Written by Paul Baize (1901-1995)

Translated from French by Nick Loveday (1981)
Several Parts Re-Translated by Andrew James (2002)
Edited by Grant Searle (1981) and
Further Updated by Andrew James (2002)


Written by Andrew James

In my own opinion, the following paper is one of the best ever produced on the subject of double star colours. It is unusual because it discusses an avenue of double star observation that was not very much in vogue at the time. This paper concentrated on the ones that show the most significant colour contrasts, like the famous β Cyg / Beta Cygni (Alberio) and γ And / Gamma Andromedae, and attempts to explain why these colours appear so spectacular and so pronounced.

I do not know the origin of this paper, but it was given to me as a photocopy while I and about two dozen of its members were conducting the Astronomical Society of New South Wales (ASNSWI) Double Star Estimates Program” during 1979 to 1982. The original article was in French and was kindly first translated into English by my friend Nick Loveday in 1981. After many hours searching I cannot find the source of the paper. As it is now nearly fifty years old, (1) Im sure I have no real fears of breaching copyright and have been very careful to acknowledged this sufficiently. If do you recognise this paper, I would truly appreciate if the reader could contact me through the link below, so I can properly acknowledge the reference and reproduce this article. If you need to formally reference this article, please kindly refer to this page in your added notes.

Paul Baize was one the most influential of the European double stars observers during the 20th Century. French by origin, he produced many observations and micrometric measures of double stars and contributed much in promoting the subject. He started his career as a physician but was a keen amateur astronomy when he started observing double star in 1924. This new study he continued to do uninterrupted until 1972, after he had in total measured some 25,000 pairs. Baize also produced one of the first red star catalogues, contributed significantly for some 140 new binary star orbits, and during 1928, was among the first modern observers to produce a decent history on double stars. (In French) This was after those old school histories like those of Crossley, et al. (1870) and Thomas Lewis (1908), but was before the interesting small histories written by giants like Robert Aitken (1935) and Wulff D. Heintz (1976).

Paul Baize sadly passed away on 6th October 1995, and the world lost one of its greatest modern double star observers.


Astronomers have in the past described double stars as celestial rubies, topazes, emeralds or sapphires; while the most interesting colourful pairs have yellow or orange primaries and blue, green or violet secondaries. Generally when looking at stars through a telescope these colours are deceptive. The lively colours, contrasts and shades described are neither as clear or as frequent as one might believe from some descriptions. Most stars produce an image quite lacking in any colour, pale and washed out. The only doubtful exceptions are double stars and the colours they appear to have are probably much in error due to the various defects inherent in our processes of visual perception. These defects include the contrast phenomenon and colour blindness.

To understand how these defects affect our perception of stars some knowledge of the structure of the eye is necessary. Rather than discuss that here, I refer the reader to any respect to the above encyclopaedia, text on optics (including Amateur Telescope Making) or texts on photography. The essential photo-receptive surface in the eye is the retina, the light sensitive elements being known as rods and cones, due to their shape. The rods provide the overall sensation of light or dark while the cones detect colour. The retina has an extra sensitive area called the macula containing mostly cones, this is the area we use most intensely for precise and acute vision. Cones however have a poorer response than rods under low levels of illumination, hence the use of indirect” vision to see faint objects in the telescope.

The mechanism by which the cones detect colour is not perfectly understood and the Young-Helmholtz theory will suffice for the purposes of this article. (2) In this theory each cone, while appearing to be a single item is really made up of three fibres. Each of these fibres can detect one of three elementary colours and there is one of each kind in a normal cone. Suppose the three elementary colours are Red, Green and Violet [,for example] (R, G, V) (3)

Unfortunately the perception of colour is also affected by the overall intensity of the light reaching the eye — the ratios of the maximum response to light of each colour is not independent of the level of brightness, and a single colour excites more than one fibre in the cone. Hence in the case of starlight which is never highly saturated (the colours are pale and there is a significant amount of light at all wavelengths in the spectrum) and which is very much dimmer than the normal level of illumination to which the eye is accustomed, gross distortions of any colour the star may have are inevitable. Three phenomena are particularly important:


1) The complementary colour effect is easily demonstrated by gazing for a few minutes at a sheet of paper having a strong colour, eg., red and then looking at a white piece of paper — it will appear to be a colour which is the complement of the first colour, in this example, green. Similarly, orange and yellow, blue and violet are complementary pairs of colours. This effect is produced by fatigue of the neurons responding to the first colour — when one looks at the white sheet of paper those neurons are tired, and the eye perceives the sheet as white minus the colour of the first sheet, I.e., The complementary colour.

This effect can occur with objects looked at in turn, as above, or simultaneously in the sense that one turns the eye to receive a star image of one star at a time on the macula, so that in the process of looking from primary star to secondary star and back again, the colour difference in a double star will be exaggerated. This is also supported by the fact that single stars (if a small enough field is chosen) rarely show any colours except orangish or white — certainly not green, violet, blue, etc. In addition, if one component of a double star is hidden from view by a thick cross hair, the other component loses all appearance of being brightly coloured. Another way to demonstrate this is to view combinations of stars and planets — Saturn appears green when next to Mars.


2) The Purkinje effect is also interesting: if a coloured point of light of variable intensity is adjusted slowly from zero to maximum brightness, the first appearances gives no impression of any colour at all, and it is only at a certain level of intensity that any colouration other than white can be detected, and it must be still higher before the colour can be recognised with any certainty.

On the other hand, Purkinje also showed that blue and violet were perceived as colours at lower intensities than other colours at lower intensities than other colours, White surfaces under low illumination appear blue-grey.


3) Colour blindness is the inability to distinguish between various colours, particularly red and green — the eye loses the ability to respond to one colour in particular; in most cases, red. John Dalton was one of the earliest people to recognise the causes of this problem; he could not distinguish between the leaves and flowers of geraniums. About 7% of all men suffer from red-green colour blindness, and a much smaller percentage of women. Most normal people see light in the range 450nm to 790nm (4), while others might be able to perceive light over a range starting well above 450nm going to well above 790nm in the ultra violet; those people would be unable to detect red colours. Others might have a range starting in the infra red, and be unable to see violet.

Without going into the explanations of these effects their existence is hardly surprising, considering the fact that stars are seen only in conditions for which the eye is not well adapted. The result is great variation in the colours of double stars as seen by various people. For instance Eta (η) Cassiopeiae was described by J. Herschel as red and green, yellow and blue by Dawes, yellow and lilac by Flammarion, yellow and reddish by the author, which corresponds to the spectra, F8 – M1. Σ648 (05h 04m 36s +31° 55′) was described as yellowish and bluish by Struve, red and red-orange by Dawes in 1842 and in 1846 white-white again by Dawes, yellow-yellow by Secchi in 1857 and again white-blue by Secchi in 1858. The companion of Mu (μ) Cygni was seen as blue by Struve from 1826-1833 and remarkable red” in 1836; Dawes noted it as blue, Engelmann red, Duner red, Perrotin orange, Lewis blue, etc. Since 1921 the author saw it as white. There are countless other examples. (5) In addition if one person sees his right eye then the left eye he will probably also see different colours. (6)

Overall, we may thus conclude that the colours
seen are highly subjective and unreliable.

Following this conclusion regarding colour effects the next step is to try and find an objective method of determining stellar colours. A brief glance at the sky shows that some colours are real. Aldebaran and Betelgeuse are definitely different to Rigel and Sirius. On the whole stars are yellowish, with some white and a still fewer orange or reddish. Various devices called colorimeters were devised by Secchi and Zellner but they have been abandoned because they did not compensate properly for the personal effects. In fact there are still two rigorous methods that can be used : the determination of effective wavelength and the measurement of the colour index.

(a) Most readers will be familiar with, direct vision spectroscopes, and the kinds of spectra that can be seen ; line spectra such as those produced by elements or simple molecules and continuous spectra, such as those produced by filament lamps. In addition there are dark line spectra, due to a cold gas or vapour being located between a source of a continuous spectrum and the observer — the gas absorbs radiation at the frequencies it would normally produce bright lines if it were excited.

Through a small spectroscope of low resolution a star shows a continuous spectrum and while the stars emit radiation across the whole of the spectrum. Cooler stars emit more strongly in the red and feebly in the blue, while hotter stars emit more strongly in the blue-violet part of the spectrum. If one studies the starlight from any star with the aid of filters and bolometer using precise photometric standards, that star will have a particular wavelength which is more intense than any other in the visible spectrum — the dominant colour. This wavelength can be determined by Wiens Law, which states that the most intense wavelength is a function of only the temperature of the star.

However, the star also radiates at all other wavelengths by varying amounts and so the colour is not saturated, in fact it is very washed-out by the presence of all the other wavelengths in the spectrum. The only astronomical objects for which this is not true are nebulae, which are composed of gas and dust particles and produce brighter spectra (emission nebulae) or dark line spectra (absorption nebulae and reflection nebulae). Emission nebulae are often strongly coloured because they emit light in only a few bands of the spectrum.

Franks has done a study of stellar colours spectra; the results are summarised below:

O White, sometimes a bit yellowish 12
B Very White 242
A White 1190
F0-F5 White-yellowish 545
G-K Yellow 1238
K2-K5 Yellow-orange 109
M Orange 150
N Red-orange 11

(b) Needless to say, because of the differences between the response of the eye to the various colours it, the spectrum and the responses of photographic emulsions, stars often appear on photographs as having magnitudes distinctly different to their visual magnitude. The colour index is defined as the visual magnitude minus the photographic magnitude. As examples Betelgeuse and Aldebaran have photographic magnitudes of 3 or 4.

The reference for the colour index is selected in the middle of the spectrum (white stars) see that white stars (e.g. Sirius, Vega) have a colour index of zero, and the range of spectra B, A, F, G, K, M correspond to the colour index range -0.4m to +1.6m in 0.4 steps :

Spectral Class B A F G K M
Colour Index (CI) −0.4 0.0 +0.4 +0.8 +1.2 +1.6

This has been refined: (King, Parkhurst & Jordan, Schwarzschild, etc.)

Spectrum Colour Index Description

B0 -0.42 White-Blue
B2 -0.21
A0 0.00 White
A5 +0.21
F0 +0.42 White-yellow
F3 +0.63
G0 +0.84
G5 +1.05 Yellow
K0 +1.26
K5 +1.47 Orange-yellow
M0 +1.68 Reddish

The description of colours among double stars is interesting. W. Struve found the following from a catalogue of 596 pairs :

Both Components Same Colour

White 295
Yellow-White 75
Reddish-orange 5

Components of Different Colours

White /yellow 43
White / blue 58
Yellow / blue 104
Green / blue 16

In the past some observers have used rather exotic descriptions of colours. eg. W. Struve described Alpha (α) Orionis as divine — deep ruby and Webb used a large range of rather subjective terms — indigo, grey-white, olivine, pale greyish-rose, bluish-red, brownish, etc. etc., and more bizarre ones are included.

Personally, my studies of many double star spectra (the 3919 stars in the Draper Catalogue) confirm that the table above made by F.G.W. Struve is quite close to describing the kinds of double stars visible quite accurately. Orbiting doubles are particularly frequently advanced types of F5-M5 and occasionally M (i.e. yellow or perhaps reddish) and on the whole the secondary is the same colour as the primary. Lick Observatory Bulletin No. 343 contains spectral measurements for 238 double stars, measured by Leonard, who found that in the vast majority the spectra of primary and secondary were almost identical. In the cases where the two spectra are not alike, the companion is almost always more advanced (redder) than the primary. For example, Eta (η) Cassiopeiae (F8 and M1), Xi (ξ) Bootis (G5-K4), 70 Ophiuchi (K0-K4), 61 Cygni (K5-K8), Alpha Centauri (G0-K5), Sigma (σ) Serpentis (A0-A5), Zeta (ζ) UMa / Mizar (A0-K5) etc.

Also interesting are double stars with identical spectra where observers see the components as having different colours. In this class are, Phi (φ) Draconis (yellow / lilac, F8-F8) and Alpha (α) Canes Venaticorum. (yellow / lilac, A0- A0).

Doubles where the companion is less advanced than the primary include Epsilon (ε) Bootis (G8-A1), ο Ceti (Mira, M0-A0) Epsilon (ε) Hydrae (F9-F5), but in most of these cases the primary star is two magnitudes difference making them giant stars. Under these conditions, this resulting and relatively uncommon situation, is especially when the stars are of very unequal brightness. This produces the phenomena of contrast which I have insisted on at the beginning of this article. It is the companion’s vicinity to the primary that produces the deep colours and the nuances like green or bluish. These are sometimes very beautiful but are to be stripped of any objective reality. Such are the pairs are well-known by amateurs.

The following doubles are well known to most amateurs and should prove interesting. Try looking at the double with both stars in the field of view, then with only one star visible at a time.

Table 1

h m
o ′
v mag.

v. mag.


Sep. Yr.
Σ163 AB Cas 01 51.4 +64 22 6.80 or 9.13 bb 34.7 / 00 K5IaO-a B5
Σ205A-BC / γ And 02 03.9 +42 20 2.31 jj 5.02 ve 9.6 / 00 K3IIb B9
Σ307 AB / η Per 02 50.7 +55 54 3.76 jj 8.50 b 28.3 / 98 K3Ib C/M3Ib-IIa
OΣ67 Cam 03 57.1 +61 07 5.25 or 8.06 b 1.7 / 91 K3I-II A0
OΣ72 Tau 04 08.0 +17 20 6.10 jj 9.71 b 4.6 / 98 K5IIIb
BU 87 Tau 04 22.4 +20 49 6.21 jj 8.60 b 2.0 / 83 B3V K3II
Σ654 / ρ (3) Ori 05 13.3 +02 52 4.62 or 8.50 b 7.0 / 95 K0
Σ997 / μ CMa 06 56.1 -14 03 5.27 or 7.14 b 2.8 / 99 G5III A2
HJ 3945 07 16.6 -23 19 5.00 or 5.84 b 26.8 / 91 K3Ib dF0
Σ1268 / ι Cnc 08 46.7 +28 46 4.13 j 5.99 b 30.5 / 00 G8Iab A5
Σ1441 AC Sex 10 31.0 -07 38 6.51 j 10.14 b 62.4 / 95 K5
Σ1657 Com 12 35.1 +18 23 5.11 or 6.33 b 20.3 / 96 K2III A7m
Σ1877 AB / ε Boo 14 45.0 +27 04 2.58 jj 4.51 bb 2.9 / 01 K0II-III A1
α Sco / Antares 16 29.4 -26 26 0.96 var or 5.4 bve 2.8 / 96 M1Ib B2.5V
Σ2140 / α Herculis 17 14.6 +14 23 3.48 or 5.40 ve 4.8 / 00 M5Ib-II F9
Σ43 Aa-B / β Cygni 19 30.7 +27 58 3.37 jj 4.68 ve 34.4 / 98 K3II+ B8V
H 84 Sge 19 39.4 +16 34 6.38 or 9.46 bv 28.4 / 00 K4Ib A?
HJ 599 AC / 54 Sgr 19 40.7 -16 18 5.42 jj 7.65 b 45.6 / 91 K3III F8V
Σ2727 / γ Del 20 46.7 +16 07 4.36 j 5.03 bb 9.2 / 00 K1IV F7V
Σ2877 AB Peg 22 14.3 +17 11 6.65 or 9.23 ve 21.0 / 00 K4 IV G5
Σ58 AC / δ Cep 22 29.2 +58 25 4.21 var j 6.11 b 40.9 / 94 F5Iab F7V

A quick examination of the preceding list shows that in the majority, the companions stellar spectrum could be determined. Star in isolated cases would have their corresponding colour matching their spectral class, and it is this that brings together the changes like the yellow stars, giving them sometimes remarkable appearances by their complementary colours. That this does not prevent the enthusiastic observers from turning their telescopes towards these pairs; and importantly the objectivity or the subjectivity of phenomenon. From the point of view of the artist, is this phenomenon not beautiful? However, the astronomer has the right and the duty to say that they are only by appearance.

j = jaune, yellow
ve = vert, green
b = bleu, blue
or = orange, orange
v = violette, violet
Double letters mean particularly strong colour


  • (1) I have looked at the French, Ann. DAstrophys. Nd. It isnt in this.
  • (2) Comment: The information is a little out-of-date. (Nick Loveday)
  • (3) The information here does not describe the correct mechanism but principles do work in similar ways.
  • (4) Baize expresses the units for nanometres (nm) as ×1012 Hz
  • (5) Comment: As demonstrated by Andrew James DSCE I, II and III. (Nick Loveday)
  • (6) Comment: This is quite true in my case. (Nick Loveday)


1.) "Universe" 28, 3 (1981) [Jan] and 28, 4 (1981) [Mar]


Last Update : 4th October 2012

Southern Astronomical Delights © (2012)

For any problems with this Website or Document please e-mail me.