Home Dealers Calendar Articles Fine Art Database About AFA Login/Register
Home | Articles | Two Gaming Tables: A Comparison

By Gregory J. Landrey

Gaming table, maker unknown, 1755-1775, possibly later. Mahogany, white oak, white pine. H: 30-1/8, W: 32-1/4, open D: 31-1/2 in. 1960.1055.1.

A contemporary definition of the word pair is, 'Two identical, similar, or corresponding things,' whereas the Royal English Dictionary published in 1804 defines a pair as 'two things suiting one another.' Two turret-cornered carved Philadelphia gaming tables in the collection of Winterthur Museum and Country Estate, acquired by Henry Francis du Pont sometime between 1925 and 1950, were considered to be a pair by Mr. du Pont, who displayed them as such. Whether they should be classified as a pair may depend on which definition we choose to accept.

On display as a pair in Winterthur's Stamper Blackwell Parlor since the house opened as a museum in 1951, the tables were presumably built in Philadelphia between 1755 and 1775.1 A design for a form similar to these card tables can be found in the 1760 publication of Ince and Mayhew's Universal System of Household Furniture.2 Furniture scholars Alan Miller and Luke Beckerdite have identified Nicholas Bernard and 'the Garvan high chest carver' as the possible artisans responsible for the carving on two tables that relate to the Winterthur examples, but firm attributions for cabinetmakers or carvers of the museum's tables have not been made.3 Some assessments have in fact suggested that one of the two tables may be of a later date.4

Gaming table, maker unknown, 1755-1775. Mahogany, white oak, white pine, H: 30-3/8, W: 32, D: 31-1/2 in. 1960.1055.2.

Unfortunately, little is known about the provenance of the Winterthur gaming tables from the time of their construction until their purchase by Mr. du Pont, whose early acquisition records do not indicate whether they were purchased at the same time or from whom they were attained. The catalogue American Rococo, 1750-1775: Elegance in Ornament (1992) notes 'there is but a handful of these tables, all apparently made in pairs, and all with more or less similar carved designs.'5 Related examples are published in Horner's 1935 Blue Book, Philadelphia Furniture,6 and are in the Chipstone Collection, the Colonial Williamsburg Foundation, the Kaufman Collection, and Bayou Bend Collection. The Kaufman and Williamsburg tables are identified as mates and understood to be purchased from the estate of John Penn in 1788 by Joseph Parker Norris and later descending in the Vaux family, all of Philadelphia.7 The overall form and central front apron 'shell' of the Winterthur tables is most similar to the Easby family table seen in plate number 234 in Horner's Blue Blook.8


Studies of other forms of card tables traditionally classified as pairs have revealed differences between them.9 The assessment of the two Winterthur gaming tables that follows is intended to determine their degree of 'pairness.'


Both tables are made from mahogany, white cedar, white oak, and white pine.10 The techniques of construction for both include a dovetailed box, mortise-and-tenon swing legs, and round tenons on the front legs that fit into corresponding sockets in the corner turrets. The overall appearance of the carving and the dimensions are also similar. Both tables have evidence of repairs that are not unusual for this form, including reworking of hinges, reattachment of the tops, rebuilding of a damaged front leg joint on table 2, replacement of sections of carving on both tables,11 and reinforcement of the drawer dovetails on table 2. Empty nail holes in the top leaf of table 1 and corresponding nail remnants in the swing leg indicate that this table was fixed in the open position for a period of time prior to its acquisition by Mr. du Pont.


Table 1, open swing leg. Inner surfaces without added color. Partial pocket to receive case and carved rail, bottom left side of the image. Side rail beveled, right side of the image.

Joinery: Both swing legs are joined to a rail with pegged double mortise and tenons. However, X-rays taken of the joints highlight a distinctly different technique of chopping the mortise and shaping the tenon. The floor of the mortises in the swing leg of table 1 have an atypical domed profile, which is outside of the norm but most likely of handwork origins. The bottom, or floor, of the mortise in table 2 has an irregular, choppy profile that is consistent with similar Philadelphia cabinetwork.12 The tenons are chamfered on table 1 but not on table 2. The mortise-and-tenon profiles in the two tables are the product of significantly different techniques, suggesting that different craftsmen with different approaches to joinery and tools cut these joints.

Swing legs and side rail juncture: The oak stock for the rear swing rail is 1/8 inch thinner and 3/4 inch narrower on table 1 than the corresponding part on table 2. The upper portion of the mahogany swing leg is fashioned on both tables to receive the corresponding corner end of the side rail and apron when the table is closed. However, table 1 has only a partial allowance for the fitting of the side rail into the swing leg and requires significant chamfering of the mahogany side rail to allow for it to slip into the semi-closed housing. The swing leg on table 2 is built to completely surround the end of the side rail as the two parts come together when the table is closed, allowing the end of the side rail to fit snugly into the housing without the need for any chamfering on the rail.

Table 2, open swing leg. Inner surface with added color, presumably original. Complete pocket to receive case and carved rail, bottom left side of the image. Side rail without bevel, right side of the image.

Coloring of secondary woods: It was common for the interiors of the swing leg on card tables to be colored so that these structural parts blended with the primary woods when the table was in the open position. Table 1 is lacking in coloring, and so the lightness (and lesser quality) of the oak and pine are apparent when the leg is open. The interior faces of the swing leg and the rear rail on 2 are colored to look like mahogany, providing a more finished appearance when the leg is open and the top is unfolded. This variation on the handling of the appearance of visible secondary woods with these two tables indicates a difference in the presentation of the tables in the open position.

Underside of table 2. Support block behind carved apron cut at a 90 degree angle. Wide knee return laps over on to the inner rail.

Underside of table 1. Support block behind carved apron cut at an 80 degree angle. Narrow knee return located on the outer rail does not lap over on to the inner rail.

Interior framing: The structure of the interior varies with the two tables. Examples include the drawer housing which is open on table 1 and boxed on table 2. In addition, the plane of the rails behind the lower apron carving is angled on table 1 at approximately 80 degrees, whereas the corresponding part on table 2 is set at 90 degrees, or perpendicular to the top.

The gaming surface: Neither table has any surviving fabric on the playing surface, although the top boards were skillfully relieved to receive a baize or similar fabric. The surface where the fabric should be on table 1 is plain and without tooling or blemish other than hinge repairs at the center edge. The surface of table 2 has the marks of the toothing iron or plane where fabric was once attached. The oval banding that defines cups for counters used in gaming are separate entities on table 1 and do not connect with the outer banding that serves as a perimeter for the gaming area. On table 2 the oval banding around the cups and the outer banding are connected. Interestingly, the cups on table 1 are positioned to the left of the sitter and are to the right of the sitter on table 2. This could be a random variation, an unintentional reversal that occurred when a pattern was transferred from one table to the other, inattentiveness by the maker, or the desire to have both manners of card playing represented in the two tables.

Aesthetic: The carving on table 1 is shallower than that on table 2. Measurements indicate that the extent of relief on table 1 in many areas is about one-half that on table 2. As an example, an area of leafage on table 1 is approximately 2 mm above the surface, while the same pattern on table 2 rises 4.5 mm above the surface. The design on the turrets also differs in the layout, particularly with the diapering. There are only two complete diamonds on table 1 (one shown), but table 2 has six complete diamonds in the diapering pattern on the turret (four shown). It is interesting that the central carved foliate 'flip' on the front apron points to the proper right on both tables. A common rococo theme for objects made as a pair was to orient such ornamentation in opposite directions, resulting in a balance in the asymmetry.
LEFT: Table 1 corner turret, with shallower carving; shown is one complete diamond in the diaper pattern. RIGHT: Table 2 corner turret, with deeper carving; shown are four complete diamonds in the diaper pattern.

Ball-and-claw feet: The carving of the ball-and-claw feet is presented differently as well. The finger and knuckle segments on table 1 are more or less perpendicular to the floor, with the lower part of the 'finger' flaring outward at its termination, where the almost invisible talon turns under the ball. The carving of the foot on table 2 is created with a knuckle projecting outward at the midpoint of the foot generating the visual of the fingers clutching the ball.

Finish: Table 2 was treated in 1976 with a system designed to remove linseed oil and other coating materials. The finish appears cleaner, more transparent and thinner than that on table 1,13 which has been left largely as it was when Mr. du Pont acquired it, and as a result, appears darker, more opaque, thicker, and of greater age. So, although there is a significant difference in the surfaces of the two tables, it is largely irrelevant for the purposes of this study and cannot be used as an indicator of age.14

Central skirt pendant, split image. (LEFT): Table 1; (RIGHT) Table 2. Table 1 shows indiscernible punch work; C scrolls terminate in a complete circle. Table 2 has prevalent punch work in a cross-hatch form; C scrolls terminate in a spiral or volute form. The lower section is restored.


It is clear from a comparison of the construction and carving of these tables that the artisans took a different approach in almost every step of design and execution, even though the overall end products are similar. I'd like to suggest three possible positions that might explain the physical evidence.

1. The tables were made at the same time in the same shop by different craftsmen. This scenario is possible, though not very likely, since it is inconsistent with the expected practices and efficiencies in a preindustrial cabinet shop producing objects of this caliber. For reasons of efficiency, such a shop would use the same patterns, stock, and/or measurement points when laying out the tables as a pair. If these tables were constructed as a pair in the same shop, the rear swing leg assembly on the tables would follow similar dimensions, housing and construction techniques; the rendering of the counter cups and strap-work on the top would be alike; the joinery work would be consistent; and the carving details would match closely.

Ball-and-claw feet, split image. (LEFT): Table 1; (RIGHT): Table 2. Table 1 shows squared off, upright knuckles. The lower digit is tucked in slightly. Table 2 has an outward projecting central knuckle. The lower digit is tucked in sharply as though clutching the ball.

2. One table was made earlier than the other, and the second was built by a craftsman employing preindustrial techniques, who used the first table as a reference. The physical evidence is consistent with a craftsman taking the measurements and details from a previously constructed object and then creating the second object. Such an explanation could account for the differences in carving and construction.

3. One table was made during the twentieth-century collecting period. This seems less likely as there are numerous repairs to both tables and indications of use suggesting that they existed for many years prior to their purchase by Mr. du Pont. In addition, all observed tool marks are consistent with preindustrial techniques. It is possible in the industrial era to make an object with traditional tools. However, while the patterns of workmanship on table 2 are more consistent with workmanship of the 1755-1775 period than those found on table 1, nothing was observed that objectively identifies table 1 as an industrial-era creation.


When provenance is lacking - as in the case of the Winterthur gaming tables - sorting through the physical evidence is essential to establishing the relationship between objects. In this case, some of the evidence is more objective, some less so. For example, the observation that the profiles of the mortises are distinctly different is both relevant and objective; the observations about the overall quality of the carving are important, but are more subjective. In the case of these two tables, both objective and subjective evidence
Finish of proper right side apron, split image. (LEFT): Table 1; (RIGHT): Table 2.

clearly indicates that these tables are different in numerous details, despite their general appearance of being alike. In the end, it may never be possible to state conclusively which of the three scenarios presented here most accurately represents the relationship of the two tables and whether or not they should be described as a true pair. However, the physical evidence supports the theory that the tables were not made as a pair, and that one was made with the other as a point of reference. How much time elapsed between the construction of table 2 and table 1 is unknown. But although the pair of tables do not fulfill the current requirement for being 'identical,' they certainly fit the 1804 description of a pair as 'two things suiting one another.'

Gregory J. Landrey is director of the Library, Collections Management and Academic Programs Division, Winterthur Museum and Country Estate, Winterthur, Delaware. All photography by Jim Schneck. Courtesy of Winterthur Museum and Country Estate.

1. Object file, 1960.1055.001 and .002, registration office, Winterthur Museum and Country Estate. See object file for condition, treatment, and analytical records referred to in the article.

2. William Ince and John Mayhew, The Universal System of Household Furniture (London, 1760), pl. LII.

3. Luke Beckerdite and Alan Miller. 'A Table's Tale: Craft, Art and Opportunity in Eighteenth Century Philadelphia.' American Furniture 2004 (Hanover: Chipstone Foundation, 2004), 12-13.

4. Object file, 1960.1055.1

5. Morrison H. Heckscher and Leslie Greene Bowman, American Rococo, 1750-1775: Elegance in Ornament (New York: Metropolitan Museum of Art, 1992), 195.

6. William Macpherson Horner, Jr., Blue Book, Philadelphia Furniture: William Penn to George Washington (Washington, D.C.: Highland House Publishers, 1935), pls. 234-35.

7. One of the tables from this pair is photographed in Horner's Blue Book, pl. 235.

8. The Easby table is also published by Luke Beckerdite and Alan Miller in American Furniture 2004 (Beckerdite and Miller), 12-13.

9. Mark J. Anderson, Gregory J. Landrey, and Philip D. Zimmerman, Cadwalader Study (Winterthur, Del.: The Henry Francis du Pont Winterthur Museum, 1995), 17, 23.

10. Object file, 1960.1055, microanalysis report, Gordon Salter, January 19, 1976. Salter records white oak and eastern white cedar as present. It is not clear from the report which of the two tables was sampled for the analysis.

11. The proper left side carved apron is likely a replacement made of cherry on table 1. The lower section of the pendant on the front apron is replaced on table 2.

12. Cadwalader, 21, 35, fig. 32.

13. See a similar treatment of a pair of chairs published in Conservation of Wood and Paintings in the Decorative Arts, Preprints of the Contributions to the Oxford Congress. (London: International Institute for Conservation, 1978), 59-61.

14. Due to the compromised nature of the surface on table 2, microanalysis was not conducted for this study. It is possible that selective cross-section analysis could generate some comparative information should remnants of an earlier surface survive on table 2.

Antiques and Fine Art is the leading site for antique collectors, designers, and enthusiasts of art and antiques. Featuring outstanding inventory for sale from top antiques & art dealers, educational articles on fine and decorative arts, and a calendar listing upcoming antiques shows and fairs.