Carsten Burkhardt's Web Project Paeonia - The Peony Library

ano_finnland7 index 0602 boreas2

Paeonia-Literatur 2006

Peony rockii and Gansu Mudan

Will McLewin Dezhong Chen






1.1 Paeonia


1.2 Gansu Mudan and Paeonia Gansu Group; 'rockii hybrids'


1.3 Paeonia rockii


1.4 ZibanMudan


1.5 Names of cultivated hybrids






3.1 The evolution of the name

3.2 A compromise proposal

3.3 Observed Characteristics and Distribution of Paeonia rockii

3.4 Evidence from Molecular Analysis












Plants (like people) have innate physical characteristics but do not intrinsically have names. People give plants names for their own purposes, most importantly for meaningful discussion and exchange of information, also in some cases for self-gratification. Misuse or careless use of names defeats their purpose and renders them useless or worse, but what constitutes misuse is, in many cases, itself controversial. This book is about a particular group of cultivated mudan (shrubby peonies) for which the appropriate common name, in our view, is Gansu Mudan and the formal name we propose is Paeonia Gansu Group. A basic problem when discussing objects or accounts of events that have been widely incorrectly described is that many of the names involved have been corrupted by their misuse but nevertheless have still to be used. Consequently, before individual plants in this group and the group as a whole can

be presented and their origins, virtues, characteristics and availability described it is necessary to discuss various names and terms used and to explain the position of these plants in the context of mudan in general.

1.1 Paeonia

The genus Paeonia as a whole is uncontroversial in the sense that there are no serious disputes or problems about which of the earth's and of humanity's plants are peonies. Paeonia was formerly included in the family Ranunculaceae but is now the sole member of the family Paeoniaceae. Within Paeonia there are two well-defined main sections (and the unnecessary small third section for the two very similar North American herbaceous species). One section consists of peonies with permanent over-wintering woody stems. The common name is 'tree peonies', which is misleading because they are much more shrubs than trees. The name used by botanists for this




section is Moutan, an early transliteration from Chinese for the compound character now spelled Mudan in pinyin. The established, unambiguous and modern Chinese pinyin word Mudan is easy to say* and to spell. We use it throughout and recommend its use generally. (Incidentally and usefully 'mudan' is both singular and plural. See 9.2). The other section consists of herbaceous peonies, with non-woody stems which do not over-winter and grow anew each year, The pinyin word for these plants is Shaoyao. (There are some details about the origin of these Chinese terms in Chapter 4.) In the last few decades an intermediate section of Itoh or intersectional peonies has been developed, initially by crossing examples of mudan with examples of shaoyao. These peonies have blurred the distinction between the two main sections and as more examples are produced the once simple morphological distinction will be obscured by examples of intersectional hybrids which are not clearly intermediate but very close to one section or the other.

* (Consonants as in English; mu long as in 'moon'; dan short, like 'can' and with a slight stress.)

1.2 Gansu Mudan and Paeonia Gansu Group; 'rockii hybrids'

With mudan as the term for woody peonies in general the immediate question that needs comment is why Gansu Mudan or Gansu Group is the most appropriate name to use for this group or subset of woody peonies a whole. Perversely we begin by explaining why alternative names that have been used are inappropriate. The (wild) species now called Paeonia rockii is definitely the ancestral species most prominently involved and genetic material from P. rockii is undoubtedly present, albeit in almost all cases not directly. In addition, the plants themselves exhibit several of the properties that characterise P. rockii, most notably the dark blotch at the base of the flower petals and the vigour and tough versatility of P. rockii. Perhaps 'P. rockii hybrids' would seem an obvious choice for the group as a whole. However the word hybrid carries with it certain implications which do not necessarily apply here, particularly when it comes attached to a (wild) species name. (The use of the terms 'species' and 'hybrid' is discussed in Chapter 2).

If some plants are produced by crossing a true plant of P. rockii with another peony, say P. xx it is reasonable to describe them as P. rockii hybrids (describing is not the same as naming). Of course it is equally appropriate to describe them as P. xx hybrids. (It can be argued that which of P. rockii and P. xx is the female parent should determine how their offspring are named but there is neither agreement about nor general usage of this convention and in any case it is not helpful here.) Most Gansu Mudan are not P. rockii hybrids in this precise sense. If subsequently a P. rockii hybrid in this precise sense is crossed with P. xx are the resulting plants still P. rockii hybrids? And so on. Like most groups of cultivated plants of hybrid origin Gansu Mudan include hybrids of hybrids of hybrids......, with many plants of both known and unknown identity involved. (This analysis explains why a name of the form '<genus name> <species name> hybrids' is unsatisfactory in any genus and why it is not allowed in the current rules for plant nomenclature. See also below.) A further complication discussed in more detail in Chapter 3 is that what P. rockii itself is or is not is a far from simple question, especially in the context of past use in hybridisation.

There is a great deal of inaccurate (or worse) and misleading nomenclature involved with this group of plants. Frequently if they are white or pale pink they are called P. rockii for the prestige of this name when, of course, they are simply not species P. rockii. Or they are called something like 'Paeonia Joseph Rock' or 'Rock's Peony' which is (almost always) incorrect for reasons explained in Chapter 5.

Using the Chinese province of Gansu for the name of this group of plants is both accurate and appropriate and has no misleading connotations. True P. rockii is found in other provinces, Sichuan and Shanxi for example but Gansu is where it is mainly found.

Additionally, although Gansu Mudan are now widely distributed in China and there have been examples from ancient times the emergence of the present cultivars as a major new group is effectively due to the work of Chen Dezhong at Peace Peony Nursery in central Gansu.




The term Gansu Mudan was first proposed in 1992 by Professor Li Jaijue of Lanzhou University as a more precise name than the familiar (in China) but imprecise name Ziban Mudan. This last name represents all mudan with flowers that have purple (or dark) blotches and includes the wild species (both the distinct-or-not forms of P. rockii and others and the doubtful P. yananensis), those Central Plains or suffruticosa hybrids that have dark blotches and the mudan cultivars developed by Chen Dezhong. Gansu Mudan as a name, both collective and individual, has several advantages from a Western point of view. It is short, easy to spell and to say and unlike many names for plants of non-Western origin that have been given Western names this name honours the Chinese origin of the plants. So we propose the simple unambiguous elegance of Gansu Mudan for the plants, with Paeonia Gansu Group as the group name for formal taxonomic purposes.

(For people who enjoy this sort of thing we should point out that strictly, what is proposed here is Paeonia Gansu Cultivar Group, a collective name for an assemblage of individual cultivars (sharing some particular characteristic or characteristics) although the word 'Cultivar' is not formally included in the name. However the term cultivar itself allows "an assemblage" of plants (sharing some particular characteristic or characteristics) and the term 'Group' is sometimes used to make explicit the assemblage that is allowed under a single cultivar name. Crucially 'Cultivar', and 'Cultivar Group' both allow inclusion of unnamed (or named) plants that have arisen or will arise in the future with the appropriate characteristics. There is a basic problem here that essentially comes from the difference between the definition of the concept and the definition of what can be recognised as a legitimately published example. Since the word 'group' has a well-known meaning as an everyday word the distinction between 'group' and 'Group' is crucial. It is perhaps unkind to say that a Group of plants is a group of plants that has been validly named as a Plant Group.)

From the point of view of the formal validity of the name Paeonia Gansu Group according to the International Code of Nomenclature for Cultivated Plants four questions or comments arise. Something like Paeonia rockii hybrids cannot legitimately or sensibly be used as explained above. The use of 'Gansu' has occurred previously, possibly without all the formal details some would desire but if there were doubts about whether it has been unambiguously proposed and validly published they are hereby now resolved. The genus name or a transliteration of it cannot be used as part of the name because it involves a tautology. Paeonia Gansu Group avoids that problem but nobody should be expected to say of a plant "This [woody] peony belongs to the Paeonia Gansu Group". "It is a Gansu Mudan" or "They are Gansu Mudan" is so much simpler. When a plant group name is coined initially, the extant members are not always clearly defined, and as more plants are produced there will always be some ambiguity or uncertainty about whether particular plants are members of a particular group. This is not in itself a fatal flaw provided the concept is reasonably clear and the name meets what should be the most important criterion which is to be useful and helpful. Gansu Mudan, the plants that comprise or should be assigned to Paeonia Gansu Group, in one sense define themselves but they are described or delineated in more detail in Chapter 4, by their properties in Section 2 and by their pictures in Section 3.

The contrast between Gansu Mudan and the well-known shrubby peonies, suffruticosa hybrids, is instructive here as in other places. P. suffruticosa, until fairly recently, was an example of the insistence in horticulture/botany of persisting with mistakes in initial publications and nomenclatural designations. P. suffruticosa was described (by Western botanists) as a species of peony on the basis of an imported plant which was actually and definitely a cultivated hybrid. Recently the name P. spontanaea has been partially accepted as the valid (ie first legitimately published) name for the ancestral species from which the suffruticosa hybrids were most prominently (supposedly) derived. (There is a complication with a more recently designated species name, P. jishanensis, see Chapter 2.) If P. suffruticosa had ever been appropriate as a species name and was still a valid name then the objections outlined above to using it for a group of cultivated hybrids would apply with equal force, but now that the familiar, and in some sense established, name P. suffruticosa has been shown not to represent a true, wild species it becomes at least an acceptable candidate as the name for that group of hybrids. (Except of




course that 'suffruticosa hybrids' is formally unacceptable and for most people obliged to use it unnecessarily awkward to say and to spell and meaningless.) All that said, the Chinese name for those plants is Zhongyuan Mudan meaning central plains woody peonies (hybrids) and this would be a more appropriate name, but it is awkward for Western ears and suffruticosa is probably too entrenched to change. Whatever would be most appropriate, it is not our concern here, although when we need to refer to them we will usually call them Central Plains hybrids

1.3 Paeonia rockii

There is no doubt whatsoever that there are populations of wild Mudan with certain characteristics (bi-pinnate leaves with several to many small leaflets, white-petaled flowers with a dark basal blotch, etc.) for which an identifying species name is appropriate. Unfortunately the introduction of 'rockii' in the designation of P. suffruticosa ssp rockii by Haw and Lauener in 1990 compounded the error inherent in P. suffruticosa and involved a false premise. It was, in a large part, based on a cultivated hybrid/intermediate (the Highdown plant, see Chapter 5) and a poor herbarium specimen with no field data to support it and it was (and is still) impossible to determine to which wild plants this new name then applied. It is entirely appropriate that Chinese botanists proposed the epithet P. rockii based on true wild plants, or to put this another way, that the wild mudan involved were regarded as an independent species (raised to specific status). Incidentally, it is only fair to point out that Li Jiajue had already said that this change was necessary in his first book on purple-blotched mudan in 1988. To add to the confusion there are (or at least were) some plants resulting from seed sent by J. F. Rock in 1925 which came from plants in cultivation. While it is conceivable that the seed was true P. rockii seed contrary alternatives seem to us more plausible. Such questions in general may be answerable with molecular analysis but in this particular case this is now impossible. There are very few plants for which there is actual evidence that they germinated from Rock's seed and as we explain in Chapter 5 they are Gansu Mudan and not true P. rockii. Other examples we know of labelled P. 'Joseph Rock ' or something similar

and seedlings propagated from them are also not P. rockii but simply examples of Gansu Mudan.

More recently there has been considerable field work by Chinese botanists and much more and more precise information is available. Unfortunately the Chinese botanists involved have not yet arrived at a consistent consensual formulation, so their valuable work has not yet resulted in a clear picture nor with a set of names which can easily and accurately be applied. The elements which contribute to the complexity of the situation can be separated into two types. Some are familiar to taxonomists, namely, the extent to which species status depends on morphology only (ie appearance) or on phylogeny (ie parentage etc.) or on a mixture of the two. The interaction of these two ideas determines the extent to which different populations are lumped together under a single species name and the extent to which plants of unknown origin can be given an established species name because the appearance of all of them fits within a published description, and on the other hand the extent to which plants of varied appearance can be given the same species name because they are all present in a wild population. Also uncertain is the degree of morphological distinction required for different species or subspecies designation. Such questions are settled in practice by the acceptance or not of a particular taxonomic proposal or scheme. At present there is no consensus but Hong Deyuan's fairly recent revision of the genus paeonia in China published in Flora of China will probably lead to this becoming basically the accepted scheme irrespective of the merits or this or other schemes and in spite of more recent revisions by Hong Deyuan. There is some agreement that there are two forms of wild 'rockii-type' plants, initially called P. rockii and P. rockii ssp linyanshani by Hong Tao et al but now called P. rockii ssp rockii and P. rockii ssp taibaishanica by Hong Deyuan et al (but the other way round) The difference between the two is in the lobing or not of the leaflets. The form most common in the wild, Hong Deyaun's P. rockii ssp. rockii, has mostly unlobed leaflets, although the size and number of leaflets varies considerably between populations (see Chapter 3). (These plants were called P. rockii ssp.linyanshani by Hong Tao because in the paper by Haw and Lauener that raised some purple-blotch peonies




(although it is unclear which) to sub-species status as P. suffiuticosa ssp. rockii their illustrative leaf diagram was of a cultivated hybrid with lobed leaves.) They now seem likely to remain P. rockii ssp. rockii. This classification, largely accepted at present, is in fact much less clear-cut than the various papers and authors supporting it imply and is discussed in more detail in Chapter 3.

In addition to these, perhaps academic, considerations there are further problems. Mudan has been the subject of intensive cultivation and breeding for many centuries, long before the current approaches to classification were established. That most wild populations are now remnant relics increases difficulties but is not itself a great problem. (At least not for 'rockii-type' mudan. For P. spontanaea and so-called P. ostii it seems to us it is a severe problem because the existence of authentic wild ancestral plants is very doubtful.) A much bigger problem for the taxonomy of wild Mudan arises from the long-established use of the root bark (danpi) as a medicine. The fact that danpi has a monetary value has meant not only that wild plants have been collected and destroyed but that Mudan have been planted in wild places, so that the truly wild nature of some Mudan is doubtful and impossible to ascertain. This is particularly the case with some plants in Yanan, Shanxi Province where there are several interesting examples, for some of which some botanists have proposed specific status. There are also wild rockii- type plants with pink and purple-pink coloured flowers. Without careful analysis and compelling evidence to the contrary there is no justification in regarding such plants as other than Gansu Mudan and the direct or indirect result of human activity. Classification difficulties in no way detract from the splendour and horticultural value of the plants involved but in our view they increase the need for an appropriately designated name, namely Gansu Mudan (formally Paeonia Gansu Group) for the group of cultivated hybrid plants involved and make it imperative that the name P. rockii is not applied to plants without appropriate wild provenance.

1.4 Ziban Mudan

In China, this term is applied to all mudan that have dark blotches or marks at the base of the petals irrespective of the colour of the blotch and the petals and of the identity and origins of the plant. Thus a central plains hybrid with dark basal blotches, presumably the result of the presence of genetic material from a rockii-type plant, could be called a Ziban Mudan. Additionally the two species or doubtful species or 'wild hybrids' or 'hybrids, once cultivated, found in the wild' P. yananensis and P. baokangensis are Ziban Mudan. Consequently the term Ziban Mudan is not exactly equivalent to the term Gansu Mudan or to P. rockii (whatever that name means). The term/ expression/word Ziban itself is misunderstood in the West due to a rather simplistic translation. It is the case that zi is the word in Chinese for purple but it is also used in this context to mean simply 'dark'. The plants to which the term Ziban Mudan is applied include a wide range of blotch colour from darkish red through purple to black and in fact the most common blotch colour in Gansu Mudan and their wild P. rockii ancestors is black. In passing we note that some of the plants in cultivation described as P. 'Rock's Variety' or something similar have purple or purplish blotches, casting further doubt (although none is needed) on the validity of their being true species plants. (It is immediately clear from the pictures of Gansu Mudan in section 3 that blotch colour is not a simple matter.) There is no doubt that the development of Gansu Mudan (or the Paeonia Gansu Group) as a whole has involved true wild plants collected by Chen Dezhong, but also Ziban Mudan used in the sense so that the term P. rockii hybrids while undoubtedly having a desirable cachet to Western horticulturists and gardeners is inappropriate.

1.5 Names of cultivated hybrids

There is considerable confusion about English language names for individual Gansu Mudan varieties, where there should be none. Simple adoption of the pinyin transliteration of the Chinese character name is the designated




Gansu Mudan at Peace Peony Nursery

valid approach and is greatly preferable to the incorrect practice of choosing Western names for commercial purposes. Using the piny in version of the Chinese name is straightforward and legitimate and requires only that the remaining inherent reluctance to do so be overcome. To encourage this appropriate approach we have given this topic more attention than is usually the case, see Sections 3 and 4. Of course for a genuinely new cultivar or an established clone without a validly published name (see Chapter 5) different considerations apply. There are also complications arising from the use of Western colour names and the much more general question of accurate identification, discussed in chapters 9.1 and 9.3.

Apart from these questions about the cultivar or variety name one thing that is clear is that the word 'rockii' should not be a part of it. If the cultivar name is, say, Hong Lian then the correct name for the plant is Paeonia 'Hong Lian' or Peony 'Hong Lian'. Peony rockii 'Hong Lian' is invalid and incorrect because Gansu Mudan are not selections from true P. rockii.

It may be helpful to point out here that the International Code of Nomenclature for Cultivated Plants explicitly allows in cultivar names both the latin form and the common name of a genus, so Paeonia and Peony are both valid and are to be regarded as equivalent.




Before discussing mudan species and hybrid groups it is necessary to review some of the terms involved and to point out that this is not intended as an in-depth academic discussion, but to provide enough basic clarification to render the remarks about mudan categories useful and unambiguous and to increase insight through awareness of the difficulties instead of, as is often the case, increasing misunderstanding by glossing over them.

By species plants we always mean wild plants whose form and characteristics are the result of natural evolution uninfluenced by human intervention in any direct sense. By cultivated plants or groups of cultivars or hybrid groups we mean plants that do not naturally occur in the wild and whose existence is the result of human activity, deliberate or otherwise. For cultivated plants the term Group now has a formal meaning roughly equivalent to that of species for wild plants. While each of these concepts is simple neither is completely easy to use in practice. Among botanists, and horticulturists and media people who write or comment on such matters everyone knows what is meant by a species but there is no agreed definition and everyone knows something different to everyone else. Bear in mind that wild plants are not intrinsically divided into different sets, each called a species. In practice a particular set of plants is a species if botanists/botanical taxonomists on the whole agree or at least accept that it is. This bizarre situation can easily be defended and certainly has substantial advantages but it is at the same time profoundly unsatisfactory. Perhaps we should say it is profoundly unsatisfactory to a scientific frame of mind but uncontro-versial to, say, a theological frame of mind. At the heart of the problem is the question of whether species is a phylogenic concept (ie determined by parentage/ancestry) or a morphological concept (ie determined by appearance). In many cases the two do more-or-less coincide, which, paradoxically at first sight, is largely why there is a problem.

Most formal definitions of the species concept involve phylogeny much more than morphology, but almost all classification (formal definitions of different species) is done in terms of morphology, for fairly obvious reasons. Comparisons with animals are misleading: for familiar animals the idea of species is more like the idea of genus in (familiar) plants. For higher levels of classification there is only rarely controversy; most plant genera are defined quite well by morphological characters and in most cases members of one genus will not breed with members of another. Also for lower levels of classification, such as form, the criteria are simply morphological and often not maintained in offspring.

In between the extreme and well-defined examples of wild plants and cultivated plants there is a large complicated grey area. Of course a wild species plant that is collected and grown in a garden remains the same wild species plant. If it then produces seed the plants grown from that seed should generally not be properly regarded as true examples of the species in question (unless the species is apomictic). In practice they will almost always be labelled with the species name either innocently or because the seed has been produced for commercial reasons but they will almost always not be true examples of the species but cultivated hybrids as they result from fertilisation of the true species plant (determined by its provenance!) by pollen from another plant which is not a true species plant. There is scope in cultivation for rigorously ensuring that a true species plant is fertilised only by another such plant but this is rarely what happens. In such cases and of course in a wild population the offspring may not be homogeneous in appearance and may not exactly resemble either of the parents but they are by definition part of the species population and are part of its morphological diversity. The appearance (morphological) of offspring is, in a sense, irrelevant because when plants from two




different (accepted) species are crossed some or all of the offspring may look exactly or more-or-less like one of the parents or like an example of a third (accepted) species. A further complication quite often occurs when 'species' plants are grown from seed in cultivation. Assume that the original parent plants are true examples of the species and there is persistent and rigorous prevention of external fertilisation. If extreme examples are repeatedly selected (that exhibit some particular feature or colour for example) then although all can be regarded as true species plants successive generations become increasingly unrepresentative of the wild population that defines the species. Indeed rather few generations may be needed to produce examples that cannot be (or at least have not been) found as wild plants and which do not satisfy the morphological description of the species. An exactly parallel problem arises with a named cultivated plant. Its offspring from seed will in most cases vary and differ from the parent whether it is self-pollinated or not. The extent to which such offspring can properly be labelled with the culti-var name or cultivar group name depends to some extent on how the name has been defined but in most cases the offspring should not be given the parent name. Almost always of course they are, in both cases. It is perhaps unkind to point out that the detailed and complex (and admirable) rules on plant nomenclature which define whether or not any particular name for a species or a cultivar group can be valid nowhere involve a definition of the two concepts, but necessary because no outsider would believe such a thing.

The word hybrid is widely used and its meaning assumed to be clear. It is nevertheless a problematic concept and is generally used in an unhelpfully loose way, particularly when a name is attached to it. The implication is that two plants with differences in genetic make-up have combined sexually (by whatever means) to produce offspring, and because the parents are genetically different the offspring as a whole will be a mixture of and different to both. Some plants are self-infertile (there is evidence that (some at least) mudan are) so that seed can only be produced by two genetically different parents and so in a sense all resulting offspring must be hybrids. However when the parents are both from the same species (the implication being that there are sufficient genetic differences for fertilisation but not enough to render them different species) then the offspring are not hybrids but more members of the same species. This leads to perhaps the most common element of definitions of the species concept, that its members breed true within the prescribed limits of morphological variation. This satisfying idea is routinely abused in practice because the limits are necessarily defined by the wild population(s) and botanists are rarely able to do sufficient fieldwork to accurately determine them. Most cultivar plants are hybrids in the sense that they result from seed where not both the parents are true examples of the same species although some are simply exceptional/abnormal selections from species plants. Most have very complex ancestry and are hybrids of hybrids of hybrids... . Two further observations add to the complexity of the situation. A truly wild plant will usually be referred to or regarded as a hybrid if its appearance is intermediate between that of two (designated/accepted) species. Its hybrid origin is actually an historical event that is assumed but not known and even if somehow the 'cross-fertilisation' could be established the term hybrid only follows because the two parents are regarded as members of different species. The current practice (unchanged since Linnaeus 1707-1778!) is to assign such plants to one or other of the two species or to a new species or to argue that the existence of such plants necessitates the combining of the two species to one with a broader morphological range. Much the same applies to cultivated 'hybrids' albeit with slightly different 'parameters'. A general adoption of the term 'intermediate' both as a description and as a designation and much more careful and restricted use of the term 'hybrid' would have much to recommend it. Accurate, as opposed to merely confident, use of all the concepts involved requires knowledge of the immediate and distant history of the plants involved and for most plants that most people encounter this is absent and rarely remedied by the attachment of a label. In practice the term 'hybrid plant' effectively means 'not a (wild) species plant', which is not really the same thing. Offspring of a self-fertile parent may be hybrids in one sense (for example if the parent is itself, for the sake of argument, a hybrid) and may well have varied morphology but are not hybrids the in the sense defined above. In many cases the name adopted




involves a species name and this makes already unsatisfactory conventional usage worse because the involvement of the species cited may be either many generations earlier or simply erroneous. As there is no accepted term other than hybrid to imply 'not a true species plant' in the sense we have defined it and as certain groups of plants are known as so-and-so hybrids we are obliged to make use of the term.

The complications outlined above with what are superficially simple but actually very complex concepts apply in varying degrees to most plant genera. With the mudan section of pa-eonia they are compounded by three extra factors, one of which is unique to mudan.

The first is that Western (and some Chinese) botanists have defined species and coined names for them with very little background information and experience and little or no field-work, and in most cases based on single or very few specimens often dried and in herbaria some of which were in fact cultivated hybrids. P. papaveracea and P. suffruticosa are the classic examples in the present context. This has been standard practice and continues today. Where the plants involved are obscure and come from homogeneous populations it is, perhaps, defensible and may well be helpful. In the case of mudan the plants have been deeply involved in and cultivated in a civilisation much older than that of the western botanists and the consequence has been endless inaccuracy and confusion. There are more details of the basic history and the errors outlined in Chapter 3, and on so-called 'Rock's peony' in Chapter 5.

The second involves more recent studies. Chinese botanists have investigated wild mudan and the origins and characteristics of cultivated hybrids and published many papers with, not unreasonably, only summaries in English. Unfortunately the aggregate of these extensive and very welcome publications, scattered across various academic journals, is characterised by contradictions and competition between eminent authors and by frequent revisions. Chinese botanists have the particular handicap of having to accommodate as best they can the names and taxonomic mess left by earlier Western botanists, and continued in some recently published books. Such comments could be made about most plant genera endemic to China and are made here as matter of fact and not as derogatory criticism (well, mostly not).

At the present time an agreed accepted overall description of species and hybrid mudan has not been reached but the basic features outlined below seem secure. Further field-work and more extensive molecular studies will undoubtedly lead to further revision but the main elements of the picture presented by, for example, Hong Deyuan in the new Flora of China in 2001 based on morphological studies are likely to prevail if only because this is the most readily accessible and most authoritative reference. This in spite of the fact that he has already since published articles which modify his account there. It is corroborated more-or-less by a subsequent molecular study and does not differ greatly from an earlier survey. For more detailed accounts of the genus than that presented here the five references below, are together perhaps the best starting point and the gateway to many others, and no doubt there are many more to come.

The third complication is exceptional and has far-reaching consequences. For hundreds of years the root bark of woody peonies has been a significant element of traditional Chinese medicine. It is called 'danpi'. (Here pi means skin or bark and dan does not mean simply red but signifies something of value.) Unlike, for example, the bark of Quercus suber which can be harvested for cork products without killing the tree the root bark of shrubby peonies cannot be harvested easily without destroying the plant. The obvious and actual consequence has been the destruction of many wild shrubby peonies. We are not able to be precise about the extent but it is certainly so great that ancestral species of some hybrid groups are, as far as is known, now essentially non-existent. (See, for example P. spontanea below). The implications for mudan taxonomy are obvious and are a significant factor in much of the present confusion and controversy. In places where species mudan once-grew there are none and for some species the places where they can still be found are obscure and relatively inaccessible. But another serious consequence of the monetary value of danpi is that mudan have been planted in wild locations by people without their own land. As this would




above left and right: Various forms of Paeonia delavayi

generally have been secretive the locations are unknown and the extent to which it has happened impossible to guess. As this practice has likely taken place over the many hundreds of years during which mudan have been gathered from the wild and actively cultivated and hybridised, the inferences that can be drawn from wild populations (particularly small ones) will rarely be as secure as botanists would wish them to be. We should point out that several authors have published confident, eloquent even, revisions of Mudan that do not share our reservations. Such expositions are certainly more comforting (to the uncritical reader) than many of our observations.

With these preliminary observations to be borne in mind we can now present a basic overview of the mudan section of the genus paeonia. Simple alphabetic lists of species names, and even binary keys, are the standard way to describe a genus and an easy option for the writer. However they are less helpful to the reader than they appear to be because they do not provide an insightful overall perspective and their linear nature gives all names equal weight. As the purpose here is not a comprehensive exposition but a basic, pragmatic summary a less formal approach is more appropriate.

While it is crucially important to be and to remain clear about the distinction between wild species plants and cultivated plants the most helpful way to approach mudan as a whole is to begin by separating them,




Paeonia ludlowii

rather crudely, into different groups each consisting of certain species and certain cultivated hybrids, although in truth very little can be described as certain. (Here the term group is used in its general sense and not with the particular meaning defined (well actually not explicitly defined) in the International Code of Nomenclature for Cultivated Plants. It is used there partly to avoid the complications arising from the ambiguous term hybrid.) Also, although there are the four distinct groups of plants outlined below, each centred on a particular species and geographical area, plus a fifth for convenience, this does not mean that every particular woody peony belongs unequivocally to one of them. A direct hybrid between two of the base species could be regarded as belonging to both the corresponding groups. As almost all the hybrid plants have a much more complex, distant and usually unknown, species ancestry each group is an imprecise but nevertheless useful aggregate.

Delavayi group

The simplest group conceptually is the complex species or complex of species called Paeonia de-lavayi and the relatively few cultivated hybrids derived from them. In the past several species or sub-species have been described, again based on relatively few examples. Now the woody peonies that had potaninii, lutea, trollioides etc. as part of their name have been subsumed (by some authors) into the greatly variable species P. delavayi. Individuals range from a half to two metres in height, with generally small flowers from yellow to black-red. Leaves have generally fewer leaflets than in the other groups but individual leaflets are generally more lobed and dissected. Plants in this group show varying degrees of stolonicity in their growth. In spite of the great differences they are regarded as a single species because most known wild populations are heterogenous with a continuous spectrum of different forms. Another species in this group is P. ludlowii, a larger plant. It was previously designated a sub-species of the now subsumed P. lutea although it is morphologically and geographically distinct. In China this group of species and forms has not been greatly involved in breeding cultivated varieties, presumably because the flowers are small and probably because they do not cross readily with other groups. This is regrettable because most of the plants in this group are frost hardy and this character could have been developed in those other groups of hybrids in which are not so reliably hardy. Plants of the P. delavayi complex and P. ludlowii have been used more in Europe and the United States in the creation of hybrids and although the names recorded for the plants used should be treated with caution this is probably why yellow flowered examples are much more frequent than in the groups of Chinese hybrids. In fact the familiar, P. ludlowii, called large yellow mudan in China, regarded as totally hardy in Europe, requires protection to survive the fierce winters of, for example, Lanzhou in Gansu.

Suffruticosa group

The plants of the most familiar group have been known in the West for many years as 'P. suffruticosa hybrids'. The name derives from one of the earliest mistakes by Western botanists who coined the species name for a double-flowered cultivated hybrid and the name has caused confusion ever since. Subsequently other mudan, now regarded as distinct species, were assigned the status of sub-species of P. suffruticosa. This is the main group of cultivated Chinese mudan and together they represent many hundreds of years of experiment and development. There are now over 1000 named cultivars, which may be forms or clones, and there are many books of selections. ,Chinese Tree Peony' edited by Wang Lianying is arguably the best. The Chinese name for this group of plants is




P. jishanensis type without wild provenance; the faint purple pink colour at the base of the petals is more pronounced when the flower first opens. Paeonia jishanensis

Central Plains Mudan, reflecting the main area of development, now most strongly concentrated on Luoyang in Henan and Heze in Shandong. Various sub-groups such as the South-West cultivar group, are recognised, albeit overlapping and imprecisely delineated. The long history of these plants is understandably very erratically recorded despite their great importance in past and present Chinese culture, so various mysteries and uncertainties about their origins are only to be expected. In particular the ancestral species most prominently involved is unclear and the most likely species is now very uncommon. This is P. jisha-nensis, (called short mudan in China) which seems now to be accepted as a distinct species and does at least have a Chinese name. It may well be the same as P. spontanaea as has been argued, even though it is basically white-flowered and P. spontanaea was named for a red-flowered plant, possibly not a true wild plant. This itself is strange because there are old reports of hillsides of red-flowered mudan but none can be found now. Their one-time existence is at least plausible because red (pink through red to claret-purple) is the dominant colour in the Central Plains group. Presumably we have to accept that the wild red-flowered mudan have been collected to extinction. Indications of the involvement of several now-recognised species can be seen in the Central Plains group (and in any group of hybrids) and while indications are not the same as knowledge it is hard to believe that any wild species can have totally avoided making a contribution during the centuries of work on this group.

Feng Dan group

A third group is Feng Dan peonies, also known as the Southern Yangtse group. These plants are the most prolific source of danpi because their roots, although relatively few, are thick and fast growing and with a thick and easily removed root bark. They are cultivated on a vast scale in parts of China south of the Yangtse river for the production of danpi. The plants are morphologically fairly distinct from other mudan and usually white flowered (Feng Dan Bai). The white flowered plants were proposed as a species in 1990 by Hong Tao and with the name P. ostii. Both proposals were bizarre. The assumption that all the plants commercially cultivated, both overtly and covertly, for centuries must have had a wild ancestor is not unreasonable, but whether any truly wild plants have been found is uncertain. Also




A 'wild' form of P. 'Feng Dan Bai' found in Shennongjia.

the choice of name seems perverse unless either largesse or a subtle sense of humour is involved. As it is all the plants grown in farms and gardens that once had the traditional and well-known (in China at least) name of Feng Dan Bai have suddenly become species P. ostii plants. They have apparently not been much used deliberately in the development of overtly ornamental hybrids. Pink flowered forms exist (Feng Dan Fen) and purplish-pink forms (Feng Dan Zi) and white with a purplish-pink centre. Some examples of Central Plains hybrids and of Gansu mudan have morphological features that seem clearly derived from Feng Dan peonies.

Gansu Group

The fourth group is Gansu Mudan, also sometimes referred to as the North-Western group. Here the crucial ancestral species is P. rockii in one form or another. There is evidence that Lanzhou, now a largely industrial city was in ancient times a centre of mudan production. It lies on the Silk Road so mudan are likely to have been taken to and from there. The appearance of some central Plains hybrids show some of the characteristics of Gansu Mudan. They have a dark area at the centre of the flower, a stain or smudge or flare if not the pronounced and striking blotch that is a distinctive characteristic feature of most Gansu Mudan. A few Central Plains hybrids are relatively tall and some have compound leaves quite like those of most Gansu Mudan and some are more hardy than others so the genetic influence of P. rockii or Gansu Mudan seems likely. The emergence of Gansu Mudan as a substantial and important group of cultivated hybrid peonies is a relatively recent phenomenon and is essentially due to Chen Dezhong's passion and inspiration. As such, although there are (or were) some ancient culti-vars that can be assigned to this group, we can say with certainty that most of the earlier modern cultivars in this group are the result of deliberate crossings of true wild P. rockii (collected as simply Ziban Mudan, purple-blotched mudan) with central plains mudan. Subsequent cultivars use the earlier cultivars among the parent plants so we have the usual situation where the most recent additions are hybrids of hybrids of hybrids etc. The development of this group and the questions surrounding the various names attached, accurately or loosely, to plants in this group, notably P. rockii and Gansu Mudan but also papaveracea, linyanshani and taibaishanica are discussed in more detail in Chapters 3, 4 and 6.


The fifth group has no botanical or morphological coherence. It is the remaining identified mudan taxa grouped together simply for convenience and to give them an appropriate perspective in mudan as a whole, in particular the apparent absence of significant influence by them in the major hybrid groups. There are two accepted and significant species, P. szechuanica/decomposita and P. qiui.




Small plants of P. szechuanicaldecomposita with short, broad leaflets regarded as ssp. rotundifolia and on the left parts of leaves where the leaflets are intermediate.

P. szechuanicaldecomposita. An example with narrow leaflets.

above and below: Baokang Mudan: plants found in the wild but presumably of cultivated origin.

Paeonia qiui

P. qiui type without wild provenance.

It is, in our view, greatly to be regretted that eminent Chinese botanists have mistakenly interpreted the rule of precedence in botanical nomenclature as obliging them to replace the fairly well established name P. szechuanica by the dreadful name P. decomposita coined by Handel-Mazzetti in 1939. Szechu-an is the earlier Western spelling of Sichuan, the province where this species is mainly found. It has single pink flowers and is most notable for its complex leaves with very small individual leaflets. As far as we know it has not been used deliberately in the development of the cultivated hybrids but the small leaflets of some Gansu mudan suggest that it may possibly have been involved. Some examples with shorter, more rounded leaflets are designated ssp. rotundifolia. P. qiui seems to be less rare in the wild than was first believed, although many of the few plants in cultivation have the appearance of hybrids with the species as it has been described.

There are also several minor proposed species all of which have intermediate morphology between plants of the species or groups already mentioned. These include P. yananensis, P. baokangensis and P. ridleyi. There is at least one significant and apparently wild population of each of the first two but whether the plants 'discovered' are regarded as natural hybrids or intermediates or are cultivated escapes depends mostly on the point of view of the author concerned and unless further significant populations are discovered it hardly matters.

Basic references (see also Chapter 5) that will lead to many others:

Deyuan Hong and Kaiyu Pan (1999). A revision of the paeonia suffruticosa complex - Nordic Journal of Botany 19 (3) pp289 - 299.

Wu, Z., Raven, P. H. and Hong, D. Y. Hong (eds) (2001) Flora of China, Vol. 6 (Paeonaicae), Science Press, Beijing & Missouri Botanical Garden Press, St. Louis.

Zhao Xuan, Zhou Zhi-qin, Lin Qi-bin, Pan Kai-yu, Hong De-yuan (2004) Molecular evidence for the interspecific relationships in Paeonia sect. Moutan: PCR-RFLP and sequence analysis of glycerol-3-phosphateacyltransferase (GPAT) gene - Acta Phytotaxono-mica Sinica 42 (3) 236-244.4

Xiangyun Zhu &Tau Hong (August 2005) Validation and neotypification of Paeonia rockii subsp. linyanshanii (Paeoniacae) - TAXON 54(3: 806-807

Wang Lianying et al (1997) Chinese Tree Peony, (English edition) China Forestry Publishing House.




3.1 The evolution of the name

3.2 A compromise proposal

3.3 Observed Characteristics and Distribution of Paeonia rockii

3.4 Evidence from Molecular Analysis

3.1 The evolution of the name

It is crucial to bear in mind that in China the wild plants, mudan with purple-black blotched white flowers, have been familiar plants for centuries. There, the different forms of what are now called P. rockii and other basically similar plants, which may or may not be other species or hybrids introduced into the wild, are known collectively as Ziban Mudan and distinctions that may be made between individual examples are done

so partly on the place of origin. The following summary of the emergence of the name P. rockii is a summary of the application of Western botanical nomenclature by first Western and latterly Chinese botanists.

Bear in mind also that the emergence or establishment of a species name in literature is not the same as the identification and classification of plants in the wild or in cultivation.

1804 Andrews: P. suffruticosa. The new species name given to a cultivated hybrid (without blotched petals, actually a Central Plains Mudan).

1807 Andrews: P. papaveracea. The new species name given to a cultivated hybrid (with blotched petals, actually a Gansu Mudan).

1890 Pratt: specimen, presumed wild, collected; no field notes. This important but unfortunately

ambiguous specimen appears to be the earliest example known of wild collected (if it was) mudan, at least in a herbarium (Kew) in the West. Stern later accepts this specimen as an example of P. suffruticosa with 'leaflets deeply and incisely divided'. Some leaflets do have a small tooth. This specimen probably is an example of P. rockii, but may be a Feng Dan plant or a hybrid.

1914 Farrer: wild plants (of P. rockii) seen and described (extravagantly, but with inadequate botanical detail) and regarded/referred to as a form of P. suffruticosa.

1922 Licent: presumed wild specimen collected, apparently of P. rockii (but originally named P. obovata).

1925 Rock: sees plants in cultivation, unidentified and now unidentifiable but very likely examples of both Gansu Mudan and Central Plains Mudan. Seed from these plants, in two different locations, is not actually collected by Rock but is sent by him to the Arnold Arboretum and subsequently distributed.




1936 Stern: receives from Cleveland Morgan, Montreal a plant derived somehow, but probably

indirectly, from Rock's seed, and which is an example of Gansu Mudan, not a wild species plant in any sense.

1946 Stern: in his famous monograph discusses 'forms of P. suffruticosa', makes some incorrect

statements about Rock's seed and uses the expression 'Rock's peony', but not as a formal name.

1959 Stern: refers to his plant as P. suffruticosa 'Rock's variety'.

1971 Reath: lists grafted plants of 'P. suffruticosa (Rock's)' propagated from material received (directly or indirectly) from Stern.

1988 Li Jiajue: in his book Linxia Mudan explains why the purple-blotched wild mudan should be separated from the suffruticosa complex but assumes, incorrectly but understandably, that the plant named P. papaveracea is an example.

1990 Haw and Lauener: propose the name P. suffruticosa ssp rockii for Stern's plant and for Farrer's

herbarium specimen (and other herbarium specimens) erroneously assuming they are the same. The source of their leaf drawing of P. suffruticosa ssp rockii, with lobed leaflets, is not given but it is apparently of Stern's hybrid plant. Farrer's specimen is incomplete but on balance, and maybe with hindsight, has unlobed leaflets. Curiously they do not even mention Pratt's readily available specimen.

1992 Hong Tao: proposes P. rockii as a separate species, but describes the leaflets as lobed, presumably influenced by Haw and Lauener's error.

[correction Burkhardt 2006:

1993 Pei Yanglong proposes P. rockii ssp lanceolata as a new subspecies of P. rockii with unlobed leaflets in his dissertation]

1994 Hong Tao: proposes P. rockii ssp linyanshanii as a new subspecies of P. rockii with unlobed leaflets.

1998 Hong Deyuan: reverses Hong Tao's formulation and names the form with unlobed leaflets P. rockii ssp rockii and the form with lobed leaflets P. rockii ssp taibaishanica, ostensibly correcting Haw and Lauener's error and Hong Tao's development from it.

2005 Zhu Xiangyun and Hong Tao: rummage in the nomenclatural entrails and introduce the unusual spectacle of a mysteriously missing herbarium specimen (holotype) in an attempt to resuscitate linyanshanii as a subspecies name.

This bare outline omits many other contributions from these and other authors. Almost all share three characteristics: a selective use, possibly inadvertent, of what evidence is available; fieldwork that is inadequate for the complexity of the reality or is superficial or nonexistent; an apparent desire to present a simple picture without the qualifications, uncertainties and caveats that are needed for an accurate picture. We might add an absence of sustained experience of growing the various plants involved in cultivation where continuous direct comparisons can be made free of environmental variation. So it must be said that Chen Dezhong, with his substantial knowledge and experience of wild plants could have usefully contributed more than generously helping researchers and visitors.




Wild Paeonia rockii flowers from various locations, Xinglong, Tianshui, Tianshui




above and right: Hu Xiaoling and flowers of linyanshanii form of P. rockii at Phedar Nursery.

above, center and right: P. rockii fromTaibaishan: flower, leaf and fruit.

Left: Small plant of P. rockii in Tanshui (E): note both blue-green and yellow-green leaves; also strong new growth and flower/ fruit despite the adverse growing conditions.

Right: P. rockii in Tanshui (E) showing new shoot from the base.




P. rockii in Taibaishan showing new shoot from the base.

Linyanshanii form of P. rockii (from Shennongjia, Hubei) growing at Phedar Nursery, in mid April.

As we have already emphasised, the overall situation for the classification of wild mudan is genuinely difficult. The remnant wild populations are widespread and not easy of access. That is why they have survived.

That Said it is disappointing that so often authors have indulged in the traditional shortcomings of botanical taxonomy when the obvious complexity demands much greater care. The herbarium specimens, of which too much is made, never have information about the size of the population they are taken from nor about its homogeneity and the extent to which the specimen is truly representative. This information may well be difficult to obtain but even comments pointing out its absence might induce some much needed caution. Some are from plants in cultivation. A striking feature of the herbarium sheets at Kew is the annotations by subsequent botanists about the name of the taxon involved. In many cases they seem to us correct, in some sense, but no reasons are ever given and in some cases the critical evidence cannot be discerned from the actual specimen, filament colour for example. The botanist making the annotation may never have seen a wild plant let alone have experience of the source population but nevertheless feels able to comment on a wild-collected specimen. This demonstrates a familiar trait, a punctilious regard for the minutiae of nomenclatural rules (absent from some places in this text) and much less regard for solid, consistent, interlocking botanical evidence. Careful study of much of Hong Deyuan's extensive output when coupled with some experience of growing examples of the plants in question reveals superficiality and a lack of consistency. Arguments put forward in one context are not made elsewhere where they are equally valid. A particularly relevant example concerns plants with intermediate morphology. Lumping together all the very diverse plants of the P. delavayi complex because homogeneous populations have not been found and there is a continuous spectrum of flower colour and leaf and plant form is at least reasonable. But then the separation of two subspecies of P. rockii and of P. decomposita/szechuanica on the basis of differences in leaflet form when no evidence is given of population homogeneity and when intermediate leaf forms can be easily found seems perverse. Different criteria for species designation in different genera is unavoidable but within the same genus it begins to look like the taxonomy of Babel. Perhaps if less of Hong Deyuan's fieldwork had been done by his students and he was less determined to refute Hong Tao's work his own work on mudan taxonomy would be more consistent and more convincing. (Incidentally, other observers have reported seeing some homogeneous populations of P. delavayi.)




Wild P. rockii leaves from various locations with line drawings of leaves and herbarium sheets for comparison.

Tanshui (E), Wen Xian, Tianshui (E), Taibaishan, Baokang

Hong Deyuan'sline drawing of P.rockii ssp. rockii leaf

Hong Deyuan's line drawing of 'P. rockii ssp. taibaishanica' leaf.

A herbarium sheet at Kew of P. rockii from Baokang

The herbarium sheet at Kew of Pratt's specimen.




The herbarium sheet at Kew of A herbarium sheet at Kew of G. Licent's specimen. P. Baker's plant.

Two of the herbarium sheets at Kew of Stern's plant at Highdown.

The herbarium specimen at Edinburgh assumed to have been collected by Farrer.

Two herbarium sheets at Kew of P. rockii from Shaanxi and Henan.

The notes on Farrer's herbarium specimen at Edinburgh.

Above: Four leaves of P. szechuanica l decomposita showing extreme and intermediate forms.




3.2 A compromise proposal about Paeonia rockii.

The present summary overview of ziban mudan growing in wild places: 'two subspecies of P. rockii based solely on different leaflet shape and all others regarded as hybrids in one sense or another' seems to us unsatisfactory because we know of wild populations where the leaflet shapes are neither one form nor the other. However it will probably be accepted for the foreseeable future until a concerted programme of much more thorough fieldwork and population analysis can be carried out coupled with equally thorough molecular analysis.

We suggest the following compromise proposal for the species Paeonia rockii as a helpful and more accurate representation of the reality and which enables the main protagonists to depart with their 'honour' intact: Paeonia rockii is defined and characterised below (and in the illustrations) but within the varied morphology two 'extreme' forms can be distinguished: linyanshanii as described by Hong Tao with all or almost all leaflets unlobed, and taibaishanica as described by Hong Deyuan with all or almost all leaflets significantly lobed.

3.3 Observed Characteristics and Distribution of Paeonia rockii

The description and comments that follow, and the pictures of herbarium and live true wild specimens, are not presented as a revision. They illustrate the comments above and in some cases add to what is 'well-known'. The description of P. rockii is based entirely on wild plants that we have seen. The locations and the allied comments are all and only based on first-hand experience. None are extracted from other publications so in that sense they are self-evidently an incomplete part of the whole picture.

Wild P. rockii plants have single flowers at the tips of branches/twigs with two, or very rarely three, whorls of petals. Normally there are 10 petals called, if necessary, primitive petals. Occasionally there are flowers with 11 or 12 petals. The petals are thick, usually almost round or heart-shaped, sometimes fan-like, about 7-11 cm long and 6-10cm wide. In cultivation and growing in 'better' conditions, the number of petals of the flower sometimes increases. Instinctively the question 'How frequently is sometimes?' arises and the best answer we can offer is 'Certainly not often enough to be expected but often enough not to be a great surprise'. A more precise answer needs many authentic plants from many locations and several years of testing and observation. The flower colour is white and at the base of the petals there are distinct but variably sized blotches, usually purple-black. The filaments and the stigma and the sheath are white or cream-white or very faintly green-white. These characteristics are critical. The sheath initially encloses or almost encloses the carpels and recedes as they develop. Leaves are bi-pinnate with 15 to 40, rarely up to 70 leaflets. Leaflets are mid- to dark green, possibly bluish green above, below they are grey-green with sparse fine hairs. Leaflets (collectively) vary from almost entirely lobed or toothed to entirely unlobed or entire. The plants can be over two metres tall.

This basic description needs significant qualification. The flowers may be pale pink in bud and open to white, but this, and the leaf colour, does vary with the environmental conditions. The filament and sheath colour is critical, but of course there are plenty of Gansu Mudan with white filaments and a white(ish) sheath. Leaf structure varies from leaf to leaf, with usually less developed leaves with fewer leaflets nearer the tip of the branch/twig. In many cases there are complex or ambiguous leaflets with the appearance of two or three simpler leaflets fused together. Leaves tend to be larger when growing conditions enable the plant to grow vigorously. In shady situations there may be fewer, larger individual leaflets. Shoots from the base, sometimes said to be absent in true P. rockii, do occur in wild plants, albeit not as prolifically as with some Gansu Mudan. In some areas Ziban Mudan with pale to dark pink or bluish-pink flowers have been found in apparently wild situations.

Wild Ziban Mudan (Paeonia rockii) are found in the mountain areas 1,000 to 2,800 metres above the sea level in the mid and western parts of China. There are traces of wild Ziban Mudan in regions such as Song County in Henan in the east, Qilian Mountaneous areas in the north-west (Lanzhou, Lintao and Linxia), woodlands in Shennongjia in the south and high mountain woodlands in Ganquan County, Yan'an in the north.




Diagrammatic map of China with place names mentioned indicated.

1) Wild species Ziban Mudan (P. rockii) in Xinglohg Mountain (Maxian Mountain), Gansu: leaves have about 15 leaflets mostly well separated and unlobed, the underside of leaflets is noticeably hairy; white flowers. These P. rockii are one of the parents of the earlier 'modern' Gansu Mudan cultivars.

2) Wild species Ziban Mudan (P. rockii) in Wen Xian region, Gansu: leaflets variable, mostly about 25, unlobed, sometimes small and narrow. Leaves can, rarely, have as many as 70 leaflets; white flowers.

3) Wild Ziban Mudan in Yan'an region, Shaanxi: leaves have about 30 leaflets, mostly unlobed; flowers can be white or pink or reddish.

4) Wild species Ziban Mudan (P. rockii) in Shennongjia, including in Baokang, Hubei: leaves have mostly about 30 leaflets usually long and narrow, mostly unlobed, occasionally with small usually blunt lobes or teeth; white flowers; flower sepals and bud are slightly big and thick. Other much more ambiguous plants are found in this region.

6) Wild species Ziban Mudan (P. rockii) in Tianshui in S. Gansu): leaves with up to thirty leaflets, mostly unlobed to variably mixed lobed and unlobed, white flowers.

6) Wild species Ziban Mudan (P. rockii) in Taibaishan, Shaanxi: leaves with mostly about twenty leaflets, mostly clearly lobed, white flowers.

In these summary descriptions 'white flowers' does not preclude flowers briefly pale pink on opening, and the number and form of leaflets is always complicated by leaflets, especially the terminal leaflet, which appear to be two or three simpler leaflets shortly fused at their base.



034 035

Paeonia Gansu Group 'Qi Lian Nong Xia'


4.1 On the term 'Mudan'

4.2 Historical Development of Gansu Mudan

4.3 Gansu Mudan: Current State of Propagation Horticultural Properties

4.4 Comparison of basic morphological features of plants of P.rockii, Gansu Mudan and Central Plains (suffruticosa) mudan

4.1 On the term "Mudan'; in particular on 'MUDAN' and/or SHAOYAO and/or PAEONIA

We think there is no doubt that shaoyao means peonies or paeonia in general and herbaceous peonies in particular1 and mudan means woody peonies. However it is not a simple matter of translation because Chinese characters/prose can not be treated like words/prose of eg another European language, and in what might be called commercial literature nowadays the two words in pinyin are not always used carefully.

There is precise evidence in Professor Li Jiajue's 1988 and 1999 books (in Chinese) 'Linxia Mudan' and 'China Mudan and Shaoyao'. The first part of the main text of the latter is a careful review of references to peonies in ancient Chinese texts. His first sentence translates as 'Mudan and shaoyao are both plants of shaoyao belonging in shaoyao category /office'. In other words 'Mudan and shaoyao are sections of the genus shaoyao (paeonia) in family shaoyao (paeonaicea)'. Note that 'shaoyao belonging' is to be read as a single entity like 'shaoyao-belonging' not as 'shaoyao that belong to'.

In his evidence on the emergence and meaning of 'mudan' Li Jiajue refers to several ancient texts. A critically important one is the famous 'Compendium of Materia Medica' by Li Shizhen in the Ming Dynasty (1368-1664). In it he says Mudan was named because it "is mostly red.

It bears seed and produces shoots from its root. So it is referred to as Mu, the colour of the flowers is red so it is called Dan", Mu in this context meant 'it could be propagated with nutrient'.

Zheng Qiao in the Song Dynasty (960-1279) wrote in his 'Tong Zhi: Kun Chong Cao Mu Lue' (General Records: on Insects and Grasses (herbs) and Woody Plants) "What has been named woody shaoyao in olden times and today is mudan. As for mudan, its flowers are lovely like shaoyao, its root is woody, so it got its name as woody shaoyao. Mudan did not have a name at the beginning so it got its name from shaoyao."

Zheng Qiao quotes from Tu Lian Fa' (Methods of Making and Applying/Taking Medicine) by An Qisheng (Qin Dynasty 221-207 BC) who said "there are two varieties of shaoyao: one is jin shaoyao and the other is mu shaoyao. The jin shaoyao is white in colour and contain fat; the woody shaoyao is purple in colour and has a lot of veins. This can be confirmed by checking the roots." An Qisheng was from Langya (today's Linyi in Shandong Province).

Li Jiajue concludes that "It can be seen that during Qin and Han (206 BC - 220 AD) dynasties, the names, woody shaoyao and mudan, occurred almost simultaneously. This also shows that it has been more than 2000 years since our ancestors realised the medical significance of mudan."




A flower and leaves from the Gansu Mudan at Ness E.G. Cheshire.

The characters for shaoyao include 'pieces' of characters associated with medicine and grass (and spoon!). The current mu of mudan is a rare character with implications of earth/soil and propagation but when used with li, 'muli' means oyster. Part of the character implies male or ox. Dan is also a rare character, literally an archaic word for red (the usual, prosaic word is hong), but which is also used to mean 'of value' or 'special', rather like gold is used in English.

Li, Jiajue, 1988. Linxia Mudan. Beijing: Science and Technology Publishing House. ISBN 5304-0371-0/S-26. (In Chinese)

Li, Jiajue, 1999. China Mudan and Shaoyao. Beijing: China Forestry Publishing House. ISBN 7-5038-2107-8. (In Chinese)

4.2 Historical Development of Gansu Mudan

Basic historical background

The Gansu Plain is one of the places traditionally and historically associated with the production of Mudan as ornamental plants. They have been cultivated there for more than one thousand years. Gansu Province is also one of the areas for growing the Chinese medicine danpi (the root bark of the tree peony) although in effect any region where Mudan grow (or once grew) in the wild is associated with danpi. Of the eighty-one

counties and cities in the province more than half I grow Mudan with Lanzhou, Linxia, Lintao and! Longxia as the centres of production both now! and in the past.

There are references to peonies and peony I growing in texts from about two thousand years I ago and archaeological evidence from a thousand years earlier still. However it was in the two Han I Dynasties (206-220), possibly earlier, that people I started to pay serious attention to the medical use I of Mudan and to appreciate the beauty of Mudan. I When Emperor Yang of the Sui Dynasty (581-618) ordered the planting of Mudan in imperial gardens, varieties of Mudan were collected in Luoyang and became the earliest Central Plains group. In the Tang Dynasty that followed (618-907) politics, the economy and culture developed rapidly and people in the capital city of Changan became obsessed with Mudan. In the Northern Song Dynasty (920-1127) there was great progress in production skills and varieties began to be propagated by grafting. In the Ming Dynasty (1368-1644) there was another widespread surge of interest and development and the propagating centre moved to Haozhou (Haoxian City in Anhui Province today) and then on to Caozhou (Heze, Cao County in Shangdong Province today) and the capital city Beijing. The propagation of Mudan also developed in places like Lanzhou and Linxia in Gansu in the Northwest.

Lanzhou was called "Gold City" in the Qin and Han Dynasties and was an important city along the ancient Silk Road. Mudan would have been traded along the Silk Road with Gansu Mudan travelling eastwards and meeting Central Plains Mudan. According to the written record of "New Gansu Chronology" published near the end of the Qing Dynasty (1644-1911), Mudan were grown "in all the government buildings and the most popular ones were from Lanzhou with all sorts of colours". This showed that Lanzhou, the ancient Gold City, has been a centre of Mudan cultivation.

ChenDezhong's extensive experimentation and breeding programme at Peace Peony Nursery has in a sense restored Lazhou's position. His development of new cultivars involved Central Plains varieties and true wild species plants now




Established Central Plains (suffruticosa) hybrids in front and established Gansu Mudan behind at Peace Peony Nursery.

called P. rockii. It also involved the traditional varieties, or at least existing examples, of Gansu Mudan and effectively subsumed them in his extensive range of new cultivars.

Current state of propagation

The First Lanzhou Mudan Exhibition was held at Peace Peony Nursery in May 1991. The large-plants and over 200 varieties of Gansu Mudan with their tall branches, abundant flowers, variety of colours and wonderful fragrance attracted a lot of attention. Subsequently Peace peony Nursery plants started to appear in various plant shows all over China and won many awards. They began to be propagated by other nurseries and to be exported abroad. This development has accelerated in recent years with growth of the Chinese economy so now they are widely available from many nurseries and dealers few, if any, of whom will ever acknowledge the original source of the plants. Peace Peony Nursery has already distributed about 150,000 named cultivars and unnamed seedlings, mostly in China.

Horticultural Qualities of Gansu mudan

Gansu Mudan have now been grown easily and successfully in many different environments. In China, in addition to Gansu, they have been thoroughly tested in Beijing, Shanghai, Inner Mongolia, Tibet, north-eastern provinces, Xingjiang and Hubei. During these tests plants experienced floods, drought and sandstorms and most coped well. These formal experiments have been mainly to establish their suitability for greening barren land in harsh conditions and selection continues to find particular varieties best suited to extreme situations. As ornamental plants in gardens and municipal planting they have grown well in many other areas. It is nicely ironic that while Gansu Mudan, at least the single, white form, remained one of the most sought after ornamental garden plants in western horticulture and were regarded as mysterious and almost mythical, they were being routinely tested in various harsh environments in China.

Although Gansu Mudan originate from a classic, moderately severe, continental climate in Gansu, exported plants have proved to be extremely adaptable and flourish equally well in Western, Mid-West and Eastern States of North America, and throughout Europe including Scandinavia and the Balkans. Their cultural requirements can be summarised as: reasonably nutritious soil that does not become waterlogged and an adequate water supply, particularly in Spring while they are becoming established. They will grow well in shade, within reason, and are undamaged by severe winters and hot summers. Persistent high humidity is far from ideal for them but in almost any situation they are worth trying.




4.4 Comparison of basic morphological features of plants of P.rockii, Gansu Mudan and Central Plains (suffruticosa) mudan growing at Peace Peony Nursery, Lanzhou, Gansu.

Wild P.rockii from Xinglong Mts

Gansu Mudan/ Paeonia Gansu Group

Central Plains Mudan


overall tall and big

80% big, 12% medium, 8% small

relatively small and short


long & thin

long & thick

short & thick


hairy 15 leaflets (more in other locations)

80% hairy, 85 % with 15 (or more) leaflets, 15% with 9 leaflets

not hairy, 9 leaflets


narrow & small, mostly unlobed

very varied, mostly lobed

round & big, lobed


small & pointed

mostly rather round & big

round & big

flower type

single, 11 cm in diameter

flower head mostly upright; various types; over 11 cm in diameter

hanging; various types; over 11 cm in diameter

flower colour

white, often pink at first

many colours

many colours



colour and form varies greatly, often bleeding out a different colour

some plants have a dark basal blush

back blotches

occasionally, small & faint

large & strong in some cultivars

occasionally, faint



usually strong and pleasant

faint, usually pleasant


filaments white, normal anthers

varied colours and forms

varied colours & forms


5 carpels

0-13 carpels, usually 5

0-13 carpels, usually 5-9



varied colours

varied colours



varied colours

usually brownish red


small but many

varied quanitiy and size

mostly few

flowering period

about 15 days

about 24 days

about 25 days

For most characters it is possible to find specimens that are exceptions to these descriptions. When grown with Central Plains hybrids the flowering period of Gansu Mudan is usually about a week later.



5. Rock'S PEONY'* - THE TRUE STORY Will McLewin and Chris Sanders

A Long Preamble: the Myth, the Reality and an Explanation

Part 1: Rock's Expedition to Gansu and the Background to It; Joseph F C Rock

Part 2: The Peony Seed Sent by Rock and its Distribution; Charles S Sargent

Part 3: Germination of Rock's Seed and the Resulting Plants; F Cleveland Morgan ,

Part 4: The Highdown Plant(s) and propagation of 'Rock's Peony'; Frederick C Stern The Present Situation; Paeonia 'Highdown' and Paeonia 'Joseph Rock' Postamble and References

A Long Preamble: the Myth, the Reality and an Explanation

The title of this chapter is deliberately perverse. The early twentieth century botanist, explorer and plant collector Joseph Rock was involved in the introduction of Gansu Mudan to Europe and North America. There is a true wild species peony now called P. rockii which commemorates his name but he in no sense discovered or introduced it nor ever saw a wild plant. There are plants in cultivation which have labels that involve his name but they do not constitute a clone or even a Group in the current formal sense and in most cases their provenance is uncertain to a greater or lesser degree.

In essence there is no such thing as 'Rock's peony', or rather there are many different plants labelled or referred to as 'Rock's peony', or 'Rock's variety' or Peony 'Joseph Rock', or Peony suffruticosa 'Joseph Rock ' or Peony suffruticosa 'Rock's form' or some other variation on these names. None of these are Paeonia rockii (unless

by some extraordinary mischance). The different names do not necessarily refer to different and distinct plants or groups of plants and no one of these names, as used up to now, represents a particular clone. Exceptions to this, for example one individual propagating one plant vegetatively and labelling them with one of these names, does not preclude the same name being used elsewhere for different plants, nor does it preclude plants which could or should properly be given one of these names actually being labelled with an entirely different name. All are simply examples of Gansu Mudan with more-or-less white flowers, of which there are many more examples with Chinese names or no particular name at all. Aside from the many cultivars illustrated in Section Three which have originated in China, there are particular clones which have originated in Europe or North America for which the use of a clonal or cultivar name is appropriate. The two most important emerge from the details of the true story. Alternatively, all such plants with their various

*In the term 'Rock's peony' the single quotation marks are used ironically. As we have pointed out 'Rock's peony' is an ambiguous expression and within the Gansu Mudan context it is misleading and almost meaningless. This use of single quotation marks is entirely distinct from their recommended use for validly published cultivar names.




names could collectively be regarded as a cultivar group based on morphological characteristics, but there is little virtue in attaching a cultivar group name to a subset of Gansu Mudan on the basis of the historical accident of having been grown in the West and previous inappropriate nomenclature. In addition the impossibility of describing such a (sub)group in a useful and usable way is immediately apparent from the first few pages of pictures in Section Three.

Our use of 'true' is admittedly provocative. 'The truth' is an elusive entity that can only be sought, never found (with absolute certainty), particularly in contexts such as this one. However, as in this case we have sought it more diligently and with more success than others who have pronounced on the subject, we feel justified in using the word.

The myth

The story of the so-called discovery in China of 'Rock's Peony' by the American plant collector Joseph Rock has been told a number of times in books and magazines. A recent version, similar to many others, is as follows: "He [Rock] was at the Zhoni (sic) lamasery, a Buddhist monastery in Gansu Province, for the winter of 1925-26 and saw the wonderful peony growing in the lamasery garden. He photographed the plant and collected its seeds, later distributing them to botanical gardens on both sides of the Atlantic A fine specimen of P. rockii grown from Joseph Rock's original seed is still there [in Frederick Stern's garden at Highdown in Sussex]".

Leaving aside the obvious question of what Rock could have seen of a wonderful peony in winter most of this account is wrong. In particular, the fine specimen at Highdown is not P. rockii.

The Reality

The basic outline of the true story of 'Rock's peony' is that Rock sent three separate batches of shrubby peony seed to Charles Sargent at the Arnold Arboretum (plus another batch, described by Rock as shrubby peony seed but actually herbaceous peony seed and not relevant to this story, but see below). These were almost certainly collected for him in his absence from shrubby peonies in cultivation in two separate locations. Some of these seeds were then sent to certain botanic gardens and to certain individuals, not including Stern. Some recipients germinated some seeds and subsequently seeds and small plants were distributed. Ten years after the initial distribution of the seed sent by Rock, Stern received a plant (very probably but not certainly the now famous one at Highdown) from Cleveland Morgan in Canada and subsequently exhibited flowers and distributed propagating material, seedlings and seed. There is no evidence to suggest that any of the many plants involved are other than cultivated hybrids, Gansu Mudan in fact, and all examples known to us are clearly Gansu Mudan and not P. rockii.

This outline indicates that in addition to Rock and Stern there are two other leading participants, Sargent and Morgan. They are almost never mentioned in other accounts but their involvement is crucially important. Our account of the historical section of the story is divided into four successive chronological periods which although they overlap and interact, mainly cover successive horticultural considerations. In each one of these four parts one of the major participants is prominent, and we give some brief biographical details there, although all are mentioned in other places. In addition to these four 'important' people whose documents are saved in archives and who are written about in articles and historical accounts and who, in fact, between them left the basis of all the subsequent misunderstanding, there are some other 'unimportant' people. They appear briefly in the account but their small items of evidence, preserved by chance, are crucial to uncovering the true story.

Before beginning our detailed account of the history of 'Rock's peony' we emphasise that comments like 'there is no evidence for' are, of course, to be read as 'we have found no evidence for', but also that where statements we regard as inaccurate have been made, and in most cases frequently repeated, no evidence has been given that adequately supports them.




An Explanation

As to why misinformation about a group of plants should have become so widespread, the simple answer is that overall the situation is genuinely complicated as explained in other chapters of this section, and everyone involved has been less than careful. The first mudan to be seen in the West were hybrids, the products of centuries of cultivation, but were named and described incorrectly as if they were true species plants (P. papaveracea, which is a Gansu Mudan, and P. suffruticosa and subsequent similar examples, which are Central Plains hybrids.) Farrer is sometimes credited with discovering P. rockii, in 1914 which is absurd. The plant had been well known to Chinese horticulturists for hundreds of years and evidently involved in hybridisation. But aside from this botanical imperialism, on rinding P. rockii in the wild Farrer apparently did not make careful and detailed field notes to describe the plants. Instead, in 1914 in The Gardeners Chronicle and in 1917 in On the Eaves of the World he produced the two much quoted versions describing, as Stern puts it, "in his own characteristic way" his act of discovery and his emotions at the time. ("Through the foaming shallows of the copse I plunged........that single enormous blossom, waved and crimped into the boldest grace of line, of absolutely pure white.......the boss of golden fluff at the flower's heart......" and so on.) Farrer did collect at least one example of the flower and some leaf material, presumably intended for a herbarium specimen sheet. Unfortunately this specimen (and others of herbaceous peonies) was received from Mrs Farrer at Edinburgh only after a considerable lapse of time. It was identified (as P. suffruticosa) in 1941 by William Stearn who was working at the time on Stern's seminal and still important 1946 monograph A Study of the Genus Paeonia where, curiously it is said that "unfortunately he [Farrer] did not collect specimens".

Most authors appear to have based their versions of the story directly or indirectly on Stern's account in his monograph and on very little else. Stern's information was, in turn, mostly derived from a letter he received from Rock in 1938 much of which is quoted in his monograph.

It is reproduced in its entirety here because what is not said is as significant as what is said. Rock recounts how he collected seeds from plants growing in the courtyard garden of the Yamen (official residence) of the lamasery of Choni (correct modern pinyin spelling Zhouni) in SW Gansu which he says he "occupied... for about a year". (In fact, for almost two years, he used it as a base from which to explore the surrounding area.) No date is given for the collection in the letter, although Rock does mention that the lamasery was "entirely destroyed and the lamas all killed in 1928 by the Mohammedans, so the plant (sic) does in all probability not exist any more as the entire Lamasery was burned to the ground". It is rather curious therefore that Stern should have stated in his monograph that "some time after 1932 [Rock] sent seed of a tree-paeony to the USA", although by the time he published A Chalk Garden in 1960 he had revised this to 1926 (closer, but still incorrect). However, this is just another of the errors, ambiguities and half-truths attributable to both Stern and Rock, which subsequent authors have uncritically repeated. To be fair to Stern, he did his best to find out about his own plant, but without complete herbarium specimens of wild plants and without first hand knowledge of the Chinese context and hampered by the passage of time and thel 939-45 war he failed to realise his plant was a cultivated hybrid. Stern's mistake was repeated, much less excusably, in 1990 by Haw and Lauener. They proposed a subspecies, ssp rockii, of P. suffruticosa that included Stern's plant, but without fieldwork or ever having grown or even seen true wild plants and partly on the basis of Farrer's very poor herbarium specimen. It should not be surprising that their detailed description of ssp. rockii also is erroneous in that while it fits Stern's hybrid it does not fit true wild P. rockii. Stern's main problem was the dearth of information concerning the collection and dispatch of the seed from Zhuoni, its arrival at the Arnold Arboretum and its subsequent distribution to other interested parties. Some twelve years were to elapse between the original receipt of the seed at the Arnold by Sargent and Stern's first sight, in 1938, of flowers similar to those that had made such an impression on Farrer 25 years before. Only then was Stern galvanised into trying to trace




the details of the origin of his plant. Enquiries to the Arnold Arboretum no doubt received the same reply then as we have today - there are no records of any tree peony seed received from Rock. This even though meticulous records of every other wild collection he made, together with his field notes were and are available. To be fair to Rock, he was exceptionally resourceful and the only person of all those involved thoroughly familiar with Chinese culture and language, but he was apparently not interested in peonies. He was probably familiar with large-flowered mudan in cultivation in China and attached no great importance to the mudan seed he sent to Sargent. Indeed we guess that after its dispatch it had never crossed his mind until Stern's letter arrived.

So, although both Rock and Stern, subsequent to events, wrote clearly inaccurate comments which have certainly helped to create the misconceptions about 'Rock's peony', they carry no more blame than the many commentators since who, with one or two notable exceptions, have seemed determined to demonstrate the truth of Roy (Lord) Jenkins' observation that "meticulous research is almost always the enemy of good anecdote". Most bizarrely of all perhaps, Rock himself in 1953 presented RBG Kew with a specimen of 'Rock's peony', which is nothing of the kind, and not even a Gansu Mudan.

Fortunately, we have discovered enoug hevidence from other sources, notably Rock's original diaries at RBG Edinburgh, previously unpublished correspondence between Rock and his sponsors now available on the Arnold Arboretum web-site, Stern's papers at RBG Kew as well as the accession records in the archives of both Kew and Edinburgh, to be able to piece together most of the story of the introduction of Gansu Mudan into western gardens.

Part 1: Rock's Expedition to Gansu and the Background to It

Dr Joseph Francis Charles Rock (1884-1962)

Josef Franz Karl Rock was born in 1884 in Vienna, Austria. He appears to have had an extraordinary linguistic ability and, according to the obituary by A. K. Chock, had by the age of sixteen mastered a number of foreign languages including Arabic, Chinese and Latin. In 1905, a year after graduating from Vienna University, he emigrated to the United States. As a young man Rock suffered from tuberculosis and presumably as a consequence went to Hawaii in 1907. (Hawaii had been annexed by the United States in 1899.) He regained his health there and in 1913 became a naturalized US citizen and anglicized his name to Joseph Francis Charles Rock. During this stage of his career he thoroughly explored the islands and developed a comprehensive herbarium of the native flora and published a number of authoritative books and scientific papers (in his adopted language, English). To his friends and colleagues there he was known as Pohaku, the Hawaiian word for rock. Before being appointed Professor of Systematic Botany at the College of Hawaii in 1919 he undertook plant introduction journeys to Malaysia in 1916 and 1919 and to Southern California in 1917 and a round the world trip in 1913/14.

With a secure reputation as a scientific botanist and plant explorer Rock finally left Hawaii in 1920. He was to spend the next three decades in active research and exploration in Asia. During this period he introduced literally thousands of Asiatic plants (including 439 species of Rhododendron alone) to the United States and gathered many more thousands of botanical, ornithological and zoological specimens, as well as taking over forty thousand photographs and mapping previously unknown regions. Numerous plants, birds and mammals commemorate his name. From his base in Li-chiang/Lijiang in the Chinese province of Yunnan he made a comprehensive study of the Naxi people, translated some of their literature and later published a dictionary of their language.




Rock is a truly heroic figure yet is not as well known, at least in the UK, as some of the other plant hunters of the first half of the twentieth century. There is no doubt that he was their equal in many respects and clearly superior in others.

The Arnold Arboretum Expedition to Kartsu/Gansu

Note: for Chinese place names we use the modern* pinyin transliteration, but in some places where the text is essentially quoting from historical accounts we use the spelling there followed by the modern spelling e.g. Kansu/Gansu. However the question of spelling place names in the latin alphabet is complicated when the local population is not Han Chinese but one of the many semi-autonomous ethnic minorities, who may well use their own transliteration scheme distinct from pinyin. Zhuoni (see below) is one place where this complication arises. Rock spelt it Choni; the pinyin spelling is Zhuoni; the indigenous (Naxi) transliteration is Jone.

During the early years of his travels in Yunnan and neighbouring parts of Burma, Siam and Indo-China, Rock's chief sponsors were the US Department of Agriculture and the National Geographic Society. In the summer of 1924, he returned to Washington D.C. and visited the elderly Professor Sargent, Director of the Arnold Arboretum at Jamaica Plain, Boston, Massachusetts. Sargent had previously employed the by then famous English plant collector Ernest ("Chinese") Wilson to collect seeds of woody plants for the arboretum, mainly in the more southerly provinces of Hupeh/Hubei, Szechuan/ Sichuan and Yunnan. Many of these introductions did not prove hardy enough to withstand the cold Massachusetts winters, so Sargent was keen to find tougher species or forms from the mountain ranges farther north. Wilson, who had suffered a nasty injury whilst returning down the Min Gorge in Sichuan after collecting thousands of bulbs of Lilium regale in September 1911, had long since retired from active exploration and was now working at the Arnold. Sargent saw that Rock was the ideal person to undertake an expedition to the Amne Machin/Animaqing/A'nye Maqen Shan and Datong Shan ranges on the borders of Gansu and Tibet (both are now included in the province

of Qinghai). The Arnold Arboretum expedition, to be sponsored by Harvard University, was to take three years and Rock (salary $500 per month and a similar amount for expenses) immediately began to purchase essential equipment and instruments prior to setting sail from San Francisco bound for Shanghai in the first week of October 1924.

Rock arrived at Yunnanfu/Kunming, the capital city of Yunnan, essentially the starting point of the expedition, in early November 1924. This was an unsettled period in Chinese history and fighting and brigandage were rife in the province. It was December 13th when Rock and his caravan of twenty-six mules, porters, trained Naxi staff and heavy military escort set out on their long journey north.

The four-month danger-ridden journey initially followed the ancient caravan route via Chaotung/Zhaotong and Suifu/Shuifu to Chengtu/Chengdu. Rock had to delay there for a month until the route became safe to continue on via Mienchow/Mianyang and across the Gansu border to Kaichou/Wudu, Minchou/Min Xian and Choni/Zhuoni. The details are recorded faithfully in Rock's copious diaries and also in the numerous letters he wrote to Sargent and others at the Arnold. Only someone of his stubborn determination, tenacity and bravery would have contemplated let alone successfully accomplished such a venture. He and his caravan eventually reached Zhuoni on April 21st 1925. Zhuoni, then a village of some four hundred families and roughly 2000 inhabitants, according to Rock, lies in the valley of the Tao He, a tributary of the great Huang He (Yellow River), at an altitude of 2500 metres. It was the centre of a semi-independent Tibetan principality of the same name which was ruled over by a hereditary prince called Yang Chi-ching. Rock described it thus: "The village is by far the best situated spot in Kansu Province, and the Prince's territory, which I traversed from north to south and east to west, is the choicest bit of land. Nowhere else in Kansu are there such forests, and the scenery is unsurpassed". Prince Yang (his actual title in the local Naxi dialect was t'ssu - chief or headman, tusi in pinyin) was in his mid thirties and was a cruel and rapacious leader by all accounts who treated his subjects abominably (he was later horribly murdered by his own




The photograph of the garden at the lamasery with peonies, mostly herbaceous, in flower.

people after Rock had left), but he nevertheless proved a good friend to Rock, helping him with letters of introduction, procuring supplies and pack animals and even lending him money. As he was also the Grand Lama of the nearby lamasery he was able to provide Rock with comfortable accommodation there. The lamasery was situated on a plateau about five hundred feet above and slightly to the west of the village. It was said to be the third largest of its kind and was almost a village in itself, with 172 buildings plus a further ten large and small chanting halls within its walls. Around 700 "malodorous" monks were in residence at the time, although more had been present in earlier years. The Prince and his family resided in his own palatial yamen down in the village itself.

Rock had anticipated being well clear of the troubles affecting Yunnan and Sichuan, but found that fighting between the Mohammedans and the local Tibetans and hostile tribes such as the Tebbus and the Ngoloks, severely hampered his plans to get to the mountain ranges he had come to explore. (The province of Tibet at that time included much territory that is now included in the modern provinces of Qinghai, Gansu and Sichuan.) Rock, therefore, decided to make Zhuoni his base for the remainder of the expedition, returning there at the end of each foray into the surrounding mountains and valleys. He spent the next two winters there, writing up his field notes and sorting, labelling, packing and then dispatching specimens and the collected seed to Sargent at the Arnold Arboretum. He left Zhuoni for what was to be the last time on March 10th 1927.

(The articles by Rock published in the National Geographical Magazine and S. B. Sutton's book In China's Border Provinces among the references at the end of this chapter provide more details about the expedition. Rock's diaries are particularly rec-mended.)

Part 2: The Peony Seed Sent by Rock and its Distribution

Professor Charles Sprague Sargent (1845-1927)

Unjustified assumptions about the mudan seed that Rock sent, or rather unquestioning acceptance coupled with fanciful extrapolation from Rock's letter to Stern, are a major reason for misconceptions about 'Rock's Peony', so it is important to examine what is known about the seed and where it came from.

The Lamasery Courtyard Garden and its Peonies

In his letter to Stern dated 30th August 1938 Rock states that "The seed of the Paeonia about which you enquire I collected from plants which grew in the Yamen of the Choni Lamasery...". Rather confusingly also he refers to "the plant" and several times to "it" as if there was only one plant, at least of this particular kind. On the other hand he says that "There were no double-flowered ones, all were single", without making it clear whether they were all the same type. Further evidence of the presence of peonies there occurs in one of Rock's articles for the National Geographic - "Life Among The Lamas Of Choni", in which he refers to "a tiny courtyard filled with choicepeonies, lilacs and other flowers".




J. F. ROCK 48 SHIH CH'IAO P'U K'UN-MING Yunnanfu Aug 24/38

Major F.C.Stern Highdown

Goring by Sea, Sussex England.

Dear Major Stern,

Your letter addressed to Dr. Godspeed was forwarded to me here for reply.

The seed of the Paeonia about which you enquire ( collected from plants which grew in the Yamen of the Choni lamasery elev. 8500 ft in S.W. Kansu. I occupied the Yamen in that lama sery for about a year. In the court of the Yamen grew a very beautiful single-flowered Paeonia. There were no double flowered ones, all were single, I remarked at the time that it looked to me like a wild species. The lamas told me it came from Kansu but whence, the exact locality, they did not know. I never came across it in a wild state. It had been kept for years in the lamasery. I took a photo of it growing in the court and I enclose a copy with


The lamasery has been entirely destroyed, and the lamas all killed in 1928 by the Mo-hammadans, so the plant does in all probability not exist any more, as the entire lamasery was burned to the ground.

I hope this information will be of interest to you. If you have any seed to spare of this beautiful Paeonia, I would appreciate if you could send me a few seed, by registered letter post to assure their arrival. I should like to plant them in my garden. With best regards

Sincerely yours, J. F. Rock.

P.S. If you should wish to publish the enclosed picture please do so. The legend might read: Growing in the Yamen court of K'ang-ting ssu (lamasery), Choni, Southwest Kansu elev. 8500 ft. Photo by J.F.Rock, 1925. J. F. R.

Rock's letter to Stern





Paeonia Gansu Group 'Jiao Mei'



The photograph that accompanied the letter was subsequently used by Stern to illustrate an article entitled "The Moutan Paeony" in the Journal of the Royal Horticultural Society for December, 1939. It shows the whole garden, which appears to be no more than perhaps 60 square metres. According to Rock's diary, it is likely that this photograph was taken on May 18th 1925. It is not possible to make out much detail, but most of the plants do look like peonies, some of which are in flower. However, there appears to be one plant only, which by its shape and stature and the poise of its flowers could be a Gansu Mudan or P. rockii tree peony. It is not clear whether the other plants are mudan or herbaceous kinds. The photograph is missing from Stern's file at Kew, but the original print, reproduced here, is held at Edinburgh (along with 20,000 others).

The question of the number and kind of plants from which the seed was collected and whether other different mudan were also present is critical. It has an important bearing on whether the plant or plants could have been true P. rockii to start with and whether the seed could have been true P. rockii seed. Rock gave his opinion of the plant in the letter to Stern: "I remarked at the time that it looked to me like a wild species. The lamas told me it came from Kansu, but whence, the exact locality, they did not know. I never came across it in a wild state. It had been kept for years in the Lamasery". Stern twice embellished this: in a note in the Journal of the Royal Horticultural Society in 1959 he wrote "The Lamas told him that it came from the mountains of Kansu, known as "Min Shan", which DR. ROCK tells me are limestone" and in The Chalk Garden (1960) "The lamas told him that it came from the mountains of that district [SW Kansu]". Stern and others, including Haw and Lauener who described and named P. rockii (as P. suffruticosa ssp rockii) for the first time in 1990, seem to have accepted without question that plants in cultivation raised from the original seeds sent back by Rock belong to the wild species. It is conceivable that the original plant(s) in the lamasery were transplanted there from a wild source and could, therefore, have been true P. rockii. It is conceivable that if there was more than one actual plant all were true P. rockii. It is conceivable that there were no other kinds of tree peony present either in the same

small garden or any other gardens nearby in the lamasery complex for cross pollination. Our view is that other possibilities are much more plausible. There is substantial evidence, at least from plants in cultivation, that isolated plants (or more than one clonal plant together) are self sterile. If this is the case and if there was only the one mudan in the garden, as appears likely from the photograph, then some other mudan must have been present nearby to pollinate it. So even if the plant in the picture was actually true P. rockii any seed collected from it is likely to have been hybrid seed. And in spite of what Rock said in his letter it is by no means certain that any of the few plants that were grown from his seed actually came from seed of the mudan in the lamasery garden (see below). What is much more likely to have been the case in our opinion is that the plant(s) were of hybrid origin to start with. Bear in mind that mudan had been cultivated and hybridised in China for centuries and these included Ziban Mudan (literally purple-blotched mudan) in Gansu, and that such plants were not uncommonly planted near temples and monasteries. Rock's comment that "it looked to me like a wild species" is in our view not significant. He would surely have come across Central Plains hybrids in cultivation in China which would mostly if not all have been doubles so the phrase "like a wild species" means, in our view, nothing more than the flowers were single. Incidentally, it is curious that Rock nowhere refers to the blotches on the petals that are a prominent and characteristic feature of Gansu Mudan and P. rockii but not of other peonies.

The Prince's Garden

In all the accounts of 'Rock's peony' that we have read, no one has ever referred to the garden of the Prince of Zhuoni or the role that the peonies there might have played in the story of 'Rock's peony'. Other than as an addendum to his field notes of wild plants (we presume), the only mention by Rock himself that we have found was in a letter to Sargent dated December 13th 1925. Right at the end of a long account of the seeds collected that autumn he casually remarked "I have also seeds of several Paeonia grown by the Prince of Zhuoni. These are tall plants and he says they come true to seed".




Rock's field notes for the various seed collections received at Kew and Edinburgh are preserved in the accession records at both places. They are typed, presumably from his handwritten originals. Three separate notes that refer to the mudan seed appear at the very end of the list and, in contrast to all the rest, do not have collection numbers. Rock's notes for these items read:

(1) "Paeonia sp. Shrub 4-5 ft tall, fls. very-large, white. Grown in the garden of the Choni prince's Yamen, Choni. elev. 8500ft. October 1925"

(2) "Paeonia sp. Shrub 3-4 ft. in the Lamaserie garden of Choni. — large, colour? October 1925"

(3) "Paeonia sp. Shrub 4-5 ft. fls. very large red, grown in the Choni Prince's Yamen. 8500ft. October 1925"

As is evident, collections (1) and (3) refer to the Prince's garden in Zhouni itself, while collection (2) is from the courtyard garden in the lamasery 500 ft. above the village. In spite of the omission in the description of the latter, which we guess is due to the inability of the typist to decipher Rock's handwriting, this seed is presumably the subject of Rock's comments in his letter to Stern referred to earlier. The " elev. 8500 ft." for Zhouni is presumably simply an error, since Rock, in his letter to Stern, says that the lamasery also was at this height. ( Estimations of altitude were not Rock's strong point - he once claimed that one of the peaks in the Amne Machin Range rivalled Everest.) The significance of the two collections from the Prince's garden increases later on when we discuss the plants known to have resulted directly from the seed sent by Rock (which do not include Stern's famous plant at Highdown).

There is another crucial point that should be highlighted here. In his letter to Stern, Rock stated that he collected the seeds from the plant (or plants) in the lamasery garden. The collection date given in the notes for all three of the tree peonies was simply "October 1925". However, Rock's diary clearly shows that he was away from Zhuoni from August 13th until December 3rd, so it is highly unlikely that he, Rock, could have collected any of the seeds personally. This may not be of great significance, since Rock used a number of trained Naxi collectors, but by using the phrase "Rock collected seed [of the peony]" with its added veracity other authors appear to have misled themselves as well as their readers. It also begs the question of how Rock, on receipt of Stern's enquiry about the origin of the seed, could have known which plant(s) it had come from. (Quite apart from Stern's dubious assumption that his plant had come directly from the peony seed that Rock had sent.)

Incidentally, it is usually assumed that Rock, in his letter, was replying to an enquiry directly from Stern. In fact Stern had written to Professor Goodspeed at the University of California, the sponsors of Rock's 1932 expedition (to Western China and Tibet) believing that Rock had collected the mudan seed on that expedition. Stern's letter was forwarded from there to Rock in Kunming.

There was one other packet of peony seed in Rock's 1925 consignment. This was from plants in the wild (Tao river basin, 9000ft) and had the collectors number 13593. The plant was described as 1 to 2 feet high with flowers large, white, pink, red, and although Rock listed it as a shrubby peony it was later identified by Rheder and Kobuski as an example of the variable and widespread species P. anomala.

Rock recorded in his diary that he completed the labelling of seeds harvested that autumn together with corresponding herbarium specimens and despatched them to the Arnold Arboretum on December 15th 1925 (not in 1926 as reported by Stern in 'The Chalk Garden'). The mudan seeds were sent at the same time, but as far as is known unfortunately without any specimens, dried or otherwise, of the plants.

Arrival and Departure of the Mudan Seed

Rock's seeds, including the mudan seed, arrived at the Arnold Arboretum on the 12th February 1926. They must have been dealt with very promptly, because a letter addressed to the Director of Kew E.G. dated the 19th of that month and signed by Charles Sargent, advising of the dispatch of 151




packets of Rock's seeds, was date stamped at Kew on the 28th. The same letter was sent to Edinburgh, the only difference being that they were sent 148 packets. Presumably all the other recipients were treated similarly. Enclosed with each letter was a typed copy of Rock's field notes and a handwritten check-list with details of the collection numbers of all the seeds in that consignment. All that is, except for the mudan seeds, which had no numbers and were added on at the end. In fact, the Arnold does have the international distribution records of this and other Rock seed consignments, together with the total number of packets sent to each recipient, but not the details of exactly which seeds were sent where, so it is not possible to tell from the Arnold list who received some of the mudan seed. Stern, in his monograph, rather strangely we think, included a rather vague list which he and others have repeated in subsequent accounts. Strange, because we know he was in possession of a more extensive and more precise list at the time. The evidence for this is contained in a copy of a letter sent to him from Montreal by someone who was one of the original recipients of the seed. This was F. Cleveland Morgan, another major player in the story but one who has remained unmentioned in other accounts. Stern must have contacted Morgan early in 1938 seeking information as to who else might have received some of the original seed. Morgan, in turn, wrote to the Arnold and received two letters in reply dated 5th and 6th May 1938 from William H. Judd which he sent on to Stern. Judd worked at the Arnold for thirty years from 1920 to 1950, for at least part of that time as propagator and is commemorated in Prunus x juddii and Viburnum x juddii. These two letters are among the most crucial bits of evidence. In the first letter he mentions that Charles Sander, who was Sargent's gardener, had raised two plants "from the same lot of seed as yours" and that they had flowered in Sargent's garden. He also wrote that the colour of the petals was "pale lilac" and that there were two sheets from these two plants in the herbarium. In this letter he said that he did not know of anyone else who had received seed, but the very next day he wrote again to say that he had found others who had also been sent seeds.

The list given in the second letter was:

Royal Botanic Gardens Kew,

UK Royal Botanic Gardens Edinburgh, UK

Berlin-Dahlem Botanic Garden, Germany

Bergianska Botanical Garden Stockholm, Sweden

Professor A. P. Saunders, Clinton, New York

T. A. Havemeyer of Long Island, New York

Lionel de Rothschild, Exbury, UK

F. Cleveland Morgan, Montreal, Canada

Professor C. S. Sargent

This is not the complete list of recipients of seed from Rock's December 15th 1925 consignment. According to the Arnold's international distribution records botanic gardens or similar institutes in Nanjing, Hokkaido, Paris, Illinois, Leningrad and Ottawa also received seeds. Judd, in his letter, does not say how he found his list of recipients twelve years after the event so we can only guess whether it contains everyone to whom Sargent sent or intended to send the mudan seed. In fact there was at least one other recipient, namely Kenneth McDouall of Logan, Scotland.

An interesting question concerns whether Rock and all the recipients of the peony seed were aware of Farrer's account. It seems certain that Rock was not. He makes no suggestion in his diaries that he attached any great significance either to the packets of peony seed he sent or to the plants themselves. There is no evidence that any of the recipients of the seed, with the likely exceptions of Sargent and Saunders, made the connection or that the seed was regarded with any special attention.

The absence of a record at the Arnold of the three packets of mudan seed, and hence the role of Sargent in this part of the story had puzzled others before it puzzled us. It perplexed Stern for years, even up to 1967 when he died, aged 82. There is a letter amongst his papers at Kew dated in that year that shows it still troubled him then. In that letter, replying to an enquiry from one Leo J. Armatys, Stern refers to the Judd letters and





Paeonia Gansu Group 'Lan He'



bizarrely includes himself among the recipients of the seed. (Equally bizarre is Stern omitting, we assume, to examine the acquisition records at Kew, which would at least, as it did with us, have given him a basically correct chronology.) In fact Judd's first letter contains the vital piece of information that provided an explanation of the mystery of why there are no records at the Arnold of Rock's mudan seed or of the plants which surely the arboretum would have raised. (We discount the possibility that the records were among correspondence of the Arnold from 1936 to 1946 destroyed by the then director E. D. Merrill in a "fit of temper".) The first hint of an answer came from the check-lists that accompanied the consignments sent from the Arnold to Kew and Edinburgh where the three lots of mudan seed do not have collection numbers. This explains why they are missing from the original handwritten, numbered notes that Rock sent back with all the other collections. Apparently, Rock took with him a set of notebooks with pre-numbered pages in which to record the details of each collection. The mudan were not included, presumably because they were collected from a cultivated rather than a wild source. However, we know from the typed version of the notes that were sent out from the Arnold to Kew, Edinburgh and presumably the other recipients, that Rock did send brief notes about the peonies, which must have been separate from his field notes. What happened to those brief notes? A plausible explanation is that Sargent took some of the mudan seeds home and took Rock's brief notes along with the seed. He was known to have taken home seeds that he considered were of horticultural rather than scientific interest and when Judd's letter came to light confirming that Charles Sander, Sargent's gardener, had raised two plants from the original seed there was an at least plausible answer to the mystery. There is another possibility differing slightly from this explanation. The information about the mudan seed could have come to Sargent not with Rock's field notes sent with the consignment of seed on December 15th 1925 but as a note included with his letter of December 13th. Then, when arranging for the subsequent repackaging and distribution of the consignment, Sargent also arranged for the recipients of the mudan seed to be given the information he had about it.

Part 3: Germination of Rock's Mudan Seed and the Resulting Plants Frederick Cleveland Morgan (1881 -1962)

We consider in turn each of the known recipients of Rock's mudan seed.

Edinburgh and Kew

As would be expected, different consignments of seed were sent to different recipients. For example, the list of contents sent, along with the 151 packets of seed, to Kew shows only two of the three batches of mudan seed whereas Edinburgh was sent all three among its 148 packets. There seems to be no information on the amount of mudan seed Rock sent to the Arnold or how much of it was sent by Sargent to the various recipients.

Surprisingly, we can find no evidence that any plants were raised at either Kew or Edinburgh. At Kew, there is no record of sowing, although other collection numbers from the same consignment do appear in the propagation records. At Kew there are several plants erroneously labelled Paeonia rockii or something similar, but their accession records confirm that none have any direct connection with Rock's seeds. At Edinburgh, there is the pencilled note "Did not germinate", against one of the lots on the contents list but that is all we could find. At Edinburgh, there is an interesting plant labelled P. suffruticosa ssp. rockii, believed to have originated at Logan (see below), which may possibly be a clonal propagation from an original Rock seedling, but is probably a later seedling. There is no doubt that at least some of the seed was viable since some plants were raised in other places. As the seed coats would have been dry and hard on receipt, it is doubtful if any would have appeared above ground until the following spring at the earliest, so perhaps the seed containers were thrown out too soon; but one would have thought botanic garden staff were experienced enough to know that. However, it should be borne in mind that literally thousands of packets of seeds were pouring in to these places each year and it is understandable that some may have been neglected or not even sown at all.




Rothschild, Saunders and Havemeyer

We have no knowledge that plants were raised by Rothschild, Saunders or Havemeyer. It might have been expected that Professor Saunders, an acknowledged authority and breeder of both tree and herbaceous peonies would have been successful, but in his many contributions to the American Peony Society Bulletin right up to the time of his death in 1953 there is no sign that he was. If he had raised plants, he surely would have used them in his breeding programme and there is no evidence of this as far as we know.

Charles Sargent

The two plants that Charles Sander raised and flowered in Sargent's garden at Wollaston, Massachusetts have given rise to their own mysteries. Sadly, Sargent died in 1927 so would not have known of Sander's success. The two plants must have first flowered in 1932 or 1933, as Judd, in his letter to Morgan of the 5th May 1938, refers to having raised a batch of seedlings from them "which are not yet large enough to flower." (Stern, in his monograph, inferred that all the original seedlings first flowered in 1938, another error that has been repeated down the years). Judd's comment about the colour of the petals being pale lilac is interesting, but possibly of limited significance. He may only have been referring to the colour on first opening which is often pale pink before changing to white. On the other hand the flowers may have been bluish pink. In which case this is the only instance of a bluish tinge reported from plants from Rock's seed although it is not especially rare in Gansu Mudan as a whole. Also, we note in passing that Sargent's death was a disaster for Rock because no further sponsorship by the Arnold was forthcoming.

The fate of Judd's (second generation) seedlings has eluded us. There appears to be no record of any finding their way into the Arnold arboretum. It is possible that they were destroyed by rodents, but we have not been able to confirm this. Sargent's estate was sold after his death and has since been built over. The fate of the original two plants is revealed in the letter dated 16th February 1967 in Stern's file at Kew from an amateur peony breeder Leo J Armatys, of Sampson & Armatys, Attorneys

at Law, from Central City, Nebraska. At Stern's suggestion he had written to the Director of the Arnold Arboretum "for information as to Rock's Variety", and mentions that he had received "several interesting letters from them". He goes on to say "Their records do show two tree peonies grown from seed sent by Rock from Tibet. One was 481-29, from Sargent's estate and the other was 907-36 from Charles Sanders (sic) - but both plants died". There seems little doubt that these were the two plants mentioned above. The letters from the Arnold to Armatys could be interesting, but we have been unable to find out whether the Arnold still has them.

The Bergius Botanical Garden, Stockholm

We have not visited the Bergius Botanical Garden and, again, have experienced some difficulty in obtaining the definite proof that we would like to see regarding the fate of the seeds sent to them in 1926. For much of what we know we are indebted to Tom La Dell, a landscape designer from Kent who worked at the Bergianska as a student in the 1970's. La Dell became interested in this subject following the article by Stephen Haw in The Plantsman (September 1991) where the petals of "Paeonia suffruticosa 'Rock's Variety'" were described as "more or less pure white, sometimes slightly tinged pink" (this was based on the plant at Highdown which Haw &j Lauener wrongly believed to be P. rockii and an original from Rock's seed). La Dell remembered that plants he saw growing in the Bergianska in his student days and which were said to have been raised from Rock's seed, varied from light to dark pink in colour and none was white. This concurs with several reports we have received about pink-flowered plants of 'Rock's peony' in other gardens in both Sweden and Norway, although we suspect that 'pink' could equally mean 'not pure white' or 'blended pink and white'. La Dell returned to the Bergianska in October 1991 and examined the accession records which showed that six plants had been raised from seed received from the Arnold Arboretum in 1926, of which four were still living. The most recent (2005) information we have is that only two, or possibly three, still survive. Critically, the records also show the origin of the seed to have been the "Prinsens av Choni tradgarden"




(the Prince of Zhuoni's garden). We have seen only copies of the (typed) records, which are not actual contemporaneous records, so we are unsure whether the information about the Prince's garden did come with the packets of seed sent from the Arnold to the Bergianska. However, it is the most likely answer and we have no reason to doubt that it did. Then of course it is no surprise that the resulting flowers were various shades of pink because two of the three seed lots were from plants in the Prince's garden with 'red' and 'white' flowers. Like the plant(s) in the lamasery garden, but even more so because of the different colours, the plants in the Prince's garden were very probably Gansu Mudan and not P. rockii.

Whether they were derived from the batch of seed from the red-flowered plant (which was likely to have been some shade of pink), or from the white-flowered plant the two (and there may well have been more than just those two) would almost certainly have cross pollinated. Karl Evert Flinck, the respected Swedish horticulturist, in a letter to Tom La Dell dated 21st May 1992, reports that "I have planted probably a thousand seeds (from the Bergianska plants) over the years and never had any white flowering plants. There have been plants that have flowered very pale but not dear white as the Rock form does (sic)". This would be curious if the information, in Rock's brief notes, about the flower colour of the seed parent plants could be taken at face value because even if the white colour was recessive, one would expect it to appear eventually. But just as 'red' might not mean what we might instinctively understand then so might not 'white'. We have evidence from a catalogue of the Bergianska nursery dated 1939 that plants raised from seed taken from the six originals have been distributed since at least that year and this no doubt accounts for many of the the pink-flowering Gansu Mudan to be found in Scandinavian gardens today.

Berlin-Dahlem Botanic Garden

Our information about over thirty Gansu Mudan at Berlin-Dahlem has come from our friend Irmtraud Rieck who contacted, among others, Gartnermeister Rolf Marquardt, a retired former curator of the botanic garden. It seems that the wooden shelves carrying what remained of files of historic records after extensive wartime damage became infected, probably with dry rot, and in 1974 both the shelves and the files were burnt. There appears to be no documentary evidence at Berlin to connect the mudan there with Rock's seed. Marquardt's opinion, received, he says, from his predecessor Stenzel, is that they are (directly) from Rock's seed. According to Marquardt, "It has always been the case that the plants here today, at least seventy years old, are of the first generation Rock provenance and none are of the further offspring sown later in European, American or New Zealand gardens". If this view was accepted then Berlin would have not only have had highest germination rate of any of the recipients of Rock's seed, but also has by far the highest number of surviving plants today and by far the widest range of flower colour. A detailed report prepared by Marquardt in May 2002 indicates that the thirty or so plants with single flowers and with the characteristic blotches vary from pure white to wine-red; the colour of the anther filaments also varies from white to 'red' and there are examples where the sheath colour is 'red'. The wide variation in colour indicates extensive hybridity, possibly involving Feng Dang group plants.

Whatever the details of their origin and whether or not they are all over seventy years old we have no doubt that all these plants, like those at the Bergianska are not P. rockii but examples of Gansu Mudan.

[Remark Burkhardt, June 2006: At least one other possiblity for the large number and different colours of the Mudan in Berlin is the collection of Camillo Schneider, who arrived back from its China trip after the end of WW1, where he spent the war years at Arnold Arboretum, working together with Sargent and Rehder. (1) The Späth Nursery of Berlin was one of the main sponsors of the plant collecting trip of Camillon Schneider and Heinrich Freiherr von Handel-Mazzetti, and in the same way one of the major sponsors of the Botanical Garden in Berlin. And – additionally, (2) Camillo Schneider returned back from Vienna to Berlin to start to work on the famous garden magazine 'Gartenschönheit'. And, contrary to Rock, (3) Schneider was very much intersted in peonies, as one article written during his journey in China shows: 1914: In der Heimat unserer Gartenpäonie. Österreichische Garten-Zeitung, 9. Jg., 281-283.]

Cleveland Morgan

Of the characters in this saga, Frederick Cleveland Morgan is, in horticultural circles, the least known; yet the significance of the part that he played cannot be over-estimated. Even Stern, despite receiving his plant from him never referred to him by name in any of his published writings. In fact the only clue as to his identity comes from a handwritten note on Stern's garden record card for the plant, then referred to as Paeonia suffruticosa.

Cleveland Morgan was a wealthy businessman, a son of James Morgan and eventually vice-president in the family business of Henry Morgan and Company (known as Morgan's). He had a country estate at Senneville on the Island of Montreal, Canada much of which




became in 1945, due to Cleveland Morgan's initiative, about half of the Morgan Arboretum of McGill University. He collected fine art from all over the world and was largely responsible for establishing the Montreal Museum of Fine Arts, and for most of its exhibits. He studied botany, embryology, geology and zoology at Cambridge University and might have gone on to a career in natural sciences had not his eyesight been inadequate for microscope work due to a childhood accident. He was well known as a horticulturist, particularly for his interest in rock gardening and as a breeder of bearded iris and orchids. It may have been through irises that he initially made contact with Stern. Exactly when they first met is not known, but they were both present at a conference on, ironically, "Rock Gardens and Rock Plants" in London on May 5th, 6th and 7th 1936, organised jointly by the Royal Horticultural Society and the Alpine Garden Society, where Morgan gave a paper entitled "Rock Gardening in the Province of Quebec". It was probably at this conference, or during the dinner on the Tuesday evening, or perhaps in the hotel bar afterwards, that Stern first learned of Morgan's involvement with 'Rock's peony'. Stern had presumably by then begun work on what would become his famous monograph and it is perhaps not too fanciful to imagine his astonishment as, out of the blue, he suddenly discovers that growing in a garden in Canada is what he assumes is Farrer's legendary peony and additionally that small plants had been produced. Whatever actually took place it cannot be a mere coincidence that he duly received a specimen of the peony "from a friend in Canada as quite a small plant in the autumn of 1936" (JRHS, 1959). Also present at the conference was G. P. Baker, a leading figure in horticultural circles at the time, a former President of the Alpine Garden Society and an expert on Iris, which would have linked him with Stern. He too received, presumably from Morgan, a mudan plant, which proved to be similar or identical to Stern's plant. The note on Stern's own record card confirms that the donor was "Cleveland Morgan of Montreal" and the date as November 1936. He also noted on his record card that "Dr Rock tells me he collected it in summer of 1926"!

Details of the outcome of Rock's seed sent from the Arnold to Morgan in 1926, remain obscure. It is hard to believe that Morgan kept no horticultural diaries or notebooks, but despite help from several people in Quebec none have as yet been located, although there are plenty of records and correspondence of his fine art activities. However, by good fortune and the interest and diligence of Celine Arseneault, the librarian at Montreal Botanical Garden, a letter dated January 27th 1941 from Morgan to Henry Teuscher, the garden's designer and first curator, was found in the Teuscher archives there. The two were apparently in the habit of exchanging plant material and in this letter Morgan wrote "The few seeds of Peony suffruticosa are worth trying as it is very rare and only now becoming known. I grew my one plant from seed sent me by the Arnold Arboretum and collected by Rock in Tibet over ten years ago". This shows Morgan probably knew the correct date of his receipt of the seed and it seems unlikely that he would not have given Stern accurate information, which makes Stern's subsequent errors about the original seed from Rock all the more inexplicable. Here again the solution of one mystery leads to another. There is no mention by Morgan of the flower colour of his plant although we assume it must have been white, more-or-less. On the other hand the two seedlings raised by Sander and said by Judd to be "from the same lot" were described as lilac coloured. If there was literally only one plant how was it pollinated to produce viable seed? Morgan's granddaughter Elizabeth Morgan, who still lives in the old family home, 'Le Sabot' in Senneville, has confirmed that the plant no longer exists there. The "quite small plant" that Morgan sent Stern is believed to be the large shrub still flourishing in Stern's old garden at Highdown, near Goring-by-Sea, Sussex.

Kenneth McDouall (1870-1945) of Logan, Scotland

Logan, a sub-tropical garden a few miles north of the Mull of Galloway, became part of RBG Edinburgh in 1969; prior to that it had been owned privately. The garden was effectively created by Kenneth McDouall and his brother Douglas during the first half of the last century. McDouall was an outstanding plantsman, well-




known in his day. He sponsored the expeditions of contemporary plant hunters such as E. H. Wilson and would have acquired new plants from many sources. McDouall was not on the list of recipients supplied by Judd and the only published indication of his involvement comes from a remark by Stern in the December 1939 article he wrote for the JRHS. In his letter to Stern Rock says "If you have any seed to spare of this beautiful Paeonia, I would appreciate if you would send me a few seeds by registered letter to ensure their arrival. I should like to plant them in my garden." In his article Stern relates that he did send Rock some seed, not from his own plant but from McDouall, but omits any information about the source of McDouall's seeds. Instead he writes "It may be interesting to record that seeds of this Paeony kindly sent to me by Mr K. McDouall of Logan, were sent to Dr Rock in China to replace the plants destroyed in the Lamasery at Choni and were safely received by him, so let us hope the Paeony will again bloom in the lamasery in years to come to bring peace and joy to the monks.". This flight of fancy may have some appeal for uncritical romantics but is actually nonsensical. The destruction of the Lamasery, however thorough and even by fire would not necessarily and not even probably have destroyed the peonies in the garden. Much more plausible is that if they had been destroyed it would be because they were dug up for danpi.

The Plant Database of the Royal Botanic Garden, Edinburgh, under the name Paeonia suffruticosa S. G. Haw (the author should be Andrews), includes five specimens under their accession number 19599781 which are stated to have been wild collected in China by Joseph F. Rock (another error). Four of these are growing at Logan and the other at the main garden in Edinburgh. This latter plant, which is said to have originated at Logan, was moved to its present site in 1987 from the garden of Professor D. M. Henderson, former Regius Keeper, when he retired. We have been told that when RBG Edinburgh took over Logan there were no records relating to the garden and its plants and that the only evidence of the origin of the mudan came by word of mouth from the older staff there. There is no evidence in their dispatch records to suggest that Edinburgh was the source,

either as seed or seedlings, of McDouall's mudan and hence to connect the Logan plants with the Arnold Arboretum and Rock. Having seen the plants at Logan in flower in 2003, we thought they were slightly pinker than usual and turned to the Bergianska as a possible source, but still could not find the link. Finally, a misfiled letter in Stern's papers at Kew provided the answer. It was the second of two letters from Kenneth McDouall to Stern dated May 9th and 16th 1938. The first letter accompanied two blooms of "Rock's Peony (sic) show colour variation". In the second letter McDouall wrote "The seed of this peony came from Wilson, from America. It was collected by Rock whom I remember described it as a beautiful species with a large flower. The seed was taken from a plant growing in a monastery garden which Rock came across on a collecting expedition in China. I do not know the year but it was some time ago. I do not think the seed had a number when it came here." Wilson was working for Sargent at the Arnold when Rock's seeds arrived and obviously knew McDouall. In fact, Barrie Unwin, the current curator at Logan, has confirmed that Wilson is known to have visited Logan. Here yet again an answer leads to more questions. How could McDouall have remembered in 1938 that Rock described 'the peony'? Did Wilson send seed to McDouall with Sargent's knowledge; which batch or batches did it come from; and did Wilson send any elsewhere? McDouall's May 9th letter appears to be in response to an enquiry from Stern and, together with his subsequent supply of seed to Stern, seems to imply that at that time McDouall had at least two fairly mature plants. After McDouall's May 16th letter Stern wrote to Edinburgh E.G. seeking information and only after drawing a blank there wrote his letter to Goodspeed which was forwarded to Rock.

We have seen the four plants at Logan and in our opinion, judging by their present size and locations, they are unlikely to be the originals raised from Rock's seed. The one at Edinburgh looks from its size to be much older, but is still most unlikely to be an original plant. It may have been a seedling from an original, or even perhaps a layer or rooted cutting from one, but it is unlikely that this can ever be determined. One thing is fairly clear though, Kenneth McDouall at




P. Highdown at the Highdown Garden in late May. Chris Sanders with the original plant of P. 'Highdown'. P. 'Highdown' at the Highdown Garden. An atypical leaf with rounded leaflets and

no red colouring taken from the original. plant of P. 'Highdown' at the Highdown. Garden. Individual flowers on the original plant of P. 'Highdown' at the Highdown Garden. Bottom row: Two flowers and a leaf taken from the plant of P. 'Highdown' in Chris Sander's garden




Logan has the honour of being the only person in the UK to grow and flower plants from Rock's original seed.

This review of the few plants known to have resulted from the original seed distributed by the Arnold Arboretum shows that only the two or three survivors at the Bergianska in Stockholm can be considered to be 'originals' although at least some of those at Berlin very probably are. In the absence of any absolute proof of vegetative propagation from an original seedling all the other plants in cultivation today known as Rock's Variety or something similar are from second or subsequent generations (or totally different sources). They must therefore be regarded as Gansu Mudan, belonging to Paeonia Gansu Group, as indeed so must the few originals. We hope we have proved conclusively that there is no justification for the use of the name Paeonia rockii for such plants and it should not, therefore, be used in future.

Part 4: The Highdown Plant(s) and propagation of 'Rock's peony' Frederick Claude Stern (1884-1967)

Stern, eventually Colonel Sir Frederick Stern, is chiefly remembered for the Highdown garden and for his two best-known works, A Study of the Genus Paeonia, published in 1946, and Snowdrops and Snow/lakes published in 1956, both of which are now much sought-after collectors' books. However his early life in particular was more colourful than this might suggest. As a young man he spent some time big-game hunting in Africa and also rode as an amateur steeplechase jockey. After he married in 1919 he and his wife Sybil began creating a garden in an old disused chalk pit at Highdown in Sussex on a south-facing slope of the South Downs. They experimented with many kinds of plants to find out which would do well on a chalk soil and he wrote about their experiences in A Chalk Garden (1960). Lady Stern left the garden to Worthing Borough Council for the enjoyment of future generations of gardeners. Many of the trees, shrubs, perennials and bulbs they planted are still doing well nearly forty years later.

Stern wrote many articles on diverse genera, but he was particularly interested in Lilies, Irises, Peonies and Snowdrops. Peonies, in particular, proved to be very successful on the Highdown chalk and many have survived up to the present day, including 'the Highdown plant', the (assumed) original plant received from Cleveland Morgan, which grows near the western edge of the garden. It first flowered in 1938, less than two years after planting and specimens taken from it on May 10th that year are preserved in the Kew herbarium. Haw & Lauener (1990) refer to "two specimens grown from Rock's seed cult, in the garden of F. C. Stern at Highdown" as if there were two plants involved, but this is clearly another mistake. There are actually three herbarium sheets at Kew and as there is no evidence that Stern had acquired other Gansu Mudan at that time they must be assumed to be from the same plant. By 1943 the plant was six feet high and "smothered with bloom" according to Stern, who submitted a vase to the RHS at Vincent Square on May 18th which was awarded a First Class Certificate. The records show a second vase was shown on the same occasion by G. P. Baker of Sevenoaks, which was jointly awarded the FCC. Both were submitted under the name of P. suffruticosa. The only information on Baker's entry form reads "It is said that it comes from Tibet". Presumably the two entries were at least very similar for the committee to make a joint award. There is a herbarium specimen taken from Baker's plant, dated 26th May 1940 at Kew. In Stern's file at Kew there is plenty of correspondence between Stern and Baker on irises, peonies and other, mostly bulbous, subjects, but no mention of mudan. While there is no documentary evidence that Baker received his plant from Morgan our guess is that both he and Stern were promised a plant at the same time in 1936 and both plants arrived later that autumn. Baker's old garden in Sevenoaks has long since been built on and we have found no information about the fate of the peony.

One important detail that remains unresolved is whether the plants which Morgan sent to Stern and Baker were seedlings from the single plant raised by him or clonal material such as rooted cuttings, layers, divisions or even




grafts. As ten years had elapsed since Morgan first received seed from the Arnold there would have been just enough time for his plant to have produced seedling offspring that would have been old enough to have flowered themselves by 1938. We know from the letter to Teuscher referred to earlier that seed was set and this seems to be the most likely source of the young plants sent by Morgan. This would make the Highdown plant two generations removed from the seed collected by Rock, thus, in our view, eliminating the faintest possibility of its having species status. In fact, there is additionally ample morphological evidence to indicate that the Highdown plant is a Gansu Mudan and not P. rockii. Firstly, the filaments of the stamens are heavily flushed with a similar blackish-purple colour to that of the blotches at the base of the petals. In all the pictures of true P. rockii and the actual flowers we have seen the filaments and the sheath are white or pale yellow and all contemporary Chinese descriptions we have seen of wild plants of P. rockii, including that in the new Flora of China, confirm this character. On the other hand in Feng Dan/P. ostii, P. spontanea (P. jishanensis), P. cjiui and many Central Plains mudan the filaments and sheath are entirely or mostly dark purplish-red. Secondly, there is a tendency for one to several narrow extra petals to be produced in some blooms and this is a further indication, by no means conclusive in itself, of the hybrid status of the Highdown plant.

'Rock's Peony' in North America

According to Don Hollingsworth and Roy Klehm most if not all the plants of 'Rock's peony' that have been propagated in North American nurseries derive originally from scions imported from Highdown by the late David Reath of Reath's Nursery, of Vulcan, Michigan. We have not so far been able to discover the exact date of the importation, but plants (probably grafted) were first offered for sale (as P. sujfruticosa var Rock's) by Reath in 1971, so it was probably sometime in the mid-late 1960s. Their catalogues at that time said it was the hardiest of all their mudan, coping well with winter temperatures down to minus 29 degrees Centigrade. There is an interesting, but probably misleading statement by Reath in a list of Peonies for Hybridising dated 1979 which included 'Rock's peony'. In this list 'Rock's Peony' was said to be "extremely fertile by selfing". As mentioned earlier there is a widely held belief in the UK, and confirmed elsewhere, that isolated plants do not produce seeds, although we have occasionally heard reports to the contrary. It is very difficult bordering on impossible to prove conclusively that plants are not self-infertile and! equally, in many of the cases where seed has been produced, to prove that cross pollination could not have occurred somehow.

Between 1969 and 1972 scions were supplied by Reath to the Klehm Nursery of Barrington, Illinois, who began selling plants grafted onto roots of herbaceous peonies around 1975. Klehm received at least two clones from Reath, one with single and one with double flowers. The emergence of Klehm's double form and the fact that Reath obtained seed from his plants suggests to us that in the scions sent to Reath several clones were likely to be involved that could have come | from 'random' seedlings at Highdown.

Propagation from the Highdown plant(s)

There seems to us little doubt that some, possibly I a quite high proportion, of the mudan grown in I British gardens today as 'Rock's Variety', 'Joseph Rock' etc., or more recently and regrettably as I Paeonia rockii, have descended directly or indirectly from the plant(s) at Highdown, mostly by means of successive generations of seedlings. The first mention by Stern that he raised seedlings does] not appear in print until he stated in A Chalk Garden (1960) "Seed of this plant has been raised and I comes true." By "true" he presumably meant that the resulting flowers were all basically white with dark blotches, not that they were all identical. It is also probable that he kept and sowed some of I the seeds sent to him from Logan in 1938, indeed this is the only possible source for which there is evidence of Gansu Mudan plants to pollinate the Highdown plant (it could have been pollinated by Central Plains hybrids of course). There are several Gansu Mudan growing at Highdown today which, from their appearance, could be seedlings from the | Highdown plant or from McDoualTs plants at Logan. There are two different plants growing at Kew I which the accession records show were donated by him, one in 1943 and the other in 1953. Neither are




the same as the Highdown plant and both are presumably seedlings. Stern was known to be a generous gardener and would have given away spare plants to friends and visitors. In this way plants would have been spread around and no doubt successive generations of seedlings have been raised from them and further distributed. From 1940 to 1960 Thomson and Morgan catalogues sometimes offered plants and/or seed usually listed as XR suffruticosa - Moutan Peony . The descriptions fit that of Gansu Mudan and sometimes are explicit about variable colour and form. Some of the seed offered is believed to be from the Highdown plant. Terry Dick in Aylesbury has two flourishing plants from this source, both white, one noticeably double, and he has distributed seedlings germinated from the seed they set.

All the seedling plants that have been produced, in gardens or commercially, will show differences from one to another. Whilst most will retain the essential characteristics of a so-called 'Rock's peony, namely large, white or pink-flushed or pink flowers, single or with varying degrees of doubling, with dark basal blotches differing in colour and shape, some will clearly show Central Plains/Suffruticosa characteristics. All those with appropriate features should be referred to as Gansu Mudan or, more formally, as members of Paeonia Gansu Group. None are P. rockii. We are sure that this rather sweeping statement applies much further afield than we have been able to see at first hand. We are conscious of labels saying P. rockii or 'Rock's Peony' in gardens, botanical and otherwise, in thousands of places unknown to us. Our challenge to all their custodians is "Prove it; but whether or not there is any presumed connection with Rock, read this account first".

The Highdown plant has been propagated vegetatively by cuttings. There is a note on Stern's record card that two cuttings were rooted in 1956 and planted out in 1958. Neither of these appears to be present in the garden today. The late Graham Stuart Thomas grew a plant in his last garden at Briar Cottage, Kettlewell Close, Woking which he told one of us (C.S.) originated as a cutting from the Highdown plant and on subsequent comparison it appeared that the two were identical. Contrary to what is often written, it is possible, with care, to root cuttings, (see Chapter 7).

On the other hand there is a good deal of evidence that the Highdown plant has been increased by grafting, at least in the UK. Notcutt's Nursery of Woodbridge, Suffolk was probably the first to be involved. According to Frank Knight, a former managing director of the company and later a director of the Royal Horticultural Society's garden at Wisley from 1955-1969, in a letter published in the NCCPG Newsletter No. 5 for Autumn 1984, scions from the Highdown plant were grafted using the roots of herbaceous peonies as understocks in February 1951. The agreement was that the resulting plants were Stern's and in due course were despatched to recipients of his choice. How long this arrangement continued is not clear but there is evidence that Notcutt's continued to propagate the plant for their own customers in small numbers until at least the early 1960s. Knight mentions in his letter that seedlings of P. delavayi were also used as understocks, but that there were problems with suckering. It never found its way into the firm's catalogues, however, and the demand always exceeded the supply. Ivan Dickings, a retired former head propagator for the company continues to graft a few plants from time to time and often uses P. delavayi seedlings as understocks.

There is an old specimen of 'Rock's peony' growing in the Dry Garden at the Savill Garden in Windsor Great Park and their records show that it was received as propagation material from Highdown in 1960. The following year plants were sent from Windsor back to Highdown and to Kew although as far as we know these do not exist at either garden today. Because of the short time-span we can assume that the plants must have been grafted, even in the absence of any direct evidence. At the RHS Garden at Wisley, Surrey, there is a large plant growing in the wild garden which is recorded as having been received from Windsor in 1973, although there is apparently no corroborative record at Windsor. Scions from this plant were grafted onto herbaceous peony roots by Ivan Dickings in the mid 1990s and two plants were returned to Wisley in 1996 and are now growing well in the Hot Borders. We have examined these plants in flower and all appear to be identical with the Highdown plant.



A leaf of P. 'Joseph Rock'

Individual flower of P. Joseph Rock

Above left and right: Individual flowers of P. 'Joseph Rock'




The Present Situation:

Paeonia 'Highdown' and Paeonia'Joseph Rock'

Probably the only person producing grafted plants of the Highdown clone commercially in the UK at the present time is Peter Catt of Liss Forest Nursery in Hampshire, albeit on a very small scale. His stock plants originate from the Graham Stuart Thomas specimen referred to earlier. Rather unusually, Peter Catt uses root sections of P. delavayi as understocks to avoid problems from suckering. A ten-year-old plant from this source is growing strongly in Chris Sanders' Eccleshall, Staffordshire garden. It is currently nearly two metres high and bore over thirty flowers in 2005. It has never produced viable seed.

The Klehm Nursery, now run by Roy Klehm and called Klehm's Song Sparrow Perennial Farm and Nursery has its main site at Avalon, Wisconsin. Their catalogues have listed their single flowered clone under various names, first as 'Rock's Variety', later changing it to 'Joseph Rock' and then P. rockii (syn P. 'Joseph Rock') and in the illustrations they have used the appearance of the flower has varied, just as the appearance of flowers of the Highdown plant in different publications has varied. Most recently it has been listed simply as P. 'Joseph Rock'. There have been several thousands of examples of this clone produced which probably makes it the Gansu Mudan clone most extensively known to have been vegetatively propagated. (It is impossible to say with confidence what might similarly have happened in China.) It would be a neat conclusion to this complicated account if the Highdown plant and Klehm's 'Joseph Rock' plant were the same clone but they definitely are not. Our first-hand experience of the Klehm plant is mostly limited to two young plants received from Roy Klehm that have not yet flowered, and two plants that remain from a batch Klehm sent to Sissinghurst in the early 1980's one of which is now growing in a garden in the Cots wolds. All have an identical and distinctive foliage characteristic, notably broader, thicker and somewhat overlapping leaflets which lack the early bronze-red tinting present in the Highdown plant and many others. Some pictures of the Klehm plant flowers closely resemble a photograph which appeared in an article by Sir

Peter Smithers in The Garden for November 1992 which he described as "the (sic) plant distributed in the USA, propagated vegetatively...." but there was no discussion or illustrations of the foliage. Minute examination of the edges of petals and blotches of individual flowers of Gansu Mudan will always reveal variations but attaching great significance to such variations is misguided.

The relationship between the Highdown plant and the Klehm plant is unclear and will remain so even if more detailed information about what material was sent to Reath and subsequently what material was sent to Klehm is ever found, and in any case it is essentially inconsequential. Both clones are simply examples of Gansu Mudan, two amongst many hundreds, and as the pictures in Section Three show, not such exceptional examples. The factors that make them both important are their places in the story of Gansu Mudan, especially the Highdown plant, and the extent to which they have been vegetatively propagated and distributed, especially the Klehm plant. These factors lead to the necessity of validly published cultivar names.

Paeonia 'Highdown' and Paeonia 'Joseph Rock': 'New' Cultivar Names for the Highdown and Klehm Plants

The Highdown plant has been referred to by various authors in the past as 'Rock's Variety' and as 'Joseph Rock', but neither appears to have been validly published as its cultivar name. As a result, these names have been used indiscriminately for anything that vaguely resembles a 'Rock's peony', i.e. a Gansu Mudan with 'white' flowers and dark basal blotches. Since the general acceptance of P. rockii as a species, albeit with complications discussed in Chapter 3, this name, depressingly, seems to have taken precedence over both cultivar names possibly for commercial reasons. As horticultural journalists utilise the romance or mystery of the inaccurate versions of the story of 'Rock's Peony' to make a more interesting article and fail to separate it from true P. rockii (which they have almost certainly never seen) the kudos of the species name increases.




The earliest publication of the name 'Rock's Variety' that we have found was in a note from Stern in the JRHS for August 1959. However, although it is clear that he was writing about the original plant, he did not describe it, merely stating that "it seemed to be exactly like the wild tree paeonies described by Reginald Farrer" even though Farrer did not actually describe the crucial characteristics of the wild plants. In any case the word 'Variety' contravenes Article 19.19 of the International Code of Nomenclature for Cultivated Plants (I.C.N.C.P.) 2004, thereby invalidating the name. Exactly when or who first coined the name 'Joseph Rock' is not clear but it has been in use at least since 1984 when an article by Diana Grenfell entitled Paeonia suffruticosa 'Rock's Variety' appeared in the Spring Newsletter of the NCCPG. In it she used the former name interchangeably with the latter and further confused the issue by referring to several plants in other gardens as well as the Highdown plant without making it clear if she was treating them as a single clone or not. Cameron Carmichael, in the Spring 2001 issue of Plant Heritage, did his best to clarify matters by proposing that "the only plant that should be called 'Joseph Rock' is the FCC plant." However, the description accompanying the article is incomplete and there is no photograph or reference to a herbarium specimen, so we doubt if this constitutes proper establishment of the name under the I.C.N.C.P. and in any case Klehm's use of the name predates this article. As we have described above, the name 'Joseph Rock' has been used in the USA for the most extensively clonally propagated clone, and this clone is distinct from the Highdown plant. To use 'Joseph Rock' for anything other than the single-flowered clone propagated and distributed by Klehm would add yet another layer of confusion to an already chaotic situation, so we propose Paeonia 'Joseph Rock' as the cultivar name for that plant. For the Highdown plant, in our view the most satisfactory way out of the nomenclatural tangle is to wipe the slate clean and start again with a new cultivar name. The most obvious and logical choice of name is Paeonia 'Highdown' so we propose that this name is given to what is generally assumed to be the original Gansu Mudan at Highdown and to all and only its vegetative descendants.

For a publication to validly establish a cultivar name a written description is required even though such a description (in this contexfl at least) can be used only negatively, to denjn the name to a particular specimen, although it could be argued that the published description effectively defines a 'group cultivar name' - i.e. if a specimen fits the description it can take tha name whatever its origin. On the other hand even clonal examples could differ in appearance ifi grown in different circumstances. This problem is discussed elsewhere and will continue to ba a problem whatever is said and decided about it. We avoid it here by stating that the primary criterion for valid use of Paeonia 'Highdown' and Paeonia 'Joseph Rock' is provenance - the certain knowledge that they are clonally propagated from authentic specimens, and that the pictures ana descriptions here are additional to that criterion.

It may be helpful to repeat here that in a cultivar name and the formal descriptions below the latin form of a genus name, here Paeonia, anq an unambiguous common name of a genus, here Peony, are both valid and are to be regarded as equivalent.

Description of Paeonia 'Highdown'

The basic description is a white, single flowered Gansu Mudan, category 1.1. in Section 3. However neither 'white' nor 'single' here are absolute sol 'category 1.1 or just possibly 1.2 or 9.1 or 9.2' is more helpful. From a horticultural point of view there are several features which, although each is present in many other Gansu Mudan, collectively are a useful aid to its identification, at least in the negative sense: the young foliage in spring is more or less suffused with a reddish-bronze colour which gradually reduces as the season progresses, but is still present to some degree afl flowering time (probably reduced if the plant is growing in shade); the colour of the outer surface of the petals in the bud stage is pale pink and this shade shows on the inner surface on opening, but quickly becomes white (less quickly in shade); there are 10 or 11 normal petals, about 12 cm wide by 10 cm long borne in the usual two whorls; in addition there may be one to several extra strap-




shaped staminoidal petals; the inner blotches are oval-shaped, deep blackish-purple in colour, feathered lighter at the edges, up to 4.5 cm long by 3 cm wide; there are faint back blotches (see 6.1 (vi)); there are up to 300 filaments borne in a dense mass; the inner, most visible filaments are heavily stained almost the same colour as the blotches, except for a short white base and a little more white near the anthers; the outer or lower filaments are more or less white; the 5 green carpels have creamy-white stigmas and are initially enclosed by a cream-coloured sheath which splits open as they mature; flowers are moderately and pleasant scented, reminiscent of roses. Once established, P. 'Highdown' can grow quite rapidly and is capable of forming a rounded shrub up to 2 m high and 1.5 m wide in ten years.

Description of Paeonia 'Joseph Rock'

The basic description is a white, single flowered Gansu Mudan, category 1.1 in Section 3. Particular features which, although each is present in other Gansu Mudan, collectively are a useful aid to its identification, at least in the negative sense are: the foliage is almost always uniformly deep green, the leaflets are almost all lobed and pointed and the lower leaflets nearest the rachis on each of the main pinnules overlap; there are usually 10 normal white petals borne in the usual two whorls with no or very little pink colour; the petals are quite broad at the base so the blotches on the outer whorl are mostly obscured; the blotches are oval-shaped, deep blackish-purple colour, feathered lighter red-purple at the edges, up to 4 cm long by 3 cm wide, there are usually faint back blotches; the filaments are mostly white, red-purple at their base; the 5 green carpels have cream-white stigmas; the sheath is cream-white; flowers are moderately, pleasantly and typically scented. Once established on its own roots P. 'Joseph Rock' can grow strongly to form a slightly upright rounded shrub about 2 m high in fifteen years.


Our investigation into the story of 'Rock's Peony' began with plants - hundreds of Gansu Mudan, none of which were P. rockii, and fewer, but still many, labelled or claimed to be P. rockii all of which were Gansu Mudan. This was not clear at first of course. What started as a suspicion and rapidly became a total certainty was that what botanists and horticulturalists were saying about P. rockii and 'Rock's peony' was inconsistent and self contradictory. As we tried to extract scraps of evidence from all the received wisdom and uncritical and unthinking repetition of nonsense, two experiences became familiar. One was that nothing, no matter who had said it, could be accepted without testing and corroboration. The other was that even in the most dubious statements there was likely, if we came back to it often enough, to be a loose flap of darkness which could be lifted up to find some light

Some remaining uncertainties seem capable only of speculative clarification but we believe that evidence may well exist about:

i) Cleveland Morgan's plant: a description of it; what was propagated from it and hence what was sent to Stern and Baker; what happened to it; was there really only the one plant and if so what pollinated it.

ii) Reath's propagating material from Highdown: who actually collected and sent it and when; what did it consist of; what did Reath produce from it and how, and what was distributed.

iii) Judd and his seedlings: what was the source of Judd's list of recipients; did any of Judd's second generation seedlings flower and what happened to them.

iv) Recipients of Rock's mudan seed: are there others we have not discovered and are there any more original seedlings that we do not know about?





A number of people have kindly helped us in various ways to unravel the tangled threads of the story of Rock's peony. We would especially like to thank the following:

Celine Arseneault, Chris Beardsley Chris Brickell, Cameron Carmichael, Peter Catt, Anne Chambers, Paul Cook, Sarah Cook, Veronica Cross, Lindsay D'Aoust, Alexis Datta, Terry Dick, Ivan Dickings, John Elseley, Mark Flanagan, Karl Flinck, Maurice and Rosemary Foster, Chris Grey-Wilson, Don Hollingsworth, Reiner Jakubowski, Roy Klehm, Sybil Kreutzberger, Hermann Krupke, Tom La Dell, Alan Leslie, Kenneth Lorentzon, Ron Macbeath, Rolf Marquardt, Diana Miller, Elizabeth Morgan, Rolf Nilsson, Martin Page, Irmtraud and Gottlob Rieck, Eric Schmitt, Claudia Schroer (APS), Pam Schwerdt, Bill Seidl, Mike Sinnott and Barry Unwin. Also the staff at the RHS Lindley Library, the Arnold Arboretum, RBG Edinburgh, RBG Kew and the Bergius Botanical Garden, and the Trustees of RBG Edinburgh and RBG Kew for permission to reproduce archive and herbarium material.


Allen, T. (1984). The Pollination of Paeonia suffruticosa - NCCPG Newsletter No.5

Armatys, L. (1967). Letter to Stern (No.5) in Stern, Paeonia mss. at RBG Kew.

Carmichael, C. (2001). Joseph 'Rock's peony' - Plant Heritage 8 (1): 8.

Chittenden, F. J. (ed.) (1936). Rock Gardens and Rock Plants. Report of the Conference Held by the RHS and ACS May 5th, 6th & 7th.

Chock, A. K. (1963). J. F. Rock (1884-1962) - Taxon Xll (3): 89-102.

Farrer, R. (1914). The Gardeners' Chronicle 1448, 26th September 1914

Farrer, R. (1917). On the Eaves of the World, London, Arnold

Grenfell, D. (1984). Paeonia suffruticosa Rock's Variety - NCCPG Newsletter No. 4

Halliwell, B. (1984). Paeonia suffruticosa 'Joseph Rock' - NCCPG Newsletter No. 5

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. & Lauener, L. A. (1990). A Review of the Infraspecific Taxa of Paeonia suffruticosa Andrews - Edin. Journal ofBot. 47 (3): 273-281.

Haw, S. G. (1991). Tree Peonies: A Problem Resolved (sic) - The Plantsman 13 (2): 94-97.

Haw, S. G. (1993). Letter in The Garden 118 (4): 176.

Haw, S. G. (2001). Tree Peonies - A Review of their History and Taxonomy - The New Plantsman 8 (3):


Judd, W. H. (1938). Letter to F. C. Morgan (No. 180) in Stern, Paeonia mss. at Kew.

Judd, W. H. (1938). Letter to F. C. Morgan (No. 181) in Stern, Paeonia mss. at Kew.

Knight, F. (1984). Propagated at Notcutt's - NCCPG Newsletter No. 5.

McDouall, K. (1938). Letter to Stern (No.227) in Stern Miscellaneous mss. at Kew.

Morgan, D (1992). The Morgans of Montreal 209pp - Private publication, Toronto.

Morgan, F. C. (1938). Letter to Stern (No. 179) in Stern, Paeonia mss. at RBG Kew.

Morgan, F. C. (1941). Letter to Henry Teuscher in Teuscher at Montreal BG.

Palmer, The Hon. Lewis (1967). Col. Sir Frederick Claude Stern QBE, MC, FLS, VMH - Journ. Roy. Hort. Soc. 92: 379-381.

Rheder, A. & Kobuski, C.E. (1933). Journ. Arnold Arboretum (Arnoldia).

Rock, J. F. (1924-1927). Various unpublished mss. and numerous letters from Rock

Rock, J. F. (1924-1927). Original diaries covering the expedition to Kansu in the archives of RBG Edinburgh. Rock, J. F. (1925). Field Notes in the accession records at RBG Kew and RBG Edinburgh.




Rock, J. F. (1925). " Experiences of a Lone Geographer: An American Agricultural Explorer Makes His Way Through Brigand-Infested Central China En Route to the Amne Machin Range, Tibet." - National Geographic Magazine XLV111: 331-347

Rock, J. F. (1928). " Life among the Lamas of Choni: Describing the Mystery Plays And Butter Festival in the Monastery of an Almost Unknown Tibetan Principality in Kansu Province, China." - National Geographic Magazine L1V: 569-619.

Rock, J. F. (1930). "Seeking the Mountains of Mystery: An Expedition on the China-Tibet Frontier to the Unexplored Amnyi Machen Range, One of Whose Peaks Rivals Everest" - National Geographic Magazine LV11:131-185

Rock, J. F. (1933). "Land of the Tebbus" - National Geographic Magazine LXXX1:108-127.

Rock, J. F. (1938). Letter to Stern (No. 179) in Stern, Paeonia mss. at RBG Kew.

Rock, J. F. (1939). Letter to Stern (No. 210-211) in Stern, Paeonia mss. at RBG Kew

Sargent, C. S. (1926). Standard letter to all recipients of Rock's seeds in accession records at RBG Kew and RBG Edinburgh.

Smith, W. W. (1941) Letter (from Edinburgh B.C.) to W. T Stearn and copy sent to F. C. Stern

Smithers, P. (1992). 'Rock's peony' - The Garden 117 (11): 519-521.

Smithers, P. (1993). Letter in The Garden 118 (4): 176.

Stern, F. C. (1939). The Moutan Paeony - Journ. Roy. Hart. Soc. 64: 550-552

Stern, F. C. (1956). A Study of the Genus Paeonia - London, RHS.

Stern, F. C. (1959). Paonia suffruticosa Rock's Var. -Journ. Roy. Hort. Soc. 84: 366

Stern, F. C. (I960). Tree Paeonies - Journ. Roy. Hort. Soc. 85: 295-299

Stem, F. C. (I960). A Chalk Garden - Edinburgh, Nelson.

Sutton, S. B. (1974). In China's border provinces; the turbulent career of Joseph Rock, botanist-explorer – New York, Hastings House.

Wagner, J. (1992). From Gansu to Kolding - The Expedition of J. F. Rock in 1925-1927 and the Plants Raised by Aksel Olsen - Dansk Dendrologisk Arsskrift X: 19-87.

Wagner, J. (1993). Letter in The Garden 118 (4): 176.

Wu, Z., Raven, P. H. and Hong, D. Y. (eds), Flora of China, Vol. 6 2001 (Paeonaicae), Science Press, Beijing & Missouri Botanical Garden Press, St. Louis





To examine the relationships between some of the cultivated and wild Mudan discussed in Chapters 3 and 5 a genetic fingerprinting study of 13 examples was carried out at the Jodrell Laboratory, Royal Botanic Gardens, Kew by Ben Davis and Mike Fay.

Genetic fingerprinting involves the production of a set of fragments which reflects the genetic constitution of an individual plant. These fragments can be characterised as bands in gel electrophoresis using radioactivity, silver staining or fluorescent dyes to label the DNA.

Earlier methods of genetic fingerprinting required relatively large amounts of DNA and, hence, also a large quantity of leaves, but techniques developed in the 1990s incorporating the amplification technology of the polymerase chain reaction (PCR) allow fingerprinting studies to be carried out with much smaller initial samples of DNA. In 1995, a technique called 'amplified fragment length polymorphisms' (AFLP) was developed by Keygene Inc. (Vos et al. 1995) and is the most sensitive fingerprinting technique currently available suitable for use with plants for which specific techniques have not been developed. The technique has several advantages over other currently used fingerprinting methods:

1. It is fast (the technique has been automated);

2. It requires relatively small initial quantities of DNA;

3. It provides 10-100 times more interpretable fragments and is thus more sensitive than some other fingerprinting techniques;

4. It is highly reproducible.

For these reasons, AFLP was chosen to investigate the genetic diversity in the Mudan samples in a blind study with the examples numbered randomly. In the AFLP fingerprints 98 bands could be identified. Of these 70 (71.4%) were present in every sample, indicating that all the samples are closely related. The remaining 28 bands were absent from 1-12 of the samples. Parts of representative traces are shown opposite together with a tree derived using the neighbor joining cluster algorithm showing the relationships between the samples. Branch lengths in the tree give an indication of genetic distance.

All the samples had unique fingerprints at this level of sampling. The three samples from plants of wild origin, 8, 12 and 13, were the most genetically distinct and were distinct from the samples from plants in cultivation. The six samples directly or indirectly from Highdown were found to be genetically very close but P. 'Highdown' and P. 'Joseph Rock' were found to be distinct. Samples 1,4 and 5 were indistinguishable as were samples 9 and 11. Sample 10 possessed one band only not found in 9 and 11. It should be pointed out, as the traces suggest, that the technique requires some active interpretation together with skilled judgement about which peaks to accept. The reliability of the technique is probably higher than the 95% claimed by Vos et al and so the study confirms the thrust of the previous Chapters in Section 1.





Sections of sample traces from the AFLP reactions, showing some shared and some varied DNA fragments (represented by the peaks). The numbers on the scale are the length in base pairs of the DNA fragments. Some fragments varying between samples are indicated by arrows.

Provenance of samples:

1. The large Gansu Mudan at the north end of the order beds at R.B.G. Kew believed to be a seedling from P. 'Highdown'

4. Chris Sanders' Gansu Mudan believed to be vegetatively propagated from P. 'Highdown'.

5. F.CStern's original plant of P. 'Highdown' still flourishing at Highdown.

9. Young grafted plant of P. 'Joseph Rock' from Klehm's nursery.

11. Young grafted plant of P. 'Joseph Rock' from Klehm's nursery.

10. The mature plant of P. 'Joseph Rock'; one of the batch sent to Sissinghurst in 1980's.

6. The white double flowered Gansu Mudan distributed by Klehm.

2. A Gansu Mudan at Picton's Nursery near Worcester.

3. A mature Gansu Mudan at Phedar Nursery received as 'Siemans form of P. rockii'.

7. The mature Gansu Mudan at Ness Botanic garden (see page 38).

8. P. rockii from Wen Xian

12. P. rockii from Tianshui (E)

13. P. rockii from a location in Tianshui (E) distinct from that of 12.

Vos, P., Hogers, R., Bleeker, M., Rijans, M., Van de Lee, T, Homes, M., Frijters, A., Pot, J., Kuiper, M. & Zabeau, M. (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research 23: 4407-4414.


A neighbor joining tree showing relationships between the 13 mudan samples.



Paeonia Gansu Group 'Yu Long Bei'