F.C. Stern A Study of the Genus Paeonia

chapter I



A .Stems woody, perennial, petals thin, much longer than sepals.

section MOUTAN

B. Disc produced as a thin leathery sheath which at first completely envelops the carpels; lower leaves bipinnate with some hairs on lower surface ; pinnae entire or more often palmately 2-3 lobed.

subsection vaginatae *

B. Disc produced as conspicuous fleshy lobes around the base of carpels ; lower leaves biternate, glabrous ; leaflets pinnatipartite, with pinnae deeply lobed or toothed.

subsection delavayanae f

AA . Stems herbaceous, annual, dying down at end of each growing season.


C.Petals coriaceous, not larger or only slightly larger than sepals; disc produced as fleshy lobes at the base of carpels.

section ONAEPIA

CC. Petals thin, much larger than sepals ; disc not produced as conspicuous outgrowths.

section PAEON

D. Lower leaves cut to axis or nearly to axis into 9-16 (up to 23 in P.Broteri and sometimes up to 21 in P.arietina var. orientalis and even more in P.rhodia) distinct leaflets which are entire.

subsection foliolatae ^

DD. Lower leaves much dissected, cut to axis into many divisions (usually 25 or more), which are often lobed or toothed, or both.

subsection dissectifoliae §

* Sect. Moutan DC. Prodr. i, 65 (1824), subsect. Vaginatae F. C. Stern, subsect. nov., disco tenuiter coriaceo carpellis primum invaginante insignis. Typus—P.suffruticosa Andrews.

f Sect. Moutan DC. subsect. Delavayanae F. C. Stern, subsect. nov., disco ad basim carpellorum in lobis carnosis conspicuis producto distincta. Typus—P.Delavayi Franchet.

^ Sect. Paeon DC. Prodr. I, 65 (1824), subsect. Foliolatae F. C. Stern, subsect. nov., foliis infimis biternatis vel foliolis bifurcatis ad 16 (ad pluribus in P.Broteri et P.rhodia Steam) integris recognoscenda. Typus—P.mascula Miller.

§ Sect. Paeon DC. subsect. Dissectifoliae F. C. Stern, subsect. nov., foliis infimis biternatis sed foliolis in segmentis plerumque 25 vel ultra saepe lobatis et dentatis dissectis diagnoscenda. Typus—P.officinalis L. emend. Willd. [end page 01]


Sect. moutan DC. Prodr. 1, 65 (1824) ; Baillon, Adansonia, 4,56 (1863) ; Prantl in Engler & Prantl, Nat. Pflanzenfam.3, ii, 55 (i888).

Syn. Sect. Suffruticosae Salm-Dyck, Hort. Dyck, 365, 366 (1834).
Sect. Moutania Reichenbach, Report. Herb. s. Norn. Gen. PI. 191 (1841).
Sect. Palaearcticae § Fruticosae Huth in Engler, Bat. Jahrb. 14, 272 (1891).
Sect. Mutan Ascherson & Graebner, Syn. Mitteleur. Fl. 5, ii, 558 (1923).

Subgen. Moutan (DC.) Seringe, Fl. Jard. 3, 187 (1849) ; Baker in Gard. Chron. 21, 779(1884) ; Lynch in Journ. Roy. Hort. Soc; London, N. Ser. 12, 431, 432 (1890).

Moutan (DC.) Reichb. in Moessler, Handb. Gewächsk. Ed. s, 1, P.IX (1827).

Sect. onaepia Lindley in Bot. Reg. 25, sub. t. 30 (1839) ; Reichenbach, Repert. Herb. s. Norn. Gen. PI. 191 (1841),

Syn. Sect. Nearcticae Huth in Engler, Bot. Jahrb. 14, 273 (1891).

Subgen. Onaepia (Lindl.) Lynch in Journ. Roy. Hort. Soc., London, N. Ser. 12, 431, 433 (1890).

Sect. paeon DC. Prodr. 1, 65 (1824) ; Prantl in Engler & Prantl, Nat. Pflanzenfam. 3, ii, 55 (i888) excl. P.Brownii; Ascherson & Graebner, Syn. Mitteleur. Fl. 5, ii, 558 (1923).

Syn. Sect. Albiflorae Salm-Dyck, Hort. Dyck, 365, 366 (1834).

Sect. Compactae Salm-Dyck, I.e.

Sect. Corallinae Salm-Dyck, I.e.

Sect, Laciniatae Salm-Dyck, I.e.

Sect. Lobatae Salm-Dyck, I.e.

Sect. Macrocarpae Salm-Dyck, I.e.

Sect. Microcarpae Salm-Dyck, I.e.

Sect. Paeonia Reichenbach, Repert. Herb. s. Norn. Gen. PI. 191 (1841) ; Lemaire in Fl. Serres, 4, sub. t. 308 (1848).

Sect. Eupaeonia Baillon, Adansonia, 4, 56 (1863). Sect. Tripaeonia Baillon, I.e.

Sect. Palaearcticae § Herbaceae Huth in Engler, Bot. Jahrb. 14, 265 (1891).

Subgen. Paeonia proper Baker in Gard. Chron. N. Ser. 21, 779 (1884) excl. P.Brownii

Subgen. Paeon (DC.) Seringe, Fl. Jard. 3, 193 (1849) excl. P.Brownii et P.californica; Lynch in Journ. Roy. Hort. Soc., London N. Ser. 12, 431, 434 (1890).

Ser. Chinenses Komarov in Komarov, Fl. U.R.S.S. 7, 25 (1937). Ser. Corallinae (Salm-Dyck) Komarov, I.e. 28.

Ser. Dentatae Komarov, I.e. 33. Ser. Fissae Komarov, I.e. 34.

Ser. Obovatae Komarov, I.e. 26.


the species of the genus Paeonia are composed of woody and herbaceous perennial plants. The species are divided into three sections :—

section MOUTAN

Stems woody, perennial ; petals much longer than the sepals, disc produced as fleshy lobes around the base of the carpels or as a sheath which envelops the carpels

section ONAEPIA

Stems herbaceous ; petals not longer than the sepals, disc produced as fleshy lobes at the base of the carpels.

section PAEON

Stems herbaceous ; petals much longer than the sepals, disc not produced as conspicuous outgrowths.

The paeonies of the Section MOUTAN are located in eastern Asia ; these have been placed in two subsections, vaginatae and delavayanae, which are separated mainly by the disc. In the subsection vaginatae this disc entirely envelops the carpels with a leathery sheath, while in the subsection delavayanae the sheath is absent but the disc is produced as conspicuous fleshy lobes around the base of the carpels.

The subsection vaginatae consists of one species and a variety distinguished by the difference in the shape of the lobes of the segments of the leaflets.

The subsection delavayanae contains three species and a variety ; P.Delavayi is distinguished from the other species by the conspicuous involucre with the greater number of bracts and sepals ; P.lutea and P.Potanini are separated by the larger segments and lobes of the leaves and larger flowers in P.lutea. The leaves of P.Potanini vary considerably in the breadth of the segments and in the length of the leaves in different specimens, some of which may be natural hybrids between the species of this subsection. Stebbins101 (1939) considers all the species in this subsection to be varieties of P.Delavayi as they hybridise freely together. No doubt they are closely allied and as they are all diploids, they would naturally hybridise with each other. [end page 08]

P.Delavayi is easily separated from the other species. P.lutea and P.Potanini are, in the majority of specimens, easily separable and only in a few specimens do they come close to each other. As garden plants they can be readily distinguished and recognised.

The section onaepia contains herbaceous species confined to western North America, the only paeony species found on that continent. These species are unique among the herbaceous paeonies in having the disc produced as fleshy lobes at the base of the carpels.

The section paeon is composed of all the herbaceous species found in Europe, North Africa and Asia spreading from Spain to Japan. These species have been divided into two subsections, foliolatae, those with distinct entire leaflets, and dissectifoliae, in which the leaflets are much dissected and cut into many divisions. These herbaceous paeonies have been divided into species and their varieties by the shape and number of leaflets or leaf-segments on a lower leaf together with other outstanding taxonomic characters. Where two paeonies differ by only one character such as whether the leaves or carpels are hairy or glabrous, one of the paeonies has been named a variety or subspecies of the other. Again, when some character occurs which may be only a geographical variation, such as a form with very large leaflets, this form has been given a varietal name. The species with the oldest name is constituted the species and the other is called a variety of that species. The term ' variety' and not' subspecies' has been used in order to indicate the close affinity of the two plants to each other, and because subspecific names tend to become in practice and in horticulture the specific names.


Carpels.—Among herbaceous paeonies there are many occurrences of ' pairs' of species with little or no distinction between them other than the presence or absence of hairs on the carpels. Stapf (1931) discussed this character and gave four instances of such pairs, the members of which are found in the same region and differ only in the presence or absence of tomentum on the carpels. So many similar cases have now been met with that it would not be surprising to find this feature common to all herbaceous paeonies. The members of a pair of paeonies are always found in the same geographical area, sometimes in the same districts, but more often in separate districts.

Dr. Stapf considered these forms to be highly stable since no intermediate states occur even in cultivation, nor is there any record of gradual change from one form to another; throughout the genus there is a sharp distinction between species with glabrous and tomentose carpels, though parallel forms like those enumerated are not universal. Therefore he argued these types should be accepted as specifically distinct.

This argument does not appear to be a very strong one, for as a matter of fact this phenomenon of parallel forms does certainly occur in eight pairs of herbaceous paeonies, each pair found near each other in one geographical area ; it has also been reported in three other species— P.obovata and P.Veitchii (Handel-Mazzetti" (1939)) and P.arietina (Thiebaut (1936)).

The presence or absence of hairs on the carpels where there is no other distinction between two paeonies does not appear to me to entitle each paeony to specific rank but only to justify [end page 09] classification as a variety. In the following list the species and varieties arranged in pairs are differentiated only by the presence or absence of hairs on the carpels :—

carpels glabrous.

carpels hairy.


P.Wittmanniana var. nudicarpa

P.Wittmanniana ..

silky villose.

P.Russi var. Reverchoni




P.lactiflora var. trichocarpa


P.emodi var. glabrata



P.humilis(slightly pilose)

P.humilis var. villosa



P.anomala var. intermedia

densely tomentose.


P.Mairei var. ..

densely covered with short.


P.oxypetala var...

golden hairs

It is a curious fact that in nearly every case where the carpels are glabrous the underside of the leaves of that species is pubescent.

The shape of the carpels is the same in nearly all the different species with the exception of the Obovata and Mairei groups where the carpels are conical, attenuated to the stigma.

Some botanists have considered the shape assumed by the matured follicles to be important, but from experience this would seem to be an uncertain character. The follicles are larger and more numerous in cultivated specimens ; a further difficulty is that mature follicles are not always represented in herbarium specimens.

The number of follicles borne by a flower varies greatly in different species. P.Cambessedesii has the greatest number of any species, up to eight (and in cultivation even more) and P.emodi the smallest number—usually one follicle to a flower.

Hairs.—The hairs on the carpels, leaves and stem of paeonies are a distinctive feature but not a determining character. In nearly every species the hairs on the carpels are of a characteristic form—white, yellowish or yellow, long or short, bristly or downy, densely matted or almost absent, with variations of these types.

With regard to the hairs on the back of the leaves, Anderson (1817) in his monograph divided the herbaceous forms into those with glabrous leaves and those with pubescence on the back of the leaves and in some cases on the stem and petiole. Huth (1891) found this character inconstant, which is also my own experience. The back of the leaves and the stem of some species, for instance P.arietina, have a far heavier covering of hair on specimens found in some districts than in others. There are, however, few herbaceous species which are completely glabrous, but where this is normally the case, pubescence of the leaf or stem would become a distinction of importance, though not justifying separation as a species. For instance, P.coriacea is glabrous, but the form found in Algeria has hairs on the underside of the leaf and is in consequence considered a variety—P.coriacea var. atlantica. Linnaeus wrote in his Critica Botanica, " Hairiness is a distinguishing character which may very easily become misleading, since it often disappears under cultivation ; if a plant exhibits any clearer or more unmistakable distinguishing character, I hold it to be very wrong to take refuge in ' hairiness ', that is unless one species is altogether glabrous and another very hairy and provided that the latter has never been known to lose its covering of hairs." [end page 10]

Several herbaceous paeonies possess hairs or bristles along the nerves on the upper side of the leaves. P.lactiflora has along the nerves minute hairs which are apt to disappear in old herbarium specimens as also is the case with P.emodi. P.anomala and P.Veitchii and their varieties have minute bristles along the nerves ; these bristles are not so pronounced in P.anomala as in P.Veitchii, but more prominent in P.Veitchii var. Woodwardii which is the only paeony species with hair both on the upper and lower surface of the leaf.

The Flower.—In nearly all the species the flower opens widely but there are certain exceptions. The strongly concave petals of P.peregrina form the flower into a cup which is a. distinctive feature of that plant in the garden, while P.japonica, P.Potanini var. trollioides, P.Brownii and P.californica hardly ever open their flowers widely ; these four species are all diploids.

The date of flowering of the species in cultivation seems to be a useful guide to their determination. The very early flowering of P.Cambessedesii in April in the open ground distinguishes it from all the varieties of P.Russi which flower in May ; and again the late flowering of P.Veitchii and its variety in June separates them from P.anomala and its variety which flower in May. This habit remains constant even after several generations have been raised from seed.

Number of Flowers to a Stem.—Among the herbaceous species three—P.Veitchii, P.lactiflora, and P.emodi and their varieties—bear several flowers on one stem. This character appears to be of taxonomic value ; all specimens of these species in the herbarium or the garden that have been examined have invariably more than one flower to a stem, or if there is only one flower there are in every case abortive buds showing in the axils of the leaves. All the species and varieties with this character are indigenous to Asia. Saunders and Stebbins (1938) consider the paeony species with several flowers to a stem to be the more primitive species. This may be so and possibly the fact that all these species are diploids may lend support to such a theory.

Sepals.—Stebbins (1939) describes how the shape and number of the sepals have proved valuable taxonomic characters. As he says, this character is difficult to observe on herbarium specimens; it has not been possible to examine this character in this study.

Disc.—The disc has been used in the same way as by Huth 46 (1891) and other writers in the definition of the three sections of the genus.

Seeds.—Stebbins101 (1939) considers the size, shape and surface markings of the seeds are fundamental characters separating the three subgenera of the genus, and he draws attention to the peculiar wrinkling of the seed-coat in many of the species having entire leaflets.

Seeds of all the species have not been observed, but from those examined it will be seen that:

(1) The seed of P.Delavayi (measuring about 13 X 12 mm.) is the largest, of P.lutea (13 x 10 mm.) the next largest, of P.Potanini var. trollioides (about 14 X 7 mm.) next, and of P.Potanini forma alba about the same, and of P.Potanini (about 10 X 8 mm.) the smallest. [end page 11]

(2) The next largest seed, but round to ovoid, about 10 mm. by 10 mm., belongs to P.emodi and P.emodi var. glabrata.

(3) The seed of P.Brownii is cylindrical ovoid, about 12 mm. long by 7 mm. broad.

(4) The remainder of the species in the- section PAEON have small round seed about 7'5 mm. in diameter. Stebbins's observation that the seed of species with entire leaflets have wrinkled seed coats deserves further examination. The correlation holds in the limited number of species I have been able to check.


By the courtesy of the Director, the John Innes Horticultural Institution has been kind enough to study the different species of paeonies from the cytological point of view and inform me of their chromosome numbers.

It is difficult to be sure that a plant referred to by different writers under a certain name is always the true plant. This is especially the case with paeony species. In order that no mistakes could occur with the chromosome counts, I supplied the John Innes Horticultural Institution with plants of each species under the names used in this study.

The cytology of paeony species has been studied by several research workers in this field of enquiry—Stebbins (1938 and 1939), Saunders and Stebbins (1938), Langlet (1927), Dark (1936) and Stebbins and S. Ellerton (1939).

The basic chromosome number throughout the genus is 5 and the size and morphology of the chromosomes differ little in the old world species, but Stebbins and S. Ellerton (1939) state that the American species are distinguished cytologically from the Old World species by the terminal localisation of pachytene pairing and extensive reciprocal interchange of chromosome segments. Some paeony species have 20 as their somatic chromosome number instead of 10, that is, some species are tetraploids (2n = 20), and some are diploids (2n = 10). The chromosome numbers in this work have been used as a taxonomic character to differentiate between the species and have often been of great value when used in this way. For instance, the two paeonies P.japonica and P.obovata var. alba are near neighbours geographically in Eastern Asia and difficult to separate morphologically ; there are only slight external differences between them and these are not easy to describe. When it was found that P.japonica was a diploid and P.obovata var. alba was a tetraploid this character confirmed the evidence furnished by the morphological data and became the determining factor in maintaining them as two distinct species.

The same occurs with regard to the diploid P.Cambessedesii and the tetraploid P.Russi var. leiocarpa; both these species come from the same geographical area and are only separated by slight morphological differences, so here again the chromosome numbers became the determining character. [end page12]

The chromosome numbers of the different species determined by the John Innes Horticultural Institution are given below. Where the chromosome number of a species has been obtained from another authority, the name and date of that authority is given in brackets against the name.

anomala var. intermedia






Broteri (Stebbins 1939)


Brownii (Stebbins 1939)














emodi var. glabrata




lactiflora (Stebbins 1938) ..










obovata var. Willmottiae


obovata var. alba(Langlet 1927)


officinalis (Saunders and Stebbins 1938)




Potanini var. alba




Russi var. Reverchoni


Russi var. leiocarpa


X Smouthii..


suffruticosa, garden form


suffruticosa. Rock's form








Veitchii var. Woodwardii


Wittmanniana var. nudicarpa




Information as to the possibilities of hybridisation between species of paeony will be found in the paper by Saunders and Stebbins. These investigators found that the diploid species do not cross easily with each other but that tetraploid species intercross naturally when placed side by side in the garden. The diploid species P.lactiflora is difficult to cross with other diploid species but crosses easily with the tetraploid species. The diploid P.Mlokosewitschi and the tetraploid P.Wittmanniana, though taxonomically close to each other, cross with each other with such difficulty that out of fifty-four pollinations only eight F1 plants were obtained.

Attention is drawn by Stebbins to the complete inter-fertility of P.Delavayi and P.lutea, and also between P.Veitchii and P.Veitchii var. Woodwardii. The plants of each pair are without doubt closely related, but in each case the morphological differences between them are sufficient to separate them ; in the case of P.Delavayi and P.lutea the differences are great enough to constitute them different species, but in the case of P.Veitchii and P.Veitchii var. Woodwardii, where the differences are less, they are sufficient to make Woodwardii a variety of the same species.

No triploid species have been discovered but triploid plants have been found among garden hybrids.[end page13]



0. F. I. Langlet. " Beiträge zur Zytologie der Ranunculazeen." Svensk. Bot. Tidstkr. 21, pP.1-17.


I. Manton. " Problem of Biscutella laevigata." Zeitschrift für Induktive Abstammung, 67.


E. Anderson and K. Sax. " Cytological Mon. of the American Species of Tradescantia." Bot. Gaz. 97, No. 3.


S. 0. S. Dark. " Meiosis in diploid and tetraploid Paeonia Species." Jour. got. 32, pP.353-372.


A. P.Saunders and G. L. Stebbins, Jr. " Cytogenetic Studies in Paeonia I." Genetics, 23, pP.65-82.


G. L. Stebbins, Jr. " Cytogenetic Studies in Paeonia II." Genetics, 23, pP.83-110.


G. L. Stebbins, Jr. "Notes on some systematic Relationships in the Genus Paeonia." University of California publications in Botany, 19, No. 7, pP.245-266.


G. L. Stebbins, Jr., and Sydney Ellerton. " Structural Hybridity in Paeonia Californica and P.Brownii." Jour. Gen. 38, Nos. i and 2, pP.1-36.


H. N. Barber. " Evolution of the Genus Paeonia." Mature, 148, 227. [end page14]