John S. O’Connor was one of this country’s foremost glass cutters during the nineteenth century. His accomplishments were many and were not limited to the design of cut-glass patterns. Nevertheless, when his name is mentioned, it is this aspect of his work that one thinks of first, and the mind’s eye conjures up his highly regarded Parisian pattern of 1886. A re-examination of this pattern and its patent is long over-due. The complete patent (no. 16,676) is available at the USPTO Web site.

From the time of renewed interest in American cut glass — which is usually thought of as beginning with the publication of Dorothy Daniel’s book on the subject in 1950 — the claim has often been made that the Parisian’s patent was the first one in which curved miters were cut. This claim is only partly true because when the Parisian pattern is cut on horizontal (flat) surfaces — and its designer surely intended for it to be used in this fashion — its miters are straight. Only on surfaces with a vertical component are the pattern’s miters curved. This dichotomy — a pattern that can be both rectilinear and curvilinear — is why O’Connor’s Parisian pattern is truly innovative, why it has been misinterpreted and why this file undertakes an in-depth re-examination of the pattern.

Background: Straight- and Curved-Miter Cuttings

The difference between straight- and curved-miter cuttings, from the cutter’s point-of-view, is explained by the English authority E. M. Elville as follows: Straight-miter cutting is produced “when the article is held firmly in line with the edge of the [cutting] wheel. If the article, while being cut, is deflected out of line with the edge of the wheel, however, a cut with a curved outline will be the result” (Elville 1967, p. 71). It follows from this that a miter cut that is equidistant from the center of a dish, such as the 360 degree miter cut around the rim of a dish — as required by many row-type patterns — is a straight-miter cut, not a curved cut, because the dish is always “in line with the edge of the wheel”. An example of this are the four concentric circles that are shown in Philip MacDonald’s patent of 1882 (no. 12,982).

Elville concludes that cutting the curved miter is “a much more difficult technique than the more usual straight-line cutting, and is one that is attempted only by highly experienced craftsmen.” This helps explain why straight-line cutting dominated America’s early and middle periods of cut glass with the curved miter cut playing a relatively minor role until late in the nineteenth century. By the 1880s the ability of cutters had reached a relatively high level of expertise, however, and, largely in response to the demands of that affluent part of the public that wanted increasingly elaborate cut-glass patterns (and could afford to buy them), patterns containing curved miters were introduced, supplementing the many patterns that used straight-miter cuts exclusively.

The Role of Perspective

It will come as no surprise to anyone familiar with the composite Dorflinger catalog that there are many examples of the Parisian pattern that have straight, rather than curved, miters. The following image of ice-cream tray no. 1030, probably cut about 1890, is one example. The miter cuts are straight where the tray is perfectly flat, an area that comprises most of the tray’s surface. But, along the upper and lower edges, where the tray curves sharply, the miters are curved. This curvature is not real, however. It is an illusion brought about entirely by the fact that in these areas the pattern has been cut on curved surfaces. Because of perspective, the camera “sees” the straight miter cuts as curved cuts. Over most of the tray, however, perspective is essentially zero. Consequently, the miters are straight. L = 14.5″ (36.8 cm), W = 8″ (20.3 cm) Sold at an eBay auction in 2000: price not recorded (Image: Internet).


The patent’s single illustration, which is reproduced below on the left, also shows curved miters, this time on a dish. In this case, however, the curvature is real, not illusory. It is probably this image, more than any other, that has convinced many dealers, collectors, and research workers during the past 50+ years that the Parisian patent is exclusively concerned with curved miters. Nevertheless, it is difficult to believe that these gentle curves could have inspired a major development in American cut-glass design, as is often claimed. Had O’Connor chosen to illustrate his patent with the plate below, on the right, or for that matter with the ice-cream tray, in addition to the dish with its curved miters, the pattern’s dichotomous nature would have become obvious, and the “curved-miter” claim tempered. The plate, probably Dorflinger’s no. 233, is a nearly flat underplate with a diameter 6″ {15.2 cm) that is part of a punch-cup-and-saucer set.

dp16676 parisian3

The Parisian Pattern’s Dichotomous Characteristic

It should be noted that, although the title of the Parisian patent is “Design for a Dish,” the patent’s specification adds “and other Articles of Table-Ware” (note 1). In the context of the patent these articles must all be, to borrow a term from silverware, “hollowware”: vases, decanters, stemware, etc. — shapes derived from a cylinder — because when the pattern is cut on these shapes it has curved miters. On the other hand, when the pattern is cut on “flatware”: plates, trays, saucers. etc. — shapes derived from a disk — its miters are straight.

Note that the patent’s specification contains expressions like “intersecting angular lines” and “intersecting lines meet at angles” which imply linear constructions. There is no mention of curved miters per se (note 2). Yet, there is no doubt that O’Connor intended that his pattern be cut on both holloware and flatware. Perhaps this hollowware/flatware dichotomy is the reason why O’Connor worked on the pattern for at least four years before submitting it to the Patent Office (Feller 1988, p. 203). If so, it is a disappointment that he did not choose a more distinctive example of hollowware — a bowl perhaps — and then proceed to illustrate it with top and side views (as he later did in his patent for the Tuxedo pattern). When O’Connor recognized that the Persian pattern is dichotomous, he should have added a representation of it on flatware. As it stands, the patent is incomplete.

Transformation: Straight Miters to Curved Miters

Ice-cream trays such as the one at the beginning of this file were often ordered with matching individual ice-cream plates. The following image is of plate no. 1036, 6″ (15.2 cm) on a side, c1890, which was sold with that tray (Image: Internet). Note that the design in the center of the plate, a flashed double-mitered hobstar, is the same as that on the matching ice-cream tray, greatly reduced in size and simplified. The Parisian pattern, in other words, is nothing more than a flashed double-mitered hobstar, beaded, with fans, strawberry (fine) diamonds, and stars added to the pattern’s interstices if there is sufficient space. In this, its most basic form, the Parisian pattern is clearly seen to be linear.


This plate is of particular interest because it has a compound shape: one that is composed of a flat base (disk) and a rim derived from a cyclinder. It is, therefore, possible to cut the Parisian pattern two ways on the same item: with straight miters in the center of the plate, as on flatware, and with curved miters on the plate’s rim, as on hollowware. The straight and curved miters on the plate are determined solely by the plate’s geometry. Perspective is not a significant factor; the miters, therefore, are real.

The transformation of the square plate’s straight miters to curved miters is demonstrated in note 3. Real-world examples of this graphic construction are also illustrated by several examples of Hawkes’s Grecian pattern in the HAWKES PHOTOGRAPHIC FOLIO (ACGA 2003) — particularly those on pp. 12 and 175. Because of its vesicas, Hawkes’s Grecian pattern is curvilinear on both hollowware and flatware, and, therefore, the pattern as a whole is not dichotomous. However, the vesicas are extensions of the pattern’s basic hobstar design which is dichotomous: The cited examples show that the hobstar’s miters are straight on plates and curved on stemware (p. 12). The champagne jug on p. 175 — one of Hawkes outstanding achievements — also clearly shows the curved miters of the hobstar when the Grecian pattern is cut on hollowware.

Our final example of the Parisian Pattern (below, left) is compared with a contemporary pattern, Cobweb by Hawkes (below, right). Both illustrations are taken from the composite Dorflinger catalog issued by the ACGA in 1997 (p. 13). The two patterns have much in common, as well as basic differences, and are worth a brief discussion.

parisian5 cobweb5

Uncertain dating is a problem with both patterns. Different dates have been suggested for the invention of the Parisian pattern: for example, 1880 William Dorflinger and 1883 Dorothy Daniel, to which one can add 1882 (John Quentin Feller): “Company records indicate that O’Connor was cutting the pattern at least four years earlier [than the patent date, 1886]” (Feller 1988, p. 203). Another relevant quote from this source is the following: “There are no pattern names for the formative years [of the Dorflinger company] between 1852 and 1882, when John S. O’Connor designed the ‘Parisian’ pattern that was patented only four year later” (p. 108). Unfortunately, Feller provides no hard evidence for his date of 1882. Note also Daniel’s remark (in the foregoing link) that when the Parisian pattern was patented in 1886 the company “began cutting the pattern immediately,” in spite of her own contention that the pattern was “designed” in 1883. These conflicting dates suggest that, although the Parisian pattern could possibly have been cut as early as 1880, 1882, or 1883, it was probably not extensively cut until C. Dorflinger & Sons had secured the patent rights in 1886.

The Cobweb pattern, which was never patented, was introduced by Hawkes sometime between 1882 and 1885, at least a year before the Parisian pattern was patented. It could, therefore, pre-date Parisian, but, dating continues to be a problem. Cobweb was originally called pattern No. 643, but its name was changed to Cobweb “by 1888” (Spillman 1996, pp. 175-6, p. 184).

The Parisian and Cobweb patterns were the most popular patterns of the late 1880s that were among those that were ordered by C. Dorflinger & Sons from, and cut by, T. G. Hawkes & Company. This unusual arrangement was initiated by a strike at the Dorflinger factory early in 1888 and continued, at least sporadically, for several years thereafter (Spillman 1996, p. 92). Because rims with fan scallops are not present on the above illustrations of these patterns, it is impossible to determine whether they were cut by Hawkes or by Dorflinger. Dorflinger had written, however, that “. . . as we respect your rights in the Cobweb cutting . . . we have never cut any here . . .” (quoted by Spillman), even though the Dorflinger catalog contains examples of the Cobweb pattern where it has the name Hawkes originally gave the pattern, No. 643.

Both the Parisian and Cobweb patterns are curvilinear when cut on hollowware and rectilinear when cut on flatware. Parisian’s linearity has been demonstrated earlier, but no such representation of the Cobweb pattern is known to the writer. Therefore, it has been necessary to sketch the pattern based on its appearance on the above carafe. The sketch shows the pattern’s linearity when cut on flat surfaces. Thus, Cobweb joins Parisian in dichotomy-land. Incidentally, one can see that the “web” is nothing more than a triple Brunswick star:


The Cobweb pattern is fully comparable to the Parisian pattern in terms of quality and curved miters, and, as indicated, it could have been cut earlier than Parisian. Yet it is the Parisian pattern that has received more attention over the years. Interestingly, the complete Cobweb pattern, but with a triple hobstar in place of the triple Brunswick star, appears as part of Hawkes’s Pillars & Rich Star pattern, which was introduced, again, “by 1888” (Spillman 1996, p. 185) where it forms the rich-star motif. A 6″D (15.2 cm) saucer in this pattern, shown on p. 190 in Spillman’s book has remarkably detailed cutting in the saucer’s small, central area (note 4). It appears to the writer that both the Parisian and Cobweb patterns would have required the extensive use of a marking tool prior to their being cut, especially on hollowware.

Today, much is made of the Parisian pattern. It has a respectable rating in Pearson’s “Relative Value Guide” (Pearson 1978, p. 250) — a price guide that is often referred to by sellers on eBay — while the Cobweb pattern is not even listed, probably because it was rarely seen in the 1970s when this list was assembled. Not being on it has not helped Cobweb’s cause, at least among those collectors who base their buying habits on “the numbers” in Pearson’s list.

There is no doubt that the Parisian pattern has benefited from public relations to a much greater extent than has the Cobweb pattern, beginning with William Dorflinger’s address to the American Association of Flint and Lime Glass Manufacturers in 1902. But it must also be noted that Parisian is the more versatile of the two patterns, having been used on a wider variety of shapes and having additional versions that use single and triple miters respectively. This type of comparison is somewhat unfair, however, because Cobweb lacks the considerable amount of catalog material that is available for the Parisian pattern and which is the basis for much that we now know about the pattern.


Ultimately, we can ask: Are the Parisian and Cobeweb patterns curvilinear? The answer: yes and no. Yes, if cut on hollowware and no, if cut on flatware. And what about the Parisian pattern’s patent? Answer: Yes, it is curvilinear but only because the patent illustrates the pattern as cut on hollowware. Had flatware been represented instead, the patent would be rectilinear.

A more appropriate question might be: Which American patent was the first to have curved-miter cut on both hollowware and flatware? The answer would be: Hawkes’s Grecian pattern of 1887 because of its vesicas; they appear on both flatware and hollowware. Next in line: “Baker’s Gothic” by Clark of September 1888, followed by O’Connor’s first patent with curved miters on both hollowware and flatware, the Florentine pattern of November 1888, first cut by Dorflinger then by O’Connor. The following month saw the patenting of Straus’s popular Venetian pattern, which is also curvilinear on both types of tableware. In the which-was-first curved-miter sweepstakes, all of these patterns are trumped, however, by earlier domestic and foreign cut glass that used “difficult” curved-miter cuttings long before these patterns were even thought of. But the earlier patterns were not patented.

Contrary to what has been written, and based on the foregoing analysis, the writer believes that the Parisian pattern itself did not set in motion a rage for patterns containing curvilinear miter cuts. (And this is probably the main point of this file.) Patterns with this characteristic appeared essentially simultaneously within the American cut-glass industry during the middle and late 1880s. This seems to have occurred because proficiency in glass cutting had reached a high level of expertise and because cut glass had become increasingly popular with affluent consumers. The resulting intra-industry competition sparked the search for “something different” and the curved-miter cut answered this search admirably. In doing so it helped usher in the middle, or mature, phase of the brilliant period in American cut-glass design. Dorflinger’s Parisian pattern played its part, but so did other patterns, not all of which were patented.


1. There is a typographical error in the patent’s specification on line 11: “disk” should read “dish”.

In an earlier draft of this file the writer stated that the patent’s illustrated dish was derived from a disk. Wrong. The dish was derived from a cyclinder. Consequently, its miters are curved, not straight. In addition, it was originally thought that this curvature was illusory, the result of perspective. Wrong again!

2. Because perspective is present in practically all of the illustrations that accompany the design patents at the Patent Office, not too great a weight should be placed on the wording found in the patents’ specifications. They are written by lawyers in reference to illustrations and instructions provided by the patentees. In most cases these are perspective views of the objects described (although frequently called “plan” views), and, therefore, they have distortions, including “curved” miters.

It should be emphasized that a false conclusion can easily be reached when perspective is ignored. This is what happened to Dorothy Daniel in her discussion of Libbey’s patent no. 17,072, the Victoria pattern. We know that she had access to the patent because, not having the pattern’s catalog name, she assigned the name “Middlesex” to it. This is the county in Massachusetts where the patent’s designer, William C. Anderson, lived at the time of the patent, and this information is readily available only in the patent’s specification. The specification includes the phrase “crossed and overlapped arches,” a feature clearly shown in the patent’s illustration. But in both cases — text and illustration — Daniel failed to realize that these arches are illusions. It is, therefore, not surprising that she erroneously concluded that the Victoria pattern “is one of the first of the curved miter patterns ever cut” (Daniel 1950, p. 197). In his discussion of the Victoria pattern, the late Carl U. Fauster, an authority on the Libbey company and its products, passes along Daniel’s error without correction (The Hobstar, Vol. 8, No. 7, pp. 9-10 [May 1986]). Any curved miters seen on catalog representations of the Victoria pattern — and there are many in the composite Libbey catalog (ACGA 1996) — are illusions.

3. The following rough sketches illustrate what happens when a hobstar in the base of a plate (flatware) is cut on the plate’s rim (hollowware), as in the case of the square plate. In the second sketch, the double-mitered hobstar of the Parisian pattern, flashed and beaded, is simplified to form a 24-pt hobstar, half of which is shown. The linear miter running from point 0 to point 9 has been chosen to be cut on the plate’s rim. Radials are drawn from the center of the hobstar’s hobnail to points 0 to 9, and the distance from the edge of the hobnail to each of the points, 0 to 9, as projected onto the 0-to-9 miter, is measured and plotted on the graph in the first sketch. The curve that results is the linear 0-to-9 miter of the original 24-pt hobstar transformed into a U-shaped miter that is tangent to the hobnail at 4.5, the point along the 0-to-9 linear miter that is equidistant from points 0 and 9. The two sketches are placed in reverse order to reflect the relative positions of the plate’s upper rim and base to which they can be compared.

mitergraph2 mitergraph1a
4. The triple hobstar in the rich-star motif is sometimes replaced with a double hobstar when this motif is used in the “Russian & Pillar” pattern, a pattern-name that has not yet been confirmed.