To Scrape or Not to Scrape?
By Patrick Baty
The following article appeared in Traditional Paint
News, Vol.1, No.2.
One really would have thought that in the treatment
of historic houses the paint scrape had had its day. Yet
only last week a book dealing with the restoration of Uppark has
appeared, in which the making of scrapes "as pioneered by John
Fowler" still seems to be regarded as a valid method of paint
In spite of twenty-five years of articles and technical
papers cautioning against the misuse of this technique, large numbers
of intelligent people, some in positions of great responsibility,
allow themselves to be fooled by its practitioners.2
If one were to suggest to a curator that he carry
out the documentary research necessary for a particular project
using just one book taken off the shelves of a provincial public
library, one would be regarded as stupid, or, at best, mad. Yet
the same curator will often be quite happy to employ an interior
designer to scratch away at the paintwork with a scalpel, in the
belief that this simple action will reveal the nature of an earlier
painted colour scheme.
Admittedly, it is quite possible that the information
contained in that one library book would broadly convey a certain
amount of information that would be of use. However, with no corroborative
evidence, and no use of primary source material, a very incomplete
picture would inevitably be given.
A full technical analysis of the paint, on the other
hand, might be compared to research in one of the major repositories
of information, with access to original documents and the latest
published works. Furthermore, the report that followed it would
allow the investigative process itself to be examined, and supplementary
research could be added to the work at later stages, the whole building
up a fuller picture of the building concerned.
The reporting procedure following a scrape is, by
definition, sketchy. If lucky, the appearance of a layer, and its
apparent position in the stratigraphy will be explained. In the
recent, and scandalous example of a major London house, however,
colour boards of the layers relevant to the "restoration"
were apparently all that were produced.
Analysis, on the other hand, consists of a number
of different stages, and requires that the information obtained
is tested at all stages of the procedure.
Let us understand that Paint Analysis will
invariably consist of at least three processes, and will take place
only after an examination of the documentary information available
on the room/building.3
The sequence will be roughly as follows:
1) The taking of paint samples and the making of cross sections...these
will then be drawn or photographed;
2) With the stratigraphy of the cross section made clear by a drawing,
or photomicrograph, a scrape will then be made. The scrape must
then be exposed to UV light for a period of time in order to clean
up the colour.
3) Pigment and media identification will be carried out with the
aid of a number of microscopical, and microchemical techniques.
4) The evaluation of the cleaned-up colour may be made by comparison
with trial batches made up using the identified pigments.
The analyst will usually be accredited by a "competent
authority", and will have a background in the study of early
house-paint. The procedure will result in the submission of a written
report with drawings, photographs, and technical data on the materials
There is no doubt that the analysis of a painted
scheme is a lengthy procedure, but one that can do far more than
reveal the colour of an earlier scheme. Indeed, very often an approximate
date to each of the surviving layers can be given. As a result,
the work of a particular architect / designer can sometimes be identified.
Equally, if appropriate, the scheme that existed when the room was
occupied by an historic figure can be established, and an accurate
restoration to an earlier period attempted. The paint colour, in
this case, being as important as the furnishings and the objects
that the room once contained.
On the other hand, a Paint Scrape can be attempted
by virtually anyone, and usually is, with a clutch of architectural
historians, curators and interior decorators in the vanguard. There
are no professional qualifications, nor courses run in this technique,
and little opportunity for peer review.
There is no doubt that the action of scraping a painted
surface will reveal some of the underlying layers. But apart from
giving one a very rough idea of the colours used, and an approximate
idea of the number of schemes, what else is revealed ?
Will a scrape identify the pigments used in a scheme,
and therefore provide a clue to the date when it was applied ? Will
it reveal the relationship of the scheme in one room with that in
Would a scrape have revealed that the scheme that
survived in the Saloon at Uppark was the one applied by Humphry
Repton in ca. 1815 ? Could I have known that the freshly revealed
green scheme of a late seventeenth century gazebo was indeed blue
if I had used my eyes and scalpel alone ?5
Why is it that the evidence of the scraper's handiwork
is usually one of the first things to be painted out in any redecoration
? A recent visit to the Empress Josephine's house at Malmaison,
has demonstrated very clearly the problems with leaving these tell-tale
The Music Room6 retains several scrapes
from the 1983 restoration. In each case the exposed first scheme
is now significantly lighter than the surrounding "restored"
paintwork Experience has shown, however, that when freshly exposed
the original colour would have appeared much darker and more yellow
than it does now. The reason for this phenomenon has been known
in certain quarters for many years, but ignorance of this, or its
rejection, has been responsible for the past often appearing murkier
than was often the case.
In an attempt to measure the effect caused by years
of exclusion from light by the application of later paints, a recent
experiment was carried out by the author.
A 1930s edition of a Gliddens paint catalogue
containing a large number of painted colour chips was examined.
The colour names provided the first clue to the change that had
taken place in many of the samples during the sixty years that the
book had remained closed. "Sky Blue" had become a mid
green, and "Cream White" a dirty putty colour, for example.
The colours resembled those seen in freshly executed paint scrapes,
and it was felt that the exclusion from light in this (closed) book
had had a very similar effect to that of the overpainting of earlier
paint on a wall.
Measurements were made of five colours with a spectrophotometer,
and the paint samples were then left in an east-facing window under
diffuse daylight. Repeat measurements were made every two or three
days for the next four months, and the results recorded on a graph.
The colours changed significantly during the period that they were
exposed to light, with both the hue and the shade changing. The
type of change that took place was similar in all colours, with
the cleaned up examples being much lighter and much less yellow.7
In order to plot the changes that took place the colour
difference DE*ab was recorded. These reveal
an average DE of 9.92 units, with pale colours experiencing a DE
of 14 units, and darker colours a DE of 6.63 units.8
For the first three months the rate of change continued at about
a DE of .15 units each day.
When one considers that a colour difference, in a
pale colour, of more than a DE of .50 units is clearly discernible,
one will appreciate the problem.
The futility of relying on scrapes alone has been
known for many years, yet still a number of people cling on to the
nonsense. It appears that as well as being considered the easy option,
this is a manifestation of the "shock of the new" phobia
that afflicts a number of custodians; where technical analysis is
automatically equated with brash new colours in faded surroundings,
and where anything that smacks of science is, by definition, at
variance with an aesthetically acceptable result.
The opening paragraph to this paper reveals, perhaps,
a further reason behind the retention of scrapes in the curator's
repertoire. It is not simply a question of "taste", it
is, to a certain degree, snobbery. The late John Fowler is still
often credited with having had the definitive approach to the treatment
of historic interiors, and so methods practised by him thirty years
ago are still advocated by his devotees and by aspiring "cognoscenti",
eager to gain recognition.9
By listening to and understanding the importance of
the following words of Gervase Jackson-Stops it will be seen that
the paint scrape no longer has a place in the treatment of
our cultural heritage:
Among the lessons learned from the tragedy of the
Uppark fire has been the importance of recording not only the
past history of the house but what we ourselves have done to it.
We may make mistakes - in fact as humans we are bound to - but
future generations are unlikely to blame us as long as we tell
them what we did and why.
Rowell and John Martin Robinson, Uppark Restored. London:
The National Trust, 1996, 172.
the following: Morgan Phillips and Christopher Whitney, "The
Restoration of Original Paints at Otis House". Old Time
New England. Boston, Mass.: The Society for the Preservation
of New England Antiquities, vol.LXII, no.1, Summer 1971: 25-28.
Morgan Phillips, "Problems in the Restoration and Preservation
of Old House Paints", in Preservation and Conservation,
Principles and Practices. Washington, D.C.: The Preservation
Press, 1972: 273-285. Ian C. Bristow, "Repainting Eighteenth-Century
Interiors", ASCHB Transactions 1981 (vol. vi, 1982),
procedures of paint analysis have been discussed in greater depth
elsewhere. See, for example: Patrick Baty, "The Role of Paint
Analysis in the Historic Interior." The Journal of Architectural
Conservation. March 1995: 27-37.
very distinctive yellow pigment, orpiment, found in the second (salmon)
scheme in the Large Library at Goodwood House, was also found in
the first (green) scheme of the Tapestry Drawing Room. This information
(together with other details) supported the belief that James Wyatt
was responsible for both schemes, and gave weight to the theory
that an earlier architect, possibly either Sir William Chambers
or Matthew Brettingham, had originally designed the Large Library.
(Patrick Baty, Goodwood House: A Report on the First Scheme,
Following an Examination of the Painted Surfaces in the Large Library
and in other Areas. 27th May 1996.)
(Baty 1995, 33) for a full account. Briefly: what appeared to be
green, having made a scrape, was revealed as being a discoloured
Prussian blue scheme in cross section.
Music Room at Malmaison is best known by the watercolour by Auguste
Garnerey. This watercolour was begun in 1812 and finished by his
sister in 1832. It appears as Fig. 157 on page 182 of Mario Praz's
An Illustrated History of Interior Decoration, in the first
paperback edition published by Thames and Hudson Ltd. London, 1994.
lightest colours were nos. 1 (Cream White) and 2 (Light Buff). The
darkest colour (Lemont Stone) was no. 3, and the two mid tone colours
were no. 4 (Colonial Yellow) and no. 5 (Buff).
CIE L*a*b* system of notation is
the one most often used when measuring paint colour with a spectrophotometer.
The L* indicates lightness, the a* and b*
are the chromaticity coordinates, where +a* is the red
direction, -a* is the green direction, +b*
is the yellow direction.
refreshing to see signs of questioning taking place at Uppark, where,
amongst other things, the 1970s pink (Fowler) scheme in the North
Corridor was replaced with a stone colour, as used by Humphrey Repton
in the early 19th century.
and Robinson 1996, 178).