Frank Lloyd Wright furniture: A technical study of

objects from the Darwin Martin House

Pamela Kirschner, Peebles Island Resource Center

A BSTRACT : The purpose of this study was to investigate the upholstery,

construction and finishes of three pieces of furniture from the Frank Lloyd Wright-
designed Darwin Martin House in Buffalo, New York. These pieces include a
table, chair and sofa. Along with examination for physical evidence, methods
of analysis included X-radiography to determine construction techniques,
Fourier Transform Infrared Spectroscopy (FTIR) to identify the organic mate-
rials present in the finishes, and with ultraviolet light and microscopy to aid in
determining the finish. Samples were taken from the back, arms and sides of a sofa,
a chair splat and the top and base of a table. Historical research was also done on
the materials used and the manufacture of furniture during this time period.

Introduction

T
HE FINISHES THAT ARE THOUGHT TO statement by Wright that “Furnishings should be
be original on this furniture are very frag- an extension in the sense of the building in which
ile and not easily identified. The design they furnish.”
and manufacture of Frank Lloyd Wright furniture
during this period (early twentieth-century) is The principles of the Craftsman furniture were
consistent with the furniture made in the style not far from Wright’s and were based upon hon-
of the Arts and Crafts. A large amount of experi- esty and simplicity. The inventor of Craftsman
mentation for finishing techniques and materials furniture, Gustav Stickley, felt that the furniture
occurred at this time. Finishers were following the being turned out in great quantities by the
particular aesthetic of the Arts and Crafts period, factories was “badly constructed, over-ornate and
and what may have been thought to be stable meaningless.” He sought to make strong, simple
finishing materials are showing signs of consid- and comfortable furniture. Stickley chose white
erable deterioration. It is imperative to understand oak for its sturdiness and beauty, and the furniture
what these materials are to treat and preserve them was constructed to last several lifetimes. The first
properly. Experimentation with the construction of the Craftsman furniture was built in 1898, and
of furniture occurred as well at this time. These was presented to the public in 1900. Catalogs of
objects are made with techniques that are much Stickley furniture from 1904 describe finishes
more complex than would be expected with the giving a choice of “Craftsman Fumed Oak, Silver
Arts and Crafts philosophy. Gray Maple and Mahogany.” The “Fumed Oak”
finish is stated as giving the wood the look of age,
Materials and Techniques Background without injuring its natural qualities. It was felt
Frank Lloyd Wright favored minimal finishing of that the raw wood lacks “mellowness” which is
wood and admired the look of Japanese finishes. In obtained over time and weathering. The beauty
one quote he says, “strip the wood of varnish and of the grain is preserved as well as enhanced, and
let it alone—stain it.” He also felt that the silken it accentuates the “watered” pattern-like effects
texture of wood was vulgarized by varnish having which run through its texture. The wood should
an unacceptable sheen. He stated, “wood best pro- be treated so that there is little evidence of an
tects itself and that a coating of clear resinous oil applied finish. Oak should be ripened by
would be enough.” Another reference included a fuming with ammonia which has an affinity for

Kirschner: Early Frank Lloyd Wright Furniture

the tannic acid in the wood. It darkens the soft similar to what was used by the Matthews brothers
areas of wood as well as the dense rays, coloring on the Frank Lloyd Wright furniture. The owner,
in an even tone. Darwin Martin, kept a detailed journal which
included recipes for mahogany and fumed oak
The Craftsman catalog includes a description of the finishes. There are descriptions of the application
method of fuming oak. After it has been moistened of stains, both water and oil-based, sanding, and
to open the pores, the wood is placed into an air- the application of shellac or varnish.
tight box or closet with a dish of ammonia. Usually
48 hours is enough to color the wood depending Historical Background of the Darwin
on the size of the compartment. The more tannin Martin House
present, the darker the result will be. A “touch The prominent twentieth-century architect, Frank
up” may be necessary by mixing a brown aniline Lloyd Wright, was responsible for designing nine
dye dissolved in alcohol with German lacquer “ba- buildings in Buffalo, NY between 1902 and 1927.
nana liquid” (amyl acetate). Cheese cloth is used Darwin Martin, an executive of the Larkin Soap
to blend the stained areas with the fumed areas. Company, was responsible for introducing Wright
Afterwards, a touch-up coat of lacquer is applied to many of his Buffalo commissions. These in-
made of 1/3 white shellac and 2/3 German lacquer. cluded a home for Martin’s sister-in-law and then
It is advised to apply two coats of lacquer, each con- a home for Martin and his family, in which a
taining a little color, to sand in between each coat conservatory and stables were built as well as a
and to apply one or more coats of prepared “floor walkway extending from the main house to his
wax.” A last rubbing of “Craftsman Wood Luster” sister-in-law’s house. Frank Lloyd Wright’s design
is suggested, which is likely wax. This gives a soft of the Martin House included furniture as well as a
satiny luster to the surface of the wood. furniture layout.

A Sherwin-Williams Company ad from a 1906 The Martin family left the house after Darwin
Craftsman catalog gives directions on “How to Martin died in 1935. By 1940, it had started
Finish in True Craftsman Style” interior woodwork to show signs of deterioration. The house and
and furniture. They named all their finishing prod- property were sold at a tax auction to the city of
ucts “Sherwin-Williams Handcraft Stains.” Several Buffalo for $394.00. In 1950, the house was
colors were available: Weathered Oak, Cathedral purchased by a Buffalo architect, Sebastian Tau-
Oak, Tavern Oak, Flemish Oak, Brown Oak, Old riello, who saved the main house. The walkway,
English Oak, Fumed Oak. Their directions for conservatory and garage were not saved. The house
securing the true Craftsman finish suggest that the was acquired by the State University of New York
wood should be clean and smooth. The “Hand- at Buffalo in 1967. Since then, there have been
craft” stain should be applied to bare wood with a ongoing efforts to preserve the house. In 1994 it
soft brush and allowed to stand overnight. Next it was designated an Historic Site, making it the 35th
should be lightly gone over with steel wool or fine such site in the state of New York.
sand paper. A coat of Sherwin-Williams Mission-
Lac (what the company refers to as a superior sub- Sixty pieces of furniture for the Darwin Martin
stitute to shellac as well as being cheaper) should House designed by Wright, as well as some Stick-
then be applied. This will dry hard in three to five ley pieces approved by Wright for the house, have
hours. Finally, the surface should be sanded lightly been brought to Peebles Island Resource Center to
again giving a rich, soft, velvety effect. receive conservation treatment. The collection is
significant, for the objects contain areas of original
The Matthews Manufacturing Company of Mil- finish. The furniture suffered abuse over the years
waukee, Wisconsin, founded in 1857, was referred but has held up well with only some loose veneer,
to several times in letters between Wright and grime and replacement upholstery.
various contractors. They were responsible for
almost all of the interior trim and cabinet work as Objectives
well as the furniture in the Darwin Martin House. The object of this study was to examine the
It is likely that the finishes described above were construction and finishes on three pieces of

1998 WAG Postprints—Alexandria, Virginia

Frank Lloyd Wright furniture from the Darwin B. Sampling
Martin House in Buffalo, New York and determine Finish was sampled from each area on the furniture
construction techniques as well as finishing that fluoresced differently under ultraviolet light.
techniques and materials. This knowledge will The samples were taken with a scalpel and were as
aid with the preservation and treatment of other small as possible. Scrapings were taken of the finish
objects from the Darwin Martin house as well as for FTIR analysis, and small cuts were made in the
similar Frank Lloyd Wright furniture from this finish down to the wood for cross section samples.
period. The samples were either placed in small glass vials
with lids or mounted in polyester resin for cross
Procedures, Materials and Methods: sections and were labeled appropriately.
A. Design, Description and Construction:
Table Samples were taken of both the top of the table
The table was originally located in the library of and the base. The table top appears darker and
the house. It stands 29" high and 21" wide. There redder than the base. The chair was sampled on
are four casters, one under each foot, made up the back board which appears similar to the table
of a ball in a cup allowing it to roll freely. Con- top. The sofa was sampled on the arms, the side
struction found in the table using solid cores and panel and the back panel. The back panel appears
thick oak veneer surrounding the core is similar to more glossy and redder and is also less sensitive
construction described in the L & JG Stickley to the solvents tested.
Craftsman Catalog. The top appears more red and
more glossy than the lower portion of the table. C. Measurement and Data
The analytical techniques used to analyze
Chair the samples included X-Radiography, Fourier
The chair was designed for the dining room of Transform Infrared Spectroscopy (FTIR), hand-
Wright’s Heath House, also in Buffalo, but was held ultraviolet light, and microscopy. The FTIR
later moved to the basement of the Martin House analysis was performed in the Winterthur Con-
when the architect Tauriello lived there in the servation Department. All other analysis was
1960s. The chair is 43" high and 16" wide. The performed at Peebles Island Resource Center in
joinery is mortise-and-tenon, the back is a solid Waterford, New York.
oak board, and there are metal glides on the bot-
tom of each foot. The upholstery is a printed pile The X-Radiography analysis was used to under-
fabric replacement show cover on a slip seat and stand the construction techniques of the back
is very faded and worn. The under-upholstery is panel. The back panel was removed from the
original cotton batting, moss, horse hair and web- sofa and placed onto an X-ray machine by CGR
bing, which was sagging. Medical Corporation modified for flat objects.
The X-ray was taken at 60 KV and 50 MA for four
Sofa minutes. The FTIR analysis was performed using
The sofa, one of a pair, was designed to go in the the Analect RFX-65 with the XAD microscope.
library of the Martin House. It is 32" high and 73" The FTIR data aided in the identification of the
wide. The frame is constructed of solid oak as principal organic components in the finishes. The
well as veneer. The upholstery is a replacement samples were run with 200 scans with a gain of
gold velvet cover with two small loose pillows. either 2 or 2R. The FTIR microscope was used, and
The brackets on the sides of the back panel were therefore only very small samples were necessary.
loose, and they were removed to reveal construc- The samples were placed onto the diamond cell
tion as well as upholstery techniques. The brackets under a microscope and were flattened using a
themselves are attached with two dowels and glue. metal roller. The diamond cells were then placed
After these were removed, screws were revealed under the FTIR microscope, focused, and an area
running through the arm at a diagonal into the back was chosen with the aperture. A background test
panel. These unscrewed easily, and the back panel was run on a clean area of the diamond cell, and
lifted out of the tongue and grooves. then the sample spectra were run. The gain and

Kirschner: Early Frank Lloyd Wright Furniture

scans were adjusted as needed. Once the spectrum color. The other areas thought to be refinished
was acquired, the baseline was corrected and spec- as well—the arms, table top and chair rails—all
tra were plotted out. Ultraviolet light analysis was fluoresced a bluish-white color.
performed using a hand-held black light (365 nm).
Microscopy was performed using white light as well B. X-Radiography
as ultraviolet light under 100X and 200X using an The X-ray was taken of the proper left top corner
Olympus Vanox microscope. The samples were of the back panel. It revealed boards making up
mounted in polyester resin, allowed to dry under 5-ply lumber core as well as separate mitered
a tungsten bulb for four hours and then sanded boards connected to the lumber core with tongue-
and polished using Micromesh. and-groove construction. There are two dowels
joining the mitered boards together. Moss, horse
D. Analysis and Evaluation hair and upholstery nails were also apparent in
The FTIR spectra obtained were compared to known the X-ray. The upholstery understructure was
reference spectra of finishes. The additional data attached with nails to the back board equal dis-
were also compared to the research of materials and tances apart from each other. (fig. 1)
techniques used in finishes of this period.
C. Microscopy
David Bayne of Peebles Island was present when Cross sections of areas that were thought to be
the samples were taken. Examination with ultra- original, such as the table base and the side panel
violet light was done in conjunction with solvent of the sofa, show a creamy-white fluorescing finish
tests first because it is non-destructive. This infor- (possibly two applications) with a thin layer of dark
mation was compared to the analysis obtained
from FTIR and cross sections. The sofa was too
large to fit into the X-ray room. The back panel,
however, was easily removed which provided
an abundance of information. Samples of finish
were taken from an identical sofa, also part of
the Darwin Martin House collection, and were
found to be the same. These samples provided
stronger spectra, providing more information.

The project was supervised by David Bayne,

Furniture Conservator at Peebles Island. Janice
Carlson, Senior Scientist, Winterthur Museum,
aided in the use of the Fourier Transform In-
frared Spectroscopy. She also assisted in the
interpretation of the data gathered.

Results and Discussion

A. Ultraviolet Light Analysis
When examined with ultraviolet light, the table,
the sofa side panels and the chair rails, thought
to carry the original surface coating, fluoresced
in a characteristic greenish-white as well as red
undertones. The coatings from these areas were
all similarly soluble in polar solvents such as
alcohol and water as well as non-polar solvents
such as mineral spirits. The back of the sofa and
the lower side panels appeared to be refinished
because the finish was more stable, was not sol- Figure 1: X-ray of Proper Left corner of the back panel of
uble in mineral spirits and had a slightly orange the sofa.

1998 WAG Postprints—Alexandria, Virginia

particles on top. The sample from the back of the Shellac
sofa showed two applications of a fluorescing yel-
lowish-white finish as well as a thin layer of dark
particles on top. The thin layer of dark particles
on each sample was either a dirt layer or a thin,
non-fluorescing stain applied on top of the finish.

Absorbance
Samples from the arm of the sofa showed a yellow
fluorescing layer, a bluish-white fluorescing layer,
a brownish-blue layer, and dark particles on top.
The table top also showed a bluish-white layer,
but with particles incorporated into it and a dark
layer underneath.

D. Fourier Transform Infrared

The spectra acquired from the supposedly original
areas were all similar and rather complex. The Wavenumber (cm-1)
primary component appears to be a natural resin,
possibly shellac. There was also evidence of oil, OL 1195 Angelo Granta Shellac
likely a drying oil. Several of the samples exhibited
a group of bands around 1060 as well. This region
of the infrared spectrum is particularly difficult to
interpret because a number of materials—organics
such as ethers and inorganics such as silicates,
phosphates and sulfates—strongly absorb here.
Absorbance

In addition to natural resin and oil components,

the original finishes all contained a cellulosic com-
ponent. In particular it was found in the coating
from the sofa side panel. The cellulosic bands were
so prominent that a computer spectral search
came up with cellulose itself. Cellulosic materials
have been used as the primary starting materials
for such coatings as nitrocellulose and cellulose
acetate. However, these materials were not used
commonly until the 1920s, and the typical peaks
Wavenumber (cm-1)
did not show up in the spectra. One explanation
is that it could be a cellulosic component added Figure 2: Comparison of shellac versus Angelo
to the stain as a binder, such as gum arabic. (see Granta shellac on Table base/old FLW92B.
fig. 2, 3 and 4)
finish, one must use a technique such as Gas-
Infrared spectra of what was believed to be re- Chromatography-Mass Spectrometry to sepa-
finished areas showed them to be quite different rate these components first. FTIR does show us,
from the areas of presumed original finish. The however, that the original finish on these objects
coating from the splat of the chair produced an contained a mixture of natural resin, oil and a cel-
infrared spectrum which compared favorably to a lulosic component as well as inorganic materials.
polyurethane in both computer and manual spec-
tral searches. The material found on the table top
was identified as a silicone varnish. Conclusions
The finishes on the sofa side panel and the table
FTIR can only provide general classifications base were found to contain similar materials as
for organic compounds. In order to identify the found in the historical references for finishes at
specific natural resin or oil component of the the turn of the twentieth century. They contained

Kirschner: Early Frank Lloyd Wright Furniture

 Shellac/Cellulose

A
B
S
O
R
B
A
N
C
E

-1
Wavenumber (cm )
OL1195 Angelo Granta Shellac

A
B
S
O
R
B
A
N
C
E

-1
Wavenumber (cm )
Figure 3: Comparison of shellac/cellulose versus Angelo Granta shellac on sofa/side
panel/old 6B.

a mixture of oil, natural resin, possibly a cellulosic ing the materials used in the manufacturing of
binder and inorganic materials (likely pigments). this furniture, and this knowledge gives insight to
The chair back, table top and the sofa arms were understanding the deterioration of the finish. It
found to have later synthetic finishes, both silicone will aid in the treatment of similar objects to allow
and polyurethane. cleaning without removing original materials.

Most Frank Lloyd Wright furniture as well as Revealing the construction of the furniture allows
interior moldings have been refinished with no us to place the objects in context with a particu-
attempt at saving the original coating. The Dar- lar time period and the Arts and Crafts style. It
win Martin house is an excellent resource to aid in also gives us more insight into the practices of
understanding the original appearance of Frank building Arts and Crafts furniture that do not
Lloyd Wright Prairie Houses. The data obtained necessarily match the philosophy of the period
from this study is extremely helpful in identify- such as plywood, veneer and dowels. We are able

1998 WAG Postprints—Alexandria, Virginia

 Shellac /V. High Cellulose

A
B
S
O
R
B
A
N
C
E

-1
Wavenumber (cm )
OM620 Kayocel 1000L

A
B
S
O
R
B
A
N
C
E

-1
Wavenumber (cm )
Figure 4: Comparison of shellac/v. high cellulose versus OM620 KAYOCEL 1000L
on sofa/side panel/old FLWS2B.

to understand the structural wear or lack thereof End Notes

in these objects, as well.
1. Patterson, Terry L. 1994. Frank Lloyd Wright
and the Meaning of Materials. University of Okla-
homa: Van Nostrand Reinhold.
Acknowledgments
I would like to thank the following for their help 2. Ibid.
and support: David Bayne, Furniture Conservator, 3. Wright, Frank Lloyd, 1960. Writings and Build-
Peebles Island Resource Center; Jan Carlson, ings. New York: New York Horizon Press.
Senior Scientist, Winterthur Museum; Jim Gold,
Director, Peebles Island Resource Center; John 4. Gustav Stickley’s Craftsman Catalog, 1904.
Lovell, Assistant Director, Peebles Island Re-
source Center; Michael Podmaniczky, Furniture
Conservator, Winterthur Museum.

Kirschner: Early Frank Lloyd Wright Furniture