this was drawn as a result of a
corresponding observation on the remains of the wall, which are
admittedly very damaged. Such a course of the walls is also probable in
and of itself.
The initially strange phenomenon that there are
two outer walls (a and c) on the south side can be explained by a later
expansion of the settlement to the south. First c, later a was the
boundary wall. A double period is thus established for the 1st layer,
just as we shall also learn about three southern enclosing walls for
the 2nd layer and thus a repeated expansion. The northern border wall,
marked with the letter m, (p.45) belongs to the more recent enclosure
of the Ist layer, because, as can be seen in the section on Plate VIII,
it stands only on the rubble covering the oldest walls. In earlier
descriptions the wall was erroneously attributed to the oldest period
of layer I. Whether an older border wall stood at the same place
remains undecided.
The wall s south of a, of which only a very
small, heavily sloping part came to light during the excavation in D 5
in one of the rooms of layer II, must be an even more recent extension
of the I layer or a first ring wall of the II strata. A specific
decision does not seem possible to me for the time being. However, it
must be recognized that it may not be a real defensive wall at all, but
only a reinforcement or paving of the hill. In fig.7 I have
cross-hatched it in order to be able to distinguish it at first sight
from the older walls a and c.
The sections of wall n, o, p, q,
and r uncovered on the northern slope of the hill have received the
same type of hatching, although they hardly date from the same period
as s. The deviating drawing is only intended to indicate that its
affiliation with the walls of the first period is doubtful. Their exact
age and meaning are unknown because too little of them has been
uncovered. Given their altitude, they certainly belong to the first
class.
 On the other hand, wall i must certainly be counted as
part of the second or younger period, because it goes beyond f, g and h
and can therefore only have been erected after the latter had been
destroyed. Their design and altitude can be clearly seen in the
photographic illustration in fig.8. Above you can see large stones of a
wall of the II level, below our wall i of the I level and still lower
the wall f, which certainly belongs to the first period of layer I.
Fig.8:
Walls of the two periods of layer I: a=lst Period, d = 2nd
period of layer I, b = layer II; c and e =
rubble. (p.46)
I
cannot say anything more about the very narrow wall b between a and c.
I have only seen a few stones of her before; now she is completely
destroyed. It probably belonged to the younger ring wall a. In 1882 I
estimated the thickness of the latter to be about 3m and that of the
older ring wall c to be about 2.50m, based on the remains preserved at
the time; now they can no longer be measured because they are badly
damaged and moreover covered with earth again. In the section on Plate
VIII, both curtain walls could only be drawn with dotted lines, for the
same reasons. But it has earlier been seen with certainty that both
increased in thickness downwards, and that their southern extremities
were cambered. However, the size of the original embankment cannot be
specified, since both walls have been severely disfigured over the
centuries under the weight of the upper layers.
Walls d to L are
considerably thinner than a and c and therefore certainly formed the
inner walls of the settlement. Its thickness varies between (p.46) 0.60
and 0.90 m, so it fits very well with ordinary house walls. They will
all have belonged to a building of several rooms which formed the main
building of the I branch. The appearance of the walls in 1893 is
illustrated in the picture on Photo 5 (p. 32), which is based on a
photograph taken to the south. The debris and foundations of the second
layer can be seen on the right and left, while in the middle several
parallel inner walls of the oldest settlement catch the eye.
Fig.9:
Wall of Level I with stones laid at an angle (p.47).
All walls of the 1st layer are built of small stones with
earthen mortar. The flat, almost unworked pieces of stone are on
average 0.20 to 0.30m in size and show the construction of ordinary
quarry stone masonry, as was common in ancient times and still is
today.
Only wall g shows a different construction. Its
individual stones are not laid horizontally but at an angle; namely
they are inclined to the left in one layer and to the right in the
next. This resulted in a masonry which can be clearly seen in the
photographic illustration in fig.9 and which is evidently similar to
the Roman opus reticulatum.
That such a peculiar construction occurs in the oldest stratum of Troy
voi is striking and very noteworthy. It hardly repeats itself in the
younger layers, at least I know of a wall of a similar kind only from
layer V (in square G 5 in front of building VI C).
The two walls
d and f (Figure 7) are slightly slanted inwards and correspondingly
sloping outwards. One might try to trace this back to a later
subsidence or inclination of the wall under the load of the upper
strata of the earth, but a former weak embankment seemed more probable
on the spot. A small embankment is understandable for the outer wall of
the inner building.
Since no other transverse walls have been
found apart from f and h, the floor plan of the building must have
consisted of long, narrow rooms. (p.48) Whether these served as
dwellings for people or as stables for animals or as open courtyards
does not appear to be determinable at first. One would most like to opt
for dwellings, and this is confirmed by an observation made immediately
after the excavation. Schliemann mentions (see Ilios,
p. 241) a well-smoothed clay plaster, which could be noticed on some of
the walls; in 1882 only remnants of this clay plaster were preserved on
some of the inner walls at the bottom. In my opinion, such plastering
of the interior clearly indicates inhabitation, for in a stable or
courtyard the walls would hardly have been smoothed over by wall
plaster. The numerous tools and other objects of everyday life also go
very well with an apartment, as well as the remains of meals that were
found between the walls (cf. R. Virchow, Proceedings. of the Berlin Anthropol. Society 1890, p. 338) .
In
determining the purpose of the complex, the fact must also be taken
into account that no door opening has been observed in any of the
surviving sections of wall. It would, however, be incorrect to conclude
from this that there were no doors at all between the rooms. Because on
the one hand there may still be doors in the parts of the walls that
have not yet been excavated, which would come to light during further
excavation, and on the other hand there may have been door openings in
the uncovered walls, which were initially overlooked due to the poor
state of preservation of the masonry and are now no longer visible are
recognizable. Another possibility, namely that the doors were situated
higher above the ground than the walls have been preserved, is ruled
out by the fact, already mentioned, that the wall plaster reaches down
to the foot of the walls. Until the contrary is established, I take it
for granted that the various rooms had doors, and think it most
probable that these are still hidden underground further to the east.
With
the low expansion of the walls known to us, belonging to the first
settlement, only an imperfect image of the shape and the purpose of the
entire construction system can be designed. We see an apartment
consisting of several elongated rooms, which was surrounded by a
defensive wall. There was probably an open courtyard between the outer
wall and the apartment, which could be used as a lounge for the cattle
in the emergency trap. By setting up a new ring wall and conversion of
the apartment, the settlement has certainly been expanded once, perhaps
even twice.
How the walls of the first layer have decreased has
not been determined. It cannot have happened in a large general fire
because they were only very slight fire remains on the walls and
between them (p.49) and could not be compared with the strong burnt
residues that were found in stratum II. Only small remnants of
charcoals came to light in layer I. In terms of quantity they
occur as very small pieces in the layers of earth and fill in the space
between the walls. The settlement must therefore either have been taken
by the enemy and destroyed without great fire, or it was left for any
other reason one may imagine, and then gradually fell into complete
decay.
[Continue to Chapter 2a, part 2]
[Return to table of contents]
|
|