CRIZZLING - A PROBLEM IN GLASS CONSERVATION
Robert H. Brill, The Corning Museum of Glass, Corning, N.Y. 14830, USA
1. INTRODUCTION
Crizzled glasses are familiar to all curators or collectors who have handled much early English or European glass,
the word 'crizzled' being used to describe the appearance of certain glasses which have a diminished transparency
owing to very fine surface crazing. A typical example is illustrated in Figure 1. Under the microscope, a network
of fissures can be seen clearly on the surfaces of such glasses, as shown in Figure 2. On those glasses which are
most severely afflicted, small bits of glass may have spalled away. The effect is the result of unstable chemical
compositions which render the glasses susceptible to chemical attack by water, usually in the form of water
adsorbed from the atmosphere.
Crizzling is most often associated with glasses made in the seventeenth and' eighteenth centuries, but earlier
examples may exist, and nineteenth century examples are by no means rare. The glasses come from rather
widespread areas encompassing England, Western Europe (including Venice), Scandinavia, North America, and
China. There are numerous pieces, which are known to have been crizzled for a long time, throughout all of the
large collections dealing with glass of this period.
Crizzling is one variant of what has often been called 'sick glass' - a persistent, but not very precise term
referring to various forms of surface deterioration. These range from a slightly cloudy or dulled appearance, to
crizzling, to occasional occurrences of 'weeping glasses' or 'sweating glasses', which repeatedly generate slippery
surface films or droplets if exposed to humid conditions. All are related in that they are manifestations of
hydrolytic attack by atmospheric water, but we will be concerned here primarily with the crizzling phenomenon.
Perhaps the best known crizzled glasses are those made by George Ravenscroft, the English glassmaker credited
with being the first to capitalize on the properties of lead glasses. During his early experimentation, Ravenscroft
made some pieces which crizzled within a few months of manufacture, a deficiency which he quickly attempted
to correct [11 .
Several treatments have been suggested for crizzled glasses. Most of these involve impregnations with
polymeric materials which consolidate the surface and restore transparency and a glassy reflectance. The first
application of this type of treatment known to the author was that by Professor J. Arvid Hedvall of Uppsala
University [21. Actually, however, none of the treatments we have seen have yielded entirely satisfactory results.
Because storage under proper conditions and careful handling seem to forestall further damage to already crizzled
glasses, the outlook of The Corning Museum has been to postpone treatment until a completely safe and
satisfactory treatment has been perfected. This is still our outlook as far as already crizzled glasses are concerned,
since these objects seem to be stable for the present.
However, it has recently become known that there is a related and most urgent problem concerning crizzling
[31. There exists a category of glasses in early stages of crizzling, which have taken up water over the centuries
but which do not exhibit very obvious signs of deterioration. Only when they are handled and turned around
under correct lighting conditions can the telltale silvery rays of beginning crizzling be seen. These glasses are of
the same general origins and dates as the already crizzled glasses. They have become hydrated and 'equilibrated'
over the centuries with humid environments. When brought into centrally-heated galleries or apartments, they
may be subjected abruptly to very dry conditions during the winter months. For example, in our galleries the
relative humidity frequently drops into the 20 - 30% range during midwinter, and upon occasions has even fallen
below 15% on extremely cold days. Under such conditions these hydrated glasses become dehydrated, and the
crizzling fissures, which were not very evident to begin with, open up and propagate a network of hairline cracks
across the surface. Thus, these glasses can become severely crizzled - and this may happen in the course of only a
few months. We have observed one serious example of this effect in our museum and suspect that a half-dozen or
so other objects have deteriorated to a less marked degree over the years. In addition, we have heard several
reports of similar occurrences throughout North America, Europe, and Britain.
While most of the occurrences of this accelerating deterioration which have come to our attention involve
vessel glasses, the author has examined one roundel of fifteenth century German glass (a glass bearing painting
and silver-stained regions) which has developed an effect akin to crizzling over the past decade or so. Since it had
hung adjacent to a heating register during that period, it seems very likely that a dehydration process was the
cause of its accelerated deterioration.
By way of definition we have coined the term 'incipient crizzling', to describe this dangerous metastable state
of deterioration. It designates glasses which are hydrated and in the initial stages of crizzling (the damage being
apparent only under careful inspection) and which are susceptible to a rapid surface deterioration should they be
exposed to dehydrating conditions. The object of the research which follows was to arrive at recommendations
for the safe storage of glasses with incipient crizzling so as to prevent their disintegration into a severely crizzled
state. We are less concerned for the moment with already crizzled or 'weeping' glasses.
Based mainly upon intuition and a few preliminary observations, we recommended earlier that glasses which
show incipient crizzling should not be exposed to low relative humidities, that the temperature and humidity
should be maintained at relatively constant levels, and that localized high temperatures, such as could result from
installations with dramatic lighting, should be strictly avoided. Since then, some further progress has been made
towards making firm suggestions as to safe exhibition and storage conditions based upon actual laboratory
studies. The general approach to this research has been:
1. To obtain fundamental chemical information on susceptible glasses.
121
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
Table 1. Analyses of some crizzled glasses
Venetian(?) Venetian(?) French French French Silesian.. American American American Jamestown"
c.1709 c. 1709 c. 1750 1600-50 1725-50 1710-30 1790-1810 19th C. late 19th C. 1971
colourless colourless colourless grayish colourless colourless colourless colourless light blue green
457 461 449 1050 1498 4011 1487 1488 1630 1823
Si02 dig ~ 76 ~ 67 76.0 70.7 ~ 74 ~ 73.5 ~ 69 ~ 77 ~ 81 ~ 67
Na20 a 0.13 0.99 0.89 17.7 0.29 0.39 2.62 5.00 12.7 17.5
CaO a 1.30 0.30 0.22 1.53 2.87 1.61 3.19 2.72 1.96 4.66
K20 a 18.7 18.4 18.7 3.42 19.9 16.6 24.2 12.7 2.24 6.82
MgO a/s 0.04 0.06 ~ 0.1 ~ 0.1 0.22 0.03 0.20 0.32 0.12 0.10
~ ~ ~ ~Al20 3 a/s 0.1 0.05 0.5 0.5 1.95 1.03 0.55 1.51 0.44 2.61
..... ~ ~
N Fe203 a/s 0.05 0.1 ~ 0.05 ~ 0.1 0.03 0.002 0.05 0.01 0.03 0.28
N
MnO a/s ~ 0.01 ~ 0.05 0.45 0.38 0.51 0.04 < 0.01 O.OOX O.OOX 0.01
PbO q/s 2.22 12.8 0.002 0.01 nf 5.88 < 0.01 nf 1.5 0.05
As20 s q 1.66 0.24 ~ 0.5 ~ 0.2
Notes:
"All glasses except 1823 and 4011 a - atomic absorption. Also found in varying trace levels:
are heavily crizzled. 1823 is q - quantitative, various methods. Ti02, CuO, CoO, Ag20, BaO, SrO,
hygroscopic but uncrizzled; s - emission spectrography. LGO,Rb20,B203,Zr02'P20s
4011 shows incipient crizzling. g - gravimetry. contents uncertain but generally
d - Si02 by difference. less than 1.0%. Where no As20 s
Analyses by R. H. Bell and ~ - denotes spectrography. given, values are uncertain but
C. A. Jedlicka of Lucius Pitkin, nf - sought but not found. less than 1.0%.
Inc., NYC, and P. B. Adams of
Corning Glass W~rks.
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
2. To characterize the crizzling process physically and chemically.
3. To evaluate the chemical durability of the susceptible glasses relative to other types of glass.
4. To duplicate crizzling in the laboratory on both synthetic glasses and actual pieces of early glass.
5. To identify the primary factors controlling the chemical mechanism of crizzling. (Although it would be
desirable to establish the details of the mechanism in the sense of a true reaction and kinetic study, this is
not feasible just now.) .
6. To arrive at recommendations as to how to prevent susceptible glasses, those showing incipient crizzling,
from deteriorating into severe crizzling.
2. CHEMICAL COMPOSITIONS OF CRIZZLED GLASSES
It has generally been held that crizzling is caused by the attack of atmospheric water, possibly augmented by the
uptake of carbon dioxide, and that the susceptibility of the glasses to hydrolysis is the result of a deficiency in
lime, the most common stabilizer in glasses. While we also held that view - and still do - it seemed worthwhile to
confirm it by chemical analyses. Therefore, chemical analyses were made of several early crizzled ·glasses, eight of
which are reported in Table 1. To simplify the table, only major and minor constituents are reported. Data on
trace elements have been excluded. Also included in the table are an example of a glass showing incipient crizzling
(no. 4011) and one of a modern glass (no. 1823). Both of the latter were used in experiments aimed at
duplicating crizzling in the laboratory. (Descriptions of the glasses analysed are given in the Appendix.)
For working within the system Na20:CaO:Si02 (soda-lime-silica glasses), a lime content of less than about 4%
by weight would be regarded as very unstable by present-day standards. Even among ancient glasses, lime
contents are ordinarily greater than 5.5-6.0%, and many of them are heavily weathered. It will be seen, however,
that only two of the early glasses in Table 1 (nos. 1050 and 1630) are actually soda-lime glasses. The others are in
the K20:CaO:Si02 system (potash-lime-silica glasses) or are glasses containing some lead (PbO). The minimum
concentration of lime required to stabilize a potash glass may not be just the same as the 4% level cited above for
the soda-lime glasses, but it is certainly not very different from that level.
It is quite apparent from the table that all of these glasses are indeed seriously deficient in lime and should be
expected to be prone to hydrolysis. Only nos. 1487 and 1823 might be judged to have borderline lime contents
and these both have such excessively high total alkali contents that they would also be expected to be very
unstable. (A total alkali content of 20% or greater would not be tolerated by modern commercial standards, and
ancient glasses with that much alkali are usually very heavily weathered.) Nos. 461 and 4011 are the only glasses
of this group which really qualify as lead glasses, and while we have not yet completed our literature search on
durabilities within the system K2 0:PbO:Si02 , the potash levels are substantially greater than would be used in
modern manufacture.
The explanation of the compositional imbalance of these glasses is of some interest to the historian of
chemical technology. Whenever glassmaking is in an ascendency, the glassmaker seeks technological means for
improving the quality of his wares. In attempting to make more nearly colourless glass, 'cristallo' in the case of
Venice, it was recognized 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 lqwering the concentrations of calcium, magnesium, and aluminum 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 purified alkali was used for making glass, the resulting product would have been deficient in lime, magnesia,
and alumina. 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 additional lime in the form of some other ingredient. The
net result in the interim would have been the production of purer and possibly more nearly colourless 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 regarding the purification of alkali as well as in the story of
Ravenscroft's perfection of his lead-glass compositions [41. We have under way some experiments with barilla
plant ashes to verify this.
This effect would have been magnified if a greater percentage of alkali were used in the basic 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.
In connection with the chemical compositions, it is worth commenting upon a common misunderstanding.
One often hears the remark that it is the pink glasses which tend to crizzle, and there are, indeed, a great many
crizzled glasses which have a pronounced pinkish colour. However, this pink colour is confined to the crizzled
surface of the glass and the objects were initially colourless. Since manganese was often used as a decolorant, or
might even have been present in some of the plant ashes, many early glasses contained manganese. When these
glasses are attacked by water and alkali is leached out, the structures are opened up and the manganese becomes
oxidized, yielding a pink colour in the crizzled layer. That this is true has been verified by the fact that when a
crizzled layer is removed from one of these pink objects, the remaining glass is perfectly colourless.
3. ELECTRON MICROPROBE ANALYSES
Electron microprobe analyses of five of the glasses appearing in Table 1 were carried out by W. T. Kane and T. W.
Bierweiler of Corning Glass Works. The analyses were made on small samples of glass removed from each of the
objects. The samples were mounted in cross-section so that both the crizzled layer and the unaffected glass were
exposed for analysis. Individual analyses of both regions on all samples were made for silicon, sodium,
123
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
potassium, calcium, and lead (where present). Concentration maps for each element on each glass were obtained
so that it is possible to see how the concentration of each element varies over an area including the glass, the
crizzled layer, and their interface. The maps covered areas of about 100 x 25 p.m. Profiles for each element were
also constructed to see how their concentrations varied along a continuous path from the unaffected glass out
through the crizzled layers. The concentration maps and profiles established the following points, to which are
added our interpretations.
1. In the four glasses which were heavily crizzled (nos. 449,461, 1050, and 1498) the sodium and potassium
were very greatly depleted in the crizzled layer, indicating that these two elements had been thoroughly
leached out of the glasses and probably remove!i to a large extent from the surface by washing in the past.
2. The silicon appears somewhat richer on a percentage basis in the crizzled layer than in the unaffected glass.
The enrichment is the result of the depletion of the alkali, and it is likely that the total quantity of silica
present was about the same before and after the crizzling attack took place.
3. The calcium and lead contents appear to have been unaffected by the leaching process because they, as
within the silica, are less soluble than the alkalis. (It would have been difficult, however, to detect changes
in the calcium level because the concentration was low to begin with.)
4. A void space could be detected in the four crizzled specimens which clearly marked the separation of the
crizzled layer from the glass beneath. The analysts likened the appearance of the profiles to a 'step
function', showing extensive depletion of the alkali outside the fissure in the crizzled layer. In two instances
a second fissure was revealed beneath the crizzled layer, but this underlayer does not show complete
depletion of alkali, only a tendency towards lower concentration. This indicates that the hydration and
leaching reactions have proceeded onward into the glass beyond the first fissure separating the depleted
layer from the glass. This serves as a reminder that we are dealing with a dynamic system, and that the
process does not stop when a flake of crizzling products appears.
5. In a few places, highly localized concentrations of alkali showed up indicating that alkali had redeposited
within the fissures. We believe we have also seen such deposits during microscopic examinations of various
crizzled glasses. This would be the fate of alkali which had been exchanged by water but had never been
washed off the glass during its history of human handling.
6. In sample no. 4011, a glass which showed only incipient crizzling, the alkalis have not been entirely
removed from the glass, and the profile does not show the step-like appearance of the crizzled glasses. The
alkali concentration shows a gradual depletion gradient sloping. down from the unaffected glass to the
surface. The profile has the character of a diffusion front. There is some evidence of the beginning of a
fissure within the hydrolysing layer. (In anticipation of suggestions which might be prompted by the
demonstrated presence of some alkali within this layer, we regard it as doubtful that the process could be
reversed successfully and the glassy structure of the layer reconstituted by reintroducing alkali.)
7. The overall thicknesses of the reaction zones as seen through a light microscope measure between 45 and
100 11m, but the microprobe profiles yield additional information. From the spacing of the fissures in the
profiles, the thicknesses of the separated layers (the crizzled layers) can be estimated. They range from 25
to 55 11m. In three of the four 'already crizzled' glasses, the layer which had separated is totally depleted in
alkali. In the fourth glass, the front of the reaction zone is still within the crizzled layer, and the
undersurface of the crizzled layer is quite rich in alkali. In the three instances where the reaction zone had
advanced beyond the fissure separating the crizzled layer, the depletion/hydration fronts had penetrated
between 40 and 80 11m from the 'original' surface. On the piece showing only incipient crizzling, the
depletion/hydration front had penetrated 23 11m into the glass.
If one takes the thicknesses of the crizzled layers and ages of the glasses at face value, the hydration/alkali
depletion rate of these five glasses averages out to about 0.3 11m per year. Compared to ancient glasses which are
heavily weathered (those suitable for layer-counting), this is not as rapid a rate of hydration. The latter glasses
hydrate at a rate of about 0.5-3.0 11m per year. However, lightly weathered ancient glasses, such as those showing
iridescence, hydrate more slowly than the crizzled glasses. Iridescent Roman and Islamic glasses hydrate at a rate
of only about 0.1 11m per year or less.
4. SYNTHETIC CORROSIBLE GLASSES
In order to learn more about the crizzling process a series of synthetic glasses was prepared. The compositions of
the glasses, which are given in Table 2, were patterned like those obtained from the chemical analyses of four of
the early glasses reported in Table 1. The glasses were prepared by A. A. Erickson of Corning Glass Works and his
co-workers. The compositions given in Table 2 are theoretical values calculated from batch compositions, but
chemical analyses of the glasses match them closely. These glasses were used for several types of experiments.
124
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
Table 2. Compositions of synthetic early glasses
Glasses susceptible to crizzling
XS XT XU XV PP
Si02 73.5 75.5 75.0. 65.0. 75.8
Na20 22.0. 0..1
K20 20..0. 20..0. 20..0. 18.7
CaO 2.0. 2.0. 1.0. 1.0. 1.30.
MgO 1.0. 1.0.
Ah 0 3 1.0. 1.0. 1.0. 1.0. 0..2
PbO 3.0. 13.0. 2.25
MnO 0..5 0..5
As20 3 1.65
Prototype 1050. 449 457 461 457
(French) (French) (Venetian) (Venetian) (Venetian)
Table 3 contains compositions of synthetic mediaeval stained glasses also prepared by Dr Erickson. There are
seventeen such glasses in the table. Their prototypes were groups of glasses, analysed by the author, from York,
Niirnberg, and the Zerek Camii [5]. Series were also prepared to test various compositional effects on properties
related to durability and weathering characteristics. The glasses represent variations in concentrations of
Na20/K20, P20 S , MgO, and CuO; the effects of colour on both stable and durable glasses; and the duplication of
an extremely interesting group of lead-containing mediaeval glasses. Some of these glasses have been used by Mr
Bettembourg* in his research. We have carried out weathering and durability tests on several of these synthetic
stained glasses, but they will not be discussed here, since they are beyond the scope of this paper.
5. WEATHERING AND DURABILITY TESTS
Samples of the synthetic corrosible glasses (XS, XT, XU, and XV) were subjected to three types of testing. These
were a miniature powder test, weathering cabinet exposure and autodaving. All the testing was carried out by P.
B. Adams, H. V. Walters, and M. L. Nelson of Corning Glass Works using their routine testing procedures.
Since the details of the individual tests are rather complicated, we shall only outline the procedures here, but
further information can be obtained by writing to Mr Adams. ** The powder test consists of a water treatment of
powdered glass at 9QoC for four hours. The weight of alkali extracted is then determined by chemical analysis.
The weathering cabinet test consists of exposing small pieces of the glass to an environment of 98% relative
humidity at 5Qo C, and making periodic inspections for surface damage, which is then expressed in a standard
terminology. The autodaving was carried out at 121°C, at 15 psi, for various lengths of time, with the glasses
immersed in the liquid phase. Small polished plates of the glass were used for autoclaving. (Other variations of the
conditions for the tests were also used for exploratory purposes.)
The first two tests were used for evaluation of the durabilities of the glasses. The conditions parallel standard
glass-testing procedures, so the results can be used for comparison with more familiar types of glasses.
The autodaving was undertaken to look for chemical cracking and for the investigation of hydrationdehydration
effects. For the latter purpose, autodaving served to introdu~e water into the glasses. Different
samples of the glasses were then stored under different relative humidity conditions (at room temperature), to see
what levels of relative humidity might successfully inhibit the dehydration which produces crizzling. (As will be
shown below, this hydration-dehydration treatment produces a kind of deterioration wHich is indistinguishable
from the crizzling seen on actual early glasses.) The results of the tests are summarized in Table 4.
Some of the findings of the tests are listed below, accompanied by our interpretations.
1. In the powder test, glasses XS and XU, the soda glass and the potash glass with the light lead concentration,
showed very poor resistance to hydrolysis. Glass XV, the more heavily-leaded glass, shows a considerably
better resistance; and XT, the potash-lime glass, is still better. In fact, the durability of XT approaches that
of some of the poorer present-day commercial glasses. It is somewhat surprising that XT holds up as well as
it does, but it indicates that the addition of even the small increment of CaO (from 1.0. to 2.0.%) over XV
brings about a marked increase in stability. In comparing the durabilities of XS and XT, it should be noted
that although the concentrations of soda and potash are not very different on a weight-percentage basis, in
terms of mole percent the alkali content of the potash glass is markedly less (13.7% vs. 21.5%). This
undoubtedly accounts for much of its increased durability. The glasses appear to rank as follows, in order
of increasing durability: (very poor) XS ~ XU «:; XV <XT (fairly stable).
2. In the weathering cabinet exposures, XS was by far the poorest, and XT again the best. While some
commercial bulb glasses do not fare much better than XS in this test, container glasses are much more
edurable. The glasses appear to rank as follows in order of increasing durability: (very poor) XS <XU <XV
<XT (fairly stable).
*Laboratoire de Recherches des Monuments Historiques, Chateau de Champs/Marne, France.
**Corning Glass Works, Corning. NY 14830.. USA.
125
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
Table 3. Compositions of synthetic stained glasses
QQ QR QS RC RD RE RF VS VW vx VY VZ FE FF FG FH FI
Si02
K20
Na20
Cao
MgO
Al20 3
P20 S
58.5
8.5
3.0
16.0
7.0
1.0
4.0
55.5
18.5
0.5
19.0
2.5
1.0
1.0
66.0
1.5
14.0
10.0
1.4
2.0
60.0
20.0
20.0
54.0
18.0
18.0
48.0
16.0
-­
16.0
48.0
16.0
16.0
63.0
18.0
18.0
63.0
18.0
18.0
56.7
16.2
16.2
10.0
56.7
16.2
16.2
10.0
56.7
16.2
16.2
50.0
17.0
3.0
16.0
7.0
1.5
4.0
48.4
16.4
2.9
15.4
6.9
1.5
3.9
63.0
12.5
1.5
12.0
4.0
1.5
4.0
61.0
12.0
1.5
11.6
3.9
1.5
3.9
69.5
1.5
14.0
12.0
1.5
....N
0\ Fe203
MnO
CuO
CoO
1.0
1.0
1.0
1.0
1.0
1.0
3.0
0.1
3.0
1.0 1.0 0.9 0.9 0.9
10.0
0.5
1.0
0.5
1.0
3.0
0.1
0.5
1.0
0.5
1.0
3.0
0.1
0.5
1.0
PbO 10.0 20.0 17.0
Prototype York Niirn­
berg
Zerek
Camii
K20
base
glass
PbO
effect
PbO
effect
PhO
effect
K20
base
glass
Na2O­
K20
subst.
P20 S
effect
MgO
effect
CuO
effect
Un­
stable
Un­
stable,
blue
Durable Durable,
blue
Na2OK2
0
subst.
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
3. The autoclaving produced chemical cracking [6] in XS even after only short durations. XU and XV
developed a surface haziness, while XT showed no apparent change. The glasses appear to rank as follows,
in order of increasing durability: (very poor) XS <XU <XV ~ XT (apparently stable).
Table 4. Results of durability tests
XS XT XU XV
Powder test*
Alkali extracted
(Wt -% of glass)
Na20 0.48
K 2 0 0.057 0.71 0.16
Alkali extracted
(MmolellOO g glass)
Na20 7.7
K2 0 0.61 7.5 1.7
Weathering cabinet exposure* *
Time required to breakdown 2 wks 8 wks 4 wks 6 wks
Chemical cracking***
Days required 1 none ~8 ~15
occurs
* Sample wt ~1 gram, adjusted to constant surface area from density;
-40 to +50 mesh; 5 ml H20. Run for 4 hours at 90°C.
** 98% RH, 50°C. Breakdown =Obvious damage under inspection without magnification.
*** 2.5 x 2.5 x 0.1 cm polished plates in 285 ml H20, 95°C, 15 psi, for varying times.
6. INFRARED ANALYSES
Infrared reflectance and transmiSSIOn spectra were taken for several samples of autoclaved glasses and their
untreated counterparts. The spectra were run by H. Hoover and B. Butler of Corning Glass Works.
Reflectance spectra in the 8-12 11m region established clearly that the autoclaved glasses had been depleted in
alkali because they showed a sharper reflectance peak than the anhydrous glasses. The shift indicates less
disturbance of the Si-O-Si bonds by alkali or other modifiers. It is noteworthy that this was true even of the
glass XT, which showed no visible evidence of alteration after the autoclaving.
Infrared transmission spectra in the 2-5 11m region indicated the presence of large quantities of beta-hydroxyls,
or loosely-bound water. A sample of glass XV which had been hydrated through autoclaving was crizzled by
heating it to 135°C for 1'h hours. A spectrum of the crizzled glass fell in between the spectra for the hydrated
glass and an anhydrous control, indicating that although a significant amount of the beta-hydroxyl was removed,
a significant amount still remained after the crizzling was induced.
Infrared spectra of an autoclaved sample of XV (30 minutes at 121°C) were taken progressively as the
hydrated surfaces were polished away. By this means it was possible to estimate that this particular hydration
process had hydrated this glass to a depth of approximately 20 11m, somewhat less than the thicknesses of the
crizzled layer on early glasses.
Further infrared work is being conducted in an attempt to identify the different states of binding of water
within the hydrated glasses [7].
7. HYDRATION-DEHYDRATION EXPERIMENTS
Several experiments have been carried out with a view to duplicating crizzling in the laboratory. Among these
were preliminary experiments which demonstrated that by heating certain types of actual historic glasses (but not
all types) to temperatures between 100° and 150°C, severe surface crizzling can be produced. These glasses were
of diverse origins, including Roman glass from Karanis, some specimens of mediaeval stained glasses and
eighteenth century vessel fragments. Other investigators have had similar experiences [8]. With this in mind,
several pieces of synthetic glasses having corrosible compositions were held over water or immersed in water at
85°C for varying times and then dehydrated. It was found that the dehydration could be accomplished either by
warming the glass to 50-65°C or by placing it in a desiccator. By various combinations of experimental
conditions, several such glasses developed surface crizzling identical to that observed on early objects of crizzled
glass. It might be added that some of the synthetic mediaeval stained glasses behaved similarly under these
treatments.
We next set out to determine what levels of relative humidity, at room temperature, would inhibit the
dehydration of hydrated glasses so as to arrest the development of crizzling. Experiments were carried out with
the four synthetic glasses XS, XT, XU, and XV. In addition, two actual early objects were sacrificed for this
127
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
experiment. These experiments were carried out by M. L. Nelson and H. V. Walters. A similar series of
experiments was conducted by the author in the Museum.
Small polished plates of the four synthetic glasses were hydrated by autoclaving at 121°C and 15 psi with total
immersion in water. Various autoclaving times were used. After drying in air, the hydrated glasses were placed in
glass chambers (actually desiccator jars) which maintained constant internal relative humidities established by
saturated salt solutions in the bottoms of the·,chambers. The relative humidities were monitored with an electric
hygrometer. The humidities were: 1%, 25-30% (ambient laboratory conditions), 40%, 72%, and 100%. (The
samples in the 100% chamber soon became covered with droplets of condensed water.) Although the samples
selected for the experiment were autoclaved for different times, all were similar in that after the autoclaving they
showed no surface damage detectable by the unaided eye. After varying intervals of time, extending from a few
days up to several weeks, many of the glasses developed crizzling. At the time of writing (March 1975), some
samples still have not crizzled. The results are summarized in Table 5.
Table 5. Effect of storage conditions on hydrated glasses
Percentage of glasses which ultimately crizzled *
Glass l%RH 25-35%RH 40%RH 72%RH 100%RH
XS 100% 17%** 0 0 o
XT 100 40 40 0 o
xu 100 100 100*** 50*** o
XV 100 100 12 0 o
% crizzled 100% 65 38 12 o
irrespective
of composition
Duration of experiment to date: 11 Sept. 1974 - 17 Jan. 1975
Total number of samples used: R:110
* The times required to develop crizzling varied widely. Some samples
crizzled withiri a few days, others required several weeks.
** Of the balance, they might not all have been hydrated.
*** Hydration might have been excessive.
Some other experiments
Glass Treatments Results
XS 1. Imm. water, 85°C, 78 hrs Little surface change
2. Storage at: O%RH Crizz. after few days
31% No change after 35 mos.
75% No change after 35 mos.
100% No change after 35 mos.
3. Storage at: 0% All show crizz. after 9 hrs
XS, XT, 1. Imm. water, rm temp., 35 mos. XS, XU - Incip. crizz.
XU,XV XT, XV - No change
2. Storage at: O%RH XS, XU, XV - Crizz. after 9 hrs
XT - Start crizz.(?) after 2 wks
1823 1. Imm. water, rm temp., 35 mos. No change
2. Storage at: O%RH Start crizz. after 90 min.
Crizz. after 16 hrs
12%RH Start crizz.(?) after 2 wks
1823 1. Storage at: 100%RH
Rm temp., 35 mos. No change
2. Second storage at: 0% Crizz. after 16 hrs
4011 1. No laboratory hydration
2. Storage at: O%RH Crizz. within 24 hrs
12% Start crizz.(?) after -about 4 wks
21 %, 34%, 52%,
70%,100% No change yet after 11 wks
A similar set of hydration - dehydration experiments was started in 1971, but most of the apparatus, glasses,
and records were destroyed in the flood which struck the Museum jn June 1972. Fortunately, there were some
survivals which now show interesting results. Samples of glass XS were hydrated by immersion in water at 85°C
128
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
for 78 hours. The samples showed little change after the hydration step. They were then stored for eleven days at
humidities of 0%, 31%,75%, and 100%. The piece stored at near-zero percent crizzled severely; the others did not
show much change. They were left in the same chambers for 35 months and little further damage occurred. They
were then transferred to a near-zero humidity chamber and within nine hours all developed crizzling.
In another experiment, pieces of glasses XS, XT, XU, and XV were given no high temperature hydration, but
just left to hydrate in distilled water at room temperature for 35 months. XS and XU developed some chemical
cracking and incipient crizzling, respectively; but XT and XV showed no surface alteration at all. However, when
placed in a near-zero humidity, all except XT developed severe crizzling within nine hours. (XT again proved to
be surprisingly durable, as it did in certain other experiments.)
The experiments carried out on the two early glasses, nos. 1823 and 4011, offer a different approach to the
problem and avoid the difficulties-posed by the use of autoclaving to hydrate the glasses.
Glass no. 4011 is an engraved Silesian goblet made in about 1710-30. The piece shows typical incipient
crizzling (Figure 3). The foot of the goblet had been broken in the distant past, and the piece was not considered
to be of outstanding historical importance. Therefore, Mr Jerome Strauss, who owned the object for some years,
donated it to the Museum for use in our experiments, feeling that the information to be gained from sacrificing
parts of the glass would prove to be more valuable than the glass itself. Samples were sawed from the rim above
the engraving. They were carefully photographed under reproducible lighting conditions, and photomicrographs
were taken to record the original state of the incipient crizzling. One small piece was used for chemical analysis,
and several of the others were placed in chambers for storage at known humidities.
Once again the humidities of the chambers were controlled by saturated salt solutions, except for a desiccating
chamber employing silica gel for a near-zero humidity, and one containing distilled water for 100% humidity. No
attempt was made to control the temperature. Instead, ambient room temperatures prevailed. The humidities, as
measured with an electric hygrometer, are: R:<O%, 12%,21%, 34%, 52%, 70%, and 100%.
It was found that the piece of object no. 4011 placed in the chamber at near-zero humidity became severely
crizzled within 24 hours (Figures 4 and 5). The piece in the chamber at 12% seems to have begun to show signs of
developing crizzling at the time of writing, which corresponds to an elapsed time of about four weeks at 12%.
None of the samples at other humidities tlave changed.
Glass no. 1823 is a small green bottle made in 1971 at the Jamestown, Virginia, factory. It was sent to the
author for examination because it developed a slippery surface feeling shortly after manufacture. After a chemical
analysis showed an excessive alkali content to be the cause, parts of this bottle were also sacrificed for
hydration-dehydration experimentation. Fragments were stored for four months at ambient conditions in
Corning and then stored in a 100% humidity chamber for 35 months. Another fragment of the same glass was
immersed in distilled water for 35 months. Neither glass showed any visible signs of surface deterioration at the
end of these hydration steps. Samples of these same fragments were then placed in a near-zero humidity chamber,
and others in a 12% chamber. Within 90 minutes of being placed in the near-zero chamber, the samples showed
beginning stages of crizzling. They were left overnight in the chamber and had become severely crizzled by the
next morning after a total of 16 hours at near-zero humidity. One of the samples in the 12% chamber appears to
be developing slight crizzling after about two weeks, at the time of writing.
It is conceded that the logic which is about to be followed is weakened by a few significant defects, but even
with these in mind it is worth pursuing. In some of the experiments outlined above, the hydration conditions
were chosen arbitrarily and were not always the same for each glass, so care must be exercised in comparing the
behaviour of one glass with another. Moreover, the autodaving used to hydrate the glasses was a rather drastic
treatment. It can be argued that the hydrated state induced by autodaving might not be the same as that brought
about by the slower 'natural' hydration produced by the action of atmospheric water vapour over centuries at
room temperature. But nonetheless, these are the only data we have, and in effect this same defect marks the
interpretation of any accelerated ageing test, so it is familiar to all conservators.
From the results in the first part of Table 5, it could be concluded that hydrated glasses of this type will
crizzle rapidly if stored at near-zero humidity, but will not dehydrate (not rapidly) at humidities of about 70% or
greater. In the range of 25-40%, many such glasses will crizzle; between about 40-70%, some glasses will crizzle.
The findings presented in the second part of Table 5 represent longer term experiments and different
hydration processes, some of which eliminate the onus of accelerated ageing. From the results of these
experiments, it can be inferred that glasses hydrated by exposure either to hot water for short times or to cool
water for a matter of years will retain their moisture for long periods when stored at humidities between
30-100%, but that they rapidly give up that moisture and become severely crizzled if transferred to a near-zero
humidity environment.
The experiment most closely simulating actual museum conditions, that involving the Silesian goblet,
demonstrated unequivocally that storage of a glass showing incipient crizzling at a near-zero humidity will
provoke severe crizzling within a matter of hours. The data on the Jamestown glass substantiate this finding. It
remains to be seen whether relatively rapid crizzling can also develop on susceptible glasses in the range of about
25-40% relative humidity, and we have set up long-term experiments to answer this question.
8. STORAGE CONDITIONS AND CASE DESIGN
From all of the foregoing, it is apparent that dehydration is the cause of the advancement of incipient crizzling
into full crizzling, and it is obvious that very low humidity conditions must be avoided if susceptible glasses are to
be protected. The next point to be established is just what constitutes an optimum range for safe storage.
It is somewhat disquieting that dehydration occurs at relative humidities as high as 40-60% and that only levels
129
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
of 70% or greater appear to hold hydration water in the experimental glasses. It is likely, however, that this is
because the water had been driven into the glasses under rather drastic conditions. It is doubtful that early glasses
with incipient crizzling dehydrate significantly in the 40-60% range. After all, such glasses must already have
spent much of their existence in humidities within that range and have not gone over to full crizzling. We are
prepared to assume, then, that the 40-60% range is safe.
Storage at very high humidities with a view towards keeping water in might then accelerate the hydration of
the glasses and contribute to their ultimate deterioration, to say nothing of the impracticalities posed by possible
mould growth, condensation, etc. We conclude, therefore, that from the viewpoint of protection against
dehydration:
1. The 70-100% range is unnecessarily high and undesirable.
2. The 40-60% range is safe.
3. The 25-40% range remains open to question and should be studied in terms of years-long storage
experiments on actual early glasses.
4. The 0-25% range is dangerous, in fact very dangerous at the lower end.
The situation, then, can be reduced to questioning whether there is anything else that can go wrong in the 40-60%
range.
In 1959 Organ and Bimson [9] dealt with a related - but not identical - problem to that being discussed
here. They recommended that for safe storage unstable glasses should be kept dry, specifying that by dry they
meant at a relative humidity below 42%. The reasoning was that at humidities below 42%, potassium carbonate
(the ultimate form of alkali leached out of potassium glasses) is not hygroscopic. This reasoning is certainly sound
for the types of glasses they were considering, that is, 'weeping' or 'sweating' glasses, and the validity of their
recommendation seems to have been borne out by actual experience. In the intervening years, large numbers of
'weeping' glasses from the collections of the British Museum were stored in cases where the humidity was kept at
or below 42%. In a recent visit to the British Museum, the author accompanied Ms Bimson and Mr Hugh Tait on
an inspection of the glass involved. It was found that few of the glasses in the cases had the slipperiness to touch
which characterizes 'weeping' glasses. (Those examples which were found to be somewhat slippery were Venetian
glasses which might have been soda glasses instead of potash glasses.)
A troublesome question then arises as to whether the dehydration rate of glasses with incipient crizzling might
become fast enough to provoke full crizzling if they are stored at humidities below 42%; or, indeed, could the
same thing happen to 'weeping' glasses? It is probably artificial to treat these as two entirely different kinds of
glasses. Mr Tait and Ms Bimson did not recognize among the glasses in the British Museum cases any which they
could say had become more deteriorated during the years of storage in those cases. However, since so many
glasses were involved, and since no photographic records of the original states of the objects were at hand, it
could not be determined with certainty whether or not any of these glasses had become more crizzled.
With this in mind, we can turn to the practical problem faced by The Corning Museum of Glass, the protection
of glasses which show incipient crizzling. There are perhaps 50 such objects in our collections. Our primary object
is to arrest any further deterioration brought on by dehydration and, if possible, to minimize the continued
uptake of moisture. (It is recognized that improving storage conditions will not correct the damage already
inflicted upon the glass.) Ideally, one woul? assume the safest condition to be a humidity level where a steady
state could be established under which neither the hydration nor dehydration reactions would progress. Since it is
not likely that we could now arrive at a single level which would satisfy steady-state conditions for so many
different glasses, we have elected to err on the side of hydration, feeling that this offers less hazard to the glasses
than dehydration. It might further be guessed that, to ensure that dehydration will not occur, the humidity
should be in approximately the range to which the glasses had become accustomed in their early histories.
For these reasons, we have decided to aim initially for case control at the 45-55% relative humidity range, with
provisions for sufficient adaptability to allow the humidity level to be shifted somewhat in either direction. (For
example, if persistent slipperiness developed upon the surfaces, we might want to drop slightly below the 42%
level.)
A committee of conservators was convened to offer advice on our system design and case construction. The
committee consists of Robert Feller, Lawrence Majewski, Susanne Sack, and Nathan Stolow. The designs have
not yet been completed, but several pivotal decisions have been made. The resultant system contains three main
components: the cases, controlling devices, and a regenerating unit.
The cases themselves will be sealed, with access through one door. One face, at least, will be glass or plexiglas
for exhibition purposes. Proper gasketing will be very important. The case will contain some humidity-buffering
material to offset variation caused by changes in temperature. Since the temperature of our galleries can be
controlled, no special temperature control will be introduced in the cases.
Special care will be given to lighting the cases. The lighting will be external and properly situated so as not to
cause a heat build-up, and infrared absorbing shields will be used. The deleterious effects of spotlighting objects
for dramatic exhibition are appreciated by conservators but not by all curators or collectors. The effects of heat
build-up can be very serious on certain types of glass, for example, glasses with incipient crizzling, painted
surfaces, or weathering products. In an experiment conducted in our museum, flat plates of glass were placed on a
wooden pedestal painted black and lit with two 12-volt automobile lamps at a distance of about one metre. The
temperatures of the glasses were measured by thermistor probes placed in hollowed-out depressions in the glasses.
At one point within a piece of glass, a temperature of 67°C was measured: Temperatures of this level could
seriously endanger glasses of the sorts mentioned and should never be tolerated.
The cases will be connected to controlling units containing trays of silica gel which will have been equilibrated
previously to the target relative humidity, perhaps 45% to start with. It was decided that silica gel would be used
130
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
as the controlling medium because of its inertness, low sensitivity to temperature change, regenerative qualities,
and adaptability to different levels of humidity control. Circulating devices with proper safety controls will be
used to assure adequate circulation between the control units and the cases. A remote recording thermohygrograph
will monitor the relative humidity and temperature through sensing probes. A separate re-equilibrating unit
in our workroom will be used to regenerate the silica gel as required.
It is expected that, if the sealants are good and the cases are not opened frequently, the system will remain
stable for some months at a time. One of our major goals is to install a system which will operate with a minimum
of maintenance and will be as free as possible of annoyances caused by failures of mechanical or electrical parts.
For this reason, which may appear paradoxical to some readers, the Museum has elected not to install a
fully-automated system. Our belief is that the more elaborate a system becomes mechanically, the more things
there are to go wrong.
It is hoped that, through proper design and construction of these cases, our glass with incipient crizzling will
be safely preserved throughout the coming years while we attempt to devise treatments which will restore
chemical stability to these glasses, instead of just maintaining a delicately-balanced metastable state.
ACKNOWLEDGEMENTS
The author thanks all of the persons cited in the text for their contributions of experimental work. In particular,
P. B. Adams, H. V. Walters, and M. L. Nelson have devoted much time to the planning of the work and
interpretation of results. Special thanks are due to those who donated glasses which could be used in our
experiments. D. P. Lanmon, Curator of European Glass at The Corning Museum of Glass, and P. N. Perrot and K.
M. Wilson, both formerly of the Museum, are especially thanked for their concern and assistance over the course
of several years. This research was partially funded by a grant from the National Endowment for the Arts.
REFERENCES
1. There are numerous accounts of Ravenscroft's early experimentation. One is: Hartshorne, A., 'Antique
Drinking Glasses', Brussel and Brussel, New York, 1968, pp.24Q-241. The term 'crizzle' was applied to glasses
as early as Ravenscroft's time. The common usage of the word was to describe roughening of surfaces, or the
crackling of sunburnt skin on the face or hands, an apt parallel to the appearance of these glasses.
2. Hedvall, J. A., Jagitsch, R. and Olson, G., Uber das Problem der Zer-Storung antiker Glaser, 'Transactions of
Chalmers University of Technology' No.118, Gothenburg, Sweden, 1951. We are not attempting here to
compile a complete bibliography on this subject.
3. Brill, R. H., The scientific investigation of ancient glasses, in 'Proceedings of the Eighth International Congress
on Glass', London, 1968, pp.47-68; Incipient crizzling in some early glasses, 'Bulletin of the American Group
of the International Institute for Conservation of Historic and Artistic Works', 12 (1972).
4. There are many editions of Antonio Neri's 'L'Art Vetraria'. The first edition was printed in Florence in 1612;
the first English translatio~, by Merrit, was in 1662. R. M. Organ has also described this chemical effect [9].
5. Brill, R. H., Scientific studies of stained glass, 'Journal of Glass Studies', 12 (1970),185-192.
6. Adams, P. B., Crack propagation of annealed glass during exposure to water, in 'Tenth International Congress
on Glass', No. 11, The Ceramic Society of Japan, Kyoto, 1974, pp.16-23.
7. Other experiments have also been attempted or are as yet incomplete, so they are not discussed here. These
include various approaches to determining water contents, and electron microscopy.
8. Bimson, M. and Werner, A. E., The danger of heating glass objects, 'Journal of Glass Studies', 6 (1964),
pp.148-150.
9. Bimson, M. and Organ, R. M., The safe storage of unstable glass, 'Museum News', 46 (1968), pp.39-47.
131
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
APPENDIX: CATALOGUE OF SAMPLES
449 Wineglass, French, c. 1750.
Colourless glass, pink surface, heavily crizzled.
457 Beaker, Venetian(?), c. 1709.
Colourless glass, heavily crizzled.
Donated by Gudmund Boesen; Rosenborg Collection, Inv. no. 109.
461 Cut pendant from candelabrum, Venetian(?), c. 1709.
Colourless glass, heavily crizzled.
Donated by Gudmund Boesen; Rosenborg Collection, Inv. no. 27-43.
1050 Wineglass, French, 1600-50.
Grayish glass, heavily crizzled.
Donated by William J. Young.
1487 Pane in bookcase, American, 1790-1810.
Colourless glass, heavi~y crizzled.
Donated by Dwight Lanmon.
1488 Pane from frame of painting, American, nineteenth century.
Colourless glass, heavily crizzled.
Donated by the Henry Francis duPont Winterthur Museum.
1498 Wineglass, French, 1725-50.
Colourless glass, pink surface, heavily crizzled.
Donated by the Henry Francis duPont Winterthur Museum.
(Figure 1 in text.)
1630 Rim fragment, petal and loop design, American, late nineteenth century.
Light blue, heavily crizzled.
•
1823 Reproduttion of early bottle, Jamestown, Virginia, 1971.
Green glass, 'weeping'.
Donated by James Haskett.
4011 Engraved goblet, Silesian, 1710-30.
Colourless glass, incipient crizzling.
Donated by Jerome Strauss. (His no. S538)
(Figures 3, 4, and 5 in text.)
132
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
Figure 1. Typical example of early crizzled glass. French,
1725-50, no. 1498 in this study.
Figure 2. Photomicrograph of surface of wineglass in
Figure 1. Magnification approx. 6X.
Figure 1.
Figure 2.
133
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org
Figure 3. Typical example of glass with incipient crizzling.
Silesian, 1710-30, no. 4011 in this study.
Donated by Mr Jerome Strauss for
experimentation.
Figure 4. Rim sample of glass in Figure 3. No hydration
treatment. Traces of incipient crizzling can be
seen along lower left edge.
Figure 5. Same sample as in Figure 4. Severe crizzling was
produced by room-temperature storage at nearzero
RH within 24 hours.
Figure 3.
Figure 4. Figure 5.
134
Brill, R. H. "Crizzling - A Problem in Glass Conservation." Conservation in Archaeology and the Applied Arts, Stockholm Congress, 1975. pp. 121-134.
© International Institute for Conservation of Historic and Artistic Works (IIC). Used with permission.
Rakow Research Library, The Corning Museum of Glass - http://www.cmog.org