The fragrance of Old Roses by Robert Calkin

Part of our fascination with roses stems from their extraordinary variety. No other group of plants in horticulture seems capable of producing such an array of wonderful colours and forms. Yet what is sometimes overlooked is that this applies equally to the diversity of their fragrance.

Famous old varieties such as ‘Lady Hillingdon’, ‘Mme Isaac Pereire’, ‘Desprez à Fleur Jaune’, ‘Sophie’s Perpetual’, ‘Belle de Crécy’, ‘Mme Alfred Carrière’, ‘Ispahan’, ‘Seagull’, and ‘Splendens’ each have fragrances which are not only entirely different, but the quality of their fragrance has given each rose an enduring place in cultivation quite apart from its visual beauty. Having spent my working life as a perfumer I never cease to wonder at the extraordinary beauty of fragrance to be found in roses such as these.

The chemical composition of the rose fragrance, in all its variety, is immensely complicated. Analysis of the fragrant essential oil extracted from the flowers of R. damascena ‘Kazanlik’ – the rose most widely grown for the production of rose otto used by the perfumery industry – has so far disclosed nearly 400 identified constituents.

‘Head space’ analysis can now capture and analyse the fragrance as it leaves the flower, giving a truer representation of the fragrance as we actually smell it. Hundreds of roses have been studied in this way, and although many details have yet to be published, it is clear that the total number of ingredients found across the whole range of species and hybrids is considerably greater than the number found in the ‘Kazanlik’ rose.

If we examine the composition of this rose in more detail we find that approximately 85% by volume of the fragrant oil is made up of only four materials, another ten represent approximately 10%, with the remaining several hundred constituents in the final 5% of the oil. This type of composition is found not only in the scent of roses but in many other flowers such as jasmine, narcissus, and lavender.

In R. gallica and the groups descended from it, the Damasks, Centifolias and Albas, the major components are the so-called rose alcohols: phenylethyl alcohol, citronellol, geraniol and nerol, which occur in different proportions from one rose to another.

Phenylethyl alcohol has a soft petal-like character typical of the lighter coloured Gallicas and is the main ingredient of commercially produced rose water. Citronellol has a wonderfully warm and vibrant character perhaps best smelled in some of the Rugosa hybrids such as ‘Blanc Double de Coubert’. Geraniol is similar but with a somewhat sharper character reminiscent of geranium leaves, while nerol is the harshest of these and fresher.

All these materials, which are widely used by the perfumery industry, are essentially rosy in character and together form the basis of the typical ‘Old Rose’ fragrance of the European roses and Rugosas. On their own, however, they would make a poor fragrance, for it is the hundreds of other materials, many of which are quite unrose-like in character and some of which are intensely strong, that provide the individuality, depth and carrying power.

Why this type of composition, which evolved for the ‘delight’ of insects, should appeal to our own sense of the beautiful is not understood, but it is equally important in the creation of fine perfumes.

The fragrance of R. gigantea, the ancestral species from which the Tea-scented roses were derived, has a similar composition with two types of material making up the heart of the fragrance. One of these, dimethoxy toluene, which has a slightly tarry and humid character (as in a greenhouse), represents 50% of the fragrance and is unique to this rose and its descendants. The other, dihydro-beta-ionol, which represents about 10%, has an earthy, violet character. This material is also found in smaller amounts in R. chinensis, another presumed parent of the Tea roses and a grandparent of the Tea Noisettes.

This and related materials, the ionones, can best be smelled in the wonderful Hybrid Musk rose ‘Buff Beauty’, and in the Banksian roses. (From the olfactory evidence I suspect that the Banksians, with their intense violet character, may also have been involved somewhere along the line in the Tea rose ancestry.)

Another of these materials, dihydro-beta-ionone, is partly responsible for the raspberry character of many modern hybrids such as the famous ‘Queen Elizabeth’ rose, the fragrance of which is closely related in composition to that of the Tea-scented roses, with almost no European influence.

Some of the most remarkable fragrances have resulted from crosses between two species or near species. For example, the Autumn Damask, ‘Quatre Saison’, to my nose one of the most beautifully scented of roses (if sunshine had a smell this would be it!) is a wonderful blend of the translucent old rose quality of R. gallica with the exotic spicy character of R. moschata.

Other examples include the exquisitely simple fragrance of ‘Stanwell Perpetual’ (R. pimpinellifolia Autumn Damask), a scent which perfectly matches the colour and form of its flowers, and ‘Agnes’, the result of the somewhat unlikely cross between R. rugosa and R. foetida, with its intense verbena character. In these examples the resultant fragrance can be seen at least in part as a blending of those of the two parents with a certain amount of rebalancing.

Occasionally, however, the combined chemistry of the two parents will throw up something entirely new. This occurred in two roses, ‘Ruga’ and ‘Splendens’, resulting from crosses between R. arvensis and a variety of R. chinensis, probably ‘Old Blush’. Both ‘Ruga’ and ‘Splendens’ have a fresh anise character due to the dominance of a single material 4-vinyl anisole. This material has reappeared in many of the English Roses of David Austin, beginning with ‘Constance Spry’ and reaching new levels of intensity in such varieties as the beautiful ‘Scepter’d Isle’ and ‘St Cecilia’.

(Such roses are usually described as myrrh-scented, and although this has now become an established convention the smell has little to do with that of real myrrh. Something approaching a true myrrh character can be found as part of the fragrance of the Hybrid Musk ‘Penelope’.)

One of the most important marriages in the genealogy of roses was that between ‘Quatre Saisons’ and ‘Old Blush’. This took place in the Reunion, or Bourbon Isles, which subsequently became one of the main centres for the production of raw materials for the French perfumery industry. Not only did this introduce the gene for continual flowering into the European rose (as opposed to the single remontancy of the Autumn Damask), but it was also responsible for the typical Bourbon fragrance, combining the brilliance of the Damask scent with a rich fruity character.

Not all the early crosses were olfactorily ‘successful’. In some, involving members of the Synstylae in which the fragrance is largely confined to the stamens, their double-flowered descendants (with few stamens) have little or no fragrance.

The lack of fragrance in some modern Floribundas, descended from R. multiflora , is probably a result of this effect. A similar failure occurred in many of the R. arvensis seedlings, although this, the more correctly named ‘musk rose’ of Shakespeare, has one of the most pungent and diffusive of fragrances. The ‘Musk Rose’ R. moschata, another member of the Synstylae, has a delicious clove-like scent.

All these roses have the wonderful ability to fill the air with fragrance across a garden. The word `musk’ comes from the Himalayan Musk Deer whose scent, known in Europe since the Middle Ages and widely used in classical perfumery, is similarly diffusive. Sadly this extraordinary character carried in the stamens of many of the old roses, such as the marvellous Hybrid Musks descended from ‘Trier’, is absent in most modern roses.

‘Blush Noisette’ is another example of a less than immediately happy coupling – to my nose at least. (Although I have not smelled ‘Champneys’ Pink Cluster’ of which it was a seedling, the parents of this rose were R. moschata and R. chinensis ‘Old Blush’). ‘Blush Noisette’s powerful fragrance combines an intense green character, which can be described as smelling of cut grass and banana skins, due to the presence of hexenols (mainly cis-3- hexenol and cis-3-hexenyl acetate) inherited from R. moschata and R. chinensis, with an over sweet, somewhat marshmallow-like character coming from phenylethyl acetate, which makes a minor contribution to the fragrance of R. moschata.

Although ‘Blush Noisette’ is frequently described as clove-scented, this character, which also comes from its moschata parent, only appears spasmodically, just as the stamens reach maturity. However, whatever its own shortcomings, this remarkable rose was the parent of the incomparably perfumed Tea Noisettes, which inherited most of their fragrance from their Tea-scented ancestor ‘Park’s Yellow China’, as well as ‘Deprez à Fleur Jaune’ with its amazing apricot and jasmine fragrance which probably owes more to its chinensis grandparent.

Apart from the Synstylae many other roses produce a fragrance in their stamens, which is often quite different to that of the petals. One of the most remarkable is the widely grown ‘Roseraie de l’Haÿ’ in which the typically sumptous Rugosa fragrance of the petals contrasts with the fresh cucumber-like smell of the stamens. Incidentally, this combination of smells is used also by perfumers. In R. gallica officinalis , as the flowers mature, the pure Old Rose fragrance produced by the petals gives way to a light musk note, coming from the stamens.

A summary of the main types of fragrance found in old roses is given in the accompanying table [readers need to consult the original journal to see the table – Ed]. This is backed up, where available, by the results of head space analysis.

Inevitably, as in any such classification, there are roses which don’t fit comfortably into any of these categories, and this is particularly true of the old China roses. After well over a thousand years of hybridization both in the wild and in the ancient gardens of China, China roses are a fascinating and mysterious subject in their own right.

It is frequently said that the fragrance of most modern roses, although sometimes strong, lacks the exquisite beauty and depth of their early ancestors. No doubt many of the poorer old varieties have long since disappeared from our gardens, so the comparison may not be strictly fair. Nevertheless, many of the old roses, being closer to the original species, retained the successful combinations and balances of components which were ‘discovered’ by nature over millions of years of evolution.

Repeated hybridization has produced increasingly random mixtures of sometimes incompatible fragrances, which, like a badly made perfume, lacked either a well defined character or ‘lift’. As with colour in painting the excessive mixing of pigments can lead to a general effect of ‘muddiness’.

The Hybrid Perpetuals were the product of a frenzied period of hybridization between European and Chinese roses and now in some Hybrid Perpetuals, but by no means all, the quality of their fragrance seems to show a deterioration, though this is not necessarily true of the strength. Varieties such as ‘Hugh Dixon’ and ‘Ulrich Brunner’ have, to my nose at least, a hardness to the quality of their fragrance which fails to draw one in like the wonderful ‘Reine des Violettes’ and ‘Georg Arends’, which are closer in style to their Bourbon and Damask ancestors.

I suspect that the introduction into the gene pool of ‘Slater’s Crimson China’ with its extraordinary beer-like fragrance may also have had a detrimental influence. I have only once smelled this rose, in the conservatory at the home of Maurice and Rosemary Foster, so my assessment is open to confirmation. Perhaps it was having a bad day !

But it would be wrong to write off the fragrance of modern roses such as the Hybrid Teas simply as being inferior to that of their ancestors. Many, too many to give individual samples, have remarkable and beautiful scents which make a wonderful contribution to our gardens.

In the development of David Austin’s English Roses, by introducing some of the species and early hybrids into the breeding programme, not only have some of the great fragrances of the past been recaptured, but some exciting new ‘directions’ have been introduced. For instance, the soft almondy-lilac character of ‘Cottage Rose’, the wild strawberry note in ‘The Countryman’, the delicious white wine character of ‘Sharifa Asma’, and the intriguing blend of the ‘tea’ and ‘myrrh’ scents in ‘The Pilgrim’.

Although purists may object to some of these innovations it would be a dull world if all new roses smelled the same as their ancestors, however beautiful these may have been. One of the great attractions of the rose lies in its ‘infinite variety’.


1. Joichi, A. et al. Shiseido Research, Japan.
2. Schmaus, G. Dragoco Research, Holzminden, Germany (unpublished).
3. Yomogida, K. Shiseido Research, Japan.
4. Brunke, E-j. et al. Dragoco Research, Holzminden, Germany.

Robert Calkin has spent 40 years in the perfumery industry and is the co-author of Perfumery, Practice and Principles. Recently retired he has been studying the fragrance of roses.

This article appeared in the Spring 1999 issue.