Indigo – a colour of the rainbow?

From time to time I come across web pages and groups of people who get irrate about indigo being in the rainbow. There is even a facebook group called “Get Indigo out of the rainbow”. It was Newton who suggested that the rainbow contains seven colours: red, orange, yellow, green, blue, indigo and violet. It has been suggested that, at the time, Newton was trying make some anology with the musical scale and the octave (with its seven intervals) and hence was keen to identify seven colours in the rainbow or visible spectrum. Many modern commentators claim that only six distinct colours can be observed in the rainbow.

Interestingly, the facebook group referred to above would like to eject indigo from the spectrum on the basis that it is not a primary or secondary colour but rather a tertiary colour. The group shows the following colour wheel:

colour wheel

In this so-called painters’ wheel the primary colours are red, yellow and blue and the secondary colours are orange, green and violet. It is argued that since six of the colours in the rainbow are primary or secondary colours in the colour wheel and indigo is not, then indigo has no right to be there. This is wrong on so many levels it is hard to know where to start.

The first thing I would have to say is that this argument seems to ignore the difference between additive and subtractive mixing. Additive mixing – http://colourware.wordpress.com/2009/07/13/additive-colour-mixing/ - describes how light is mixed and the additive primaries are red, green and blue. The additive secondaries are cyan, magenta and yellow. Orange is not in sight – and yet surely if we are to make an argument for inclusion in the spectrum based on primaries (and/or secondaries) then it is the additive system that we should be using since the spectrum is emitted light.  

The optimal subtractive system primaries are cyan, magenta and yellow (with the secondaries being red, green and blue) though the artists’ colour wheel (which is like the painters’ wheel above) has red, blue and yellow as the primaries. 

In my opinion there is nothing special about the colours that we see in the spectrum. Indeed, orange is clearly a mixture of red and yellow and does not seem to me to be a particularly pure colour. I just do not think that arguments to exclude indigo from the spectrum based upon colour wheels or primary colours is valid. That said, I have already mentioned that many people believe that indigo cannot be seen in the spectrum as a separate colour; but this is a phenomenological observation not dogma. I am one of those who believe that indigo and violet cannot be distinguished in the spectrum and therefore I agree with the aims of the facebook group even if I do not agree with their arguments.

The really interesting question is why we see six (or even seven) distinct colour bands in the spectrum when the wavelengths of the spectrum vary smoothly and continuously? I have postulated some possible reasons for this in an earlier post – http://colourware.wordpress.com/2009/07/20/colour-names-affect-consumer-buying/ - but it is far from a complete and convincing explanation. It may explain why we see distinct colours in the rainbow, but why six and why those six in particular. Comments on this would be very very welcome.

46 thoughts on “Indigo – a colour of the rainbow?

  1. Perhaps this, like the conventional (red-yelow-blue) primary colours, reflects the subconscious influence of the opponent hues. Colour is an experience that lies between the poles of redness vs greenness and yellowness vs blueness, so it doesn’t seem at all surprinsing that those four colours should have been singled out by Newton. Theodoric of Freiburg saw these four colours as making up the rainbow, and Aristotle came up with something very similar (red, green, purple, plus yellow by “contrast” of red and green). There’s a complication in that Newton’s “blue” doesn’t seem to be our blue. I know that this question is debated, but Helmholtz for example regarded indigo-blue as having the hue of ultramarine pigment ( which in turn is the hue of our modern RGB “blue”), and introduced the name cyan-blue for the greenish blue part of the spectrum included in Newton’s “blue”. Accepting this correlation, Newton’s three remaining spectral colours would constitute three of the four possible combinations of the unique hues – yellow-red (Newton’s “orange”), green-blue (Newton’s “blue”/cyan) and red-blue (Newton’s “violet”). In opponent terms the seven-hue spectrum is thus R – YR – Y – G – BG – B – RB. I don’t know why yellow-green didn’t get its own name in the spectrum, but there isn’t much in the way of a good name for this hue in general, and perhaps its position at the peak of the photopic sensitivity curve makes us tend to see it as an overlap of yellow and green, instead of a distinct intermediate hue?

    Cyan and (“indigo-”) blue look to be reasonably distinct entities in the spectrum to me, despite all that has been wrtitten about Newton making up an extra colour to fit his musical scale. It seems plausible that cyan didn’t get established in the spectrum because there was no room for it in the conventional “artist’s colour wheel”, where there is room for only one “blue” (again, due to the subconscious influence of the opponent hue categories). Now that the name has acquired widespread currency as the blue-green subtractive primary, perhaps we will start seeing it in the spectrum again!

  2. I red your post today and was surprised. I was born and grew up in Ukraine (former Soviet Union), and was taught in school that colors of rainbow are: Red, Orange, Yellow, Green, Sky Blue (or azure we have a specific name for it, it is similar to #007FFF),Blue, Violet. I know that the same is in Russia.So it’s interesting to me that in some Countries we are taught to see indigo and in others Sky Blue

  3. Hi Stephen

    I will start shortly a whole series on the rainbow/spectrum on my blog at mycarta.wordpress.com. One of the things I will try to explain and possibly resolve (with more perceptual rainbow colormaps) is the presence of those bands of nearly constant hue. Thanks
    Matteo

  4. Well this helped answer my question a little bit. But I am hoping I can find further help on the matter. I teach art to elementary school children and some of them have learned or will learn about the SEVEN colors of the rainbow in science classes. When I was a kid, I have no memory of indigo ever being included in the rainbow. This could just be my bad memory, I’m not sure.

    So one, is there a simple way to explain to young children why some people say the rainbow has seven colors and some people say it has six?

    Also, why indigo? What made indigo get to be the only intermediate (or tertiary) color in the rainbow? Why not yellow-green?

    Thanks
    Meg

  5. If I was teaching a group of elementary school children about this I think I would wait until there was a great rainbow outside and then take them outside and look at it and ask them what they see. Do they six colours or do they see seven? Let them decide for themselves rather than impose some sort of dogma about what is meant to be in the rainbow. This also teaches scientific method. Rather than accept something as a ‘fact’ why not observe what is going on and draw your own conclusions.

    I have never come across any reasonable explanation yet, as to why we see distinct colours in the rainbow when in fact the wavelength of light varies continuously and smoothly from short wavelengths (bluish purple) to long wavelengths (reddish purple). So why do we see distinct colours and why those particular 6 or 7 colours? I am afraid there is no easy answer to this.

    1. The reason for the colours we see is to do with the response of the eye, which contains colour-sensitive dyes which absorb light of different frequencies. Our colour perception depends upon the amount of light absorbed by each of the dyes.

      1. Indeed, but this does not even begin to explain why we see six or seven distinct colours in the spectrum when the spectrum varies smoothly and continuously in wavelength. The question is why is colour vision categorical in this situation.

  6. I can’t wait for a rainbow, unfortunately. I’ve never seen one anywhere near the school, and But I can show them pictures. Maybe I can create one in the room using a prism. I’ve told them that some people use indigo in the rainbow, but when we color them, we don’t have indigo colored pencils so they can either leave it out or mix the colors themselves. I kind of wanted to use the ROY G BIV song. I guess I’ll just give them a choice.

  7. I was taught to remember the colors in a rainbow by “ROY G BIV”….Red Orange Yellow Green Blue Indigo and Violet…Pluto will always be a planet to me and indigo will always be in my rainbow:)

  8. We see distinct colours by the way our receptors work. Most people, as you know, have three sets of cones with a specific peak frequency which determine which colours we see. It’s the combination of the sum AND the difference which gives us the six (primary+secondary) colours. The very basic sums can be seen in a television colour bar chart – usually eight bars with the addition and subtraction of all light – black and white. This, done electronically with our subjectivity taken away.

    It may be that Newton describing the initial colours of the split light, named cyan as blue and blue as indigo to differentiate between the two in that part of the spectrum. I’m not sure that ‘cyan’ would have been widely in use at that time. This concurs with David’s comments and particularly Sasa’s comments. Next door in Russia, Blue is Galuboy and Cyan, Cyny.

    I find it interesting that the cyan colour in the painter’s chart is very dark and indistinct, yet cyan (or sometimes aqua) seems quite ‘bright’ and is used widely for products through its ease of printing! We in the west seem to have a blind spot for it. Even some website explaining colour, print a spectrum without cyan which looks odd when compared with a photograph of a spectrum. The Twitter logo seems Cyan to me with Facebook seeming a ‘real’ blue.

  9. Thanks for the comment. I still think you are not quite getting the point though. Yes, the photoreceptors underpin our colour vision and each wavelength of light in the spectrum gives a unique ration of these cone responses. This allows us to distinguish one wavelength from another. At best we can distinguish differences of just 1 nm. Towards the long and short ends of the spectrum, of course, our performance falls off and we can only distinguish changes of about 5 nm. However, that still leaves about 200 different hues that we can see. The real question I am asking is why we see bands (why do we see bands of red and yellow, for example, with quite sharp cut-offs?) rather than a continuous variation of hue. I don’t think this is easy to explain from quantum catches of cone responses. Quantum catches (and particularly Rushton’s principle of invariance from 1972) do explain why red, green and blue lights add together to form cyan, magenta and yellow which may be the point your are making. But that’s quite different to which colours we see in the spectrum.

    Great comment about cyan.

    1. Thanks for the reply. I have to say that my knowledge of quantum catches is far less than yours(!) I can only make an analogy with other models. I will certainly read up on it, thanks for the information (on Rushton).

      From my limited perspective at this moment, I would hypothesize that the cone’s response curve determines the ‘bandwidth’ as a strength of that particular single colour; more specifially, how the cone’s output reduces in amplitude but NOT frequency as the incoming wavelength moves away from the centre peak. it may be the interaction with other cones which determines the different hues – particularly with yellow since there is a major overlap of response between the red-green cones. This summing of responses would also explain the hump at yellow on the L* graph on Matteo’s page entitled ‘The rainbow is dead – long live the rainbow.’

      Furthermore, the shape (or bandwith) of the response curve of each cone may determine the width of the band and the relationship with other cones. The overlap between red and green is quite pronounced, but the response curve for blue is well down on the other two, and the overlap between blue-green seems much less pronounced, perhaps leading to the cultural ‘blindness’ to cyan. On the L* graph, Cyan is about 50, fairly close to the red response of 50. (Nothwithstanding, the eye’s rod response also peaks in that part of the spectrum.)

      A test of this might be to pass different hues through a single coloured plate to a photoreceptor and note the output. E.g. to observe the hues from a television through a single channel, e.g., Red. What other hues might be seen in this single channel? Probably only the narrow range or ‘band’ of reds? Would orange also be seen in this channel?

      Apologies if this seems a little simplistic. but I’m open to other ideas.

  10. @swestland (sorry I don’t know your name) – more specifically: Regarding cone response. Let’s take one cone with a peak response centred at 630 nanometres(nm) representing RED. Let’s say for discussion purposes only, that the curve slopes down to half the output (50%) from the cone between wavelengths of 650 nm and 610 nm.

    That is the BAND width. For REDness.

    Let’s say the level of signal also falls to a 10% of the peak signal at 670 nm AND 590 nm, ie the slope is pretty steep between these 10% points and the 50% point. That will give you the perception of a sharp cut off.

    So between 670 nm and 590 nm, the signal from the RED cone to the brain will increase in AMPLITUDE to the peak at 630 nm and then decrease back down again.

    The output signal from the RED cone to the brain varies in amplitude but NOT in hue. Since the ‘voltage’ at 670 nm OR 590 nm is the same level (at 10% of the peak.) It’s only when a simultaenous signal from another cone is processed in the brian will another hue be perceived.

    Therefore the bandwidth of the cones determines the bands within colour perception.

    regards,

    Steve :-)

  11. Just a little comment on this:

    “It may be that Newton describing the initial colours of the split light, named cyan as blue and blue as indigo to differentiate between the two in that part of the spectrum. I’m not sure that ‘cyan’ would have been widely in use at that time.”

    Newton at one stage actually used the term cyan (cyaneus) for the “blue” interval of his published works. In Part II of the revised version of his Cambridge lectures he first describes a five-hue spectrum of red, yellow, green, blue (caeruleus) and “purple or violet”, but later he refers to more finely divided spectra of eleven hues (Lecture 3) and ten hues (Lecture 8) in which the place of caeruleus is taken by two “gradations”, cyan (cyaneus) and indigo (indicus). Indigo reaches its canonical position as a hue BETWEEN blue (caeruleus) and violet in Lecture 11, where Newton gives his (ostensible) reasons for adding just it and orange to “the five more prominent” colours, and then goes on to introduce his musical correlation.

  12. @matteo I like your page ‘The rainbow is dead…long live the rainbow! – Part 3′

    From your Gregory quote: “where the response curves have their steepest opposite (one going up, the other down) slopes … [and] … we should thus expect hue discrimination to be exceptionally good around yellow – and indeed this is so’.

    That’s how I understand it. In addition, I also agree with Gerald in his October 9th post about the green-blue boundary (Cyan – again!).

    As you say, it’s the shape of the discrimination curves of each cone that determines the wide bands of Red, Green and Blue and the narrow infills of yellow and cyan.

    In fact my Paintshop Pro software program does something similar.

    I’ve attempted to show the original Gregory ‘analogue’ and Matteo’s ‘ digital’ spectrums along with the Paintshop Pro’s palette to show the differences for comparison.

    This is a TEMPORARY web page which I shall pull down as soon as parties have seen it – say a max of 14 days: http://www.odyssey.ltd.uk/colour.htm

    I can certainly understand why one would want to ‘linearise’ the spectrum for discrimination of colour mapping purposes.

    regards,

    Steve

  13. Thanks Steve! :-)

    I’ve just updated the page with television colour bars to show how the colours are distributed digitally (you’ll be able to work out the underlying logic).
    The colours look muted because analogue teleivsion sent the information at about 90% saturation.

  14. Indeed, very cool.

    And thanks for expanding on Gregory’s quote. That’s exactly how I would have put it. Great example.
    My favorite example is the flat green band.
    You can see it Gregory’s rainbow but it’s more evident in the CIE 1931 spectrum palette –
    here is a digital version:
    http://mycarta.files.wordpress.com/2011/11/natural-spectrum1.png
    When I then look at the three color response curves here
    http://www.atmos.ucla.edu/~fovell/AS3/theory_of_color.html
    I see that the green curve has a rather flat apex (peak response), with gentle gradients on the side, and towards the blue there is no overlap until down to about 500 nm. Hence the flat band.

  15. As an aside a few things I’ve read on Newton and colors:

    – George Biernson, in 1972, wrote in the American Journal of Physics the paper Why Did Newton See Indigo in the Spectrum? and “hypothesizes that Newton saw seven reasonably distinct colors in the artist’s paint mixture color circle (red, orange, yellow, green, blue, violet, and purple) and therefore assumed he could also see seven distinct colors in his crude spectral projections”. Many others have argued Newton was trying to add a seventh color to match the seven notes of the western world’s musical scale, as Stephen pointed in the post.

    - Goethe’s color wheel was made of only 6 colors.
    http://www.homodiscens.com/home/ways/perspicax/color_vision_sub/art_color_theory/index.htm
    Newton fiercely criticized it, to the point of ridiculizing Goethe. The story is discussed at length in Chapter 6 of Jan Koenderink’s wonderful book Color For the Sciences, and his Goethe Edge Colors palette is wonderful, and elegantly derived. You can also read about it in here:
    http://home.earthlink.net/~johnrpenner/Articles/GoetheColour.html

  16. Hi Matteo,

    I have to admit, I had not seen Goethe’s colour wheel before.

    I thought I would find out what colours they ‘actually’ are according to my Paintshop Pro programme.

    I isolated the most representative Goethe’s colour samples of each segment from the wheel on ‘…Homodiscens’s…’ page and analysed it for hue (and saturation).

    Although the record of the wheel has aged and possibly yellowed, it is interesting to see the hue values of each colour. In the odd case, I’ve blured the image to ‘mix’ the underlying two colours he attempted to mix to portray the colour he observed. Some of the colours are a bit wishywashy because of the saturation, and there is a different in brightness because of the way the different programmes portray colours but I think it was worth the effort.

    Rather than to ‘tell’ you what colour, I thought I would update the page and to allow you to see each colour under analysis: http://www.odyssey.ltd.uk/colour.htm

    ‘Homodiscens’ digital version looks a bit odd in my opinion, as expected, the cyan segment has been transcribed as ‘blue’

    Goethe, interestingly senses Cyan and Indigo….

  17. That’s really good material Steve, thanks for sharing. With the actal colors now we can try and make a palette and play with it. We just need some good test images… :-)

    This guy wrote an interesting post on color wheels abd came up with an intriguing one, wheter right or wrong:
    http://blog.asmartbear.com/color-wheels.html
    The post generated interest and quite the controversy in the comments.

    And this is a fantastic page on human color vision:
    http://www.handprint.com/HP/WCL/color1.html

  18. Hi Steve

    Thank you for sharing, this is great to be able to see the colors in Photoshop, and to replicate them. Now we can make interpolated palettes and try them on some fun test images.

    More on color palettes:
    http://blog.asmartbear.com/color-wheels.html
    This is an interesting article. The author propose a 4-primary-color wheel that looks intriguing to me.
    His ides generated quite the controversy, so make sure you read through the commens section too.

    More on human color vision:
    http://www.handprint.com/HP/WCL/color1.html
    This is a fantastic read.

  19. Hi Matteo,

    I particularly value the ‘handprint’ web page, particularly the eye resonse curve to UV light.labelled: ‘prereceptoral filtering in an adult eye’. People who have had an operation to remove the cornea/lens report seeing perception of extended UV.

    Jason Cohen has attempted to ‘reinvent the wheel’ by transposing the Opposing Colour Theory (Ewald Herring) onto a wheel. Worthy of discussion but probably not worthy of being called the ‘right system’ for it depends on WHICH system one wants to use in the CONTEXT required.

  20. No one has yet mentioned culturally shaped perception of color. Some cultures “see” color that includes a physical texture component. Across cultures people “see” different arrays of color, People raised in particular cultures have been experimentally shown to have differing perception of actual wavelength color. Not just a naming difference but apparent perception difference as to when two colors are identical or obviously not identical. Similar to the way people raised with a language can not differentiate sounds that are easily distinguishable to speakers of a different language. Brain mediation and interpretation can’t be left out of a discussion of perception of color. Don’t have tight references at hand, but check anthropological and psychological studies of color perception. Also a recent Nova or National Geographic program documenting color naming in which people in some groups easily and accurately discern five shades of green where other groups see only one.

  21. The important question is do you actually get seven colours from three primaries? Whether they are the colours of light or the colours of matter there’s still only three primary colours which can only make three secondary colours. There’s also only the three cone receptors in the eye for the primary colours. So how can it be anything but six colours of the rainbow?

  22. Indigo does not seem to be a primary or a secondary colour whether you look at the six colours of light or the six colours of matter. If you said indigo was a colour of the rainbow then you would have to add all the other tertiary colours as well. Either way seven colours of the rainbow is not possible.

  23. hey, if someone bothers to answer then here is the question: Why say that the rainbow has 6 colors only because human’s eye, mostly, can’t identify one?

  24. Well put, though you raise more questions and do not answer the original question other than alluding to the idea that the colors are arbitrary. However, I agree with your logic.

  25. This is all fascinating! I just stumbled across this whole discussion after searching online for an answer as to why indigo seems to be so frequently omitted in depictions of rainbows in children’s books. I had no idea that there was actually a controversy over the matter!

    The science of how many nanometers are in each color’s wavelength or how the eye functions is over my head, but I would like to throw a few things into the fray.

    First off, in Joseph Chilton Pearce’s book, The Biology of Transcendence, he makes the following statement: “Twenty years ago the average subject could detect 350 shades of a particular color. Today the number is 130.” He makes similar claims about the decline in the sensitivities of our other senses, but I wanted to mention his findings on color. Pearce lays the blame mostly on how television screens have affected our eyes, which seems logical to me.

    If Pearce’s claim is true, how does this factor into the debate? It does seem to be the case that different cultures in the modern world think about, label and actually perceive color differently, but human eyes may have also changed (deteriorated?) over the years as well. Is it possible that Goethe, Newton, et al., simply had a different capacity for perceiving color than we have today?

    I have two other random observations. The first one comes about as a result of the fact that I am a yoga teacher who is fascinated by the study of the subtle energy centers known as chakras. The system of yoga which I learned teaches that there are 7 chakras in the human energy system and each one is connected with a color from the spectrum of natural light, going right in sequence with good old ROY G BIV. The ajna chakra, or “Third Eye,” is associated with the color indigo. (Not all systems of yoga agree with this color schematic.) I don’t know as much about this, but there are other systems of yogic chakra anatomy which say there are only 6 human chakras. I can’t verify it, but I would bet money that such other systems leave out a connection with the color indigo.

    Lastly, I was doing some reading recently about the mantis shrimp, a little crustacean which holds the distinction of being able to see the most colors of any animal. They have fascinating and complex eyes! (Check out their Wikipedia listing at https://en.wikipedia.org/wiki/Mantis_shrimp.) I wonder what these little critters would see if they were looking at a rainbow!

    Anyway, there’s my ramble. Thanks for reading!
    Jim

    1. I offer a couple of thoughts Jim. The physiology of the eye has probably not changed, but our internal sense of colour may be influenced by the reduced or limited colour gamut of modern visual displays that we appear to spend a lot of time in front of! However, I’d like to see any data on that but it is plausible. Maybe we are getting used to the combination of colours so we can recognise a broader RELATIVE shade (red to green to blue etc.) rather than its absolute uniqueness. i.e. Jade, beige, pink, umber, teal, crimson, sage, international klein blue, etc, etc.

      However, in real life, switching to 16 bit colour on your display will probably show up significant degradation compared with 32 bit colour settings. Alternatively a visit to a paintshop with their charts could be interesting. You may find this link useful:
      https://en.wikipedia.org/wiki/Color_depth

      I think colour is a superb cognitive way of identifying different parts, segments or single elements.. In addition to your Chakras, Electrical resistor components are also encoded by colour: http://www.circuitstoday.com/working-of-resistors

      Black…0
      Brown…1
      Red…2
      Orange…3
      Yellow…4
      Green…5
      (Light) Blue (In many cases It’s cyan)…6
      Violet…7
      Grey…8
      White…9

      so a 120 Ohm resistor would have three rings; Brown Red and Black (other rings are for other purposes)

      Regarding Chakras, there appears to be a close correlation with the Endocrine System and the seven major hormone producing organs, which are essential regulators of the body’s functions, but that is another story.

      Yes, the mantis shrimp would find our world pretty dull!

  26. My theory is that he called “blue” what we call “cyan” today, and what we call “blue” today he called indigo.

    I think his choice of words might have been influenced by the fact that indigo dye was one of the most valuable products around at the time.

Leave a Reply

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


nine − = 3

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>