TSO logo

Sponsor this page

For information about how you could sponsor this page, see How You Can Help


Article from Bean's Trees and Shrubs Hardy in the British Isles

Recommended citation
'Rosa' from the website Trees and Shrubs Online ( Accessed 2022-08-09.


  • Rosaceae


Made up or consisting of two or more similar parts (e.g. a compound leaf is a leaf with several leaflets).
Protruding; pushed out.
Plant originating from the cross-fertilisation of genetically distinct individuals (e.g. two species or two subspecies).
Flower-bearing part of a plant; arrangement of flowers on the floral axis.
Small dry indehiscent fruit that has a single seed (as in e.g. Polylepis).
Fused with a different part by having grown together. (Cf. connate.)
Attached singly along the axis not in pairs or whorls.
(pl. apices) Tip. apical At the apex.
Bluish or greyish waxy substance on leaves or fruits.
Female reproductive organ of a flower. Composed of ovary style and stigma. Typically several carpels are fused together in each flower (syncarpous). The number of them can be of taxonomic significance; it can often be assessed by counting the stigma branches or the chambers in the fruit.
Firm and tough but flexible; gristly.
Made up or consisting of two or more similar parts (e.g. a compound leaf is a leaf with several leaflets).
Term used here primarily to indicate the seed-bearing (female) structure of a conifer (‘conifer’ = ‘cone-producer’); otherwise known as a strobilus. A number of flowering plants produce cone-like seed-bearing structures including Betulaceae and Casuarinaceae.
Having a rounded surface.
The inner whorl of the perianth. Composed of free or united petals often showy.
In form of corymb.
(of a plant or an animal) Found in a native state only within a defined region or country.
With an unbroken margin.
Protruding; pushed out.
A group of genera more closely related to each other than to genera in other families. Names of families are identified by the suffix ‘-aceae’ (e.g. Myrtaceae) with a few traditional exceptions (e.g. Leguminosae).
Lacking hairs smooth. glabrescent Becoming hairless.
Bearing glands.
A collection of preserved plant specimens; also the building in which such specimens are housed.
Cup-shaped or tubular structure at the base of a flower (‘floral cup’) formed by enlargement of the receptacle and/or the bases of the floral parts.
Odd-pinnate; (of a compound leaf) with a central rachis and an uneven number of leaflets due to the presence of a terminal leaflet. (Cf. paripinnate.)
Flower-bearing part of a plant; arrangement of flowers on the floral axis.
Loose or open.
Leaf-like segment of a compound leaf.
Taxonomic account of a single genus or family.
Stalk of a single flower.
Leaf stalk.
Lying flat.
Enlarged end of a flower stalk that bears floral parts; (in some Podocarpaceae) fleshy structure bearing a seed formed by fusion of lowermost seed scales and peduncle.
(sect.) Subdivision of a genus.
Single segment of the calyx. Protects the flower in bud.
Lacking a stem or stalk.
(of a leaf) Unlobed or undivided.
Like a slender tapering cylinder.
With three leaflets.
(var.) Taxonomic rank (varietas) grouping variants of a species with relatively minor differentiation in a few characters but occurring as recognisable populations. Often loosely used for rare minor variants more usefully ranked as forms.


There are no active references in this article.


Article from Bean's Trees and Shrubs Hardy in the British Isles

Recommended citation
'Rosa' from the website Trees and Shrubs Online ( Accessed 2022-08-09.



the species

Botanical Characters


Select Bibliography


roses in the garden

by Graham Stuart Thomas, obe, vmh, dhm, vmm

Historical Introduction



Classification of the Garden Hybrids


References for Illustrations



In the great family of Rosaceae, which gives to gardens more beautiful hardy deciduous trees and shrubs than any other, no genus stands out with greater distinctness than the one from which it derives its name. Writing at the end of the 16th century, L’Obel and Pena hailed Rosa as ‘Orbis Deliciae’, the Darling of the World. Yet even at that time the garden roses were not wildlings but the product of long centuries of selection by man. Many were hybrids of which even now we do not know the history or even the parentage. Today the situation is different and yet essentially the same. Most of the wild roses of the world have been brought into gardens at one time or another, yet few have gained a permanent footing and fewer still have ever been much grown outside collections. At the same time the store of garden roses has been enriched by sorts of a beauty and variety undreamed of even two centuries ago. There is no need to justify the omission from this work, as from previous editions, of the great race of Hybrid Teas and Cluster-flowered roses (Floribundas), a subject so well covered in many excellent and low-priced books that to intrude on it here would be presumptuous, even if they could be regarded as shrubs in the normal acceptation of the word. At the same time there is a large and now increasing number of roses available to gardeners, some very old and others of recent raising, that are undeniably shrubs or woody climbers and would have been treated in the present edition, albeit inadequately, even if Mr Graham Thomas had not so generously offered to undertake this task himself.

The section devoted to Rosa has therefore been divided into two parts. The first is botanical in the main, and is an expansion of the treatment contained in previous editions, with the main difference that proportionately more space has been given to the species that have contributed most to the formation of modern garden roses. The second part is entirely the work of Graham Thomas and, in addition to descriptions of some three hundred cultivars, it also contains sections on the history of the rose in gardens, as well as its cultivation and propagation. The two parts are complementary and linked by cross-referencing.



At a rough estimate, something over 3,000 names of specific rank have been published in this genus, but the number of good species is comparatively small – probably no more than 150. An authoritative estimate is impossible, for no monograph on the genus has been published since Lindley’s in 1820 and needless to say that is not of much relevance today.

The genus is mainly confined to the cooler parts of the N. Hemisphere and only three or four extend south of the Tropic of Cancer. As is the case with so many other woody genera, Rosa is most richly represented in E. Asia – especially in China, where the majority of the climbing species are to be found, as well as most of the most beautiful and individual of the shrubby species. But the rhododendron-rich, very rainy parts of the Sino-Himalayan region are poor in species of Rosa (most of the Chinese roses in cultivation were introduced by Wilson from Szechwan and Hupeh). Central Asia seems to be the main home of the groups centring around R. xanthina and R. webbiana. Europe and S.W. Asia can claim as their own the large and complex section Caninae, but is not otherwise at all rich in species. The majority of the N. American species belong to the wide-ranging section Cassiorhodon or to the closely related endemic section Carolinae; apart from these it has the endemic subgenus Hesperhodos and two outliers of mainly Old World groups – R. gymnocarpa and R. setigera.

No rose attains the dimensions of a tree, though a few have attained it in cultivation, e.g., the famous Hildesheim rose, an ancient specimen of the common brier, which is said to have attained a height of 45 ft and a girth of 2 ft. But that grew against a wall of the cathedral; and the ‘tree’ of R. moschata at Isfahan in Persia, named R. arborea by Persoon, probably had some support. The majority of the species are erect or bushy shrubs, and rarely attain a height of more than 10 ft. The climbing roses all belong to the section Synstylae or to the small E. Asiatic sections Banksianae, Bracteatae, Chinenses, and Laevigatae. No species is constantly prostrate, though some may become so in certain habitats, e.g., R. wichuraiana and R. pimpinellifolia.

All species of Rosa are armed, though in a few species, notably R. blanda and R. banksiae, the armature may be very sparse. The typical rose ‘thorn’ is the hooked or curved prickle of the common brier and most climbing species. But the prickles may be erect and straight, and with these as one extreme there is a series of types ranging through the awl-shaped prickle to the needle (acicle), the flexible bristle, and finally to the erect hair, usually gland-tipped. In cross-section the prickle may be terete or more or less flattened – even cartilaginous and translucent in the pteracanthous form of R. sericea. The armature may be uniform, or a mixture of various types, and may vary from one part of the plant to another, strong growths from the base often having an armature markedly different from that of the branches. As a general rule the flower-bearing branchlets are the least armed part of the plant, and unfortunately many herbarium specimens consist only of such pieces. Even a cultivated plant, if it lacks vigorous growths, may give no idea of the true armature of the species to which it belongs.

The leaves of the roses are imparipinnate (except in subgenus Hulthemia, where they are simple). The number is rarely in excess of nine (four pairs of lateral leaflets) and only in a few species is it reduced to three, except immediately under the flowers (even R. laevigata, the most constantly trifoliolate species, sometimes has five leaflets, at least on cultivated plants). The shape of the leaflets varies even in the same species and is rarely of much use for identification. The marginal toothing may be simple or compound. Compound toothing is nearly always associated with the presence of glands on the back of each major tooth and may occur in species where the toothing is commonly simple. On the other hand, the presence of numerous aromatic glands on the undersurface of the leaflet is a constant feature of some species of the sections Pimpinellifoliae and Caninae.

Stipules are present in all species except R. persica (subgenus Hulthemia). In most groups they are adnate to the petiole for most of their length and persistent, but deciduous and largely free in the sections Banksianae, Bracteatae and Laevigatae. The actual shape of the stipules is not of much value in distinguishing related species.

The flowers are borne on laterals springing from wood of the previous season, or on extensions from existing laterals, though R. rugosa and some American species can produce flowers on the current season’s growths. The ability of garden roses such as the Hybrid Teas to produce their flowers in flushes on the summer growths is not an inheritance from any wild rose but the result of mutations or new gene combinations which cause the extension growths to go over immediately to the flowering condition instead of producing flowers in the succeeding year as would normally be the case. The main source of remontancy in modern roses is the Chinese garden race descending from R. gigantea and the spontaneous form of R. chinensis.

Solitary flowers borne early in the season on short laterals are a feature of a few species, notably the Pimpinellifoliae. At the other extreme are the Synstylae, which mostly have elaborate inflorescences opening after the longest day. But most non-climbing species bloom around midsummer and bear their flowers singly or few in a cluster. Only in a few shrubby species is the standard number of flowers in excess of five. Many-flowered inflorescences are sometimes seen in species where three or five is the norm, but these are really the result of the merging of distinct, contiguous inflorescences. Bracts (reduced leaves) are usually present in the inflorescence except in some Pimpinellifoliae, and demarcate the pedicels from the lower part of the inflorescence, whose branches belong to the vegetative part of the plant. The pedicel is usually glabrous in the sense that it rarely has a coating of hairs (except in the Bracteatae and some Synstylae). But it frequently bears stalked glands or glandular bristles, and when these are present they usually extend to the receptacle. Indeed, pedicel and receptacle are intimately related, to such an extent that the upper part of the pedicel may become fleshy in the fruiting state.

In several genera of the Rose family, e.g., Rubus and Potentilla, the carpels are numerous and inserted on a convex torus. But in Rosa, the sole member of the tribe Roseae, the carpels line the inner walls (or the base only) of a concave receptacle (hypanthium, urceolus), on which the sepals, petals, and stamens are borne. This receptacle, which usually becomes fleshy in the fruiting stage, is variable in shape even within the same species and does not have the taxonomic value given to it by some of the earlier students of the genus. In the subgenus Hesperhodos the receptacle is open at the apex, but in other members of the genus is more or less closed by the development of a disk. In R. pimpinellifolia the disk is a mere rim. In most species it partly seals the opening of the receptacle, leaving an aperture through which the styles protrude. The disk may be flat, or built up into a mound or truncated cone.

The sepals equal the petals in number and alternate with them. They may be narrowly triangular in outline, i.e., broadening from the very base, as in the section Chinenses and most members of the section Cassiorhodon (Cinnamomeae), or be slightly constricted at the base, as in the section Synstylae and many species of the Caninae. After narrowing towards the apex the sepal often expands again into a prolonged ‘tail’. The margins of the sepals may be quite entire (as those of the two inner sepals always are) or bear lateral appendages. These may be few, narrow, and inconspicuous, but are strongly and often elaborately developed in some of the Caninae and even more strikingly in the old cultivated hybrids – R. alba, R. centifolia, R. damascena, and R. richardii; also in some forms of R. gallica. This feature seems to be associated with polyploidy, though the correlation is not constant. Whether the sepals fall soon after the flowering stage or persist until the ripening of the fruit is generally agreed to be an important taxonomic character, and so too is the posture of sepals, if persistent. The Caninae are variable in both these respects.

The corolla in Rosa is on the average larger than in other members of the Rose family, a width of 2 to 3 in. being the norm, which is exceeded by some E. Asiatic climbing species such as R. gigantea and R. laevigata. Only in some species with many-flowered inflorescences is it as little as 1 in. wide. The normal number of sepals and petals is five; they are reduced to four only in one species – R. sericea – and the standard number is never exceeded in wild plants except as an abnormality. The predominant colouring is white, creamy white, blush, pink, or crimson; pure red is seen only in R. chinensis and R. moyesii. The more vivid shades of yellow in modern roses derive from R. foetida of the section Pimpinelliffoliae and all the other truly yellow-flowered species belong to this section. The soft yellow and cream shades of the old Tea roses is the contribution of R. gigantea.

The petals themselves do not provide any diagnostic character of value, except that they are silky on the outside in R. longicuspis and R. setipoda.

The very numerous stamens add much to the beauty of all single and semi-double roses. Rarely are there fewer than thirty and the median number varies from one species to another. Boulenger names five in which the minimum number is too or more: R. bracteata, R. palustris, R. roxburghii, R. rugosa, and R. setigera.

The styles (one from each carpel) protrude through the aperture, their stigmas forming usually a sessile dome-shaped or conical head. But in the Synstylae the styles are agglutinated into a slender column which is exserted well above the disk and may be as long as the stamens. Exserted styles, though to a less marked degree, are also a feature of some Caninae, of the E. Asiatic R. davidii, and often of R. gallica.

After the petals have fallen the receptacle matures into the hep (or hip) so characteristic of the genus, enlarging and usually becoming fleshy but retaining the original shape and whatever outgrowths it had. Only in the anomalous subgenera Platyrhodon, Hulthemia, and Hesperhodos is the fruit (like the receptacle) always armed, but in a few species, such as R. pimpinellifolia and the R. eglanteria group, small prickles (as distinct from stalked glands) occur on the fruits. Only in the section Bracteatae is the fruit hairy. The size ranges from almost 112 in. in width in R. gigantea and R. rugosa to barely 14 in. in some species with many-flowered inflorescences such as R. cymosa or R. multiflora. Red, crimson, or scarlet are the predominant colours, but darker and more muted shades are common in the Pimpinellifoliae, and the subgenera Hulthemia and Hesperhodos, while R. roxburghii (subgen. Platyrhodon) is unique in its greenish fruits.

The sepals are deciduous from the ripe fruit, or earlier, in the Synstylae and other climbing species, in R. gallica and its derivatives, and in the Carolinae, usually persistent in the section Cassiorhodon (Cinnamomeae), while the section Caninae varies in this respect as in so many others.

The ripe carpel is a bony achene, containing a single seed.


Subgen ROSA

For characters, see generic introduction.

sect. Chinenses. – Robust climbing shrubs in their wild state, with hooked prickles as their only armature. Stipules narrow, adnate to the petiole for most of their length, persistent. Leaflets glossy, glabrous, mostly five or seven, with more than twenty-five teeth on each margin. Flowers solitary, sometimes paired or in threes. Sepals entire or almost so, tapered from the base, deciduous in fruit. Styles free, slightly exserted. The two E. Asiatic species in this section – R. chinensis and R. gigantea – have contributed more to the formation of the modern garden roses than any other group.

sect. Banksianae. – See R. banksiae. The leading characters of this small E. Asiatic group are the tall, lax, or climbing habit, sparse, simple armature of prickles, the slender free, deciduous stipules and corymbose inflorescence. Apart from R. cymosa, mentioned under R. banksiae, the only species is R. collettii Crép., described from the Shan Hills of Burma; it is allied to R. cymosa.

sect. Laevigatae. – See R. laevigata, the only species. It resembles the Banksianae in the stipules, which are free from the petiole for most of their length, the free part deciduous. But the flowers are solitary and larger, and the leaflets three only (sometimes five).

sect. Bracteatae. – See R. bracteata, which is the only species apart from the closely related R. clinophylla. The leading characters of this very distinct group are the almost free, deeply incised stipules, the large inflorescence bracts, and the woolly receptacle.

sect. Synstylae. – Climbing, sprawling or prostrate shrubs, with a simple armature of hooked prickles. Leaflets three to nine. Stipules adnate to the petiole for most of their length, persistent, entire, toothed, or (as in R. multiflora) laciniate. Inflorescence in most species a much-branched paniculate or corymbose cluster with numerous flowers, but sometimes few-flowered, as usually in R. arvensis. Petals usually white. Sepals with a few slender lateral appendages (larger and more numerous in R. phoenicea), deciduous from the ripe fruit. Styles more or less agglutinate, forming an exserted column.

Apart from the American R. setigera all the species are natives of the Old World, mostly in E. Asia. Those treated here, arranged geographically, are:

Europe: R. arvensis, R. sempervirens.

S.W. Asia: R. moschata, R. phoenicea.

Himalaya and N.E. India: R. brunonii (also in China), R. longicuspis.

China and continental N.E. Asia: R. anemoniflora, R. cerasocarpa, R. filipes, R. gentiliana, R. helenae, R. henryi, R. lucens, R. maximowicziana, R. mulliganii, R. multiflora (varieties), R. rubus, R. soulieana.

Japan: R. luciae, R. multiflora, R. watsoniana (garden origin), R. wichuraiana.

There is an interesting extension of the Synstylae into the northern tropical zone. R. leschenaultiana (Thory) Wight & Arn. is a native of the Nilgiris and other hills of S.W. India (Madras), while R. abyssinica R. Br. inhabits the Horn of Africa and has two close relatives in the Yemen.

sect. Pimpinellifoliae. – Shrubs of diverse habit. Prickles straight or curved, sometimes flattened or even thin, cartilaginous and wing-like, often mixed with needles or bristles. Leaflets commonly more than seven (up to seventeen in R. sericea). Flowers rarely other than solitary, white or yellow. Petals normally the usual five, four in R. sericea. Sepals entire or with a few slender lateral appendages, erect and persistent in fruit. Styles free, included. A few species in the Old World.

R. ecae, R. foetida, R. hemisphaerica, R. hugonis, R. pimpinellifolia (the only European species), R. primula, R. sericea, R. xanthina.

sect. Gymnocarpae. – This small section is allied to the Pimpinellifoliae, but has the peculiarity that the top of the receptacle as well as the sepals is deciduous from the ripe fruit. The type-species, R. gymnocarpa, is W. American; the other species are Asiatic, those treated here being R. albertiana, R. beggeriana, and R. willmottiae.

sect. Cassiorhodon (Cinnamomeae). – Shrubs, often suckering. Prickles when present usually in pairs or clusters at the nodes (infrastipular); needles and bristles are often present, especially on strong growths (turions) and are the predominant form of armature in some species. Leaflets mostly five or seven (sometimes nine). Flowers in some shade of pink or red, solitary or few, sometimes fairly numerous in a corymbose cluster. Sepals entire or with a few slender lateral appendages. Styles included. Fruits crowned with the erect persistent sepals.

As understood by Crépin and those who have adopted his classification this is a large group, comprising most of the shrubby species of the subgenus Rosa apart from the ‘cryptohybrid’ Caninae. The following arrangement is mainly based on Boulenger’s classification, in which only the first group belongs to the Cinnamomeae as he understood it:

1. Arranged geographically the species in this subgroup are:

Widely distributed in the N. Hemisphere’: R. acicularis (N. America, N.E. Asia, Siberia, parts of European Russia, Sweden).

Europe: R. glauca (rubrifolia); R. majalis (cinnamomea), which ranges into Siberia.

E. Asia: Those treated here are R. rugosa, R. setipoda, and its two close allies R. caudata and R. hemsleyana.

N. America: R. arkansana and its allies, R. blanda, R. californica, R. nutkana, R. pisocarpa, R. woodsii.

2. The leading species in this subgroup is R. webbiana, to which R. bella, R. fedtschenkoana, and R. sertata are closely allied. The others treated here are R. elegantula, R. forrestiana, R. latibracteata, R. multibracteata. This group, which is near to the Pimpinellifoliae, is mainly confined to China but extends into the Himalaya and Central Asia.

3. This subgroup largely consists of the European R. pendulina and its allies – R. oxyodon (Caucasus), R. macrophylla (Himalaya and probably S.W. China), R. wardii (S.E. Tibet), and the Chinese R. banksiopsis, R. davidii, R. moyesii, R. prattii, R. saturata, R. sweginzowii.

sect. Carolinae. – A small group confined to eastern and south-eastern N. America, near to the section Cassiorhodon (Group 1) but with the sepals deciduous in the fruiting stage. In setting up this group Crépin gave as the leading character that the achenes are inserted at the bottom of the receptacle only, but later workers have found this criterion unreliable. The species concerned are R. carolina, R. foliolosa, R. nitida, R. palustris, R. virginiana.

sect. Caninae. – To this section belong all the roses of Europe with the exception of the two Synstylae (R. arvensis and R. sempervirens), R. pimpinellifolia, R. glauca (rubrifolia), R. pendulina, R. majalis (uncommon outside Scandinavia), and R. acicularis (Scandinavia only). These are comparatively uncomplicated species, and it is the Caninae, a group of the most intricate variability, that was the chief preoccupation of the 19th-century rhodologists. A vast number of species was made out of the character-combinations that occur in this group, and even when the number of recognised species was reduced to a moderate number it remained the practice to inventorise and name every variation within it. For example, in Wolley-Dod’s Revision of the British Roses (1931) five species are recognised in the subsection Caninae (R. canina aggregate), but these are subdivided into in all sixty-six varieties, many with two or more formae. In Boulenger’s treatment of the same group, in Roses d’Europe, two species are recognised with thirty-three varieties between them, but the listing of the synonyms of this comparatively limited number of categories takes up sixty-three pages.

As is now well known, the species in this group are polyploids of a peculiar sort. The majority are pentaploids (35 chromosomes – five times the base-number of 7). In the reduction-division that leads to the formation of the sex-cells only 14 of these form pairs in the normal way, the other 21 remaining unpaired. In the pollen the unpaired chromosomes are lost, so that the pollen grains carry only seven chromosomes, as in a normal diploid. The ovules, on the other hand, contain all the unpaired chromosomes plus half the pairing set – 28 in all. Thus the pollen contains only a small part of the gene-complement of the plant and its influence is weak. The bulk of the genes that determine the characters of the offspring are contained only in the ovules. Inheritance is therefore predominantly maternal, and minor variations are able to perpetuate themselves with little alteration from one generation to the next. The small number of species now recognised in this group are highly complex groups of microspecies and no attempt is made in modern floristic works to take account of every character-combination. It is likely, in any case, what with modern methods of agriculture, and afforestation, that many of the forms and hybrids that delighted the old rhodologists have long since departed with the rose-rich hedges and thickets in which they grew. Crépin’s favourite hunting-grounds in Belgium had largely been destroyed by the time the centenary of his birth was celebrated in 1930.

It is now thought that the species of Caninae are what have been termed ‘cryptohybrids’, i.e., they are the product of ancient hybridisations between species that no longer exist, at least not in their primitive state. Dr Hurst’s experimental crossings suggested that there may be a rose of the Synstylae in their ancestry, since hybrids of which a member of the Caninae was the pollen-parent showed Synstylae characters. For Boulenger’s theories, see below.

The leading characters of the Caninae are: Stems erect or arching, armed with hooked or straight prickles. Leaflets mostly five or seven. Inflorescence a few-flowered corymb. Flowers white or light pink. Outer sepals with conspicuous lateral appendages. Styles free except in the Stylosae, sometimes slightly exserted. Sepals usually persistent until the fruit ripens, sometimes falling earlier.

subsect. Caninae (Eu-caninae). Dog roses. – prickles hooked or curved. leaves glabrous or downy, not or only slightly glandular.

Species treated (see R. canina): R. afzeliana, R. canina, R. coriifolia, R. corymbifera (dumetorum), R. obtusifolia.

Boulenger considered that this group derives from R. majalis (cinnamomea) and R. acicularis. He therefore associated it with these species and their allies (see sect. Cassiorhodon Group 1) in his group Cinnamomi-Caninae.

subsect. Villosae (Vestitae). – Prickles more slender than in the subsection Caninae, straight or curved. Leaflets with glandular-compound teeth, their undersurface usually densely woolly and often glandular. Pedicels bristly, as usually are the receptacles and fruits.

Species treated: R. jundzillii, R. mollis (very closely allied to R. villosa), R. sherardii, R. tomentosa.

Boulenger associated these species with R. pendulina (alpina) and its allies (see sect. Cassiorhodon, Group 3), holding that R. pendulina was ancestral to them. This group he named Alpinae-Vestitae.

subsect. Rubiginosae. Sweet Brier Group. – Prickles hooked, sometimes mixed with needles and bristles on the flowering branches. Underside of leaflets never woolly, often glandular.

Species treated: R. biebersteinii (horrida), R. eglanteria (rubiginosa), R. micrantha, R. pulverulenta (glutinosa), R. serafinii, R. sicula.

In Boulenger’s classification these species are associated with the Pimpinellifoliae, considered to be ancestral to them, forming his group Pimpinelli-Suavifoliae (the Suavifoliae of Crépin was a group of “species” all long since reduced to synonymy, mostly under R. eglanteria).

subsect. Stylosae. – One species. See R. stylosa, mentioned under R. canina.

A species usually included in the sect. Caninae, but out of place there, is R. elymaitica (q.v.).

sect. Gallicanae (Rosa). – This section has a single species, R. gallica. Its other two components – R. damascena and R. centifolia – are of hybrid origin and not known in the wild state (the hybrid R. alba, sometimes placed in the Gallicanae, is really nearer to the Caninae).

The leading characters of R. gallica are the suckering habit, the mixed armature of needles and bristles, the reduced number of leaflets (three or five), and the reflexed, deciduous sepals.

R. centifolia has been proposed as the type-species of the genus Rosa. No worse choice could have been made, since this rose is certainly of hybrid origin but, if it is accepted, the proper title of the sect. Gallicanae would be sect. Rosa, according to the rule that a subdivision of a genus takes the name of the genus if it contains its type-species.


Bark peeling. Receptacle with a torus (fleshy emergence) at the base, on which the ovules are inserted (these in Rosa are normally inserted at the bottom and on the walls of the receptacle). Only one species, R. roxburghii. This is sometimes included in the subgenus Rosa as the monotypic section Microphyllae.


Receptacle without a disk. Fruits globose, prickly. According to some authorities the ovules are inserted on a torus as in the subgen. Platyrhodon, but others have found no trace of it. Two or three species in western N. America. See further under R. stellata.


Leaves simple, without stipules. Flowers yellow with a maroon centre. Fruits globose, prickly. One or two species in S.W. and Central Asia. See further under R. persica.


Andrews, H. C. – Roses. London, 1805-28.

For dating of plates see Journ. Arn. Arb., Vol. 18 (1937), or Stafleu, Taxonomic Literature (1967), pp. 7-8.

Boitard, M. – Manuel Complet de l’Amateur de Roses. Paris, 1836.

Boulenger, G. A. – ‘Les Roses d’Europe de l’Herbier Crépin’ [Vol. I] in Bull. Jard. Bot. Bruxelles, Vol. 10, pp. 3-417 (1924-5); [Vol. II] in op. cit., Vol. 12, pp. 3-542 (1931-2).

A study of the European species and natural hybrids, as represented in Belgium, Britain, France, Germany, Holland, and Switzerland. The hybrids are treated in Vol. II of this work (Vol. 12 of the Bulletin), pp. 257-506. A synoptic key to the species is provided in the same volume, pp. 507-11.

– – ‘Revision des Roses d’Asie …’, Bull. Jard. Bot. Bruxelles, Vol. 9, pp. 203-79 (1933) [Synstylae]; op. cit., Vol. 14, pp. 274-8 [Synstylae, supplement]; op. cit., Vol. 13, pp. 165-266 (1935) and Vol. 14, pp. 115-221 (1936) [Other species].

– – ‘Introduction à l’étude du Genre Rosa’, Bull. Jard. Bot. Bruxelles, Vol. 14, pp. 241-73 (1937).

Bunyard, E. A. – Old Garden Roses. London, 1936.

Byhouwer, J. T. P. – ‘An Enumeration of the Roses of Yunnan’, Journ. Arn. Arb., Vol. 10 (1929), pp. 84-107.

Clapham, A. R., Tutin, T. G., and Warburg, E. F.Flora of the British Isles, Rosa, ed. 1 (1952), pp. 516-27; ed. 2 (1962), pp. 405-13.

Crépin, F. – Primitiae Monographiae Rosarum.

A series of papers on the taxonomy of Rosa, published in Bulletin de la Société Royale de Botanique de Belgique, 1869-82, later published separately and reprinted in Plant Monograph Reprints, ed. J. Cramer and H. K. Swann, 1972.

– – ‘Sketch of a New Classification of Roses’, Journ. Roy. Hort. Soc., Vol. 11 (1889), pp. 217-28.

– – ‘Nouvellc Classification des Roses’, Journal des Roses, Nos 3-5, 1891.

– – ‘La Question de la Priorité des noms spécifiques envisagés au point de vue du Genre Rosa’, Bull. Herb. Boiss., 3rd series, Vol. 5 (1897), pp. 129-63.

De Pronville, A. – Nomenclature Raisonée des Espèces … du Genre Rosa Paris, 1818.

De Pronville’s other work, Sommaire d’une Monographie du genre Rosier (1822), is based on Lindley’s monograph.

Duhamel du Monceau, H. L. – Traite des Arbres et Arbustes …, Ed. 2 (‘Nouveau Duhamel’), Vol. VII (ed. Loiseleur-Deslongchamps), pp. 17-60.

Dumont de Courset, G. L. M. – Le Botaniste Cultivateur, Ed. 2, Vol. 5 (1811), pp. 465-89.

Falconer, R. W. – ‘The Ancient History of the Rose’. A lecture read before the Botanical Society of Edinburgh in 1838 and printed in Loudon’s Gardener’s Magazine, Vol. 15 (1839), pp. 379-89.

Ferguson, Jessie. – ‘A Botanical Study of Rose Stocks’, Journ. Roy. Hort. Soc., Vol. 58 (1933), pp. 344-7.

Fernald, M. L. – Gray’s Manual of Botany, Ed. 8 (1950), pp. 868-74.

Erlanson, Eileen W. – ‘Cytological Conditions and Evidences for Hybridity in North American Wild Roses’, Botanical Gazette, Vol. 87 (1929), pp. 443-506.

– – ‘Experimental data for a Revision of the North American Roses’, Botanical Gazette, Vol. 96 (1934), pp. 197-259.

Gleason, H. A. – Illustrated Flora of the Northeastern United States …, Vol. 2 (1952). pp. 322-7.

Herrmann, J. – Dissertatio Inauguralis Botanico-Medica de Rosa. Strasbourg, 1762.

Hitchcock, C. L., et al. – Vascular Plants of the Pacific Northwest, Part 3 (1961), pp. 164-71.

Hurst, C. C., and Breeze, Mabel. – ‘Notes on the Origin of the Moss-Rose’, Journ. Roy. Hort. Soc., Vol. 47 (1922), pp. 26-42. reprinted in G. S. Thomas, The Old Shrub Roses, pp. 57-94.

For a discussion of Dr Hurst’s account, see the series of articles by C. Hartman Payne in Gard. Chron., Vol. 72 (1922), indexed in that periodical under ‘Rose, Moss’.

Hurst, C. C. – ‘Origin and Evolution of Garden Roses’, Journ. Roy. Hort. Soc., Vol. 66 (1941), pp. 73-82, 242-50, 282-9. Reprinted in G. S. Thomas, The Old Shrub Roses, pp. 57-94.

Ibn al-‘awwam. – Kitab al-filahah, translated as ‘Le Livre d’Agriculture’ by J. J. Clement-Mullet. Paris, 1864 (Roses in Vol. I, Chap. VI, Art. 28.61).

Joret, C. – La Rose dans l’Antiquité et au Moyen Age. Paris, 1892.

Juzepchuk, S. V. – Rosa/i>, in Fl. URSS (Flora of the Soviet Union), vol. 10 (1941). pp. 431-506.

Klastersky, I. – Rosa, in Fl. Europaea, Vol. 2 (1968), pp. 25-32.

Lindley, J. – Rosarum Monographia. London, 1820.

Lawrance, Mary. – A Collection of Roses from Nature. London, 1799.

Mandenova, I. P. – ‘A Revision of Rosa in Turkey’, Notes Roy. Bot. Gard. Edin., Vol. 30 (1970), pp. 327-40.

Melville, R. – Rosa, in C. A. Stace, ed., Hybridisation and the Flora of the British Isles (1975), pp. 212-27.

National Rose Society. – See Royal National Rose Society.

Nilsson, O. – Rosa, in P. H. Davis, ed., Flora of Turkey, Vol. 4 (1972), pp. 106-28.

Paul, William. – The Rose Garden. London, 1848 and later editions.

Prevost fils. – Catalogue Descriptif … des Espèces … du Genre Rosier. Rouen, 1829.

Redouté, P. J., and Thory, C. A. – Les Roses. Paris, 1817-24 (Vol. I, 1817-19; Vol. II, 1819-21; Vol. III, 1821-4).

The plates are unnumbered; references are to the page facing the plate. For further details see Stafleu, Taxonomic Literature (1967), pp. 376-7 and 379. The third edition (1828-30) and the similar fourth edition of 1835 contain additional plates and text. A complete facsimile folio reprint of the original edition was published by De Schutter 1974-1977. An additional volume with commentaries and unpublished plates appeared in 1978.

Regel, E. – ‘Tentamen Rosarum Monographiae’, Act. Hort. Petrop., Vol. 5 (1878), pp. 287-398.

Rehder, A. – Rosa, in Pl. Wilsonianae, Vol. II (1915), pp. 304-44.

Rivers, T., jun. – The Rose Amateur’s Guide … London, 1837 and later editions.

Roessig, C. G. – Die Rosen nach der Natur … (with a French translation by de Lahitte). Leipzig, 1802-20.

– – Oekonomisch-botanische Beschreibung … der Arten der Rosen. Leipzig, Part I, 1799, Part II, 1803.

Royal National Rose Society. – The Rose Annual, 1907-.

Seringe, N. C. – Rosa, in De Candolle, Prodromus, Vol. II (1825), pp. 597-625.

–– ‘Critiques des Roses déséchées de N. C. Seringe’, Mélanges Botaniques, Vol. I (1819), pp. 3-63.

Täckholm, G. – ‘Zytologische Studien über die Gattung Rosa’, Act. Hort. Berg., Vol. 7 (1922), pp. 97-381.

Thomas, G. S. – The Old Shrub Roses. 1955 and 1980.

–– Shrub Roses of Today. 1962. Revised edition 1980.

–– Climbing Roses Old and New. 1965. Revised edition 1978.

Thory, C. A.Prodrome de la Monographie des Espèces … du Genre Rosier. Paris, 1820.

Trattinick, L. – Rosacearum Monographia. 4 vols. Vienna, 1823-4.

Wylie, Ann P. – ‘The History of Garden Roses’ (Masters Memorial Lecture, 1954), Journ. Roy. Hort. Soc., Vol. 79 (1954), pp. 555-71, Vol. 80 (1955), pp. 8-24, 77-87.

Willmott, Ellen. – The Genus Rosa. London, 1910-14. 2 vols (Vol. I, 1910-11, Vol. II, 1911-14).

The technical descriptions and the introductory chapter were contributed by J. G. Baker, formerly Keeper of the Kew Herbarium (cf. R. A. Rolfe, Gard. Chron., Vol. 62 (1920), p. 124).

Wolley-Dod, A. H. – ‘A Revision of the British Roses’, Journal of Botany, Supplement, 1930-1 (also printed separately).

From the Supplement (Vol.V)

Some recent works dealing at least partly with shrub roses are:

Griffiths, Trevor.The Book of Old Roses. London, 1984.

Harkness, Jack.Roses. London, 1978.

Paterson, Allen.The History of the Rose. London, 1983.