(OCR-Scan from The New Plantsman, 2001 8 (3) :156-171)

Tree Peonies

A Review of their History and Taxonomy

STEPHEN G. HAW

Attention! This publication is subject to copyright. The Copyright belongs to the author!

Tree peonies began to be cultivated in China well over one thousand years ago. When the first tree peonies were seen by Westerners, they were not wild plants but ones growing in Chinese gardens. The earliest introductions of tree peonies to the West were of plants obtained from cultivation in China. They were described scientifically and given botanical names, but their wild origins were completely unknown. At the time, it was impossible to ascertain whether they had been developed from a single wild species or were of hybrid origin. This later gave rise to considerable confusion when wild tree peonies began to be discovered and described (see Haw (1986)). In recent years, this confusion has begun to be resolved, but a number of problems have persisted, which this revision aims to clear up.

History of Tree Peonies in China

The earliest references to tree peonies in Chinese literature are to be found in texts more than two thousand years old. These very early references are to the use of tree peonies in Chinese medicine, however, and there is nothing to suggest that tree peonies were cultivated so long ago. Though there are some indications that they may sometimes have been grown in Chinese gardens as early as about AD 400, it seems likely that they became common garden plants during the Sui and Tang dynasties, at some time after AD 580. There are various references to tree peonies in Chinese literature of the Tang dynasty which indicate that they were probably widely cultivated at the time. However, few of these references make mention of distinct tree peony cultivars, nor are they sufficiently detailed to allow identification of any species in particular.

During the Song dynasty (AD 960-1279), they became extremely popular garden plants in China, so popular that the word 'flower' is said to have become more or less synonymous with 'tree peony'. As early as AD 986 a monograph which described some 32 different kinds of tree peonies is recorded as having existed: unfortunately, only its preface is extant today. This loss is at least partly compensated for by the fact that the Luoyang Mudan A [Record of Tree Peonies in Luoyang], written by Ouyang Mu about AD 1034, has survived to the present. It records 24 different cultivars and gives limited descriptions of them all. Some of the varietal names are still used for cultivars that exist today, though it is impossible to be certain that the name is still applied to the identical plant.

Ouyang's work tells us that by his time there were already tree peonies with fully double flowers (called 'thousand-petalled') as well as with semi-double ('many-petalled') and single blooms. A wide range of flower colours also existed, including yellow, various shades of red, purple and white. One cultivar is described as having 'semi-double' flowers that are white when they first open, gradually reddening as the days pass, so that when they fall they are deep red'. Interestingly, Ouyang also says that the people of Luoyang were able to distinguish the various varieties just by looking at the bushes, indicating that there were differences in habit and foliage as well as in colour and form of flower. This suggests that those cultivated at this period may well have been of hybrid origin.

The Chinese name for tree peonies, as used in the texts referred to above, is Mudan (formerly often transliterated as Moutan). This word is used today (and in the past) as a general name for all tree peonies: it does not refer to any particular species. The cultivars described by Ouyang Mu seem likely to have belonged to the group of cultivars included under the botanical name Paeonia suffruticosa Andrews. There is good evidence to suggest that these cultivars are the result of hybridisation of several wild tree peony species: this evidence will be reviewed below.

Ouyang Xiu's record of tree peony cultivars must have been less than comprehensive. Writing only some 50 years later, about AD 1082, Zliou Shihou listed no less than 109 cultivars in his Luoyang Huamu Ji. These included 10 cultivars with double yellow flowers, 34 with double red flowers, 10 with double purple, four with double white and one with double crimson: those with semidouble flowers included 32 reds, 14 purples, three yellows and one white. Zhou classified the cultivars into 37 groups, according to their varying forms.

An early illustration of a tree peony is found in a Chinese pharmacopoeia [bencao] printed in AD 1249 and, fortunately, a copy of the original printing has survived. The Chongxiu Zhenghe Jingshi Zhenglei Beiyong Bencao was an official work produced under Imperial patronage and a revised edition of a book first published in AD 1108. The illustrations may well date back to this earlier edition. The plant illustrated is said to come from ChuMou (now in eastern Anhui province, close to the border with Jiangsu, on the north side of the Yangtze River, north-west of the city of Nanjing) and appears to have bitemate (or perhaps sometimes tritemate) leaves. It is not really possible to be sure which tree peony this picture was intended to depict, but P. ostii or a single-flowered variety of P. suffruticosa would seem closest. Today, P. ostii and cultivars derived from it, such as 'Feng Dan BaP, are still commonly cultivated for medicinal use in Anhui province, particularly near Tongling and Ningguo, immediately south of the Chuzhou region.

Tree Peonies continued to be developed in China. The Mudan Shi [History of Tree Peonies], written about 1600 by Xue Fengxiang, records the names of more than 260 cultivars. Today, several hundred different cultivars are grown in China. They are still cultivated on a large scale in several areas, including Heze in Shandong province, Luoyang in Henan province, south-astern Anhui province and southern Gansu province. While much cultivation is for the root- bark of the plants, which is used in Chinese medicine, their ornamental value is not overlooked. Several Chinese nurseries are now supplying plants for export as well as for the home market.

Tree Peony Taxonomy

During the last dozen years or so, a great deal of new information has been published regarding tree peonies. New taxa have been discovered and described and the status of known taxa has been revised, in some cases more than once. Most of this work is summarised in an article by D.Y. Hong and K.Y. Pan (1999). (This article is in Chinese with an English abstract. Another article by the same authors appeared in English in the same year and duplicates most, but not all, of the content: Hong, D. Y. and Pan, K. Y. (1999a).) Unfortunately, Hong and Pan entirely ignore some recent taxonomic work on tree peonies, particularly that of Josef J. Halda (1997). They also accept or propose some dubious or erroneous hypotheses.

Hong and Pan recognise eight species, three of which each include two subspecies. These are (1) Paeonia suffruticosa Andrews subsp. suffruticosa and subsp. yinpingmudan D. Y. Hong et al., (2) P. jishanensis T. Hong & W. Z. Zhao, (3) P. qiui Y. L. Pei & D. Y. Hong, (4) P. ostii T. Hong & J. X. Zhang, (5) P. rockii (S. G. Haw & L. A. Lauener) T. Hong & J. J. Li subsp. rockii and subsp. taibaishanica D. Y. Hong, (6) P. decomposita Hand.-Mazz. subsp. decomposita and subsp. rotundiloba D. Y. Hong, (7) P. delavayi Franch. and (8) P. ludlowii (Stern & Taylor) D. Y. Hong. They also recognize two species of hybrid origin, P. x papaveracea Andrews (pro sp.) and P. x baokangensis Z. L. Dai & T. Hong. Since the Section Moutan DC. has only eight species, they see no reason to subdivide it and therefore do not recognise EC. Stern's subsections Vaginatae and Delavayanae.

I consider that there are few problems with Hong and Pan's species concepts. However, it seems extraordinary that they reject the opinion that P. suffruticosa is of hybrid origin and accept the supposed subsp. yinpingmudan as the sole wild ancestor of the species. Dr Osti (1994: 203) wrote that: 'everybody agrees that P. suffruticosa was first described on the basis of a plant which was a hybrid cultivar.' He has been in contact with both Chinese and western botanists for a number of years. Wang Lianying (1997: 8) suggests P. spontanea, P. rockii and P. ostii as the parents of the main grouP. of tree peony cultivars. Yu, Li and Zhou (1987) analysed the karyotypes of P. suffruticosa var. papaveracea (i.e. P. rockii) and P. suffruticosa var. spontanea (i.e. P. spontanea) and compared them with the karyotypes of several cultivars of P. suffruticosa. They concluded that some cultivars appeared to be close to P. rockii and some close to P. spontanea, suggesting either that the cultivars were derived from different ancestors or were of hybrid origin. Unfortunately however, neither karyotype analysis nor studies of DNA have yet provided very substantial evidence for the ancestry of P. suffruticosa. Sang, Crawford and Stuessy (1997: 1130) state that: 'Within subsection Vaginatae of section Moutan, however, morphological divergence apparently exceeds DNA sequence divergence. Three species studied by DNA sequences are morphologically distinct and allopatrically distributed, but have identical ITS sequences.' It is very unfortunate that Y. P. Zou, Cai & Wang (1999) did not include P. suffruticosa in their otherwise very useful study.

The external morphology of P. suffruticosa cultivars strongly suggests hybridity. Many cultivars have more or less distinct basal blotches on their flower petals, indicating the likelihood of P. rockii being involved in their parentage, but their leaves are clearly distinct from those of P. rockii and closer to those of either P. ostii or P. spontanea. Most cultivars produce basal suckers, (very rare in R rockii), even if they have flowers with blotched petals. The shape and number of leaflets and the growth habit of P. suffruticosa cultivars vary considerably. Seedlings from a single plant often show a wide range of variation. It is also significant that the cultivars are often self-fertile: tree peony species are usually strongly reluctant to self-pollinate. For example, Zhou, Hong and Pan (1999) report total self -incompatibility in P. jishanensis (P. spontanea). The reluctance of P. rockii to set seed unless it is cross-pollinated is well-known. I have personally reported that P. ostii does not seem to set viable seed when grown in isolation from other tree peonies (Haw (2000: 16 1)). In view of this strong tendency not to self-pollinate and the long period during which tree peonies have been cultivated in China, it would be surprising if P. suffruticosa cultivars. were not of hybrid origin. The strong likelihood is that P. suffruticosa is the result of hybridisation between P. rockii, P. ostii and P. spontanea, with back-crossing and selection over a period of several centuries.

Paeonia suffruticosa subsp. yinpingmudan

D. Y. Hong, Pan and Xie (1998) described P. suffruticosa subsp. yinpingmudan on the basis of only two plants, one of them cultivated. One of the plants has pale red-purple flowers and the other white. The plant with white flowers, from which the holotype was collected, was formerly usually identified as P. ostii, e.g. by Osti (1994: 199). by Hong Tao (personal communication) and by Wang Lianying (1997: 5). The other plant is said to have been brought into cultivation from the wild more than 30 years ago. However, in view of the very long history of cultivation of tree peonies in China. almost any apparently wild plant could be an escape from cultivation. It must be doubted whether this particular plant is genuinely of wild origin. Hong, Pan and Xie ( 1998) give absolutely no indication of how their P. siiffruti~.osa subsp. yinpingmudan differs from P. ostii. This is very surprising, in view of the previous identification of their type plant with the latter species. The leaf of the type illustrated in their article (Fig. 2. p. 519) appears to be within the possible range of variation of P. ostii foliage. The photographs accompanying the article (Plate 1: 2, 3) also seem to show plants closely similar to P. ostii, as far as it is possible to judge. Their description of the new subspecies is extremely brief: they say only that it differs from the typical subspecies by its single flowers ('A subspecies typica floribus singulis differt'). As there are single-flowered cultivars of P. suffruticosa, this is hardly sufficient tojustify the description of a new subspecies. This supposed subspecies must be rejected.

This article contains other very doubtful statements. The authors attempt to show that P. jish(itiensis (P. spontanea) cannot be the ancestor of P. suffruticosa because of a number of supposed differences. Some of the differences they cite (Table 1, p. 516) are spurious, however. For example, they claim that leaf shapes are different, but leaves of P. suffrutit.osa cultivars vary considerably and in some are quite similar to those of P. spontapica. The claim that P. suffruticosa reproduces only by seeds and never produces any root suckers is absolutely wrong. Most P. suffruticosa cultivars produce suckers occasionally.

Paeonia jishanensis (P. spontanea)

The recognition of P. jishanensis T. Hong & W. Z. Zhao as a distinct species seems to me to be correct. There is, however, a problem regarding the name of this taxon. When Hong and Zhao published this name in Hong, T. et al. (1992: 225 - 226, 232), they designated the specimen Hong Tao 915010 as its holotype. After describing the species and giving diagnoses in Latin and Chinese, they went on to say: 'On the basis of specimen no. 338 collected by W. Purdom in 1910, 50 li west of Yan'an, A. Rehder, the American taxonomist of woody plants, published a new variety of tree peony (i.e. P. suffiruticosa Andrews var. spontanea Rehd.). The original Latin diagnosis of this variety described the flowers as pink, sometimes with petaloid stamens (Journal of the Arnold Arboretuni 1: 193, 1920). In 1990, the British scientists S.G. Haw & L. A. Lauener raised A. Rehder's variety to a subspecies, P. suffruticosa Andrews subsp. spontanea (Rehd.) Haw & Lauener, and included the wild tree peony with white flowers from Majiagou, Ji Shan, Shanxi province in this subspecies. The authors' opinion is that P. suffruticosa subsp. spontanea has petaloid stamens, which are an important characteristic developed after domestication of a wild tree peony, so that therefore P. suffruticosa subsp. spontanea should be reduced to the status of a cultivar, 'Spontanea'.' (translated by S. G. Haw from the original Chinese.)

This clearly indicates that Hong and Zhao included the type of P. suffruticosa subsp. spontanea (i.e. Purdom 338) in their concept of the species P. jishanensis (this has been confirmed by personal communication with Hong Tao). In other words, they were not describing a new taxon at all, but changing the rank of a previously-recognised taxon. They were raising P. suffruticosa subsp. spontanea to specific rank. They seem to have thought that because they considered Purdom 338 to represent a cultivated plant differing slightly from wild plants, it could not be the type of their new species, and therefore designated Hong Tao 915010 as the holotype of their P. jishanensis. This, however, is an error. There is no reason at all why a cultivated plant cannot be the type specimen of a taxon. The type of this taxon must be Purdom 338. The name P. jishanensis, typified by Hong Tao 915010, is, therefore, superfluous and illegitimate. In correspondence with Chinese botanists, I suggested that this might be the case. Subsequently, T. Hong and Osti (1994: 238) published the name P. spontanea (Rehd.) T. Hong & W. Z. Zhao, with P. suffruticosa Andr. var. spontanea Rehder, P. suffruticosa Andr. subsp. spontanea (Rehd.) S. G. Haw & L. A. Lauener and P. jishanensis T. Hong & W. Z. Zhao as synonyms. Unfortunately, Hong and Osti did not explain that P. jishanensis was an illegitimate name, so it appeared that the name P. spontanea was merely a later synonym of P. jishanensis.

Paeonia delavayi and P. ludlowii

Many gardeners and some botanists may find it surprising that all taxa formerly recognised in Subsection Delavayanae have been reduced to synonymy with P. delavayi Franch., except for P. lutea Delavay var. ludlowii Stern & Taylor, which has been raised to specific rank. Good reasons have, however, been put forward for these changes. D. Y. Hong, Pan and Yu (1998) analyse several populations of P. delavayi sensu lato and show that it is impossible to discern clear distinctions between them. D. Y. Hong (1997) makes a convincing case for raising P. ludlowii to the rank of species. These conclusions are supported by RAPI) analysis, as reported by Y R Zou, Cai and Wang (1999). This treatment therefore seems to be correct. Previously recognised taxa within P. delavayi, which are distinct for horticultural purposes, will need to be given informal or cultivar names.

Paeonia X papaveracea (P X yananensis)

DX Hong and KM. Pan (1999 and 1999a) consider that P. yananensis T. Hong & M. R. Li is obviously a hybrid between P. rockii and P. spontanea. This opinion seems to be very probably correct. However, they further consider that P. yananensis is identical with P. papaveracea Andrews. Haw and Lauener (1990) treated P. papaveracea as a cultivar of P. suffruticosa. Although the flowers of P. yananensis and P. papaveracea show considerable similarity, there are very clear differences in the leaves. Those of P. yananensis often have 11-15 leaflets, while those of P. papaveracea never have more than nine. The shape of the leaflets is also different. The foliage of P. papaveracea is absolutely typical of P. suffruticosa, closely similar to that of most P. suffruticosa cultivars. It does not, therefore, seem reasonable to consider these two taxa identical. P. yananensis is probably the result of hybridisation only between P. rockii and P. spontanea, while P. papaveracea is a cultivar of P. suffruticosa, a complex hybrid with not only R rockii and P. spontanea but also P. ostii in the parentage. The name for the hybrid between P. rockii and P. spontanea should, therefore, be P. x yananensis T. Hong & M. R. Li (pro sp.).

Finally, although the number of species in Paeonia sect. Moutan is small, the division into subsections reflects a clear difference between two very distinct groups of species. The analysis of Y. P. Zou, Cai and Wang (1999) supports this division. The subsections were created at a time when there were fewer species of tree peonies recognised than today. It is not unreasonable to maintain these subdivisions of the genus. Indeed, Halda (1997) has raised their rank, the section Moutan becoming a subgenus and the subsections becoming sections. Though much of Halda's work is questionable, this treatment seems entirely reasonable and is accepted here.

Taxonomic treatment

Paeonia L. subgenus Moutan (DC.) Seringe, Fl. Jard. 3: 187 (1849).

Sect. Moutan DC., Prodr. 1: 65 (1824); sect. Suffruticosae Salm-Dyck, Hort. Dyck, 365, 366 (1834); sect. Moutania Reichenbach, Repert. Herb. s. Nom. Gen. Pl. 191 (1841); sect. Palaearcticae subsect. Fruticosae Huth in Engler, Bot. Jahrb. 14: 272 (1891); sect. Mutan Ascherson & Graebner, Syn. Mitteleur. Fl. 5 (2): 558 (1923); Paeonia subgen. Moutan (DC.) Lynch, Joum. Roy. Hort. Soc. 12: 432 (1890); Moutan (DC.) Reichb. in Moessler, Handb. Gewachsk. Ed. 2 (l): 55 (1827), pro gen. Typus: P. x suffruticosa Andrews

KEY TO SPECIES

1. Flowers solitary, erect; disc leathery, sheathing the carpels to at least half their length (Section

Moulan DC.): 2

1. Flowers usually 2 or 3, more or less pendent; disc fleshy, surrounding only the base of the carpels (Section Delavayanae (F. C. Stern) J. J. Halda): 6

2. Carpels 2-5, glabrous, sheathed to no more than two-thirds their length by the disc; leaves much divided into (29-)33-63 leaflets:

5. P. decomposita

2. Carpels 5(-7), tomentose, more or less entirely sheathed by the disc; leaves biternate or 2-3-

pinnate, leaflets 9-33: 3

3. Small shrubs to not more than c. 1.2 m tall; leaves more or less biternate, leaflets usually 9, sometimes 11-15, ovate to suborbicular; flowers comparatively small, c. 10 cm in diameter: 4

3. Shrubs often at least 1.5 m tall; leaves more or less 2 - 3-pinnate, leaflets 11-33, ovate to lanceolate; flowers comparatively large, more than 10 cm in diameter: 5

4. Leaflets deeply 3-lobed and toothed, green above; petals unblotched: 1. P. spontanea

4. Lateral leaflets often entire, terminal leaflets usually shallowly 3-lobed, often reddish above; petals often with a red basal blotch: 2. P. qiui

5. Leaflets rarely more than 15, entire or deeply 2-3-lobed; petals unblotched: 3. P. ostii

5. Leaflets (17-) 19-33; petals with a prominent dark purple basal blotch: 4. P. rockii

6. Carpels usually 2-5, sometimes more; petals, filaments and stigma often not entirely yellow; plants usually less than 2 m tall: 6. P. delavayi

6. Carpels usually single, occasionally 2(-3); petals, filaments and stigma always yellow; plants 13-3.5 m tall: 7. P. ludlowii

NOTE: Tree peony leaves vary in size and number of leaflets on every shoot. Numbers of leaflets referred to throughout this article are those on the best-developed leaves, usually the second or third from the base of the shoot.

Section Moutan DC.

Prodr. 1: 65 (1824).

[sect. Moutan DC. subsect. Vaginatae Stern, Stud. Gen. Paeonia 1 (1946); Paeonia sect. Moutan ser. Suffruticosae Kem.-Nath., Trudy Tbil. Bot. Inst. 21: 14 (1961A.

Typus: P. x suffruticosa Andrews

DESCRIPTION. Flowers solitary, more or less erect; disc leathery, partially to more or less entirely sheathing the carpels.

1. Paeonia spontanea (Rehder) T. Hong & W. Z. Zhao, Bull. Bot. Res. (Harbin) 14 (3): 238 (1994). Typus: China, Shaanxi, "50 Ii W of YenanfW' [Yan'an], 1910, Purdom 338 (lectotype 1 A n.v., isolectotypes E, K).

P. suffruticosa var. spontanea Rehder, Journ. Arnold Arb. 1: 193 (1920), p.p. excl. specim. Taipeishan (Mt. Taibai), Purdom s.n.; Stern, Stud. Gen. Paeonia 43 (1946); K. Y. Pan, Fl. Reip. Pop. Sin. 27: 45 (1979); P. suffruticosa subsp. spontanea (Rehder) Haw & Lauener, Edinb. Journ. Bot. 47 (3): 278, fig. It (1990); P. jishanensis T. Hong & W. Z. Zhao, Bull. Bot. Res. (Harbin) 12 (3): 225, fig. 2 (1992), nom. illegit.; P. suffruticosa subsp. spontanea (Rehder) Haw & Lauener var. jishanensis (T. Hong & W. Z. Zhao) J. J. Halda, Acta Mus. Richnov., Sect. Nat., 4 (2): 30 (1997).

DESCRIPTION. A small shrub to about 1.2 m tall. Leaves usually with 9 leaflets or sometimes 1115, ovate or broadly ovate to almost circular, usually rather bluntly lobed. Flowers single, rather small (c. 10 cm diam.), purplish to pink or white, unblotched, disc dark reddishpurple.

DISTRIBUTION. Shanxi and northern and central Shaanxi.

2. Paeonia qiui Y. L. Pei & D. Y. Hong, Acta Phytotax. Sin. 33 (1): 91, fig. 1 (1995). 'Iypus: China, Hubei, Shennongjia, Songbai [Sunbail Town, alt. 1650-2010 m., 5 iv 1988, J. Z Qiu PB88034 (holotype, PE n.v.).

P. ridleyi Z. L. Dai & T. Hong, Bull. Bot. Res. (Harbin) 17 (1): 1, fig. 1 (1997); P. suffruticosa subsp. spontanea (Rehder) Haw & Lauener var. qiui J. J. Halda, Acta Mus. Richnov., Sect. Nat., 4 (2): 31 (1997)

DESCRIPTION. A small shrub to about 80 cm tall. Leaves bitemate, leaflets 9, reddish-purple above, ovate to almost circular, lateral leaflets often unlobed, terminal leaflets usually trilobed. Flowers single, rather small (c. 10 cm diam.), petals pink or rose, often with a red basal blotch, disc dark reddish-purple.

DISTRIBUTION. Hubei (Shennongjia mountain, region) and Henan (Xi Xia).

3. Paeonia ostii T. Hong & J. X. Zhang, Bull. Bot. Res. (Harbin) 12 (3): 223, fig. 1 (1992). Typus: China, Zhengzhou Institute of Aeronautical Industrial Management, Arboretum of Rare and Endangered Trees, introduced from Mt. Yangshan of Songxian, Henan Province, 10 v 1990, T Hong 905010 (holotype, CAF n.v.). P. ostii T. Hong & J. X. Zhang var. lishizhenii B.

A. Shen, Acta Phytotax. Sin. 35 (4): 360 (1997); P. suffruticosa subsp. ostii (T. Hong & J. X. Zhang) J. J. Halda, Acta Mus. Richnov., Sect. Nat., 4 (2): 30 (1997); P. suffruticosa subsp. yinpingmudan D. Y. Hong, K. Y. Pan & Z. W. Xie, Acta Phytotax. Sin. 36 (6): 519, fig. 2, pl. 2.

DESCRIPTION. Shrub to about 1.5 m tall. Leaves bitemate or bipinnate with three pinnae, leaflets 9-15, more or less lanceolate with acuminate tip, entire or sometimes 2-3-lobed, greyish-green, especially when young. Flowers large (c. 15 em diam.), single, petals white, often flushed with purplish-pink, unblotched, disc dark reddish-purple.

Recorded as growing wild in Henan, Anhui, Hunan and Shaanxi, but possibly now extinct as a genuinely wild plant except in western Henan. Perhaps still common in cultivation, especially in Anhui, but many cultivated plants similar in appearance to this species may be hybrids between it and P. X suffruticosa. Plants cultivated under the name 'Feng Dan Bai' (='Phoenix White') may be the true species, but at least some appear to be hybrids.

4. Paeonia rockii (S. G. Haw & L. A. Lauener) T. Hong & J. J. Li ex D. Y. Hong, Acta Phytotax. Sin. 36 (6): 539 (1998). Typus: China, Kansu [Gansul, 'probably near Wutu [Wudul, (Farrer's Chieh Jo)', [probably iv 19141, Farrer (no. 8?) (holotype E).

P. rockii T. Hong & J. J. Li, Bull. Bot. Res. (Harbin) 12 (3): 227, fig. 4 (1992), nom. inval. (lacking exact page reference of basionym); P. suffruticosa subsp. rockii S. G. Haw & L. A. Lauener, Edinb. Journ. Bot. 47 (3): 279, fig. la (1990); P. suffruticosa sensu Stern, Stud. Gen. Paeonia 40 (1946), p. p. & sensu Fang Wenpei, Acta Phytotax. Sin. 7 (4): 313 (1958), p. p.; P. papaveracea sensu Anon., Icon. Cormophyt. Sin. 1: 652, fig. 1303 (1972); P. suffruticosa var. papaveracea sensu K. Y. Pan, Fl. Reip. Pop. Sin. 27: 45, pl. 3 (1979), non Kerner; P. suffruticosa 'Rock's Variety' ('Joseph Rock') hort.

DESCRIPTION. Shrub to about 2 m tall. Leaves more or less triternate or 2-3-pinnate, with 19-33 leaflets, rarely less than 19 but always more than 15. Flowers very large (15-20 cm diam.), single; petals white, sometimes flushed with pink, with a conspicuous deep purple basal blotch, disc whitish.

KEY TO SUBSPECIES

Leaflets lanceolate to ovate-lanceolate, often entire or sometimes 2-3-lobed:

4a. P. rockii subsp. rockii

Leaflets ovate to suborbicular. mostly 2-5-lobed:

4b. P. rockii subsp. taibaishanica

4a) P. rockii subsp. rockii

P. rockii subsp. lin.yanshanii T. Hong & G. L. Osti, Bull. Bot. Res. (Harbin) 14 (3): 237, fig. 1 & 2 (1994); P. suffruticosa subsp. rockii S. G. Haw & L. A. Lauener var. linyanshanii (T. Hong & G. L. Osti) J. J. Halda, Acta Mus. Richnov., Sect. Nat., 4 (2): 30 (1997).

DESCRIPTION. Leaflets lanceolate or ovate- lanceolate, entire or sometimes 2-3-lobed.

DISTRIBUTION. Southern Gansu, southern Shaanxi (on the southern slopes of the Qinling Mountains), Henan (Funiu Mountains) and 1 western Hubei.

4b) P. rockii subsp. taibaishanica D. Y. Hong. Acta Phytotax. Sin. 36 (6): 542, fig. 2 (1998). Typus: China, Shaanxi, Mt Taibai, Shangbaiyun, alt. 1750 m., in broad-leaved deciduous forest on cliff, 24 v 1985, D. Y Hong & X. Y Au PB85061 (holotype, PE n.v.).

DESCRIPTION. Leaflets ovate to suborbicular, mostly 2-5-lobed.

DISTRIBUTION. Found only on the northern slopes of the Qinling Mountains in southern Shaanxi and Gansu. Reported to have been common on Mt Taibai in the 1960s, but now scarce.

5. Paeonia decomposita Hand.-Mazz., Acta Hort. Gotoburg. 13: 39 (1939); D. Y. Hong et al., Taxon 45 (1): 68 (1996); D. Y. Hong. Kew B u 11. 52 (4): 957-963 (1997). Typus: China, NW Sichuan, Chosodjo, H. Siniffi 4641 (holotype, UPS- n.v., photo. K).

P. szechuanica Fang, Acta Phytotax. Sin. 7 (4): 315 (1958); K. Y. Pan, Fl. Reip. Pop. Sin. 27: 45, fig. 4 (1979); D. Y. Hong in L. K. Fu (ed.), China Plant Red Data Book 1: 536 537 (1992).

DESCRIPTION. Shrub up to 1.5 m tall. Leaves much divided into usually about 3363 leaflets. Flowers large (1215 em diam.), single, rosepink, with the white disc enveloping the carpels only to about half to twothirds of their length.

KEY TO SUBSPECIES

Carpels usually 5, occasionally 46; leaflets narrow, lanceolate to ovatelanceolate:

5a. P. decomposita subsp. decomposita Carpels 25, often 3 or 4; leaflets broad, ovate to suborbicular: 5b. P. decomposita subsp. rotundiloba D. Y. Hong

5a. P. decomposita subsp. decomposita

DESCRIPTION. Leaflets narrow, lanceolate to ovatelanceolate. Carpels usually 5, occasionally 4 or 6.

DISTRIBUTION. Found only in north-west Sichuan, in the valleys of the upper Dadu River (Da Jin Chuan, Chuosijia River) and its tributaries. The holotype and most other specimens come from localities within Barkam and Jinchuan Counties [xian], but there is a specimen from Danba County and also one from northern Kangding County.

5b. P. decomposita subsp. rotundiloba D. Y. Hong, Kew Bull. 52 (4): 961, fig. IA (1997). Typus: China, NW Sichuan, Lixian, D. Y Hong, Y B. Luo & Y H. He H95033 (holotype, PE .n.v., isotypes GH1A, K, MO, US all n.v.)

DESCRIPTION. Leaflets broad, ovate to suborbicular. Carpels 2-5, often 3 or 4.

DISTRIBUTION. Also restricted to north-west Sichuan, but in the valleys of the upper Min River and its tributaries, separated from subsp. decomposita by mountains some 4000 m high. Most specimens come from Maoxian County, with a few from near Lixian and Wenchuan, a couple from eastern Heishui County and one from near Zhenjiangguan in southern Songpan County.

Hybrid taxa:

Paeonia X suffruticosa Andrews, pro sp., Bot. Rep. 6: t. 373 (1804); Stern, Stud. Gen. Paeonia 40 (1946); Fang Wenpei, Acta Phytotax. Sin. 7 (4): 313 (1958), p.p. excl. syn. P. decomposita Hand,Mazz.; K. Y Pan, Fl. Reip. Pop. Sin. 27: 41 (1979), p.p. excl. syn. P. decomposita Hand, Mazz. Typus: AndreWplate, Bot. Rep. 6: t. 373 (1804).

P. arborea Donn, Hortus Cantabrig., 3rd. ed., 102 (1804), nomen nudum; C. C. Gmel., Hort. Mag. Duc. Bad. Carlsruh. 192 (1811); P. suffruticosa var. purpurea Andrews, Bot. Rep. 7: t. 448 (1807); P. papaveracea Andrews, Bot. Rep. 7: t. 463 (1807); P. moutan Sims, Curtis' Bot. Mag. 29: t. 1154 (1808); P. fruticosa Dum. Cours., Bot. Cult., ed. 2, 4: 462 (1811); P. moutan var. anneslei Sabine, Trans. Hort. Soc. 6: 482, tab. (1826); P. suffruticosa f. anneslei (Sabine) Rehder, Journ. Arnold Arb. 1: 194 (1920); P. yunnanensis Fang, Acta Phytotax. Sin. 7 (4): 306, pl. 61: 2 (1958).

This taxon originated centuries ago in China and is believed to be the result of hybridisation between P. spontanea, P. ostii and P. rockii.

Paeonia X yananensis T. Hong & M. R. L!, pro sp., Bull. Bot. Res. 12 (3): 226, fig. 3 (1992). l~rpus: China, Shaanxi, Yan'an, Wanhua Shan, in Platycladus forest, 15 v 1991, T Hong 915013 (holotype, CAF .n.v.)

P. suffruticosa subsp. rockii S. G. Haw & L. A. Lauener var. yananensis (T. Hong & M. R. Li) J. J. Halda, Acta Mus. Richnov., Sect. Nat., 4 (2): 30 (1997); P. x papaveracea sensu D. Y Hong & K. Y Pan, Acta Phytotax. Sin. 37 (4): 366 (1999T, non Andrews.

This taxon is intermediate in morphological characters between P. spontanea and P. rockii and is considered to be a hybrid between them. Both these species are present in the area where it was found. It has quite small flowers (10- 12 cm diam.) with white or pale purplish-rose petals with a dark purple basal blotch and deeP. purple disc. The leaves are quite similar to those of P. spontanea, but with narrower leaflets. D. Y Hong and K. Y Pan (1999) considered T. Hong and M. R. Li's species to be identical with Andrews' P. papaveracea. There are very obvious differences between the leaves, however: those of Andrews' P. papaveracea have no more than 9 leaflets with entire margins or few lobes (rarely more than 3), while those of Hong and Li's P. yananensis have up to 11 usually 3-lobed and toothed leaflets. P. yananensis also seems to be a much smaller shrub than P. papaveracea. The latter is almost certainly a cultivar of P. x suffruticosa.

Paeonia X baokangensis Z. L. Dal & T. Hong, pro sp., Bull. Bot. Res. (Harbin) 17 (1): 2, fig. 2 (1997). Typus: China, Hubei, Baokang County, Houping Township, 2 v 1996, Z. L. Dai, D. Y Ran & Q. D. Li 96047 (Baokang Forestry Research Institute. n.v.)

DESCRIPTION. Intermediate between and believed to be a hybrid of P. qiui and P. rockii (see Hong, D. Y. & Pan (1999a): 297-299).

Section Delavayanae (F. C. Stern) J. J. Halda

Acta Mus. Richnov., Sect. Nat., 4 (2): 29(1997).

[sect. Moutan DC. subsect. Delavayanae Stern, Stud. Gen. Paeonia 1 (1946); Paeonia sect.

Flavonia ser. Lutea Kem.-Nath., Trudy Tbil. Bot. Inst. 21: 14 (1961); Paeonia sect. Paeon ser. Delavayanae Kem.-Nath., Trudy Thil. Bot. Inst. 21: 14 (1961)] Typus: P. delavayi Franch.

DESCRIPTION. Flowers usually 2 or 3 (3), nodding or more or less pendent; disc fleshy, surrounding only the base of the carpels.

6. Paeonia delavayi Franch. in Bull. Bot. Soc. Fr. 33: 382 (1886); Huth, Bot. Jahrb. 14: 273 (1892); Stern, Stud. Gen. Paeonia 44 (1946); K. Y. Pan, Fl. Reip. Pop. Sin. 27:47-48, pl. 5 (1979); D. Y. Hong, K. Y. Pan & H. Yu, Ann. Missouri Bot. Gard 85: 554- 564 (1998). Typus: China, NW Yunnan, Likiang [Lijiang], Delavay 1142 (holotype P. n.y., isotype K).

P. lutea Delavay ex Franch., Bull. Soc. Bot. Fr. 33: 382 (1886); Huth, Bot. Jahrb. 14: 272 (1892); P. delavayi var. lutea (Delavay ex Franch.) Finet & Gagnep., Bull. Soc. Bot. Fr. 51: 524 (1904); P. delavayi var. lutea f. superba Lemoine, Rev. Hort. 14 (1906); P. lutea var. superba (Lemoine) hort. ex Gard. Chron., Ser. 3, 44: 50 (1908); P. potaninii Kornarov, Not. Syst. Herb. Hort. Petrop. 2: 7 (1921); P. trollioides Stapf ex Stern, Journ. Roy. Hort. Soc. 56: 77 (1931); P. potaninii var. troffloides (Stapf ex Stern) Stern, Stud. Gen. Paeonia 50 (1946); P. delavayi var. angustiloba Relider & Wilson in Sarg., Pl. Wilson. l: 318 (1913); P. delavayi var. atropurpurea Schipcz., Not. Syst. Herb. Hort. Petrop. 2: 47 (1921); P. delavayi var. alba Bean, Trees and Shrubs 3: 265 (1933); P. potaninii f. alba (Bean) Stern, Stud. Gen. Paeonia 49 (1946); P. x franchetii J. J. Halda, Acta Mus. Richnov., Sect. Nat. 4 (2): 31 (1997); P. x handel-mazzettii J. J. Halda, Acta Mus. Richnov., Sect. Nat. 4 (2): 31 (1997).

DESCRIPTION. Shrub or subshrub, 0.24.8 m tall, glabrous throughout. Roots often fusifornfly thickened, widely suckering. Lowermost 2-3 leaves bitemate, the leaflets divided into 3-11 primary segments, these in turn each more or less deeply cleft into 2-11 lobes; segments and lobes linear to narrowly lanceolate, entire or occasionaally dentate. Flowers 1 or 2-3 (3) on each shoot, 4- 10 cm diam.; petals yellow, yellow with a red or purplish basal blotch, red or purplish-red, or sometimes white, orange, greenish-yellow, or yellow with a red margin. Disc green, yellowish, yellow, red to dark red. Carpels 2-4(-8).

DISTRIBUTION. W China (west & south-west Sichuan, W Yunnan, SE Tibet).

Many gardeners, and perhaps some botanists, will find it hard to accept that the several formerly recognised species and varieties listed as synonyms above should now all be included in the single species P. delavayi, with no infraspecific taxa. There are, however, good reasons for this, which have been carefully researched and explained by D. Y. Hong, Pan and Yu (1998). They have shown that characters that have previously been used to separate taxa now all included within this species, such as flower colour, presence or absence of a conspicuous involucre and width of leaf segments, are all very variable and show little or no correlation with each other and with geographic distribution. In my opinion they therefore cannot be considered taxonomically significant and only one, variable, species can be recognised, although this view may not be shared by some of my taxonomic colleagues. For horticultural purposes, it would be useful and appropriate for clones within this species to be given forma or cultivar names (this will be tackled in a forthcoming paper).

7. Paeonia ludlowii (Stern & Taylor) D. V. Hong, Novon 7 (2): 157, fig. 1 & 2 (1997). Typus: China, SE Tibet, Kongbo Prov.. Miling (Mainling), Tsangpo Valley, 28 v 1938, Ludlow, Sherrj & Tavlor 4540 (holotype, BM).

P. lutea var. ludlowii Stern & Taylor, Journ. Roy. Hort. Soc. 76 (6): 217 (195 1 ); P. ludi(;vt,ii (Stern & Taylor) J. J. Li & D. Z. Chen, Bull. Bot. Res. (Harbin) 18 (2): 154 (1998), isonym; P. lutea Delavay ex Franch. subsp. ludlowii (Stern & Taylor) J. 3. Halda, Acta Mus. Richnov., Sect. Nat., 6 (3): 234 (1999); P. delavayi var. lutea sensu Marquand, Journ. Linn. Soc. Bot. 48: 158 (1929); P. lutea sensu Stern, Stud. Gen. Paeonia 46 (1946), p. p., quoad plantas a Ludlow, Sherriff & Kingdon Ward in Tibet lectas, P. lutea. Tibetan form, Stern, Journ. Roy. Hort. Soc. 72: 394, fig. 157 (1947).

DESCRIPTION. Shrub to 3.5 m. Roots not fusiform, suckering close to the stem bases. Lowermost 2-3 leaves more or less biternate; leaflets usually cleft almost to the base into 3 segments, each segment 3-lobed more or less to the middle, the lobes in turn entire or with 1 or 2 teeth; segments and lobes acuminate at the apex. Flowers usually 3 or 4 on each shoot. 10- 12 cm diam.. nodding or somewhat pendulous; petals always entirely yellow. Disc yellow. Carpels 1 or occasionally 2(-3).

DISTRIBUTION. Endemic to south-east Tibet (Nyingchi, Mainling and Lhunze counties). It seems entirely appropriate for P. ludlowii to he regarded as a distinct species. The fact that the 1 flowers, although usually larger than those of P. delavayi, have only one carpel, or sometimes two, is a very significant difference. The congested growth habit of P. ludlowii and its geographic separation from P. delavayi also support its elevation to specific rank.

Intersectional hybrid taxon

Paeonia x lemoinei Rehder, Journ. Arnold Arb. 1: 194 (1920). Typus: plate in Gard. Chron., ser. 3, 57: supplementary plate (1915), (lectotype, here designated).

P. saundersi Haworth-Booth, The Moutan or Tree Peony: 30 (1963), nomen nudum.

This hybrid is only known to exist in cultivation. The cross between P. delavayi (with yellow flowers: 'P. lutea') and P. suffruticosa was first made by the French nursery company Lemoine et fils in about 1900. There are now several named cultivars of this hybrid, the oldest being 'L'Esperance' and 'La Lorraine'. The name must also apply to all crosses between any plant belonging to P. delavayi sensu lato, irrespective of flower colour, and P. x suffruticosa. When Relider named the hybrid, he did not designate any acceptable holotype, saying only that 'The form 'L'Esperance' may serve as the type of this hybrid though he provided no reference to any specimen or illustration of this cultivar. He did, however, cite references to two illustrations of 'La Lorraine'. The best and most easily available of the illustrations cited is the colour plate designated as lectotype above. This hybrid has frequently been overlooked by taxonomists of tree peonies. Stern (1946) makes no mention of it, though he does note that a white-flowered plant of 'P. potaninii' had hybridised with P. delavayi (in his narrow sense) in his own garden. No Chinese author has included this name in any account of this subgenus, as far as 1 am aware.

Doubtful taxon

Paeonia suffruticosa subsp. atava (Brühl) S. G. Haw & L. A. Lauener, Edinb. Journ. Bot. 47 (3): 280 (1990). Typus: Tibet, Chumbi, Tuk Chang, vi 1884, King's collector 549 (holotype, K).

P. moutan subsp. atava Brühl, Arm. Roy. Bot. Gard. Calc. 5 (2): 114, t. 126 (1896).

All the comments on this subspecies in Haw and Lauener (1990) still hold good. D. Y. Hong (1997) reports finding no tree peonies in the Chumbi area and suggests that this taxon should be identified with P. rockii, which is cultivated in Tibet. However, examination of the type specimen shows this to be highly unlikely. The flower of the specimen is small and has unblotched petals, quite unlike those of P. rockii. Moreover, the collection location is uncertain and may in fact not have been in the Chumbi valley. P. rockii and cultivars derived from it (some possibly hybrids with other tree peonies) are also cultivated in Bhutan and the specimen may actually have been collected in western Bhutan. While it is indeed likely that this specimen was collected from a cultivated plant, or one that had escaped from cultivation, there can be no certainty of this. As the available material is rather poor and information about its collection location unreliable, it is unfortunately impossible to resolve the uncertainty about this taxon.

BIBLIOGRPAHY

Andrews, H. C. (1804). The Botanist's Repository, vol. 6:London.

Andrews, H. C. (1807). The Botanist's Repository, vol. 7: London.

Fang, W. P. (1958). Notes on Chinese Paeonies. Zhiwu Fenlei Xuebao/Acta Phytotaxonomica Sinica 7 (4): 297-323, pl. LM- LMIL

Halda, J. J. (1997). Systematic treatment of the Genus Paeonia L. with some nomenclatoric [sic] changes. Acta Musei Richnoviensis, Sect. Nat., 4 (2): 25-32.

Halda, J. J. & Horacek, L. (1999). New descriptions and combinations. Acta Musei Richnoviensis, Sect. Nat., 6 (3): 234 - 238.

Haw, S. G. (1985). Mudan: the king of flowers. The Garden 110 (4): 154-159.

Haw, S. G. (1986). A problem of Peonies. The Garden 111 (7): 326-328.

Haw, S. G. (2000). Paeonia ostii in Britain. The New Plantsman 7 (3): 160-163.

Haw, S. G. & Lauener, L, A. (1990). A review of the

infraspecific taxa of Paeonia suffiruticosa Andrews. Edinburgh Journal of Botany 47 (3): 273-28 1.

Haworth-Booth, M. (1963). The Moutan or Tree Peony: London, Constable.

Hong, D. Y (1997). Paeonia (Paeoniaceae) in Tibet. Novon 7 (2): 156-161.

Hong, D. Y (1997a). Notes on Paeonia decomposita Hand.-Mazz. Kew Bulletin 52 (4): 957 - 963.

Hong, D. Y (1998). Paeonia rockii and its one new subspecies from Mt. Taibai, Shaanxi of China. Zhiwu Fenlei Xuebao/Acta Phytotaxonomica Sinica 36 (6): 538 - 543.

Hong, D. Y. & Pan, K. Y (1999) Taxonomical history and revision of Paeonia sect. Moutan (Paeoniaceae). Zhiwu Fenlei Xuebao/Acta Phytotaxonomica Sinica 37 (4): 351-368.

Hong, D. Y & Pan, K. Y (1999a). A revision of the Paeonia suffruticosa complex (Paeoniaceae). Nordic Journal of Botany 19 (3): 289-299.

Hong, D. Y, Pan, K. Y & Pei, Y L. (1996). The identity of Paeonia decomposita Hand.-Mazz. Taxon 45 (1): 67-70.

Hong, D. Y_ Pan, K. Y. & Xie, Z. W. (1998). Yinpingmudan, the wild relative of the King of Flowers, Paeonia suffruticosa Andrews. Zhiwu Fenlei Xuebao/Acta Phytotaxonomica Sinica 36 (6): 515-520.

Hong, D. Y, Pan, K. Y. & Yu, H. (1998). Taxonomy of the Paeonia delavayi complex (Paeoniaceae). Annals of the Missouri Botanical Garden 85 (4): 554-564.

Hong, T. & Dai, Z. L. (1997). Study on the Chinese wild woody Peonies (III): New taxa of Paeonia L. sect. Moutan DC. Bulletin of Botanical Research (Harbin) 17 (l): 1-5.

Hong, T, & Osti, G. L. (1994). Study on the Chinese wild woody Peonies (H): new taxa of Paeonia L. sect. Moutan DC. Bulletin of Botanical Research (Harbin) 14 (3): 237- 240.

Hong, T. et al. (1992). Study on the Chinese wild woody Peonies (I): new taxa of Paeonia L. sect. Moutan DC. Bulletin of Botanical Research (Harbin) 12 (3): 223-234.

Li Shizhen (1590). Bencao Gangmu. Reprinted 1977: Beijing, Renmin Weisheng Chubanshe.

Osti, G. L. (1994). Tree Peonies revisited. The New Plantsman 1 (4): 195-205, 247.

Guyang XW (c. 1034). Luoyang Mudan A. Reprinted in: DU Weirno and Chen Xin (eds.) (1982). Ouyang Mu Wen Xuan: 43-56: Beijing, Renmin Wenxue Chunbanshe.

Pan, K. Y. (1979). Paeonia, in Flora Reipublicae Popularis Sinicae, 27: 37-59: Beijing, Kexue Chubanshe.

Pei, Y L. & Hong, D. Y (1995). Paeonia qiui - a new woody species of Paeonia from Hubei, China. Zhiwu Fenlei Xuebao/Acta Phytotaxonomica Sinica 33 (1): 91-93.

Rehder, A. (1920). Paeonia suffruticosa var. spontanea, var. nov. Journal of the Arnold Arboretum 1: 193-194.

Sang, T., Crawford, D.J. & Stuessy, T. F. (1997). Chloroplast DNAphylogeny, reticulate evolution and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84 (9): 1120-1136.

Stern, F. C. (1946). Study of the Genus Paeonia: London, RHS.

Tang Shenwei (ed.) (1249). Chongxiu Zhenghe Jingshi Zhenglei Beiyong Bencao. Photographic reprint 1957: Beijing, Renmin Weisheng Chubanshe.

Wang Lianying (ed.) (1997). Zhongguo Mudan Pinzhong TuzhilPictorial Record of Chinese Tree Peony Varieties: Beijing, Zliongguo Linye Chubanshe.

Wang Yuhu (ed.) (1964). Zhongguo Nongxue Shu Lu: Beijing, Nongye Chubanshe.

Xue Fengxiang (c. 1600). Mudan Shi. Reprinted 1983: Hefei, Anhui Renmin Chubanshe.

Yu, C. Y, Li, S. E & Zhou, J. Y. (1987). Karyotype analysis of Paeonia suffruticosa var. papaveracea and P. suffruticosa var. spontanea. Xibei Zhiwu Xuebao/Acta Botanica Boreali-Occidentalia Sinica 7 (1): 12-16.

Zhou Shihou (c. 1082). Luoyang Huamu A. Reprinted in: Gujin Tushu Acheng, 3rd. edition, Shanghai, Tongwen Shuju, 1898.

Zhou, S. L., Hong, D. Y. & Pan, K. Y. (1999). Pollination biology of P. jishanensis T. Hong & W. Z. Zhao (Paeoniaceae), with special emphasis on pollen and stigma biology. Botanical Journal of the Linnaean Society 130 (l): 43-52.

Zou, Y R, Cai, M. L. & Wang, Z. P. (1999). Systematic studies on Paeonia sect. Moutan DC. based on RAPD analysis. Zhiwu Fenlei Xuebao. Acta Phytotaxonomica Sinica 37(3): 220-227.

Stephen G. Haw, Worton Edge, Over Worton, Chipping Norton, Oxon., OX7 7EW.