F.C. Stern A Study of the Genus Paeonia

Section Paeon Subsection Foliolatae

wittmanniana group.—The Wittmanniana Group includes all the yellow-flowered herbaceous species which are indigenous to the Caucasus and northern Persia ; it consists of P.Wittmanniana, its two varieties, and P.Mlokosewitschi.

Other paeony species are also to be found in this region, including P.daurica (syn. P.triternata) and P.tenuifolia, which are referred to elsewhere.

There are two main difficulties with regard to these yellow Caucasian paeonies, firstly, their nomenclature, and secondly, the correct locations of the different species. Most of the specimens in herbaria come from the Tiflis Botanical Garden or from other gardens which had been furnished from the garden at Tiflis.

With regard to nomenclature, the original description of P.Wittmanniana was made by Lindley (1846) when he described the carpels as tomentose ; two years later (1848) Steven described a form with glabrous carpels under the same name, P.Wittmanniana, obviously in ignorance of Lindley's prior work. The plant with glabrous carpels has been known by Steven's name of P.Wittmanniana ever since that time, while Lindley's species with tomentose carpels has been named P.Wittmanniana var. tomentosa by Lomakin (1897) and also as P.tomentosa by Busch (1903) and Stapf (1931). Under the International Rules of Botanical Nomenclature it would appear that Lindley's name of P.Wittmanniana has the priority and should be adopted for the yellow-flowered paeony of the Caucasus with tomentose carpels, while the variety with glabrous carpels should be known as P.Wittmanniana var. nudicarpa, the name given by Schipczinsky "in 1921.

The reason why Steven's name was maintained by later writers for the paeony with glabrous carpels was probably because both he and Lindley write of a paeony to be found in the district of Abkhazia, at the west end of the Caucasus range near the Black Sea, where Wittmann was said to have collected. It has been stated by Busch (1901) and Stapf (1931) that the glabrous-carpelled paeony came from the west end of the Caucasus range and that the home of the tomentose-carpelled form was at the east end only of the range in the district of Talysch, near the coast of the Caspian Sea. This has now been proved not to be the case. Busch had already, in the Addenda (1902) to Flora Caucasica critica, said that specimens with tomentose carpels had been found in Abkhazia and this has been confirmed by specimens sent to Kew by the Leningrad Herbarium ; three of these specimens with tomentose carpels were collected in the district of Abkhazia. So it would appear that a yellow-flowered paeony with tomentose carpels is found in Abkhazia, the district where the glabrous carpelled form is more usually found. This paeony with tomentose carpels is no doubt the paeony described by Grossheim (1930) as P.abchasica Mischenko, which only differs from the description of P.Wittmanniana by the larger leaflets and slightly different hairs on the underside of the leaf. The Leningrad Herbarium sent to Kew their specimen of P.abchasica (which had been grown in the Tiflis Botanical Garden) and this specimen differed in no way from P.Wittmanniana.

Specimens of the paeony from the Talysch district with tomentose carpels are rare in herbaria and most of the specimens come from the Tiflis Garden. The only wild specimen known [end page 21] to me is in the British Museum and had been collected by Woosnam in the Elburz Mountains in Persia, on the south side of the Caspian Sea. It has also been collected by Buhse at Masula, which is in Persia.

P.Mlokosewitschi Lomakin, the other yellow-flowered Caucasian species, differs from P.Wittmanniana and its variety in that it has oblong to oval leaflets with the apices subacute to rounded and having only slight short pubescence on the underside of the leaf. Further, P.Mlokosewitschi has proved to be a diploid and the other two, P.Wittmanniana and its variety, are tetraploids. This paeony is found in a small district in the eastern central Caucasus near the village of Lagodekhi.

Another paeony with white or yellowish flowers was described by Albow (1895) from Abkhazia, which is the most southerly station of P.Wittmanniana var. nudicarpa, and called by him P.corallina var. Wittmanniana forma macrophylla. Albow described the carpels as glabrous, but Lomakin (1897), referring to this species, says: "carpels . . puberulous or glabrous." There is a specimen of this paeony at Kew collected by Albow, and I have also seen one lent by the Leningrad Herbarium, but in both cases the dried specimen is composed of leaves only which are very large, 24 cm. long by 15 cm. wide, resembling in shape those of P.Wittmanniana. Busch and the Flora U.R.S.S. (1937) describe this paeony as having glabrous carpels. Busch's name [end page 22] of P.Wittmanniana var. macrophylla is here adopted. There is a yellowish-flowered paeony at Highdown which came from the Tiflis Garden via Kew, with the name P.macrophylla. This plant has large leaflets 22 cm. long by 15 cm. wide on a lower leaf when in flower, but the carpels are tomentose and it differs from Wittmanniana only by the large leaflets.

To sum up the information cited, the names of the yellow-flowered Caucasian species and varieties and their distribution are as follows :— P.Wittmanniana At the eastern end of the Caucasus range near the Caspian Sea, extending down the mountains near the coast of the Caspian Sea to the Elburz mountains in northern Persia, and also in the district of Abkhazia

P.Wittmanniana var. nudicarpa In the district of Abkhazia, at the western end of the Caucasian range near the Black Sea.

P.Wittmanniana var. macrophylla In the most southerly station of P.Wittmanniana var. nudicarpa.

P.Mlokosewitschi Eastern central Caucasus, near the village of Lagodekhi.

the Russi group.—The four species which compose this group are confined to the islands of the western Mediterranean, Corsica, Sardinia, Sicily and the Balearic Islands. These species are grouped together by the shape and thick papery texture of the leaflets and by the number of leaflets, usually nine on a lower leaf.

P.Russi, which is found in Corsica, Sardinia and Sicily, is easily distinguished from the other species in this group by its densely tomentose carpels and the undersides of the leaves, which are only slightly hairy or pilose. In southern Corsica and in Sardinia occurs another paeony described by Le Grand (1899) under the name of P.Russi var. Reverchoni, which differs from P.Russi in having glabrous carpels and tomentose hair on the back of the leaf. This is another instance of two paeonies indigenous to the same geographical area whose main distinction is the presence or absence of tomentum on the carpels. In the centre and south of Corsica there is [end page 23] another paeony which also has glabrous carpels, but in this plant the leaves are larger than in P.Russi or its variety Reverchoni and the backs of the leaves are shortly tomentose. This paeony was named P.Corsica by Sieber and described by Tausch (1828). When this plant is seen growing, one might at first think it sufficiently different from P.Russi and its variety to be considered a distinct species, but on critically examining the plant it is seen to differ from the variety Reverchoni in the larger leaflets and deeper green of the leaves which are somewhat flushed with red as are also the carpels. In these characters this plant approaches the paeony of the Balearic Islands, P.Cambessedesii, but it is distinct from P.Cambessedesii in that the latter is quite glabrous, has larger leaflets, coriaceous and deep purple at the back, and also a greater number of carpels which are purple. As Sieber's paeony is more closely connected with P.Russi than with P.Cambessedesii, I have considered it a variety of P.Russi and given it the varietal name of P.Russi var. leiocarpa, following Cosson's (1850) name of P.corallina var. leiocarpa. P.Russi and its two varieties have a character in common which is lacking in P.Cambessedesii—the distinct petiolule of the terminal leaflets. P.Russi var. Reverchoni and P.Russi var. leiocarpa have been examined cytologically and found to be tetraploids, whilst P.Cambessedesii is a diploid, a finding which clinches the morphological evidence of their separation.

mascula group.—The name P.mascula of Miller must replace the name P.corallina Retz., owing to the priority of Miller's name. Huth (1891), in his monograph, brought a number of species together under the name P.corallina. They included P.mascula, P.daurica, P.banatica and P.Broteri ; these species were all connected together because of their entire leaflets, tomentose carpels and their unusually glabrous leaves ; with them he also included the Russi grouP.Although the four paeonies mentioned above have certain characters in common, they can be easily separated. Shortly the differences are : P.mascula, a tetraploid, has usually nine leaflets on a lower leaf with the terminal one sometimes bifurcating, making the number of leaflets up to thirteen, broadly elliptic with the apex acute ; P.daurica, a diploid, has usually nine leaflets which are oval to orbicular with undulate margins and the apex rounded to truncate ; P.banatica, chromosome number unknown, has the lower leaf like that of P.mascula but the upper leaves oblong or narrow elliptic after the style of P.arietina ; and P.Broteri, a diploid, possesses elliptic or broad elliptic leaflets up to twenty-three on a lower leaf. These differences, which are discussed fully under the description of each species, seem to preclude them from being regarded as mere varieties; they are, therefore, treated as distinct species.

P.Broteri is so different in the number of leaflets and in their shape from the other three species that it cannot be put into the same group as the others, and it appears to be a solitary species unconnected with any other. P.Broteri is confined to southern Spain and southern Portugal.

The remaining three species, P.mascula, P.daurica and P.banatica, on the other hand, form a group connected together by certain characters they have in common. Their distribution also tends to confirm this view.

P.mascula is indigenous to the centre of France in the Department of Loir et Cher and around Reichenhall in Germany, but it is also found in several other outlying stations such as Steep Hoime Island in the Bristol Channel, the islands of Cyprus and Sicily, as well as in Armenia and Syria. The explanation of this curious distribution may be that the plants found in these outlying stations are escapes from monastery gardens, since monasteries have been known to have existed in early Christian times in all these places and it is also well known that this paeony was considered to have valuable medicinal properties and was grown in monastery gardens for this purpose.[end page 24]

P.daurica is found in an area in and around the Crimea, reaching to the western Caucasus. P.banatica is found only in one small district in the Banat. This species is most puzzling and seems to combine some of the characters of P.mascula and some of P.arietina. It seems to be found only around Bazias where there is or was a monastery. At Kew there are specimens from the vicinity of the monastery.

The main distribution of these three species stretches from the Caucasus and the Crimea north-west through Transylvania to Reichenhall and the centre of France, and there is a newly discovered species P.kesrouanensis in Syria which is like P.mascula in the shape and number of the leaves but differs from it by the glabrous carpels and long stigma coiled only at the apex.

obovata group.—There are a number of paeony species in eastern Asia, China and Japan which fall into this group of herbaceous paeonies, all characterised by their entire leaflets, which are broadly oval with the terminal leaflet obovate, and by the long attenuated carpels ; all the leaflets have a pointed apex with a cuneate to widely cuneate base.

Paeonia obovata was described by Maximowicz in 1859 from plants collected in the Amur district north of Vladivostock ; this species covers a large area in eastern Siberia, Manchuria and northern China ; both pink and white flowers have been reported but only the white form, as far as can be ascertained, is in cultivation in England. In 1916 Dr. Stapf described a form of this paeony as P.Willmottiae from garden specimens raised from seed collected by E. H. Wilson in the province of Hupeh. Stapf raised it to specific rank mainly on account of its larger leaflets,[end page 25] which are tomentose beneath, and its white flowers. On examining several specimens collected by Wilson, Henry and Giraldi in western Hupeh and the province of Shensi, there seems to be only one outstanding character in which the plant differs from P.obovata—namely, the tomentose underside of the leaf; among a number of specimens examined there is little difference between the size of the leaflets in P.obovata and this form. There is, however, a curious habit in Miss Willmott's plant that the leaflets increase in size after the plant has flowered ; measurements taken from a plant at Highdown are given under P.obovata var. Willmottiae and show how extraordinary the increase is. This habit, which is also noticeable in P.obovata, is much more pronounced in P.Willmottiae and has caused some confusion among dried specimens in herbaria, as those in fruit have so much larger leaflets than those which are in flower. This plant, which appears to be confined to one geographical area, does not seem to be more than a variety of P.obovata ; in order to keep Stapf's name and also to commemorate the original cultivator of this beautiful paeony, it has been named P.obovata var. Willmottiae. Both P.obovata and its variety are tetraploids.

P.japonica was described first by Makino (1898) as a variety of obovata and then as a species by Miyabe and Takeda (1910). This plant has been known to Japanese botanists from the remotest times, and was in the past considered to be a white-flowered variety of P.obovata. Miyabe and Takeda raised it to specific rank and distinguished it from Maximowicz's plant mainly by the flowers, which open less widely and have more concave petals and short stigmas, and by the leaflets which are usually glabrous.

It is difficult to define the differences between this plant and P.obovata—the differences in the flowers and leaves are not great. However, on examination at the John Innes Horticultural Institution it was found that P.japonica is a diploid ; this character, therefore, separates it from P.obovata and its variety, which are tetraploids. On this account the specific name P.japonica given by Miyabe and Takeda has been retained.

In the Kew Herbarium there is a dried specimen of a paeony from Japan with tomentose carpels which comes from the mountains near Hakone ; Stapf marked the sheet with the name of P.chaetocarpa and a description has been found among his papers at Kew. There is only one specimen at present available. At first from the shape of the leaves this paeony might be regarded as a variety of P.japonica with hairy carpels, but the carpels are not long and attenuated as in P.japonica and the other species in this group ; this specimen I have not described as there is a doubt in my mind as to its true status. Stapf has written at the end of his description: " The specimen was mixed up with a detached leaf and the upper part of a young stem of P.lactiflora, which name was written upon the label."

This group consists of:—P.obovata Maxim.,; Siberia and North China.

P. obovata var. Willmottiae (Stapf) Stern; Hupeh and Shensi, West China

P. japonica; Miyabe and Takeda, Japan.

mairei group.—It is difficult to place P.Mairei and P.oxypetala, the two closely allied species in this grouP.These species, of which there are few specimens in herbaria, appear to lie between the Lactiflora group, which they somewhat resemble with their elliptic or obovate-elliptic leaflets, and the Obovata group which has similar long attenuated carpels. P.Mairei and P.oxypetala both have carpels long and attenuated up to the stigma, and tomentum, when present, [end page 26] of silky golden-brown hairs unlike the tomentum of any other paeony species. Plants found in the same districts seem in some cases to have glabrous carpels, while in others they have tomentose carpels. P.oxypetala differs from P.Mairei only in the shape of the petal, which has the apex shortly acute or pointed, a character very unusual among paeonies.

The few specimens that exist in herbaria were collected by four different collectors, Wilson in 1904, Pere Farges about 1905, Maire in 1913, and Handel-Mazzetti in 1914. Their distribution is confined, as far as is known, to a district in western China, extending from Tatsien-lu in Szechwan, southward through Ningyuen also in Szechwan, to the northern borders of Yunnan.

These long attenuated carpels are found only in the Obovata and Mairei groups, but the leaves in these two groups are so different that they have been kept as separate groups.

arietina group.—P.arietina and P.Bakeri are easily recognisable by the leaflets, which bifurcate into twelve to fifteen narrow segments measuring from 7 to 14 cm. long and 4 to 6-6 cm. wide, with the apex acute or subacute. The undersurface of the leaves, petioles and stem are villous and the carpels densely pubescent. P.arietina has a wide distribution through Bosnia, Italy, Greece and Asia Minor as far as northern Armenia and Turkestan. The species from different districts are variable in the size of the leaflets and the degree of hairiness of the leaves, petioles and stem. The forms from Bosnia and Italy are less hairy than those from Armenia, which are densely hairy. The paeony from Bazias, P.banatica, seems to be a nearly related species, but is discussed under the Mascula group.P.Bakeri was described by Lynch (1890) from a plant in the Cambridge Botanic Garden. It differs from P.arietina in its shorter segments and its taller and more robust habit. There are no wild specimens of this plant in herbaria. It seeds freely and the seedlings come true to type.

P.arietina and P.Bakeri are both tetraploids.

On the islands of Cyprus and Crete, and in Syria there are paeonies which are like P.arietina in the shape of their leaves and tomentose carpels, but otherwise are quite glabrous.[end page 27]

These have been identified with P.arietina var. orientalis Thiebaut; the type locality is Feitroun in Syria.

There are, however, specimens from the Island of Rhodes of paeonies with white flowers, which have been identified as a new species by W. T. Steam and described by him under the name of P.rhodia. It resembles P.arietina var. orientalis in many respects, but differs from it by the larger number of leaflets and their thin and papery character. It has been determined as a diploid, while P.arietina is a tetraploid. As P.arietina var. orientalis is not in cultivation, it has not been possible to ascertain its chromosome number.

broteri group.—This group, which consists of one species only, is discussed under the Mascula group on page 24.

coriacea group.—P.coriacea is found in southern Spain in the Sierra Nevada and the province of Granada and in the Middle Atlas Mountains of Morocco. It is separated from the Mascula and Russi groups by the leaflets which bifurcate into fourteen to sixteen segments with the lateral leaflets practically sessile. The carpels are glabrous and the plant is a tetraploid. In the hills of Algiers a variety is found, P.coriacea var. atlantica, which differs from P.coriacea by the pubescence on the lower surface of the leaf. The African plant flowers in April and the Spanish one in June.

lactiflora group.—This group contains two species and their varieties, P.lactiflora, a native of north-eastern Asia, and P.emodi of northern India. They differ from the other groups of the subsection foliolatae by the elliptic leaflets, about 2 1/2 to 3 times as long as broad, with acuminate apices.

P.lactiflora and P.emodi and their varieties have more than one flower to a stem, which is a character shared only with one other species, P.Veitchii and its variety. P.emodi differs from P.lactiflora in that the leaflets are not scabrid along the margins, but it carries hairs along the main veins of the otherwise glabrous upper side of the leaves. The flowers are always white and the carpels are usually solitary, rarely two, in which respect it differs from all other species. Both species have forms with glabrous and hairy carpels. In the typical plants of P.lactiflora the carpels are glabrous, but there are garden forms with hairy carpels which have been named P.lactiflora var. trichocarpa. In P.emodi the carpels are hairy, but there is a wild variety, P.emodi var. glabrata, with glabrous carpels. These two species and their varieties are diploids.

P.lactiflora is found mainly in Siberia and northern China but also occurs in Mongolia, Tibet and China. It was introduced to Europe at the beginning of the 19th century and garden forms have long been in cultivation, where they are especially welcome for their sweet-scented flowers.

P.emodi is native in northern India from Kumaun to Chitral, while the variety glabrata is recorded from Kashmir and the United Provinces. It is reported that in Kashmir plants with glabrous and with hairy carpels may be found growing together (page 96).Maps a-z