Investigations of Astrolabe Metallurgy Using Synchrotron Radiation
arcs spaced with a particular symmetry,
rather than complete rings.
that is transmitted gives a measure of the
sample thickness. By scanning the sample
across a small beam and recording the
transmitted intensity at regular intervals
(e.g., every 0.5 mm), a thickness profile is
obtained along the line that was scanned.
As shown in Figure 5, a typical scan of
the x-ray transmission across the M-33a
plate reveals irregular thickness variations,
consistent with a sheet produced by ham-
mering. A typical scan across the DW0595
plate reveals a uniform thickness charac-
teristic of a rolled sheet. The fine-scale
variations probably result from the lines
incised in the plate.
destructively, in about 36 h of beam time,
for 15 astrolabe components and three
measurement types. This efficiency is an
important consideration for delicate arti-
facts that must be removed from museum
display or storage and transported to the
experimental site(s).
ꢀ
Recrystallization during heating of such
a textured microstructure can change the
texture, and thus the symmetry, of the dif-
fraction pattern.
ꢀ
Manual hammering of a sheet tends to
randomize the orientations of the crystals,
so that bright arcs appear in all directions
around the rings.
Analyses such as those reported here
are particularly interesting for compara-
tive purposes. One desires a large number
of reference measurements with which to
compare a newly measured astrolabe. We
intend to continue the types of analysis
discussed here with as many of the Adler
instruments as possible, to ascertain the
composition, microstructure, and thickness
profile characteristic of astrolabes from dif-
ferent time periods, geographic regions,
and makers. Once these data are collected,
statistical analysis will help us evaluate
measurements of particular instruments.
More generally, we wish to continue ex-
ploring interdisciplinary research oppor-
tunities, to see whether the points of view
of the physical scientist and the historian
can complement one another. In physical
science, one studies a sample to learn not
about that particular sample but about
something more general, a process or a
type of material. Variations that are char-
acteristic of a particular object are often of
little interest. Historical artifacts, in con-
trast, are of interest not only as examples
of more general classes but often specifi-
cally because of their unique attributes.
Study of those very attributes can bene-
fit greatly from the judicious application
of analytical methods from the physical
sciences. It is only because the techniques
of fluorescence analysis, diffraction analy-
sis, and scanning radiography have been
developed in a scientific context that they
are understood well enough for them to
yield understanding of a particular astro-
labe signed by Ioannes Bos on 24 March
1597. Yet a focus on a single object can pull
us away from generalities into the more
complex world that is, in the end, what we
are all studying. In simplest terms, the sci-
entist is often interested in reproducible
general phenomena, to the neglect of indi-
vidual variations. The historian may focus
on particular details and fail to see beyond
them. We think that both scientist and his-
torian can profit from exposure to the
other’s point of view.
Note that unlike fluorescence, transmission
diffraction analysis probes the interior of
the specimen. It requires a beam of suffi-
ciently high energy to avoid complete ab-
sorption in the artifact. The 68-keV beam
available at the APS easily penetrates
millimeter-thick brass sheets.
The Mystery of Ioannes Bos:
The Metal Speaks
Figure 4 shows typical diffraction pat-
terns from four representative astrolabe
components. The diffraction pattern from
the M-33a plate (Figure 4a) shows rings
consisting of random orientations of bright
arcs, consistent with thorough working
of the plate by manual hammering. The
M-33a rete pattern (Figure 4b) is very simi-
lar, indicating that the sheet of metal from
which the rete was fashioned had also
been produced by hammering. Thus, the
intricate detail of the rete was likely pro-
duced by cutting out many small regions
from a sheet, a very laborious process em-
ployed in 16th-century astrolabe manufac-
ture.14 The diffraction pattern from the
DW0595 plate (Figure 4c) is strikingly dif-
ferent. The bright spots show a highly
directional {100} ꢀ001ꢁ (cube face) texture.
We found that the orientation of the tex-
ture is identical in different areas of the
plate and, in fact, is aligned with the en-
graving on the plate. This texture is that
expected for recrystallized, cold-rolled
copper.13 This would be consistent with
the fire-gilding of a modern rolled copper
sheet. In addition to the bright spots from
the Cu, the diffraction pattern shows weak
uniform rings with smaller radii, corre-
sponding to the larger lattice constant of
Au, coming from the fine-grained micro-
structure of the surface gilding. The dif-
fraction pattern from the rete of DW0595
(Figure 4d) shows many randomly ori-
ented bright spots from the Cu, rather
than a strong texture. Evidently, the rete of
DW0595 was not cut from a rolled sheet
similar to that of the plate, but instead it
was made by casting.
Metallurgical analysis using x-rays indi-
cates that the composition and microstruc-
ture of the Adler instrument M-33a (apart
from the spacer) are consistent with the
1597 date, while those of the Harvard in-
strument DW0595 indicate a much more
recent origin. To say that the metallurgy of
M-33a is consistent with the date engraved
on it is not the same as confirming that
date. We have shown that M-33a was
made with technology available 400 years
ago: hammered sheets of low-zinc-content
brass. On the other hand, DW0595 shows
signs of a more modern technology: cold-
rolled sheets of copper. Since the process
of rolling wide sheets of metal had not
been developed by the end of the 16th
century,3 it is unlikely that DW0595 bears
an authentic date and signature.
The rete of DW0595 shows crystallog-
raphy indicating that it was cast. It is plau-
sible that the rete, by far the most intricate
component of the instrument, should be
made by casting from a mold, since
DW0595 is one of a group of identical as-
trolabes now dispersed among various
collections.5 It appears likely that the in-
struments in this group are reproductions
of M-33a, made at the same time as each
other, perhaps as attractive (and extrava-
gant) souvenirs.
Outlook
Of the three measurement types reported,
only the diffraction analysis required the
intense, highly collimated beam of a syn-
chrotron. A less brilliant laboratory source
would have sufficed for the fluorescence
analysis and the radiography. The particu-
lar advantage of the synchrotron for these
two analyses was convenience. Since the
artifacts had been removed from museums
and taken to the beamline for the diffrac-
tion analysis, it was simple and efficient to
perform fluorescence analysis and scan-
ning radiography at the same time. The
results reported here were obtained non-
Scanning Radiography
Acknowledgments
Hammered metal varies noticeably in
thickness across a distance of a centimeter
or two, because it is nearly impossible to
hammer a sheet to a uniform thickness. A
modern rolled metal sheet is highly uni-
form in thickness. For a sample of uniform
composition, the fraction of the x-ray beam
The authors wish to thank Martha
Goodway for helpful comments, and
Armon McPherson and Adam Pyzyna for
their help during the experiments. The
DW0595 astrolabe was loaned from Har-
vard University’s David P. Wheatland Col-
lection of Historic Scientific Instruments,
22
MRS BULLETIN/JANUARY 2001
Kohler, Sonnichsen
