Technical Study of a BLUE CHAIR

Full paper access

Abstract

Technical analysis was conducted on a blue japanned chair with raised and gilt decoration, which is attributed to Giles Grendey c. 1735, London, recently accessioned by Winterthur Museum (2016.0046). This study on the painted surfaces of the English blue japanned chair reveals the materials and techniques applied to it in the 18th century, which are probably originated from the historical records and recipes from Stalker’s and Parker’s A Treatise of Japanning and Varnishing in the 17th century, but the modifications and variations suggest more consistency with Dossie’s Handmaid to the Arts in late 18th century.

Key words: English blue japanning, gilding, raised decoration, sprinkling, technical analysis, Giles Grendey

Introduction

This chair was donated to the Winterthur/University of Delaware Program in Art Conservation Study Collection by Elizabeth and Wallace Gusler in 2016 and then accessioned into Winterthur Museum. The blue japanned, high-back side chair with raised and gilt decoration is a late Baroque chair, attributed to Giles Grendey, c. 1735, London. It features many characteristics of this period, such as cabriole legs, ball-and-claw feet, caning, and japanned surfaces. While little is known about this piece in particular, similar pieces by Grendey have been found and researched.

Based on the examination, documentation and research conducted in 2016 by Madison Brockman and Victoria Wong, technical analysis of the japanned surface was conducted in 2019 to identify the materials and techniques employed on this japanning surface, and then relate them to the historical documents and recipes about the eighteenth century English/European japanning, trying to ascertain the provenance of the chair, and provide sound basis for future conservation and research.

Historical Context

There was further expansion of the British Empire during the late Baroque period under the Queen Anne and George I periods (Whiton 1974). This expansion reinforced trade, making England one of the economic powerhouses of the time. Late Baroque, or Queen Anne, style furniture is defined by the abandonment of the rectilinear shapes of the William and Mary period for S-curves. This transition was due to the influx of Chinese export furniture, brought in by the British East India Company. Other characteristics of this period are cabriole legs, ball-and-claw feet, caning, and japanned surfaces (Vandal 1990). English japanning was most commonly carried out in black, but other English japanned pieces exist in red, green, and blue. English furniture in this period was typically made of English walnut (Anderson 2016).  The characteristics discussed are representative of the Chinese influence during this period, and are seen in numerous chairs of this time, such as those by Giles Grendey (Brockman and Wong 2016).

Giles Grendey was a well-known English cabinet-maker. Numerous pieces that do not bear a Grendey label or inscription have been attributed to him if they closely match pieces that do bear a Grendey label or inscription (Bowett 2015).  This chair closely resembles pieces made for the Duke of Infantado that are red. Two Grendey chairs, formerly at the Colonial Williamsburg Foundation, are nearly identical to this piece, except that they are green and feature large claw-and-ball feet (Gusler and Gusler 2016). This may be due to ageing varnish casting a yellow tint on a blue ground (Thompson 2016). (Brockman and Wong 2016)

The word japanning came into use during the seventeenth century, when lacquerware from China and Japan rose to the height of fashion—black and gold lacquer from Japan becoming the most admired and sought-after. Since that time, the words lacquer and japan have been used generally and indiscriminately to describe varnishes and paints that produce a glossy surface. In short, japanning is the Western re-creation of Far Eastern lacquerware. By the end of the seventeenth century, European craftsmen were aware of the techniques and materials used in Far Eastern lacquerware (Ballardie 1998).

In many European historical publications that described the techniques, materials, and designs used for japanning, A Treatise of Japaning and Varnishing by John Stalker and George Parker (1688) is probably the best known. Stalker and Parker include white japan recipe in his publication, as well as blue japan (For details please see appendix). Two further versions of the recipes may be found in the 18th century literature. Although one—in the Polygraphic Dictionary (1735), under the heading “Japan”—is almost identical to that of Stalker and Parker, the second—in Dossie’s Handmaid to the Arts (1796:314)—shows modifications and variations (Ballardie 1998).

The literature has provided great insight into the materials, techniques, and designing ideas for English japanning in the 17th and 18th century. However, it could not entirely reflect the common practice of craftsmanship in that period. These recipes wouldn’t be directly employed without any modification and variation. Therefore, for this specific case study, technical analysis will help to reach a better understanding about the common practice for japanning around mid- 18th century and provide supplementary information for future research.

Experimental

Cross-section analysis was conducted first to reveal the stratigraphy of the japanned surfaces, verify the materials identified with instrumental analysis, trace them down in the layer structure of the japanned surfaces, and deduce the possible techniques that craftsman employed.

In this study, the pigments, varnishes of resins, binders, metals, glues/sizes, are major components appeared in the literature or cross-section analysis to be identified or confirmed with instrumental analysis. The analysis was conducted in the Scientific Research and Analysis Lab at Winterthur Museum, Library and Garden, with the supervision of Catherine Masten, in 2019.

For visible colors, mainly are blue, red and white, are seen by naked eye. To be specific, the blue pigments are likely to be Prussian blue, smalt; the red pigments are likely to be red lead, vermilion; the white pigment is possibly lead white. The varnishes used in historical recipes are mainly natural vegetal resins, including sandarac, copal, elemi, mastic, Venice turpentine, benzoin, rosin. Shellac is mentioned in some recipes as varnish or finish, which is exuded from insects. These materials need to be determined with instrumental analysis. Gum-water (gum Arabic in water), isinglass and drying oil were possibly used in this chair as binders for pigments and grounds. The thin layer of gilding is likely to be gold leaf, and the possible regilding/retouching needs to be identified. The glue/size is possibly animal glue, isinglass to be specific, or drying oil as gold size.       

To identify these materials, the main analytical techniques employed in this study are XRF, FTIR, SEM-EDS, Raman and GC-MS.

Results

Cross-section Analysis

Three rounds of sampling were conducted by Madison Brockman and Victoria Wong, and the author, for the WUDPAC documentation project in December 2016, cleaning course in October 2018, and technical analysis in May 2019 (For details and results, please see appendix). Due to the friability of the surface, the first two rounds of sampling didn’t obtain an intact sample that can depict the entire layer architecture of the japanned surfaces, but the samples provided preliminary information and sound basis for the study. In May 2019, consulted with Christine Thomson, 5% (w/v) B72 in acetone was used to consolidate the sampling spots firstly, then the third round of samples were taken, to verify the materials identified with instrumental analysis, trace them down in the layer structure of the japanned surfaces, and deduce the possible techniques that craftsman employed.

Four samples were taken in the third round. Although the gilt decoration layers still separated from the blue japanned layers, they kept intact instead of pulverization and were embedded in the same polyester cube for observation. Sample A and C showed the stratigraphy of the blue japanned and gilt decoration.

The diagram above shows the architecture of the japanned and decoration surfaces. They include, the base coating, blue japanning, raised decoration and gilding, partial regilding likely and finish (or refinish).

Base Coating

Stalker and Parker suggested a gesso ground for both blue and white japanning. The gesso ground is a mixture of isinglass size and whiting (CaCO3) to the consistency of light cream, then apply three coats. When it’s dry, rub back the surface until very smooth. While Dossie maintains that the gessoed ground should not be applied because it causes the japanned ground to “crack and fly offin flakes.” Applied in its place is a layer of “clear-coat” or a “clear-col,” which is a hot size with the addition of a small amount of whiting.

As we can see in several cross-section samples, there is a clear, but pale blue layer, fluoresce pale yellow under UV light, in between the blue pigment layer and wood substrate. This layer sometimes soaks into the wood, which seems to be consistent to Dossie’s “clear-coat”, when dry, it is rubbed smooth and, although it fills only the pores of the wood, can produce an excellent surface (Ballardie 1998).

Blue Japanning

The cross-section samples that obtain the blue pigment layers show there are at least two layers of blue pigment. The lower layer is lighter than the upper layer. They seem to be blue pigment mixed with lead white in different proportion to build up the depth of blue color. According to literature, the blue pigment is mainly smalt, while Prussian blue, blue verditer were also available, or used as mixture. The particles of the blue pigment in the cross-sections didn’t appear to be angular shaped, irregular, or degraded blue color, which are the main optical microscopic characteristics of smalt.

Above the blue pigmented layers, there is a layer with angular, irregular shaped and almost translucent particles, likely to be a layer of smalt only with binder, to build up the depth of the color, and give lust to the surface. The binder for this smalt layer is isinglass size according to Stalker and Parker, while later in the18th century, Gum elemi and poppy oil were often heated as binder. There is also possibility that the angular shaped particles are quartz, or plaster for some preparatory or ground layer.

After the blue pigmented layers were built, a thick layer of brownish translucent varnish was applied. In Stalker’s and Parker’s recipe, such best white varnish will be up to seven or eight layers. However, the blue japanned surface may also have been covered with colored varnish in later practice, like golden, to produce translucent green colored surfaces (Ballardie 1998).

Raised Decoration and Gliding

Above the blue japanning, the chair was decorated with raising figures in relief and gilded as well. For some of the samples, the raised decoration is a very thin layer of white material, while for the others, the raising part is thick, white and solid, raising and tapering can be seen.

Stalker and Parker suggested using whiting and red bole to be ground together with gum Arabic as binding media. In later recipe, whiting and pigments were advised, and then gesso. In this case, it’s possibly gesso or whiting since the color, and the gold size above the raised decoration.

After the raised decoration was done, a thin layer of gold size was applied to it, then there’s another thin layer with red pigment on which the gold leaf was adhered. Two or three layers of varnish were applied after gilding. The application of white raised decoration, gold size and red pigment, may suggest the gilding is later than Stalker’s and Parker’s publication

Possible Partial Regilding/Retouching

In some of the samples taken from the center of the splat, where heavily decorated, several more layers of varnish were applied above the gilding. A layer of metal shavings or chippings is in the middle of the varnish layers. The metal shavings or chippings seem to be crumpled thin metal leaf. The varnish layers weren’t polished to smooth as the others before, even fused with varnishes below. It’s likely to be a later campaign of regilding/retouching.

Refinish

On the top of these samples, a layer of refinish can be found, filling into the cracks of varnishes below. The lower part of this finish layer has orange fluorescence under UV light, likely to be shellac. As the historical recipes mentioned, several layers of shellac or resin varnish can be applied and polished as finish. In this case the finish is later than the other painted work due to the cracks formed in the varnishes below.

The interpretation of the cross-section samples reveals the relation between the layer structure of the japanned surfaces and the historical literature. The japanning and decoration on this chair are probably later than Stalker’s and Parker’s publication, which is consistent to its attributed date. The identification of materials such as pigments, binders, varnishes, metals, with scientific instruments, will develop deeper insight into the painted surfaces, reveal the recipe and techniques employed on this chair.

XRF

Seven spots of two areas on the splat were taken for XRF analysis. One is in the center part of the splat, the other one is on the side rail. The XRF analysis was conducted and interpreted with the supervision of Dr. Marcie Wiggins.

The seven spots for XRF analysis focused on the blue japanning and gilding on the raised decoration respectively. The two areas, one is on the center part of the splat, the other one is on the side rail of the splat. The spectrum of the four spots on the blue japanning all show significant amount of lead, iron, and small amount of calcium and barium. The three spots on the gilding and raised decoration contain fair amount of copper, zinc, gold and mercury, besides the lead, iron, calcium and barium detected in the other four spots mentioned above.

It’s noteworthy that on the area of center part of the splat, all the four spots show significant amount of cobalt, nickel, potassium, manganese, and trace amount of arsenic, whose spectra may be covered by that of lead.
In summary:

The iron presents in all spots suggests the use of Prussian blue, while further confirmation is still required to ascertain the material.

The significant amount of lead indicates the use of flake white, that is lead white, used together with blue pigment, or white pigment alone. Another possibility is red lead, for the gold size.

Calcium and barium can be found in whiting, or gesso.

Mercury is detected only in the gilding area, suggesting vermilion was used as red pigment for gold size.

Copper and zinc, components of brass, could be used as the partial regilding/retouching.

Cobalt is the indicator for smalt, but also could be used as drier for oil. But in this case, several elements, such as nickel, arsenic, manganese, and potassium (component of glass), were detected along with cobalt, which suggest the presence of smalt. Since they only appeared on the center part of splat, it’s possible that smalt was used only for the heavy decoration areas where desire more lust and depth of blue japanning. It’s probably used after two layers of Prussian blue with lead white were applied, in another layer with large angular shaped particles, which are the typical form of smalt particles.

SEM-EDS

Two cross-section samples taken in the second round were analyzed with SEM-EDS. Both samples were taken from the side rail of the splat. Sample B1 only contains blue japanning layers and wood substrate. Sample G2 contains the other layers except the blue japanning and substrate, focusing on the gilding and decoration.

The backscattered electrons (BSE) image and elemental mapping of sample B1 at 360x confirm the results from XRF analysis. The blue pigments contain significant amount of iron and lead, which suggest Prussian blue and lead white. Calcium mainly is located in the varnish layer, indicating the whiting in the varnish.

The BSE image and elemental mapping of sample G2 at 267x also depict the distribution of the elements detected by XRF analysis. Calcium appears mainly in the varnish and raised decoration, which may suggest the whiting used in the varnishes. Gold leaf was used for gilding and the overlapping of the leaves can be seen. In the gold size, there is rich content of lead and mercury, suggesting red pigments of red lead and vermilion, which confirms the earlier assumption from literature, cross-section analysis and XRF analysis.

It’s noteworthy that the campaign of regilding/retouching layers are actually copper-zinc fragments suspended in the varnish. There is no clear boundary between these layer and earlier gilding decoration, even appears to be wet-on-wet application. And the cracks of this generation go through the varnishes below.

In the literature, there is a technique of sprinkling, or strewing or speckles in Stalker’s and Parker’s, which sprinkles the metal powder or dust onto the wet varnish, and then use another layer of varnish to seal it. This technique is to enhance the sheen and color for the gilding. It’s believed in this case, sprinkling was applied onto the gilding decoration. The BSE image supports the sprinkling technique other than the assumption of regilding/retouching from the cross-section analysis. Copper and zinc alloy, probably brass dust, was the common choice in the historical recipe for this technique.

For the layer with angular shape particles on the bottom of this sample, it’s supposed to be a layer of smalt with binder to add the depth and lust to the blue japanning. However, cobalt wasn’t detected in the particles, where contains mainly silicon. Silicon is the major component for glass, which is also ingredient of smalt. The reason why there is no cobalt is probably because of the degraded material and less content of cobalt to be detected. This is consistent to the XRF analysis, on the side rail of the splat, there is no presence of cobalt. Maybe in this less decorative area, such layer of smalt is not so important that often to be left out.

In sample G2, the elemental mapping doesn’t show a dominant element in the raised decoration part. It’s possible that in this particular sample, the raised decoration was not included or too thin to be analyzed.

For further confirmation of the use of smalt and materials for raised decoration, another sample taken from the center part of the splat is recommended to conduct SEM-EDS analysis.

FTIR

The four samples came from the side rail of the splat. Each sample was scrapped and supposed to be taken from the different layers of varnishes or binders: the top dark brown varnish, the second brownish varnish/binder with the gilding decoration, the darker varnish on top of the blue japanning, and the lighter varnish/binder in blue japanning layers.

The result of FTIR analysis of the dark top varnish content matches Kireelac in the spectral library, a natural resin from India, exuded from insects.

The result of FTIR analysis of the second brownish varnish/binder possibly for gilding matches perfectly to sandarac resin.

The result of FTIR analysis of the darker varnish on top of the blue japanning suggests there are animal glue, lead white, and Prussian blue, which is consistent to previous analysis and possibly is the mixture of blue japanning materials.

The result of FTIR analysis of the light varnish/binder in blue japanning layers contains lead white (in linseed oil), Prussian blue, and likely animal glue, which may suggest the use of drying oil in the varnish/binder.

GC-MS

GC-MS was used to determine multiple organic compounds more specifically. Eight samples were taken from the lower part of the side rail of the splat. Six of those were scraped from the top layer to the bottom. They were supposed to cover each layer of the painted surfaces and analyzed for oils, resins, varnishes, and waxes. The other two samples were taken at different spots for protein (animal glue) and polysaccharide (gum water, starch) analysis. The GC-MS analysis was conducted and interpreted by Dr. Chris Peterson.

The six samples analyzed for oils, resins, varnishes, and waxes came out consistently. All of them contain pine resin and variations, drying oil and mineral wax. From the top layer to the bottom, the content of each components is less, and the drying oil and pine resin are less degraded compared to the sample from the top layer. In sample 3, there is small amount of sandarac, which is consistent to the results from FTIR analysis, but mostly pine resin.

The sample for polysaccharide analysis came out no result. Gum water or starch were not detected in the samples acquired.

The sample for protein analysis only shows trace amount of material, possibly animal glue. It’s not clear if animal glue was used as binder. It is likely in the 18th century, drying oil and resins replace gum water and isinglass size to be binders for japanning and decoration.

Raman

Sample G2 was analyzed with Raman to confirm the presence of lead white, Prussian blue, calcium carbonate, vermilion, and smalt. Except for smalt, other results are positive to the materials mentioned above.

Conclusion and Discussion

Based on the literature review and cross-section analysis, the technical analysis bridges the materials and techniques employed on this English blue japanned chair, which can be described as the following diagram.

On the wood substrate, a layer of “clear coat” was applied first to make a smooth and excellent ground for japanning. It’s possibly hot size and whiting in historical recipe, but the technical analysis didn’t get result of animal glue, this layer could also be drying oil with whiting.

Prussian blue and lead white were mixed in different proportion, possibly in binder of isinglass size (trace amount detected with instrumental analysis), or drying oil, or resins, for blue japanning. Cross-section samples showed two applications.

Then a layer of smalt, possibly with a little whiting, mixed in isinglass size, was applied to enhance the lust and depth of blue. A coat of white varnish was applied and rubbed to smooth on the blue japanning. The varnish is mainly pine resin, with small amount of sandarac.

For the materials used for the raised decoration, it’s still unclear due to the location of the samples analyzed. From the cross-section samples, it seems to be chalk, or gesso. Further analysis and confirmation are necessary.

The gilding decoration consists of a clear coat of gold size, possibly dry oil and resin, then a layer of red lead and vermilion to enhance the gold, and a thin layer of gold leaf.

Several layers of varnishes, which are mainly pine resin and sandarac, with a little whiting, were applied after gilding.

When the varnishes were dry, another layer of varnish was applied for sprinkling. In some places the previous varnish wasn’t dry totally, causing fusion of the layers. Brass dust, which is copper and zinc alloy, was sprinkled on the wet varnish, then sealed with another layer of varnish.

The finish is probably a later restoration work. Kireelac was applied and it flowed into the cracks of the previous generation of coatings. Mineral wax was also detected on the surface, possibly used as finish.

This japanned blue chair conforms to its attributed date, ca. 1735. The materials and techniques employed on this chair are apparently originated from Stalker’s and Parker’s publication, however, the simplification of the applications like less layers of paints, and change of material, such as gum water and red bole were no longer used, multiple variations like the replacement of gesso ground and pine resin with little sandarac for varnish, all indicate this chair is more consistent to the 18th century literature of English japanning.

This technical study provides preliminary investigation of materials and techniques of the English blue japanned chair, c. 1735. In order to develop better understanding of the craftsmanship about English blue japanning during that period, comparison with other similar examples, or even American blue japanning, will definitely contribute to the research on this subject. In terms of that, further identification of the binders and the raised decoration, as well as confirmation of smalt for this case are still necessary.