A discussion with
Catherine Metzger
Conservator specializing in Flemish and Italian Renaissance Paintings
at the
National Gallery, Washington, D.C.

I spoke with Catherine at the National Gallery by phone and we had an extensive and very positive exchange. I did not tape the conversation so as to directly quote Catherine, but did take written notes.
Below is a summary of our talk in my own words and with added commentary.
Therefore the statements below do not necessarily represent the ideas of Ms. Metzger or the National Gallery.

I have had several people email me and challenge the definition of painting "Fat-over-Lean." Therefore, I first asked her to define that phrase: She confirmed that "fat" refers to the fatty drying oil used as a binder and in mediums. Therefore, a fatty layer of paint has more fatty oil (linseed, poppyseed, walnut, etc.) in it than does a lean layer, or one containing less fatty oil. To paint "Fat-over-lean" is therefore defined as painting in layers of paint which contain successively greater quantities of fatty drying oil. (fat is fat - lean is less fat: the same as at the meat market.)

The purpose of this method of layering oil paints is due to the fact that oil paints do not dry by evaporation, but by oxidation, during which time the paint film flexes and moves. By adding more oil in each layer of paint you insure that the lean under-layers will dry before the top layers to prevent cracking. Whereas a lean layer painted over a fatty layer would completely oxidize and become inflexible, causing it to crack when the underlying paint layer flexes during oxidation. This effect can be demonstrated by looking at cement sidewalks: The cement is inflexible, but the earth beneath moves when freezing and thawing occurs. Since the top layer of cement is unable to flex with the earth, it cracks and heaves.

Of course, painting on a flexible support, such as canvas, creates flexibility within the structure of the painting which can, itself, be damaging. Any painting on cloth which received a blow (whether from being bumped from behind or dropped, etc.) will cause a certain type of cracking which is usually visible in a circular pattern on the painting's surface. There is only one way to attempt to avoid this - the same way as one does with cement. When pouring cement, a craftsman will always cut through the layer to create pathways for cracking. We readily see this pattern of cutting in sidewalks where it is meant to allow the heaving earth a place to push without cracking the cement. The sidewalk thus acts as a series of tiles floating on the earth. The most crack-free way of painting is the same: paint in a series of daubs of color separated so as not to create a continuous paint film. the flexing support can then bend between the brushstrokes without cracking them. However, this style of painting is very limiting creatively.

Most artists desire to create a continuous paint film across the support and have, for centuries, done so in varying thicknesses of paint: thin wash imprimatura, covered with transparent shadows and bold, thick highlights, glazed with transparent glazes and scumbled with thin layers of translucent opaques. The key is then in knowing the amount of fatty oil in each color. Each pigment has different properties, including particle size, which determine how much oil is required to turn it into a paste. Craftsmen of past ages, who ground their own paints, knew which pigments were high in oil and which were low. The underpainting would then be produced in lean colors and successive layers used fattier pigments, or else more oil (in the form of a medium) was added to altar the paint film's elasticity. Lead white is a very lean color - it requires very little oil to produce a paste for painting. Catherine also mentioned that each pigment chemically reacts with the oil to effect drying. Lead white reacts in a way that speeds drying times by increasing polymerization. It should, therefore, be used in underpainting or for painting over washes of color, but is more likely to crack if layered over a very fatty color such as a thick layer of umber or carbon black. These colors do not have sufficient metals to dry the oil and may not dry completely for 50 to 100 years. These darker areas of older paintings suffer the most damage when cleaning, she said.

As a general rule earth colors, especially those containing heavy metals, dry faster than do organic or synthetic pigments which have finer particle sizes.

Because of the yellowing of linseed oil, many artists and paint manufacturers throughout the ages have sought to use other binders or vehicles for grinding pigments into paste, especially lighter colors and cooler colors which may be effected by yellowing. Rubens and others of his day ground whites and blues in walnut oil because it yellows less.

Today, many paint manufacturers, creating paints for the direct painting methods of the impressionists and plein air painters, use poppyseed, sunflower, or safflower oils. However, none of these are suited to layering paint. Catherine discourages the use of poppyseed oil because it is brittle and cracks more easily. Safflower is also not a good substitute because it never dries and is thus not truly a "drying oil." Most paint manufacturers using these oils add dryers to the mixture to force consistent drying. But there is little that they can do to prevent the inherent brittleness of these paint films. The very chemical within linseed oil that makes it yellow is also the thing that makes it flexible.

We went on to discuss the problems of cracking and how to avoid these pitfalls. However, the first point of mutual agreement was that ALL oil paintings crack: there is nothing that can be done to stop it.

Good craftsmanship includes :

1.) reducing the amount of cracking,

2.) reducing the speed at which it occurs, and

3.) controlling the type of cracking.


We discussed the writings of Eastlake and others of his era who translated centuries-old manuscripts and determined that pigments were ground in vehicles consisting of a combination of sun dried or boiled drying oil immixed with resins (amber, sandarac, mastic, etc.) and drying agents. Catherine noted that Eastlake and his contemporaries have been found to have inaccurately translated old Latin texts which are full of abbreviations, causing them to be subjectively interpreted.

The National Gallery, London, and other museums have engaged in gas chromatography to determine the chemical makeup of master works in their collections. They are finding that paintings throughout Europe have a vehicle (binder) consisting of partially polymerized fatty drying oil (linseed or walnut oil) produced by sun drying (similar to stand oil or sun thickened oil of today). Polymerization of drying oil gives it properties similar to resins. But no actual resins have been found in the paint layers investigated. This would imply that either the recorded recipes of the past were not actually used in painting, or else paintings with resin in them have not survived.

Despite reports from his day that Rubens used resin in his paint to avoid the need for varnish on top and to speed drying, no resins have been found in his paintings. They contain drying oil only: linseed in most colors, with walnut oil used in his whites.


When painting, paints should not be excessively thinned with turpentine or mineral spirits which evaporate. excessive thinning overextends the binder. When the diluent evaporates the particles of oil wrapped pigment are spread out in a disconnected manner that prevents them from forming a solid paint layer and flaking occurs.

Thinned paints CAN be used in washes for sketching in the preliminary paint layer. In this use, the surface remains porous, allowing for good adhesion of successive paint layers which will completely cover the wash, holding it in place. Otherwise, paints should be thinned with a medium.

Following the Fat-Over-Lean principle, each layer of paint should be mixed with a different medium, each one containing more fatty oil than the last. Next, the medium should be added to the paint in a controlled manner. I set up my palette of colors and then add one to two drops of medium per 1/4 to 1/2 inch of each color and mix it in with a palette knife. Any further thinning of the paint that needs to occur to make it more easy to handle or to increase flow and reduce viscosity should be done by the addition of a diluent, not more medium!

One of the worst mistakes that artists make is to set up a cup of oily medium on their palette or stand and dip into it indiscriminately with their brush. If one medium uses 2 parts oil and another uses 3 parts oil, but you apply twice as much of the medium with 2 parts oil, then you have just added 4 parts of oil to your paint, which will be fatter than the layer with 3 parts that follows it. You will not, then, be painting fat-over-lean. You must apply the same amount of medium to your paints for each color and for each layer.

On this point Catherine was overjoyed to hear my methods. She said that if I could do nothing else in my career, if I could at least make artists understand the differences between a medium and a diluent, she would be eternally grateful.

So here is it:

A medium is used to alter the properties of a paint film.
A diluent is used to dilute (thin) paint.

A medium adds a new and permanent element to the paint film.
A diluent evaporates and leaves no remaining element.


A medium must be employed in measured amounts.


A diluent may be added in varying amounts, and more may be added as it evaporates in order to keep the paint fluid.


The safest medium is the most simple - a drying oil.
But since artists often want to effect the paint's properties and handling, other ingredients are often added.

Resins have been used to speed drying and reduce the viscosity of paints, producing greater flow while maintaining a continuing paint film. Because hard resins are too brittle and tend to yellow greatly, the safest and most popular resin in use today is Damar. Damar varnish does yellow and crack, as all varnishes do, but less so than others.

Damar varnish will molecularly crosslink with linseed oil which allows it to be used in mediums with linseed. But its use as a final varnish can cause problems for this same reason. As it ages and becomes more brittle, it also engages in this crosslinking with underlying paint layers, making it harder to remove and endangering the underlying paint layers during harsh cleaning.

Because damar is resoluble with turpentine, excessive amounts should not be used in the paint film. Experts agree that ratios of damar less than 50 percent, and preferably less than 25 percent, are safe. When mixed with the paint and drying oil in this percentage, the dry paint film structure is not resoluble.

My medium contains damar and stand oil mixed in a ratio of 1/1 and thinned with turpentine up to 7 parts. When a couple of drops are mixed with tube oil colors containing a linseed binder, the ratio of damar in the paint film is well within safe limits.

The resolubility issue is of greatest concern when using damar in a medium. But when used in a controlled manner, it is safe.

The issue of cracking when using damar in a medium was of no concern to Catherine.

Since damar can only be dissolved using turpentine (not mineral spirits) some may wish to eliminate damar from their studios for safety reasons. If so, Catherine advises using linseed oil only, but in a less polymerized form than stand oil if one wishes a more fluid paint. The trade off would be a slower drying paint than with stand oil and damar.


Catherine was unfamiliar with claims that the old masters added minnium (red lead) to boiled oil in order to speed drying, but thought that it should work well, since lead is a natural dryer. The painters of old seem to have ceased their use of it, however, because it caused excessive drying and eventual cracking. She says that Rembrandt added smalt (cobalt glass ground into powder) to his oils to speed drying.

Bouguereau was said to add cobalt dryers and other additives to his paint.

I asked her about the use of alkyds to speed drying. Despite a relatively brief time in use, she felt (as do others) that they are perfectly safe to use mixed with traditional oils and all oil mediums, being a modified oil themselves. Alkyds dry in as little as 6 to 8 hours and remain as flexible or more flexible than traditional oils.

Among white pigments, lead is the most flexible and best for use in layered techniques. However, it is also the most translucent white, increasingly so as it ages. Titanium white is most opaque, but is also more brittle. Zinc is most brittle and translucent.

The toxicity of lead cause many to avoid it, though it is safe when handled with care. The solution for layered painting may be in the use of an alkyd binder and titanium white. This combination gives opacity, flexibility, and quick dry times. For these reasons, and the lower expense relative to lead white, I have been using Winsor & Newton's Griffin Alkyd Titanium White, with traditional oil paints for the rest of my palette.

For expensive commissions, I still use the lead white. After investigating concerns of incompatibility between lead and sulfurous pigments (cadmiums and ultramarine), I concluded that the mixtures are safe in an oil binder, the oil preventing chemical reaction from occurring. This has been the opinion of other experts that I have consulted, and Catherine concurred that no compatibility issues should arise. The National Gallery hand grinds their own pigments and they use lead white with ultramarine (as did the masters) and cadmiums without ill effect.


All grounds cause problems for permanence:
Cloth flexes with changes in weather and can rip.
Wood panels can warp and crack.
And metal can bend and dent.

Nonetheless, here are some hints.

Copper is considered one of the most permanent grounds upon which to paint in oils. However, if dropped or bent, it is very hard to straighten out without greatly damaging the painting. Its weight and difficulty in transportation make large images impractical.

Linen is stronger than cotton canvas because of a longer fiber structure. But canvas can hold up well if it is made of a heavier weight (12 oz. or above). Sailcloth is good to use if you can find it untreated. [most sailcloth is water repellent for use on boat sails.]

Wood panels can last if they are properly engineered to avoid warping and cracking as the wood ages. A good modern alternative is pressed masonite panel. The newer panels no longer contain oil that was a problem in the past, and they can hold up well if not exposed to excessive moisture.

Synthetic panels of solid material may be good because they do not flex with humidity and temperature changes. Polyesters are especially good to use and will last for 100s of years.


As mentioned earlier, damar varnish can be difficult to remove from a painting without damage when aged for a long period. This becomes especially true when large amounts of damar have been used in the painting itself.

For safety sake, if one wishes to use damar in one's medium, then a synthetic varnish is advisable. Catherine highly recommended Gamblin's "GamVar" varnish. Other synthetics may still crosslink with the paint film underneath. But GamVar is a hydrocarbon based varnish that does not crosslink and removes easily with mineral spirits (which do not dissolve damar). Also, it does not yellow or crack.

Another benefit is that GamVar is flexible enough to be applied as soon as the painting is dry to the touch, instead of waiting 6 to 12 months. This means that artists can apply a full final varnish before rushing paintings out the door to galleries. Catherine says that no varnish seals out air from a painting, and that immediate varnishing will not hinder oxidation of the paint layers beneath.

GamVar is now available premixed in 5.4 and 2.0 oz. jars.

The gloss of GamVar is meant to be comparable to damar, but some say that it is shinier. To reduce gloss, add solvent (mineral spirits) or wax medium.

No resin protects against oxidation or atmospheric moisture. Its sole purposes are to protect against dirt and atmospheric pollution, and to even out the sheen of a painting, adding depth to color.

After our discussion I ordered a batch of GamVar to prepare a group of paintings for a show. The varnish flowed evenly over the paintings when applied and coverage was exceptional. I mixed the varnish 2 parts to one part Turpenoid and found the gloss a light sparkle and not at all excessive. I was able to apply it as soon as paintings were dry to the touch. And best of all, it dried over night, whereas damar varnish has often taken the better part of a week.


Return to Top

History, Definitions, and Techniques | Drying Oils and Mediums
Resins and Varnishes | Pigments Past | Pigments Present | Pigment Chemistry
Supports for Painting | Grounds on Canvas | Techniques of Past Masters
Discussion with National Gallery Conservator