Studies on the flower colours in Rosa
Kenichi Arisumi

Arisumi, K. 1963 Studies on the flower colours in Rosa, with special references to the biochemical and genetic analyses and to the application of those results to the practical breeding I. Sci. Bull. Fat. Agr. Kyushu Univ., 20 : 131-149 (in Japanese with English summary)


1. The anthocyanins in flowers of garden rose 'Frensham' were isolated and identified by paper chromatographic methods. The anthocyanins identified were cyanin, pelargonin, chrysanthemin, and ρ-coumaroyl cyanin which was assumed to be cyanidin 3-ρ-coumaroylglucoside 5-glucoside.

2. Chrysanthemin of ordinary garden roses appear after anthesis and increase with flower age. But chrysanthemin of 'Frensham' was unusual in synthesizing mechanism. It occured before anthesis and practically did not fluctuate with flower age.

3. Replacement of the ordinary 3:5-diglucosides with the 3-glucosides of 'Frensham' type, e.g., cyanin with chrysanthemin was considered to afford more brilliant flower colours. And also the unfading anthocyanin colours might be realized from the coupling of two enzymatic systems, one concerning with the formation of the 3-glucosides of 'Frensham' type and the other of the light induced 3-glucosides of ordinary garden roses.

4. In stock, Matthiola incana, the shift towards the bluer shade appeared to originate from the acylation of anthocyanin molecule. If this situation is true, the ρ-coumaroyl cyanin might offer the third clue to so-called blue roses.

Table 1. Fluorescence and colour reactions in anthocyanins found in garden roses and as well as in species roses.

Spot Fluorescence NH3 FeCl3 Pb-acetate Na2CO3 Identification 1)
1 blue violet violet blue blue cyanin
2 blue gray violet violet blue blue gray chrysanthemin
3 rosy orange blue dull rosy red purple blue paeonin
4 yellow violet dull brick red dull red violet pelargonin
5 grayish lavender dull salmon pink rosy red violet callistephin

1) Identification was also made by means of the co-chromatography with authentic specimen.


1. Chromatographic studies have shown that the anthocyanins of cyanic forms in genus Rosa are cyanin, chrysanthemin, pelargonin, callistephin, paeonin, and its probable 3-monoside, i.e., paeonidin-3-monoglycoside. Besides these there also appeared some other anthocyanins which have not been identified but would be duly regarded as the cyanidin derivatives from their certain chemical behaviours. Irrespective of wild or cultivated forms, cyanin was the commonest anthocyanin in roses, but the others were usually confined to some definite special forms, such as strains, varieties or to a certain species.

2. Hybrid Teas and Floribundas, both composing the major part of garden roses, have shown the remarkable difference between anthocyanin constitutions. Firstly, the capacity of synthesizing 3-monoside, which was distinctively found in their common ancestral species, R. chinensis, and which is quite different from the 3-monoside synthesis occurring in the other species, had been eliminated in the Hybrid Tea throughout their breeding history, while it remained intact within the Floribunda. Most Floribundas have a tendency to intensify the flower colour at the later stages of flowering derived from the synthesis of 3-monoside of anthocyanidin. Secondly, the pelargonin which is considered to appear first in Polyantha class, the ancestral garden strain of Floribunda, was scarcely detectable in the Hybrid Tea and the forms which contain the same anthocyanin were strictly confined to the hybrids derived from the crosses between the Hybrid Tea and the Floribunda. On the contrary, most Floribundas were pigmented by pelargonin.

3. The introduction of R. foetida into garden roses had afforded a revolutionary change in flower colours of the roses. R. foetida proper was ascertained to be latently different anthocyanin constitution from garden roses, because one of its variety, bicolor shows a quite different anthocyanin constitution. But if we compare the constitutional pattern of anthocyanin before and after the introduction of the above form, any noticeable change could not be found. On the other hand, with the distribution pattern of carotenoid the situation was quite different. A considerable amount of carotenoid appeared for the first time after the introduction. Therefore, it may be concluded that the revolutionary change in flower colour resulted from the introduction of R. foetida would be derived by adding a new enzymatic system which can produce a large amount of carotenoid and not by changing the constitutional pattern of anthocyanin pigmentation.

4. Under the co-existence of the carotenoid and cyanin the flower colour in the Hybrid Tea showed a considerable modification accompanying various orange shades. But in the Floribunda the similar modification had been induced through the co-existence of pelargonin, though there could be noticed a certain different nuance as compared with the former co-occurrence of cyanin and carotenoid.

5. A tremendous development of Floribunda roses, that has been attained in these thirty or forty years, would be attributable to their many outstanding characters which have not been realized in the Hybrid Tea, including a remarkable character derived from the participation of pelargonin. Thus it becomes to be an urgent problem to introduce the pelargonin into the Hybrid Tea in order to expand their colour range more extensively. And moreover, it has been quite impossible to find out the co-occurrence of a fair amount of pelargonin and carotenoid so far as the present investigations have gone. Because the pelargonin has a rather unique favourable shade than the cyanin, and the more brilliant colour variation was achieved through an addition of carotenoid to the ground colour associated with cyanin, it would be reasonable to expect that the co-existence of pelargonin and carotenoid might afford the most remarkable flower colour that have not been found among our cultivated roses. With the view of the above breeding project, the present author has been engaged in the present experiments since 1952, and succeeded in combining these two pigments through hybridizations between the acyanic yellow varieties of kaempferol type and those highly pigmented by pelargonin which were selected out by the detailed analysis of flavonoid constitutions. The flower colours thus realized, for the first time, were so brilliant in their shades that scarcely ever be seen in garden roses.

6. Characteristic to form the 3-monoside derived from R. chinensis is very outstanding in its effect of bringing about the various changing colours within the same flower-truss or scattered on one tree. But the examples of excellent and successful performance with this character are rather rare, even in the Floribunda which shows the most frequent occurrence of such 3-monoside among the garden roses. Therefore, it will be the urgent future project to realize the complete forms of Floribunda having this colour-change character, and, moreover, to introduce this preferable feature into the Hybrid Tea and other garden strains. And further, the ingenious introduction of callistephin formation, which has not been realized until now, will be taken as an another future subject, because the callistephin is a derivative of pelargonidin and could afford the more brilliant colour-change than with chrysanthemin.

7. Paeonin was frequently found in many rose species. This anthocyanin realizes, as a rule, much stronger and redder tone than that of cyanin. Extensive breeding efforts have hitherto been directed to eliminate the bluish tone of cyanin colouring from our old roses. If we compare the colours of old garden roses and that of "Poinsettia" which is considered to be one of the most brilliant success in scarlet roses, the above situation is quite perceivable. Therefore, the replacement of cyanin by paeonin in cultivated roses might afford much more brilliancy compared with those of cyanin colouring in the whole series of colours, ranging from pink, rose, cerise, red to dark red. Moreover, the above situation is quite same with the 3-monosides which were occasionally found in wild species roses, and differ in the synthesizing mechanism from those 3-monosides of R. chinensis and of most Floribundas. Because 3-monosides usually show much redder tone than those of their corresponding 3:5-dimonosides.

8. To make up the blue rose was our old practice and now it is yet our present problem. The careful paper chromatographic analysis, however, could not reveal the existence of any trace of delphinidin or myricetin and their derivatives. So it may be considered to be a very difficult task to breed out the true blue rose. The so-called blue roses, such as "Veilchenblau", "Belle de Crecy", and as "Cardinal de Richelieu", etc., rapidly develop violet shades after the full bloom, though any noticeable change did not accompanied in their anthocyanin constitution. There could be noticed, however, several rare instances of blue colouring under the only participation of cyanidin derivatives. Matthiola incana and Centaurea cyanus are taken as such typical examples. Modification towards blueness is said to occur under the conditions such as copigmentation, the change of pH towards alkaline side, or as the formation of pigment-metal complex. In roses it is not clear, at the present, which of those conditions or some other alternatives would participate. Another line of approach to the so-called blue roses have been substantiated by "Grey Pearl" which was introduced in 1944. The pigment participated showed complete peculiarities in their chemical behaviours, as they could not been extracted either by the hydrochloric acid or the petroleum ether. Repeated treatments with these solvents could dissolve out the ordinary anthocyanin and carotenoid from the petals of "Grey Pearl", so that the remnant has shown the beautiful bluish tinge. But it was rather lavender and far from the true blueness. From the offsprings of "Grey Pearl" we have reached to "Sterling Silver", which is considered to be the most successful performance of this colour range, but there is no discrimination between the intact petals of "Sterling Silver" and those of "Grey Pearl", from which co-existing anthocyanin and carotenoid were fully removed. Thus we may conclude that the endeavours to produce true blue rose have been directed to eliminate the contaminative pigments from "Grey Pearl", which disturb the effect of the above undefined lavender pigment. It would, no doubt, be a very hard task to get the true blue rose. In any way, it is necessary to make clear the precise nature of this peculiar pigment and also to reveal the processes through which "Veilchenblau" attains the bluer tone, in order to get a definite clue of approaching the true blue rose. And moreover, the elucidation of the latter situation is considered to be an indispensable step for attaining the much more brilliant and preferable flower colours in the future rose breeding, because there usually prevail the considerable variations in tones of flower colours in spite of the only participation of the same one anthocyanin.

9. The practical rose breeding has hitherto been carried out through the breeder's personal knowledges or his experiences. Implementation of the scientific procedures into the practical rose breeding has been considered to be a very troublesome task on account of the several reasons presented as follows: the tetraploidy prevailing throughout the garden roses; the very heterozygous genetic composition of the most garden varieties; the remarkably wide range of variation, mainly derived from the complicated participation of noteworthy asexual reproduction and as well as the multiple phylogenic origin of formation. Under those intricate circumstances it has been possible to provide a number of scientifically approved projects for future rose breeding through the systematic application of biochemical analysis.

Arisumi, K. 1964 Studies on the flower colours in Rosa, with special references to the biochemical and genetic analyses and to the application of those results to the practical breeding II. Sci. Bull. Fat. Agr. Kywhu Univ., 21: 169-184 (in Japanese with English summary)


1. Some of the Floribunda roses and R. chinensis change their flower colour remarkably with the progress of anthesis. "Masquerade" and R. chinensis var. mutabilis are the typical examples. Such colour change was revealed chromatographically to be derived from the formation of the anthocyanins which have the 3-glycosidic configuration such as chrysanthemin and callistephin. These 3-monosides were scarcely detectable in flowers at their early stages of anthesis, but after full bloom a large amount of these pigments became to be recognizable. Any one example could not be met with, where such 3-monoside appearing in the flower bud stage was followed with entirely new formation of the 3:5-dimonoside towards the full blooming stage. Therefore, with regard to the glycosilation there could be noticed the existence of a law that the formation of the 3:5-dimonoside is always preceded by the formation of the 3-monoside. Moreover, if the pre-existing 3:5-dimonoside was exclusively cyanin, the 3-monoside formed after full bloom was exclusively chrysanthemin, and when pelargonin prevailed against cyanin, callistephin predominated over chrysanthemin. With the gradual change of the 3:5-dimonoside constitution from cyanin to pelargonin, the corresponding shift from chrysanthemin to callistephin was revealed. Thus, there has been established a persistent rule that the anthocyanins having the same hydroxylating pattern for B-ring show co-occurrence in this characteristic colour change.

2. Concerning the biosynthetic pathway towards the formation of the above-mentioned 3-monoside, the following two alternative possibilities have been presented. Firstly, the 3:5-dimonoside is considered to be the direct precursor of the 3-monoside and the splitting off the glucose residue at 5 position of the former becomes to form the resultant 3-monoside. Secondly, both the 3:5-dimonoside and the 3-monoside are derived from a certain common precursor, i.e., first following the same pathway to a certain stage of the biosynthetic processes, and since then diverging into the different pathways. The former scheme will due to the case, where we ought to expect the following phenomena; (i) The disappearance of the 3:5-dimonoside in roses showing the characteristic colour change with the progress of efflorescence, (ii) The accumulation of the 3:5-dimonoside in the ordinary roses which are lacking such a characteristic colour change, because those roses are more or less deficient in an enzymatic system which catalyze the splitting off the glucose residue at 5 position of the 3:5-dimonoside. But above these were not the cases. And moreover, this colour change is strongly affected by light intensity. The flower parts which develop the 3-monoside are confined to the parts which have received the full sunshine. The chromatograms obtained from those materials were presented elsewhere in Fig. 3, and according to the former hypothesis we can not duly explain the situation where the relative concentration of the 3:5-dimonoside stayed unchanged throughout whole the materials obtained under different environmental conditions of light. On account of the above reasons and others the latter hypothesis seems to be rather preferable. But the possibility of the linear pathway, the former hypothesis, could not be entirely excluded if we suppose the intervention of a certain colourless substance and under several additional assumptions.

3. The formation of callistephin, i.e., pelargonidin-3-monoglucoside, seems to be an enzymologically noteworthy phenomenon in connection with that of the 3-monoside. The former pigment did not occur in forms of R. chinensis which has the faculty for synthesis of 3-monoside. And in the Polyantha roses which are capable of forming a large amount of pelargonidin derivatives, the anthocyanin usually occurred in this category was confined only to pelargonin, i.e., pelargonidin-3:5-diglucoside. Callistephin occurred very rarely and even if it happened the amount was so small that it could hardly determine whether it existed originally or was derived from the partial hydrolysis of pelargonin during the preparation of sample or the executing of analysis. A large amount of this pigment was confined exclusively to the Floribunda roses. Such situations would be duly interpreted as follows; the coupling of these two enzymatic systems, one concerning with the formation of the 3-monoside and the other of the pelargonidin derivatives, will operate complementally with each other and will become to open the steady route to callistephin. This would also suggest the possibility of occasional formation of the entirely new substances under the crossings within the genetically distant forms or among the species having different enzymatic systems.

4. The situation which reveals the co-occurrence of anthocyanins having the same B-ring configuration was also made clear by T. Yoshitake (unpublished) on his studies concerning the inter-relationship between anthocyanins and flavonols. Thus, in roses two different patterns of co-occurrence of flavonoid components, i.e., cyanin—chrysanthemin—quercetin and pelargonin—callistephin—kaempferol, were confirmed. And from the following experiments such situations were clearly proved not to be a fortuitous coincidence, but to be derived under a certain genetic background. In the crossing experiments between a Floribunda rose and several acyanic yellow varieties, i.e., between a form containing pelargonin and the forms having different flavonol constitution, the following facts were ascertained that the crosses with yellow roses of kaempferol type could entirely afford the offspring having pelargonin, while the cross with quercetin type produced almost exclusively the pelargonin-free individuals. With the shift of flavonol constitution from quercetin to kaempferol in the used paternal yellow varieties, the resultant progeny showed the corresponding shift, i.e., both the percentage occurrence of pelargonin and also its relative concentration vs. cyanin became progressively higher. And moreover, irrespective of the cross-combinations those individuals containing pelargonin were of rigid kaempferol type or nearly so, but the pelargonin-free ones were of quercetin or of intermediate type, respectively. These results suffice to show that the inferring the genetic behaviours of certain forms will be quite possible from their flavonoid constitutions and also that with the genetically very complex ornamentals such as garden roses the deliberate selection of suitable materials for the new breeding is easily practicable.

5. In the author's previous paper, the several breeding projects were presented concerning the future flower colours of garden roses. Among those practical projects the introduction of pelargonin and, moreover, the realization of co-existence of a large amount of pelargonin and carotenoid, were assumed to be one of the most fascinating and urgent projects. And in the present experiment such breeding object was more or less completely and rapidly attained, through the crosses between the yellow varieties of kaempferol type, whose flavonol constitutions are completely quercetin-free, and those of pelargonin type, which show the highest relative concentration of pelargonin as compaired with cyanin.