CHANGES IN GLOSS OF MARBLE SURFACES AS A RESULT OF 
METHYLCELLULOSE POULTICING 
Julie A. Lauffenburger, Carol A. Grissom and A. Elena Charola 
Abstract-A range of aqueous methylcellulose mix- 
tures was applied to polished Carrara marble tiles for 
four hours, 24 hours, and until the poultices had dried 
to films. The specular gloss of each test area was 
measured before the poultice was applied and after its 
removal. When removed wet, poultices made with 
hard tap-water or deionized water containing a large 
excess of calcium carbonate or bicarbonates produced 
little or no change in surface gloss. On the other hand, 
poultices made with deionized water alone produced 
significant decreases in specular gloss; this effect was 
slowed but ultimately not diminished by the addition 
of ammonium hydroxide to pH 9. After only four 
hours, AB57 made with EDTA produced severe dis- 
solution of the marble, while AB57 made without 
EDTA did not appear to alter the surface. Mechanical 
removal of calcite crystals occurred when poultices 
were allowed to dry to films. When bulking agents 
were used, there was little or no reduction in surface 
gloss. Comparison of the effects of poultices made 
with methylcellulose of three different molecular 
weights showed no discernible difference. 
1 Introduction 
Cellulose ethers have been widely used by con- 
servators as adhesives, consolidants and thick- 
eners [1]. For treatment of stone, they have been 
used to thicken aqueous cleaning poultices. One 
major method has been to incorporate cellulose 
ethers in poultices which remain on the stone for 
a period of time, but are removed while still wet. 
Compositions range from methylcellulose poul- 
tices with no special additives to complex mix- 
tures such as AB57, used especially in Italy for 
removal of 'black crusts' from calcareous stone 
[2]. Sodium carboxymethylcellulose is specified 
for use in AB57, but some conservators have 
substituted methylcellulose in the formulation to 
avoid retention of sodium salts by the stone. A 
second major method for using cellulose ethers 
has recently gained popularity around the world 
as a more effective means of stone cleaning: thick 
aqueous mixtures of sodium carboxymethyl- 
Received 23 July 1991 
Studies in Conservation 37 (1992) 155-164 
cellulose [3, 4] or methylcellulose [5-7] are 
applied and then removed after drying to films. 
These cellulose ether mixtures often include plas- 
ticizers to aid in film removal, as well as other 
additives such as surfactants. Both methods are 
generally regarded as safe for use on reasonably 
sound surfaces and, while a few analyses of the 
effects of actual treatments on stone have been 
published [6-8], no systematic study has been 
reported. The purpose of this paper is to deter- 
mine by means of reflectance measurements the 
effects on polished marble of a range of methyl- 
cellulose poultices, removed both wet and dry. 
2 Experimental 
2.1 Specular gloss measurements 
Specular gloss measurements were taken on the 
surface of marble test tiles using a Dr Lange 
Labor-Reflektometer. ASTM Standard Test 
Method for Specular Gloss (Designation 
D523-85) [9] was used as a basis for testing. 
Gloss measurements are based on the ASTM 
standard of a highly polished, plane, black glass 
with a refractive index of 1 567 for the sodium D 
line, which is assigned a gloss value of 100. For 
this particular instrument, a black glass working 
standard with a gloss value of 99-3 was used. 
Initial testing with a 60? geometry indicated that 
the surfaces had a high gloss more suitable for a 
20? geometry, and the latter was therefore used 
throughout. 
Multiple specular gloss readings were 
recorded before and after treatment for each test 
area. Each time, five gloss readings were taken at 
different points for the experiments of set 1, 
while 15 gloss readings were taken at one centi- 
meter intervals in an easily reproducible pattern 
for the second and third sets of experiments. 
Control readings were taken on untreated 
areas adjacent to each type of poultice at the 
same time as readings were taken on poulticed 
areas before treatment and after all treatments 
had been completed. For the first set of experi- 
155 
Julie A. Lauffenburger, Carol A. Grissom and A. Elena Charola 
Table 1 Poultice formulations 
Exp. no. Viscosity in cps. Poultice formulation 
(METHOCEL A type) per 400ml water 
1/1 400 (4C) Tap water 
1/2 1500 (15C) Tap water 
1/3 400 (4C) Deionized water (DI) 
1/4 1500 (15C) DI 
1/5 400 (4C) DI + NH40H (pH 9) 
1/6 1500 (15C) DI + NH40H (pH 9) 
1/7 400 (4C) DI + excess CaCO3 
1/8 1500 (15C) DI + excess CaCO3 
1/9 400 (4C) AB57 without EDTA [11]: DI + 12g NH4HCO3 + 20gNaHCO3 
+ 2-5ml Triton X-100 (10% solution) 
1/10 1500 (15C) AB57 without EDTA 
2/1 400 (4C) Tap water (pH 6-5) 
2/2 1500 (15C) Tap water (pH 6-5) 
2/3 400 (4C) DI 
2/4 1500 (15C) DI 
2/5 400 (4C) DI + NH40H (pH 9) 
2/6 1500 (15C) DI + NH40H (pH 9) 
2/7 400 (4C) DI + 8ppm CaCO3 (pH 9-2) 
2/8 1500 (15C) DI + 8ppm CaCO3 (pH 9-2) 
2/9 400 (4C) AB57 without EDTA 
2/10 1500 (15C) AB57 without EDTA 
2/11 400 (4C) AB57 with EDTA [2]: DI + 12g NH4HCO3 + 20g NaHCO3 
+ 2-5ml Triton X-100 (10% solution) + lOg disodium EDTA 
3/3 1500 (15C) Tap water (pH 6-5) 
3/4 1500 (15C) Tap water + fumed silica 
3/5 1500 (15C) DI + 15ppm CaCO3 (pH 9-2) 
3/6 400 (4C) Tap water (pH 6-9) 
3/7 4000 (4M) Tap water (pH 6.9) 
Viscosity in cps.: viscosity in centipoises for a 2% aqueous solution at 20?C; pH measured with a Model 404 Orion 
Research/Specific Ion Meter 
ments, two control areas were measured for each 
type of poultice and there were 20 control runs 
in total. For the second and third sets, one area 
was measured for each type of poultice and there 
were 16 control runs in total. 
2.2 Selection of marble test tiles 
Carrara marble tiles were selected for the experi- 
ment because Carrara marble is commonly used 
for sculpture. For the first set of experiments, 
white Gioia Carrara tiles were purchased from 
Morris Tile of Tuxedo, Maryland, and for the 
two subsequent sets of experiments, from Sita 
Tiles of Capitol Heights, Maryland. The samples 
measured 300 x 300 x 18mm and had been 
commercially prepared and polished. Inconsis- 
156 
tencies in gloss and texture were noted among 
tiles to the extent that reflectance means varied 
before treatment from 41-3 to 84-8 gloss units. 
Variability in surface within single tiles is 
reflected in the spread of standard deviations. 
This is most evident in the third set of experi- 
ments, for which the tiles had highly irregular 
surfaces. 
2.3 Choice and preparation of poultices 
Aqueous mixtures prepared for the experiments 
are listed in Table 1. For most of the experiments 
Dow METHOCEL premium-grade methyl- 
cellulose powders of two different viscosities 
were used: METHOCEL A4C (400 centipoises) 
and A15C (1500 centipoises); METHOCEL 
Studies in Conservation 37 (1992) 155-164 
Changes in gloss of marble surfaces as a result of methylcellulose poulticing 
Table 2 Specular gloss measurements for poulticed areas 
Poultice MC Exp Removed wet Removed wet Removed dry 
no. at 4 hours at 24 hours at 2-3 days 
n XB SB XA SA XB SB X s X B S XA sA 
Tap water 
DI 
4C 1/1 
2/1 
3/6 
15C 1/2 
2/2 
3/3 
4M 3/7 
4C 1/3 
2/3 
15C 1/4 
2/4 
DI + NH4OH 
(pH9) 
DI + CaCO3 
excess 
8ppm 
excess 
8ppm 
15ppm 
DI + AB57 
w/o EDTA 
w/ EDTA 
4C 1/5 
2/5 
15C 1/6 
2/6 
4C 1/7 
2/7 
15C 1/8 
2/8 
3/5 
4C 1/9 
2/9 
15C 1/10 
2/10 
4C 2/11 
5 
15 
15 
5 
15 
15 
15 
5 
15 
5 
15 
5 
15 
5 
15 
5 
15 
5 
15 
15 
5 
15 
5 
15 
15 
73-8 
80-3 
62-9 
76-3 
79.4 
1-5 
2-9 
2-2 
1.9 
2-8 
74-1 
82-9 
63-4 
71-8 
78-0 
2-2 
2-1 
2-1 
1-6 
3-8 
62-2 1-9 59-3 1-4 
78-1 
71-3 
62-6 
59'6 
59.3 
81-0 
63-2 
77-1 
51-9 
82-3 
54-6 
78-6 
66-2 
54.3 
69-8 
55-8 
70-1 
69-6 
1'0 
7-1 
1-7 
4-2 
4-5 
3.7 
1-3 
6-1 
2-0 
3-1 
2-3 
2-5 
84 
4-2 
4-6 
24 
4-4 
9.7 
65'4 
63-1 
47-5 
53-5 
54-2 
77-1 
62-0 
74-1 
52-7 
73-0 
54.5 
69-0 
51-6 
57-0 
69-6 
55.4 
69-9 
4-1 
1-2 
7'0 
1-0 
4-1 
3-6 
2-9 
1-3 
4-5 
2-7 
4-0 
2-8 
3-1 
8-4 
3.7 
2-3 
2-9 
4-5 
0-6 
51-5 
72'7 
68'8 
53-9 
77-4 
6-8 
4-4 
4-5 
6-0 
2-8 
52'4 
73-8 
72'4 
53-2 
76-8 
5.5 
3-6 
4-6 
6-8 
2-7 
71-1 6-5 67-6 2-6 
57-3 
72-7 
69-5 
71'4 
49-8 
63'4 
52-0 
59'8 
59-5 
68'6 
60-8 
71-2 
84'8 
63-5 
57-1 
68'4 
55-0 
79'2 
7-8 
1-0 
3-8 
1-3 
2.4 
9.4 
1-0 
4-0 
3-4 
1-9 
0-9 
2-1 
1-9 
1 2 
6-1 
4.3 
3-9 
3-6 
40-0 
54-6 
48-0 
53'4 
29'7 
51-7 
35-7 
42-7 
56-5 
52-6 
58'8 
50-6 
65'8 
60-0 
54-2 
67'4 
56-7 
3.4 
4-5 
1-5 
1-0 
2-7 
0'8 
6-0 
1-2 
2-4 
2-2 
2-5 
2-0 
34 
2-5 
2-9 
5-4 
1-5 
4.7 
0.4 
15C 3/4 15 41-3 6-1 35-2 6-5 
MC: METHOCEL A type; n: number of pairs of observations; XB: mean of specular gloss measurements before treatment; 
XA: mean of specular gloss measurements after treatment; SB: standard deviation of mean before treatment; SA: standard 
deviation of mean after treatment. 
A4M (4000 centipoises) was used for one experi- 
ment. Viscosity is proportional to molecular 
weight: A4C and A15C are considered to have 
relatively low molecular weights as conservation 
materials, while A4M is relatively high. METH- 
OCEL A4C and A15C were substituted for 
sodium carboxymethylcellulose in the AB57 
mixtures (see Table 1), and Triton X-100 for 
Desogen, which is not available in the USA. 
In most cases additives were first mixed in 
water and the gels were then prepared by adding 
Studies in Conservation 37 (1992) 155-164 
5% methylcellulose using the 'dispersion in hot 
water technique' [10] outlined by Dow. In a few 
cases the 'dry blending technique' [11] was used 
with greater success because of the high propor- 
tion of added solids compared to METHOCEL 
powder, e.g., for some AB 57 poultices; however, 
this change of technique is not believed to have 
influenced the results. 
2.4 Application and removal of poultices 
The poultices were individually applied to 
157 
73-8 
81'4 
82-1 
48-8 
76'4 
70'2 
80-8 
49-1 
77-2 
73-3 
74'6 
61-1 
70-7 
50-7 
71-5 
60-7 
78-5 
61-1 
77-3 
66-4 
65-5 
70-9 
67-7 
63-3 
84-7 
6'4 
3-1 
1-6 
2-0 
5'1 
2'6 
1-'7 
0'5 
3-6 
4-3 
7-4 
2-3 
3-2 
8-9 
4-1 
0-8 
3-2 
0-8 
4-8 
10-1 
2-7 
4-5 
2-6 
7-8 
4.7 
44-1 3-8 
81-8 2-9 
42-5 7.4 
32-0 2-6 
46-0 3-8 
27-3 4-0 
33-6 7.5 
31-1 2-8 
45-2 2-0 
52-6 6-8 
45'1 2-7 
38-5 5-3 
46-1 3-6 
27-4 9-3 
43-3 3-0 
58-4 2-1 
51-1 3-4 
56-3 2-2 
51-0 2-0 
56-7 7-0 
65-6 4-1 
70-8 4-5 
65-6 3.4 
62-5 7-3 
3-1 0.4 
DI + SiO2 
Julie A. Lauffenburger, Carol A. Grissom and A. Elena Charola 
Table 3 Specular gloss differences before and after poulticing 
Poultice MC Exp n Removed wet Removed wet Removed dry 
no. at 4 hours at 24 hours at 2-3 days 
AD SSD AD SSD AD SSD 
Tap water 
Deionized 
water 
(DI) 
DI + NH4OH 
(pH 9) 
DI + CaCO3 
excess 
8ppm 
excess 
8ppm 
15ppm 
DI + AB57 
w/o EDTA 
DI + AB57 
w/ EDTA 
4C 
4C 
4C 
15C 
15C 
15C 
4M 
4C 
4C 
15C 
15C 
4C 
4C 
15C 
15C 
4C 
4C 
15C 
15C 
15C 
4C 
4C 
15C 
15C 
1/1 
2/1 
3/6 
1/2 
2/2 
3/3 
3/7 
1/3 
2/3 
1/4 
2/4 
1/5 
2/5 
1/6 
2/6 
1/7 
2/7 
1/8 
2/8 
3/5 
1/9 
2/9 
1/10 
2/10 
5 
15 
5 
5 
15 
15 
15 
5 
15 
5 
15 
5 
15 
5 
15 
5 
15 
5 
15 
15 
0-3 40 09 131 29-7 
2-6 26 1.1 41 0-4 
0-5 2-2 36 46 39-6 
4-5 37 07 136 16-8 
1-4 3.4 06 28 30-4 
42-9 
2-9 1-7 (3-5) 50 47-2 
12-7 
8-2 
15-1 
6-1 
5-1 
3-9 
1 2 
3-0 
08 
9-3 
0-1 
9-6 
14-6 
5 
15 
5 
15 
2-7 
0-2 
04 
0-2 
23 
7-1 
30 
42 
87 
3-4 
28 
5-4 
50 
3-6 
5-4 
28 
8-5 
84 
3-7 
5-7 
4-5 
17-3 
18-1 
21-5 
18-0 
20-1 
11-7 
16-3 
17-1 
3-0 
16-0 
20 
20-6 
19'0 
3-5 
29 
1-0 
1-7 
13-5 
1-3 
5-9 
2-1 
3-8 
7-9 
23 
3-2 
6-1 
22 
3-3 
2-8 
22 
4-7 
5-8 
68 
4.4 
18-0 
32-0 
20-7 
29-5 
22-6 
24-6 
23-3 
28-2 
23 
27-4 
4-8 
26-3 
9-7 
0'1 
0-1 
2-1 
08 
4C 2/11 15 65-5 6-9 75-8 2-6 81-6 3.4 
DI + fumed 
silica 
15C 3/4 15 61 6-4 
MC: METHOCEL A type; n: number of pairs of observations; AD: absolute difference of before and after means (Xb - Xa); 
SSD: statistically significant difference using a two-tailed ttest at a 99% confidence level [tV/(sB2 + sA2)/n]; bold type indicates 
that the AD > SSD + 1 gloss unit. 
75 x 150mm sections on the marble tiles after 
the surface had been wiped with ethanol, using 
lint-free cotton pads. A rubber spatula was used 
to apply the poultices to a thickness of 
approximately 6mm, and the tiles were left 
in a laboratory with controlled environment 
(approximately 50% RH and 21?C). For each 
mixture, a poultice was removed wet with a rub- 
ber spatula at four and 24 hours, and a third 
poultice was allowed to go to dryness (between 
two and three days later). Most of the dry poul- 
tices pulled away from the stone by themselves, 
but those which did not were carefully peeled 
away. No residue of solid material incorporated 
in the mixtures, such as fumed silica and calcium 
carbonate, was apparent on the surface after the 
Studies in Conservation 37 (1992) 155-164 
11-2 
30 
5-4 
4.9 
4-5 
3-4 
5-5 
4-3 
29 
12-1 
56 
8-7 
3-4 
19-3 
3-6 
3-4 
3-3 
3-5 
3-7 
8-8 
7-4 
4-5 
64 
7-6 
158 
Changes in gloss of marble surfaces as a result of methylcellulose poulticing 
100 
90 - 
80- 
70- 
60 - 
50 - 
40- 
30 - 
20- 
10 - 
0 .. . 
. . . .. .. 
Tap DI DI + 
water water NH OH 
en 
c. 
c 
4- 
o (u 
u 
0 
c 
a) 
CLe 
IX 
DI + DI + DI + DI + 
Bppm Excess AB57 AB57 CaCO CaCO w/EDTA w/o EDTA Ca3 
2/3 2/ 2n / a2/11s Di DI DI+ DI + DI+ DI + + 
water NH4 OH 8ppm Excess AB57 AB57 
CaCO3 CaCO w/EDTA w/o EDTA 3/ ET 
Figure I Percent decrease in gloss after poulticing for 
24 hours. 
poultice had been removed. All marble surfaces 
were wiped with tap-water and then with ethanol 
and allowed to dry before readings were taken 
with the glossmeter. 
3 Results 
Data are presented in Tables 2 and 3, and statis- 
tical procedures are detailed in Appendix 1. 
Mean values and standard deviations for the 
specular gloss measurements are given in Table 
2. The absolute difference (AD) of the mean 
values measured before and after treatment are 
given in Table 3, for comparison with the statis- 
tically significant difference (SSD) calculated for 
each test pair. A significant change in surface 
gloss is indicated when the AD is greater than 
the SSD. When the AD is more than one gloss 
unit greater than the SSD, it is shown in bold 
type. 
The magnitude of significant change in mea- 
sured gloss gives an indication of the physical 
degree of change that has occurred. In general, 
it increased with application time of the poultice 
and corresponded with visual observations 
regarding the amount of change on the surface. 
Changes in gloss after treatment with a repre- 
sentative selection of poultices are illustrated in 
Figures 1 and 2. Percent decreases in gloss for 
surfaces from which poultices were removed wet 
after 24 hours are shown in Figure 1, and for 
surfaces from which poultices were removed 
when dry in Figure 2. In these figures, each bar 
represents the difference between mean 
reflectance measurements before and after the 
Figure 2 Percent decrease in gloss when poultices dry 
to films. 
specified treatment, divided by the mean mea- 
surement before [(Xb - Xa)/Xb]. Note that cal- 
culations for these graphs do not take into 
account standard deviations. 
In general, the glossmeter seemed slightly 
more sensitive than the human eye in 
differentiating the relative degree of change. 
However, whenever a significant change was 
measured it could also be perceived by the 
human eye, and vice versa. 
Instrument repeatability was excellent, as 
indicated by measurements of control areas 
made at the same time as those for poulticed 
areas before and after treatment. With one small 
exception, there was no significant difference in 
the control measurements at the 99% confidence 
level, and generally the AD was much less than 
the SSD. 
4 Discussion 
The following discussion is based on a com- 
parison of data listed in Table 3 and visual obser- 
vations made when the experiments were carried 
out. All decreases in gloss can be attributed to 
two different mechanisms: chemical dissolution 
of the marble and mechanical removal of calcite 
crystals. Chemical dissolution is indicated by 
change in gloss measurement for poultices 
removed wet after four and 24 hours. Chemical 
dissolution may continue until poultices dry, but 
mechanical removal of calcite apparently domi- 
nates when there is substantial change in gloss 
measurement between 24 hours and dryness. 
Figures 1 and 2 illustrate these effects. 
Studies in Conservation 37 (1992) 155-164 
U) 
CU 
,- 
0 J- 
c 
a) 
-, 
C) C-1 
159 
Julie A. Lauffenburger, Carol A. Grissom and A. Elena Charola 
An interesting observation for poultices 
removed after four and 24 hours is the difference 
between poultices prepared with deionized water 
and with common tap-water. When made with 
deionized water, significant differences in surface 
gloss were measured even for poultices removed 
after four hours. This can be attributed to the 
solubility of calcite in water (about 1-4 x 
10-4M or 14ppm [12]; in the presence of CO2 
this value increases [13]). After 24 hours this 
effect was more pronounced, as would be 
expected from a dynamic situation that has not 
reached equilibrium. When made with tap- 
water, no significant changes in gloss were mea- 
sured for poultices removed wet. This can be 
attributed to the common ion effect provided by 
the relatively high average concentration of cal- 
cium ion (1 0 x 10-3 M) as well as other ions in 
local tap-water [14]. 
The addition of ammonium hydroxide to the 
deionized water to adjust the pH to 9, which in 
theory should decrease attack on the marble sur- 
face, seemed to slow the etching effects of the 
pure deionized water. No significant change in 
gloss measurement occurred after four hours, 
but after 24 hours there was a substantial reduc- 
tion in gloss, similar to that with the pure 
deionized water. This can be attributed to the 
fact that at the higher pH the initial dissolution 
rate is diminished, but over longer periods dis- 
solution is not affected. The dissolution mech- 
anism of calcite between pH 6 and 9 is mainly 
independent of pH, and the rate of dissolution is 
surface controlled [15]. 
Powdered calcium carbonate was added to 
poultices made with deionized water to see 
whether adding material of the same com- 
position as the marble itself would have a 
beneficial effect. Three different amounts were 
used: 8ppm, 15ppm, and a large excess of cal- 
cium carbonate. Poultices made with 8 and 
15ppm produced reflectance decreases similar to 
deionized water, apparently because the solu- 
tions were not saturated. While the concen- 
tration of calcium ions in a saturated solution 
of calcite is normally given as 1-4 x 10-4M 
(14ppm), empirical testing showed that 8ppm 
was the maximum amount of calcium carbonate 
that readily dissolved and that, when 15ppm was 
added to the water, a slight excess of solids 
remained visible because of slow dissolution. By 
contrast, poultices made with a large excess of 
160 
calcium carbonate did not produce any change 
in gloss when removed wet, indicating that no 
chemical dissolution had taken place. If the solu- 
tion were not fully saturated, any attack on cal- 
cium carbonate would be concentrated on the 
high specific surface area of the large excess of 
dispersed powder rather than the marble sur- 
face. Poultices made with a solution of sodium 
and ammonium bicarbonates with excess bicar- 
bonate solids and a detergent (the AB57 mixture 
for mural paintings) did not change the gloss of 
the marble surface. The fact that no etching 
occurred can be attributed to the high concen- 
tration of bicarbonate ions. Under these condi- 
tions, the equilibrium with carbonate ions is 
shifted so that the dissolution of calcium carbon- 
ate is inhibited. 
A poultice made with the same bicarbonate/ 
detergent mixture plus EDTA (the AB57 mix- 
ture for stone cleaning) etched the polished 
marble surface severely compared with all other 
mixtures. The effect was noted after four hours 
of poulticing and increased very little thereafter. 
The rapidity of the etching is attributed to the 
rate at which EDTA chelates calcium at pH 8-5. 
Etching of the polished marble was expected, yet 
the rapidity and severity were surprising. 
The second mechanism of damage, mechani- 
cal action, was evident when poultices dried to 
films. Small crystals adhering to the underside of 
the dry poultice could be felt by touch and were 
visible to the naked eye throughout, with higher 
concentrations at the perimeter of the film. The 
crystals were confirmed as calcite by scanning 
electron microscopy accompanied by energy 
dispersive X-ray analysis. In the most extreme 
cases, losses were visible on the stone at low 
magnification, corresponding to concentrations 
of crystals on the poultices. No difference was 
observed between poultices which detached 
themselves and those which were peeled away. 
When comparisons are made between the data 
for poultices removed wet at 24 hours and those 
removed after drying, substantial changes 
confirm mechanical damage (Figures 1 and 2). 
This is most notable in the case of tap-water 
poultices, which produced significant decreases 
in gloss when allowed to dry, although little or 
no change occurred when the poultices were 
removed wet. 
The addition of bulking and other materials 
eliminated the mechanical removal of calcite 
Studies in Conservation 37 (1992) 155-164 
Changes in gloss of marble surfaces as a result of methylcellulose poulticing 
crystals as the films dried. Most notably, a 
methylcellulose poultice containing fumed silica 
and tap-water produced no statistically 
significant change in surface gloss on drying, 
although a decrease occurred when the poultice 
was made with tap-water alone. Poultices made 
with excess calcium carbonate and bicarbonate 
salts (AB57 mixtures), which apparently acted as 
bulking materials, also produced little or no 
change on drying. Compared to poultices made 
with bicarbonates alone (data not reported), 
detergent-containing AB57 produced more slip- 
pery poultices and irregular films. 
Comparison of data for poultices made with 
the two low-molecular-weight methylcelluloses 
showed similar patterns of significant change 
after treatment. Tap-water poultices made with 
methylcelluloses of all three molecular weights 
also produced similar results. 
5 Conclusions 
Measurement of marble surfaces with a gloss- 
meter proved to be a reliable and relatively easy 
method of assessing the effects of a range of 
aqueous methylcellulose poultices used for 
cleaning marble. Polished marble tiles provided 
a good standard from which to evaluate changes 
in gloss resulting from poulticing. Subtle 
changes were both observed and measured 
which might not have been apparent on eroded 
surfaces. In actual practice, the effects of poul- 
ticing on surfaces might be somewhat different. 
For example, a methylcellulose film does not 
adhere well to a dirty surface and may therefore 
be less likely to remove calcite mechanically; 
conversely, poultice cleaning might do more 
damage to a deteriorated surface than to pol- 
ished marble. Nevertheless, we believe that our 
results are relevant to conservation practice. 
Deionized water, frequently used for cleaning 
marble, decreased surface gloss from four hours 
onwards, but dissolution was averted by includ- 
ing large excesses of calcium carbonate or bicar- 
bonates. Poultices made with small amounts of 
calcium carbonate produced approximately the 
same degree of etching as deionized water alone. 
Excellent results with a large excess of calcium 
carbonate suggest that its addition to a poultice 
made with deionized water might be easier and 
more reliable than the common practice of soak- 
ing marble chips in deionized water before use. 
Studies in Conservation 37 (1992) 155-164 
A bicarbonate/detergent mixture (AB57 for wall 
paintings) also did not change the reflectance of 
the polished marble, and use of ammonium 
bicarbonate is recommended as an alternative to 
calcium carbonate. However, use of sodium 
bicarbonate is not recommended because of the 
possible retention of sodium salts by the stone. 
The addition of ammonia to deionized water to 
adjust the pH to 9 did not appear to reduce the 
solubility of the marble, but the use of ammonia 
in stone cleaning mixtures may be warranted for 
dirt removal. 
Poultices made with local tap-water and 
removed wet did not cause dissolution of the 
marble, apparently because of its high calcium 
content. However, we do not recommend the use 
of tap-water to clean marble because of the 
many other undesirable ions in solution, includ- 
ing sodium, iron, copper and chloride. 
Results showed severe etching by the AB57 
mixture containing EDTA (AB57 for stone 
cleaning), and this mixture is emphatically not 
recommended for use on marble, with the excep- 
tion of a surface which has been altered to a 
'black crust' principally composed of calcium 
sulfate, soot and silica. A marble surface after 
treatment with the EDTA mixture may appear 
'whiter' but this may be attributed to increased 
scattering of light as a result of etching, not to 
cleaning. 
Mechanical removal of calcite crystals 
occurred when test poultices were allowed to 
form films, unless the mixtures had been 
modified with excess carbonates, excess bicar- 
bonates, or fumed silica. In an effort to reduce 
mechanical removal of calcite crystals by dry 
poultices, conservators have typically modified 
mixtures with plasticizers such as glycol; greater 
flexibility and reduced adhesiveness of poultice 
films are reported [3, 6, 7]. Although we did not 
test plasticizers, we would expect them to dimin- 
ish or eliminate mechanical damage. Conser- 
vators have also reported that low-molecular- 
weight methylcelluloses produce more flexible 
and less damaging films than high-molecular- 
weight ones [6, 7] but no difference was observed 
in our limited experiments. More work needs to 
be done, particularly with methylcelluloses of 
4000 centipoises and above. 
When using methylcellulose poultices, several 
other factors should be considered which were 
not tested in this project. Methylcellulose poul- 
161 
Julie A. Lauffenburger, Carol A. Grissom and A. Elena Charola 
tices left to dry are extremely efficient in 
removing dirt from porous surfaces, and the 
degree of cleaning desired should be established 
by preliminary testing. Conservators have also 
noted, particularly when application is uneven, 
that residues of methylcellulose film can be 
tedious to remove. Finally, although no residues 
of admixed materials such as calcium carbonate 
were evident on the polished test tiles-nor are 
they expected, because of the adhesive properties 
of the methylcellulose gel and the manner in 
which it dries-each surface to be cleaned should 
be first tested with the poultice mixture to be 
used, and examined for any deleterious results. 
In conclusion, we believe that a methyl- 
cellulose poultice made with deionized water, 
properly modified with calcium carbonate or 
ammonium bicarbonate and materials to reduce 
the adhesion of a dried film, can be safely used 
for cleaning sound marble surfaces. 
Acknowledgements 
The authors would like to thank Lawrence Becker, 
currently objects conservator at the Virginia Museum 
of Fine Arts, for making the first observations that led 
to this study while he was at the Metropolitan 
Museum of Art; Robert J. Koestler, who is in charge 
of the scanning electron microscopy facilities in the 
Department of Objects Conservation of the Met- 
ropolitan Museum of Art, for performing analyses on 
preliminary test samples; Timothy J. Vitale, senior 
paper conservator at the Conservation Analytical 
Laboratory, for his assistance throughout the project, 
especially with computerization of data; and the Sam- 
uel H. Kress Foundation for their support of this 
interlaboratory collaboration. The experiments were 
carried out by Julie Lauffenburger for a research 
project during her postgraduate conservation intern- 
ship at the Smithsonian Institution Conservation 
Analytical Laboratory. 
Suppliers of materials 
Calcium carbonate C-64: Fisher Scientific, 711 Forbes 
Ave., Pittsburgh, PA 15219, USA. 
METHOCEL powders: Dow Chemical Company, 
Midland, MI 48640, USA. 
Triton X-100 wetting agent: Fisher Scientific, 711 
Forbes Ave., Pittsburgh, PA 15219, USA. 
Appendix: Statistical methods 
The means of specular gloss measurements (X) and 
162 
the standard deviations (s) were calculated by the 
glossmeter, according to the formulae: 
zxi 
/(X- Xi)2 X = i= s= x x , where i= 1, 2, ... n n n- 1 
observations. 
To evaluate whether a statistically significant change 
occurred after treatment, a two-tailed t test was 
applied at a 99% confidence level. The paired t test for 
two samples with equal sample sizes is: 
X1-X2 t = 
l/n (B2 + SA2) 
where the subscripts (B and A) represent before and 
after treatment values and the degrees of freedom ()) 
for the test were nB + nA- 2. At a 99% confidence 
level, the t values from statistical tables are as follows: 
for Set 1 experiments with 5 pairs, i.e. 10 obser- 
vations, t99(4,,=8) = 336 
and for Sets 2 & 3 experiments with 15 pairs, i.e. 
30 observations, t99o(<4= 28) = 276 
By algebraic manipulation of the formula given above 
we obtain: 
X - X t/(SB2 + A 2) X X t n 
and we designate 
(s2 + sA2) 
nt = SSD, the statistically significant n 
difference. 
This quantity (SSD) can then be compared to the 
measured difference of the before and after treatment 
mean (AD) so that a simple visual inspection of the 
data shows when a treatment produced a significant 
effect. Note that for convenience we have used the 
two-tailed t test, although we have reported the abso- 
lute difference of the means. 
The t test is strictly applicable only when mea- 
surements possess the same inherent standard devi- 
ation, i.e., when the data belong to the same 
population as indicated by the F test for variance 
homogeneity. The F test compares the ratio of the 
variances from two sets of data (s2), 
2 
F = 
S2 
with the critical value of F given for the F distribution 
from statistical tables. If the calculated value is 
smaller than the critical value of F, the two sets cannot 
be said to arise from different populations. 
The F values for variance homogeneity were calcu- 
lated at the 5% level, and most samples passed the 
test. Rejections can be readily explained. All samples 
poulticed with the AB57 mixture containing EDTA 
produced reflectance measurements with very low 
Studies in Conservation 37 (1992) 155-164 
Changes in gloss of marble surfaces as a result of methylcellulose poulticing 
standard deviations after treatment and were thus 
rejected; apparently the surfaces had been evenly 
etched by the EDTA. A few rejections were also 
obtained for samples poulticed to dryness, apparently 
because of the erratic mechanical removal of calcite 
crystals. Since it is well known that the two-tailed t test 
is resistant to non-extreme heterogeneity of variance, 
we chose to apply this t test uniformly to all our 
experiments. 
References 
1 FELLER, R. L., and WILT, M., Evaluation of 
Cellulose Ethers for Conservation, The Getty 
Conservation Institute, Marina del Rey, CA 
(1990). 
2 URBANI, G., ed., Problemi di Conservazione, 
Editrice Compositori, Bologna (1973) 57, 
341-343. 
3 MEL'NIKOVA, E. A., and LEBEL, M. N., 'Applica- 
tion of polymer films removing surface con- 
taminations from sculptures made of different 
materials' in ICOM Committee for Conser- 
vation 5th Triennial Meeting, Zagreb (1978) 
78/10/8. 
4 AGRAWAL, O. P., SINGH, T., and JAIN, K. K., 
'Study and conservation of spotted red sand- 
stone of Mathura' in Case Studies in the Con- 
servation of Stone and Wall Paintings, IIC, 
London (1986) 165-169. 
5 LEHMANN, J., 'The methodology for the cleaning 
and desalting of stone objects in Goluchow 
Castle Museum' in ICOM Committee for Con- 
servation 8th Triennial Meeting, Sydney (1987) 
487-491. 
6 GRISSOM, C. A., POWER, T., and WEST, S., 
'Methyl cellulose poultice cleaning of a large 
marble sculpture' in Proceedings VIth Inter- 
national Congress on Deterioration and Conser- 
vation of Stone, Torun (1988) 551-562. 
7 GOLDBERG, L. A., 'A fresh face for Samuel Gom- 
pers: methyl cellulose poultice cleaning', Jour- 
nal of the American Institute for Conservation 
28 (1989) 19-29. 
8 LAZZARINI, L., ASMUS, J. L., and MARCHESINI, 
L., 'Lasers for the cleaning of statuary: initial 
results and potentialities' in 1st International 
Symposium on the Deterioration of Building 
Stones, La Rochelle (1972) 89-94. 
9 Annual Book ASTM Standards, 06.01 (1987). 
10 How to Prepare a Solution of METHOCEL 
Cellulose Ethers, product literature, form no. 
192-958-88, Dow Chemical USA. 
11 MORA, P., MORA, L., and PHILIPPOT, P., Conser- 
vation of Wall Paintings, Butterworths, Lon- 
don (1984) 342. 
12 WEAST, R. C., ed., CRC Handbook of Chemistry 
and Physics, CRC Press, Boca Raton, FL 
(1984), B-82. 
13 BATHURST, R. G. C., Carbonate Sediments and 
their Diagenesis, Elsevier Scientific Publishing 
Company, Amsterdam (1976) 231. 
14 Data taken from routine analysis of water for the 
Potomac Water Filtration Plant, revised Feb- 
ruary 1990. 
15 SJOBERG, E. L., 'A fundamental equation for 
calcite dissolution kinetics', Geochim. Cosmo- 
chim. Acta 40 (1976) 441-447. 
JULIE LAUFFENBURGER, BA in art history, Cornell 
University (1985); MA and Certificate of Advanced 
Study in Art Conservation, State University College 
of Buffalo, New York (1989). She is currently Assis- 
tant Objects Conservator at The Walters Art Gallery. 
Author's address: The Walters Art Gallery, 600 North 
Charles Street, Baltimore, MD 21201, USA. 
CAROL GRISSOM, BA in art history, Wellesley College 
(1970); MA in art conservation, Oberlin College 
(1974). Since 1984 she has been chief objects conser- 
vator at the Conservation Analytical Laboratory of 
the Smithsonian Institution. Author's address. 
CAL/MSC, Smithsonian Institution, Washington, DC 
20560-0001, USA. 
A. ELENA CHAROLA, PhD in chemistry, University of 
La Plata, Argentina (1974). She is currently a consul- 
tant for ICCROM and coordinator for the Easter 
Island Program at the World Monuments Fund. 
Author's address: 500 East 77th Street, Apt. 1239, New 
York, NY 10162, USA. 
Resum--On a applique, pour polir des carreaux de 
marbre de Carrare, des solutions aqueuses de methyl- 
cellulose pendant quatre heures, puis 24 heures, puis 
jusqu'a ce que les cataplasmes sechent et pelliculent. 
Le brillance speculaire de chaque surface traitee a ete 
mesuree avant et apres l'application du cataplasme. 
Quand ils restent humides, les cataplasmes con- 
fectionnes avec de l'eau dure du robinet ou de l'eau 
demineralisee contenant du carbonat de calcium ou 
des bicarbonates en exces produisent peu ou pas de 
difference dans le brillant de surface. D'autre part, les 
cataplasmes prepares avec de l'eau demineralisee 
seule produisent une decroissance significative de 
l'effet speculaire; cet effet est ralenti, mais finalement 
pas diminue si l'on additione de l'hydroxyde d'am- 
moniaque jusqu'a obtention du pH 9. Apres seule- 
ment quatre heures, l'AB57 melange d'EDTA pro- 
duit une dissolution importante du marbre, tandis qui 
l'AB57 sans EDTA ne parait pas alterer sa surface. 
L'usure mecanique des cristaux de calcite se rencon- 
Studies in Conservation 37 (1992) 155-164 163 
Julie A. Lauffenburger, Carol A. Grissom and A. Elena Charola 
tre chaque fois que le cataplasme est laiss6 jusqu'au 
sechage complet. Lorsque des agents de charge sont 
employes, on constate une reduction legere ou nulle 
au niveau du brillant de surface. Enfin des cata- 
plasmes prepares avec du methylcellulose de trois 
masses molaires diff6rentes ne montrent pas de 
difference visible. 
Zusammenfassung-Eine Reihe von Methylcellulose- 
praparaten wurde in wassriger Losung an polierten 
Carrara Marmorplatten erprobt. Ihre Anwendung 
erfolgte in Form von Gelen, die vier Stunden, 24 
Stunden oder solange einwirkten, bis sie zu einem 
Film abgetrocknet waren. Vor und nach der Anwen- 
dung wurden Glanzmessungen auf den behandelten 
Oberflachen durchgefuhrt. Gele, die mit hartem 
Leitungswasser oder deionisiertem Wasser, das Cal- 
ciumcarbonat oder 
-hydrogencarbonat im iiberschuB 
enthielt, angesetzt worden waren, veranderten- 
sofern sie in noch nassem Zustand abgenommen 
wurden-den Oberflachenglanz nur wenig oder iiber- 
haupt nicht. Wurde auf einen Carbonatzusatz ver- 
zichtet, nahm der Glanz deutlich ab. Hieran anderte 
selbst ein Zusatz von Ammoniumhydroxid (bis zu 
einem pH von 9) nicht viel. Bereits nach vier Stunden 
loste AB57 mit einem Zusatz von EDTA die Ober- 
flache des Marmors stark an: Wurde auf einem 
Zusatz von EDTA verzichtet, konnten dagegen keine 
Veranderungen beobachtet werden. Trockneten die 
Gelschichten zu einem Film ab, konnte beobachtet 
werden, daB dieser bei seiner Abnahme Calcitkristalle 
aus der Oberflache mitriB. Falls Dickungsstoffe 
zugesetzt waren, wurde der Oberflachenglanz nur 
geringfugig oder gar nicht beeintrachtigt. Das Mole- 
kulargewicht der Methylcellulose scheint keinen mer- 
klichen EinfluB auf die Wirksamkeit der Gele in der 
geschilderten Anwendung zu haben, was an Hand 
dreier verschiedener Molekulargewichte ausgetestet 
wurde. 
Studies in Conservation 37 (1992) 155-164 164