A Monograph of Codonopsis and Allied Genera (Campanulaceae)

By De-Yuan Hong

Codonopsis and its allied genera, are a group of plants which are important in economy and horticulture. A Monograph of Codonopsis and Allied Genera (Campanulaceae s. str.) offers its audience comprehensive knowledge of these plants including palynology, cytology, population biology, morphological description, geographical distribution with vouchers cited, excellent ink illustrations, and color photos, and keys to genera and to species.

This excellent work will facilitate identification of relevant plants, use of plant resources, assessment of endangered states, the development of conservation strategies, and will promote systematic and evolutionary research of this group.

• Provides comprehensive descriptions and classifications of Codonopsis and allied genera
• Richly illustrated with line drawings and high-quality color photographs
• Delineates and clarifies the relationships of Codonopsis and its allied groups based on the analyses using data from external morphology, pollen morphology, chromosomes, and molecular biology

• Language: English
• Publisher: Academic Press
• Release date: Sep 16, 2015
• ISBN: 9780128019412

De-Yuan Hong

Hong De-Yuan graduated from Institute of Botany, CAS in 1966. He is currently a professor of Institute of Botany, CAS, elected as member of CAS in 1991 and member of Third World Academy of Sciences (TWAS) in 2001. He is Vice-Chairman of the Botanical Society of China, Correspondent member of the Botanical Society of America, Correspondent member of the Botanical Society of Japan, and Director of the Department of Life Sciences, National Natural Science Foundation of China (NSFC). Professor Hong has been working on systematics and evolution of plants (Campanulaceae, Paeoniaceae, Scrophulariaceae) more than 45 years, found 8 new genera and more than 50 new species. He published 260 papers and 15 books, including 4 English titles. He is the vice co-chair of the editorial committee, Flora of China.

Book preview

A Monograph of Codonopsis and Allied Genera (Campanulaceae)

De-Yuan Hong

State Key Laboratory of Systematic and Evolutionary Botany, Institute of Botany, Chinese Academy of Sciences, Beijing

In association with

Qiang WANG and Kai-Yu PAN

Table of Contents

Cover image

Title page

Copyright

Dedication

Synopsis

Preface

Acknowledgements

Chapter 1. Introduction: The Taxonomic History and Questions to be Addressed

Discoveries in Codonopsis and Allied Genera

Circumscription of Codonopsis

Systematic Position of Codonopsis and Allied Genera

Taxonomical Revisions of Codonopsis and Allied Genera

Chapter 2. Biology

Gross Morphology

Seed Morphology

Pollen Morphology

Chromosomes

Chapter 3. Systematics and Evolution

Circumscription of Codonopsis and Its Relationships with Allied Genera

A Three-Tribal Classification System of the Campanulaceae s. s. and the Systematic Position of Codonopsis and Allied Genera

Evolution and Biogeography of the Cyanantheae

Chapter 4. Taxonomic Revision of Codonopsis and Allied Genera

Codonopsis Wall.

Himalacodon D. Y. Hong & Q. Wang

Pankycodon D. Y. Hong & X. T. Ma

Pseudocodon D. Y. Hong & H. Sun

Cyclocodon Griff. ex Hook. f. & Thomson

Echinocodon D. Y. Hong

Bibliography

Index to Specimens Cited

Index to Botanical Names and Synonyms in Codonopsis and Allied Genera

Index

Copyright

Academic Press is an imprint of Elsevier

125 London Wall, London EC2Y 5AS, UK

525 B Street, Suite 1800, San Diego, CA 92101-4495, USA

225 Wyman Street, Waltham, MA 02451, USA

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

Copyright © 2015 China Science Publishing & Media Ltd. Published by Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-801933-7

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

For information on all Academic Press publications visit our website at http://store.elsevier.com/

Publisher: Janice Audet

Acquisition Editor: Simon Tian

Editorial Project Manager: Simon Tian

Production Project Manager: Julia Haynes

Cover Designer: Victoria Pearson

Typeset by TNQ Books and Journals

www.tnq.co.in

Printed and bound in United States of America

Dedication

Dedicated to my parents:

Brave and diligent father, Guan-Hua HONG (1894–1959);

Intelligent and diligent mother, Shi-Ru HU (1901–1971)

Synopsis

The author has worked on the Campanulaceae s. s. since 1970’s. Recently after he had partially finished his work on Paeonia worldwide he has been concentrating his energy on studies of Codonopsis and its allies, which is a group very controversial among authors in taxonomy and systematics. He and his assistants made expeditions to the Himalaya (S and SE Tibet and N Pakistan), to the Hengduan Mountains (NW Yunnan and SW and W Sichuan), and to E Qinghai, where most members of this group are distributed. During the field work they conducted population observations, population sampling, and collection of materials for chromosome counting and DNA analyses in addition to collection of vouchers and herbarium specimens. The author examined and identified all specimens available in 27 herbaria. Based on the material collected in the field and studied in herbaria they observed gross morphology, seed morphology, pollen grains, and chromosomes, and made intensive DNA sequence analyses. As a result, they have established three new genera, described seven new species and two new subspecies, clarified the circumscription of Codonopsis and its relationships with allied genera, and created a new three-tribal classification system of the Campanulaceae s. s. The monograph assembles the results achieved in the author’s work, and discusses evolution of the group, particularly the evolutionary trends of characters such as ovary position, fruit morphology, and basic number of chromosomes. The major part of the monograph is the taxonomic revision of Codonopsis and allied genera: Codonopsis (46 spp.), Cyclocodon (3 spp.), Echinocodon (1 sp.), Himalacodon (1 sp.), Pankycodon (1 sp.), and Pseudocodon (8 spp.). It clarifies the relationships between genera and between species, provides an useful key to genera and to species for each genus with more than one species, and gives a concise but diagnostic description for each species. Every species is provided with a fine line illustration, and most species with photos taken in the fields. Specimen citation and distribution map are presented for each species. Forty-three collections are designated as types (lectotypes, isolectotypes, or neotypes). There is a list of all the herbarium specimens examined and their species designations.

Preface

Nearly all species of Codonopsis are medicinal plants. For example, Codonopsis pilosula (Franch.) Nannf., Dang Shen in Chinese, is one of the most famous Chinese traditional medicinal plants, and is widely cultivated in China. The raw extract of its roots promotes digestion, strengthens the immune system, dilates peripheral blood vessels and inhibits adrenal cortex activity. Codonopsis and its allies are also important in horticulture. According to Grey-Wilson (1990), 32 species of Codonopsis in the broad sense (64% of the entire genus) are cultivated throughout the world as ornamentals for their beautiful flowers, even though the great majority of species are foetid.

The taxonomy of Codonopsis, however, has been extremely controversial on several issues.

First, there is no stable classification system of the Campanulaceae s. s. and thus how many subdivisions should be recognized, three or four, is still in dispute. Closely related to this issue is the systematic position of Codonopsis and allied genera, which until 1997 had been placed with Wahlenbergia, but the former has colpate pollen, whereas the latter possess porate pollen. Takhtajan’s (1997) classification system of the Campanulaceae s. s. recognized four subfamilies: Campanuloideae, Cyananthoideae, Canarinoideae and Ostrowskioideae. He separated Codonopsis and its allies from Wahlenbergia, which is a good step towards a natural classification of the family, but at the same time he separated Codonopsis from some of its allies, e. g. Canarina and Cyclocodon. Therefore, the systematic position of Codonopsis and allied genera, and relationships among these are still unclear.

Second, the circumscription of the genus Codonopsis has been controversial. Ten generic names have been involved in the taxonomic history of the complex: Codonopsis, Campanula L. Campanumoea Bl., Campanopsis Kuntze, Cyclocodon Griff. ex Hook. f. & Thomson, Echinocodon D. Y. Hong, Glossocomia D. Don, Leptocodon Lem., Numaeacampsa Gagnep., and Wahlenbergia Schrad. ex Roth. The recent literature focuses on two questions. One is whether Campanumoea, Cyclocodon and Leptocodon should be kept separate from Codonopsis as genera or merged with Codonopsis (Moeliono, 1960; Lammers and Klein in Hong et al., 2011, in notes), while the other is whether Codonopsis, even in the narrow sense (Hong, 1983; Hong et al., 2011), is homogeneous or heterogeneous (Eddie et al., 2003; Haberle et al., 2009).

Third, the taxonomy of a number of species complexes in Codonopsis and its allies has been questionable or controversial. For example, in the Codonopsis pilosula (Franch.) Nannf. complex, eight species have been described, but their identities and relationships are unclear. The taxonomy of Codonopsis convolvulacea Kurz in subg. Pseusocodonopsis is even more problematic and has been very controversial among authors. The number of species recognized in the complex has changed from one, three, four, six to nine (Shen & Hong, 1983: Hong in Hong et al., 2011; Hong & Ma, 1992; Grey-Wilson, 1995; Lammers & Klein, 2010). Which treatment is most rational? Grey-Wilson (1995) expressed the difficulty: "no treatment of Pseudocodonopsis is going to be entirely satisfactory and any treatment will have its critics". Therefore, taxonomic revision of Codonopsis and its allies is a great challenge.

Fourth, some of the taxonomic controversies of Codonopsis and its allies mentioned above come from the evaluation of characters. For example, Kolakovsky (1987 & 1994) established a new subfamily, Canarinoideae with two genera, Canarina and Campanumoea, just because these two genera share baccate fruit. A justifiable answer to the question whether Campanumoea should be merged with Codonopsis or kept separate from it as an independent genus depends on how to evaluate the systematic significance of the baccate fruit. The position of ovary, dehiscence of fruit, and pollen morphology, apertures in particular, have been used frequently in the taxonomy of the Campanulaceae s. s.

In addition, the distribution patterns of the platycodonoid group (= the tribe Cyanantheae), where Codonopsis and its allies are placed, are very interesting and a number of questions arise from this.

The great economic importance of Codonopsis and allied genera, a number of challenging questions mentioned above, and the interesting distribution patterns attracted my attention. In the present study I took an integrative approach to taxonomy, making extensive field observation and collection in the Himalaya and its adjacent Hengduan Mts., covering N Pakistan, S and SE Tibet, NW Yunnan, W and SW Sichuan, E Qinghai, and a number of other sites, and critically examining thousands of specimens in 27 herbaria in Austria, France, Germany, India, Japan, Nepal, Pakistan, Sweden, Switzerland, the United States, and the United Kingdom, in addition to China. With great help from my assistants and colleagues I observed the gross morphology, pollen morphology, seed-coat sculpture, and chromosomes, and sequenced four chloroplast genes and one nuclear gene, and subsequently constructed phylogenetic trees based on the molecular data. In doing so, my intention was to address the questions mentioned above. As a result, we proposed a three-tribal classification system of the Campanulaceae s. s. and established six subtribes for one of three tribes, the Cyanantheae, so the systematic position of Codonopsis and its allies has been clarified. Based on molecular phylogenetic analyses, combined with data from gross morphology, pollen morphology and cytology, we have proposed three new genera and clarified the circumscription of Codonopsis and its relationships with allied genera. Combined molecular data, population observations and sampling, critical examination of a large amount of specimens and statistical analysis have resulted in improved taxonomy of the Pseudocodon (Codonopsis) convolvulaceus complex and other groups. We have traced evolutionary trends of the characters which were often used for studies on phylogeny and taxonomy. Finally we also inferred the origin and the formation of distribution pattern of the tribe Cyanantheae.

Acknowledgements

I could not have completed this complex work without assistance from numerous persons. First, I should send my sincere gratitude to Dr. D. Geltman in the Komarov Institute of Botany, Russian Academy of Sciences, for providing us with DNA material of Ostrowskia magnifica, and to Prof. Dr. J. Murata in the Botanical Garden, Tokyo University, for adding localities of three species in Japan on the distribution maps. My special thanks are due to Dr. D. Boufford of Harvard University and Dr. M. Gilbert at the Royal Botanic Gardens, Kew, for significantly improving the English. Mr. Bing LI provided us with DNA material of Echinocodon draco (Pamp.) D. Y. Hong (Echinocodon lobophyllus D. Y. Hong). Mr. Xin-Tang MA helped me as head of my expeditions to Tibet, and took many beautiful photos. Miss Ai-Li LI prepared all the illustrations, which are of a high standard from both scientific and aesthetic point of view. Miss Xiao-Lin GONG processed the manuscript. Sincere thanks are due to Mr. Min LI and Miss Yan-Li XUE who assisted in preparation of colour plates. I am grateful to Professor Hang SUN, Dr. Huan-Chong WANG, Mr. Yang YANG and Mr. Ji-Pei YUE for their help in field trips to NW Yunnan in 2010, to professor Xing-Jin HE and Mr. Xiang-Guang MA for their help in the field work in SW Sichuan in 2011, and professor Shang-Wu LIU and Mr. Qing-Bo GAO for their help in the expedition to Datong, Qinghai, in 2012. I should send my gratitude to Professors Hang SUN and Dun-Yan TAN, Drs. Fu-Sheng YANG, Shu-Ren ZHANG, Xiao-Hua JIN, and Mr. Lian-Zhong FU, for generously providing me with photos. My sincere thanks are due to curators of the herbaria A, B, BM, CAL, CAS, CDBI, E, G, GXMI, HAST, HNWP, IFP, ISL, K, KATH, KUH, KUN, LD, LE, P, PE, PMNH, SM, SZ, TI, W, WU, for their permission to examine their specimens, and particularly to curators of the herbaria A, K, B, P, KUN, CDBI, SM and SZ for sending me many important specimens on loan. I should also thank many anonymous persons who have contributed to this work. The present work was financially supported by the National Natural Science Foundation of China (grant no. 31170175), and the National Publication Foundation of China.

Chapter 1

Introduction

The Taxonomic History and Questions to be Addressed

Abstract

This chapter describes the circumscription of Codonopsis, which has been greatly changed and is still in dispute. In addition, the systematic positions of Codonopsis and its allies in the Campanulaceae s. str. are discussed, which involves the classification system of the family as a whole and the phylogeny of the tribe Cyananteae in particular. Finally, the great controversies about the taxonomic treatments of certain species complexes are described, particularly for species related to Codonopsis convolvulacea Kurz (= Pseudocodon convolvulaceus).

Keywords

Circumscription; Codonopsis; Codonopsis convolvulacea Kurz; Systematic position; Taxonomic treatments

Readers need to pay attention to three issues when reading this chapter. One is the circumscription of Codonopsis, which has been greatly changed and is still in dispute; second is the systematic position of Codonopsis and its allies in the Campanulaceae s. s., which involves the classification system of the family as a whole and the phylogeny of the tribe Cyananteae in particular, and the third is the great controversy over taxonomic treatments of certain species complexes, particularly those species related to Codonopsis convolvulacea Kurz (= Pseudocodon convolvulaceus).

Discoveries in Codonopsis and Allied Genera

Wallich was the first botanist to report on Codonopsis when he established it as a new in 1824, and described three species, C. viridis, C. purpurea and C. thalictrifolia, all from the Himalaya.

Roxburgh (1824) described Campanula lancifolia from what is now Bangladesh as new. The species was transferred to Cyclocodon by Kurz (1872), transferred to Campanumoea by Clarke (1881), and treated as a synonym in Codonopsis by Moeliono (1960). With the restoration of Cyclocodon Griff. ex Hook. f. & Thomson, it is now placed back in Cyclocodon (Hong & Pan, 1998).

In 1825, Don described a new genus and species from Nepal, Glossocomia tenera, which was, however, the same taxon as Wallich’s Codonopsis thalictrifolia.

Another new genus, Campanumoea, with two new species, C. javanica Bl. and C. celebica Bl. was described by Blume (1826) from Indonesia. The status of Campanumoea, its circumscription, and its relationship with Codonopsis in particular, has been a controversial issue. Campanumoea javanica Bl. as a species has never been questioned, whereas the fate of C. celebica is quite different. It was transferred to Codonopsis by Miquel (1835), and then into Campanula by Dietrich (1839), but placed back in Codonopsis as a subspecies of Codonopsis lancifolia by Moeliono (1960). It was transferred by the present author to Cyclocodon (Hong in Hong & Pan, 1998). This action was accepted by Lammers (1998) and Morris and Lammers (1999), but they treated it again as a subspecies, Cylcocodon lancifolia subsp. celebicus (Blume) K. E. Morris & Lammers.

De Candolle (1830) published two new species from India, Codonopsis parviflora Wall. ex A. DC. and Codonopsis truncata Wall. ex A. DC., based on plants discovered by Wallich. The former was transferred to Cyclocodon by Hooker and Thomson (1857), then placed in Campanumoea by Bentham and Hooker (1876), but finally returned to Cyclocodon by Hong (Hong & Pan, 1998). Codonopsis truncata was transferred to Cyclocodon by Hooker and Thomson (1857), then placed in Campanumoea by Diels (1901), but was treated as conspecific with Campanumoea celebica by Clarke (1881), with Campanumoea lancifolia by Tsoong (1935) and Hong (1983), and with Codonopsis lancifolia by Moeliono (1960). Finally, it was treated as a synonym of Cyclocodon lancifolius (Roxb.) Kurz by Hong (Hong & Pan, 1998; Hong et al., 2011). De Candolle (1839) described another species, Campanumoea roylei, which was reduced to the synonymy of Codonopsis ovata by Hooker and Thomson (1857).

Bentham (1834) described two species as new from the western Himalaya and Karakoram, Codonopsis rotundifolia and C. ovata. The former has never been questioned as a distinct species, while the latter has been recognized by all authors except Nasir (1984), who treated it as conspecific with C. clematidea.

Campanumoea lanceolata Sieb. and Zucc. (1835) was transferred to Codonopsis by Trautvetter in 1879, while Lindley’s (1839) Codonopsis lurida is a synonym of C. rotundifolia. Schrenk (1841) discovered a new species of Codonopsis in Middle Asia, but he placed it in Wahlenbergia, as W. clematidea. Codonopsis albiflora was described by Griffith (1854), but was reduced by Clarke (1881) to the synonymy of Campanumoea celebica. In 1856, Hasskarl described Campanumoea cordata from Sumatra, Indonesia, but it was found to be another name for Campanumoea javanica.

Hooker and Thomson (1855) described two new species, Codonopsis gracilis and C. inflata, but one year later Lemaire (1856) separated the former as a new genus, Leptocodon Lem., which was subsequently accepted by Hooker and Thomson (1857).

The year 1857 is important in the history of studies on Codonopsis. Hooker and Thomson published an article containing four distinct new species, C. affinis, C. benthamii, C. foetens and C. subsimplex, all from the Himalaya. In addition, the article includes a description of Griffith’s (1854) unpublished genus, Cyclocodon Griff. ex Hook. f. & Thomson, which contained two species with baccate fruit. This genus has been variously treated since then. Clarke (1881) treated it as a section of Campanumoea, also with baccate fruit. Moeliono (1960) merged all the species of Campanumoea with Codonopsis, but Hong (in Hong & Pan, 1998) separated Cyclocodon from Campanumoea and restored it to generic status. Also in 1857, Ruprecht described two new species; Glossocomia ussuriensis (Ruprecht & Maximowicz, 1857), from Manshuria, the Far East, which was transferred to Codonopsis by Hemsley (1889), and Glossocomia hortensis Rupr. (Ruprecht 1857), which was treated as a synonym of Codonopsis lanceolata by Hemsley (1889). Again in 1857, another new species, Codonopsis leucocarpa, was described by Miquel (1857), but it was treated as a synonym of Campanumoea celebica by Clarke (1881).

In 1863 Morren published Campanumoea japonica Sieb. ex Morr., another synonym of Codonopsis lanceolata.

Kurz (1873) described Codonopsis convolvulacea from southern Yunnan, which is distinct from other members of Codonopsis in having the corolla cleft nearly to the base, and thus rotate, and in having globose or ellipsoid root tubers. Komarov (1908) later established a new subgenus, Codonopsis subg. Pseudocodonopsis Kom., based on this species.

Unaware of the name Campanumoea japonica Sieb. ex Morr. (1863), Maximowicz described Campanumoea japonica Maxim. in 1867, which is, of course, an invalid name. This plant was then treated as a variety of Campanumoea javanica by Makino (1908), Campanumoea javanica var. japonica Makino. It was transferred to Codonopsis as a subspecies, Codonopsis javanica subsp. japonica by Lammers (1992).

The year of 1881 is also important, for it was then Maximowicz described Codonopsis viridiflora from southeastern Gansu, and Clarke described two new species from the Himalaya, Wahlenbergia dicentrifolia and Codonopsis griffithii. Although the latter species is a superfluous name of Codonopsis viridis Wall., the first species is very distinct. It was transferred to Codonopsis by Smith (1913a), and treated by Hong (1980) as the single member of Codonopsis subg. obconicicapsula D. Y. Hong based on its obconical capsule and compressed seeds.

Three years late Franchet found a medicinal plant cultivated in the vicinity of Beijing which he described as a new species of Campanumoea, C. pilosula (Franchet, 1884). It was transferred to Codonopsis by Nannfeldt (1930) as Codonopsis pilosula (Franch.) Nannf. This is one of China’s most commonly used medicinal plants and is widely cultivated in China.

Oliver (1891) described two new species from western Hubei of China, Codonopsis henryi and C. tangshen. The former has never been questioned as a distinct species. The specific epithet ‘tangshen’ is derived from the name of a famous Chinese medicinal plant, Dang Shen, the same name given to Codonopsis pilosula. I tend to consider them to be two forms of the same species, Codonopsis pilosula.

Two very important articles were published in 1908 after a silence of nearly 20 years. In a review of the entire genus Komarov (1908) recognized a total of 23 species, including five new ones, C. cardiophylla Diels ex Kom., C. cordifolia, C. subscaposa, C. tubulosa, and C. vinciflora. Subgenera were proposed for Codonopsis for the first time: Eucodonopsis (= subg. Codonopsis) and Pseudocodonopsis Kom. Within subg. Codonopsis Komarov grouped all the species with twining stems into series Volubiles with three subseries, while the others with erect stems were placed in series Erectae. Later that year Chipp (1908) also provided a taxonomic revision of Codonopsis, which recognized 22 species, including four new, C. deltoidea, C. micrantha, C. mollis, and C. pilosa. The first two are still maintained as distinct species whilst C. mollis is now treated as a subspecies of C. thalictrifolia and C. pilosa is a synonym of C. tubulosa. Interesting is that he described a new variety, Codonopsis ovata var. nervosa Chipp based on a collection from Kangding (Tachielu), Sichuan, China, far from the distribution area of C. ovata (NW Pan-Himalaya and S Middle Asia, see Map IV 1–8). That is actually Codonopsis foetens subsp. nervosa (Chipp) D. Y. Hong. He did not accept Komarov’s subgeneric divisions, nor make his own subdivision.

Pampanini (1910) described a new species, Codonopsis draco Pamp. from Hubei of China, which was ignored until 2014. This plant was rediscovered in Hubei over 70 years later and described as a new genus, Echinocodon lobophylla D. Y. Hong (1984). The new combination, Echinocodon draco (Pamp.) D. Y. Hong, has been made (Hong, 2014b).

Hayata (1911) described Codonopsis kawakamii from Taiwan, the only species of the genus in the narrow sense in Taiwan.

Diels identified five further species, one of which was published by Komarov (1908), Codonopsis cardiophylla. The remaining four species were published by Diels (1912): Codonopsis macrocalyx, C. forrestii, C. bulleyana and C. meleagris. Codonopsis macrocalyx was reduced to a synonym of C. benthamii by Hong (2010); C. forrestii has been treated variously by previous taxonomists. Ballard (1939) and Shen (1983) treated it as a variety of Codonopsis convolvulacea, Hong and Ma (1992) and Hong (Hong in Hong et al., 2011) treated it as a subspecies of C. convolvulacea, whereas Grey-Wilson (1995) and Lammers and Klein (2010) restored it to specific status. There has been no dispute among taxonomists on the other two species.

Smith (1913a) transferred Wahlenbergia dicentrifolia to Codonopsis, but it is very isolated in that genus. In the same year he described two new species, both from Yunnan, Codonopsis subglobosa and C. efilamentosa (Smith, 1913b). The latter is a member of Codonopsis subg. Pseudocodonopsis Kom.

Interestingly, Léveillé described eight species from Guizhou and Yunnan under Codonopsis and other generic names between 1904 and 1916 (Léveillé, 1904, 1914, 1915, 1916). In a review of the material in Léveillé’s herbarium, Chamberlain (1977) synonymised all but one of these, C. graminifolia H. Lév. (1915). The status of C. graminifolia has been controversial. It was treated as a synonym of C. limprichtii Lingelsh. & Borza (1914) by a number of authors, e.g. Lammers & Klein (2010), but we (Hong & Ma, 1994; Hong in Hong et al., 2011) found it to be distinct in Codonopsis subg. Pseudocodonopsis Kom., covering all forms with linear leaves (see section 4 in this chapter for more).

Lingelsheim and Borza (1914) described Codonopsis limprichtii as new from Dali, Yunnan Province. It was treated as a variety of C. convolvulacea by Anthony (1926), but Lammers and Klein (2010) argued for its status as a good species.

Nakai (1915) described Codonopsis minima from the Korean Peninsula, which was reduced to a variety of C. ussuriensis by Lee (1986). According to my observations of the holotype (TI) and isotypes (E & K), it is a simple synonym of C. ussuriensis.

Handel-Mazzetti (1924) published four varieties of Codonopsis based on his own collections from SW Sichuan and NW Yunnan, C. foetens var. major, C. macrocalyx var. coerulescens, C. limprichtii var. hirsuta, and C. limprichtii var. pinifolia. The third has been raised to specific rank, C. hirsuta (Hand.-Mazz.) D. Y. Hong & L. M. Ma (1992), while the last one is a synonym of C. graminifolia.

Anthony (1926) provided a key to the world species of Codonopsis, treating 35 species but not including four species not seen by him. In the article he described two new species, C. farreri and C. chimiliensis from the border between China and Myanmar. The two species are both distinct, and no later author has challenged their status as species.

Nannfeldt (1930, 1931, 1936, 1940, 1950) contributed much to the study of Codonopsis. In his first publication Nannfeldt (1930) described C. modesta from N Sichuan as new and recognized C. nervosa (Chipp) Nannf. (= C. ovata var. nervosa Chipp) from NW Sichuan. He found the second species to be very common in W Sichuan and N Yunnan. More importantly, he transferred Franchet’s Campanumoea pilosula to Codonopsis, creating the new combination, Codonopsis pilosula (Franch.) Nannf. In his second publication, Nannfeldt (1931) described two new species, C. alpina and C. macrantha, both from NW Yunnan. The first is distinct, whereas the second is merely a form of C. nervosa with larger flowers. Nannfeldt (1931. fig.1) emphasized with drawings the value of the shape of both calyx lobes and corolla for delimitation of some species. In his third publication Nannfeldt (1936) recorded 17 species of Codonopsis from China without describing any new species. In the fourth publication Nannfeldt (1940) described three species, C. canescens from W Sichuan, and C. glaberrima and C. volubilis, both from Shanxi, but unfortunately only C. canescens is distinct, the other two are synonyms of C. pilosula (Franch.) Nannf.

P. C. Tsoong, my supervisor when I was a graduate student from 1962 to 1966, was the first Chinese author to produce an overall taxonomic revision of the Campanulaceae in China (Tsoong, 1935). For Codonopsis he recognized 20 species, including two new ones, C. argentea from Guizhou (Fanjing Shan) and C. cordifolioidea from NW Yunnan (Fugong). Both have proven to be distinct species with very narrow distributions.

Gagnepain (1948) described a scandent herb from Laos (A. F. G. Kerr 21186, P) as a new species in a new genus, Numaeacampa kerrii Gagnep., and stated that it was closely related to Campanumoea. Takhtajan (1997) and Mabberley (2008) treated Numaeacampa as a synonym of Codonopsis. Lammers (2007b), however, considered the identity of this taxon to be a mystery; he stated, based on Gagnepain’s description, that the plant is unlike anything else in the Campanulaceae. Therefore, he writes: "Numaeacampa Gagnep. = ? and Numaeacampa kerrii Gagnep. = ? ". Thanks to the images of this specimen from P, we determined it as Lycianthes lysimachioides (Wall.) Bitter, a species of the Solanaceae (Wang & Hong, unpublished). The ovary of the genus Lycianthes looks to be inferior, but is actually superior, surrounded by a cupular calyx.

There were few studies on the taxonomy of Codonopsis from 1940 to 1965. Wu (1965) broke the silence when he published his report on the floristic survey of Yunnan (Campanulaceae), recording five species of Codonopsis, of which two were new, C. chlorocodon from Dêqên County and C. gombalana from Gongshan County. They were both recognized at specific rank by Shen & Hong (1983) and by Hong et al. (2011), but later I found that the second one is actually a synonym of C. farreri J. Anthony.

Ludlow, who spent many years on plant collection in the Himalaya, particularly in SE Tibet and Bhutan, described Codonopsis bhutanica as new based on eight collections from the border area between Bhutan and Tibet (Ludlow, 1972). My C. xizangensis D. Y. Hong (1980) from Cona, S Tibet, proved to be a synonym of this species.

Shen and his colleagues (Shen et al., 1975) revised the taxonomy of Codonopsis in Sichuan, recognizing 24 species, of which one, C. levicalyx from N Sichuan, was considered to be new. Codonopsis levicalyx was recognized by Shen & Hong (1983) and Hong et al. (2011), but after critically examining the holotype and a large number of specimens it is here treated as a synonym of C. henryi.

Hara (1978) described Codonopsis nepalensis from Nepal, but after critically examining a large number of specimens of C. rotundifolia ranging from N Pakistan to Nepal, I found Hara’s species to be within the wide variation range of C. rotundifolia.

In 1980, I described three new species from Tibet, Codonopsis longifolia, C. xizangensis, and Leptocodon hirsutus (Hong, 1980). The former two proved to be synonyms, since I was then unaware of Ludlow’s (1972) and Hara’s (1978) publications, while the last was transferred to Codonopsis by Lammers (2001).

Huang (1984) described Codonopsis pianmaensis and C. farreri var. grandiflora from NW Yunnan, but the latter has been reduced (Hong in Hong et al., 2011), while the former is an extreme form of the variable C. benthamii Hook. f. & Thomson.

Studies on the taxonomy of Codonopsis appears to have entered another active period since 1990. First, Grey-Wilson (1990) published an article "A survey of Codonopsis in cultivation", where he used a broadly defined Codonopsis, including Leptocodon and Campanumoea. He recognized and described 33 species, of which two from central Nepal, C. bragaensis and C. nepalensis Grey-Wilson (1990, non Hara, 1978), were new. The first species was said to be very distinct, while the second was renamed as C. grey-wilsonii by Shaw [1996, = Pseudocodon grey-wilsonii (J. M. H. Shaw) D. Y. Hong].

Wang and Xu (1993) described two species as new, C. microtubulosa from N Sichuan and C. retroserrata from S Sichuan. They both were reduced to synonymy in ‘Flora of China’ (Hong et al., 2011). However, due to examination of more materials, field observation, and molecular analysis, they were found to be distinct and thus are restored in the present work.

Lammers and Klein (2010) revised Codonopsis subg. Pseudocodonopsis Kom., adding one more species of uncertain provenance to the genus: Codonopsis macrophylla. It appears to me a form of the very variable Codonopsis convolvulacea subsp. forrestii [= Pseudocodon convolvulaceus (Kurz) D. Y. Hong & H. Sun subsp. forrestii (Diels) D. Y. Hong].

Dash and Mao (2011) published Codonopsis vadsea from the Yarlung Zangbo-Brahmaputra region in the eastern Himalaya, which, based on their figure and description, I treat here as a subspecies of C. tubulosa.

The most active period in the study of Codonopsis and allied genera was from 1900 to 1949, when 33 new species, nearly half of the total, were described. Although only 15 new species were described in 60 years from 1950 to 2011, several of them are distinct. Among them, Codonopsis chlorocodon C. Y. Wu (1965), C. nepalensis Grey-Wilson (1950) (=C. grey-wilsonii J. M. H. Shaw) and C. bhutanica Ludlow deserve particular mention.

Recently a number of new species have been described; Hong (2014a) published six new species from the Pan-Himalaya and one new subspecies from W Sichuan. They are C. bomiensis, C. campanulata, C. elliptica, C. hemisphaerica, C. lixianica, C. reflexa, and C. cardiophylla subsp. megaphylla. At the same year, Wang and Hong (2014) described a distinct species as new from Gaoligong Shan, NW Yunnan, Codonopsis gongshanica.

From Wallich (1824) to the present, a total of 101 species have been described under Codonopsis or first under other generic names and later transferred to Codonopsis. In addition, numerous subspecific and particularly varietal taxa have been published. Ten generic names have been used to accommodate these 101 specific names besides Cyananthus and Melothria, which were wrongly used by Léveillé for two of his new species. It is clear from the above that frequent changes have been made in the taxonomic history of Codonopsis and its allies and many generic names have been involved in the taxonomy of the group. These facts reflect the intricate relationships among Codonopsis and its allies and imply that to clarify the circumscription of Codonopsis is a challenging but very important task.

Circumscription of Codonopsis

Since the establishment of the genus Codonopsis by Wallich in 1824, its circumscription has been questioned and changed greatly, as mentioned at the end of the previous section. Ten generic names have been involved in the circumscription of Codonopsis.

Don (1825) used a new generic name, Glossocomia, to accommodate his new plant, Glossocomia terera from Nepal. This was treated as a synonym of Wallich’s Codonopsis thalictrifolia by Hooker and Thomson (1857). Glossocomia ussuriensis Rupr. & Maxim. (1857) was transferred to Codonopsis by Hemsley (1889), while Glossocomia hortensis Rupr. (1857) was treated also by Hemsley as a synonym of Codonopsis lanceolata. Since then no one used the generic name Glossocomia for any new species.

De Candolle (1839) described Wahlenbergia roylei, Schrenk (1841) described W. clematidea, and Clarke (1881) used the name Wahlenbergia with a question marker for his new species, W. (?) dicentrifolia. All three of these species were transferred to Codonopsis, the former two by Clarke (1881) and the last one by Smith (1913a). Since then, no species of Codonopsis or its allies have been described under Wahlenbergia.

Sprengel (1825) transferred Wallich’s three species of Codonopsis, C. viridis, C. purpurea and C. thalictrifolia, to Campanula L., but his action has never been followed.

Campanumoea was established by Blume (1826) for species with baccate fruit. Siebold and Zuccarini’s (1835) Campanumoea lanceolata from Japan lacking this was transferred to Codonopsis by Trautvetter (1879). Campanumoea pilosula Franch. (1884) was transferred to Codonopsis by Nannfeldt (1930). Campanumoea japonica Sieb. ex Morr. (1863) was treated as a synonym of Codonopsis lanceolata by Komarov (1908). All these three taxa have a loculicidal capsule. The relationship between Campanumoea sensu Blume and Codonopsis has been in dispute. Clarke (1881) expanded the circumscription of Campanumoea by including not only Codonopsis inflata Hook. f., but also Cyclocodon Griff. ex Hook. f. & Thomson (1857). He then divided the newly defined genus into two sections, sect. Campanumoea and sect. Cyclocodon (Griff. ex Hook. f. & Thomson) C. B. Clarke (1881). The two sections share baccate fruit. Clarke’s (1881) concept of Campanumoea was followed by Tsoong (1935) and Hong (1983). Moeliono (1960) merged Campanumoea sensu Clarke into Codonopsis. This action was followed by Lammers (1992a & b), who transferred Campanumoea javanica subsp. japonica (Makino) D. Y. Hong into Codonopsis and made a new combination accordingly, stating that a single character, berry or capsule, is insufficient to justify generic status. In another direction, however, Hong & Pan (1998), based on LM and SEM observations of pollen and seed coats, revealed that the five species of Campanumoea sensu Clarke (1881) form two distinct groups that correspond with Clarke’s (1881) two sections. The pollen in sect. Cyclocodon is 3-colporate and the seed coat has irregularly shaped lumina with rope-like secondary ornamentation on the radial walls, in addition to the plants having erect stems. In sect. Campanumoea the pollen is 5–8-colpate, the seed coat has regular and polygonal lumina with bead-like secondary ornamentation on the radial walls, and the habit is twining. Based on these observations, Hong & Pan (1998) restored Cyclocodon to generic status. Furthermore, Cyclocodon was considered to be more closely related to Platycodon than to Campanumoea. That is, Codonopsis as circumscribed by Moeliono (1960) and his followers is polyphyletic. The action of Hong & Pan (1998) were followed by Lammers (1998 & 1999). Recent molecular data (Borsch et al., 2009; Wang et al., 2013) clearly indicate that Cyclocodon is more closely related to Platycodon than to Campanumoea and Codonopsis. The status of Campanumoea in the narrow sense, whether it should be maintained or merged with Codonopsis, is still controversial (see Fl. China 19: 505. 2011).

Leptocodon (Hook. f. & Thomson) Lem. (Lemair, 1856) was established based on Codonopsis gracilis Hook. f. (Hooker, 1855). Its generic status was recognized by Hooker and Thomson (1857), Clarke (1881), Tsoong (1935), and Hong (1983). Grey-Wilson (1990), however, merged Leptocodon with Codonopsis, a move followed by Lammers (1999 & 2001). The pollen of Leptocodon is similar to that in core Codonopsis (Erdtman, 1952; Avetisjan, 1986; Morris & Lammers, 1997; Hong & Pan, 2012).

Pampanini’s (1910) Codonopsis draco has been transferred to Echinocodon, and the new combination made (Hong, 2014b).

In Codonopsis s. s., Komarov (1908) separated two species with rotate corolla as subgenus Pseudocodonopsis; Hong (1980) treated Codonopsis dicentrifolia (C. B. Clarke) W. W. Smith as the only representative of subgenus Obconicicapsula D. Y. Hong, based on its obconical ovary and fruit and compressed seeds. As a result of rather extensive observations on pollen morphology of Codonopsis and its allies Morris and Lammers (1997) recognized four types of pollen morphology: Type I for pollen of Codonopsis subg. Codonopsis, Leptocodon and Campanumoea sect. Campanumoea, Type II for pollen of Campanumoea sect. Cyclocodon (= Cyclocodon), Type III for pollen of Codonopsis subg. Pseudocodonopsis (= Pseudocodon), and Type IV for pollen of Codonopsis subg. Obconicicapsula (= Himalacodon). Hong and Pan (2012) made even more extensive observation on the pollen morphology of this group. Our results confirmed the conclusions of Morris and Lammers’ (1997) and showed that the pollen of Campanumoea s. str. differs from that of subg. Codonopsis in having shorter colpi (colpus length/polar axis length = 0.54–0.57 vs. ≥ 0.7) and a granular colpus membrane (vs. spinulose). Thus Hong and Pan (2012) defined this type of pollen as the Codonopsis Type (Type I) Campanumoea subtype. Their observations also revealed that Codonopsis subg. Codonopsis is heterogeneous in pollen morphology, i.e. the pollen of C. purpurea and C. chimiliensis is basically similar to Morris and Lammers’ (1997) Type IV, but with a smooth colpus membrane and they thus assigned it to the Obconicicapsula Type (= Type IV) Purpurea subtype.

According to Shen & Hong (1983) and Hong (in Hong et al., 2011), Codonopsis comprises three subgenera, subgs. Codonopsis, Obconicicapsula D. Y. Hong and Pseudocodonopsis Kom. In the ITS trees presented by Eddie et al. (2003) and Haberle et al. (2009), Codonopsis, even in Hong’s (in Hong et al., 2011) narrow sense, is not monophyletic; Codonopsis dicentrifolia, the only species of subg. Obconicicapsula, forms a clade with Cyananthus lobatus Wall. ex Benth. Therefore, Codonopsis is polyphyletic from palynological and molecular data.

According to the above review of the history, further multidisciplinary studies should be conducted on Codonopsis and its allies to address the following questions.

1. Should Campanumoea and Leptocodon be included in or kept separate from Codonopsis?

2. Is Codonopsis in the narrow sense monophyletic or polyphyletic?

3. What are the phylogenetic relationships within Codonopsis and its allies?

4. How should Codonopsis be circumscribed?

Systematic Position of Codonopsis and Allied Genera

There has been a common view that Codonopsis and allied genera belong to the Campanulaceae. The circumscription of the Campanulaceae, however, has involved different views. De Candolle (1839) treated it in a narrow sense, separating it from the Cyphiaceae (including Lobelia), but including Pentaphragma with a question marker. Bentham and Hooker (1876) were the first to circumscribe the family broadly, including Lobelieae and Cyphieae. From then on botanists have been divided into two schools. Endlicher (1841), Baillon (1880), Wettstein (1924), Hutchinson (1973), Thulin (1975), Kovanda (1978), Dahlgren (1983), Takhtajan (1987, 1997), Lammers (1992), Cosner et al. (1994), and Gustafsson and Bremer (1995) hold the narrow view, whereas Schönland (1889), Bessey (1915), Wagenitz (1964), Cronquist (1981), Thorne (1992), Lammers (1998, 2007), and Brummitt (2007) hold the broad view.

According to Gustafsson and Bremer (1995) the Campanulaceae clade includes five families: Campanulaceae, Nemacladaceae, Cyphiaceae, Cyphocarpaceae and Lobeliaceae. Lammers (1998, 2007), however, treated these five families as five subfamilies in the Campanulaceae s. l.

Regardless of whether one accepts the family broadly or narrowly, the Campanulaceae (Gustafsson & Bremer, 1995) or the Campanuloideae (Lammers (1998, 2007)) is a monophyletic group. Therefore, it is a matter of taste to take either view. As we prefer to recognize the family in the strict sense, Codonopsis and its allies are members of the Campanulaceae s. s. Codonopsis and its allies have been placed in the Campanulaceae since de Candolle (1830, 1839).

Phylogenetic relationships of the genera within the Campanulaceae s. s., however, are far from well established. As stated by Kovanda (1978), The family Campanulaceae……is rather natural and homogeneous but its subdivision presents serious problems……. De Candolle (1839) divided the family into three tribes, Wahlenbergieae Endl., Campanuleae G. Don and Mercieae A. DC., and placed Campanumoea, Codonopsis, Canarina and Platycodon together with Wahlenbergia in Wahlenbergieae. His Codonopsis is actually Cyclocodon, while he treated true Codonopsis, including the first described three species of Codonopsis, i.e. C. viridis, C. thalictrifolia and C. purpurea, as members of Wahlenbergia. Two issues have been involved in the systematic position of Codonopsis and allied genera. First, since de Candolle (1839), Codonopsis and its allies have been placed together with Wahlenbergia primarily because they share a loculicidal capsule. Second, the circumscription of Codonopsis and its phylogenetic relationships with its allies have been unclear. They were distantly separated and placed in different groups. Schönland (1889) divided the tribe Campanuleae (= Campanulaceae s. s.) into three subtribes, with Codonopsis, Campanumoea, Leptocodon and Cyananthus along with Wahlenbergia in the Wahlenberginae, Platycodon, Microcodon and Musschia in the Platycodinae, and Canaria and Ostrowskia together with Campanula and Adenophora in the Campanulinae. Some authors went even further. For example, Kolakovsky (1987 & 1994) partitioned Codonopsis and its allies into two groups and placed Codonopsis, Leptocodon and Platycodon, all with capsular fruit, together with Wahlenbergia in the Wahlenbergioideae, while he placed Canarina and Campanumoea with baccate fruit in the separate subfamily, Canarinoideae. Although Takhtajan (1997) was the first botanist who separated Codonopsis, Leptocodon, Campanumoea, Cyananthus, Platycodon from Wahlenbergia, he also separated Canarina and Ostrowskia as two independent subfamilies, Canarinoideae and Ostrowskioideae. Recent molecular work (Eddie et al., 2003; Haberle et al., 2009) have provided insights into the phylogenetic relationships within the Campanulaceae s. s., but unfortunately sampling in those studies was not adequate enough to clarify the systematic position of Codonopsis and its relationships with its allies, which remains our present task. Chapter III provides a detailed discussion of this issue.

Taxonomical Revisions of Codonopsis and Allied Genera

The first significant taxonomic revision of Codonopsis and its