The Flemish method

Van EyckThere are several current painting methods whose practitioners claim that they are working just like early Flemish painters did. In the early 1400’s, Flemish and Netherlandish painters (Jan van Eyck, Robert Campin, Rogier van der Weyden, and so on) invented oil painting as it is practiced today. In the 19th and early 20th centuries, art historians spent a lot of time speculating about how the first oil painters practiced their art. Unfortunately, they could only look at paintings and read old manuscripts. As it turns out, looking at paintings doesn’t tell you a whole lot about the methods by which they were made. And the early oil painters didn’t, so far as we know, write down their valuable trade secrets for scholars hundereds of years later to study.

So most of what you will see about “Flemish” painting methods, although sometimes used by very skilled artists, is basically what some 19th century German academics erroneously thought Northern European painting was. There is one school that alternates layers of egg-oil emulsion white with layers of oil glazes in primary colors. Another school calls for each painting to be made in seven layers, each of which dries for seven weeks (are we artists or numerologists?).

The reason why people are so fascinated by early Flemish paintings is that they are so remarkable. First, although they are very, very old, they are often in much better shape than oil paintings that are hundreds of years younger (many early 20th century paintings have deteriorated more than paintings from the 15th century). Second, the paintings glow with a sense of vitality that is seldom seen in more recent paintings. They have a quality that is often described as “jewel-like.” This is especially evident if you see them in person; it’s easy to pick out the Flemish paintings from those done in other periods, or those from the same period done in other places. The precise detail and gorgeous colors are marvelous to look at.

The thing is, we now know a lot more about early Flemish painting methods than we used to, because modern technical analysis provides a lot more information. So I can, with reasonable authority, describe actual “Flemish” painting methods, which were generally much simpler than those taught in special classes where the “Secrets of the Old Masters” are revealed.

The Northern European paintings we have are almost entirely done on panel, not because canvas wasn’t used, but because panel paintings are the ones that have lasted. The panels were made of local hardwoods such as oak. They are often quite small; the large ones were made with several planks fastened together. The panels were planed to a smooth finish and then seasoned (usually for several years). They were then coated in hide glue, which is a gelatin that becomes liquid when heated. Then an initial priming was applied that consisted of warm hide glue mixed with chalk (calcium carbonate). This priming was applied in several layers, allowed to dry, then sanded or scraped smooth. The modern word for a priming layer like this is “gesso,” although technically gesso was the Italian version, made from hide glue mixed with gypsum (calcium sulfate). Modern acrylic primer is often labelled “gesso,” but that’s totally incorrect. Anyone caught referring to acrylic primer as gesso will be required to stay after school and write “gesso is not made with plastic” 1000 times on the chalkboard.

There were very few pigments available at the time to make paint with, although that limitation is not obvious from looking at old master paintings. The available colors included ultramarine blue (very expensive as it had to be imported from Afghanistan and then laboriously refined), azurite blue (sort of like a cobalt blue), red lake (similar to alizarin crimson), lead tin yellow (similar to Naples yellow), vermillion (similar to cadmium red), bone black (similar to what is now, incorrectly, called “ivory” black), flake white, copper green (similar to viridian), and various earth colors (similar to modern ochres, siennas, and green earths). Paint was made by hand, usually that morning or the day before, by apprentices and studio assistants. If you haven’t tried oil paint made fresh by hand, it is more fluid than the stuff you get today in tubes.

In the early 15th century, oil painters probably didn’t use solvents such as spirits of turpentine or oil of spike. This demostrates that it is quite possible to apply oil paint with very fine precision without having to thin it down. It’s easier to do this kind of work with handmade paint than with tube oil colors.

This method involves multiple layers that are allowed to dry in between applications. In some cases, lower layers were done in egg tempera, switching to oil later on in the sequence. Flesh tones were painted very thinly, in one or two layers. The modern oil painting approach (mostly developed in Florence in the early 16th century) emphasizes thin, transparent darks and thick, opaque lights. By comparison, early Northern European painting was primarily a glazing method in which darks were established with thick layers of transparent paint. This allows the darks to have a sense of depth and color that is not possible in direct (single layer) painting. Glazing is typically done over a layer of opaque paint of similar color; red lake over vermillion, for example. Black, since it tends to produce dull mixtures, is not used except to darken earth colors (most artists think that the idea of avoiding black was invented by the impressionists, but no). Instead, darks are created by thick glazes and by what we would today recognize as complementary color mixtures (red lake could be darkened with ultramarine, for example).

Here’s one possible reconstruction of the painting sequence:
  1. Make a detailed drawing on paper and transfer it to a gessoed panel. Reinforce and elaborate the drawing with ink or dark paint. Northern European artists usually made very detailed underdrawings.
  2. The initial layer of paint, called the primuersel, is used to define the dark areas of the painting. The primuersel color is made by mixing black with earth tones such as red or yellow ochre; it is therefore similar to the verdaccio used for flesh tones in the traditional Italian egg tempera method (but without a green earth underpainting and not just in flesh tones). This layer was probably often done in egg tempera or tempera grassa. Either way, this layer should be used to define form, edges, shadows, and other darks throughout the painting.
  3. Now apply the basic colors to each area of the painting, starting to work up modeling of forms with opaque colors but avoiding fine detail. This stage is called “dead coloring.” As with primuersel, some artists used egg tempera or tempera grassa for this layer. Flesh tones can be begun with mixed tones using appropriate brown or pink mixtures of white, vermillion, and earth colors. In light areas, keep the paint as thin as you can; you want the white of the gesso to show through. X-ray scans show that lead white was often used only to emphasize the brightest highlights of flesh tones, with a very thin toning of the gesso used for most light areas of flesh. This approach keeps light areas bright and avoids later yellowing by minimizing the amount of oil. In shadow areas, don’t worry so much about paint thickness, but keep the surface of the paint smooth. In areas that will later be glazed, keep light areas lighter than the intended final effect, since glazes will darken what they cover. When you have taken this stage as far as you can, let the paint dry once again.
  4. If you began the painting in egg tempera or tempera grassa, you will switch to pure oil paint no later than when dead coloring is completed. Use paints ground in linseed oil. When switching from tempera to oil paint, you may choose to apply a very thin layer of oil to prevent excessive absorption by the tempera underlayers. From now on, you will let each layer dry thoroughly before further painting. If the layers are thin, this will take up to several days each time, depending on the pigments you use and whether the paint contains siccatives. It may be helpful to wet sand in between layers as needed to maintain a surface that is thin and smooth.
  5. You will now work each area of the painting toward the intended final finish. Where desired, you can either thin the paint slightly with medium (oil mixed with a varnish or balsam) or put a very thin layer of medium onto the surface and paint onto that. Apply fine detail to light areas. Areas of dark or bright colors, especially those around the main areas of interest, can be glazed. So, for example, a red robe, after a primuersel of black mixed with red ochre and then initial opaque modeling with vermillion and white, could then be glazed with a transparent red lake, using fingers, a rag, or a soft brush to make the glaze very thin over the lights and thicker over the darks. A little ultramarine or indigo could be glazed in to neutralize and reinforce the darker shadows. Multiple glazes can be applied (allow the paint to dry and wet sand in between layers) until a clear sense of three-dimensional depth has been achieved, generally with inky darks, intense midtones, and bright lights. This process can be very time consuming, especially if you are not using siccatives, but it produces a striking effect that is immediately noticeable when comparing Flemish oil paintings to other work of the period.
  6. The final stages would be used primarily for detail work and to apply thin scumbles of opaque color mixed with white where needed over lighter areas of glazed color. Again, sand between coats. It is possible that some fine details were applied with egg tempera, worked into wet oil paint. This development of detail would continue until a very high degree of finish was obtained.
You might also be interested in these posts

Tags: , , ,

  1. Peter Berndt’s avatar

    Very nice account of the Flemish painting process. Very Interesting!
    Peter Berndt

    Reply

  2. Dorothea von Eckhardt’s avatar

    Interesting, Informative lesson in flemish painting. I am an oil painter that may combine the two mediums.

    Reply

  3. Kate Wilson’s avatar

    Fascinating! Thanks for sharing!

    Reply

  4. TIm Jee’s avatar

    According to my book on 15th Flemish painting in the National Gallery, x ray’s reveal that a lot of the paintings didn’t have highly developed underdrawings or underpaintings. some of the drawing under the painting is quite loose and hastily done which only goes to make the process by which the highly finished painting was done, even more baffling. These paintings are often referred to as being “jewel’ like and used to justify the heavy use of glazes to duplicate these paintings. Anybody who has seen these paintings in the flesh cannot fail but to notice the curious OPACITY in the colours which adds to the sense of formal integrity. One can see this opacity in the headress of Robert Campins “woman”. The methodology of the final fine detailing is also baffling. Try duplicating it with just linseed oil (as this is often claimed as the sole medium by many current art conservators) and see how far you get.

    Reply

  5. David’s avatar

    Tim,

    Which book is that?

    Reply

  6. Greg’s avatar

    Its worth looking at this chap’s work when talking about the flemish style, a Russian who currently sticks to a very structured Flemish process, interesting seeing the comparisons to your write up.

    http://www.antonovart.com/

    A free workshop;

    http://artpapa.net/content/view/28/54/

    Reply

  7. David’s avatar

    Greg,

    I just can’t get past the whole seven layers, each of which takes seven weeks to dry thing. Artists are not numerologists; nor were they in the early 15th century. They were professionals attempting to deliver a marketable product in a manner that allowed them to earn a living. There was no room for such silliness in their business model.

    Antonov is not doing anything like what modern art materials scientists understand early Flemish painting to be. He appears to be doing an approximation of what one school of 19th century academics thought early Flemish painting was. They were taking their best guess based on very limited information, but they were wrong.

    That’s not to say there’s anything wrong with his painting method, if you like being really anal about the way you paint. The work on his website seems kind of nice. But it’s not early Flemish—it’s modern Russian.

    Reply

  8. Greg’s avatar

    The mainstream would probably label any time consumed in the pursuit of longevity being anal in this age of cookie cutter abstracts so I suppose I was assuming that the more anal you get the longer the painting lasts. I am personally trying to make a balance between Flemish conservation and 19th century speed of execution. Being interested in current international and political events, the seven, seven ideal just isn’t cricket. On the subject of conservation, does an extensive amount of layers have a strong effect on longevity either way beyond the adherence to fat and lean? Can’t seem to stumble across anything on it.

    Reply

  9. David’s avatar

    Greg,

    Direct painting, in one layer, creates a simpler and therefore more stable paint film than painting in many layers. The more layers added, the greater the risk of one layer failing to adhere reliably to the layer beneath it. So fewer layers, at least in theory, is likely to produce longer-lasting paintings.

    Here’s a post from this weblog on making paintings last a long time.

    Reply

  10. Tim Jee’s avatar

    @David -
    Hello David,
    Forgot I’d posted the comment so sorry for the delay in replaying….(that’s if you get to read this!!!)….the book is “The Fifteenth Century Netherlandish paintings” by Lorne Campbell. It’s published by National Gallery catalogues (London). Its a very good book but expensive. I haven’t got the time at the moment to go into detail, but the 7 layer method is nonsense…I mean historically. Compare the tightly painted feeble paintings produced by modern practioners of this method with the masters of the flemish tratiditon and you’ll see a world of difference. There are a lot of myths like this about. Rather like the theory that Egg Tempera paintings were all painted over highly completed ink or monochrome paintings. Once again, modern x ray analysis shows that this ( in most cases)simply isnt true. Sorry if this sounds opinionated but there are a LOT of misconceptions about the most likely methodology of renaissance painters.

    Reply

    1. David’s avatar

      @Tim Jee -
      Thanks, Tim. That book is on my “want to buy” list. In terms of historical painting methods, the “7 layer” approach is certainly hogwash.

      And you are of course right that the assertion (originating with Daniel V. Thompson) that Italian tempera technique relied on a detailed underdrawing is completely inaccurate. (Though in the 1930’s it was a reasonable hypothesis.)

      Reply

  11. Tim Jee’s avatar

    @David - I think I ought to qualify what I wrote in the above. Highly finished ink underpaintigs may have been practiced
    by some artists…well at least by Cenini!..and I think the Daniel Thompson book on tempera painting is a marvelous little book …one of my favourites. The problem is however that people latch on to these methods as if they are the one and only “correct way” to paint, or that one has to use this method to replicate the look of paintigs from certain periods. Certainly, I’ve seen plenty of tempera paintings from C15th that don’t appear to have any ink or dark underpainting. However , as a method in its own right, why not try it. I’m currently experimenting with different coloured underpaintings in my own tempera paintings. The other great book on Tempera painting is the Robert Vickery “new tecnniques in egg termpera ” (no longer in print but easily available). I imagine his more improvised approach to the medium is closer to Andrew Wyeths.
    Flemish paintings have fascinated me for a while because of their particular paint quality. Eastlake seems fairly convinced that they used hard resins and certainly they have a look similar to some of Holman Hunt’s paintings who definitely used Copal resin. ( Oh there’s another myth biting the dust…Copal being banished these days because it turns the painting yellow and causes cracking…Hunt’s paintings are over 150 year’s old and are in beautiful condition)…
    The book on Netherlanish paintings that I have is very detailed from the conservators point of view. I wouldn’t get it unless you’re a completist..it’s too expensive ..it’s really for those of a nerdish dispositon who enjoy the minutae of investigations into paintings, (that’ll be me!). I have the other National gallery books “Italian paintings before 1400” and ” From Giotto to Durer” which are paperback, cheaper, and more general but full of interesting stuff on the current understanding of these paintings. You may have already seen them. I’m currently after “The Panel paintings of Masolino and Masaccio: The art of Technique”.
    And now, back in the real world…an acquaintance of mine picked up a painting in an junk/antique shop that looked like an early Raphael..Not quite as good but beautifully done with superb subtle paint handling around the face and hands. It really looked genuine. However I managed to bring to bear the knowledge I’ve acquired over the years to discern that it was indeed a fake. What were the tell-tell signs? you might I ask. Well. probably the words “Winsor & Newton” printed in large letters on the back!

    Reply

  12. David’s avatar

    @Tim Jee -

    Tim,

    I think we’re both clear on the role of underdrawings and underpaintings in Renaissance Italian tempera paintings. Thompson was a brilliant guy, but he didn’t have the benefit of modern technical analysis. I have not read Vickrey’s book; I should probably pick it up sometime. I do like much of his work quite a lot.

    Mayer did a great job trashing copal. Of course, there are many different kinds of copal, some of which probably do cause the painting to turn yellow, crack, explode, and corrupt children to the dark side of the force.

    I am certainly of the nerdish disposition when it comes to 15th century painting methods. I have the other books, except for Masolino/Masaccio. That one’s on my list.

    I read somewhere that Winsor and Newton were very active back in the Renaissance, or Victorian period, or one of those olden times. So maybe it’s real…

    Reply

  13. jeff’s avatar

    This description of the method is very interesting, David. I have been doing something similar without really knowing the pitfalls. I have found that the underdrawing is really not really that important in terms of its detail with this approach. The gradual layering means that it simply gets lost and refined anyway. The gradual nature of the approach means that it becomes quite easy to get strong realism from something that is originally very crude.

    I do have a question or two though.
    1. With wet sanding: is that just a case of lightly refreshing the surface and giving it tooth or a reasserting of the lights as well? How necessary is it?
    2. You don’t mention it but I have found significant problems with oil absorption into the ground that regardless of layering means that the final painting needs varnishing. Is varnishing important in this method or am I missing something?

    Reply

  14. Playskull’s avatar

    How did they manage to achieve such smooth surfaces which have a car enamel finish? Raw Linseed oil and pigment grinded together tend to produce brush strokes. The addition of stand oil helps because it levels out brush strokes due to its viscosity but stand oil cannot be grinded with. Maybe it had to due with the method of application. Is it even possible to make a gradation in oil without any turpentine?

    Reply

  15. David’s avatar

    @Playskull -

    Playskull,

    Have you ever made your own fresh oil paint? It’s not the same stuff you get from a tube. It is much more manageable with a brush, and you can adjust it’s working properties by adding different amounts of oil, grinding for different amounts of time, etc.

    Stand oil wasn’t available in the 15th and early 16th centuries. Heat-bodied oil, which has properties that are similar, but not identical, was. There were many recipes for preparing oil that certainly had an effect on its characteristics.

    Reply

  16. jeff’s avatar

    I’d have to agree with David. Homemade oil paint is not necessarily subject to brushmarks. As for needing turpentine to get a graduation, that is not at all needed. A drop of oil is all it takes to make the paint blend easily.

    Reply

  17. Playskull’s avatar

    Very true. Its certainly not the same as tube paint. Tube paints have a uniformity across different pigments. Hand grinded paints have less uniformities across different pigments. Ultramarine is stringy, bone black is gritty, lead white is magically buttery.
    I always look at my paintings glare from an angle to see how smooth my paints look in a glare. I always have trouble getting bone black smooth. I tried grinding it twice as long as other pigments. I am currently experimenting with lamp black and its certainly smoother :)
    As for the brush stroke leveling issue. I have had some success with sun thickened oil. It levels paint a bit less than boiled stand oil etc. I spread my paint on acid free paper to soak out extra oil and replace it with some sun thickened or stand oil
    My newest issue is dust. My paintings are mostly under 15 inches so dust is an issue. By the way fellas, never electric shave your face before painting. I learned the hard way. @David -

    Reply

  18. Koren’s avatar

    A few technical questions about the method you outline here. First, when you refer to the primuersel as a mixture of black and an earth, I’m assuming here that you’re describing essentially a monochrome approach (one earth added to black, to modify the hue and/or value) with no white added at this stage, correct? Second, when moving from the primuersel stage to the body color stage, is the painting allowed to dry completely between those two phases? If not, doesn’t the body color become somewhat muddy with the (at least partial) mixing of the earth/black mixture and the local color? Or is the earth/black mixture confined to only the darkest value areas of the painting, so that the body color is used to model the mid- and high-value areas and there is relatively little mixing between the two areas? To me, these are fine but essential distinctions…

    Reply

  19. David’s avatar

    @Koren -

    Koren,

    The initial detail layer was often done in monochrome. This would often be on top of a lightly applied monochrome toning layer to change the white of the gesso panel to a grey. It may be that this toning layer was also used to reduce the absorbency of the gesso to make it easier to apply the next layers of paint.

    No one absolutely knows how the layering worked, but my assumption is that they allowed layers to dry completely before applying more paint.

    Reply

  20. Tim Jee’s avatar

    @jeff - Jeff,

    If you’re painting on a real gesso panel, a light wash of rabbit skin glue before you begin to paint, helps to seal it and reduce the absorption of the oil.

    Reply

  21. David’s avatar

    @Tim Jee -

    Tim,

    I, too, generally use a layer of hide glue to prepare traditional gesso for oil paint. It’s unnecessary for tempera.

    Reply

  22. Koren’s avatar

    Thanks for the response to my previous question. I thought as much, but it was still nice to get an answer. I have a related question that I thought I’d ask - I’ve been reading about glazing methods, and understand that the previous paint layer must be dry before a glaze can be applied. What’s not clear is HOW dry - does it need to be completely dry to the touch (as you’d get after 2-4 weeks) or completely dry period (meaning 6 months or more)? Based on what I’ve read from other artists the answer seems to vary greatly. I guess I thought initially that following fat over lean would ensure that the upper layers wouldn’t crack even if the lower layers were still oxidizing when the upper ones were applied. Any guidance here would be appreciated.

    Thanks,

    KLF Minneapolis

    Reply

  23. David’s avatar

    @Koren -

    Koren,

    If it were necessary to wait 6 months between layers, very few multi-layer paintings would have ever been completed. Oxidizing is not completed for many decades, so there is no practical point at which you can apply more paint after lower layers are done with expanding and contracting. That’s one reason why fat over lean is important.

    Wait until the under-layer is fully dry to the touch. If the underlayer is very smooth and shiny, then do something to improve mechanical adhesion, such as wet sanding or oiling out with a thin layer of medium containing a balsam.

    Reply

  24. Koren’s avatar

    Thanks much for this reply! I assumed this was true, but it’s nice to have confirmation from someone with experience.

    Also, I apologize in advance for the rant I posted on your ‘convenience paints’ page regarding painting materials and health. To anyone interested in the indirect method who would like to find a more health-friendly solvent for the ‘lean’ layers, I just discovered a new W&N Artisan thinner product that’s made from dipropylene glycol monomethyl ether - after a little research relative to both OMS and turps this appears to be lower toxicity . (I’m not affiliated with any paint company, just like to avoid anything toxic whenever possible.)

    Thanks again!

    Reply

  25. David’s avatar

    @Koren -

    Koren,

    Thanks for the info. The “rant” was perfectly appropriate.

    Reply

  26. Rita Thompson’s avatar

    would a modern example of a theme in these paintings be a political hero being in the background of a gathering

    Reply