Reference · Updated August 2026
Lightfastness Ratings Explained
The permanence reference for painters: what the ASTM categories actually mean, how the Blue Wool scale roughly corresponds to them, and a translation table for the eleven different ways paint companies print the same idea on a tube.
By the Austin Gallery editors · August 8, 2026
The quick answer
Buy ASTM I or ASTM II and nothing lower. ASTM I is Excellent, II is Very Good, and both are described by Winsor & Newton as permanent for artists' use. Categories III, IV and V do not conform to the artists' paint specification. The catch is that brands do not all print "ASTM I" on the tube: Winsor & Newton prints AA, A, B or C, Holbein prints asterisks, Royal Talens prints plus signs, and Schmincke prints stars based on a different scale entirely. The table below translates between them.
The translation table
This is the part worth bookmarking. Each row reproduces one brand's own published code set and its own published wording, taken from that brand's site or chart. We have not merged them into a single universal scale, because they are not measuring the same way and a merged column would be a fiction.
| Brand | What is printed | What each code means, in their words | The catch |
|---|---|---|---|
| ASTM (the shared standard) | I, II, III, IV, V | I Excellent, II Very Good, III Fair. Categories are assigned from measured colour difference before and after exposure. | Only I and II conform to the artists' paint specifications. III to V are non-conformant. |
| Winsor & Newton | AA, A, B, C (plus ASTM I to V and N/L) | AA Extremely Permanent, A Permanent, B Moderately Durable, C Fugitive. N/L means Not Listed. | Prints its own permanence letters and an ASTM column side by side. Both ASTM I and II are described as permanent for artists' use. |
| Golden Artist Colors | ASTM I to V, plus N/A | I Excellent, II Very Good. Shown on the label by the sunburst symbol. | N/A marks a rating not obtained under ASTM protocol. Golden's QoR watercolors use EX for ASTM I and GD for ASTM II. |
| Gamblin | ASTM I, II, III, plus NT | I Excellent, II Very Good, III Fair. | Gamblin says every color in its line is rated I with a few exceptions rated II. NT marks a color not tested by ASTM. |
| Daniel Smith | ASTM I to IV, plus NR | I Excellent, II Very Good, IV Fugitive, as printed on the Daniel Smith pigment chart. | NR means not rated by ASTM. Daniel Smith then runs its own Xenon Fadeometer test and assigns a rating. |
| Liquitex | ASTM I to V, plus NR | I Excellent, II Very Good, III Fair, IV Poor, V Fugitive or very poor. | The only brand we found that publishes year bands against each ASTM grade. NR marks the fluorescent and iridescent colors. |
| M. Graham | LF I, LF II, LF III | LF I Excellent, LF II Very Good, LF III Acceptable, footnoted as good in mass tone with tints that may fade. | The chart legend does not cite an ASTM number for lightfastness. Ratings combine historical data, accelerated testing and pigment maker data. |
| Sennelier | ASTM I, II, III plus asterisks, and nr | I three asterisks, II two asterisks, III one asterisk. nr means non répertorié, not listed. | Runs both notations in the same table, under a heading that names ASTM as the basis. |
| Holbein | Three asterisks, two, one | Three Absolutely Permanent, two Permanent, one Moderately durable. | Holbein calls the field a Permanency Rating, not a lightfastness rating. There is no fugitive tier at all. |
| Schmincke | Five stars down to one | Five is the highest, four very good, three good. | An accredited five-star system based on the Blue Wool scale rather than on ASTM. We could not read the one and two star wording, so it is not reproduced here. |
| Royal Talens / Rembrandt | +++, ++, +, ° | +++ at least 100 years, ++ 25 to 100 years, + 10 to 25 years, ° 0 to 10 years, all under museum conditions. | Four symbols, not three. The page names no test standard at all, unlike every other brand here. |
| Old Holland | No graded code set | A blanket claim that all pigments are lightfast to ASTM categories I and II. | Two named exceptions: Madder (crimson) Lake Deep Extra and Chremnitz White. |
Warning worth reading before you use this as a lookup: the same symbol means different things across systems. Two asterisks means Permanent at Holbein, while four stars means very good at Schmincke, and neither maps cleanly onto ASTM II. Never carry a symbol across brands without carrying the brand with it.
Scroll the table sideways to see every column.
What the ASTM categories actually are
ASTM D4303 is the test method. It exposes painted specimens by one of four routes, southern Florida daylight through window glass, Arizona daylight, or one of two xenon-arc devices, and then assigns a category based on the colour difference between the specimen before and after exposure, calculated with the CIE 1976 L*a*b* colour difference equation. The categories run I to V.
The words attached to those numerals come from the paint makers rather than from a single public ASTM sentence. Golden publishes that I is Excellent and II is Very Good, and that fading gets progressively worse for III, IV and V. Gamblin publishes the same three words for I, II and III. Liquitex publishes the full ladder, and is the only brand we found that attaches year bands to each grade.
| Standard | Title | What it does |
|---|---|---|
| ASTM D4303 | Test Methods for Lightfastness of Colorants Used in Artists' Materials | The test method that assigns the categories. Four methods: two natural daylight exposures (Florida, Arizona) and two xenon-arc. |
| ASTM D4302 | Specification for Artists' Oil, Resin-Oil, and Alkyd Paints | Sets composition, performance and labeling requirements, including lightfastness, for oils. |
| ASTM D5098 | Specification for Artists' Acrylic Dispersion Paints | The acrylic equivalent. |
| ASTM D5067 | Specification for Artists' Watercolor Paints | The watercolor equivalent. |
| ASTM D6901 | Specification for Artists' Colored Pencils | The colored pencil equivalent. |
| ASTM D5383 | Practice for Visual Determination of the Lightfastness of Art Materials by Artists and Art Technologists | The test you can run yourself. Uses a card of eight Blue Wool references, with 3, 6 and 7 as controls. |
Scroll the table sideways to see every column.
The important structural fact: only categories I and II conform. Golden's Just Paint states that categories I and II are for colors that conform to ASTM D4302, while III through V are considered non-conformant. The Colored Pencil Society of America applies the same rule under D6901, and adds that for an entire line to be labeled compliant, every pencil in it must rate I or II.
None of this testing is required. Golden says so directly: after describing its whole program, its own article notes that none of this testing is actually required of a manufacturer. A tube with no rating on it has not failed a test. It may simply never have been tested.
about $116 on Amazon
Gamblin publishes that all of its colors carry an ASTM lightfastness rating of I, with a few exceptions rated II. That is about as close to a whole-line guarantee as this category gets.
The Blue Wool scale, and why it is not a conversion
Blue Wool is a different instrument for the same question. Under ISO 105-B02, eight dyed wool references are exposed alongside your sample. They are built to fade in a geometric progression, each successive reference taking twice as long to fade to the same grade as its predecessor. Blue Wool 1 fades quickest; Blue Wool 8 is the most resistant.
| Reference | How sources describe it |
|---|---|
| Blue Wool 1 | Fades fastest of the eight references. |
| Blue Wool 2 | Roughly half as responsive to light as Blue Wool 1. |
| Blue Wool 3, 4 | Derwent describes this band as fair to moderate. |
| Blue Wool 5, 6 | Derwent describes this band as good to very good, and equates Blue Wool 6 with ASTM II. |
| Blue Wool 7, 8 | Derwent describes this band as excellent. Blue Wool 8 is the least responsive to light. |
Scroll the table sideways to see every column.
Say this plainly: ASTM and Blue Wool are two different test methods, and any correspondence between them is approximate, not a conversion. No standards body publishes an official mapping. What exists are manufacturer approximations. Derwent equates Blue Wool 6 with ASTM II and says results above it will not fade for over 100 years in gallery conditions. Faber-Castell maps its three-star rating to Blue Wool 7 and 8. Golden, describing the origin of the ASTM ranges, calls its own Blue Wool equivalents rough. And a 2023 study in Studies in Conservation by researchers at the National Museum of Australia and the British Museum found measurable batch-to-batch and manufacturer variation in the fading rates of the first three Blue Wools themselves, most significantly Blue Wool 2. The references are a useful comparator, not a ruler.
There is one place the two systems formally meet. ASTM D5383, the practice for artists running their own lightfastness test, uses a card of eight Blue Wool references, with numbers 3, 6 and 7 as the controls that tell you when to pull the samples and rate them.
Which pigments are still fugitive, and what to buy instead
Ratings vary by brand for the same pigment, so this table names who published each rating rather than presenting any of them as universal. Where sources disagree, we have said so in the cell.
| Pigment | Published ratings, and who published them | What to buy instead |
|---|---|---|
| Alizarin Crimson (PR83) | Gamblin ASTM III Fair. Daniel Smith IV Fugitive. M. Graham LF III. Sennelier III. Winsor & Newton watercolor: Moderately Durable. | Gamblin's own Alizarin Permanent, which the company says carries an Excellent rating. Winsor & Newton's Permanent Alizarin Crimson in oil is PR177, rated Permanent and ASTM I. Also the quinacridones: PV19, PR206, PR202, PR179. |
| Rose Madder Genuine (NR9) | Winsor & Newton, the only maker, rates it Moderately Durable in watercolor but Permanent and ASTM II in oil. Independent testing at handprint.com calls it too fugitive for professional work. | PV19. Winsor & Newton sells Permanent Carmine on PV19 as the lower cost replacement for genuine carmine, rated Permanent. |
| Carmine / cochineal (NR4) | ASTM watercolor very poor, category V, as reported by handprint.com, whose own testing agrees. | PV19 again. Watch out for cheap Carmine Hue lines, which may use a different and weaker pigment. |
| Opera Pink / Opera Rose | Winsor & Newton discloses PR122 only and rates it Moderately Durable. Some brands' equivalents disclose a fluorescent dye alongside PR122, and Mijello's own chart warns that its bright colors have lower lightfastness. | Straight PR122 or PV19. There is no lightfast substitute for the fluorescence itself, because that effect comes from a dye. |
| Aureolin (PY40) | Daniel Smith rates it II Very Good. Independent testing at handprint.com puts it at Fair and reports a grey cast after weeks of direct sun. | PY150, PY110 or PY153. Note that most paint sold as Aureolin today is a hue: Winsor & Newton's Aureolin Hue is PBr7 plus PY151, rated Extremely Permanent and ASTM I. |
| Gamboge (NY24) | The Art is Creation pigment database lists it at IV and notes it fades completely. | Winsor & Newton's New Gamboge is PR209 plus PY150, rated Permanent and ASTM II. |
| Van Dyke Brown (NBr8) | Unrated by ASTM. handprint.com's own testing puts it at Fair and advises avoiding it. Sources disagree sharply: Natural Pigments rates its Cassel Earth ASTM I, but indexes it as PBr7. | Golden's Williamsburg line replaced the pigment entirely with PBk9 plus PR101, which it rates lightfastness 1. Winsor & Newton's Vandyke Brown oil is already a substitute mix rated Extremely Permanent and ASTM II. |
| Hansa and arylide yellows | One family, four grades. handprint.com reports PY1 at ASTM very poor, PY3 and PY97 at very good with brand-dependent failure in tints, PY65 at very good, and PY74 unrated by anyone. | PY150, PY110, PY153 or PY65 within the same hue range. |
| Fluorescent and neon colors | Golden states plainly that because they are pigmented dyes, they are not lightfast, and its own testing put nearly all of them past the worst ASTM band. Liquitex labels its fluorescents NR. | Nothing. There is no lightfast fluorescent. Golden's own advice is damage limitation with a UV-filtering varnish, which it says will not make them archival. |
| Prussian Blue (PB27), the special case | Rated ASTM I, and Winsor & Newton rates it Permanent. But it is the classic reversible fader, and Winsor & Newton's own labelling key carries a qualifier for exactly this: fluctuating colour, fades in light, recovers in dark. | Nothing to replace. Just know the behaviour. Conservation literature on cyanotypes documents near-complete recovery in the dark, though that is paper, not a paint film. |
One pigment is deliberately absent from the fugitive list: Indian Yellow. What is sold under that name today is a hue. Winsor & Newton's watercolor version is PO62 plus PY139 rated Permanent and ASTM II, and the oil version is PR101 plus PY139 rated Permanent and ASTM I. The name has a fugitive history; the current paint does not.
Scroll the table sideways to see every column.
about $18 on Amazon
The direct answer to the alizarin problem, from the company that names it. Gamblin states Alizarin Crimson is the only color in its line rated III, which is why it makes Alizarin Permanent with an Excellent rating instead.
The governing rule behind that whole table, from the independent pigment database at handprint.com: the same generic pigment can receive very different paint lightfastness ratings depending on who manufactured the pigment and who manufactured the paint. The same source notes that most manufacturers test none of their paints, and that the number on the tube is often the generic pigment's ASTM rating rather than a test of that specific paint. Conservation guidance adds the medium variable: lightfastness can change depending on the vehicle, so check the rating for a specific manufacturer and medium wherever you can.
about $15 on Amazon
A quinacridone rose, from the pigment family that every source above names as the replacement for fugitive alizarin, madder and carmine reds.
If you are assembling a palette from scratch rather than replacing one color, our oil paint set guide, acrylic paint set guide and watercolor set guide all lean toward the artist-grade lines that publish this information in the first place.
What lightfastness means if you sell your work
This is the part that turns a technical spec into a commercial one. A collector who buys a painting is buying the painting they can see, and the implicit promise is that it stays that way. The Conservation Center for Art and Historic Artifacts states that damage caused by light exposure is both cumulative and irreversible, and that watercolours are more susceptible to deterioration from light than other paint mediums. Nothing you do after the fact undoes a faded tint.
The practical standard the trade uses is simple and worth adopting: Gamblin says to use only colors labeled ASTM I and II, which will ensure your colors do not change over time. Golden goes further on the labeling question, arguing that companies should adhere to the ASTM standard whenever reporting lightfastness, because only ASTM standards are peer reviewed by a broad group of manufacturers, scientists and other experts. When a brand publishes its own private scale with no standard named behind it, that is worth noticing.
Two traps for anyone selling work on paper. First, the word permanent on a marker does not mean lightfast: CCAHA notes it only means the ink is indelible in water, and that even markers labeled archival still have limited lightfastness. Second, the display conditions matter as much as the pigment. CCAHA gives 50 lux for light-sensitive material including works on paper, photographs and textiles, 150 lux for less sensitive material including oil paintings, and 300 lux for the least sensitive, with cumulative exposure calculated as lux multiplied by hours.
To be straight about the limits of this: we could not find any gallery, artist association or conservation body that imposes a duty on an artist to disclose pigment permanence to a buyer. Every obligation we located attaches to the manufacturer's label, and even those are voluntary. Treat lightfastness as professional practice and as something a serious buyer will eventually ask about, not as a regulation. If you are pricing and selling work, our guide to selling art online covers the commercial side.
about $144 on Amazon
The line whose permanence system this page spends the most time translating. Winsor & Newton prints AA, A, B or C alongside an ASTM column, so you can read both systems off one tube.
How the testing works, and where the sources disagree
Golden runs the most publicly documented program: outdoor exposure until samples reach a total solar irradiance of 1260 megajoules per square metre, roughly three months, plus a xenon arc chamber run of 410 hours that it describes as equivalent to roughly 100 years of indoor gallery lighting. The 1260 figure was originally chosen by ASTM because it gave results corresponding to the fading seen in historical paintings going back over a century. Liquitex describes 300 hours in a Q-Sun machine as the equivalent of 100 years in gallery conditions. Schmincke's outdoor test takes 2 to 2.5 years of real time.
Two honest caveats belong next to those numbers. Golden's own scientist writes that accelerated lightfastness results are not meant to be predictive, that reality is too complex and materials too variable, while still concluding that it makes sense to bet on the colorants that did better. And repeatability is hard: daily, seasonal and yearly variation in sunlight intensity, plus geography, cause tremendous variation in outdoor results.
We are not publishing a Delta E threshold table, and here is why. The colour-difference cutoffs that define the ASTM categories appear in at least three mutually incompatible versions across Golden's own properties: one set puts category I at 4 units or less and II at 4 to 8, another puts I at 0 to 2 and II at 2 to 4, and a third discusses a Very Poor band above 24. We could not resolve which is current from the sources available, and the underlying standard is paywalled. A table here would look authoritative and be wrong. If you need the exact thresholds, buy the standard.
Related references
This is one of three technical references we keep for painters. The paper weight conversion chart covers lb to gsm by paper type, and the paint brush size chart publishes measured hair dimensions by size and shape. For surfaces, see standard canvas sizes, and for protecting finished work, our museum glass and UV framing guide deals with the other half of the fading problem.
about $38 on Amazon
Golden prints the lightfastness rating on the label as a sunburst symbol and publishes exactly where each rating came from, including when it did not come from an ASTM test. That disclosure is rarer than the rating itself.
about $38 on Amazon
Daniel Smith publishes a full pigment characteristics chart with a lightfastness column per color, and tests in-house on a Xenon Fadeometer when ASTM has not rated a pigment.
Frequently asked questions
What does lightfastness mean?
Lightfastness is how well a paint resists fading or changing colour on exposure to light. Gamblin's glossary puts it as how permanent a color is, or how unaffected it is by light. The formal version is a measurement: ASTM's test method exposes painted samples to daylight or to a xenon arc lamp, then assigns a category based on the measured colour difference between the sample before and after exposure, calculated with the CIE 1976 L*a*b* colour difference equation.
What is a good lightfastness rating?
ASTM I or II, and nothing lower, if the work is meant to last. Gamblin says to use only oil colors labeled ASTM I and II, which will ensure your colors do not change over time. Winsor & Newton and Natural Pigments both describe I and II as permanent for artists' use. Golden's Just Paint states the technical version: categories I and II conform to the artists' paint specification, while III through V are considered non-conformant.
What is the Blue Wool scale?
A set of eight dyed wool references, standardised under ISO 105-B02, exposed alongside your sample so you can see relative fading. They are designed to fade in a geometric progression, with each successive reference taking twice as long to fade to the same grade as the one before it. Blue Wool 1 fades quickest, Blue Wool 8 slowest. ASTM's own artist self-test practice uses a card of these references with numbers 3, 6 and 7 as controls.
How does the Blue Wool scale convert to ASTM?
It does not convert. They are two different test methods measuring different things, and no standards body publishes an official mapping. What exists are manufacturer approximations: Derwent equates Blue Wool 6 with ASTM II, and Faber-Castell maps its three-star rating to Blue Wool 7 and 8. Golden describes its own equivalents as rough. Treat any Blue Wool to ASTM correspondence as approximate, and never as arithmetic.
Is alizarin crimson still fugitive?
The traditional pigment, PR83, still rates poorly nearly everywhere. Gamblin rates it ASTM III and says it is the only color in its line at that grade. Daniel Smith's chart marks it IV Fugitive. M. Graham lists LF III. Winsor & Newton rates its watercolor Moderately Durable. Golden's own testing found the oil masstone held up reasonably while the tint shifted several times further, and a watercolor sample turned nearly completely white in a test chamber. Buy a permanent alizarin or a quinacridone instead.
Does lightfastness matter if I sell my work?
Yes, in the sense that it decides what a buyer owns in twenty years. Conservation guidance from the Conservation Center for Art and Historic Artifacts notes that light damage is both cumulative and irreversible, and that watercolours are more susceptible to it than other paint mediums. There is no rule requiring you to disclose pigment permanence to a buyer, and we could not find any gallery or professional body that imposes one. But a collector who paid for a painting reasonably expects it to keep looking like the painting they bought, and the whole ASTM apparatus exists because that expectation is old.
Does the same pigment have the same lightfastness in every brand?
No, and this is the single most useful thing to understand about ratings. handprint.com, the most heavily cited independent pigment database, states that the same generic pigment can receive very different paint lightfastness ratings depending on who manufactured the pigment and who manufactured the paint, and that the rating on the tube is often the generic pigment's ASTM rating rather than a test of that paint. Conservation guidance adds that lightfastness can change with the vehicle, so check the rating for a specific manufacturer and medium.
How long does an accelerated lightfastness test run?
It depends on the maker, and their equivalences do not agree. Golden exposes samples in a xenon arc chamber for 410 hours, which it describes as roughly 100 years of indoor gallery lighting, and outdoors until a total solar irradiance of 1260 megajoules per square metre, about three months. Liquitex says 300 hours in a Q-Sun machine is the equivalent of 100 years in gallery conditions. Schmincke's outdoor rooftop test runs 2 to 2.5 years in real time. Golden's own scientist adds the honest caveat that these results are not meant to be predictive.
Can I test lightfastness myself?
Yes, and ASTM publishes a practice for exactly that: D5383, for artists and art technologists, which exposes your swatches alongside a card of eight Blue Wool references. Conservation guidance from CCAHA recommends it, noting that running your own test is a good way to verify manufacturer information. Be aware that results are hard to repeat, because sunlight varies daily, seasonally and by location.
Sources
Ratings are attributed to the organisation that published them. Where two sources give different figures for the same pigment, both appear above.
- ASTM International: D4303 test methods for lightfastness, D4302 artists' oil, resin-oil and alkyd paints, D5098 acrylic dispersion paints, D5067 watercolor paints, D6901 colored pencils, and D5383 visual determination by artists
- Winsor & Newton, Terms Explained (the AA/A/B/C permanence system, the N/L code and the six qualifier notes), plus the composition and permanence tables for Artists' Oil Colour and Professional Water Colour
- Golden Artist Colors, Technical Specifications Explained and the technical chart; Golden's Just Paint on lightfastness testing at Golden, ASTM testing for oil paints, alizarin crimson, fluorescent colors and QoR labeling
- Golden, on fluorescent colors and Williamsburg's Van Dyke Brown reformulation
- Gamblin, Building Permanent Paintings and the Gamblin painting glossary
- Daniel Smith, how it tests for lightfastness and the Daniel Smith watercolor pigment characteristics chart
- Liquitex, What Is Lightfastness and its FAQ on the NR rating
- M. Graham oil color chart, Sennelier watercolour nuancier, Holbein watercolor chart, Schmincke on lightfastness, Royal Talens on lightfastness, and Old Holland Classic Oil Colours brochure
- Blue Wool: Q-Lab on ISO 105-B02, James Heal on the Blue Wool standards, National Library of Australia illumination policy, Derwent colour charts, and Faber-Castell lightfast ratings
- Ford, B. and Korenberg, C. (2023), "Manufacturing Variations in ISO Blue Wool Fading Standards under Microfading Exposure Conditions", Studies in Conservation, DOI 10.1080/00393630.2023.2184555 (PDF)
- Conservation guidance: CCAHA, Preventive Conservation for Artists and CCAHA on light exposure in exhibition; Mike Ware, A Blueprint for Conserving Cyanotypes (the Prussian blue reversibility mechanism, on paper rather than in paint film)
- Independent pigment databases: Bruce MacEvoy's handprint.com on rating variability, its red and violet pigment pages, its yellow pages, its blue pages; and Art is Creation
- Natural Pigments on composition and permanence and the Colored Pencil Society of America on ASTM D6901
What we left out. No Delta E threshold table, for the reason given above. No year bands for Daniel Smith grades, because the widely circulated figures do not appear on any Daniel Smith page we could read. No wording for Schmincke's one and two star grades, because that key is published as an image. No star rendering for Royal Talens, because we could only confirm the plus-sign notation. No rating for PR192, because we could not source one. And no claim that a standard requires artists to disclose permanence to buyers, because we could not find one. Product links on this page are affiliate links; prices are what we saw at publication. See our affiliate disclosure.