Edward Hopper: materials and method
A technical report from documented sources. "Uncertain" marks claims that are unverified, disputed or taken from a single object. The evidence assembled here comes mainly from his record books, a 1959 interview, his wife's diary as quoted by a biographer and the technical notes of a few museum catalogues. Much of the record-book evidence and the biographical evidence was read through search excerpts. The two 1939 museum technical accounts were also read through search excerpts; the preparatory-drawing and 1935 conservation reports were read through search summaries.
Summary
Hopper's mature oils are on commercially primed canvas, mostly linen. He bought it ready-primed from colormen and did not prepare his own grounds. He worked out the design in drawings on paper first. Then he drew on the canvas in charcoal or graphite; in one recorded case he went over the lines in blue paint. He started with paint thinned almost wholly with turpentine and added linseed oil as the work went on, "until the medium becomes pure oil," using "as little oil as I can possibly help". Examined canvases include successive layers, mostly wet over dry in one example, and thin opaque paint with localized thick passages in another. Record-book entries report several weeks of painting for individual canvases. He "rarely" glazed. He scraped paint off to correct passages and applied only a retouching varnish, never a final one. His record books name the canvas, the brand of colors, the white and the oil for most paintings from the late 1930s on, and individual pigments now and then. The documented pigments are listed in §4.
Terms. Ground/priming: the preparatory layers between canvas and paint. Single/double prime: a canvas given one or two coats of priming by the maker. Tacking margin: the canvas folded over the side of the stretcher. Squaring: a grid ruled over a drawing and over the canvas so the drawing can be copied square by square at a larger size. IRR (infrared reflectography): imaging that shows carbon-based drawing under paint. X-radiograph: an image made with X-rays, which shows lead-rich paint. Wet over dry: a new layer laid on paint that has dried. Glaze: a transparent layer. Traction crackle: wide cracks that form when a layer shrinks over paint that was not yet dry. Cross-section: a microscopic sample cut through all layers. Metal soaps: compounds that form when metal ions from a pigment react with the fatty acids of the oil.
1. Supports
- Linen, bought from colormen. In 1959 he reported buying linen from a London colorman. He did not make his own stretchers. In 1966 he recorded a "thick single white prime linen canvas" from the same maker.
- Named canvases in the record books. The entries name several commercial canvases by grade and origin: "Belgian canvas" (by 1938); "Belgian smooth, double prime canvas" (1939–41); "Winton canvas," a grade of the London colorman (late 1930s to 1943, and 1958); an American colorman's "tested artist's canvas (linen)" (1944); a Belgian colorman's "heavy linen canvas" (1944); "Irish linen primed in America"; the London colorman's "(National) double prime canvas" (from 1949); and its "Herga" single-prime linen (1956–66). One entry of 1945 reads "Linen canvas, smooth prime of doubtful character".
- Choosing canvas. For one picture he spent "several days trying to locate canvas with a surface he liked that he believed strong enough" before starting. Single account.
- Weave. Two examined canvases of 1939 are medium-weight plain weave. One is described as single-thread. Read through search excerpts.
- Formats. The Whitney holds 222 of his paintings. Of these, 121 are on canvas and six on linen. The others are small panels of paperboard, wood, composition board or canvas board, mostly about 24 × 33 cm, and most of them date from his student years, his stay in Paris (1906–07) and the late 1910s. Most of his canvases there from the 1920s on measure between about 64 × 92 cm and 102 × 153 cm; a few are smaller. One of 1939 measures 91.9 × 127.2 cm. His biographer mentions a "ten-by-twelve-inch sketch on canvas board" (1914) and "small paintbox-panel-sized studies" made one summer "for quick preliminary effects".
- Original size kept. Both examined canvases of 1939 keep their tacking margins, so they are close to their original dimensions. Read through search excerpts.
2. Grounds
- Commercial priming. "I use the ground of the already prepared canvas. I can't tell what this ground is exactly … I paint directly upon it".
- Color and thickness. On one canvas of 1939 the ground is a "thin cream-colored layer that does not obscure the weave". On another of the same year it is white. In both it runs over all the tacking margins, which shows it was applied before the painting, by the maker. Read through search excerpts. The record books distinguish single from double priming. In 1966 he preferred a thick single white priming, "smooth as double would be, but less apt to crack than double".
- Ground left visible. On one canvas, thin early layers leave the ground showing through. On another, he scraped paint away down to the ground and exposed the canvas texture. Read through search excerpts.
3. Drawing and lay-in
- Drawings first. Hundreds of his drawings in conté and charcoal survive, most of them preliminary studies. Before one canvas of 1940 his wife noted that he "is doing things with no end of preparation" and "had 2 highly finished crayon sketches". A museum report describes extensive preparation before painting and few alterations in X-rays of two canvases. Read through a search summary.
- Charcoal on the canvas. For several canvases his biographer, drawing on his wife's diary, records that he had the canvas "drawn in charcoal" before painting.
- Blue paint over the charcoal. She also noted that he "had charcoal drawing half covered by permanent blue in which he covers charcoal lines". "Permanent blue" was a trade name for synthetic ultramarine. X-radiography of a canvas of 1943 found only a partial underdrawing, with outlines of some areas in dark blue paint. Uncertain (read through a summary).
- Squaring and pencil. On a canvas of 1939, graphite squaring lines show along the edges and faintly through the paint. They suggest the design was transferred from a smaller drawing or a photograph. On another canvas of the same year, IRR shows underdrawing in two places. Pencil marks set the ends of a series of horizontal lines, and an incised line fixes one vertical edge. Read through search excerpts.
- Lay-in thin. He began "with almost pure turpentine" as the medium. On one canvas the first layers are "opaque but quite thin" and let the ground show through.
4. Pigments
The record books name the colormen for most paintings and individual pigments now and then. Many of the color words in them describe the picture rather than name a paint. He reported using about twelve or thirteen colors.
| Pigment | Evidence | Notes |
|---|---|---|
| Lead white (flake white, "silver white," "Cremnitz white") | "I use only lead white now" (1959); record-book entries from 1938 to 1966 | Bought as flake white (including "#1 & 2" in 1944), a Belgian "silver (lead) white" (1939–41) and Cremnitz white (1950s) |
| Zinc white | Early record-book entries pair zinc white with poppy oil; it is listed for a canvas finished in 1935; he also reported using it in one passage of another canvas | Given up for lead white |
| Cerulean blue | Named twice in one entry, the paint from a Belgian colorman; twice more in other entries | |
| Cobalt blue | Named in several entries | |
| Synthetic ultramarine | "Ultramarine" in several entries; "permanent blue" in his wife's diary | |
| Viridian | "veridian," "verid-ian" in two entries | |
| Cadmium yellow | "cad. yellow light" and "cadmium" in entries | Grade usually not given |
| "Cadmium deep" | Two entries | Probably a deep cadmium yellow or orange (inference; cadmium names ran from pale yellow to "orange" for darker shades) |
| Light red (a red iron oxide) | One pigment list in an entry; a second mention may be a color word | |
| Yellow ochre | One pigment list in an entry; other mentions may be color words | |
| Black | "touch of black" in one entry | |
| Burnt sienna | Two entries | Uncertain: possibly a color description, not a pigment |
- Brands. He named one London colorman as his maker in 1959. The record books show a Dutch colorman's paints in the early entries and in 1938. From 1939 to 1945 he used paints from a Belgian colorman together with the London colorman's. After that he used the London colorman's alone. When he changed to the London colorman, his wife took over his remaining Dutch paints, "including the zinc white, as Edward changed to lead white".
- A fugitive yellow. On one canvas of 1939 a green passage has turned blue where it was exposed to light. Paint under the frame rabbet keeps the original green. The cause is probably a fugitive yellow in the mixture. Inference by the cataloguers; the yellow was not identified. Read through search excerpts.
5. Media and varnish
- Turpentine to oil. "I use almost pure turpentine to start with, adding oil as I go along until the medium becomes pure oil. I use as little oil as I can possibly help".
- Linseed, earlier poppy. In 1959 he reported discontinuing poppy oil because of concerns about cracking, and using linseed oil and lead white. The record books match. Early entries pair zinc white with poppy oil. Entries of 1939–45 name poppy oil with the lead white. Linseed oil appears in 1938, linseed oil with turpentine in 1943–44, and linseed oil in every entry seen from 1947 on.
- Two oils, two whites in one painting. One entry of the later 1930s reads "1st lead white and linseed oil. 2nd Zinc white and poppy oil". Another, of 1944, reads "Started with [a New York supplier's] linseed oil, finished with [the London colorman's] poppy oil".
- Slow drying. He once wrote that a new linseed oil "seems to take a long time drying … it is still tacky about a month after I last worked on it". Single account.
- Handbooks. He said he received no instruction in materials at art school. What he knew came "from perhaps experience or having read" published painting-materials handbooks.
- Varnish. "I never use a final varnish. I use a retouching varnish which is made in France … and that's all the varnish I use." He applied it "almost at once, as soon as there is no tackiness in the pigment." Its purpose: "it brings out the areas that have dried in and gives them their proper value." A final varnish he left to later owners or restorers. The two canvases of 1939 now carry synthetic varnishes from later treatments.
6. Layers and application
- Direct and in layers. He reported painting directly on the prepared canvas. On one canvas the paint was applied "in layers, mostly wet over dry". Read through search excerpts. A conservation report describes eight or more layers in one sampled passage of a 1935 canvas. Read through a search summary.
- Thin and opaque, with a few thick passages. On one canvas of 1939 most of the paint is in "opaque but quite thin layers." One passage is built up in many thick layers, repeatedly applied and dragged with the brush into a convoluted texture. Another is laid on thickly with a palette knife. Read through search excerpts.
- Few glazes. "No, I rarely use glazes, rarely".
- Reserves. On one canvas he left unpainted reserves around some forms and then emphasized the edges by scraping away ground to expose the canvas texture. Read through search excerpts.
- Graphite over paint. On one canvas of 1939, fine lines were drawn with graphite pencil over the paint. Read through search excerpts.
7. Working sequence and reworking
- Order. Accounts describe preparatory drawings and drawing on the canvas; one diary account records blue paint over charcoal lines. He reported beginning with turpentine-rich paint and adding oil as work progressed, then applying retouching varnish once the surface was no longer tacky.
- Duration. The record books give "1 month painting" and "About 4 weeks painting" for single canvases; a third reads "1 2 month painting," probably one and a half months (OCR reading). His wife wrote of one canvas: "Each day I don't see how E. can add another stroke". On another: "Each day one thinks the picture done, but no".
- Scraping out. His wife recorded passages being scraped off and repainted. He expressed concern that covered paint might become visible through later layers as they aged.
- Few changes on the canvas. A museum report describes few alterations in X-rays of two canvases. Read through a search summary. A canvas of 1943 shows changes mainly in one position and in some outlines. Uncertain (summary).
- Late additions. One early canvas had a passage added in June 1959, decades after it was begun.
- Studio work. The record books note where each canvas was painted, usually his New York or South Truro studio. In Paris in 1906–07 he painted outdoors almost every day, on small panels of about 24 × 33 cm and canvases of about 60 × 73 cm.
8. Condition and aging
- Zinc white. He believed zinc white could scale and crack, and thought a zinc-white passage in one returned canvas had required restoration. On a canvas of 1935 whose record-book entry lists zinc white, a conservation report describes extensive old flaking, fractures through and between its layers, and zinc soaps. Read through a search summary. In oil, zinc white "tends … to form hard, often brittle films" that crack over softer layers.
- Traction crackle. In the thickest passage of one canvas, wide traction cracks expose an earlier, darker layer. It was probably not fully dry when the next layers went on. Read through search excerpts.
- Faded yellow. See §4: a green turned blue where it was exposed to light.
- Linings. Both canvases of 1939 were wax-lined and put on new stretchers: one in 1964, the other early in its museum life and again in 1980. Read through search excerpts.
- Few other problems. In 1959 he knew of only one painting that had needed attention. His own account.
9. The tubes here
The historical names in this note are mapped to these tubes: lead white (flake, silver or Cremnitz white) → lead white; zinc white → zinc white; cerulean → cerulean blue; cobalt → cobalt blue; ultramarine and "permanent blue" → ultramarine blue; viridian → viridian; "cad. yellow light" → pale cadmium; cadmium yellow → cadmium yellow; "cadmium deep" → deep cadmium (a deep yellow-orange, inferred); light red → red earth (a proxy); yellow ochre → yellow ochre; black → bone black (a proxy); burnt sienna → burnt sienna. Burnt sienna rests on an uncertain reading (§4).
10. At this easel
- Support. Plain-weave linen with ground layers laid from the tubes.
- Paint. Pigment from the tubes with an oil-medium share.
- Taking paint off. A brush picks up wet paint as it passes; that is how paint comes off the canvas.
- Everything else the easel does is in the easel guide.