Analysis of Raw Glass From Carthage HIMT

INSTITUT NATIONAL D',ARCHEOLOGIE ET 0' ART DE TUNISIE
AND BRITISH ACADEMY TUNISIA COMMITTEE
EXCAVATIONS AT CARTHAGE
THE BRITISH MISSION
VOLUME II, 1
THE CIRCULAR HARBOUR, NORTH SIDE
THE SITE AND FINDS OTHER THAN POTTERY
BY
H.R.HURST
Contributions by
C. Duhig, S.P. Ellis, S. Gibson. M. Henig, I.C. Freestone,
G. Lloyd-Morgan, M.A. Levine, D.R Pringle, R Reece,
V.A. Tatton-Brown, RF. Tylecotet. W.A. van Zeist,
A. Wheeler, J. Zaouali
.. '.
Published for THE BRITISH ACADEMY

by OXFORD UNIVERSITY PRESS
1994
, 290 CHAPTER 15
c.t. no. and type No. or uamples several components. namely iron. manganese and titanium oxides
5. deep bowl 5 (FeO. MnO and Ti02 respectively). A typical Roman vessel or
17. 19.20.22-3. conical beakers 11 window glass without "deliberate" colouration (i.e. where the blue or
25. concave beaker bases 9 green colour is incidental due to impurities in the raw materials) has
28-31. beaded-stemmed goblets II less than I% each of FeO and MnO and less than 0.3% Ti02; glasses
32-3. plain-stemmed goblets 12 32831 and 32834 would not therefore be unusual for Roman glass
34. slender splayed·stem goblets 5 (Vel de op.cil .• Velde and Gendron op.cit.). However. several of the
36-7. folded-in goblet feet 9 present glasses are exceptionally high in all of these components
(32832.32833 and 33027) with up to 0.7% Ti02' 2.7 MnO and 3.3%
The concave bases (no. 25) and the beaded- and splayed- FeO. Furthermore. there is a very marked correlation between FeO
stemmed goblets (nos. 28-31 and 34) featured in the and Ti02. and a less marked correlation between FeO and MnO.
comparable list at AvB. Discounting the goblet feet, deep The high concentrations of iron and titanium in the glasses result in
their green colouration. Greens and yellow of this type arc not
bowls (no. 5) and conical beakers (nos._17-19) emerge as
generally considered to be deliberate but the result of unintentional
more common on this site than at AvB. In the case of the incorporation of the colourant elements in the glass. Generally in
bowls, this is probably a reflection of their date. since two Roman glass iron oxide was at the level of one percent or less. and the
of the five were from Istl2nd-century contexts, suggesting colouring effects of iron were counteracted by the addition of
that this was an early type. The earliest context for a conical manganese oxide (Sayre 1963).
The explanation for the unusually high concentrations in the
beaker was phase 4.7d. generally associated with pottery of
Carthage fragments probably lies in the contamination of the glasses
1stl2nd century date, but all the other contexts were 4th- with iron and titanium by poorly selected raw material (sand) or by
century or later. ThIJS chronology does not seem to be the dissolution of the 'melting pot. This would have produced a strongly
reason these vessels were more common on the present site. coloured glass and an attempt has been made to reduce the colouration
On the other side, one or two very common forms at AvB. by the addition of excess manganese oxide (MnO). Unfortunately at
the exceptionally high concentrations of iron and titanium in these
were lacking on the present site, notably nos. 97, the concave glasses. the manganese oxide has proven ineffective as a decolourant
base now reclassified as belonging to a beaker (of which there so that the high iron-manganese-titanium ("HIMT") glasses (32832.
were some 44 examples or 8% of the entire glass assemblage), 32833. 33027) are stongly coloured relative to. for example. 32831
and nos. 98-9, cylindrical flasks or jugs with pushed-in tubular (yellow-green) and 32834 (pale blue).
base rings (27 examples or 5% of the assemblage). In other Other examples of HIMT glass of Roman date appear rare. but one of
forty-eight peices of Gallo-Roman glass analysed by Velde (l9S5) has a
respects the two collections were similar. similar composition with 0.6% Ti02' 2.2% MnO and 2.0% FeO. Given
that the present analystica1 sample was deliberately selected to include
lighter pieces to give a full range of colour. HIMT glass is likely to be
APPENDIX: CHEMICAL ANALYSIS OF even more strongly represented in the Carthage group than is indicated . .
by Table I.emphasising the unusual nature of the material.
'RAW' GLASS FRAGMENTS The fragments of glass from Carthage are likely to have been raw glass
broken up for use in glass-working. Their strong colour caused them to be
by I.C. Freestone
rejected by the craftsman and discarded. The presence of some less
strongly coloured material in this sample is not surprising. as a certain
Around twenty angular glass fragments. up to about 4 em across and amount of accidental loss and wastage in such a process in inevitable.
weighing approximately 500g in total. were submitted for examination. The presence of a group of glass fragments with a high proportion
The fragments are all transparent, and range from pale yellow to dark of deliberately discarded material strongly suggests that sorting of raw
olive green in colour. although one fragment is a distinctive pale blue. glass fragments was occurring in the vicinity. and that glass-working
As indicated elsewhere. they appear to be the result of breaking up large was occurring nearby.
pieces of raw glass; vessel fragments. drips and trails are absent.
TABLE I - Energy Dispersive X-ray Analysis of Glass Culler
ANALYSIS Sample No. 32831 32832 32833 32834 33027 33028

Location Rm7 RmB7 RmB2 Rm IS RmB9 Rm7(S
Small samples were broken from six of the samples selected to robber tr.)
represent the range of colours present. including the pale blue. They Layer 657 604 150 179 603 46
were mounted in epoxy resin. polished. coated with a thin layer of Phase 4.3Sd 4.36f 10 5.27b 4.3Sc 10
carbon and analysed by energy dispersive x-ray analysis in the scanning
electron microscope; for details see Bimson and Freestone (1988). Si02 68.5 64.2 63.7 69.4 64.8 65.9
TI02 0.2 0.7 0.5 0.2 0.7 0.2
AI203 I.S 3.3 2.8 2.9 3.2 2.7
FcO 1.0 3.3 2.0 0.5 2.1 1.0
RESULTS AND DISCUSSION MnO 0.8 1.6 2.7 0.2 2.7 1.9
MgO 0.5 1.2 I.l 1.0 1.3 1.3
The analyses are presented in Table I. The glasses are of the soda- CaO 6.4 5.0 6.5 9.5 5.2 8.6
lime-silica type. typical in this respect to most other Mediterranean Na20 19.4 18.4 18.4 15.5 IS.7 16.9
glass. Their low magnesia (MgO in the range of 0.5-1.3%) and potash K20 0.3 0.4 0.5 0.3 0.4 0.8
(K20) in the range 0.3-0.8%) are typical of glass of the Roman CI 1.0 1.0 0.9 0.9 1.1 0.9
period. as opposed to Islamic glass in which concentrations of these 0.3 0.1 0.3 0.1 0.1 0.2
5°2
oxides are higher (Sayre and Smith 1'961). Alumina (AI203) at
1.&-3.3% is also typical of Roman glass (eg. Bimson and Freestone NOles Results in weight per cent. Analyses carried out by energy
1983. Velde 1985. V!:lde and Gendron 1980). dispersive x-ray analysis in a JEOl JSM840 scanning electron
In detail. however. this group of glasses shows some unusual microscope. Other elements analysed but not detected include
features. In particular a number of the samples are unusually high in antimony, lead. tin and copper.

Analysis of Raw Glass From Carthage HIMT

Uploaded by

Copyright:

Available Formats

Analysis of Raw Glass From Carthage HIMT

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Analysis of Raw Glass From Carthage HIMT

Uploaded by

Copyright:

Available Formats

INSTITUT NATIONAL D',ARCHEOLOGIE ET 0' ART DE TUNISIE

AND BRITISH ACADEMY TUNISIA COMMITTEE

Published for THE BRITISH ACADEMY

ANALYSIS Sample No. 32831 32832 32833 32834 33027 33028

You might also like