V. PETROGRAPHIC CHARACTER OF CLASSIC MARBLES
Francis J. Turner
Department of Geology and Geophysics
University of California, Berkeley
Petrography.
Rocks are aggregates of mineral grains. Ideally each grain is a crystal
or crystal fragment with a specific chemical composition - quartz SiO2, cal-
cite CaCO3, potash feldspar KAlSi308 and so on.   Standard petrographic pro-
cedure consists in microscopic examination of thin sections of rocks - about
1 cm in diameter, 0.02-0.03 mm thick, mounted on glass. Examination is
carried out in polarized light, for this permits determination of optical
properties related to crystal symmetry and peculiar to each mineral species.
By this means two sets of petrographic characteristics of any rocks emerge;
and both give valuable information as to the origin and history of the rock:
(1) The mineralogical identity of the component grains, the principal
mineral constituents of most rocks being only half-a-dozen in number.
(2)  Their mutual geometric relations which define the rock texture.
The respective textural characteristics of volcanic rocks that have crystal-
lized from melts, of sandstones laid down in water, and of metamorphic rocks
that have recrystallized in the solid at high temperature and pressure, can
be distinguished as such almost at a glance.
Nature of Marble.
Marbles are mineralogically simple rocks formed by metamorphic recrystal-
lization of limestones (calcareous sediments, mostly originating as beds of
shell and other organic debris laid down in clear marine waters). The main
constituent is calcite, the common crystalline form of CaC03; and since stat-
uary marbles are usually selected for purity and homogeneity the great major-
ity contain more than 90 per cent, and some over 99 per cent of this single
mineral. A principal component of a few marbles, among them some from the
ancient quarries of Naxos, is dolomite, CaMg (C03)2.   During metamorphic re-
crystallization calcite crystals have grown to diameters of the order of 0.5
to 2 mm, and now build a closely interlocking mosaic. Metamorphic temper-
atures mostly lie within the range 25 0C-650?C.   Under these conditions,
maintained in rocks deeply buried for periods of 10,000 to 10,000,000 years,
mineral impurities, notably clays and silica, when present, react with the
carbonate matrix to produce new and characteristic metamorphic minerals read-
ily identified with the polarizing microscope.
Non-Diagnostic characteristics.
From what I have said it is clear that all statuary marbles have much
-34-
35
in common. Nothing of the unique character of marble from an individual out-
crop or quarry is likely to emerge from routine chemical analysis, study by
X-ray diffraction, or even from superficial microscopic examination by an
"expert" lacking wide petrographic experience.   Such techniques, valuable though
they may be in identifying other materials, will merely tell us that we are
dealing with rocks composed principally of calcium carbonate (calcite) with a
minor-element pattern (trace of lead, copper, strontium, barium and so on)
that is monotonously the same for most marbles. And this leads to a point of
significance in evaluating "expert" opinion based on examination of archaeo-
logical materials by modern sophisticated techniques. However refined or com-
plicated these may be, their use is justified only where they can yield new or
more precise diagnostic information not obtainable by simpler and more common-
place means. A dozen years ago the Fogg Museum acquired a marble statue of
Trajan. Critical as to its authenticity and history was the source of the
marble from which it was carved. It was submitted to an "expert" who carried
out detailed investigation by x-ray diffraction and spectrographic analysis
and superficial petrographic examination. From all this there emerged, as
could have been predicted by any mineralogist, the information that the Trajan
statue is composed of uniform-grained marble with the chemical composition
(including trace-element pattern) of almost any marble. Yet, on the basis of
this evidence, the material was identified as "second grade Carrara marble;"
and this pronouncement became an essential component of a lengthy reconstruct-
ed history of the statue, beginning in ancient Rome. It may well be that the
story so put together is essentially true. The marble itself may well have
come, as claimed, from Roman quarries at Carrara. But no compelling evidence
to this effect has been presented; and the marble may equally well have come
from any of a hundred sources, known or unknown, within or outside Europe.
Diagnostic Characteristics of Marble in Problems of Archaeology.
How then may petrography be of use in archaeological problems concerning
classic marble? We commonly encounter two kinds of problems relating respect-
ively to (1) the possible geographic source of a particular piece of stone, and
(2) the possible matching of fragments supposedly broken from the same piece -
especially in reconstruction of fragmented inscribed slabs.
Diagnostic petrographic characteristics that may contribute toward sol-
ution of such problems are those that reflect the imprint of the particular
local episode of metamorphism by which a particular limestone was transformed
into a particular marble.   Metamorphism, it will be recalled, is a response
to long-sustained high temperature and pressure over some period of geologic
time. For Greek marbles this was probably during the Paleozoic era - perhaps
400-300 million years ago. The marbles of northern Italy - at Carrara, west
of Florence, and at Lasa in the upper Adige Valley - were metamorphosed much
later during the growth of the Alps, beginning perhaps 50 million years ago.
The temperature-pressure conditions at any broad locality, e.g., in the
36
Pentelicon-Hymettus area west of Athens, or in the Carrara quarries of Italy,
cover a limited range.  Moreover, during metamorphism, which also involves
intense deformation and flow of marble in the solid state - just as in the
forging of red-hot metal - the pattern and degree of deformation are likely
to differ from one locality to another; but at any one locality they are
likely to be rather uniform.
The useful characteristics that we seek in marble for present purposes
fall into three categories relating to texture, fabric and mineralogy:
Texture. The most obvious textural characteristic of a thin section of
marble is grain-size. This tends to be highly variable from point to point
on a quarry face, or even within a small slab or sometimes in a single thin
section. The diagnostic value of the size criterion is correspondingly
slight. Three characteristics that reflect deformational history, and so are
more uniform in a given outcrop are degree of twinning, grain shape and con-
figuration of grain boundaries.
Twinning. Any crystal subjected to stress at high temperature and con-
fining pressure may respond by internal flow in a pattern dictated by the
regular geometric arrangement of its component atoms. Seen in the simplest
way the response can be envisaged as mutual displacement of adjacent planar
arrays of atoms by analogy with sliding of cards in a card deck. By this
means the grain elongates in one direction and becomes flattened in the di-
rection perpendicular thereto. In special cases this process of slip or glide
leads to development of thin laminae within which the atoms have become re-
arranged in a configuration symmetrically related to that in the host crystal.
Such a reconstituted layer is termed a twin.   Seen beneath the polarizing
microscope twins are visible because their optical behavior differs from that
of the host (Plate 1). Calcite is a mineral that twins readily under stress -
especially at temperatures of less than about 500?C. The degree of twinning
shown by calcite in marble tends locally to be rather consistent; it may
differ significantly from one locality to another, or even from point to
point in one quarry. Contrast the heavily twinned conditions of some grains
in a marble from the Greek island of Kos (Plate 2a) with the paucity of twinn-
ing in a specimen from Paros (Plate 2c). On the whole twinning is not pro-
fusely developed in true classic Pentelic marbles -those from ancient quarry
sites on the lower levels of Mt. Pentelicon (Hertz and Pritchett, 1953, p. 75,
77) - or in Hymettan marble from the Hymettan quarries (Plate 3a). It is
even less conspicuous in most pure white Carrara marbles (Plate 4).
Shape and outline of grains. In marbles that were strained at rather
high temperature without subsequent recrystallization, individual grains show
a tendency for local parallel elongation (Plate la). Where recrystallization
has outlasted deformation - or commoner situation - the grains show no such
elongate shape and are said to be equant (Plate 2). Grain outlines may be
37
highly irregular and interlocking (Plates 2, 3). More rarely when high-temp-
erature recrystallization followed relatively cold deformation and was long
sustained under stress-free conditions, the resulting grain boundaries tend
to be planar. In the ideal case where surface tension at boundaries has been
minimized the section shows points at which three boundaries intersect at
about 60? (Plate 4a, b). Textures of this kind are characteristic of annealed
metals. Combined properties of grain shape and outline are likely to be con-
sistent within the limits of a single slab, or even within a quarry.    Much of
the high-quality marble from the Carrara quarries has equant grains with
planar outlines - what we might call an annealing texture. Most specimens of
Greek marble that I have examined - Pentelic, Hymettan or from Paros - have
equant grains, often variable in size, with sutured or irregular boundaries.
The few specimens that I have seen from Naxos are generally coarse, locally
variable in size, equant and with some planar boundaries. Some Naxos marbles
consist largely of coarse dolomite with sutured margins.
Fabric.  A much more subtle set of structural characteristics, difficult
to measure but consistent within a slab or even a large quarry, are what
students of deformed rocks collectively term fabric. During rock flow under
stress the individual microscopic crystals tend to become aligned in some sym-
metrical pattern related to that of flow itself. By analogy we may recall
the respective patterns of logs in a flowing stream or of clouds in the steady
trade-wind sky. It is not only by external shape that elongate grains become
aligned. Even in marble whose grains are equant there is a pervasive though
invisible pattern of orientation of the principal symmetry axis of the indi-
vidual calcite grain. To bring out the pattern we must use specialized and
rather laborious techniques of microscopy or X-ray analysis (cf. Weiss, 1954;
Herz, 1955). But this is by far the most effective tool in problems relating
to matching of fragments of statuary or of inscribed surfaces. If fragments
match not only in direction and sense of inscription but also in the pattern
and orientation of crystal fabric, the chances that they belong to the same
slab are very strong indend. Conversely, failure to match can be demonstrated
with certainty.
Mineralogy. Much of the marble from the classic Dionysus quarry on Mt.
Pentelicon is streaked sparsely with greyish green aggregates of the common
metamorphic minerals chlorite, albite and epidote (Plate 3b) - all recogni-
zable at a glance in a thin section.   Occurrence of the same suite of minerals
in marble of a temple such as that of Sunion is consistent with a Pentelic
source, but is by no means conclusive proof of this; for impurities of the
same kind are known in other mables.   Some sources can, however, be excluded
with certainty.   Such is the group of quarries near Apollonia on the north-
east coast of Naxos. All Naxos marbles that I have seen are almost pure cal-
cite or dolomite. But if silicate impurities occur within them, it is imposs-
ible that they should include chlorite-albite-epidote. General experience
of metamorphic petrography shows that this mineral assemblage can form at
38
temperatures no higher than 350?. On the other hand sufficient is known of
the geology of Naxos to indicate that in the general vicinity of Apollonia
metamorphism was effected at temperatures of the order of 5000-6000C
(Schuiling and Oosterom, 1967). Chlorite, albite and epidote could not co-
exist in this environment.
I have seen very little of Egyptian statuary marbles. Several speci-
mens that I have examined microscopically contain silicates such as forster-
ite, Mg2SiO4, indicating conditions of metamorphism (local high temperature
and low pressure) completely inconsistent with classic Greek and many north
Italian sources (among these Carrara). The sources are probably local. The
Egyptian marbles could not have come from classic Greek quarries. But as
far as mineralogical evidence alone is concerned, the source could be located
(though I think the probability infinitely small) in the Mojave Desert or the
Sierra Nevada of California'
Mineralogical evidence accruing from petrography, relating to our pre-
sent problem, leads mainly to exclusion of particular alternative solutions.
Rarely we encounter evidence of a more positive kind. Some marbles are streak-
ed with red, due to iron-oxide impurity in the form of hematite. Among the
specimens collected by Herz from a classic Pentelic quarry is a red variety
that in my experience is mineralogically unique. It contains the easily rec-
ognized red-to-orange highly refractive calcium-manganese silicate piedmon-
tite. I know of no other marble from Greece or elsewhere that contains this
mineral.  Yet, by way of due caution, it seems most likely that somewhere,
and quite possibly in Greece, another such rock may exist. Any slab of pied-
montite-bearing marble found in Greece has most likely, but not certainly,
come from Herz' source on Mt. Pentelicon.
Conclusion
Petrographic investigation can contribute a good deal toward solution
of problems relating to sources or to matching of marble objects of archaeo-
logical interest. Only a start has been made in exploring the petrographic
nature and variety of marble at a few classic quarry sites. Much more must
be done, preferably by geologists. Badly needed, too, is extensive expert
petrographic work on the objects themselves. This must be done by archaeo-
logists, for only they can set up the problems and sustain interest in
their solution. Any young student of classic archaeology who might care to
include in his university training a few adequate courses in crystallography,
mineralogy and petrography could make an immense contribution in this field.
This is not so difficult a task as, for example, to master a language such
as Russian or even German.
Finally it must be realized that archaeological problems in which
the source of marble objects is crucially significant are seldom open to
unique solution, Petrographic evidence may rule out certain hypothetical
39
solutions as impossible. Others may be found consistent with the petrographic
data, but no more than that. Very rarely a particular alternative may find
petrographic support demonstrating probability almost - but never quite - to
the point of certainty. But certainty, however much it tinges much of the
opinion found in the older literature of classic archaeology and in the reports
of some museum "experts," is not the goal of the scientist's quest. The petro-
grapher, viewing from the outside problems of the kind we have been discuss-
ing, as clearly set out in the critical essay of Herz and Pritchett (1953) on
"Marble in Attic Epigraphy," finds a paradoxical situation: pervading much of
the pertinent literature of classic archaeology, as well as the reports of
some "experts," there seems to be some reluctance to apply the principles of
logic formulated - while some of the ancient quarries were operating, and
others as yet unopened - by Socrates and Aristotle.
Acknowledgment
Financial assistance from the National Science Foundation (Grant GA-
10636) is gratefully acknowledged.
Bibliography
Herz, N.: Petrofabrics and classical archaeology, Am. Jour. Sci., v. 253,
pp. 299-305, 1955.
Herz, N., and W. K. Pritchett: Marble in Attic epigraphy, Am. Jour. Archaeol.,
v. 57, pp. 71-83, 1953.
Schuiling, R. D., and M. G. Oosterom: The metamorphic complex on Naxos
(Greece), Proc. Koninkl. Nederl. Akad. Wetensch. Amsterdam, Ser. B,
v. 70, pp. 165-175, 1967.
Weiss, L. E.: Fabric analysis of some Greek marbles and its applications to
archaeology, Am. Jour. Sci., v. 252, pp. 641-662, 1954.
40
Quartz inrclusions
Fig. 3a. Radiation environment of crystalline grains embedded in the pottery clay matrix. For
grains of v, 100 microns diameter the alpha radiation from the clay is attenuated by
approximately 80%. For grains ofevmicrons the alpha radiation is almost entirely atten-
uated but now even the beta radiation from the clay is attenuated (by approximately 50%).
Crystatline inclusion
/3,   radiation
Surrounding clay matrix
Fig. 3b. Alpha radiation attenuation by large crystalline grains. R is the maximum range of alpha
particles in the crystal and a is the radius of the grain. As no alpha particles penetrate
beyond a depth, R, into the grain, the inner regions experience only beta and gamma
radiation during the burial of the pottery.
41
a)
0 4J
F,
S
Cco
0
-4
O
p
0
U
2
a)
U
a)
0)
>-4
ifo
(4-4
N0
4-i
:.<   44
0
Cto
. H
''; .W
0
m
--4
1-4
0
0
to
3
-H
"I5:
I4-
CiN
a)
4-i
0
4-i
I i
l-4
-4
a)
4-
-H
1-4
a)
N
Cco
1-4
0
0         .
1-4
a)
4--i
-4
.41
0 0
00:x:
Z'
~:0
Q    0) -4
bo *  Q:
0'H 0-
:i2 0 0>
A 0-4
bO bOt
0   0
0 Cl 0
o   *- CO
4'H a)  a
a'*O
O   *  m
4       0
a )   Cco
4-   ,i .\'
4  H , 4
:~-H
a)   4- JO
o0   0 p
a)  'H4  a)
ITLi44 0 0
- O    4
-4   'H  '0
co
co
CO0  c 0-4
O     H
'Ha))
coQ  )) a
a)-  Ci
N
a)
z
a)
4i
0
0
-4
-4
0
0
a)
0-I
Ca
a)
a)
0
coi
C))
0
-H
0
a)
1-d
C
v
42
(U)
00
4,-
co
U)
,*-I (
.~
--I
0
0 u
O ,   ,
,OO -4
N -,4 4i
U)Q
.,4 i
co  v
,-IH 4
4J
la
(U
H
U)
-4
v
co
O
s-I
U)
U)
p
0
u
,-4
v
4.4
0
(U
0
4-i
U
U)
(U
H
a
U)
4-i
(U
P-I
O
O        Cn
0
,-4
,- 44
(U)
4     ,-4 ,-
(U
O. c
-    -
U) N'
co
4-i
JJ
es )
N
cn U4
U)U)
.H
ns-I
5-4
84
-Cl
(UU)
(U
G
(U
co
U)
4
0
4J
O
co
CO
a)
,--I
rl
54
44
U)
0
0
.4
0
o
U)
O
4i
O
0
co
'-I
1-4
I ns
0
-.4
q
'-A
co
-H
co
44i
'-I
0
tl)
(U
0
4-A
0
.,4
E-4
4i
04
P
u
v,
V
00
x
(U
54
(U
C4
0
(a
_/1