Reinventing the Wheel: Why Red is not a primary color


Ever had a painting turn to mud when you try to mix colors? The good news is it is not you. Conventional color theory using red, yellow and blue does not work. It will turn your pictures to mud and frustrate you as you struggle to mix colors that are impossible with this color triad. The same goes for the split primary system where you mix with a warm and cool version of each primary. The split primary system works much better than just using the three primaries but it misses the point about how colors really mix and what the true primary colors are. What are the real primary colors? Just ask your printer, Cyan, Yellow, and Magenta. Before you write me off as a heretic, read on, try a few simple color mixing experiments at home with your colored pencils watercolors, opaque paints and see for yourself. By changing the way you think about primary colors, your ability to mix colors will improve overnight. 


Red is not a Primary Color

the color red-mised from yellow and magentaI often see red and blue included in paint sets and on color wheels as a primary color. A bright fire engine red is usually shown as the red and some form of navy blue such as ultramarine stands in for the blue. Neither of these colors are primaries. A few quick tests will prove it. First, let’s mix red from other colors. You will need a clear magenta and a bright yellow. As you start to add magenta to yellow you will see the mixture turn orange, then red. If you can mix red, it is not a primary color for pigment.


There are also colors that you can not mix using red. Let’s start with magenta itself. You can not mix the yellow out of red to create magenta. The same is true of hot pink. If you dilute red, either with water or white you do not get pink. You get light red. Hot pink comes from diluted magenta. You can not mix vivid purple by combining red and blue. The combination is a bruise colored dull purple. Vivid purple is the result of combining magenta and cyan (or blue). You need to have a good magenta in your palette. My favorite is Daniel Smith Quinacridone Pink. If you are using prismacolor pencils, the magenta color is called process red (just to add to the confusion).


Blue is Not a Primary Color Either

Blue mixed from magenta and cyanBlue is confusing because many hues get called blue. In this case we are talking about navy blue such as ultramarine. This color can be mixed by combining cyan paint with a little bit of magenta. Greens mixed from yellow and ultramarine make a dull unsaturated green. If you want to mix vivid greens, mix yellow and cyan. Cyan is the color that you need to get comfortable with. If you are using watercolor, Phthalocyanine Blue (GS), also called monastral blue, and phthalo blue is a good cyan. If you are using prismacolor pencils, look for true blue.



The Real Primary Colors

Color wheel mixed from Cyan, Magenta and Yellow.What then are the primary colors? Your color printer knows the answer: cyan, yellow and magenta. These colors mix a bright and clean spectrum. You can mix red, green and blue from these primaries. Orange and violet still  occupy their familiar locations in between red and yellow and blue and magenta respectively. Interestingly enough, the color wheel for light is the is the same as pigment only the primary and secondary colors are reversed. You mix red, blue and green light to create all the other colors.


Why not just call magenta a shade of red and cyan a shade of blue?

Those names are already used as the primary colors for light..The primaries for pigments must be distinct to differentiate them from red and blue. Second. we have a clear idea of a hue in our head when when say “red”. Color names become less useful when they are used to describe a larger arc of color across the color wheel. The word blue already has this problem. It is used on the tubes of so many colors of paint that it has lost its spesificity.To preserve the meaning of blue, use it to describe the primary color of light or the consequence of mixing cyan and magenta pigments. It is easier for us all to envision the same color when we say cyan. Just as it would be confusing to describe red as an orangeish violet, describing cyan as a greenish shade of blue makes no sense (both green and blue are secondary colors). Call cyan cyan. Name the color that you mean. Stop describing magenta as an off color of red. Instead call magenta magenta. It is red that is a yellowish magenta.

The problem with the Split Primary Palette

Many painters use a split primary palette that has a warm and cool version of each primary color. Here is one example of a split primary palette. Here is an example cool yellow: Hansa yellow light, warm yellow: cadmium yellow or new gamboge, warm red: cadmium scarlet or pyrrole red, cool red: quinacridone carmine or quinacridone rose, warm blue : ultramarine blue, and cool blue: phthalocyanine blue GS.  I think that this just confuses the issue. You can mix  a “warm blue” by combining the cool blue and a little cool red. You can mix a “warm red” by combining the cool red and a little cool yellow. And you can mix a “warm yellow” by combining the cool yellow and a little cool red. So why not just say we have three primaries here that we can use to mix the other colors? There is a very good analysis of the split primary palette on

There is nothing wrong with having more colors on your palette.

Even though I can mix red, I still have red on my palette. Just because you can mix more colors from the three primaries does not mean that you need to limit your palette to these colors only. Mixing all your colors from a handful of primaries does not make you more of an artist than someone who uses a larger range of pigments.Get to know the properties of a variety of pigments. Sometimes you will need a staining color. Other times you may want a color that you can wet and lift out more easily. Pigments also vary in lightfastness, granulation, transparency, and intensity. It is great to have a selection of colors to fit the job. Also if you are working in the field it may be very convenient to have more pigments on your palette for faster color matching.

In defense of mud.

Finally, there is nothing wrong with “mud”. Many of the colors we see around us are desaturated. If you mix all three primary colors together you will get a muddy or desaturated mixture. This is good and you will find you want to do this all the time. However, when you are trying to keep colors brightly saturated, mix using adjacent primary colors.

Would you like to learn more? See these sites:


64 thoughts on “Reinventing the Wheel: Why Red is not a primary color

  1. Tim says:

    I know this is an old thread but cyan as used by printers is nothing like Phthalo blue. Just open up your old printer ink cartridge and look. Magenta and yellow are easily matched by Quinacridone magenta and yellow light but cyan? It’s a “light blue”. I agree entirely with your article and I know this works in real life. Is this a case of the pigments in Phthalo working, more than it being “cyan”?

  2. Julia D'Ambrosi says:

    For years people have complimented my sense of colors. They’d say “your colors are so vivid”,”your paintings are so bright.” for years I’ve said my favorite color to paint with is quincadrone magenta. When I would have color wheel assignments however my results were awful and I couldn’t figure out why. (The teachers had us using red.) Now it all makes sense! I’m shocked I didn’t figure it out. Thank you so much I’m in graduate school to be an art teacher and I’m going to teach this! I’m embarrassed to only just learn this but so glad I have. Luckily I’ve always painted with magenta just not for the color wheel assignments .

  3. Amy says:

    I love this. A few questions:

    1) I tried to print your watercolor CMY color wheel and it came out with all screwed up colors (the cyan looks blue-blue and the magenta looks red). I realize that colors on the screen will not necessarily translate to the printer, I’m just annoyed that I can’t print it just as it looks on my screen. Do you know how/where I could print it to be more true-to-color?

    2) Obviously I can recreate the CMY color wheel using my own paints. And particularly for my students I will be doing so. However, where can I purchase CMY paints – both acrylic and watercolor? I hope to find some to use in my classroom to practice color mixing and I want the colors to be as close to “true” as possible.

    3) I am considering also doing some transparency work with this. I’d love to find transparencies OR colored acrylic that I could use to mix colors with. Any idea where I can find some materials like this?

    Thank you in advance for your help.

    • syncing your screen with your printer is difficult. Every screen makes colors look a little differently. On top of that every printer prints its own gamut or color space. Some printers do better jobs at printing some colors. The type of paper you use also changes the colors you will see. It is not something you are doing wrong.

      for colors: what brand of acrylic or watercolor do you use?

      For transparency: look in a stage lighting supply house or photography store. ask if they have colored “gells (gels,jells jels?)” for lighting. These are tranparent colored sheets to go over lights in all colors.

  4. John, your comments about the paint-mixing primaries and the split primary idea are of course completely correct. I have some demonstrations explaining why the ideal subtractive primaries are CMY on my site here:
    and last year added a discussion of modern and “traditional” (RYB) colour theory and the reasons for the current prevalence of the latter here:
    There is also an outline of the history of different concepts of “primary colours” here:

    My only suggestion for your article is that since modern colour theory has three different kinds of “primary colours” (additive, subtractive and psychological), it only invites pointless argument to single out any of these three as “the” “real” primary colours. Red IS one of the four psychological primaries or unique hues:

  5. Martin says:

    Good thing i have a color vision deficiency and don’t need to care about that (I don’t see any difference there).

    But it is an interesting article.

  6. alex says:

    but in fact red is only color that can NOT be produced by mixing any single wavelength lights
    violet+some cyan = blue(evenmore when i switch from 430nm violet to 460nm blue, blue appear cyan-green from start and after some miliseconds appear blue).

    • I love what you are doing. When you see the color wheels side by side it tells the story very well. Two other things I would suggest. mix the magenta and the yellow so students can see that they can mix red (as well as orange) and also see what happens when you add a little bit of the remaining primary to a mixture to tone it down a little. Thus if you take the green made by cyan and yellow and add a little magenta you can make the whole selection of natural greens, to olive, to brown (when you add even more magenta).

  7. David Smothers says:

    It depends upon the physical characteristics of your pigments. I have noticed that watercolors are the least forgiving in mixing colors, for whatever reason. I paint with oil, and have no trouble getting clear tones. It just depends on what you paint, and how intense or subdued you want your colors to be. Since I paint landscapes, it’s always a more subdued hue that I’m looking for, those pure colors are quite rare in nature. So interesting essay, but I think it’s basically splitting hair, as cyan is blue, and the magenta you have pictured is basically red. There has been a shift in recent years, Magenta used to be considered a more violet tint than what is considered magenta today.

    • Hi David, Thank you for your thoughts. I agree that I lean toward more muted hues in my landscape drawing. For botanicals however you often need those more saturated colors. For splitting these hairs, I find that the CYM system helps students (and professional illustrators) more accurately match colors. Cyan is not blue (blue is cyan+magenta) and Magenta (not the purple version) is not red (red is magenta+yellow). I think the distinction is important for the reasons described in this post.

      • Coming from a professional preprinting background, I agree that the distinctions between RYB and CYM are both important and very helpful for watercolorists – and those who work in other media – to understand. The optical and chemical properties of pigments vary from brand to brand and across color names. And honestly, excellent paints are expensive, and deservedly so. Many landscape artists prefer working with a limited palette to save space and weight in their kit, so it can only help us to identify those pigments that offer the purest colors with the widest mixing abilities.

  8. Cathy Austin says:

    Wonderful clarification of why I kept getting mud!

    I also had to laugh a little because the sounded like someone using Socrates’ Socratic method of completely confusing people with convincing others with his circular argument, that what they believed to be wrong was right. I’m so glad to have learned about the changes.

  9. Alexis Hamilton says:

    late to the party but so glad you have revealed one of the great lies of childhood for the falsehood that it is… As an elementary art teacher I’ve been frustrated with the effects the kids get with color mixing. From now on magenta will be on their palettes.

  10. Sannesthesia says:

    In Art School we also used the CMY system; though we also had to buy a seperate tube of orange as the combining of magenta and yellow couldn’t produce a fresh enough orange.

    I think the more primary colours used the better and more natural tones.
    Orange, Teal and Violet can still look off with only CMY.

    A Colourwheel with Near UV and Near IR incorporated would also look different. There are people with Aphakia that can see into UV and people who can see a bit farther towards near IR.
    There are also people who are tetrachromats and it would do justice to them as to have way more primary colours used in painting as well in tv.

  11. Riley E says:

    Actually, the three primary colors are in fact red, yellow, and blue. You are thinking of the three primary PIGMENTS which are different. The primary pigments are indeed magenta, yellow, and cyan…

    • I disagree. There are primary colors of light (RGB) and primary colors of pigments (CYM). I do not know of a distinction being made for RYB color separate from pigments. Let me know if you have a source so I can read more. Thank you.

      • Peter Cockburn says:

        Absolutely! We live in the subtractive world of colour. We remove elements of colour from white light (RGB), to create our painted and printed representation, using the complementaries CMY+black, known in the printing world as CMYK

    • Peter Cockburn says:

      Sorry, you have this incorrect. The 3 additive primaries of white light are red, blue and green..yellow is created by red and green light ( just look at a rainbow sequence). They are additive because they start from blackness and add together to make white light. The 3 subtractive primaries are magenta, cyan and yellow (the direct complementary colours), which together take colour away to make black again. Although it is not a true black, which is why the printing process uses black as well for contrast.

  12. Bob Hewitt says:

    I teach color theory for Photography. I’ve always taught both systems. RGB as primaries for light, (combined the produce white light) and CMYK for offset printing K is black. RGB was theorized by Thomas Young in the 1800’s. there is another system for light which never caught on which includes Yellow as a primary. Then again if I were printing wet color I would only use Y and M to avoid neutral density.

  13. The Crazy Art Teacher says:

    Very instructive essay. For those teachers and learners who struggle with color, there actually isn’t a “unified theory” of color. That is pure myth. It was an attempt from the 1930s to make colors “scientifically categorized” and it failed.

    There are two natural things going on with mixing colors: transparent light (sunlight) and pigments. Sunlight is pure color, but it has to be projected as light. It cannot be reproduced on paper or canvas. Pigments do that. Pigments are not “primary, secondary” etc.

    Red is not a “color,” it is a hue. There are familys of hues, such as rust red or orange; both of these are “reds.” Yellow is a hue; blue is a hue; green is a hue. These cluster in families.

    With *paint* —which is pigments made with natural materials, such as minerals or dyes— you simply have to learn by experience which pigments mix with other pigments to create a new hue. Color has *no* formulas, and there are no recipes; it is not like a cookbook.

    A red manufactured by Winsor & Newton will mix in one way, one set of properties; a red manufactured by Grumbacher or by Liquitex will mix with completely other properties.

    You also have to learn by experience how much of each pigment (“color” or “ingredient”) of hue yields a certain desired new hue. There is “grass green” and “sea green” and “emerald green” but there is no “green” that is universal. The same thing with “red” or “yellow” or “blue.”

    Primary magenta PR122, for example, only works if you are mixing together printing inks. Printing inks are transparent. If you put magenta PR122 on yellow paper, it will look orangish. If you put PR122 on white paper, it will look reddish. If you put PR122 on white paper and then put a varnish on it, it will appear to be richer than if you do not put a varnish on it. This system is purely for industrial categories that permit close but not perfect comparisons.

    Acrylic paints are most predictable. Oil paints will change over time, and depend on many chemistry factors. Watercolor also depends on many chemistry factors, and function just like printing inks.

    Blue and yellow therefore is not green! Color is an art, it is not a science.

    • Thank you Crazy Art Teacher, I agree that you need to learn each pigment and there is no substitute for playing with your colors and having the CYM framework in your head is a great model, though not a recipe. I find a good quinacridone based magenta mixes red (when combined with yellow) with watercolor, acrylic, and gouache. Magenta also mixes red with printers inks and colored pencils. I have not tried it with oil but I suspect it works the same way.

  14. idyllis says:

    fantastic! I just tried this yesterday with a yellow and a magenta, and I got red. but it’s not as vibrant as the tube paints for some reason? And also what cyan would you recommend (especially in the sennelier and golden series) because most primary cyans are dark

    • Taca, that is a really important question. The key point is that you can not make cyan, magenta, or yellow. You need these colors in your palette. Everything else can be mixed.

  15. Marcia says:

    Just came across your post,
    and found it reassuring,
    as I have been using Tom Lynch’s Holbein pac for years:
    Opera, Lemon Yellow, Peacock Blue.
    BTW, the light and pigment systems are not exactly opposite;
    although to mix green, red and yellow in Light, we do come up with white. Pigment: it’s blue, red and yellow to create an almost black.
    This is covered in additive and subtractive theories.
    Try the Holbein set, it is miraculous!

    • Thanks Marcia,
      I would suggest replacing Opera with a Quinacridone based magenta pigment. It will be much more lightfast. Also experiment with how you can make both red and blue from Tom’s primary set. If you can mix red and can mix blue (secondaries), then you can also mix any tertiary colors from the original primaries as well. I will look into Tom Lynch’s work. Thank you for putting him on my radar.

    • The Crazy Art Teacher says:

      Names like “opera” and “peacock blue” are fake names for ordinary color mixtures. My pet “peacock” is an entirely different blue than Holbein’s pet peacock. These ways of naming colors is bogus, and not very helpful. It is purely a marketing ploy.

      There is no substitute for learning how much raw sienna to mix with pthalo blue to get a purple. If you add a tiny bit of pthalo blue to raw sienna, you’ll get an olive green. If you add white to that purple, you’ll get violet. Violet is a secondary color and is a subset of purples.

      Theories are nice, but they cause confusion and they will never ever work in the studio.

      • The Crazy Art Teacher says:

        PS — warning: mixing pthalo blue with raw sienna will give you yucky mud. Just cautioning. Pthalo blue and pthalo green are overwhelming pigments, and really on work with white. Use cobalt or ultramarine for blue; whether watercolor or oil or acrylics. 🙂

        • Yes, I find those Phtalao pigments strongly staining and overwhelming. a little goes a looooooooooong way. Unfortunatly phthalo blue is also the cleanest cyan that I have found. Therefore I guedgeingly use it. It would be great to find a cyan that plays well with others. I use maganese blue hue as a non staining cyan but it is the other end of the spectrum, a very weak color. I find that cobalt and ultramarine are not as close to cyan (though cobalt is a little closer).

      • Yes!!! the common names are very confusing and not standardized. There is a standardized system for naming colors, an alpha numeric code for each pigment and the chemical name of the pigment itself. The same pigment will have different names as you move between brands. The pigment beta copper phthalocyanine or PB15:3 is called phthalocyanine blue by M. Graham and Utrecht, phthalo blue by DaVinci and Daniel Smith, primary blue by lukas and MaimeriBlu, and Windsor and Newton calls it winsor blue GS and manganese blue hue depending on its formulation. Lightfastness, staining and other properties change across brands. Very confusing. The best analysis of paints is click on any of the color family names at the top of the page and it brings you to side by side comparisons of water color paints from different manufactures. The research on this site is deeper than anything you could do yourself.

  16. Bing says:

    Hi what you had said is reasonable but not entirely correct:) The reason why you are saying so is because of the fact that you are using paint t prove your points. On the contrary, use blue red and green LIGHTS instead. On a white screen in a dark room, when u mix the different colors of light by shining them together, you will get the secondary colours magenta cyan and yellow! 🙂 Thus it us certain that red blue green are actually primary colours.

    • Hi there Bing.
      Here is a little more background to consider. There are two systems of primary colors. One for light, one for pigment. These are opposite systems. The more lights you mix the lighter the color becomes. The more pigments you add the darker the colors become. The primary colors of one system create the secondary colors of the other. The primary colors for light are, as you stated RGB that mix to the secondary colors CYM. The primary colors for pigment, CYM, mix to create the secondary colors RGB. It is the same color wheel but the primary colors are reversed. You can read more about this in this post.

      • I completely agree with you, John! Finding the right magenta/s and blues is key to a successful limited palette. I’m a fan of working in colored pencil when I want to play around with an idea and although my brand preference is Faber-Castell Polychromos pencils, I always keep a Prismacolor Process Red at hand because it’s the purest magenta I can find. In watercolor there are several good quinacridones but they will all mix differently with the yellows and cyan-leaning blues, depending on brand so I’ve found it important to experiment and back things up with color charts.

        • Kathleen, I just visited your website. Fantastic watercolor illustrations. I love your work. I recently discovered Faber-Castell Pencils and I love them too. What are your top choices for primaries in Prismacolor and Faber-Castell? For FC pencils my primaries are Phthalo Blue, Fuchsia, and Light Chrome Yellow For Prismacolor I use True Blue, Process Red and Lemon Yellow.

        • Delia says:

          John, Darcy: your 2 articles combined are amazing! Thank you so much for sharing all this information. My kids and I are exploring painting with watercolors this summer, and I was disappointed to end up with such “muddy” colors upon mixing. These 2 articles explain it all! Very excited to finally understand color & to apply what we’ve learned 🙂

  17. Laura says:

    This is incredibly useful. When I started painting 30 yrs ago, this is how we were taught. Recently I picked up the brush again when someone gifted me a set of paints. I cringed because it was based on the red, yellow, blue pallete. It is frustrating to not be able to get the colors you know you can get with the magenta, yellow, cyan. I will never understand why they put kits together based on the red, yellow, blue wheel. There ought to be a law.

  18. Jeff Ray says:

    Any three equidistant colors of equal saturation can be “primaries”. The differences represented in this article are due to the nature of pigments and the gamut available given the chemistry of those pigments. CMY has more possible colors than RGB. Several systems have been made based on more than 3 primaries, and Polaroid created a system using two. Yes, two. The majority of people who viewed the images said they appeared to be full color. This is based on how we perceive color, our eyes first find the “reddest” possible color in an image than they lock onto the most “violet” color, other colors are interpreted by the brain if not physically present.

  19. Lisa H says:

    My daughter’s Waldorf school does a lot of painting, starting in kindergarten with yellow, then blue, then red as primary colors. The paints they use are from Germany, and blend to form the familiar secondary colors, unlike any other paints I’ve seen. I haven’t worked with them myself enough to know if their red is closer to magenta, or their blue to cyan.

    • Interesting that you mention Waldorf classrooms. I was in a Waldorf school recently that was using the CMY system for color mixing and doing so successfully. To see how well the two systems work compare a violet made with blue and red vs cyan and magenta. Also see if it is a clean green as you would get with cyan and yellow vs the dull green you get with blue and yellow. If they are indeed using magenta and calling it red, I think this is not helpful. Red is a useful term and we should keep it meaning red. Magenta needs to be in our color mixing vocabulary. The link at the end of the article from the Science Education Foundation is very helpful for teachers. They have great resources for teachers to experiment with and teach this to children.

      • Christopher Scappaticci says:

        I teach in a Waldorf school and we use watercolor paints from Mercurius. The 3 basic colors we use are ultramarine (blue), lemon yellow, and carmine (red). We add (beginning in 3rd grade) Prussian blue, golden yellow, and vermillion. We mix all of the intermediate colors from these. Occasionally teachers may pre-mix green, but usually we are painting wet on wet and mixing on the paper.

        • Hi Christopher, At your prompting, I have been exploring Mercurius online. I love the fact that they are non toxic watercolors. It is hard to tell from the sample colors that they show online but it does not appear to make a clear magenta. I suspect that prussian blue is closer to a cyan than ultramarine blue. From what I see online with their color samples I don’t think you will be able to make a satisfactory purple with carmine and ultramarine (I guess it would come out dull) and you will not be able to paint a bright lime green ultramarine will make a dull muted green with lemon yellow).
          Interestingly Goethe’s colour-circle (which Mercurius claims their paints are based on) recognizes the importance of Magenta which he called Purpur and this is not included in the primary set from Mercurius.
          On their website Mercurius describes creating the color circle from RYB but as you can see in their examples, they are unable to create a bright violet, or bright green (they also would not be able to produce magenta, pink, or cyan). See If you could make a few labeled color wheels with those colors and send them to me I can give you even more specific feedback.

        • I am also a Waldorf teacher. I love the Mercurius line. Have you used heir “circle paint” line? They are the true CYM primaries. That’s what we use in Kinder -2. My class is in 2nd, and still love the variety of colors they can make. My previous class used the ultramarine, lemon, and carmine. I could never get a good purple that wasn’t muddy. Try the circle paints! My color wheels come out incredibly bright.

  20. Hannah Hinchman says:

    Jack, what an enormously useful post. I’ve known this for a long time (on an operating level) but have really blundered trying to explain it. I guess I’ve just not been confident enough to tell it like it is! Thanks!

    • This exact dilemma was what frustrated me so much in primary and high school. And even when I took watercolor classes at the American Academy of Art, there was no clear explanation of primary colors versus pigments. Once I got a job in the printing industry (in pre-press work) I had my light bulb moment and started searching out watercolor paints and colored pencils that offered a true magenta.

Leave a Reply

Your email address will not be published. Required fields are marked *