Les Iris Cultivés 1923 p. 84-92
Société National d'Horticulture de France Commision des Iris
THE RANGE AND DISTRIBUTION OF COLOR IN
R. S. STURTEVANT
I write as a close student of garden irises, not as a scientist familiar with the chemistry of color; nor even as a scientific hybridist; my inferences are drawn from careful observations of the characters of many varieties, and from records of experiments in hybridization, incidental to the production of new and worthwhile varieties for the garden.
My treatment of the subject is divided into three fairly distinct parts. First, the actual range of color as recorded from a standard color chart; second, the distribution and localization of these colors upon the various parts of the Iris flower, as suggested, both by appearance and inheritance; and third, the relative value that these facts should be given in formulating a system of classification.
Recently I compiled a full list of the colors recorded on the Data cards of between four and five hundred varieties, and in accordance with the actual comparisons with Color Standards and Nomenclature. This chart has been adopted as a standard by The American Iris Society and has proved adequate and convenient, more convenient for garden use than the Repertoire des Couleurs. It permits of the accurate definition of some 1115 colors, of which only 250 to 300 are found in Irises. Five to seven hues are given between each primary color, seven tones between white and black, and both hues and tones are also given in four gradations between pure color and grayed neutrality. I enclose a list of terms arranged according to Ridgways' Chart.
It is quickly noted, in studying this list, that two thirds of the Iris colors come between violet and red, and even less than one third among the yellows. Clear yellows are more common, comparatively, than clear violets, but furthermore deep tones of yellow are practically non-existent. The only tones of red and orange hue are of deep shades, colors found in the velvety falls of such a variety as Fro. There  are no true rose-pinks, no true blues, and even the blue-violets are rare and much neutralized. Visually, we recognize blue and pink toned Irises, but they are blue, or pink, only by comparison with other Irises. Both the yellow and violet hues become deeper as they approach red, and lighter as they tend toward green and blue respectively, and only in varieties of comparatively recent introduction do we discover these hues at the extreme limit of the range. I think that further investigation would reveal that each year the color range of garden Irises was increased, and the mere listing of the color terms suggests that new developments in coloration will be toward light tints of pink, blue, and yellow green, and toward dark shades of red. If Iris colors were simple pigments we might visualize and actually produce Iris of every hue, at least in the neutralized tones, in fact, however, certain colors are found almost without exception in certain parts of the flower, and not elsewhere, so that it is not inconceivable that certain kinds of colors are inherently «linked» with what we might call a factor of distribution or localization.
The question of real interest to the hybridizer is whether and to what extent, he may break this linkage of characters and by careful breeding recombine them. Visually, we discriminate not only between the colors themselves, but between self-colored, bicolored, and what we are accustomed to call plicata Irises. Furthermore, within the bicolored division we may distinguish what I call a «solid bicolor», such as Oriflamme, from a «velvety bicolor» like Souvenir de Mme Gaudichau, or a «veined bicolor» like Her Majesty. You will note that I now disregard mere difference in color and consider its distribution only. Distribution is less difficult of exact definition than color, for who can draw a concise line between light and dark, or red-and blue-purple?
Practically all varieties show traces of reticulation at the base of the petals; in cypriana, this is yellowed; in pallida, blue-toned; in variegata, red-toned. We have no examples of yellow venation of the blade, although often the vein color is modified by a yellow sap, or ground color, and curiously enough, we find true red only as a vein color in varieties of variegata strain. Throughout this paper, I use the word «venation» as applied to the presence of veins on the blade of the fall, and extending beyond the beard. From research through hybridization this «venation», seems to be a typical, and usually a dominant character of variegata; its presence is a proof of variegata blood. The experiments of MM. Dykes and Bliss prove conclusively that our original so-called amoena, neglecta, and to an extent our squalens groups, were all developed from pallida and variegata crosses. This, naturally, would not apply to the more  recent introductions of trojana, cypriana and mesopotamica origin, these last three, when crossed with variegata, tend to produce sterile progeny, which are rarely of variegata coloring. This last point is emphasized when we realize that but one, or two, typical variegata have been ranked higher than 80% in the Symposium recently published by The American Iris Society. Before leaving this subject of «venation», it is worth noting that a «solid bicolor» may exist unveined, in fact is usually unveined, but that a velvety quality of the blade of the fall is almost invariably due to the confluence of richly colored veins.
Broadly speaking, our modern tall bearded Irises are derived from pallida, variegata, germanica, trojana, cypriana and mesopotamica, and from the point of view of the hybridist the lines of distinction between these last four are close and their inter-relationship not yet fully understood. It would seem that variegata of germanica and trojana origin are typically lavender, the standards lighter than the falls, which very rarely show the velvety character found in so many neglecta which have their origin in variegata. The progeny of cypriana hybrids are also typically lavender, as are those of mesopotamica, but the former shows yellow tones on the haft, tones that not only soften and dull the color but that gives a sheen to the surface that reflects the light. This extra substance and finish modifies the color considerably and may be due to a factor entirely distinct from color, but it is worth noting that this fine quality is the rule among cypriana seedlings and the exception among seedlings of other origin. For example, compare Caterina and Arsace with any pallida except the variety pallida dalmatica, Princess Beatrice.
If by plicata we mean varieties the standards of which are variably feathered, edged or otherwise marked with a deeper color, we have examples, not only of original plicata like Buriensis and Mme Chereau, purple on a white ground, but also what I term variegata-plicata like Mme Chobaut and Mme Denis, where not only the ground color, but the markings and often the venation, all show a blending of tones due to an infusion of variegata blood. The plicata character is clearly a recessive, and originally was very closely allied to pallida, if not a "break" from that, stock. At least from our records, a plicata crossed with plicata has but once produced a plicata, whereas seedlings of deep-toned pallida often produce a plicata. Only by pedigree breeding through a number of generations can we hope to secure plicata in the first generation and it is proving extremely difficult to achieve the plicata markings on the standards in combination with different ground colors and different habits of growth. I have seen these markings on yellow, lavender and blended grounds, I expect to see also trojana and  cypriana-plicata in various colors, but they are still in the future. Hybridization work with Irises is not a fertile soil for scientific investigation, practically all breeders are working toward garden results rather than toward scientific proof. We are dealing with relatively small numbers of any one parentage, and with very large numbers of inherent factors and, consequently, we must draw inferences as to certain tendencies revealed by inadequate records rather than attempt to state proved facts.
As I think that inherent characters that influence the distribution of color should be considered in the formation of a classification, I shall now present the deductions that I have drawn from my familiarity with the work of Miss Grace Sturtevant and A. J. Bliss. Both these hybridists plan and record their experiments in hybridization. Miss Sturtevant is working chiefly to develop strains of yellow pallida and cypriana types of height and habit, and Mr. Bliss has done the most with plicata and his "Dominion race", but both have produced improvements in many other types and colors.
I wish that I might foresee to what extent your hybridists would agree to the following statements.
Self (standards and falls of apparently the same color, and unmarked). A dominant character in lavenders of pallida origin, a recessive in yellow and white.
Plicata (standards marked) a recessive character, the ground color variable, the markings also variable, sometimes bronze, but never yellow.
Veined bicolor (venation present on blade of falls) ground of any color, venation red or purple-toned, sometimes bronze, but never yellow.
Velvety bicolor usually, but not always, due to veins becoming confluent on the fall. (Jacquesiana and Dr. Bernice are from their appearance of similar make-up, but the first gives selfs in the first generation, the other bicolors, often veined bicolors. Hence; this quality may be a separate inherent factor).
Solid bicolor derived from many sources, colors often blended, but rarely dominantly yellow.
These distributions of color seem to be inherent, as well as apparent characters, and permit of sufficiently accurate definition to classify all possible combinations of color as found in garden Irises. The great majority of Iris varieties are either lavender selfs, or solid, or veined bicolors, the small minority (at present) white, or yellow selfs, plicatas, or velvety bicolors. Presumably, this minority represent recessive characters that, as it happens, are fairly closely linked with certain habits of growth or form of flower. For example, our yellow variegata lack height and size, qualities which  practically all breeders are attempting to get through an infusion of cypriana or trojana blood. M. Denis reports that cypriana and variegata usually produces sterile plants, and he has reported blends like Mme Durrand, rather than true yellows. Mr. J. Marion Shull has worked with trojana and variegata, but has not yet succeeded in combining the habit of one and the color of the other. Miss Sturtevant has introduced two yellows, veined with red, of cypriana blood. The color is brilliant, the venation sparce, the flower fairly large and of a modified variegata form, the height and branching distinctive. In the second generation, these produced (some 20 perhaps) all veined yellow essentially like the parents. Unfortunately, none of these set seed easily, but a third generation is on the way and theoretically at least we look for further development toward a new type of yellow.
In all this hybridization work few factors have proved invariably dominant, but we can say that two lavenders or two yellows (not yellow selfs) bring forth their like, but an amoena, or a squalens, or a plicata, may give almost anything. In one instance, in a number of cases of crossing Anne Leslie by itself (an amoena) the progeny showed the following variations in the first generation: solid and velvety amoena, selfed and veined lavenders, veined and velvety variegata, velvety blends and even a pure white.
I know of only one example of two whites producing a batch of all white seedlings, ordinarily the result is either colored or veined.
The same is true of yellow selfs. By progress breeding, Miss Sturtevant has succeeded in this to a certain extent with yellow, white, and pale pink in both selfs and plicata, and unless I am mistaken, Mr Bliss has accomplished similar results with plicata and the Dominion race. I think this line of progress breeding a necessity if we are to develop new colors and colors distributed in a novel way.
Before leaving this subject of color distribution, I wish to consider briefly the status of blends or shot shades. A large proportion of the recent introductions are in this class and not only may the colors show a preponderance of pink or yellow or blue, but they may be self-toned or bicolor veined or solid, or in some cases one color seems to be actually laid upon another. Whatever the classification adopted, their proper grouping will prove extremely difficult. Who shall say whether blue or yellow predominates in varieties like Mady CarriŹre, or Afterglow, so evenly balanced are the two tones? They are practically blended selfs. Or take Alcazar, with its rich coloring, it is a solid bicolor, but I should hesitate to specify any particular color as dominant.
One other point as to distribution I find of interest in that it suggests the development of varieties with the upper petals darker than  the lower ones, and in both variegata and plicata there is a strong tendency toward an absence of sap color at the center of the falls. Two unrelated observations but there is a fair chance of their developing a new type.
Before taking up a concrete suggestion for classification, I wish to state my opinion as to the limitation of Iris development because any system which does not look to the future cannot prove permanent. With two exceptions, yellow venation of the falls, or markings of the standards, and a velvety upper petal, I consider all other combinations of our present colors and their distribution factors perfectly possible of development. It may take years, but if our interest in Irises continues, we have but to review the advances made in the last decade to realise what the future may bring forth.
A classification, to be of value to even amateurs, must have two qualifications: first, a clearly defined and easily perceived distinction between the various classes, and second, sufficient classes and subdivisions to separate varieties into groups numbering not over thirty or forty. An even smaller number would prove a greater convenience in description and identification. Furthermore, no classification which depends upon the arbitrary placing of varieties by a small Committee can become of international importance. We must have definitions pliable enough to admit all future introductions and yet so clearly expressed that anyone who can read and observe, may classify a variety without error.
The color of the flower is our reason for growing Irises and is therefore a fit basis for a garden classification. Botanical distinctions will not be understood by the average grower; variations in height, in time of bloom, are wide under different conditions of soil or climate. Coloration at the base of the leaves has been suggested as a basis, but it also is variable and should merely be considered as an aid to identification when the plant is not in bloom. There remains color, but color also varies in response to atmospheric conditions.
Certain varieties hold their color in the hot sun, others fade; certain varieties develop splotches or veins of deeper tone, under moist conditions; in short, we can draw no sharp lines of distinction between slight variations of color. I do not mean that we should not use the terms light and dark, for example, but we should use them only as a last resort in a subdivision of very minor importance. If you accept these statements, you will agree that a good classification must make use of the distribution of color, as well as of color itself, and personally, I think that the inherent factors which govern the distribution are of first importance.
My proposed classification is based on these principles and I trust that in substance, if not in detail, it will meet with your approval  and that after full consideration of this and all other proposals, your Committee may formulate a classification that will prove acceptable to the members of the American Iris Society.
(1) Int. intermediate; DB dwarf bearded.
A PROPOSED CLASSIFICATION OF BEARDED IRISES
(with typical examples).
GROUP I. — Self-colored.
(Standards and falls of the same color or nearly so).
a) Reticulations at haft lavender: Mrs H. Darwin, Ingeborg, Int. (1).
b) » » » yellowish: La Neige, 18 in. Innocenza, 3 ft.
a) Clear: flavescens, light, 30 in. aurea, 2 ft. Ella, Int. 18 in.
b) Falls occasionally clouded: Mrs Neubronner, deep, 2 ft.
3° Lavender (violet, or purple, might be selected) (pallida);
a) As light as Celeste. Lohengrin (lilac), 3
b) Darker than Albert Victor.
c) As dark as Pare de Neuilly.
1° Blue toned Parc de Neuilly, 2 ft.
2° Red-toned: Caprice, 2 ft. Edouard Michel, 3 ft.
4° Blends or Shot Shades;
d) No color predominant.
1° Light. Mady CarriŹre, 3 ft.
2° Dark (bronze) Alcazar, 40 in.
a) Red dominant: Queen Alexandra, 2 ft.
b) Yellow dominant.
c) Blue dominant
GROUP II. — Bicolored.
(Standards and falls of different tones or colors).
1° Standards while (or nearly white) amoena.
a) Falls, veined: Edina.
b) » solid: Sybil (red), 2 ft. Rhein Nixe (purple), 3 ft.
c) » velvety: Victorine, 2 ft.
2° Standards yellow;
a) Falls veined.
b) Falls solid: Princess Victoria Luise, 30 in. Ossian, 3 ft.
1° Bordered with color of standards.
c) Falls velvety: Maori King, 18 in. Mithras, 3 ft.
1° Bordered: Marsh Marigold, 3 ft.
3° Standards lavender; falls blue toned, unless otherwise specified
a) Falls veined: Albatross, 2 ft.
b) Falls solid: Saracen, 4 ft. Shelford Chieftain, 3 ft.
c) Falls velvety: Perfection, 3 ft. Kharput (early, red) 3 ft.
1° Bordered Monsignor, 2 ft '
4° Standards blended;
a) Falls veined.
Red, yellow, or blue dominant.
b) Falls solid.
Red, yellow, or blue dominant.
c) Falls velvety.
1° Red dominant: Syphax, 2 ft. Jacquesiana, 3 ft.
2° Yellow dominant: Iris King, 2 ft. Dr. Bernice, 30 in.
3° Blue dominant. Nuée d'orage, 3 ft. Lent A. Williamson, 40 in.
4° No dominance.
GROUP III. - Plicata.
(Standards feathered or edged, style branches usually conspicuous)
1° Ground color white: or nearly white. Markings of blue or red-violet hue.
a) Standards marked only toward the base.
1° Markings light.
2° » dark: Ma Mie, 3 ft.
b) Standards heavily edged with color.
1° Blue-toned Mme Chereau 3 ft.
2° Red-toned Pocahontas, 30 in.
c) Standards completely netted all over.
1° As light as Mrs G. Reuthe.
2° As dark as Parisiana
2° Ground color yellow: Montezuma, 2 ft.
3° Ground color lavender:
4° Ground color lightly blended; 
a) Markings very light: Pancroft, 30 in. Fantasy, 30 in.
b) Markings dark.
1° No "venation" of falls apparent.
2° Falls veined.
a) Red dominant: Demi Deuil, 30 in.
b) Blue dominant: Mme Denis, 2 ft.
The word vein or venation refers only to the presence of veins on the blade of the fall, not to reticulations on the haft alone.
It is clearly evident that certain main subdivisions need greater reduction into still smaller units and these minor classes may easily be added elsewhere as new varieties are introduced. By adding height and tone after each variety listed we avoid using these unstable characters in an important place and the specification of certain varieties as light or dark tends to more clearly define our meaning.