All About Glass

You are here

Analyses of Some Finds from the Gnalić Wreck

All About Glass

Fragments of eight types of objects excavated from the Gnalić shipwreck were submitted to The Corning Museum of Glass for examination and chemical analysis.1 The specimens, which are from the collections of the Narodni Museum in Zadar, consisted of fragments of six glass objects and two pigment samples. Several of the fragments are of the types illustrated by Sofija Petricioli in the preceding article.

We were very pleased to have the opportunity to analyze these glasses since so few analyses of Venetian glasses of firmly-established dates have been done. Because the fragments themselves were sizeable, it was possible to remove and sacrifice samples of glass which were large enough for full quantitative chemical analyses. This is in marked contrast to the usual situation, for most of the early Venetian glass which has survived is in the form of complete and valuable objects. Consequently it is rarely possible to obtain a sufficient sample of glass for analysis without either inflicting unsightly damage to the objects or running the risk of breakage. We will take full advantage of this opportunity by carrying out two additional types of experiments which are expected to provide very useful information. These two experiments are oxygen-isotope2 and lead-isotope3 determinations. Unfortunately, neither of the isotope studies has been completed as this article goes to press.

The determinations of soda (Na2O), lime (CaO), potash (K2O), magnesia (MgO), alumina (Al2O3) and iron (Fe2O3) were carried out by the atomic absorption method. The copper (CuO), lead (PbO) and tin (SnO2) determinations were also by atomic absorption if the preliminary emission spectrographic analyses indicated they were present at levels of greater than a few tenths of a percent. The analyses of the other elements were by emission spectrography, except for SiO2, which was estimated by difference from 100%.

Analyses of Some Gnalić Glasses
    1764 1765 1767 1768 1769 1770
SiO2 d ≈71 ≈72 ≈71 ≈71 ≈67 ≈64
Na2O a 13.7 12.8 12.3 13.1 12.3 10.3
CaO a 6.53 8.04 8.86 7.78 7.49 8.40
K2O a 2.88 2.60 2.36 2.74 2.18 2.70
MgO a 1.76 2.50 2.58 1.96 1.76 2.00
Al2O3 a 1.60 1.18 1.11 1.44 1.48 2.72
Fe2O3 a 0.75 0.39 0.44 0.60 1.21 2.97
TiO2   0.03 0.01 0.01 0.02 0.01 0.02
Sb2O5   n.f. n.f. n.f. n.f. n.f. n.f.
MnO a 0.94 0.50 0.50 0.63 1.12 1.49
CuO a 0.01 0.01 0.01 0.01 0.28 0.56
CoO   0.005 0.001 0.001 0.001 0.16 0.15
SnO2 a 0.08 0.03 0.18 0.03 2.56 2.11
Ag2O   0.000X 0.000X 0.000X 0.000X 0.000X 0.000X
PbO g 0.28 0.20 0.25 0.072 2.34 2.25
BaO   0.01 0.01 0.01 0.01 0.02 0.03
SrO   0.18 0.18 0.22 0.22 0.15 0.20
B2O3   0.02 0.02 0.02 0.02 0.02 0.02
NiO   n.f. n.f. n.f. n.f. 0.06 0.07
Bi2O3   0.005 n.f. n.f. n.f. 0.13 0.13

a Analysis by atomic absorption.
g Analysis by gravimetry.
d SiO2 estimated by difference.
All other analyses by emission spectrometry
Sought but not found: V2O5, Cr2O3, ZnO, As2O5, Li2O, Rb2O, ZrO2 and P2O5.

The results of the analyses are presented in the accompanying table. The glasses are generally quite uniform in composition, so much so, in fact, that all appear to have been made in the same factory. We were especially interested to discover this since we wanted to know if the window glass (no. 1764) was made in the same factory as the vessels. The results of our analyses give no reason to believe that the roundels were made in a different factory. The basic compositions are of the soda-lime-silica type (Na2O:CaO:SiO2). The levels of potassium and magnesium are sufficient to confirm that the source of the soda alkali was a plant ash, probably barilla or polverine. The levels of iron are about what one might expect in an early glass-given the limitations which then existed on purification of raw materials, but the greenish tints which this much iron would be expected to cause have been quite effectively removed by the use of a decolorizer. This decolorizer was manganese, as is indicated by the presence of MnO in the glasses at a level which can be regarded as the result of an intentional addition. The marked grayish cast of the glasses is to be expected from the action of a manganese decolorizer. Beyond this, the minor and trace element levels are not unusual, except for the significant percentages of cobalt (CoO) and lead (PbO). It is tempting to associate the low, but significant, concentrations of cobalt in the glass with the manganese. Cobalt at the 0.005% level does have a noticeable color effect and is used today in conjunction with other elements to control color in some glasses.

Whereas the cobalt could have been an intentional additive, this does not seem to have been the case with the lead. At a concentration of a fraction of a percent, the lead would not alter the properties of the glass noticeably. It seems more likely to be a persistent impurity. The tin (SnO2) impurity level could be associated with the lead.

The colorants in the two strongly colored glasses (nos. 1769 and 1770) are cobalt oxide for the dark blue and a mixture of cobalt, copper, iron, and (possibly) manganese in the "black" glass. An interesting characteristic of these two glasses is that the tin and lead levels are quite high. (A possible explanation is given below.) The nickel (NiO) and bismuth (Bi2O3) impurities are also noteworthy, and might someday serve to characterize the sources or nature of the ingredients used as colorants for glasses in sixteenth-century Venice. There are several associations which could be considered among the nickel, bismuth and silver on one hand, and the copper, cobalt, tin and lead on the other; and the manganese and iron could also be involved. But to attempt to establish relationships on the basis of only two colored glasses would be too speculative at the moment.

Another fact emerges from these data which will be of interest to those concerned with the conservation of glass. Certain types of European glasses, including some from Venice, are subject to a chemical attack by atmospheric moisture. This attack ultimately produces surface deterioration on the objects, taking the form of crizzling, “weeping” or a cloudiness and loss of transparency. The susceptibility of a glass to this attack is dependent upon its chemical composition.4 Chemical analyses of crizzled or similarly afflicted glasses have established that the most common cause of this susceptibility is a deficiency in lime (calcium oxide, CaO). When the lime content falls below about 4% (by weight) the resultant glass is in danger of attack by moisture. A stable glass, if it does not contain excessive alkali, will contain about 7-11% CaO.

The analyses of the Gnalić glasses show that their compositions are stable and that the glasses would not have been particularly prone to the hydrolytic attack. This is consistent with the observation that the glasses, although heavily pitted on the surfaces, are not unduly weathered, considering that they have been submerged in water for nearly four centuries. Had the compositions been in the unstable range (and the glasses susceptible to crizzling) it might be that none of the glass would have survived at all.

For purposes of comparison, we can mention analyses we have made of some other Venetian glasses, although they are not yet published. These other glasses are from the collection at Rosenborg Castle in Copenhagen. (The samples were provided by Mr. Gudmund Boesen.) These specimens are important because, like the Gnalić glasses, they are well dated, having been acquired by Frederick IV in 1708-1709. Several of the Rosenborg glasses are heavily crizzled and our analyses demonstrate that the instability is due to low lime and high alkali contents. (One glass, for example, containing 18.4% K2O, contains only 1.3% CaO.)5

There is one chemical feature of the Gnalić glasses—the presence of a small copper impurity in all the samples analyzed—which can be interpreted in terms of technological processes. Neri6 prescribes the use of copper and brass vats for alkali preparation. The use of such implements could introduce traces of copper into the alkali and subsequently into the glass.

Regarding glass colorants, Neri makes frequent mention of cobalt (as zaffre and smalt), manganese, copper, iron and lead. These ingredients are used in various combinations to arrive at slightly different color variants which were probably used to satisfy the changing tastes and preferences (not to mention economic needs) of both the glassmaker and his customers. Tin, which is a substantial ingredient in the two Gnalić colored glasses, is not mentioned specifically in the sections Neri devotes to colored glasses. But in The Sixth Book, Neri describes a "Stuff for Enamels," which is the basic ingredient in his recipes for making enamels. This material is made by calcining together near equal weights of "fine lead" and "fine tin." Used in combination with colorants, such as manganese, cobalt, iron and copper, it yields enamels of various colors. One cannot help but notice that the proportions of lead and tin oxides in the Gnalić colored glasses are about equal, and wonder if a prepared enamel had not somehow found its way into the glass as a colorant.

The two remaining specimens, nos. 1762 and 1763, were chunks of raw pigments. We carried out X-ray diffraction experiments which confirmed, as had been suspected, that the substances are, respectively, cinnabar and white lead. The latter is a valuable specimen because lead-isotope data can now be routinely utilized for studying the geographical origins of the ores from which leads were extracted. The value of this specimen lies in its possible bearing on our lead isotope analyses of pigments in Italian paintings.7 It will also be interesting, of course, for comparison to the isotope data for the lead in our specimens of Venetian glass from the Gnalić and Rosenborg collections. The lead-isotope technique, which consumes only very minute samples of glass, is perhaps the most promising of all scientific methods for the possible differentiation of Venetian and façon de Venise glasses.



1764—Fragment of roundel of window glass. Colorless glass with grayish cast; lightly weathered. Pontil mark preserved and arc of folded rim; estimated original diam. ca. 15.6 cm. No. 42/645. Petricioli Fig. 22.

1765—Foot and stem of goblet with scratch engraved designs on foot. Colorless glass with grayish cast; lightly weathered. Somewhat similar to Perticioli Fig. 18.

1767—Base fragment of vessel, possibly a bottle or jug. Colorless base glass with remains of lattimo banded decoration; lightly weathered. Sample was of colorless glass mainly, but contained traces of the lattimo glass. Petricioli Figs. 10 or 16.

1768—Body fragment of stemmed, thin walled goblet. Colorless glass with slight greenish cast; lightly weathered. Petricioli Fig. 3.

1769—Fragments of neck of thin-walled vessel or large vial. Blue transparent glass with many elongated bubbles; lightly weathered. Petricioli Fig. 23.

1770—Fragment of neck and body region of small vessel, probably a bottle. "Black" (very dark purple) glass; lightly weathered. No. G 64. Possibly like Petricioli Fig. 13.


1762—Chunks of dense, crystalline, dark red mineral described as cinnabar.

1763—Cone-shaped chunk of dense, whitish material with gray and yellowish brown stains, described as carbonate of lead.

Figure references are to the preceding article by Sofija Petricioli.

This article was published in the Journal of Glass Studies, Vol. 15 (1973), 93–97.

1. For a general discussion of the scientific investigation of early glasses see: R.H. Brill, "The Scientific Investigation of Early Glasses," Proceedings of the Eighth International Congress on Glass, London, 1968, pp. 47–68.

2. For information on oxygen isotopes in early glasses see reference cited in note 1 and, R.H. Brill, "Lead and Oxygen Isotopes in Ancient Objects," Philosophical Transactions of the Royal Society of London, A, Vol. 269 (1970) pp. 143-164. The same material appears in The Impact of the Natural Sciences on Archaeology, edited by Allibone, London: The British Academy, 1970.

3. For information on lead isotopes in early glasses see the reference cited in notes 1 and 2 and also: R.H. Brill, "Lead Isotopes in Ancient Glass," Annales de 4e Congrés des Journees Internationale Verre, Liège, 1967, pp. 255-261.

4. R.H. Brill, "Incipient Crizzling in Some Early Glasses," Bulletin of the American Group—The International Institute for Conservation of Historic and Artistic Works, Vol. 12 (1972), No. 2, pp. 46–47.

5. A plausible explanation of why some glasses are unstable could lie in the fact that whenever glassmaking is in an ascendancy, the glassmaker is seeking technological means for improving the quality of his wares. In attempting to make more nearly colorless glass, cristallo in the case of Venice, it must have been recognized at some point that the purification of raw materials would have a beneficial effect. When leaching and recrystallization processes were introduced for the purification of plant-ash alkalis, this would have had the effect of lowering the concentration of calcium and magnesium salts in the resulting alkali, because these elements have a tendency to be present in less soluble chemical forms than the soda and/or potash. Consequently, when the purified alkali was used for making glass the resulting product would be deficient in lime and magnesia. Such glasses then would have unstable compositions until the "glass technologists" of the day became aware of this effect and compensated for it by adding more lime in the form of some other ingredient. The net result in the interim then would have been the production of purer, possibly more nearly colorless glass, which would have been susceptible to crizzling in the years to follow. There is ample evidence, in fact, that this is just what did happen in the descriptions of Neri as to the purification of alkali as well as in the story of Ravenscroft's perfection of his lead-glass compositions.
This effect would have been magnified if a greater percentage of alkali were used in the historic formulation, as might have been done in Venice in an effort to produce a glass with a longer working range, that is, a glass which would remain softened for a longer time and allow the glassmaker to perform the more elaborate manipulations required to make complicated decorative forms.

6. There are many editions of Antonio Neri's L'Arte Vetraria. The first edition was printed in Florence in 1612; the first English translation, by Merrit, was in 1662.

7. R.H. Brill, W.R. Shields and J.M. Wampler, "New Directions in Lead Isotope Research," Application of Science in Examination of Works of Art, edited by W.J. Young, Museum of Fine Arts, Boston, 1973, pp. 73-83.

Published on June 13, 2013