Phylogeny and Classification of the Bamboos (Poaceae: Bambusoideae) Based on Molecular and Morphological Data Wei-Ping Zhang Iowa State University

Home , Alvimia, Ampelocalamus, Anomochloa, Anomochlooideae, Apoclada, Arthrostylidiinae

Iowa State University Capstones, Theses and Retrospective Theses and Dissertations Dissertations

1996 Phylogeny and classification of the bamboos (Poaceae: Bambusoideae) based on molecular and morphological data Wei-Ping Zhang Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/rtd Part of the Botany Commons

Recommended Citation Zhang, Wei-Ping, "Phylogeny and classification of the bamboos (Poaceae: Bambusoideae) based on molecular and morphological data " (1996). Retrospective Theses and Dissertations. 11430. https://lib.dr.iastate.edu/rtd/11430

This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UME fihns the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reprodactioii is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, begiiming at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. BSgher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UME directly to order. UMI A Bell & Howell Information Company 300 North Zed) Road, Ann Aibor MI 48106-1346 USA 313/761-4700 800/521-0600

Phylogeny and classification of the bamboos (Poaceae: Bambusoideae) based on molecular

and morphological data

Wei-Ping Zhang

A dissertation submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Major: Botany (Systematics and Evolution)

Major Professors: Lynn G. Clark and Jonathan F. Wendel

Iowa State University

Ames, Iowa

1996

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 11

Graduate College Iowa State University

This is to certify that the doctoral dissertation of

Weiping Zhang . has met the dissertation requirements of Iowa State University

Signature was redacted for privacy. o-major Professor

Signature was redacted for privacy. or Program

Signature was redacted for privacy. For the iii

TABLE OF CONTENTS

LISTOFHGURES v

LIST OF TABLES vi

ABSTRACT vii

CHAPTER L GENERAL INTRODUCTION 1 Introduction 1 Dissertation Organization 2 Statement of Researcti and Literature Review 3 Literature Cited 9

CHAPTER 2. PHYLOGENY AND CLASSMCATION OF THE BAMBUSOIDEAE (POACEAE) BASED ON ndhF SEQUENCE AND MORPHOLOGICAL DATA 14 Abstract 14 Introduction 15 Materials and Methods 17 Results 41 Discussion 62 Taxonomic Treatment 75 Acknowledgments 80 Literature Cited 80

CHAPTERS. BAMBOOZLED AGAIN!: INADVERTENT ISOLATION OF FUNGAL rDNA SEQUENCES FROM BAMBOOS (POACEAE: BAMBUSOIDEAE) 88 Abstract 88 Introduction 89 Materials and Methods 91 Results 97 Discussion 105 Acknowledgments 112 i V

Literature Cited 112

CHAPTER 4. GENERAL CONCLUSIONS 119 General Discussion 119 Recommendations for Future Research 121 Literature Cited 123

APPENDIX. COMPLETE ndhF GENE SEQUENCE DATA MATRIX 125

ACKNOWLEDGMENTS 161 V

LIST OF HGURES

CHAPTER 2.

Figure 1. Silica body diagram 38

Figure 2. One most parsimonious tree of the complete ndtiF sequence data set 44

Figure 3. Strict consensus tree of the complete ndhF sequence data set 46

Figure 4. Neighbor-Joining tree of the complete ndhF sequence data using Kimura 2- parameter distance measure 49

Figure 5. One most parsimonious tree of the morphological and anatomical data set 52

Figure 6. Strict consensus tree of the morphological and anatomical data set 54

Figure 7. One most parsimonious tree of the reduced ndhF sequence data set 57

Figure 8. One most parsimonious tree of the combined data sets 60

CHAPTER 3.

Figure 1. Aligned data matrix of 5.8S sequences for representative ascomycetes (A), basidiomycetes (B) and angiospermous plants (P) 99

Figure 2. Strict consensus of the 15 shortest trees found in parsimony analysis of plant, fungal, and "contaminating fungal" 5.8S sequences, rooted with angiosperms 102

Figure 3. Strict consensus of 414 equally most parsimonious trees found in parsimony analysis of 5.8S sequences from basic^omycetes and putative basidiomycetous fungi associated with bamboos, rooted wi^ three representative ascomycetes 104

Figure 4. Scannmg electron micrographs of the leaf surfaces of four species of bamboos 110 V i

LIST OF TABLES

CHAPTER 2.

Table 1. Taxa, voucher specimens, and GenBank accession numbers 18

Table 2. Morphological and leaf anatomical characters and character states used in the analysis 24

Table 3. Data matrix of morphological and anatomical characters and character states used for cladistic analysis in the bambusoid clade 31

Table 4. Indels (insertions and deletions) in the Bambusoideae ndhP gene sequences 41

Table 5. Sequence variation within the complete ndhF sequence data set and the subsets of bambusoid taxa 61

Table 6. Proportion of nucleotide difference and standard errors of ndhF gene among different groups of bambusoids and other grasses 63

CHAPTER 3.

Table 1. Taxa, vouchers, and GenBank accession numbers for species studied 91

Table 2. Primers (5' to 3') used for PCR ampUfication of the ITS region 95 vii

ABSTRACT

Chloroplast ndhF gene sequences, morphological and leaf anatomical characters were analyzed separately and as combined data sets to reconstruct the phylogeny of subfamily

Bambusoideae. The analyses further confirmed that the monophyletic bambusoid clade consists of only two monophyletic lineages: the woody bamboos, and the herbaceous olyroid bamboos. Buergersiochloa was resolved as the basal lineage in the herbaceous olyroids, whereas Pariana/Eremitis was sister to the rest of the olyroids. The woody bamboos were divided into two main groups: temperate woody bamboos and tropical woody bamboos; and the tropical clade was further subdivided along geographic lines into the Old World Bambuseae and the New World Bambuseae. Puelia was resolved as the most basal Uneage of the 'higher grasses' and thus was excluded from the bambusoid clade. Streptogyneae joined the oryzoids including the Oryzeae and Ehrharteae to form another separate monophyletic clade. The results of this study clearly indicate that the current Bambusoideae is not acceptable phylogenetically.

Therefore a new circumscription of the Bambusoideae with two tribes is proposed in which only members of the bambusoid clade are included; the woody bamboos are classifed as one tribe, the Bambuseae, and the herbaceous olyroid bamboos are classified in another tribe, the

Olyreae.

In order to provide additional resolution of phylogenetic relationships within the bambusoid clade, an attempt to generate nuclear ribosomal ITS sequence data of bamboos was made. However, polymerase chain reaction amplification (PGR) led to the recovery of fungal instead of bamboo sequences. Phylogenetic analyses based on the 5.8S sequences indicated that all the sequences belonged to basidiomycetes and that none was an ascomycete. A diverse assemblege of basidiomycetes was isolated from different bamboo hosts and various fungi coexisted in the same host plant. No evidence showed that closely related fungi associated viii witJi closely related bamboo hosts. True bamboo ITS sequences were obtained only after leaf surface sterilization before DNA isolation. This study highlights the possibility of inadvertent

PGR amplification of "contaminating DNA" in molecular phylogenetic studies. The results also indicate that a close ecological association between epiphytic basidiomycetes and bamboo leaves may exist, but fiuther study is needed. I

CHAPTER 1. GENERAL INTRODUCTION

Introduction

The bamboos are known for their uses in fishing, papermaking, landscape gardening,

handicrafts, medicine, art, and food, and more recendy have become publicized as the main source of food for the endangered giant panda. Bamboos have played a very important role in

the cultural heritage of Far Eastern countries such as China, Korea, and Japan. Bamboos were

used to record historical events and cultural treasures before the invention of paper. In tropical regions, the bamboos were regarded as the "timber of the poor" because people used bamboos for house construction, and house and farm implements. The importance of bamboos is evne

more evident now as the tropical forests diminish.

The bamboos are distributed mainly in tropical, subtropical, and mild-temperate regions in south and southeastern Asia, America, and AMca. The majority of species are the woody bamboos with which most people are familiar, but there is also a group of herbaceous bamboos that inhabit the understory of tropical and subtropical forests. Although bamboos have a glorious history because of their cultural importance, and they continue to be economically valuable today, relatively littie is known about bamboo anatomy, cytology, genetics, morphology, ecology, physiology, and phylogeny. Taxonomic and phylogenetic

understanding of the bamboos, which is basic to all other areas of study, is still undeveloped.

The elaboration of vegetative morphology and reduction of floral morphology, the difficulty of access to some geographic regions where there is much bamboo diversity, the low quality of many bamboo specimens, and the long intervals between flowering in woody bamboos, have all contributed to making the study of bamboo taxonomy and basic biology more difficult. The 2 unavailability of complete morphological information has resulted in the construction of classifications that rely on a biased selection of characters. In the past few decades, more attention has been paid to the study of bamboos, especially regarding leaf anatomy, embryology, cytology, and detailed morphology, which has lifted bamboo systematics to a new stage; however, it is still not up to the same level as other grasses.

More recently, the emergence of molecular techniques has had a strong impact on plant systematics because molecular data are very powerful for phylogenetic reconstruction.

Several molecular studies already have been conducted on the grass family (Hamby and

Zimmer 1988; Doebley et al. 1990; Davis and Soreng 1993; Cummings et al. 1994; Nadot et al. 1994; Clark et al. 1995; Barker et al. 1995; Duvall and Morton 1996; Liang and Hilu 1996), and especially within the last two years, much has been learned about the phylogenetic position of taxa traditionally placed within the bamboo subfamily. Clark et al. (1995) demonstrated that the Bambusoideae, as traditionally recognized, were polyphyietic, and that certain tribes of herbaceous forest grasses formerly classified as bamboos were actually representatives of the earliest divergences within the family. The remaining core of the bamboos, however, had never been studied in detail using molecular data or cladistic techniques. The objectives of this dissertation were to test the monophyly of the bambusoid clade recovered in the previous analyses, and elucidate phylogenetic relationships within the bamboos using both molecular and morphological data.

Dissertation Organization

This dissertation is composed of two papers prepared for publication in scientific journals. An introduction to the bamboos, the subject of this doctoral research, is provided, followed by a more explicit statement of research and hterature review in the general 3 introduction (Chapter 1). Chapter 2 consists of the manuscript entitled "Phylogeny and classification of the Bambusoideae (Poaceae) based on ndhF sequence and morphological data," which is intended for submission to Systematic Botany. Chapter 3 is a manuscript entitied "Bamboozled again!: Inadvertent isolation of fungal rDNA sequences from bamboos

(Poaceae: Bambusoideae)." This has been accepted for publication to Molecular Phylogenetics and Evolution, and is currently in press pending revision. A general conclusion summarizes the findings of this research in Chapter 4. The ndhF gene sequences used in the analyses in

Chapter 2 are included in their entirety in the appendix.

Statement of Research and Literature Review

Davis and Soreng (1993; pp. 1452-1453) noted in their analysis of chloroplast DNA restriction site variation that "the placement of Bambusoideae relative to other grasses thus remains a key to our understanding of the earliest diversification of grasses." Two years later, the accuracy of that statement was confirmed by Clark et al. (1995), who, using ndhF sequence data, demonstrated that the broadly defined Bambusoideae were indeed highly heterogeneous and polyphyletic, and that three of its herbaceous tribes represented the earliest radiations within the grass family. Both Clark et al. (1995) and Duvall and Morton (1996), based on relatively limited sampling within the true bamboos, recovered a monophyletic bambusoid clade, nested well within the higher grasses, consisting of the woody bamboos and the herbaceous olyroid bamboos. However, four taxa included within the Bambusoideae in recent global treatments remained unsampled, and further study within the bambusoid clade was necessary in order to explore the boundaries of the bambusoid clade and its phylogeny, and redefine the Bambusoideae. 4

We accordingly sampled 38 species of the bambusoid clade, and also included one species each of Buergersiochloa and Puelia (two of the four taxa previously unsampled), and exemplars from other tribes previously classified within the Bambusoideae (Clayton and

Renvoize 1986; Watson and Dallwitz 1992) and other major lineages of grasses. Joinvillea ascendens (Joinvilleaceae) was used as the outgroup in analyses of the complete family, whereas one of the basal grasses was used for more detailed analyses of the bambusoid clade.

A general phylogenetic reconstruction was attempted based on ndhF sequence data, and then a combination of ndhF sequence data and morphological/leaf anatomical data was analyzed to estimate phylogenetic relationships within the bambusoid clade.

The previous resolution of Anomochloeae, Streptochaeteae, and Phareae as the basal lineages within the grass family was confirmed, and Buergersiochloa was found to resolve within the bambusoid clade. A novel and somewhat surprising result was the resolution of

Puelia as the probable basal lineage within the higher grasses. The Streptogyneae, another tribe previously regarded as herbaceous bamboos, continued to associate with the rice clade, although this placement was only weakly supported. The relationship between the oryzoid and bambusoid clades was explored further using both molecular and morphological data, as were the deeper branches within the bambusoid clade. The oryzoid and bambusoid clades were supported as divergent, monophyletic lineages. Within the bambusoid clade, two major clades were recovered: the herbaceous olyroid bamboos with Buergersiochloa in a basal position, and the woody bamboos. Each of these was strongly supported, especially in the combined data sets analysis, but while good resolution within the olyroid clade was obtained, further resolution within the woody bamboo clade was not satisfactory. The results of these analyses are described in Chapter 2.

It became apparent that a faster-evolving region, preferably nuclear, would be needed to resolve relationships within the woody bamboos. We therefore chose the nuclear ribosomal

DNA rrS regions, which were expected to provide more informative characters. A preliminary 5

attempt to amplify and sequence the ITS regions from eight woody bamboo species instead

resulted in the inadvertent amplification and sequencing of fungal ITS regions. While

uninformative from the standpoint of bamboo phylogeny, this phenomenon was interesting as

it was apparently caused by the close association of fiingi and bamboos, coupled with the use

of universal ITS primers for amplification of the DNA. All known fungal endophytes of grasses are ascomycetes, but further investigation showed that the fungal sequences obtained from the bamboo hosts were all basidiomycetes. This work is described in Chapter 3.

The Bambusoideae have been regarded as one of the most poorly understood groups in the grass family (Soderstrom and Ellis 1987). The concept of the subfamily including both woody and herbaceous bamboos has evolved over centuries with the development of the grass family classification. The subfamily concept was so heterogeneous that it varied from the inclusion of two or a few woody bamboo groups in most early classifications (Kunth 1815,

1822; Ruprecht 1839; Munro 1868; Bentham 1881; Hackel 1887; Gamble 1896; E. G. Camus

1913; A. Camus 1935; Roshevits 1946; Holtmm 1956) to as many as 13 to 15 tribes in most recent systems (Clayton and Renvoize 1986; Soderstrom and Ellis 1987; Tzvelev 1989;

Watson and Dallwitz 1992).

The concept of bamboo was first described by Rumphius (1750), but a more precise recognition did not appear until a half century later by Kunth (1815, 1822) and Agardh (1822).

Kunth (1815,1822) listed ten groups of grasses, the bamboos was one among them. Agardh

(1822) treated the grasses the same way as Kunth did, with the bamboos as one separate group in the grasses. However, it was Nees (1829, 1835) who brought much more attention to the concept by his detailed study of bamboos. Nees (1829) divided the bamboos into two, and later (1835) three, groups: the Bambusa group, the Arundinaria group, and the Streptochaeta group. Nees' later treatment (1835, not 1836) was the first time that herbaceous taxa were associated with the woody bamboos, probably based on similarities of the hermaphrodite spikelets, six stamens, and three stigmas. As a subfamily concept, it was first introduced by 6

Ruprecht (1839), who excluded Streptochaeta and put the Bambusa and Arundinaria groups

into one subfamily on a similar level with other grasses. Most of the later classifications

(Munro 1868; Bentham 1881; Gamble 1896; E.-G. Camus 1913; A. Camus 1935) just adopted

Nees' (1836) or Ruprecht's (1839) treatments with minor modifications.

Leaf anatomical and embryo studies provided a tremendous impact on the delimitation

of the Bambusoideae (Brandis 1907; Krause 1909; Avdulov 1931; Prat 1931; Michaud-Page

1947; Metcalfe 1956,1960; Reeder 1962). These studies clearly indicated the affinities

between herbaceous and woody bamboos. The combination of morphological and anatomical

information made possible the appearance of a modem system of classification for the

bambusoid subfamily. Calderon and Soderstrom (1980) reviewed in detail the history of

bamboo classification and changes in the concept of the subfamily up to that date. Clark et al.

(1995) reviewed and compared the major systems of classification of the Bambusoideae in this century.

The different patterns of bamboo infiorenscence development, especially of the woody bamboos, were used to divide the woody bamboos into two major groups (Keng 1959;

McClure 1960; Keng 1982, 1987). One group included those taxa with unrestricted inflorescences, which have buds at the bases of spikelets, and the buds continue developing to form new spikelets. Most the Old World tropical and some temperate bamboos belong to this group. The otiier group contained bamboos with restricted inflorescences, which develop in one period of growth and lack buds at the bases of spikelets. Some Old World temperate bamboos and most New World bamboos have restricted inflorescences. Using these two inflorescence types, and some vegetative characters, such as rhizome types, culm branching complements, Keng (1959) and Keng (1982, 1987) classified the woody bamboos as two supertribes: Bambusatae with three tribes, and Arundinariatae with two tribes. These systems were based on several key characters, which have limited value in phylogeny reconstruction. 7

Of most relevance for the present study are the most recent global treatments of the

Bambusoideae, in which the woody bamboos, several tribes of putative herbaceous bamboos, and the rices were all classified within the Bambusoideae. Clayton and Renvoize (1986) broadly defined the subfamily based on the presence of fusoid cells and arm cells in the chlorenchyma, the presence of papillae on the leaf epidermis, a tendency toward trimerous floral parts, and the bambusoid or oryzoid embryo. They indicated that the subfamily was broadly treated, and highly heterogeneous. Five groups were recognized loosely: 1)

Bambuseae, characterized by bambusoid leaf anatomy and embryo, arborescent habit with complex branching systems and inflorescences, trimerous spikelets, and distributed mainly in tropical and warm temperate areas; 2) Anomochloeae and Streptochaeteae, a small group characterized by bambusoid leaf anatomy, anomalous spikelets, and found in forest undergrowth in the New World tropics; 3) Olyreae, Parianeae, and Phareae, a primarily New

World tropical group, all with unisexual spikelets, the latter included in the subfamily only because of the bambusoid leaf anatomy and embryo; 4) Oryzeae, Phyllorachideae, and

Streptogyneae, a mixed group lacking die typical bambusoid leaf anatomy and embryo, but included in the subfamily because there were no clear diagnostic characters to separate them from the olyroid group; 5) Phaenospermatae, Ehrharteae, Diarrheneae, and Brachyelytreae, isolated groups with no clear affinities, but shghdy closer to the bambusoids than to any odier group.

Soderstrom and Elllis (1987) established the 'core Bambusoideae' and 'partial

Bambusoideae' concepts by using ten diagnostic morphological and anatomical characters: embryo formula, hilum shape, embryo size, number and appearance of the lodicule, seedling morphology, bicellular microhair shape, mesophyll arrangement, presence or absence of arm cells and fusoid cells, midrib vasculature, and silica body shape and orientation. The 'core

Bambusoideae' was regarded as a monophyletic group which contained two supertribes, the

Olyrodae and Bambusodae. Four herbaceous bamboo tribes, Anomochloeae, 8

Buergersiochloeae, Olyreae, and Streptochaeteae were recognized as Olyrodae, while the woody bamboos, Bambuseae, formed a separate supertribe, Bambusodae. 'Partial

Bambusoideae' was used to describe those tribes showing some relationships to the 'core

Bambusoideae' but not fitting within it: Guaduelleae, Oryzeae, Phareae, Puelieae,

Streptogyneae, and Zizanieae.

Based on numerous morphological and cytological characters, such as the woody or herbaceous habit, rhizome type, number of florets per spikelet, number of lodicules, stamens and stigmas, caryopsis morphology, and chromosome number, Tzvelev (1989) classified 14 tribes within the Bambusoideae. The woody bamboos were divided into six tribes, while ±e herbaceous tribes Atractocarpeae (Puelieae), Streptogyneae, Streptochaeteae,

Buergersiochloeae, Olyreae, Parianeae, Leptaspideae (Phareae), and Anomochloeae also were included in the subfamily.

Kellogg and Watson (1993) used 172 morphological and anatomical characters for cladistic analysis to reconstruct the bambusoid phytogeny. Five monophyletic clades in the subfamily were resolved. Olyreae, Phareae, Buergersiochloeae, and Phyllorachideae formed a clade, which was sister to the Oryzeae clade. Streptochaeteae, Streptogyneae, Puelieae, and

Guaduelleae were resolved as a monophyletic clade sister to the woody bamboo clade. The

Ehrharteae clade was interpreted as paraphyletic or basal to the clade including woody bamboo clade and Guaduelleae and allies clade.

Using a phenetic approach, Watson and Dallwitz (1992) selected from 496 listed morphological, anatomical, cytological, and ecological characters to delimit the Bambusoideae.

Based on the culm habit, number of lodicules and stigmas, and the appearance of arm cells and fusoid cells, two supertribes, Olyrodae and Bambusodae, were adopted from Soderstrom and

EUis (1987). The Olyrodae included 12 herbaceous tribes: Oryzeae, Olyreae, Centotheceae,

Anomochloeae, Brachyelytreae, Diarrheneae, Ehrharteae, Phaenospermatae, Phyllorachideae, 9

Phareae, Streptochaeteae, and Stxeptogyneae. The woody bamboo supertribe Bambusodae included three tribes: Guaduelleae, Puelieae, and Bambuseae.

Recent increasing ease of use of molecular tools made it possible to apply DNA information to explore bamboo phylogeny. Although several studies have been reported (Li

1989; Watanabe et al. 1993; Friar and Kochert 1994; Kobayashi, in press), the limited sampling or the focus on a particular groups in these earlier studies fails to contribute to an understanding of bambusoid phylogeny. Clark et al. (1995) for the first time used molecular data to reconstruct the phylogeny of the Bambusoideae. The chloroplast ndhF gene sequence data analyses indicated that the monophyletic bambusoid clade only included the woody bamboo tribe Bambuseae and herbaceous olyroid bamboos (Olyreae and Parianeae).

Anomochloeae, Streptochaeteae, and Phareae were resolved as the most basal lineages in the grass family. Diarrheneae, Phaenospermateae, and Brachyelytreae were associated with the pooid clade, while the Oryzeae, Ehrharteae and Streptogyneae were grouped together in an oryzoid clade. The Centotheceae, which were nested within the PACC (Panicoideae,

Arundinoideae, Chlorodoideae, and Centotheceae) clade, were shown to be far distant from the bamboos.

Literature Cited

AGARDH, C. A. 1822. Aphorismi Botanici. Lund: Literis Berlingianis.

AVDULOV, N. P. 1931. Kario-sistematicheskoye issledovaniye semeystva zlakov.

Prilozheniye 44 k Trudy po prikladnoy botanike, genetike I selektsii, 428 pages,

Leningrad.

BARKER, N. P., H. P. LINDER, and E. H. HARLEY. 1995. Polyphyly of Arundinoideae

(Poaceae): Evidence from rbcL sequence data. Systematic Botany 20:423^35. 10

BENTHAM, G. 1881. Notes on Gramineae. Journal of the Linnean Society (Botany) 19:14

-134.

BRANDIS, D. 1907. Remarks on the structure of bamboo leaves. Transactions of the

Linnean Society of London, 2nd Ser. Botany: 69-92.

CALDERON, C. E., and T. R. SODERSTROM. 1980. The genera of Bambusoideae

(Poaceae) of the American continent: keys and comments. Smithsonian Contributions

to Botany 44: 1-27.

CAMUS, A. 1935. Classificadon des Bambusees. Archives du Museum d'Histoire Naturelle

(Paris) 12:601-603.

CAMUS, E.-G. 1913. Les Bambusees — monographie, biologie, culture, principaux

usuages. Paris: P. Lechevalier.

CLARK, L. G., W. ZHANG, and J. F. WENDEL. 1995. A phylogeny of the grass family

(Poaceae) based on ndhF sequence data. Systematic Botany 20: 436-460.

CLAYTON, W. D., and S. A. RENVOIZE. 1986. Genera graminum. London: Her

Majesty's Stationery Office.

CUMMINGS, M. P., L. M. KING, and E. A. KELLOGG. 1994. Slipped-strand mispairing

in a plastid gene: rpoC2 in grasses (Poaceae). Molecular and Biological Evolution 11:

1-8.

DAVIS, J. I. and R. J. SORENG. 1993. Phylogenetic structure in the grass family (Poaceae)

as inferred from chloroplast DNA restriction site variation. American Journal of

Botany 80: 1444-1454.

DOEBLEY, J., M. DURBIN, E. M. GOLENBERG, M. T. CLEGG, and D. P. MA. 1990.

Evolutionary analysis of the large subunit of carboxylase {rbcL) nucleotide sequence

data among the grasses (Poaceae). Evolution 44: 1097-1108. 11

DUVALL, M. R., and B. R. MORTON. 1996. Molecular phylogenetics of Poaceae: An

expanded analysis of rbcL sequence data. Molecular Phylogenetics and Evolution 5:

352-358.

FRL\R, E., and G. KOCHERT. 1994. A study of genetic variation and evolution of

Phvllostachvs (Bambusoideae: Poaceae) using nuclear restriction fragment length

polymorphisms. Theorical Application of Genetics 89; 265-270.

GAMBLE, J. S. 1896. The Bambuseae of British India. Annals of the Royal Botanic

Garden, Calcutta 7: 1-133.

HACKEL, E. 1887. Gramineae. Pp. 1-97 in Die natiirlichen Pflanzenfamilien , Volume 2,

eds. A. Engler & K. Prantl. Leipzig: W. Engelmann.

HAMB Y, T. K., and E. A. ZIMMER. 1988. Ribosomal RNA sequences for inferring

phylogeny within the grass family (Poaceae). Plant Systematics and Evolution 160:

29-37.

HOLTTUM, R. E. 1956. The classification of Bamboos. Phytomorphology 6:73-90.

BCELLOGG, E., and L. WATSON. 1993. Phylogenetic studies of a large data set. I.

Bambusoideae, Andropogonodae, and Pooideae (Gramineae). Botanical Review 59:

273-343.

KENG, P. C. 1982. A revision of the genera of bamboos from the world. Journal of

Bamboo Research 1:1-19.

. 1987. A systematic key to the tribes and genera of subfam. Bambusoideae

(Gramineae) occurrence in China and its neighborhoods. Journal of Bamboo Research

6: 13-28.

KENG, Y. L. 1959. Flora Illustralis Plantarum Primarum Sinicarum, Gramineae. Peking.

KOBAYASHI, M. in press. Phylogeny of world bamboos analyzed by RFLP of chloroplast

DNA. Pp. xx-xx in The Bamboos, ed. G. Chapman. London: Academic Press. 12

KRAUSE, E. H. L. 1909. Ein Besserungsversuch am system der gramineen. Beihefte zum

Botanischen Centralblatt 25:421-489.

KUNTH, C. S. 1815. Considerations generales sur les Graminees. Memoires du Museum

d'Histoire Naturelle 2: 62-75.

. 1822. Synopsis Plantarum. Volume 1,491 pages. Paris; F. G. Levrault.

LI, S. 1989. Application of isozymes in Arant/i/iarfa classification. Journal of Bamboo

Research 8: 2-21.

LIANG, H., and K. W. HILU. 1996. Application of the warAT gene sequences to grass

systematics. Canadian Journal of Botany 74: 125-134.

METCALFE, C. R. 1956. Some thoughts on the structure of bamboo leaves. Botanical

Magazine, Tokyo 9:148.

. 1960. Anatomy of the monocotyledons /. Grammecte. Pp 540-713. Oxford:

Clarendon Press.

MCCLURE, F. A. 1966. The Bamboos Pp. 11-121. Boston: Harvard University Press.

MICHAUD-PAGE, V. 1947. Leaf anatomy of Streptochaeta and the relation of diis genus

with bamboos. Bulletin of the Torrey Botanical Club 74: 232-239.

MUNRO, W. 1868. A Monograph of the Bambusaceae, Including Description of all the

species. Transactions of the Linnean Society of London 26: 1-157.

NADOT, S., R. BAJON, and B. LEJEUME. 1994. The chloroplast gene rps4 as a tool for

the study of Poaceae phylogeny. Plant Systematics and Evolution 191: 27-38.

NEES VON ESENBECK, C.G.D. 1829. Agrostologia Brasiliensis . Pp. 608. Stuttgart:

Tubingen.

. 1835. Bambuseae Brasilienses: Recensuit, et Alias in India Orientali Provenientes

Adjecit. Linnaea 9: 461-494. 13

. 1836. Genera (of Grasses). Pp. 378-383 in A natural system of botany, second

edition, eds. J. Lindley. London: Longman, Rees, Orme, Brown, Green, and

Longman.

PRAT, H. 1931. L'epiderme des Graminees, Etude Anatomique et Systematique. Paris:

Faculte des Sciences de Paris.

REEDER, J. R. 1962. The bambusoid embryo: A reappraisal. American Journal of Botany

49; 639-641.

ROSHEVrrS, R. Y. 1946. Sistema zlakov v svyazi s ikh evolyutsiei. Leningrad: Komarav

Botanical Institute, U. S. S. R. Academy of Sciences. [Evolugao e sistematica das

grammeas. Boletim do Instimto de Botanica No. 5: 1-20. Portuguese translation by T.

Sendulsky, Instituto de Botanica, Sao Paulo, Brazil, 1969.]

RUMPHIUS, E. S. 1750. Het Amboinsch Kruid-boek (Herbarium Amboinense) 6.

Amsterdam.

RUPRECHT, F. J. 1839. Bambuseas Monographice Exponit. Pages 1-75, 18 plates. St.

Petersburg: typis Academiae Caesareae scientiarum.

SODERSTROM, T. R., and R. ELLIS. 1987. The position of bamboo genera and allies in a

system of grass classification. Pp. 225-238 in Grass systematics and evolution, eds.

T. R. Soderstrom, K. W. Hilu, C. S. Campbell, and M. E. Barkworth. Washington,

D. C.: Smithsonian Institution Press.

TZVELEV, N. N. 1989. The system of grasses (Poaceae) and their evolution. The Botanical

Review 55: 141-203.

WATANABE, M., M. ITO, and S. KURITA. 1994. Chloroplast DNA phylogeny of Asian

bamboos (Bambusoideae: Poaceae) and its systematic implication. Joumal of Plant

Research 107: 252-261.

WATSON, L. and M. J. DALLWITZ. 1992. The grass genera of the world. Wallingford,

Oxon, England: CAB International. 14

CHAPTER 2. PHYLOGENY AND CLASSIFICATION OF THE BAMBUSOIDEAE (POACEAE) BASED ON NDHF SEQUENCE AND MORPHOLOGICAL DATA

A paper to be submitted to the journal Systematic Botany

Weiping Zhang and Lynn G. Clark

ABSTRACT

The bamboo subfamily (Bambusoideae) historically has been circumscribed in different ways. Prior study of grass family phylogeny based on ndhF sequence data clearly indicated that the traditional Bambusoideae were highly heterogeneous, and that a monophyletic bambusoid clade consisted of woody bamboos and certain herbaceous bamboos. The primary objectives of this study were to: 1) test the monophyly of the bambusoid clade; and 2) reconstruct phylogeny within this clade using both molecular and morphological data.

Parsimony analysis of ndhF sequences of 38 species from the bambusoid clade and 21 other grasses, with Joinvillea ascendens (Joinvilleaceae) as the outgroup, resulted in: I) confirmation of the basal position of the Anomochloeae, Streptochaeteae, and Phareae (with

Leptaspis sister to Pharus): 2) resolution of f^ielia as the probable basal lineage within the higher grasses; 3) recovery of a well-supported oryzoid clade with Streptogyneae weakly associated at its base; and 4) resolution of a monophyletic bambusoid clade consisting of the woody bamboos (Bambuseae) and the herbaceous olyroid bamboos (Olyreae including

Buergersiochloa and Parianeae), with four subgroups: Olyreae plus Parianeae, temperate

Bambuseae, the New World tropical Bambuseae, and the Old World tropical Bambuseae. A 15

morphological/leaf anatomical data set was generated for a subset of 26 species from the bambusoid clade, two species of Oryzeae, one species of Streptogyneae, and a composite

"basal grass" outgroup. Parsimony analysis of this data set revealed that the Bambuseae and the Olyreae + Parianeae each formed a monophyletic clade, while their relationships with each other and the Oryzeae and Streptogyneae remained unresolved. A combined analysis of the morphological/leaf anatomical data set with a parallel ndhF data set strongly supported a monophyletic bambusoid clade including a monophyletic Bambuseae and a monophyletic

Olyreae + Parianeae, but excluding Streptogyneae. Based on these results, we propose a new circumscription for the Bambusoideae, as well as a classification for the subfamily in which the two tribes Bambuseae and Olyreae are recognized.

INTRODUCTION

Long regarded as enigmatic and taxonomically difficult, the Bambusoideae until recentiy were one of the most poorly understood groups of grasses (Soderstrom and Ellis

1987; Kellogg and Watson 1993), but also one of the most critical in terms of reconstructing the phylogeny of the whole family (Davis and Soreng 1993). Clark et al. (1995), in a study of the grass family based on ndhF gene sequence data, showed that the Bambusoideae as circumscribed in recent global treatments (Clayton and Renvoize 1986; Tzvelev 1989; Watson and DaUwitz 1992) are polyphyletic. The Anomochloeae, Streptochaeteae, and Phareae, three tribes traditionally regarded as herbaceous bambusoid grasses, were resolved as the most basal lineages within the family (Clark et al. 1995). Three other tribes often classified within the

Bambusoideae, the Diarrheneae, Phaenospermateae, and Brachyelytreae, associated strongly with the pooid clade, whereas the Streptogyneae, also usually regarded as herbaceous bambusoid grasses, associated weakly with the oryzoid clade. While each of the oryzoid, pooid, and bambusoid clades was supported as monophyletic, the relationships among them remained unresolved. The bambusoid clade consisted of the Bambuseae (woody bamboos) 16 and the Olyreae (herisaceous bamboos), but four additional putative bambusoid taxa were not sampled: Buergersiochioa Pilg. (Buergersiochloeae or Olyreae); I^elia Franch. (Puelieae or

Bambuseae); Guaduella Franch. (Guaduelleae or Bambuseae); and the Phyllorachideae.

Clark et al. (1995) reviewed the history of classification of the Bambusoideae in detail.

As shown in Table 1 of Clark et al. (1995), as few as one to as many as 15 tribes have been classified within the subfamily based on different criteria. The woody bamboos always were included within the Bambusoideae, but their unique characters, such as woody culms and complicated rhizome systems, were regarded as primitive by some workers (Y. L. Keng 1959;

P. C. Keng 1982, 1987), and as derived by others (Kellogg and Watson 1993). Although

Nees (1835) included the herbaceous Streptochaeta Schrad. within his concept of the bamboos, it was not until this century, beginning with Roshevits (1937, 1946) that herbaceous taxa were classified consistently with the Bambuseae. However, two broadleaved herbaceous Afncan genera, Puelia (Atractocarpa^ and Guaduella have been associated with the Bambuseae since they were described originally as bamboos (Franchet 1887), either as members of the

Bambuseae (Franchet 1887, 1889; Roshevits 1946; Prat 1960; Clayton and Renvotze 1986) or as separate tribes, the Puelieae and Guaduelleae (Soderstrom and Ellis 1987; Watson and

Dallwitz 1992). Buergersiochioa. whether placed in its own tribe, the Buergersiochloeae

(Fijten 1975; Soderstrom and Ellis 1987; Tzvelev 1989), or submerged within the Olyreae

(Prat 1960; Clayton and Renvoize 1986; Watson and Dallwitz 1992), was always associated with the olyroid bamboos, even when the Olyreae were excluded from the Bambusoideae. The position of the Phyllorachideae historically was ambiguous, and the tribe was not included within the Bambusoideae until relatively recently, when a very broad concept of the subfamily was adopted (Clayton and Renvoize 1986; Watson and Dallwitz 1992).

Clark et al. (1995) showed that many of the morphological/anatomical characters presumed to be synapomorphies for the Bambusoideae were actually plesiomorphic within the family. Further study of the bambusoid clade was needed to search for morphological 17 synapomorphies and to reconstruct phylogeny within the clade based on more thorough sampling. We therefore decided to extend the ndhF sequence analysis of Clark et al. (1995), in order to test the monophyly of the bambusoid clade using additional sampling within the

Bambuseae, Olyreae, and the four tribes not included in the prior study. Another objective was to explore the higher level phylogenetic relationships within the bambusoid clade using a combination of molecular and morphological data sets. Although we were unable to obtain satisfactory material of Guaduelleae and Phyllorachideae for sequencing, we here provide evidence for a monophyletic bambusoid clade including Buergersiochloa but excluding Puelia.

Based on our results, we propose a revised circumscription of the Bambusoideae and a tribal level classification for the subfamily.

MATERIALS AND METHODS

Plant Materials. In the most inclusive molecular analysis, a total of 59 species of grasses was sampled, with Joinvillea ascendens fJoinvilleaceae) as the ougroup (Table 1).

Thirty-eight species from the bambusoid clade of Clark et al. (1995) representing the

Bambuseae, Olyreae, and Parianeae (as recognized by Clayton and Renvoize 1986) were analyzed, in addition to Streptogyna americana (Streptogvneae. two accessions of one species), the Oryzeae (three species), the Ehrharteae (one species), the Phareae (three species), the

Anomochloeae (one species), the Streptochaeteae (one species), and exemplars of the Pooideae s. 1. (five species) and the PACC clade (six species). All vouchers cited in Table 1 are deposited in the Ada Hayden Herbarium of Iowa State University (ISC), except those listed as

SBG, which are deposited in the herbarium of the Sichuan Forestry School (SIFS), the collection of Streptogvna crinita which is deposited at the Gray Herbarium (GH), and the collections of Puelia olvriformis and Leptaspis banksii. which are deposited in the United

States National Herbarium (US). An unvouchered sample consisting of leaf pieces identified 18

TABLE 1. Taxa, vouchers, and GenBank accession number for the species sequenced. Classification foilows Clayton and Renvoize (1986). Abbreviations are as follows: AC = A. Carvalho; CCE = Cambridge Congo Expedition 1959; GD = G. Davidse; ISU = Iowa State University; LC = L. Clark; NB = N. Barker, NTBG = National Tropical Botanical Garden; QBG = Quail Botanic Garden; EIP = R. Pohl; SBG = Sichuan Academy of Forestry Botanical Garden; SD = S. Dransfield; WZ = Weiping Zhang; XL = X. Londono.

Taxon Voucher GenBank #

Joinvilleaceae Joinvillea ascendens Gaudich. ex Brongn. & Gris NTBG-8(X)379 U21973

Poaceae Arundinoideae Phragmites australis (Cav.) Trin. LC 1294 U21996 Molinia caemlea (L.) Moench LC 1165 U21994

Bambusoideae Anomochloeae Anomochloa marantoidea Brongn. LC 1299 U2199I

Bambuseae Alvimia gracilis Soderstr. & Londono AC 4389 Ampelocalamus scandens Hsueh & W. D. Li LC 1291 Apoclada simplex McClure & Smith LC 1207 Arthrostvlidium ecuadorense Judz. & L. G. Clark LC 1101 Arundinaria gigantea (Walter) Chapm. WZ 8400703 U21846 Bashania fargesii (E. G. Camus) Keng f. & Yi WZ9201 Bambusa aff. bambos (L.) Voss LC 1300 U22000 Bambusa stenostachva Hack. WZ 8400174 U21967 Cephalostachyum pergracile Munro WZ 8400635 U21968 Chimonobambusa marmorea (Mitford) Makino SBG 9203 U21969 Chusquea circinata Soderstr. & C. Calderon QBG s.n. U21990 Chusquea latifolia L. G. Clark LC&XL417 U21989 19

Table 1 (continued) Glaziophyton mirabile Franch. LC 1066 Guadua paniculata Munro QBG s.n. U22001 Fargesia robusta Yi WZ92II Hickelia madagascariensis A. Camus SD 1290 Melocanna baccifera Kurz. XL & LC 930 N"astus elatus Holttum SD s.n. Neurolepis aperta rMunro) Pilger XL&LC913 Otatea acuminata (Munro) C. Calderon & Soderstr. LC «&; WZ 1348 Phvllostachvs bambusoides Sieb. & Zucc. Phyllostachvs pubescens Mazel ex J. Houz. LC 1289 U21970 Pseudosasa japonica (Sieb. & Zucc. ex Steud.) WZ 8400708 Makino ex Nakai Puelia olvrifomiis (Franch.) Clayton & Renvoize CCE288 Racemobambos microphvUa (Hsueh & Yi) Wen WZ 8400604 Rhipidocladum pittieri (Hack.) McCIure LC & WZ 1349 Sasa variegata E. G. Camus WZ 9228 Schizostachvum luzonicum Gamble SD 1323 Shibataea kumasaca (ZoU.) Makino LC 1290 Yushania exilis Yi WZ 9230

Brachyelytreae Brachvelvtrum erectum CSchreb.) P. Beau v. LC 1330 U22004

Diarrheneae Diairhena obovata (Gleason) Brandenberg LC&WZ1216 U21998

Ehrharteae Ehrharta calvcina Sm. NB s.n. U21995

Olyreae Buergersiochloa bambusoides Pilg. SD 1382 Lithachne humilis Soderstr. LC 1298 U21977 Lithachne pauciflora (Sw.) P. Beauv. LC 1297 U21978 20

Table I (continued) Olvra latifoUni XL&LC911 U21971 Raddia distichophvUa (Schrad. ex Nees) Chase LC 1306 U22006 Sucrea maculata Soderstr. LC & WZ 1345

Oryzeae Leersia virginica Willd. LC1316 U21974 Maltebrunia petiolata Aug. SD 1336 Oryza sativa L. Sugiura(1989) XI5901

Parianeae Eremitis sp. nov. LC & WZ 1343 Pariana radiciflora Sagot ex Doell LC & WZ 1344

Phaenospermatae Phaenosperma globosum Munro LC 1292 U22005

Phareae Leptaspis banksii R. Br. Darbyshire 1170 Pharus lappulaceus Aubl. LC 1329 U21993 Pharus latifolius L. LC 1302 U21992

Streptochaeteae Streptochaeta angustifolia Soderstr. LC 1304 U21982

Streptogyneae Streptogvna americana C. E. Hubb. 1 RP&GD 12310 U21965 Streptogvna americana C. E. Hubb. 2 Johnston 433 Streptogvna crinita P. Beauv. Dinklage 3375

Chloridoideae Sporobolus indicus (L.) R. Br. LC 1293 U21983 Zovsia matrella (L.) Merr. LC 1174 U21975 21

Table 1 (continued) Panicoideae Sorghum bicolor (L.) Moench R. Wise, ISU U2198I Zea mays L. cv. 'B73' M. Lee, ISU U21985

Pooideae Avena sativa L. cv. 'Ogle' R. Wise, ISU U21999 Poa pratensis L. LC 1156 U21980 as Humbertochloa greenwavi. representing the Phyllorachideae, was sequenced and included in a preliminary analysis. The Bambuseae include an estimated 60-70 genera and ca. 1000 species; the Olyreae plus Parianeae include an estimated 20 genera and ca. 100 species (Clark,

1995) and the Buergersiochloeae are monotypic. Therefore we sampled approximately 40% of the genera and 3% of the species of woody bamboos, and 30% of the genera and 7% of the species of the Olyreae/Parianeae.

A subset of 30 species from the complete ndhF data set, including a composite "basal grass" as the outgroup, was analyzed based on morphological/leaf anatomical characters. A parallel matrix of ndhF sequences for these same taxa also was constructed, with Streptochaeta angustifoha designated as the outgroup. This subset included 26 species of the bambusoid clade of Clark et al. (1995), with representatives of the Bambuseae (19 species) excluding

Puelia olvriformis. Olyreae (five species), and Parianeae (two species), as well as exemplars of the Oryzeae (two species) and the Streptogyneae (one species).

DNA Sequencing. Total cellular DNA was extracted from leaves using methods detailed in Paterson et al. (1993). We used silica gel dried leaves for most species (Chase and

Hills 1991), but fresh or frozen leaves, as well as leaves from herbarium specimens, were used for DNA isolation. The DNA extracted from the mostly old herbarium specimens was cleaned one or several times using ammonium acetate precipitation. 22

The polymerase chain reaction (PGR) was used to amplify the ndhF gene for all taxa listed in Table I using the primers described in Olmstead and Sweere (1994) and Ohnstead and

Reeves (1995). To amplify the whole gene, two sets of primer combinations, I'F/B 18'R and

972^/21 lO'R, were used for most of the double-stranded DNA amplifications. For a few species, especially those for which the DNA was extracted from very old herbarium specimens, the gene was amplified using three or four sets of primer combinations: I'F/SOSR,

536'F/1318'R, 972'F/2110'R, or 972'F/1603'R, and 1318'F/21 lO'R. For double-stranded amplification, each 50 ul reaction contained 100 ng of template DNA, 1.5 mM MgCb, 200 uM of each dNTP, 0.2 pM of each primer, 5.0 ul Mg-free Tag buffer, and 0.5 unit of Taq polymerase. For single-stranded amplification, each reaction contained 100 ul of same concentration of each component as used in the double-stranded amplification, except for 5 ul double-stranded amplification product instead of the genomic DNA. The Hot Start program consisting of 2 minutes at 95°C and 5 minutes at 80°C, was used for both double-stranded and single-stranded PGR in order to get more specific amplifications. Following the Hot Start, a program of 1 minute and 30 seconds at 94°C, 2 minutes at 42°C, 3 minutes at 72°C, 30 cycles, and 10 minutes at 72°C as extension was designed to complete the double-stranded amplification.

To generate single-stranded DNA, we conducted two-stage PGR amplifications

(Kaltenboeck et al. 1992), using double-stranded PGR amplification products as the template.

Following the Hot Start, the same PGR program as the double-stranded amplification, except only 20 cycles, was used for single-stranded amplification. Single-stranded PGR products were purified using microcon 100 ultrafilters (Amicon, Beverly, Massachusetts) prior to the sequencing reaction. Standard metiiods of dideoxy sequencing with Sequenase 2.0 (United

States Biochemical, Gleveland, Ohio) and [a-^^S]dATP were used for the sequencing reactions. Electrophoresis was conducted using Long Ranger (FMG Bio-Products, Rockland, 23

Maine) polyacryiamide gels, whicli were vacuum dried prior to exposure to X-ray autoradiography.

Sequences were aligned manually with the rice fOryza sativa) sequence (Sugiura 1989) as a reference. The ndhF gene in rice is 2,205 bp in length, occupying coordinates 101,433

(3' end in our sequence order) to 103,637 (5' end in our sequence). The regions corresponding to the terminal amplification primers I'F and 21 lO'R (103,6111 to 103,637 and 101,506 to 101,525 respectively) were not included in the analysis, but all other internal primer region sequences were verified and retained in the final analysis. Therefore, sequences corresponding to rice positions 101,526 to 103,610 were analyzed, except that one small internal region, from 102,820 to 102,837, was omitted due to gel compression in some taxa.

Morphological and leaf anatomical characters.

Sixty-four morphological and leaf anatomical characters with two or more character states each (Table 2) were generated for a subset of 30 species from the fiill ndhF analysis. All of the morphological and anatomical information was extracted from the individual species included in our analyses, except for Sasa variegata. Fargesia robusta. Yushania exilis. and

Racemobambos microphvUa. for which no inflorescence information was available. The generic inflorescence descriptions were adopted for these four taxa with the assumption that they are correctly classified. Morphological character information was collected by direct observation of available herbarium specimens and living plants, as well as reference to the literature.

Because of the lack of consistent leaf anatomical data for many of the taxa, we used light microscopy to examine leaf cross sections and epidermal scrapes for the majority of the species; observations were supplemented by scanning electron micrographs for some taxa and reference to the literature. Leaf material from the same vouchers as those for the molecular analyses were used whenever possible. Fresh or dried leaves (from silica-gel dried material or 24

TABLE 2. Morphological and leaf anatomical characters and character states used in the analysis. Numbers following the character and character state are the consistency index and retention index respectively, corresponding to the analysis in Fig. 5.

1. Habit: herbaceous (0); woody (1). (1.000,1.000) 2. Lifespan: perennial(0),annual(1). (1.000,0/0) 3. Strong, well developed woody rhizome system: absent (0), present (1). (1.000, l.OOO) 4. Rhizomes pachymorph without elongated neck: present (0), absent (1). (0.167,0.545) 5. Rhizomes pachymorph with elongated neck: absent (0), present (1). (0.250,0.250) 6. Elhizomes amphimorph: absent (0), present (1). (0.333,0.333) 7. Rhizomes leptomorph: absent (0), present (1). (l.(X)0, 1.000) 8. Culms: erect (0), scandent or climbing (1). (0.200,0.429) 9. Culm branching: unbranching (0), branching (1). (1.000,1.000) 10. Primary buds per mid-culm node: none (0), one (1), two or more (2). (l.(K)0, 1.000) 11. Culm intemodes: solid (0), hollow (1). (1.000,1.000) 12. Leaves: mostly basal (0), cauline, not basally aggregated (1). (0.500,0.667) 13. Leaves: no differentiation of foliage leaves and culm leaves (0), foliage leaves and cukn leaf differentiated (1). (1.000,1.000) 14. Foliage leaf blade: not disarticulating from the sheath (0), disarticulating from the sheath (1). (0.500,0.500) 15. Foliage leaf blade pseudopetiole: present (0), absent (1). (l.OOO, l.OOO) 16. Outer (abaxial) ligule of foliage leaf: absent (0), present (1). (0.500,0.899) 17. Foliage leaf oral setae: absent (0), present (1). (0.200,0.667) 18. Foliage leaf lateral appendages: absent (0), present (1). (0.333,0.000) 19. Transverse veins in leaf blade: visible to the unaided eye (0), not clearly visible to the unaided eye (1). (0.250,0.667) 20. Branching season: no vegetative branching (0), branching in the same year as shooting (1), branching in the following year (2). (0.667,0.938) 21. Flowering: annual or frequent (0), gregarious at long intervals (1). (l.(X)0, l.OOO) 22. Inflorencences: with fully bisexual spikelets only (0), with functionally unisexual spikelets with or without staminodes or pistillodes(l). (1.000, 1.000) 23. Staminodes or pistillodes visible to the unaided eye in spikelets: absent (0), present (1). (l.OOO, 0/0) 25

Table 2 (continued) 24. Gemmiparous bracts subtending the spikelet proper: absent (0), present, with buds developing subsequently or not (1). (0.333,0.778) 25. Subtending bracts at the base of first or second order paracladia: absent (0), present (1). (0.333, 0.846) 26. Prophylls at the base of first or second order paracladia: absent (0), present (1). (0.333, 0.800) 27. Rachilla extension: absent (0), present, with or without a rudimentary floret (1). (0.500, 0.909) 28. Pedicels: slender relative to the male (or bisexual) spikelets, not broad and strongly flattened (0), broad relative to the male spikelets, strongly flattened and indurate, often connate with pedicels of similar male spikelets (1). (1.000, LOCK)) 29. Number of fertile florets per spikelet: one (0), two or more (1). (0.333,0.846) 30. Glumes: present (0), absent (1). (1.000,1.000) 31. Number of lodicules per female fertile floret: three (0), two (1). (1.000, 1.000) 32. Leimna apex: pointed (0), awned (1), blunt (2). (0.500, 0.714) 33. Mature lemma of female fertile spikelets: remaining pliable (0), becoming indurated (1). (1.000, 1.000) 34. Strongly keeled lemma: absent (0), present (1). (1.000,1.000) 35. Palea apex: bifid (0), entire (1). (0.500,0.875) 36. Abaxial surface of palea: sulcate (2-keeled) (0), rounded (1), l-keeled(2). (0.667, 0.667) 37. Number of stamens: six (0), more than six (I), three or fewer (2). (0.400,0.667) 38. Stamen filaments: free (0), fused (1). (0.500,0.500) 39. Number of stigmas: three (0), two (1). (0.333,0.833) 40. Style: glabrous (0), pubescent (1). (0.333,0.667) 41. Fruits: with thin and hard pericarp (0), with thick and fleshy pericarp (1). (0.500, 0.000) 42. Number of bladeless sheaths on the seedling: one or more (0), none (1). (0.500,0.000) 43. Lamina orientation of the first seedling leaf: horizontal (0), vertical (1). (0.500,0.500) 44. Vasculature of midrib: complex (0), simple (1). (0.143,0.333) 45. Midrib of the foliage leaf blade: projecting both abaxially and adaxially (0), projecting abaxially only (1), projecting adaxially only (2), not projecting (3). (0.250,0.250) 46. Air spaces complexed with vasculature of midrib: present (0), absent (1). (1.000, 1.000) 26

Table 2 (continued) 47. Microhairs on foliage leaf blades: present (0), absent (1). (1.000,0/0) 48. Papillae on the long cells in the stomatal zone (abaxial or adaxial): absent (0), present (1). (0.500, 0.500) 49. Papillae on the long cells in the interstomatal zone: absent (0), present (I). (0.200, 0.429) 50. Papillae in the mtercostal zone (excl. buUifonn cells): absent on both surfaces (0), present on abaxial surface only (1), present on both surfaces (2). (0.500,0.500) 51. Papillae on the subsidiary cells of the stomatal apparatus: absent (0), present (I). (0.500, 0.667) 52. Intercostal sclerenchyma in mesophyil: present (0), absent (1). (0.500,0.667) 53. Vertically tall and narrow intercostal silica bodies: present (0), absent (1). (0.333.0.750) 54. Horizontal dumbbell intercostal silica bodies: absent (0), present (1). (1.000,0/0) 55. Saddleshaped intercostal silica bodies: present (0), absent (1). (0.333,0.818) 56. Crenate (olyroid) intercostal silica bodies: absent (0), present (1). (1.000, 1.000) 57. Vertical tall and narrow costal silica bodies: absent (0), present (1). (0.200,0.500) 58. Saddleshaped costal silica bodies: present (0), absent (1). (0.333,0.500) 59. Vertical dumbbell costal silica bodies: absent (0), present (1). (1.000, 1.000) 60. Horizontal dumbbell costal siUca bodies: absent (0), present (1). (0.200,0.333) 61. Cross-shaped costal silica bodies: absent (0), present (1). (1.000, 1.000) 62. Distribution of stomata on foliage leaf blades: present on both surfaces (0), present on the abaxial surface only (1). (0.167,0.167) 63. Arm cells: poorly developed (0), well developed (1). (0.250,0.400) 64. Fusoid cells: present (0), absent (1). (0.500,0.500) herbarium specimens) were soaked in Pohl's solution (Pohl 1965) for approximately 30 min.

For cross sections, the central portion of the blade was sectioned widi a single-edged razor blade, and mounted in lactophenol/aniline blue (Sass 1958). For epidermal scrapes, the central portion of the blade was scraped gendy to remove either the abaxial or adaxial surface and the mesophyil, and the remaining epidermis was mounted in lactophenol/anihne blue.

Scaiming electron microscopy also was used to study leaf epidermal characters for some of the bambusoid taxa. Representive samples approximately 0.5 X l.O cm were excised from the middle third of the mature foliage leaves, and were either mounted directly or 27 sonicated in xylene for about 10 min to remove epicuticular wax. Specimens were mounted on brass discs with silver paste or double-stick tape, coated with Au-Pd in a Polaron E5100 sputter-coater, then viewed at 15 kV in a JEOL JSM-35 scanning electron microscope.

Photographs were taken using Polaroid Type 665 positive-negative film.

OUTGROUP SELECTION. Recent molecular studies (Clark et al. 1995; Soreng and

Davis 1995; Duvall and Morton 1996; Liang and Hilu 1996) have indicated unequivocally that the anomalous, tropical forest tribes Anomochloeae, Streptochaeteae, and Phareae comprise the basal lineages in the grass family. Although there is strong support for the clade referred to as the "higher grasses" (Clark et al. 1995; Soreng and Davis 1995; Duvall and Morton 1996;

Liang and Hilu 1996), which includes all other grass lineages above die basal ones, relationships of major lineages within the higher grasses remain to be resolved robustiy. The rices and true bamboos often appear near the base of the higher grasses with the Pooideae sister to the PACC clade (Barker et al. 1995; Soreng and Davis 1995; Duvall and Morton 1996;

Liang and Hilu 1996), but Clark et al. (1995) recovered a single, relatively weakly supported clade including the Bambusoideae, Oryzoideae, and Pooideae (the 'BOP' clade) sister to the

PACC clade, which is well supported in ail analyses (Hilu and Johnson 1990; Davis and

Soreng 1993; Barker et al. 1995; Clark et al. 1995; Soreng and Davis 1995; Duvall and Morton

1996; Liang and Hilu 1996).

Based on the ndhF data, the lack of resolution among the Bambusoideae, Pooideae, and Oryzoideae made it difficult to select confidentiy either the pooids or the oryzoids as an outgroup for an exploration of phyiogenetic relationships within the true bamboos, and the

PACC clade was considered to be too highly modified with respect to the bamboos to be appropriate as an outgroup. The basal lineages were therefore the most conservative outgroup choice. Because of the lack of certain characters in certain basal grasses (e.g., the absence of lodicules in Anomochloeae and Streptochaeteae), or ambiguous interpretations of some characters (e.g., spikelet bracts in Anomochloeae and Streptochaeteae), we created a composite 28

'basal grass' outgroup in which were combined the character states of the Anomochloeae,

Streptochaeteae, and Phareae that reasonably could be considered homologous to those of the ingroup. 'Basal grass' therefore was used as the outgroup in the morphological analysis, while Streptochaeta angustifolia was designated as the outgroup in the parallel ndhF sequence analysis.

CHARACTERS. Most of the morphological and anatomical characters included in our analysis were used previously in various studies (Wu 1958; Metcalfe 1960; Ellis 1976;

Renvoize 1985; Clayton and Renvoize 1986; Soderstrom and Judziewicz 1987; Soderstrom et al. 1987; Soderstrom and Ellis 1987; Kellogg and Campbell 1987; Judziewicz and Soderstrom

1989; Kellogg and Watson 1993; Guala 1995), but some were used the first time in this study, or with different interpretations from previous studies. Certain characters, e.g., silica body shapes or number of florets per spikelet, displayed continuous variation, and thus divisions between character states were sometimes arbitrary. Other characters, such as the presence of stomata on one surface or the other, were coded only functionally. If so few stomata appeared on a given surface that they could not be contributing significandy to gas exchange, they were coded as absent. Both binary and multiple states were used to code the characters as shown in the data matrix (Table 3). Consistency and retention indices of each character were included in

Table 2 with reference to the cladogram in Figure 5 to evaluate these morphological and leaf anatomical characters.

Habit, Life Span, and Flowering (Characters 1-2, 21): These species are annual or perennial herbs or perennials with lignified culms (woody). All taxa with rhizomes were considered perennial. McClure (1973) did not indicate clearly diat Neurolepis exhibited woody culms, and the genus was described as herbaceous by Clayton and Renvoize (1986). Through field observation, however, it has been confirmed that at least the lower portions of the culms in most species of Neurolepis. including N. aperta. are woody (Clark, pers. obs.), and thus it is scored as woody in this analysis. TABLE 3. Data matrix of morphological and leaf anatomical characters and character states used for cladistic analysis in the bambusold clade. Characters were coded as indicated in Table 2. * indicates that character 39 of Ariindinaria gigantea has two states; '0' and T, and '?' means missing data.

10 20 30 40 50 60 64 Bueraersiochloa bambusoides 0000000000 7101000010 0110000000 0100102110 0001200101 0101001001 0100 Pariana radiciflora 0001100000 0101001010 0100000100 0200111110 0000000101 0010110001 1110 Sucrea maculata 0000000000 0101000010 0100000000 0210112010 0000200111 0010110000 1110 Olvra latifolia 0000000000 0101000010 0100000000 0210102010 0000000111 0010110001 1110 Lithachne oauciflora 0000000000 0101000010 0100000000 0210102010 0001000111 0010110100 1110 Raddia distichoohvlla 0000000000 0101000010 0100000000 0210102010 0001000111 0010110101 1110 Eremitis sp. nov. 0001100000 0101001010 0100000100 0200112110 0000100000 0010110101 1010 Arundinaria aiaantea 1011010011 1111011101 1000101010 00000020*0 0001100111 0000100000 0110 ftmpelocalamus scandens 1010000111 1111011101 1001111010 0000002000 0000100111 0000000000 0110 Shibataea kumasaca 1011001011 1111011001 1001111010 0000002000 0000100111 0000100000 0110 Chimonobambusa marmorea 1011010011 1111011001 1001111010 0000002010 0000100111 0000100000 0110 Phvllostachvs nubescens 1011001011 1111011101 1001111010 0000002000 0000100111 0000100000 0110 Sasa varieaata 1011010011 1111011001 1000101010 0000000000 0000100111 0000100000 0110 Yushania exilis 1011100011 1111011001 1000101010 0000002010 0000100111 0000101000 0110 Alvimia qracilis 1010000111 1111011011 1001111010 0000002010 1001100111 0100001000 0110 Otatea acuminata 1010000011 1111011011 1000001010 0000002010 0000100111 0000000001 0010 Neuroleois aoerta 1010000001 0111010000 1000000000 0000002010 0000100111 1000001000 0100 Rhioidocladum oittieri 1010000111 1111011011 1000001010 0000002010 0001100111 0100001000 0010 Chusauea latifolia 1011010112 0111010001 1000000000 0000002010 0001100101 1000001000 0110 Guadua oaniculata 1010000011 1111010011 1001111010 0000000001 0001100111 0000001000 0011 Bambusa aff. bambos 1010000011 1111011012 1001111010 0000000001 0000100101 0000000001 0110 MelQcanns baccifera 1011100011 1111011012 1001111010 0000000000 1000100111 0000001000 0110 Ceohalostachvum oeraracile 1010000111 1111011012 1001111010 0000000001 0001000111 0000001000 0110 Racemobambos microohvlla 1010000111 1111011012 1001111010 0000000001 0000100111 0000000000 0110 Hickelia madaqascariensis 1010000111 1111010012 1000111000 0000000001 0000100111 0000000000 0110 Table 3 (continued) Nastus elatus 1010000111 1111010012 1000001000 0000000001 0000200101 0000000000 0100 Orvza sativa 0101000000 0000100010 0000000001 1101120010 0110010111 1010000110 0011 Leersia virainica 0001100000 0100100010 0000000001 100112701? 0011110111 1010000110 0001 Streotoavna americana 0000000000 0001011010 0000001010 0100002001 0110101000 0010000000 0100 Basal grass 0000000000 0000000000 0000000000 0000000000 0000000000 0000000000 0000

U) o 31

Most grasses bloom within the year of growth following germination, but even in perennial grasses, once reproductive maturity is reached flowering normally occurs on an annual basis. Olyroid bamboos generally lack a well defined blooming season, and have been described as flowering frequently (Calderon and Soderstrom 1980). At least one species may grow vegetatively for long periods before blooming (O. standlevi. Pohl 1991; Soderstrom and

Zuloaga 1989), but annual or frequent flowering is found among the species included in this analysis. Gregarious blooming at intervals of greater than two years, more usually greater than

10 years, is characteristic of woody bamboos (Keng 1982; Calderon and Soderstrom 1980;

Pohl 1991; Clark 1995). This character is scored without reference to monocarpy, which typically occurs in woody bamboos, but of course is also found in annual species.

Rhizomes (Characters 3-7): The rhizome system, die rhizome proper, and the rhizome neck were described in detail by McClure (1966). The rhizome neck lacks the prophyllate buds and root primordia present on the rhizome proper. Two main types of rhizome morphology are recognized within bamboos: pachymorph and leptomorph (McClure 1966). Pachymorph rhizomes, short and thick with solid intemodes that are wider than long, are correlated with sympodial branching and clumped culms. Leptomorph rhizomes are long and slender with typically hollow intemodes that are longer than wide, and are correlated with monopodia! branching and relatively widely spaced culms. Within the pachymorph type, some species exhibit elongated necks up to several m in length (McClure 1966; Soderstrom 198 la; Keng

1982; Soderstrom and Judziewicz 1987), resulting in widely spaced culms analogous to species with leptomorph rhizomes. Some bamboos consistently exhibit both pachymorph and leptomorph rhizomes within the rhizome systems of individual plants, and are described as amphimorph. Because there is controversy over the interpretation of homology among these growth patterns (Wang and Ye 1980; Chao et al. 1980; Stapleton, in press; Ding and Liese, in press), particularly with respect to the amphimorph condition, we have scored each of these four conditions separately based on the descriptive differences cited above. 32

Culms, Buds, and Branching (Characters 8-11, 20). Most of the species have erect culms, except several tropical woody bamboos, which have culms that are scandent on other vegetation or climb into Urees. We scored culm branching with respect to the production of additional vegetative, not reproductive, axes. Many of the herbaceous bamboos (Olyreae) produce axillary inflorescences (Soderstrom and Zuloaga 1989; Davidse and Pohl 1992), but it is rare to observe vegetative branching. Olvra latifolia does occasionally branch vegetatively, but this is not a consistent feature and thus was scored as absent. Among the species in this study, only the woody bamboos consistentiy possessed a bud or buds at each node and exhibited vegetative branching of the aerial culms. Characters 9, 10, and 20 are not independent, but character 10 includes the additional information that multiple buds are present in Chusquea. The culms of Neurolepis do not branch, but a single bud per node is often observed, especially at lower nodes (Clark, pers. obs.). Within the woody bamboos, the timing of vegetative branching in relation to shoot production can be informative, and thus this was scored as a separate character. More detailed information on vegetative branching patterns exists for some woody bamboos (McClure 1966; Clark 1989; Keng 1982, 1987), but is lacking for many taxa. Rigorous analysis of these patterns undoubtedly will be informative in assessing phylogenetic relationships within the Bambuseae, but was beyond the scope of the study.

Intemodes were scored as hollow if a well defined lacuna was present, regardless of wall thickness; solid culms can become irregularly fistulose with age, as in many species of

Chusquea (Clark 1989), but these were scored as solid. No information was available for

Buergersiochloa for this character.

Leaves (Characters 12-19). Leaves were scored as cauline if at least some leaves are distributed along the culm during vegetative growth. In woody bamboos, two kinds of leaves are differentiated (McClure 1966): fohage leaves which are photosynthetic, and culm leaves which protect the fragile intercalary meristem at the base of each intemode as the new shoots 33 elongate. Culm leaves, consisting of a sheath, blade, and inner ligule, are homologous to foliage leaves but exhibit a shift to a protective function, even if they make some contribution to photosynthesis (some culm leaves remain green for an extended period). Reproductive shoots in some olyroid bamboos may be highly modified in comparison to vegetative shoots

(Calderon and Soderstrom 1980), but dimorphic vegetative shoots are not known.

The outer ligule is a small flap of tissue present as an abaxial extension of the sheath summit (Clark and Pohl 1996), and is nearly always thicker and tougher than the inner ligule.

The pseudopetiole is a constriction at the base of the foliage leaf blade (McClure 1966) and may be only weakly developed as in Streptogyna (Soderstrom and Judziewicz 1987) or extremely weU developed and up to 50 cm long as in some species of Neurolepis (Davidse and Clark

1996). Oral setae, also known as fimbriae, may be present at the sheath summit or on lateral appendages (auricles) extending from the sheath summit (McClure 1966). Although grasses have small commissural veins connecting the major parallel veins of the leaf blade, these are usually not visible on the surface of the blade. But in many woody bamboos and some other grasses, these veins are visible and form a pattern described as "tessellate."

Inflorescences, Spikelets, Flowers, and Fruits (Characters 22-41). Inflorescence morphology in bamboos, and indeed all grasses, is complex. The traditional terminology and concepts applied to grass inflorescences are inadequate and even inaccurate, but an integrated perspective based on rigorous assessment of homology across grass lineages is only just begirming to emerge. In light of these problems, and the lack of data for many taxa of bamboos, we again chose to adopt a purely descriptive approach to inflorescence characters.

The characters presented here represent only a first approximation.

Among the taxa in this study are found many species with strictly bisexual florets in all spikelets of the inflorescence, but some taxa possess unisexual florets, although all of these are monoecious. With Character 23 we distinguish between functionally unisexual florets with obvious staminodes or pistillodes, as in Buergersiochloa (Pilger 1915; Fijten 1975) and clearly 34 unisexual florets without such obvious vestigial structures. The olyroid bamboos are scored as unisexual and lacking obvious staminodes or pistillodes, but some species, such as

Diandrolvra bicolor (Stapf 1906), Olyra cordifolia (Butzin 1965), and several other species of

Olvra. Crvptochloa. and Maclurolvra (Calderon and Soderstrom, 1973), may have highly reduced staminodes or pistillodes visible upon microscopic examination.

Woody bamboos especially are known for their bracteate, often indeterminate, difficult to interpret inflorescences (McClure 1966; Wang and Ye 1980,1988; Soderstrom 1981b; Keng

1986; Stapleton, in press). In many species the inflorescence is relatively easy to delimit, but in others, the whole plant arguably could be regarded as the inflorescence (Stapleton, in press).

For this analysis, we tried to compare equivalent structures (e.g., primary and secondary paracladia) as much as possible in order to assure homology, but for some taxa alternative interpretations may exist. We chose three characters often used in circumscribing genera, subtribes or tribes within the woody bamboos: the presence of gemmiparous bracts subtending the spikelet proper, the presence of subtending bracts at the base of the first or second order paracladia, and the presence of prophylls at the base of the first or second order paracladia.

The presence of genraiiparous bracts and prophylls is typically used to define the concept of pseudospikelet (McClure 1934; Holttum 1956; Keng 1982, 1986; Clark and Pohl 1996), but we have attempted to dissect this poorly understood strucmre into its components, as diere are suggestions that they may not always co-occur. Gemmiparous bracts are those bracts found below the spikelet proper (which may or may not have glumes) which each bear a single axillary bud that may or may not develop into another spikelet; when present, there are typically only 1-3 of these bracts. If no buds were present, these bracts would be regarded as glumes.

Subtending bracts are defined here as bracts that subtend the first or second (or higher order) paracladia within the infloresence. These could be regarded as gemmiparous bracts that now subtend branches rather than buds. A prophyll (two- or sometimes one-keeled but adaxial in 35

position) present at the base of a paracladium would also indicate the prior presence of a bud

now developed into a branch.

The presence of a rachilla extension with or without a rudimentary floret was scored as

a single character state (Character 27). The unique flattened pedicels of the male spikelets in

Pariana and Eremiris are described and illustrated in Judziewicz (1990) and Hollowell (in

press). The number of glumes and the number of fertile florets per spikelet are characters that

both exhibit considerable variation within the study group and thus their scoring was

simplified; we recognize that this may not be ideal but we did not want to overemphasize diese

characters by scoring each possible number as a separate character. The use of multistate

scoring was not feasible given the polymorphism that resulted.

Seedlings (Characters 42-43). Soderstrom (1981b) showed that the first one or few seedling leaves in many bambusoid taxa consist only of a sheath or present only a rudimentary

blade (lamina), whereas subsequent leaves have well developed blades. The orientation of the

lamina in the first leaf (or the first leaf with a well developed lamina) is either horizontal or

vertical.

Transverse Leaf Anatomy (Characters 44-46,52,63-64). Midrib vasculature was defined as simple if only one central vascular bundle was present, or as complex if one or more smaller vascular bundles were associated with the central one. Intercostal sclerenchyma was scored as present if it occurred only internal to one surface (usually the abaxial epidermis) but in Rhipidocladum pittieri it was present both abaxially and adaxially. Arm cells were scored as well developed if the invaginations were at least one-third of the height of the cell.

Epidermal Micromorphology (Characters 47-51,53-62). The presence or absence of papillae, and distribution of papillae in different parts of the costal zone, were scored as three characters based on the variation among taxa in our sample set. The stomatal zone is that part of the intercostal zone where stomata occur; the long cells tend to be shorter and wider. In relation to the mesophyll, this zone occurs direcdy beneath any given file of fusoid cells. The 36 interstomatal zone is devoid of stomata and occurs between the two stomatal zones of a given intercostal zone. This zone coincides with the area of mesophyll between files of fiisoid cells.

These features are illustrated and described in Palmer and Tucker (1981) and Clark (1990).

Variation in silica body shape is often continuous, and the divisions into character states that we recognized are illustrated in Figure 1. The presence or absence of each type of silica body shape we defined was scored as a separate character, and a distinction was made between intercostal and costal silica bodies. If stomata were present but very rare on a surface, they were scored as absent.

Phylogenetic Analvses. For all analyses, maximum parsimony was applied using

PAUP version 3.11 (Swofford 1993); because of the large size of the data sets, only the heuristic search option was used. To discover other possible islands of trees, stepwise addition and branch swapping options were employed. All characters were weighted equally in all analyses. Bremer support (Bremer 1988) (decay analysis) was evaluated for individual clades with modifications described for the different data sets. MacCIade version 3.04 was used to compare user-defined alternative topologies with the shortest trees gained from the parsimony analysis (Maddison and Maddison 1993).

ndhF SEQUENCE ANALYSES. Phylogenetic analyses for both the complete and reduced ndhF data sets were conducted on data matrices representing aligned sequences

(Appendix I). All codon positions and transformation types (transversion and transitions) were equally weighted. As to the indels, three different treatments were used in the analysis: 1) deletion of all indels; 2) retention of all indels; and 3) deletion of indels, treating each individual indel as one event, coding them as binary characters (present or absent), and addition of these characters at the end of the data matrix. Additionally, an analysis using only the 3' half of the gene was performed in order to resolve the position of Streptogyna crinita for which only that portion of the gene was sequenced. An unvouchered sequence of Humbertochloa greenwavi FIG. 1. Silica body diagram. Five different silica body shapes were described for the intercostal and costal zones: A. tall and narrow; B. horizontally oriented dumbbell (parallel to the long axis of the leaf blade); C. intercostal saddleshaped; D. crenate; E. cross-shaped; F. costal saddleshaped; and G. vertically oriented dumbbell (perpendicular to the long axis of the leaf blade). All drawings to the same scale. 38

O co°

OO LU 39 was included in a preliminary analysis, but excluded from the final analyses because of uncertainty regarding its identification.

To evaluate Bremer support for individual clades, strict consensus trees were generated at tree lengths up to four steps longer. Because of the size of the complete data set and insufficient computer memory, we could not get the strict consensus trees even up to one step longer if all terminal taxa were included. Accordingly, prior to running the decay analyses, some selected terminal taxa were deleted from the unresolved polychotomous branches based on the strict consensus tree generated from the most parsimonious trees, with retention of the same overall topology. Sequence variation was calculated for the complete data set as well as selected clades based on the total length of 2115 bp; genetic distance was calculated between selected pairs of taxa representing major clades. For distance-based phylogeny estimation, observed distances between all pairs of sequences were translated to evolutionarily "corrected"

Kimura two-parameter distances, and a neighbor-joining analysis was performed on the resulting distance matrix using MEGA (Kumar et al. 1993).

In the analysis of the reduced data set, exacdy the same terminal taxa as in the morphological and leaf anatomical analysis were included, in order to compare the molecular and morphological results, and for the combined data sets analysis. Several taxa from among the basal grass and basal pooid lineages as identified by Clark et al. (1995) were tested as outgroups in order to examine the effect of outgroup choice and potential problems due to long branch attraction. Decay analysis up to four steps longer without deleting any terminal taxa was completed to evaluate the individual clades.

MORPHOLOGICAL ANALYSIS. A strict consensus tree was generated from the most parsimonous trees with decay indices to indicate the strength of individual clades, and supporting characters were mapped on this tree for the major clades. Decay analysis up to three steps longer was also used to evaluate the support of individual clades. Prior to running 40 the decay analysis, eight terminal polychotomous taxa were removed because of the lack of computer memory, but the overall same topology was retained.

COMBINED DATA SETS ANALYSIS. Streptochaeta angustifolia for the molecular data and a synthzed 'basal grass' for the morphological/anatomical data were combined as the outgroup for this analysis. Separate analyses of the reduced ndhF and morphological/anatomical data sets were conducted first with subsequent combination of the respective consensus trees. Another analysis was performed in which the two data sets were combined initially and then subjected to parsimony analysis. Decay analysis up to four steps longer without removing any terminal taxa was conducted to estimate the support of each of the major clades.

RESULTS

Complete Molecular Data Set Analvsis. After removing regions corresponding to the beginning and the end of the gene, and the poorly resolved internal regions, sequences from

2064 bp to 2088 bp hi length were analyzed for 60 species including the outgroup. These sequences were aligned manually by introducing five indels (Table 4), resulting in a final length of 2115 bp. Indels b, d, and e were uninformative in the phylogenetic analysis. Two insertions, one of six bp shared by Raddia and Sucrea (c in Table 4), and the other of 15 bp shared by all the 'higher grasses' except Oryzeae (a in Table 4) were phylogenetically informative.

All three treatments of indels resulted in trees with the same topologies witii only slight differences in tree length, and consistency and retention indices. Usmg presence/absence codmg for insertions and deletions, the heuristic option yielded 736 equally most parsimonious trees of 1350 steps with consistency and retention indices of 0.476 (without autapomorphies) and 0.701 respectively; one of these is shown in Figure 2. The number of the trees was mainly caused by the lack of resolution within two particular clades corresponding to the temperate 41

TABLE 4: Indels (insertions and deletions) in ndhF gene sequences for Bambusoideae. Letters denote indels mapped in Fig. 2. Indel location is after the nucleotide listed, using as a reference coordinates published for ndhF from Orvza sativa (Sugiura 1989). Indels of non- Bambusoideae sensu Clayton & Renvoize (1986) were not included. Insertion 'c' is located between the second and third codons of the insertion 'a'.

Indel Code Size (bp) Position Taxa insertion a 15 101951 all except Anomochloa. Pharus. Streptochaeta. Leptaspis. Orvza. Leersia insertion b 9 101751 Puelia insertion c 6 101951 Raddia. Sucrea insertion d 6 101698 Olyra deletion e 3 101741 Buergersiochloa woody bamboos and the Old World tropical woody bamboos. The strict consensus tree is shown in Figure 3. The Streptochaeteae and Anomochloeae formed the most basal clade within the family, with the Fhareae as the next most basal clade. Leptaspis banksii appeared as sister to the two species of Pharus. The higher grass clade was strongly supported with a decay value of at least four and the presence of the 15 bp insertion (a. Table 4). A novel result is the resolution of Puelia olvriformis in a position basal to the higher grasses (Figures 2, 3).

The higher grasses formed a dichotomy between the bambusoid, oryzoid, and pooid grasses

(BOP clade of Clark et al. 1995) and the PACC clade (as defined by Davis and Soreng 1993).

The BOP clade was robust to only two steps longer in the decay analysis (Figure 2), but the

PACC clade was strongly supported by 29 base substitutions and a decay value of four or greater.

The three lineages within the BOP clade, the bambusoid, oryzoid, and pooid grasses, formed a trichotomy in the consensus tree (Figure 3). The bambusoid and pooid clades were each strongly supported by decay values of four or greater, as was the oryzoid clade above

Streptogvna (Figure 2). The two accessions of Streptogyna americana were separated by four FIG. 2. One most parsimonious tree of the complete ndhF sequence data set analysis. The tree has CI = 0.476 and RI = 0.701, with 1350 steps. Arabic numbers above branches indicate the branch length, and the numbers below the branches denote the decay indices. A decay index of "0" means that the branch only appears on all equally most parsimonious trees, but the branch collapses in trees one step longer, "1" indicates the branch still exists in trees one step longer, but collapses in trees two steps longer; and so on. A solid dot indicates that the branch only exists in some of the equally most parsimonious trees, but the branch disappears in the consensus tree. Letters denote indels as listed in Table 4; solid bars indicate insertions, and shaded bars indicate deletions. 43

Oryza sativa Leersia virginica Maltebrtmia petiolata Ehrharta catycina Strepto^im americana i Streptogyna americana 2 Arundinaritt gigantea Pseudosasa japonica Shibataea kumasaca Sasa variegata Ampelocaiamus scandens Bashama fargesii Phyllostachys pubescens Phyllostachys bambusoides Fargesia robusta Yushania exilis Chimonobambusa marmorea Buergersiochloa bambusoides Sucrea maadata Raddia distichophylla Olyra Uuifolia Uthachne humilis Uthachne pauciflora Eremitis sp. nov. Pariana radiciflora Glaziophyton mirabile Atvimia gracilis Arthrostylidium ecuadorense Rhipidocladum pittieri Apoclada simp^ Guadua paniculata Otatea acuminata Chusquea latifolia Chusquea circinata Neurolepis aperta Bambusa ;i£i.bambos Bambusa stenostachya Racemobambos microphylla Melocanna baccifera Cephahstachyum pergracite 11 Schizpstachyum luzonicum 9 Hickelia madagascariensis 15 Nastus elarus 13 Diarrhena obovata 40 Poa pratensis 1 45 . Avena sativa 12 26 Phaenosperma globosum ? 35 T- Brachyetytrum erectum Sporobolus indicus Zoysia matrella Zea mays Sorghum bicolor Phragmites australis Molinia caerulea Puelia olyriformis Pharus latifoUus Pharus lappulaceus >• Leptaspis banksii 18 Streptochaeta angustifolia Anomochloa marantoidea Joinvillea ascendens 44 base substitutions; their association with the oryzoid clade persisted only in trees one step longer (Figure 2). In the separate analysis of the 3' half of the gene, S. crinita did appear as sister to S. americana (tree not shown). The putative sample of Humbertochloa greenwavi in a preliminary analysis emerged as sister to the Oryzeae, an association strongly supported by decay analysis and the lack of the 15 bp insertion.

Within the monophyletic bambusoid clade, four clades corresponding to the temperate

Bambuseae, the Olyreae/Parianeae, the New World tropical Bambuseae, and the Old World tropical Bambuseae formed a polychotomy in the consensus tree (Figure 3). In the most parsimonious trees, the Olyreae/Parianeae clade was sister to the temperate Bambuseae, while the New World and Old World tropical Bambuseae formed another clade (Figure 2). The temperate Bambuseae were supported by a decay value of three, and the Olyreae/Parianeae clade by a decay value of four or greater, but the dichotomy between New World and Old

World taxa was supported by decay values of 0 and 1, respectively. Certain clades, such as die Guadua Kunth/Apoclada McOure/Otatea McClure & Smith and die Gla/inphvtnn Franch. through Otatea clades, as well as the Bambusa Schreber/Racemobambos Holttum clade, were strongly supported by decay values of four or greater, but otherwise little resolution within the tropical and temperate Bambuseae clades was recovered. Buergersiochloa bambusoides was basal within the Olyreae/Parianeae clade, and the remaining Olyreae and die Parianeae were each strongly supported as monophyletic sister clades by decay values of four or greater

(Figure 2).

The topology from the neighbor-joining analysis was congruent widi those from parsimony analysis, with three significant exceptions (Figure 4). First, Puelia olvriformis was resolved as sister to the BOP clade, with the PACC clade sister to this unit. Second, the

Olyreae/Parianeae appeared as sister to a monophyletic Bambuseae. Third, while the Old

World tropical Bambuseae were retained as a monophyletic clade, the New World tropical FIG. 3. Strict consensus tree of the complete ndhF sequence data set analysis. The tribal and subfamilial classification is shown according to Clayton and Renvoize (1986); PACC is an abbreviation for the clade including Panicoideae, Arundinoideae, Chloridoideae, and

Centothecoideae. Shaded taxa indicate those that form the bambusoid clade. The geographic distribution of the tribe Bambuseae were labeled as temperate. Old World tropical, and New

World tropical respectively. Oryza saliva Leenia virginica J Oryzeac MaUebrmia penoUtta Ehrharux catycina — Ehrhaiteae Stnptogyna americana 1 Sireptogyna americana 2 ] Strcptogyncae Diarrhaia obovaia — Diarrhcneac Poa praiensis Avena saliva ] Pooideae Phaetwsperma globosum — Phaenospcrmateae Brachyelytnun erectum — Brachyelytreae Arundinarui gigantea Pseudosasaptponica Amptlocalamuxsauutaa BashutidfargcsU Phyttostachjspubesceia PhjOosiadiysbambusoides Temperate Fargesiarabusta Bambuseae Yusbaaideimi Chimonobambiaa marmorea SJiibataea btmasaaz Sasavariegqu Buergerslocfiloa bambusoides Suaea maadata Raddid disdchophytta Olyreac Olyrn latyblia UthadmehumiSis Udiadmepaitcfflortt Ereimlissp.nm. Pariaiui radie^ra ] Pariancae GtaziopkyUm nUrabik AMmia gracilis ArthrosQlidiunt ecuadorense RUpidmJadum pittieri New World Apoclada simplex Guadua paniculata Tropical Otatea acuminala Bambuseae Chusguea latffolia Chusquea drdaaia Neurolepis aperta BambusaaS-bambos Bambusa stenostachya Racetnobambos mircrvphylla Old World Uelocama baccifera Tropical Cephaloslachyum pergracite Schizpsiackym liaoiucum Bambuseae HickeUa madagascariensis tfaslus eUuus Sporobolus indicus Zoysia matrella Zea mays Sorghum bicolor •PACC Phragmites amtralis Molinia caerulea Puelia olyriformis — Bambuseae Pharus latifolius Pharus lappulaceus ^ Phareae Leptaspis banksii Strtptochaeta angustifolia — Streptochaeteae Anomochloa marantoidea — Anomochioeae JoinviUea ascemlens FTG. 4. Neighbor-joining tree of the complete ndhF sequence data set analysis using the

Kimura 2-parameter distance measure. The tribal and subfamilial classification are shown according to Clayton and Renvoize (1986), and geographical distribution is indicated for tribe

Bambuseae. T.B. = temperate Bambuseae, N.B. = the New World tropical Bambuseae, O.B.

= the Old World tropical Bambuseae, 01. = Olyreae/Parianeae, Or. = Oryzeae/Ehrharteae, Po.

= Pooideae, PACC = Panicoideae, Arundinoideae, Chloridoideae, and Centothecoideae. 48

Yitshatia edits Bashaniafitriesii • FargaUt nbiata - PhyUostachys pubescau r - Phyllostachys bambosoides H ^^nptlocalamus scandais T.B. I ^Qumonabambusamarmomx —Uioia varieptta ÂnauSnaria gigantea V^Pseudosasa japonica ^Shibataeaburiasaca 1004 \Ôtatea acumrnam V-Âpociada simplex ^Gaatiiti pantadata -^GUmophyton minbilis SB. •'Alvimia gracilis • - • • Arthrostylidium ecuadorense - Rfupiihcladum pittieri I Hicketia madagascariensis Jr- Melocaruta bacdfera ^^Cephalostachyum pergradle Schizpstachyttm luzonicum I Racemobambos microphytia O.B. Samdujo stenostachya V^Bambttsa (^.bambos I ffastus elatus — Neurolepis aperta Chusquea lattfolta N.B. •Chusquea circinata — Otyra latifolia omo Uthachne humilis ^ Uthachne paucijlora ' Raddia dstichopkylla '^Sucrea maculosa 01. ' Pariana radiciflara ' Eremitis sp. nov. •'Buergersiochioa bambusoides tStreptogpia americana I Streptogyrta americana 2 Leersia virginica Or. Maitebrunia petiolaut Oryza sativa Ehrharta calydna - Brachyeiytrum erectum Phaenosperma gbbosum Poa pratensis Po. 'Âvenasativa - Diarrhena obovata 'Puelia olyriformis m p* Molinia caerulea ^ Phragmites australis Zeamay Sorghum bicobr PACC "" Zovsia man-ella * ' • ~ Sporobolus indicus ' Leptaspis banksii ^Pharus ladfolia ^Pharus lappulaceus 'Streptochaeta angustifolia •'Anomochloa marantotdea - Joinvillea ascendens 49

Bambuseae became paraphyletic with the resolution of the Neurolepis Meisner + Chusquea

Kunth clade as basal to the rest of the tropical Bambuseae.

Morphological/Anatomical Data Set Analysis. Parsimony analysis of this data set generated 383 equally most parsimonious trees of 160 steps, with a consistency index 0.412 and a retention index 0.743; one of these trees is illustrated in Figure 5. The strict consensus tree (Figure 6) was generated from the most parsimonious trees to evaluate the overall topology and character support (excluding those characters that are highly homopiasious within the

Bambuseae) is mapped onto this tree (Figure 6). No character unambiguously supported the association of Streptogyneae + Oryzeae + Bambuseae + Olyreae/Parianeae, although the disarticulation of the leaf blades (Character 14) is reversed only in the Oryzeae. The

Bambuseae (woody bamboos) and the herbaceous Olyreae/Parianeae clades each were resolved as monophyletic; however, their relationships to the Oryzeae and Streptogyneae remained unresolved as shown by the polychotomy in the strict consensus tree (Figure 6). Three homopiasious characters (32,49, and 51) supported the grouping of the Oryzeae with the

Bambuseae in the most parsimonious tree (Figure 5). The association of this clade with the

Olyreae/Parianeae was supported by five homopiasious characters (12, 35, 39,48, and 50)

(Figure 6). A decay value of three or greater supported the monophyletic woody bamboo and rice clades, but the herbaceous Olyreae/Parianeae clade was robust only in the consensus tree

(Figure 5).

The Olyreae/Parianeae clade was supported unambiguously only by Character 22, the presence of unisexual spikelets (Figure 6). The shift from pointed to awned lenunas (Character

32) united the four major clades, but was reversed in the Olyreae/Parianeae clade above

Buergersiochoa and in the Oryzeae + Bambuseae clade. However, the presence of blunt lemmas, as well as two anatomical characters (56 and 61), provided unambiguous support for the Olyreae/Parianeae clade above Buergersiochloa (Figure 6). Fused stamen filaments

(Character 38) and horizontal dumbbell-shaped silica bodies (Character 60) support the FIG. 5. One most parsimonious tree of ttie morphological and anatomical data set analysis.

"Basal grass" refers to the combination of Anomochloeae, Streptochaeteae, and Phareae, the basal lineages of the grass family, as a functional outgroup. Arabic numbers above branches indicate the branch length, while the numbers below the branches are the decay indices. A decay value of "0" means that the branch appears in all equally most parsimonious trees, but it disappears in trees one step longer; "1" indicates that the branch exists in trees one step longer, but collapses in trees two steps longer; and so on. A solid dot indicates that the branch appears in only some of the equally most parsimonious trees, but the branch does not appear in the consensus tree. 51

Buergersiochloa bambusoides

Pariana radiciflora

Eremitis sp. nov.

Sucrea maculata

Olyra latifolia

Uthachne pauciflora

Raddia distichophylla

Arundinaria gigantea

Sasa variegala

Yushania exilis

Chimonobambusa marmorea

Shibataea kumasaca

Phyllostachys pubescens

Ampelocalamus scaruiens

Guadua paniculata

Cephalostachyum pergracile

Melocanna bacciferaNasnts elatus

Bambusa aff. bambos

Racemobambos microphylla

Hickelia madagascariensis

Nastus elatus

Alvimia gracilis

Otatea acuminata

Rhipidocladum pittieri

Chusquea latifolia

Neurolepis aperta o Oryza sativa

{Z Leersia virginica

Streptogyna americana

Basal grass FIG. 6. Strict consensus tree of the morphological and anatomical data set analysis. Tribal classification is shown according to Clayton and Renvoize (1986). Supporting characters for each major clade were mapped, and the numbers above or below the arrows are the character numbers as listed in Table 2. Upward arrows indicate unambiguous characters, and downward arrows denote homoplasious characters. 53

Buergersiochloa bambusoides - Olyreae « « !J n 57 Pariana radiciflora 4—ill tiii J Pari;Parianeae Eremitis sp. nov.

Sucrea maciilata

Olyra latifolia 4rbrir Olyreae Lithachne pauciflora tHT Raddia distichophylla

Anmdinaria giganiea

Ampelocaiamus scandens

Shibataea kumasaca

Phyllostachys pubescens

Chimonobambusa marmorea

Sasa variegata Yushania exilis

Alvimia gracilis

iTiiUX Otatea acuminata

Rhipidocladum pinieri Bambuscae Guadua paniculata

Bambusa aff. bambos

Melocanna baccifera

Cephalostachyum pergracile u 45 &: Racemobambos microphylla

Hickelia madagascariensis irH- i« :i (s 5»s: Nastus elatus

Chusquea latifolia -Wrrt- Neurolepis aperta

ts W tt U S9 Oryza sativa TTTtT iiiiUiU Oryzeae 4 14 M 174144 51 &2M Leersia virginica Streptogyna americana Streptogyneae

Basal grass 54

Olyreae/Parianeae clade but exhibit reversals, and Qiaracter 60 is homoplasious within the

Bambuseae. The Parianeae was supported as monophyletic by the flattened pedicels of the male spikelets (Character 28), and the remainder of the Olyreae (excluding Buergersiochloa) was supported as monophyletic by the presence of indurated mature female anthecia (Character

33) and the reversal to free stamen filaments (Character 38).

The woody bamboo clade was supported as monophyletic by five unambiguous synapomorphies (Characters 1,3, 10, 13, and 21). The presence of an outer ligule (Character

16) unites this clade but is also present in Streptogvna americana. Characters 19 and 35 support the monophyly of the woody bamboos as reversals, but additional changes in

Character 19 occur within this clade. The woody bamboos form a monophyletic clade above

Neurolepis aperta based on the presence of branching in the aerial culms (Character 9), and the bulk of the woody bamboos are united by the presence of hollow culms (Character 11) above

Chusquea latifolia (Figure 6). Within the hollow culm clade, Shibataea kumasaca and

Phvllostachvs pubescens are united by the presence of leptomorph rhizomes (Character 7), and

Hickelia madagascariensis and Nastus elatus are united by three homoplasious characters (17,

24, and 29), but otherwise a large terminal polychotomy is recovered.

Reduced Molecular Data Set Analysis. Parsimony analysis of ndhF sequence data for the same terminal taxa as in the morphological/leaf anatomical analysis yielded 18 equally most parsimonious trees with 434 steps. Each tree (Figure 7) had a consistency index of 0.565 and retention index of 0.688. Outgroup substitution using Joinvillea. Anomochloa. and

Streptochaeta produced exactly the same topologies. When Brachyelvtrum was used as the outgroup, the temperate Bambuseae, the Olyreae/Parianeae, and the tropical Bambuseae resolved as a trichotomy, but otherwise the topology was similar; therefore, long branch attraction was not a problem. FIG. 7. One most parsimonious tree of the reduced ndhF sequence data set analysis. The same terminal taxa as in the morphological/leaf anatomical data set were selected. Tribal classification is shown according to Clayton and Renvoize (1986), and geographic distribution of tribe Bambuseae is indicated. Arabic numbers above the branches indicate the branch length, and the numbers below the branches are the decay indices. A decay value of "0" means that the branch appears in all equally most parsimonious trees, but it disappears m trees one step longer, "1" indicates that the branch exists in trees one step longer, but collapses in trees two steps longer, and so on. A solid dot indicates that the branch appears in only some of the equally most parsimonious trees, and the branch does not appear in the consensus tree. Buergersiochloa bambusoides - Olyrcae

Pariana radiciflora J Parianeae Eremitis sp. nov.

Sucrea maculata

Raddia distichophylla Olyreae Olyra latifolia

Uthacfine pauciflora Alvimia gracilis

Rhipidocladum pitteri New World Oiatea acuminata Tropical Guadua panicidata Bambuseae

Neurolepis aperta

Chusquea latifolia

Bambusa aft.bambos

Racemobambos microphylla

Melocanna baccifera Old World Cephalostachyum pergracile Tropical Bambuseae Hickelia madagascariensis

Nastus elatus

Arundinaria gigantea

Shibataea kumasaca Sasa variegata Temperate Ampelocalamus scandens Bambuseae Phyllostachys pubescens

Yushania exilis

Chimonobambusa marmorea

Streptogyna americana - Streptogyneae Oryza sativa Oryzeae Leersia virginica

Streptochaeta angustifolia 57

The Bambuseae + Olyreae/Parianeae clade was strongly supported as monophyletic with a decay value of three or greater (Figure 7). Streptogyna americana appeared as sister to this clade, with the rice clade sister to that. The association of the these three clades, which formed a trichotomy in the consensus tree (not shown), was supported by 49 base substitutions and a decay index of three or greater (Figure 7). Within the Bambuseae +

Olyreae/Parianeae clade, the temperate Bambuseae occupied the basal position, and the

Olyreae/Parianeae clade was sister to the tropical Bambuseae. The Olyreae/Parianeae clade was robust with a decay index of three or greater, with Buergersiochloa bambusoides forming the basal lineage. Within the tropical Bambuseae, a relatively weakly supported dichotomy between the New World and Old World taxa was recovered as in the complete data set analysis

(Figure 2).

Combined Data Sets Analvsis. Parsimony analysis of the combined molecular and morphological/Ieaf anatomical data sets provided better resolution by yielding two equally most parsimonious trees of 613 steps. Each tree (Figure 8) had a consistency index of 0.509, and a retention index of 0.691. The only difference between the two shortest trees are the positions of Yushania exilis and Chimonobambusa marmorea. such that one or the other species is sister to the Ampelocalamus scandens/Phvllostachys pubescens clade (solid circle in Figure 8). Two major monophyletic clades were recovered; the Bambuseae + Olyreae/Parianeae clade, and the

Oryzeae + Streptogyneae clade. The monophyletic bamboo clade was robust, surviving up to four or more steps longer in the decay analysis, while the Oryzeae + Streptogyneae clade only survived one step longer. The association between the two major clades, however, was supported with a decay value of four or greater (Figure 8).

In the monophyletic bamboo clade, the herbaceous bamboos (Olyreae/Parianeae) and the woody bamboos (Bambuseae) were well separated as the decay analysis showed that each clade survived at least four steps longer (Figure 8). Buergersiochloa bambusoides was resolved as basal in the herbaceous bamboo clade, and was distinct from the other herbaceous FIG. 8. One most parsimonious tree from analysis of a combined ndhF and

morphological/leaf anatomical data matrix. Tribal classification is shown according to Clayton

and Renvoize (1986), and geographic distribution of tribe Bambuseae is indicated. Arabic

numbers above the branches indicate the branch length, and the numbers below the branches

are die decay indices. A decay value of "0" means that the branch appears in all equally most

parsimonious trees, but it disappears in trees one step longer; "1" indicates that the branch exists in trees one step longer, but collapses in trees two steps longer, and so on. A solid dot

indicates that the branch appears in only some of the most parsimonious trees, but the branch

does not appear in the consensus tree. Shaded taxa are those of the monophyletic bambusoid clade we classify as the Bambusoideae. Our taxonomic treatment is indicated; 01 = Olyreae,

and B = Bambuseae. 59

BttmenwdUoaiBSmEusaiies — Olyrcae

PSfjmgrg^j^in »4 ] Parianeac SucreSmiatUua

Olyreac

iMimcimeipaucSlo'S-

Aruhdmarurgigantea

Sa^vmemia

Temperate PliittostiKfm^uBS^iur Bambuseae

GKm^ptiaitdma^

ARiamsmila

New World OtSeacxmmi& Tropical (JuWaJ^vaMa Bambuseae

Bg^^ifaSiBam^s;

Old World i^lgpmiaJBacalm Tropical GpKafaSmclvi^im^&g Bambuseae HS3^Mim^'gasca^

NasniF'elatus Oryza sativa Oryzeae Leersia virginica ] Streptogyna americana — Streptogyneae Streptochaeta angustifolia/Basal grass 60 bamboos. The rest of the Olyreae fonned another strongly supported clade which was sister to the Parianeae clade, with good internal resolution.

In the woody bamboo clade, two groups were resolved, one representing the temperate

Bambuseae, the other containing the tropical Bambuseae. The temperate bamboo clade showed strong support from the decay analysis, as it survived as least four steps longer, but resolution within this clade was only weakly supported. The tropical bamboo clade split into two groups: the Old World tropical bamboos, and the New World tropical bamboos, but the decay analysis indicated that the division of these two groups was not strong, with decay indices of one and zero respectively.

ndhF Gene Sequence Divergence and Genetic Distance Analvsis. Excluding polymorphisms introduced by gaps, the MEGA analysis showed that 701 of 2115 (33.1%) molecular characters in the complete data set were variable, and 448 (21.2%) were informative

(Table 5). However, the sequences of the monophyletic Bambusoideae clade showed much less variation: only 341 (16.1%) characters were variable, and 199 (9.4%) were informative.

Results for various clades within the Bambusoideae are presented in Table 5, and show low values, but it is noticable that the herbaceous bamboos (only eight species included) have more variable and informative characters (211 and 123) than the woody bamboos (29 species) (184 and 73 respectively).

The neighbor-joining tree (Figure 8) indicated clearly that the woody bamboos were the group with lowest inter-taxa genetic distances as indicated by branch lengths, compared with all other grasses. The temperate woody bamboos had the shortest branch length, less than

0.001, in the woody bamboo group. Following the temperate woody bamboos, the Old World tropical woody bamboos and the New World tropical woody bamboos were the next two groups with lowest genetic distance. The herbaceous olyroids, however, had a genetic distance similar to grasses of the oryzoid, Fooideae, and PACC clades. 61

TABLE 5. Sequence variation within the complete ndhF sequence data set and the subsets of bambusoid taxa. The aligned sequences include insertions and deletions with a total length of 2115 bp.

Taxon Variable Characters Informative Characters Complete data set (60 spp.) 701 (33.1%) 448 (21.2%) Bambusoid clade (37 spp.) 341 (16.1%) 199 (9.4%) Olyreae/Parianeae (8 spp.) 211 (10.0%) 123 (5.8%) Bambuseae (29 spp.) 184 (8.7%) 73 (3.4%) Temperate Bambuseae (11 spp.) 20 (0.9%) 6 (0.3%) Tropical Bambuseae (18 spp.) 149 (7.0%) 59 (2.8%) New World tropical Bambuseae (10 spp.) 135 (6.4%) 46 (2.2%) Old World tropical Bambuseae (8 spp.) 49 (2.3%) 8 (0.4%)

P-distance analysis illustrated a similar overall pattern (Table 6). The lowest p- distance, between Arundinaria gigantea and Bambusa aff. bambos. the representives for temperate woody bamboos and the Old World tropical woody bamboos, was only 0.0101.

Next to the temperate and the Old World tropical woody bamboos, the pairwise comparisons indicated that Bambusa aff. bambos and Otatea acuminata, representing the New World and die

Old World tropical woody bamboos, and Arundinaria gigantea and Otatea acuminata, were the two groups with the lowest p-distance, 0.0134 and 0.0149 respectively. However, the distances between the herbaceous olyroid bamboos, with Olvra latifoUa as representative, and the three woody bamboo representatives, were two times higher than the distances between the woody bamboo groups (Table 6).

DISCUSSION

Phvlogeny based on ndhF sequence data. Our results, based on more extensive sampling within the true bamboos plus Buergersiochloa and Puelia. and including additional taxa of Oryzeae and Phareae, were congruent with those obtained by Clark et al. (1995). As in TABLE 6 . Proportion of nucleotide difference and standard error of the ndhF gene sequences among different groups of bambusoid and other grasses. Distances are in the upper-right matrix, and standard errors are in the lower-left matrix. Abbreviations; Or. sat. = Oryza satiya; Po. prat. = Poa pratensis: A. gig. = Arundinaria pipantea: B. bam. = Bambusa aff. bam bos: Ot. acum. = Otatea acuminata: 01. lati. = Olvra latifolia: S. ind. = Sporobolus indicus: Z. mays = Zea mays: Ph. aus. = Phragmites australis.

Taxa Or. sat. Po. prat A. gig. B. bam. Ot. acum. 01. lati. S. ind. Z. mays Ph. aus.

Or. sat. 0.0793 0.0431 0.0431 0.0469 0.0614 0.0721 0.0629 0.0552 Po. prat 0.0059 0.0639 0.0629 0.0682 0.0749 0.0893 0.0788 0.0664 A. gig. 0.0045 0.0054 0.0101 0.0149 0.0259 0.0581 0.0462 0.0364 B. bam. 0.0045 0.0053 0.0022 0.0134 0.0288 0.0567 0.0447 0.0351 Ot. acum. 0.0047 0.0055 0.0027 0.0025 0.0341 0.0629 0.0495 0.0384 01. lati. 0.0053 0.0058 0.0035 0.0037 0.0040 0.0711 0.0581 0.0490 S. ind. 0.0057 0.0063 0.0051 0.0051 0.0053 0.0056 0.0557 0.0461 Z. mays 0.0053 0.0059 0.0046 0.0045 0.0048 0.0051 0.0050 0.0307 Ph. aus. 0.0050 0.0054 0.0040 0.0040 0.0042 0.0047 0.0046 0.0038 63

Clark et al. (1995) and Duvall and Morton (1996), the Anomochloeae, Streptochaeteae, and

Phareae (here also including Leptaspis") were resolved as the most basal lineages within the

grass family (Figures 2,3). The position of the Brachyelytreae, Phaenospermatae, and

Diarrheneae at or near the base of the pooid clade (Davis and Soreng 1993; Clark et al. 1995)

remained stable. The Ehrharteae were again closely associated with the Oryzeae, with the

Streptogyneae weakly supported as basal within the oryzoid clade. The bambusoid clade, consisting of the Bambuseae and the two herbaceous tribes Olyreae

(including Buergersiochloa) and Parianeae (as recognized by Clayton and Renvoize 1986), continued to be strongly supported as monophyletic (Figures 2-8). Perhaps the most significant novel result was the resolution of Puelia at the base of the higher grasses (Figures 2,

3) or at the base of the BOP clade (Figure 4). Another important result was the resolution of

Buergersiochloa as the basal lineage within Olyreae/Parianeae (Figures 2-8).

POSITION OF LEPTASPIS WITHIN THE PHAREAE. The association of Leptaspis and Pharus in the Phareae has been considered unambiguous and noncontroversial, based on morphological and leaf anatomical characters (Judziewicz 1987; Soderstrom et al. 1987;

Kellogg and Watson 1993). Members of the Phareae share resupinate, pseudopetiolate leaves; oblique venation in the leaf blades; terminal inflorescences consisting of paired male and female spikelets, with the females typically larger, one-flowered spikelets; symmetrical midribs with complex vasculature forming semi-circular keels; poorly developed arm cells; the lack of papillae on the long cells; and the apparent absence of microhairs (Metcalfe 1960; Clayton and

Renvoize 1986; Judziewicz 1987; Soderstrom et al. 1987). Based on the presence of persistent, nondisarticulating inflorescence branches and glumelike bracts subtending the female spikelets, Soderstrom et al. (1987) hypothesized that Leptaspis represented the basal taxon in the Phareae. The results of the present analysis strongly support Leptaspis and Pharus as sister genera, and are consistent with Soderstrom et al.'s suggestion that Leptaspis is basal.

Sampling from within Scrotochloa. the third genus of the tribe (Soderstrom et al. 1987; Clark 64 and Judziewicz, in press) would be necessary, however, in order to establish the basal lineage in the tribe with more confidence.

PUELIA AS THE BASAL LINEAGE WITHIN THE HIGHER GRASSES. Based on parsimony analysis, Puelia was resolved as the next most basal clade in the family above the

Phareae, representing the most basal lineage in the higher grasses. Puelia olyriformis was unequivocally placed within the higher grasses based on the shared 35 base substitutions and the 15 bp insertion (this lacking in the Olyreae and Ehrharteae). Its position as basal and sister to the BOP + PACC clade (the remainder of the higher grasses) was also strongly supported by a decay value of four or greater for the BOP + PACC clade, and the presence of an autapomorphic insertion (b. Table 4) in E*uelia. In a test of the classification of Puelia with the

Bambuseae, 22 additional steps were required to force Puelia into the Bambuseae (as sister to the Old World tropical Bambuseae clade), and 11 extra steps were required to place it at the base of the bambusoid clade. Placement of Puelia at the base of the BOP clade in the neighbor- joining analysis (Figure 4) is in conflict with the parsimony analysis, but indicates that Puelia is not a member of the bambusoid clade.

Puelia has not been well studied morphologically and anatomically, although some data are published (Soderstrom and Ellis 1987). Adequate material for molecular analysis was not available as most of the herbarium specimens of Puelia (and Guaduella as well) were either relatively old or did not have sufficient leaf material; the exception was R. olvriformis. which was sampled from a herbarium specimen. The association of Puelia with the Bambuseae

(Clayton and Renvoize 1986; Kellogg and Watson 1993) was based on a number of characters now considered to be symplesiomorphic: the forest habitat; pseudopetiolate, relatively broad leaves; three lodicules (when present); six stamens; three stigmas; small embryo with a linear hilum; and arm and fiasoid cells in the mesophyll. The presence of the outer ligule was used to link Puelia with the Bambuseae, but our results suggest that this character is independently derived in Puelia. the Bambuseae, and Streptogyna. The suggestion by Clayton and Renvoize 65

(1986) that Puelia was possibly derived from Nastus Gmel., an Old World woody bamboo genus with species in Madagascar and Reunion Island, is a much less parsimonious hypothesis. In a cladistic analysis using morphological and leaf anatomical characters (Kellogg and Watson 1993), Streptochaeta and Streptogvna formed a clade with Guaduella and Puelia that was sister to the woody bamboos.

Dimorphic shoots, in which some are leafy and some bear inflorescences only, are sometimes found in both Puelia and Guaduella (Clayton and Renvoize 1986). The spikelets of

Puelia consist of usually two or three glumes and several florets, of which the lower three to six are male or neuter and the uppermost one is female (Koechlin 1962). Both pistillodes and staminodes are relatively well developed in these florets, respectively (E. J. Judziewicz, pers. comm.). In Guaduella. the spikelets are similar but somewhat less specialized in that the lower one to three florets are male and the rest are bisexual (Koechlin 1962; Clayton and Renvoize

1986). Differences in leaf anatomy between the two genera are noted (Soderstrom and EUis

1987); Puelia lacks microhairs and has superposed bundles in the midrib while Guaduella has multicellular microhairs and has bundles in a single plane in the midrib. We attempted to sequence three species of Guaduella from herbarium material, but unfortunately we were unable to obtain reUable results from any of them so representatives of Guaduella were not included in the analysis. Even if Guaduella is not closely related to Puelia. the resolution of

Puelia at the base of the higher grasses in this analysis marks the transition from one bisexual

(or female fertile) floret in the most basal lineages to the presence of two or more bisexual florets per spikelet. The staminodes and pistillodes in Puelia florets support the assumption that multiple bisexual florets per spikelet was the ancestral condition in the higher grasses, or at least the BOP clade.

While our results indicate that Puelia should be excluded from the bambusoid clade, further study from both morphological and molecular perspectives is needed to estimate its phylogenetic position within the higher grasses more robustly. Likewise, the position of 66

Guaduella is not known; it may prove to be closely related to Puelia. or its relationships may lie closer to the true bamboos. For the moment, we have placed these two genera as Incertae

Sedis relative to our circumscription of the Bambusoideae (see Taxonomic Treatment).

STREPTOGYNEAE AND THE ORYZODD Q.ADE. Maltebrunia Kunth, an African genus of Oryzeae, was included in this analysis in addition to the previously sampled Oryza and Leersia (Clark et al. 1995). The Oryzeae were robustly supported as monophyletic with a decay value of four or greater. The tribe is marked by the loss of the 15 bp insertion (a' in

Figure 2); this may be a synapomorphy for the tribe, but further sampling, particularly of

7i7ania and/or Zizaniopsis. is needed to confirm this. The association of Ehrharta. which has the 15 bp insertion, with the Oryzeae continues to be strongly supported. This sister relationship has been recovered in other morphological and molecular cladistic analyses where both tribes were examined (Kellogg and Campbell 1987; Kellogg and Watson 1993;

Cummings et al. 1994). An ndhF sequence from unvouchered leaf pieces identified as

Humbertochloa greenwayi (Phyllorachideae) was included in a preliminary analysis; it lacked the 15 bp insertion and appeared as sister to the Oryzeae. We excluded this sequence from subsequent analyses due to uncertainty as to its identification, but if correct, the results support a relationship between the Oryzeae and Phyllorachideae, as suggested by Clayton and Renvoize

(1986).

The Streptogyneae were grouped with the monophyletic oryzoid clade; however, this resolution was not very robust as this association survived in trees only one step longer (Figure

2). Clark et al. (1995) noted similar weak support for the position of Streptogyneae at the base of the olyroid clade. In spite of the relatively weak support for this resolution, four extra steps were required to force Streptogyna to a basal position within the bambusoid clade, and 18 extra steps were needed to associate Streptogyna with the herbaceous Olyreae. The position of the

Streptogyneae with respect to the olyroid and bambusoid clades is further discussed in the

Morphological and Combined Data Sets sections. 67

Three samples representing the two species of the genus, Streptogyna crinita and S.. americana. were sequenced. For S. crinita. a sequence for only 3' half of the gene was obtained. In an analysis in which only the 3' half of the gene was used for all of the terminal taxa, these two species formed a monophyletic clade in the same position at the base of the olyroid clade. The overall topology of that analysis was nearly congruent with that obtained

from the analysis of the complete gene sequences; the only differences were found in the

positions of some taxa of woody bamboos within the bambusoid clade. Streptogyna americana was the only species in this study for which two accessions from geographically well separated populations were sequenced. Streptogyna americana 1 was collected from Nicaragua, while S. americana 2 was from Panama. The sequences representing these two populations differed by four nucleotides (Figure 2). Intraspecific chloroplast DNA sequence or restriction site variation have been reported (Clegg et al. 1984; Soltis et al. 1992), but we believe that this is the first example of intraspecific sequence variation in the ndhF gene.

In this analysis and in Clark et al. (1995), the three major lineages of the BOP clade formed an unresolved trichotomy (Figure 3). Because of this, Clarlc et al. (1995) left open the possibility that the oryzoid clade could be included within the bamboos. We note that although the relationships among the bambusoid, oryzoid, and pooid clades are ambiguous based on ndhF sequence data, a different topology is usually recovered in other molecular analyses.

Davis and Soreng (1993, 1995), Cummings et al. (1994), Nadot et al. (1994), Barker et al.

(1995), Liang and Hilu (1996), and Hsiao et al. (in review) typically derived the oryzoid clade as sister to a pooid + PACC clade, with at least the woody bamboos sister to all of them, suggesting that the oryzoid and bambusoid clades are rather divergent. The morphological similarities between the oryzoids and the bambusoids once again appear to derive largely from shared symplesiomorphies (e.g., the presence of arm and fusoid cells; see also the

Morphological and Combined Data Sets sections). Is rice a bamboo? We take the position that 68 the oryzoids should be excluded from our more restricted definition of the true bamboos, and that furthermore, rice and its allies probably are best recognized as a separate subfamily.

THE BAMBUSOID CLADE. The bambusoid clade was robusdy supported as monophyletic (Figure 2), including representatives of only the Bambuseae and

Olyreae/Parianeae. Within the bambusoid clade, a major clade consisting of a monophyledc temperate woody bamboo group sister to a monophyletic Olyreae/Parianeae was resolved as sister to a clade including all of the tropical woody bamboos; this latter clade was split into a weakly supported division between the Old World and New World tropical bamboos (Figures

2, 3). This overall topology was exactly congment with those recovered by Clark et al. (1995) based on ndhF. and preliminary analyses by Kelchner and Clark (in prep.) and Zhang et al.

(unpubl. data) based on the rpll6 intron. Among the 736 equally most parsimonous trees in this analysis, over 90% showed that the Olyreae/Parianeae clade was sister to the temperate woody bamboo clade, making the Bambuseae paraphyletic. However, only one additional step was needed to force the Olyreae/Parianeae clade sister to a monophyletic Bambuseae, and monophyly of both the Olyreae/Parianeae and Bambuseae was recovered in the neighbor- joining analysis (Figure 4).

Within the olyroid clade, Buergersiochloa was strongly supported as basal and sister to the Parianeae and the remainder of the Olyreae (Figures 2, 3). The Parianeae was strongly supported as monophyletic, as was the clade consisting of the remainder of the Olyreae (Figure

2). Sequence divergence of Lithachne from the other Olyreae was remarkably high (Figures 2,

4). The association of Sucrea and Raddia was also strongly supported by the compound insertion (c in Table 4, Figure 2) and a decay value of four or greater.

The temperate woody bamboos were relatively well supported as a monophyletic lineage with a decay value of diree (Figure 2), but robust resolution within the clade is lacking, and branch lengths are extremely short (Figure 4). There is no clear indication as to which taxon might be basal within this clade. The association between the two clades of tropical 69 bamboos disappears in the consensus tree, and each clade is only relatively weakly supported

(Figures 2,3). Within the New World bamboos, Apoclada. Guadua. and Htatpa form a robust clade with a decay value of four or greater, and four genera of the Arthrostylidiinae

CGIazinphvtnn- Alvimia. Arthrostvlidium. and Rhipidocladum) form another well supported clade (Figure 2). Soderstrom and Ellis (1987) classified Apoclada and Glazinphvtnn as genera of uncertain placement, while Dransfield and Widjaja (1995) classified both within the

Arthrostylidiinae; these molecular data are the first clearcut evidence for the association of

Apoclada with the Guaduinae. The association of the Chusqueinae ("Neurolepis + Chusquea) with the rest of the New World tropical bamboos is only weakly supported by the parsimony analysis (Figure 2) and is contradicted by the neighbor-joining analysis (Figure 4). The only robust resolution recovered within the Old World tropical woody bamboos is the clade including Racemobambos and the two species of Bambusa (Figure 2). Soderstrom and Ellis

(1987) left Racemobambos with uncertain placement, but Dransfield and Widjaja (1995) and

Stapleton (1995) place this genus in the Racemobambosinae along with two other genera.

GENE SEQUENCE VARIATION AND GENETIC DISTANCES. Gene sequence variation and genetic distance comparisons indicated that ndhF gene evolution in the

Bambuseae was significantly lower than in the Olyreae/Parianeae clade and other grass groups.

Possible explanations for the apparent slowdown of gene evolution in the woody bamboos are:

1) the long generation times for the woody bamboos; and 2) recent evolution of the woody bamboos, which did not allow them to accumulate sequence variation.

It was reported that the shorter generation-time grasses had higher evolutionary rates than the longer generation-time palms (Gaut et ai. 1992). It usually takes ten to as much as a hundred years or more of vegetative growth until a woody bamboo will bloom, produce seeds for the next generation, and die (Keng 1982, 1986; Pohl 1991). In herbaceous bamboos and other grasses, generation time is much shorter, and flowering occurs every yeeir once sexual 70 maturity is reached in perennials. A 'generation-time effect' can be hj^othesized easily for the slower gene evolution in the woody bamboos.

The temperate woody bamboos showed the lowest gene evolutionary rate, then the Old

World and the New World tropical woody bamboos. It was noticed from field observations by

Zhang, that it is more common to collect flowering specimens of the tropical woody bamboos than of the temperate woody bamboos, which may imply that the temperate woody bamboos have slightly longer generation times than the tropical woody bamboos. However, the lack of reliable documentation makes it hard to draw any conclusions about the generation time of the woody bamboos in different geographic regions. The other possible reason for the lower sequence variation in the woody bamboos may be that they are a recendy evolved group, such that they have not yet accumulated very much sequence variation. Unfortunately again, the lack of fossils to document this indicates that it is impossible to examine the possibility of woody bamboos as a young group.

Phvlogenv Based on Morphological/Leaf Anatomical Data. Although the association of the Streptogyneae, Oryzeae, Bambuseae, and Olyreae was robust in this analysis (Figures 5,

6), the relationships among these four clades remain somewhat obscure as no unambiguous, nonhomoplasious characters supporting the basal portion of the topology were found (Figure

6). The Oryzeae and Bambuseae were each strongly supported as monophyletic, whereas the olyroid clade including Buergersiochloa was only weakly supported (Figure 5).

MONOPHYLETIC BAMBUSEAE. A monophyletic Bambuseae was robusdy supported by five unambiguous characters: the presence of woody culms (1); strong, well developed rhizome systems (3); the presence of bud(s) at the mid-culm nodes (10); the differentiation of foliage leaves and cuhn leaves (13); and the gregarious flowering cycles at long intervals (21). The presence of an external ligule (Character 16) is homoplasious as it also occurs in Streptogvna. A blunt palea apex (35) characterizes the Oryzeae + bambusoid clade. 71

but a reversal to the bifid apex unites the Bambuseae. Characters 19, 53, and 57 are relatively

noninformative as they are homoplasious throughout the tree. Within the bambusoid clade, resolution was rather poor, although Neurolepis and Chusquea were serially basal and the presence of leptomorph rhizomes consistently united Shibataea and Phvllostachvs. Twenty- nine extra steps were needed when the topology was forced to be the same as that from the molecular analysis, indicating the incongruence between the two data sets.

Our results agree with the implicit or explicit recognition of the woody bamboos as a monophyletic group by other workers (Calderon and Soderstrom 1980; Keng 1982;

Soderstrom and Ellis 1987; Kellogg and Watson 1993; Clark 1995). In particular, Calderon and Soderstrom (1980) defined the woody bamboos based on the following morphological characters: woody and strongly segmented cauline axes; arborescent habit; well developed rhizome system; unbranched new shoots; culm leaves; the pseudopetiolate leaf blades; a commonly well developed outer ligule; and the long flowering cycles. The anatomical characters used by Soderstrom and Ellis (1987) to define the bambusoid "core" are all symplesiomorphies according to recent phylogenetic reconstructions of the family based on molecular data (Clark et al. 1995; Duvall and Morton 1996; Liang and Hilu 1996; Hsiao et al.. ui review). Regardless of rooting, Kellogg and Watson (1993) derived a well supported, monophyletic woody bamboo clade sister to the Streptochaeta/Streptogyna/Pueha/Guaduella clade, but they noted that relationships within the Bambuseae were poorly supported.

MONOPHYLETIC HERBACEOUS BAMBOOS. The placement of Buergersiochloa as sister and basal to the Parianeae and the remainder of the Olyreae is supported by the single synapomorphy of unisexual spikelets (Character 22, Figure 6). The Parianeae and the rest of the Olyreae are united unambiguously by the presence of crenate and cross-shaped silica bodies

(Characters 56 and 61 respectively). Further resolution is obtained by the unusual, flattened pedicels in the male spikelets of the Parianeae, and the indurate paleas of the female spikelets of the rest of the Olyreae. Our results are congruent with those of Kellogg and Watson (1993), 72 who also recovered a monophyletic olyroid clade with Buergersiochloa in a basal position, except that Pharus and Leptaspis nested within their olyroid clade.

Our results are in agreement with the longstanding association of the Parianeae, the

Olyreae, and Buergersiochloa. regardless of the tribal classification (Table 1 in Clark et al.

1995). Buergersiochloa has been classified either as a monogeneric tribe (Blake 1946;

Soderstrom and Ellis 1987; Tzvelev 1989) or more commonly submerged within the Olyreae

(Roshevits 1946; Prat 1960; Fijten 1975; Clayton and Renvoize 1986; Watson and Dallwitz

1992). Blake (1946) suggested the recognition of the Buergersiochloeae based on three characters: the absence of cross-veins in the leaf blades, the long-awned lemma, and the united filaments in the male spikelets. However, after careful reexamination, Fijten (1975) found that all three diagnostic characters were either not unique or uncertain in Buergersiochloa and indicated that these characters were not sufficient for the delimination of a separate tribe, even though Buergersiochloa was considered to be distant from all other Olyreae. In spite of overall morphological similarities between Buergersiochloa and the Olyreae, Soderstrom and Ellis

(1987) noted that the presence of intercostal sclerenchyma, the proximity of the vascular bundles, the small fiisoid cells, and the tall and narrow silica bodies, as well as the lack of crenate and cross-shaped silica bodies so characteristic of the Olyreae and Parianeae, suggested a closer relationship with the woody bamboos. They considered other feamres, including an adaxially projecting keel with simple vasculature and symmetrical leaves, to link

Buergersiochloa to the Olyreae, but concluded that the genus was rather isolated and should be maintained as a separate tribe. Tzvelev (1989) indicated that the structure of the gynoecium in the pistillate flowers of Buergersiochloa. with the two styles separate at the base and connate above as figured by Pilger (1914), was unique among grasses; Fijten (1975), however, had previously noted that this was an artifact of dissection due to the thinness of the stylar tissue.

The Parianeae have been treated more often as a separate tribe (Roshevits 1946; Prat

1960; Calderon and Soderstrom 1980; Clayton and Renvoize 1986; Tzvelev 1989) than being 73 submerged in tiie Olyreae (Soderstrom and Ellis 1987; Watson and Dallwitz 1992). Pariana nested within the olyroid clade of Kellogg and Watson (1993) as sister to Maclurolvra.

Although the Parianeae exhibit leptomorph (monopodial) rhizomes, non-branching inflorescences, two to many stamens, and well developed oral setae according to Calderdn and

Soderstrom (1980), more recent workers have regarded these differences as insufficient to warrant tribal recognition. The Olyreae s. s. (Calderon and Soderstrom 1980; Tzvelev 1989) were distinguished on the basis of pachymorph (sympodial) rhizomes, inflorescences typically rebranching and producing partial inflorescences, and three stamens. Our analysis indicates that the indurated female lemmas also unite this clade, although two unique types of silica bodies unite the Parianeae with the Olyreae s. s.

RELATIONSHIPS AMONG THE OLYREAE/PARIANEAE, BAMBUSEAE,

ORYZEAE, AND STREPTOGYNEAE. The analysis resolved the Bambuseae,

Olyreae/Parianeae, Orj^eae, and Streptogvna as four separate clades, but it provided no further resolution among the four groups. In some of the equally most parsimonious trees, one unambiguous character would link two of these clades, for example, the Oryzeae sometimes were associated with the olyroid clade based on the presence of an entire palea apex (Character

35). However, the Oryzeae were linked in other trees to Streptogyna based on the absence of bladeless sheaths on the seedlings (Character 43). These and other complex character distributions resulted in the lack of resolution among these four clades in the consensus tree

(Figure 6). We interpret the presence of well developed arm cells as a synapomorphy for the bambusoid clade (Olyreae/Parianeae + Bambuseae), with reversals in Buergersiochloa and some of the woody bamboos. Although there was no unambiguous character that grouped the herbaceous bamboos with woody bamboos, the McClade analysis showed that it took only one step longer to force the two bamboo clades to form a monophyletic clade, and only two steps longer to force a monophyletic bambusoid clade sister to a monophyletic Streptogyna +

Oryzeae clade. 74

Phylogeny Based on Combined Data Sets. When the consensus trees from the reduced molecular and morphological data analyses were combined, a polychotomy among the four major clades was recovered (tree not shown), with a topology similar to that of the morphological analysis, except that the woody bamboos were split into two groups (temperate vs. tropical). Much better resolution, both among the basal clades and within individual clades, was obtained when the data matrices were combined before running the parsimony analysis (Figure 8). The bambusoid clade was robusdy supported as monophyletic and sister to the Oryzeae + Streptogvna clade. Within the bambusoid clade, the woody and herbaceous bamboos were each strongly supported as sister, monophyletic clades; the weak molecular but strong morphological support for monophyly of the woody bamboos complemented the strong molecular but weak morphological support for monophyly of the herbaceous bamboos. The combined data sets analysis provides the most robust estimate of relationships within the subfamily to date, although the association of Streptogvna with the oryzoid clade remains relatively weak, indicating that it probably diverged early in the evolution of the BOP clade.

TAXONOMIC TREATMENT

With the resolution of the Bambusoideae s. 1. circumscribed in recent global treatments

(Clayton and Renvoize 1986; Watson and Dallwitz 1992) as polyphyletic according to molecular data (Davis and Soreng 1993; Nadot et al. 1994; Clark et al. 1995; Duvall and

Morton 1996; Liang and Hilu 1996; the present study), the need for a taxonomic reevaluation of the subfamily is imperative. A monophyletic bambusoid clade consisting only of the woody bamboos and the herbaceous olyroid bamboos was recovered by Clark et al. (1995) and Duvall and Morton (1996) based on limited sampling. The results of this study confirm that the monophyly of the bambusoid clade is robust based on molecular data, although no unambiguous morphological synapomorphies were found to support this clade. The 75 combination of molecular and morphological data provide strong support for the recognition of two lineages within the bambusoid clade: the woody bamboos and the herbaceous olyroid bamboos. Good resolution was obtained within the olyroid clade, but only weak support for major groups within the woody bamboos was recovered. We consider this study to represent a first approximation for phylogenetic relationships within the bambusoid clade.

We therefore propose a recircumscription of the Bambusoideae to include the bambusoid clade as supported in our analyses. We exclude the following tribes:

Anomochloeae, Streptochaeteae, Phareae, Oryzeae, Phyllorachideae, Ehrharteae,

Streptogyneae, Centotheceae, Phaenospermatae, Diarrheneae, and Brachyelytreae. The positions of the Puelieae and Guaduelleae are regarded as unresolved for the present, and thus we place these two tribes as Incertae Sedis until further study is completed. A description of the subfamily is provided below. Many of the features used to define this subfamily in the past

(Calderon and Soderstrom 1980; Soderstrom 1981b; see also review in Clark et al. 1995), such as the bambusoid embryo, linear hilum, presence of arm and flisoid cells in the chlorenchyma, pseudopetiolate leaves, and three lodicules, are here interpreted as symplesiomorphies for the grass family. These features are characteristic of but not unique to the Bambusoideae as delimited in this treatment; we base our circumscription of the subfamily on the presence a distinctive suite of characters in a phylogenetic context. The single synapomorphy we have identified for the subfamily is the presence of strongly developed arm cells, which we interpret as having been reversed in Buergersiochloa and a few woody bamboos.

We also propose a tribal level classification of the subfamily based on our results. Two major lineages are strongly supported: the herbaceous olyroid bamboos and the woody bamboos. Strong support for the monophyly of the woody bamboos was obtained, but resolution within this lineage was weakly supported at best. This leads us to adopt a more conservative approach, so we prefer to recognize a single, inclusive tribe (Bambuseae) for the 76 cladisticaily monophyletic woody bamboo lineage. Calderdn and Soderstrom (1980), Tzvelev

(1989), Keng (1987), and Zhang (1992 and references cited therein) among others have recognized from four to six tribes of woody bamboos, but we found little or no support for these divisions in our analyses. As discussed, the herbaceous olyroid bamboos have been classified into one (Olyreae), two (Olyreae and Parianeae), or three (Olyeae, Parianeae, and

Buergersiochloeae) tribes. If Buergersiochloa is included within the Olyreae, then this tribe becomes demonstrably paraphyletic if the Parianeae is accepted. In order to maintain monophyletic taxa, either one inclusive tribe (Olyreae) must be recognized, or the three tribes must be accepted. In order to mamtain parallel ranks within our phylogenetic reconstruction, and because Diandrolyra Stapf, a putatively basal genus within the Olyreae s. s. (Fijten 1975), was not included in our analyses, we have chosen to adopt a broad concept and thus we recognize a single inclusive tribe (Olyreae) of herbaceous olyroid bamboos.

BAMBUSOIDEAE Aschers. & Graebn. Syn. Mitteleur R. 2: 769. 1902.

Synonym; Olyroideae Pilg. in Engl. & Prantl p. p., Nat. Pflanzenfam. Ed. 2. 14d: 168.

1956.

Perennials with rhizomes strongly or weakly developed. Culms woody or herbaceous; erect, scandent, clambering, or vining. Leaves cauline; foliage leaf blades pseudopetiolate, articulated with the sheath, deciduous, often tessellate; inner ligules membranous.

Inflorescences paniculate or complex, frequendy bracteate. Spikelets bisexual or unisexual, the plants then monoecious; if bisexual, then genmiiparous bracts subtending the spikelet proper and higher order spikelets developing (pseudospikelets) or gemmiparous bracts absent (true spikelets); florets one to several per spikelet, rarely many; glumes 1-4 (-7), mostly 2 or 3; lemma several to many-nerved, apex blunt or pointed; palea bifid or entire at apex, dorsally two-keeled; lodicules 3, membranous, lanceolate; stamens 3 or 6, rarely 2 or many; stigmas 2 77 or 3; caryopsis usually dry, rarely with pericarp fleshy; seedling with one or more bladeless sheaths, the first leaf oriented horizontally.

Leaf anatomy: Vascular bundles with double sheaths; midrib vasculature complex or simple; arm cells well developed; flisoid cells present. Leaf epidermal micromorphology:

Bicellular microhairs present; long cells of one surface or the other bearing papillae; stomata mostly abaxial.

Chromosome numbers: x = 7, 9, 10, 11, 12.

Bambuseae Kunth, Mem. Mus. d'Hist. Nat. 2:75. 1815. —TYPE: Bambusa Schreber,

Genera Plantarum 2: 236. 1789.

Synonyms: Arundinarieae Aschers. & Graebn. in Syn. mitteleur. Fl. 2(1): 770. 1902.

Arundinariae E.-G. Camus in Les Bambusees, 15. 1913

Baccifereae E.-G. Camus in Les Bambusees, 17. 1913.

Arthrostylidiae E.-G Camus in Les Bambusees, 16. 1913.

Chusqueae E.-G Camus in Bambusees, 16. 1913.

Bambuseae verae E.-G. Camus in Les Bambusees, 16. 1913.

Dendrocalameae (Benth.) Keng in Fl. 111. PI. Prim. Sin. Gram.: 63. 1959.

Melocanneae (Benth.) Keng in Fl. HI. PI. Prim. Sin. Gram.: 31. 1959.

Bambusodae Lieu in Acta Phytotax. Sin. 18: 316-327. 1980.

Bambusatae Keng & Keng f in J. Bamboo Res. 1: 8. 1982.

Arundinariatae Keng & Keng f. in J. Bamboo Res. 1: 9. 1982.

Glaziophytoneae Keng f. in J. Bamboo Res. 3: 3. 1984.

Neurolepideae Keng f in J. Bamboo Res. 3: 3. 1984.

Oxytenanthereae Tzvelev in Bot. Rev. 55: 156. 1989. 78

Perennials with complicated, segmented, and well developed rhizome systems. Culms woody, usually hollow (solid in Chusquea and a few species of other genera), divided into cylindrical segments by the nodes. Culm leaves (modified leaves with expanded sheaths and

usually reduced blades) present on young shoots; upper half or more of culms with one or more branches (except filazinphvtnn. Greslania. and Neurolepis) per node, branch systems sometimes complex. Leaf sheaths often bearing fimbriae and/or auricular appendages at the summit; irmer ligules membranous; outer ligules present; leaf blades pseudopetiolate, articulated, deciduous, linear to oblong, often with evident cross-nerves. Inflorescences paniculate to complex, with one to several orders of branching; often bracteate. Spikelets with

1 to many florets; glumes (0-) 1-4 (-7) but sometimes very reduced; lemmas multinerved, apex pointed; paleas enfolded by lemmas, several-nerved with an even number of nerves, bicarinate, apex bifid; lodicules 3, membranous; stamens 3 or 6 or rarely 2 or many, filaments usually free; stigmas 2 or 3. Fruit usually a basic caryopsis, sometimes nut-like or fleshy.

Olyreae Kunth, Mem. Mus. d'Hist. Nat. 2: 75. 1815.-- TYPE: Olyra L., Systema namrae 2:

1379. 1759.

Synonym: Olyrineae Rechenb. in Deutsch. R. 6: 5. 1846.

Parianeae C. E. Hubb. in Hutchinson's FI. PL, Monocot. 2: 219. 1934.

Buergersiochloeae Blake in Blumea, Suppl. 3: 62. 1946.

Perennials with relatively weakly developed rhizomes. Culms herbaceous, vegetative branching restricted. Leaf blades pseudopetiolate, articulated, deciduous. Inflorescences paniculate or racemose, often rebranching to produce partial inflorescences. Spikelets unisexual, the plants monoecious; pistiUodes and staminodes sometimes present in male or female spikelets. Female spikelets 1-flowered, rachilla extension lacking; glumes 2; lemmas 79 herbaceous or coriaceous, several-nerved, blunt or with an awn (ui Buergersiochloa) at the apex; palea 3 to several-nerved, entire at apex; lodicules 3; stigmas 2. Male spikelets 1- flowered, small; glumes lacking, or 2 and minute; lemmas membranous, 3-nerved; lodicules 3; stamens 3, rarely 2 or many, filaments free or fused. Fruit a caryopsis, with a thin and hard pericarp.

ACKNOWLEDGMENTS

Financial support for this study was derived from National Science Foundation Grant

DEB-9218657. We thank R. Olmstead (then of the University of Colorado) for providing most of the initial primers. We owe gratitude to N. Barker (University of Cape Town, South

Africa), G. Davidse (Missouri Botanical Garden), S. Dransfield and S. Renvoize (Kew

Botanical Garden), E. Kellogg (Harvard University), and the Botany Department of

Smithsonian Institution for kindly sending us the leaf materials listed in Table I.

LITERATURE CITED

BLAKE, S. T. 1946. Two new grasses from New Guinea. Blumea, Supplement 3: 59-62.

BARKER, N. P., H. P. LINDER, and E. H. HARLEY. 1995. Polyphyly of Arundinoideae

(Poaceae): Evidence from rbcL sequence data. Systematic Botany 20:423-435.

BREMER, K. 1988. The limits of amino acid sequence data in angiosperm phylogentic

reconstruction. Evolution 42: 795-803.

BUTZIN, F. 1965. Neue Untersuchungen uber die Bliite der Gramineae. Pp. 1-183.

Inaugural-Dissertation, Freien Universitat, Berlin. CALDERON, C. E. and T. R. SODERSTROM. 1973. Morphological and anatomical

considerations of the grass family Bambusoideae based on the new genus Maclurolyra.

Smithsonian Contributions to Botany II: 1-54.

. 1980. The genera ofBambusoideae(Poaceae)ofthe American continent: keys and

comments. Smithsonian Contributions to Botany 44: 1-27.

CHAO, C., C. CHU, and W. XIONG. 1980. A revision of some genera and species of

Chinese bamboos. Acta Phytotaxonomica Sinica. 18:20-36.

CHASE, M. W. and H. H. EHLLS. 1991. Silica gel: an ideal material for field preservation

of leaf samples for DNA studies. Taxon 40: 215-220.

CLARK, L. G. 1989. Systematics of Chusquea sect. Swallenochloa. sect. Verticillatae. sect.

Serpentes. and sect. Longifoliae (Poaceae: Bambusoideae). Systematic Botany

Monographs 27: 1-127.

. 1990. Chusquea sect. Longiprophyllae fPoaceae: Bambusoideae): A new Andean

section and new species. Systematic Botany 15: 617-634.

. 1995. Diversity and distribution of the Andean woody bamboos (Poaceae:

Bambusoideae). Pp. 501-512 in Biodiversity and Conservation of Neotropical

Montane Forests . eds. S. Churchill et al. The New York Botanical Garden.

, and E. J. JUDZIEWICZ. In press. The grass subfamilies Anomochlooideae and

Pharoideae (Poaceae). Taxon.

, and R. W. POHL. 1996. Agnes Chase's first book of grasses. Washington D. C.:

Smithsonian Institution Press.

, W, ZHANG, and J. F. WENDEL. 1995. A phylogeny of the grass family

(Poaceae) based on ndhF sequence data. Systematic Botany 20:436-460.

CLAYTON, W. D. 1962. Studies in the Gramineae: n. Kew Bulletin 16: 247-250.

, and S. A. RENVOIZE. 1986. Genera graminum. London: Her Majesty's Stationery

Office. 81

CLEGG, M. T., A. D. H. GROWN, and P. R. WHTTFELD. 1984. Chloroplast DNA

diversity in wild and cultivated barley: implications for genetic conservation. Genetic

Research 43: 339-343.

CUMMINGS, M. P., L. M. KING, and E. A. BCELLOGG. 1994. Slipped-strand mispairing

in a plastid gene: rpoC2 in grasses (Poaceae). Molecular and Biological Evolution 11:

1-8.

DAVDDSE, G. and R. W. POHL. 1992. New taxa and nomenclatural combinations of

Mesoamerican grasses (Poaceae). Novon 2: 81-110.

, and L. G. CLARK. 1996. Two new species of Neurolepis (Poaceae: Bambuseae)

from Colombia. Novon 6: 150-156.

DAVIS, J. I and R. J. SORENG. 1993. Phylogenetic structure in the grass family (Poaceae)

as uiferred from chloroplast DNA restriction site variation. American Journal of

Botany 80: 1444-1454.

DING, Y., and W. LIESE. In press. Anatomical investigations on the nodes of bamboos.

Pp. xx-xx in The Bamboos, ed. G. Chapman. London: Academic Press.

DRANSFIELD, S., and E. A. WIDJAJA. 1995. Plant resources of South-East Asia No. 7

Bamboos. Leiden: Backhuys Publishers.

DUVALL, M. R. and B. R. MORTON. 1996. Molecular phylogenetics of Poaceae: an

expanded analysis of rbcL sequence data. Molecular Phylogenetics and Evolution 5:

352-358.

ELLIS, R. P. 1976. A procedure for standardizing comparative leaf anatomy in the Poaceae I.

The leaf blade as viewed in transverse section. Bothalia 12: 65-109.

FUTEN, T. 1975. A taxonomic revision of Buergersiochloa Pilg. (GramineaeV Blumea22:

415-418.

FRANCHET, M. A. 1887. Genera nova Gramineamm Africae tropicae occidentahs. Bulletin

Mensuel de la Societe Linneenne de Paris 1:673-677. 82

. 1889. Observation sur le genre Guaduella Franch. Journal de Botanique Paris 3:

305-306.

GAUT, B. S., S. V. MUSE, W. D. CLARK, and M. T. CLEGG. 1992. Relative rates of

nucleotide at rbcL locus of monocotyledonous plants. Journal of Molecular Evolution

35: 292-303.

GUALA, F. G. 1995. A cladistic analysis and revision of the genus Apoclada (Poaceae:

Bambusoideae: Bambusodae). Systematic Botany 20: 207-223.

HILU, K. W., and J. L. JOHNSON. 1991. Chloroplast DNA reassociation and grass

phylogeny. Plant Systematics and Evolution 176: 21-31.

HOLLOWELL, V. In press. Systematic relationships of Pariana and associated neotropical

taxa (Poaceae: Bambusoideae: Olyreae). Pp. xx-xx in The Bamboos, ed. G. Chapman.

London: Academic Press.

HOLTTUM, R. E. 1956. The classification of bamboos. Phytomorphology 6:73-90.

JUDZIEWICZ, E. J. 1987. Taxonomy and morphology of the tribe Phareae (Poaceae:

Bambusoideae) (PhD. thesis). Pp 19-199.

. 1990. Poaceae (Gramineae). in Rora of the Guianas. eds. A. R. A. Gorts-van Elijn.

Germany: Koeltz Scientific Books.

and T. R. SODERSTROM. 1989. Morphological, anatomical, and taxonomic studies

in Anomochloa and Streptochaeta (Poaceae: Bambusoideae). Smithsonian

Contributions to Botany 68: 1- 52.

KALTENBOECK, E., J.W. SPATAFORA, X. ZHANG, K.G. KOUSALAS, M.

BLACKWELL, and J. STORZ. 1992. Efficient production of single-standed DNA as

long as 2 kb for sequencing of PCR-amplified DNA. Bio Techniques 12: 164-171.

KELLOGG, E. A. and C. S. CAMPBELL. 1987. Phylogenetic analyses of the Gramineae.

Pp. 310 -322 in Grass systematics and evoludon. eds. T. R. Soderstrom, K. W. Hilu, 83

C. S. Campbell, and M. E. Barkworth. Washington, D. C.: Smithsonian Institution

Press.

and L. WATSON. 1993. Phylogenetic studies of a large data set. 1. Bambusoideae,

Andropogonodae, and Pooideae (Gramineae). Botanical Review 59: 273-343.

KENG, P. C. 1982. A revision of the genera of bamboos from the world. Journal of

Bamboo Research 1: 1-19.

. 1986. A preliminary study of the inflorescence type arising firom bamboos and its

variations. Journal of Wuhan Botanical Research 4:323-336.

. 1987. A systematic key to the tribes and genera of subfam. Bambusoideae

(Gramineae) occurrence in China and its neighborhoods. Journal of Bamboo Research

6: 13 - 28.

KENG, Y.L. 1959. Flora Dlustralis Plantarum Primarum Sinicarum. Gramineae. Peking.

KOECHLIN, J. 1962. Flore Du Gabon. 'Familie des Graminees' no. 5 : Pp.199 - 211.

Paris.

KUMAR, S., KOICHIR, T., and M. NET. 1993. "MEGA: Molecular Evolutionary

Genetics Analysis", Version 1.0, Pennsylvania State University, University Park.

LIANG, H, and K. W. fflLU. 1996. Application of the matK gene sequences to grass

systematics. Canadian Journal of Botany 74: 125-134.

MADDISON, W. P. and D. R. MADDISON. 1993. MacCIade: Analysis of Phvlogenv and

Character Evolution. Version 3.04. Sunderland, Massachusetts: Sinauer Associates,

Inc.

MCCLURE, F. A. 1934. The inflorescence in Schizostachyum. Journal of the Washington

Academy of Science 12: 541-548.

. 1966. The Bamboos: a fresh perspective. Boston: Harvard University Press.

. 1973. Genera of bamboos native to die New World (Gramineae: Bambusoideae).

Smithsonian Contributions to Botany 9: 1-148. 84

METCALFE, C. R. 1956. Some thoughts on the structure of bamboo leaves. Botanical

Magazine, Tokyo 9:148.

. 1960. Anatomy of the Monocotyledons. I. Gramineae. Oxford: Clarendon Press.

NADOT, S., R. BATON, and B. LEJEUME. 1994. The chloroplast gene rps4 as a tool for

the study of Poaceae phylogeny. Plant Systematics and Evolution 191: 27-38.

NEES VON ESENBECK, C. G. D. 1835. Bambuseae Brasilienses: Recensuit, et Alias in

India Orientali Provenientes Adjecit. Linnaea 9:461-494.

OLMSTEAD, R. G. and P. A. REEVES. 1995. Evidence for the polyphyly of

Scrophulariaceae based on chloroplast rbcL and ndhF sequences. Annals of the

Missouri Botanical Garden 82 : 176 - 193.

and J. A. SWEERE. 1994. Combining data in phylogenetic systematics: an empirical

approach using three molecular data sets in the Solanaceae. Systematic Biology 43:

467-481.

PALMER, P. G. and A. E. TUCKER. 1981. A scanning electron microscope survey of the

epidermis of East African grasses, I. Smithsonian Contributions to Botany 53: 1-72.

PATERSON, A. H., C. L. BRUBAKER, and J. F. WENDEL. 1993. A rapid method for

extraction of cotton (Gossypium spp.) genomic DNA suitable for RFLP and PCR

analysis. Plant Molecular Biology Reporter 11: 122-127.

PILGER, R. 1915. Neue und weniger bekannte Gramineen aus Papuasien. Botanische

Jahrbucher 167-169.

POHL, R. 1965. Dissecting equipment and materials for the study of minute plant structures.

Elhodora 67: 95-96.

. 1991. Blooming history of the Costa Rican bamboos. Revista de Biologia Tropical

39: 111-124.

PRAT, H. 1960. Vers une classification nahareUe des graminees. Bulletin de la Societe

Botanique de France 107: 32-79. 85

RENVOIZE, S. A. 1985. A survey of leaf-blade anatomy in grasses. V. The bamboo allies.

Kew Bulletin 40: 509-535.

ROSHEVrrS, R. YU. 1937. Zlaki: Vvedenie v Izuchenie Kormovykh Zlakov. [Grasses:

An introduction to the study of fodder and cereal grasses.] Moscow, Leningrad:

Sel'khozgiz, Gosudarstvennoe Izdatel'stvo Kolkhoznoi I Sovkhoznoi Literatury.

Agricultural Publishing House. [New Delhi: English translation published for the

Smithsonian Institution and the National Science Foundation, Washington, D. C.,

1980. NTISNo.: TTH 72-51033.]

. 1946. Sistema zlakov v svyazi s ikh evolyutsiei. Leningrad: Komarav Botanical

Institute, U. S. S. R. Academy of Sciences. [Evolu9ao e sistematica das gramfneas.

Boietimdo Instituto de Botanica No. 5: 1-20. Portuguese translation by T. Sendulsky,

Instituto de Botanica, Sao Paulo, Brazil, 1969.]

SASS, J. E. 1958. Botanical microtechnique. Iowa State University Press, Ames, Iowa.

SODERSTROM, T. R. 198 la. Sucrea (Poaceae: Bambusoideae), a new genus from Brazil.

Brittonia33: 198-210.

. 1981b. Some evolutionary trends in the Bambusoideae (Poaceae). Annals of the

Missouri Botanical Garden 68: 614-616.

and R. ELLIS. 1987. The position of bamboo genera and allies in a system of grass

classification. Pp. 225-238 in Grass svstematics and evolution, eds. T. R.

Soderstrom, K. W. Hilu, C. S. Campbell, and M. E. Barkworth. Washington, D. C.:

Smithsonian Institution Press.

, R. P. ELLIS, and E. J. JUDZIEWICZ. 1987. The Phareae and Streptogyneae

(Poaceae) of Sri Lanka: A morphological-anatomical study. Smithsonian

Contributions to Botany 65: 1 -27.

and E. J. JUDZIEWICZ. 1987. Systematics of the amphi-Atlantic bambusoid genus

Streptogvna (Poaceae). Annals of the Missouri Botanical Garden 74: 871-888. 86

and F. O. ZULOAGA. 1989. A revision of the genus Olyraand the new segregate

genus Parodiolyra (Poaceae: Bambusoideae: Olyreae). Smithsonian Contributions to

Botany 69: 1-78.

SOLTIS, D. E., P. S. SOLTIS, and B. G. MILLIGAN. 1992. Intraspecific chloroplast

DNA variation: systematic and phylogenetic implications. Pp. 117-145 in Molecular

Svstematics of Plants, eds. P. S. Soltis, D. S. Soltis, and J. J. Doyle. Chapman and

Hall, New York.

SORENG, R. T., and J. I. DAVIS. 1995. Phylogenetic structure of the Poaceae, results from

morphology, molecules and total evidence. American Journal of Botany (Abstracts)

82:163.

STAPF, O. 1906. Diandrolyra bicolor Stapf in Decades Kewenses, Plantarum Novarum in

Herbario Horti Regii Conservatarum. Decas XLII. Bulletin of Miscellaneous

Information (Royal Botanic Garden, Kew, England) 1906: 204-205.

STAPLETON, C. In press. Some aspects of morphology. Proceedings of an international

synposium on bamboos. Pp. in The Bamboos, ed. G. Chapman. London: Academic

Press.

SUGIURA, M. 1989. Orvza sativa chloroplast DNA 134.525 bp. Nagoya, Japan: Nagoya

University, Center for Gene Research.

SWOFFORD, D. L. 1993. PAUP: Phylogenetic analvsis using parsimony, version 3.11.

Champaign: Illinois Namral History Survey.

TZVELEV, N. N. 1989. The system of grasses (Poaceae) and their evolution. The Botanical

Review 55: 141-203.

WANG, Z., and G. YE. 1980. On the problems of the classification of Chinese bamboos

with creeping rhizomes. Acta Phytotaxonomica Sinica 18:283-291.

. 1988. The separation of Neomicrocalamus from Racemobambos and their systematic

position in Bambusoideae. Acta Botanica Yunnanicarum 10:212-214. 87

WATSON, L. and M. J. DALLWITZ. 1992. The grass genera of the world. Walllngford,

Oxon, England: CAB International.

WU, C. Y. 1958. The anatomical study of bamboo leaves. Quarter Journal of Taiwan

Museum 11: 349 - 370.

ZHANG, W. 1992. The classification of Bambuseae (Poaceae) in China. Journal of

American Bamboo Society 9:25-42. 88

CEEAPTER 3. BAMBOOZLED AGAIN!: INADVERTENT ISOLATION OF FUNGAL rDNA SEQUENCES FROM BAMBOOS (POACEAE: BAMBUSOIDEAE)

A paper accepted for publication in the Journal of Molecular and Phylogenetic Evolution

Weiping Zhang, Jonathan F. Wendel, Lynn G. Clark

ABSTRACT

PCR amplification of the nuclear internal transcribed spacer (US) and 5.8S regions of rDNA from woody bamboos (Bambuseae) led to the recovery of fungal instead of bamboo sequences under a variety of PCR conditions and irrespective of whether the plant DNA was extracted from fresh leaves or silica gel-dried material. Phylogenetic analyses based on the

5.8S sequences indicated that the fungi were basidiomycetes and that none was an ascomycete.

A diverse assemblage of basidiomycetes was isolated from different bamboos and various fungi coexisted in the same host plant. There was no evidence that closely related fungi associate with closely related host bamboos. Phylogenetic analysis based on 5.8S sequences showed that some fiingi were in lineages near Volvariella, Lentinula, Peniophora and

Rhizoctonia, but the insufficiency of basidiomycete ITS sequences in sequence data bases precluded more precise fungal identifications. Bamboo ITS regions were amplified only when fresh leaves were surface sterilized before DNA extraction, suggesting that the fungal associates are epiphyllous rather than endophytic. This study highlights the possibility of inadvertent PCR amplification of contaminating DNAs in molecular phylogenetic studies, particularly when using "universal" amplification primers. 89

INTRODUCTION

The ability to amplify specific sequences from small quantities of genomic DNA has had a profound and positive impact on systematic and evolutionary biology. In some cases, however, PGR amplification experiments have resulted in the recovery of non-specific amplification products or target sequences from contaminating DNAs (Sarkar and Sommer,

1990; Bobola etal., 1992; Olmstead and Palmer, 1994; Smith and Klein, 1994, 1996; Blaney,

1995; Liston etal., 1995; Liston and Alvarez-Buylla, 1995). These examples highlight the familiar and expensive errors (Roberts, 1991) that may result from non-specific primer binding, DNA template contamination, and many other causes. The sensitivity of PGR to such factors as annealing temperamre, reaction component concentrations, and cycling parameters increases the likelihood of recovery of unwanted products (Olmstead and Palmer, 1994; Rao,

1994; Roux, 1995). Optimizing PGR for particular targets generally entails manipulation of

PGR conditions and cycling parameters, and use of specific primers and purified DNA templates. Employing various additives and treatment with ultraviolet light also can improve the fidelity and quality of amplification (Sarkar and Sonmier, 1990; Frothingham et al., 1992;

Lu and Negre, 1993; Gayouette etal, 1996). PGR optimization, however, no matter how carefully conducted, may still lead to "PGR nightmares", particularly if contaminating DNAs are present and if amplification primers have been designed from highly conserved sequences.

The purpose of this note is to relate an example of this phenomenon, from our ongoing studies of bamboos, that stemmed from our ignorance of their close association with a variety of different fungi.

The impetus for molecular phylogenetic work in the Bambuseae was that determination of evolutionary relationships has been problematic due to extensive morphological reduction, especially in floral morphology, and their unusual life cycles (Glayton and Renvoize, 1986;

Soderstrom and Ellis, 1987; Kellogg and Watson, 1993). Gonflicting classifications exist because of different emphases on the importance of various characters. In early studies, the 90 woody bamboos were divided into two or thiree groups based on the number of stamens and stigmas (Monro, 1868; Bentham and Hooker, 1883), which resulted in the artificial grouping of rather distandy related taxa. Stress on the different types of inflorescence development led to the recognition of another two major groups in the woody bamboos (Keng, 1959; Keng,

1987; Zhang, 1992). Other classifications were based on ovary morphology (Holttum, 1956), leaf and embryo anatomical characters (Soderstrom and Ellis, 1987), or a combination of these characters with inflorescence development patterns (Clayton and Renvoize, 1986).

Recently, molecular data have been used in an attempt to reconstruct the phylogeny of the Bambuseae (Clark et al, 1995; Zhang and Clark, unpubl. data). Sequence data from the chloroplast-encoded gene ndhF were useful for resolving relationships within the grass family and the subfamily Bambusoideae, but the gene evolves too slowly for use within the

Bambuseae. At this level, we thought that sequences from the internal transcribed spacers

(ITS) of nuclear rDNA might prove phylogenetically informative, as they have in many other plant groups (e.g., Kim and Jansen, 1994; Baldwin et al., 1995; Downie and Katz-Downie,

1996). In addition, ITS sequence data have been used for phylogenetic purposes in grasses

(Hsiao et al., 1993, 1994, 1995).

As a pilot project, we attempted to sequence ITS directly from PCR products from eight species of bamboos, but had difficulty due to excessive polymorphism. Amplification products were subsequently cloned and sequenced, again disclosing sequence polymorphism among clones, but in addition revealing a surprising amount of length heterogeneity. Hsiao et al. (1993, 1994, 1995) indicated that the lengths of the ITS regions in grasses were quite stable, and that polymorphism among repeats was not a major concern. Nonetheless, woody bamboos, like many other grasses, are polyploid (Soderstrom, 1981; Hunziker and Stebbins,

1987), and the results of Hsiao et al. notwithstanding, the possibility remained that our observations reflected "weak" concerted evolution and hence high polymorphism among repeats within and/or between rDNA loci. Nucleotide and length variation levels in our ITS 91 sequences, however, were greater than reported for other grasses (Hsiao et al., 1993, 1994,

1995), and were so high, in fact, that alignment was not possible except within the conserved

5.8S gene. These results led us to suspect that we had recovered sequences from organisms other than bamboos, which we report on here.

MATERIALS AND METHODS

The eight species of woody bamboos (Bambuseae) used in tliis study are listed in Table

1, along with their vouchers, clone numbers, and GenBank numbers; database accession numbers for other sequences used in phylogenetic analyses are also listed. Voucher specimens are deposited in the herbaria of either the Sichuan Forestry School (SIFS), Dujiangyan, China or the Ada Hayden Herbarium (ISC) at Iowa State University.

Total cellular DNA initially was extracted from silica gel-dried or fresh apparendy healthy bamboo leaves using a modified CTAB method (Paterson et al., 1993) without any additional treatment of the leaf material. A subsequent set of extractions was performed on fresh, surface sterilized bamboo leaves foUowmg a procedure modified from Schultz et al.

(1993). Leaves were swirled in 95 % EtOH for 1 min, 5% bleach (NaOCl) for 5 min, and then reunmersed in 95 % EtOH for 30 sec, after which they were blotted dry and then immediately extracted.

Primers ITS4 and ITS5 from White et al. (1990) were initially used for PCR amplification, with die later use of other combmations of available primers and three designed during this study (Table 2). Combinations used for double-stranded amplifications were: ITS4 with ITS5new; C26A with ITS5new; ITSL with ITS4; ITSL with C26A; N18L18 with C26A; rrsw witii rrSP; rrsw witii rrSR; and ITS5 with rrSR; and rrs IF with ITS4B. The 92

Table 1. Taxa, vouchers, and GenBank accession numbers for species studied.

Taxon Voucher^ Clone GenBank# Number Fungi Ascomycota Acremonium uncinatum L20305 Candida albicans X71088 Epichloe typhina L07133 Phialophora americana U31837 Thermomyces lanuginosus M10392 Tricoderma longibrachiatwn L07957 Verticillium tricorpus L28679 Basidiomycota Cronartium flaccidum X78257 Filobasidiella neoformans L14068 Heterobasidion annosum X70022 Hyphodontia aspera Lentinula lateritia U33070 Peniophora nuda Peridermium pini X83913 Puccinia sorghi L08734 Rhizoctonia solani U19950 Uromyces scillarum L08733 Volvariella volvacea U15973 Arundinaria gigantea WZ8400703 1 U65616 2 U65617 3 U65618 4 U65619 5 U65661 Pseudosasa japonica WZ8400708 1 U65662 Bashaniafargesii WZ9201 1 U65602 2 U65599 3 U65615 93

Table 1 (continued) 5 U656I4 6 U65623 7 U65624 Sasa variegata 1 U65598 2 U65603 5 U65601 Yushania exilis WZ9230 1 U65620 2 U65600 4 U6562I Phyllostachys pubescens LCI289 I U65608 2 U65609 3 U65606 4 U65605 5 U65607 Chimonobambusa marmorea SBG 9203 I U65610 2 U65611 3 U65612 5 U65613 Shibataea kumasaca LC1290 I U65622 3 U65604 Plantae Vicia faba V0142I Oryza sativa iVI35384 Triticum vulgare M10469 Armdinaria gigantea WZ8400703 U65663 Phyllostachys pubescens LCI 289 U65665 Pseudosasa japonica WZ8400708 U65664 Shibataea kumasaca LC1290 U65666 a: collector prefixes are as follows: LC = L. Clark; SBG = Sichuan Academy of Forestry

Botanical Garden; WZ = W. Zhang. 94 internal primers ITS2, ITS3, N5.8S , and C5.8S were used for single-stranded and double- stranded DNA sequencing to obtain sequences from both strands. A variety of different PGR conditions were attempted in an effort to obtain angiosperm sequence. Our basic protocol was to use 50 Hi reactions containing 1.5 mM MgCl2^ 0.4 pM of each primer, 200 pM of each dNTP, 100 ng of template DNA and 2.5 units of Taq polymerase. For double-stranded amplifications, 35 cycles were used, as follows: annealing for 2 minutes at 45°; extension for

2 minutes at 72°; and denamration for I minute at 94°. We also tried the protocol of Hsiao et al. (1993), which differed only by the inclusion of 10% DMSO (dimethyl sulfoxide), and a number of other methods wherein we varied: primer choice (Table 2) and combination; concentration of dNTPs and MgCl2- annealing temperature (incremented from 45° to 55°); glycerol concentration (0, lor 2%); and PGR cycle number (20 - 35).

PGR amplification reactions with multiple bands were electrophoresed in 1% low melting agarose (Nusieve, FMG), visuaUzed with ethidium bromide, and excised separately.

Different size fragments of DNA were recovered using GeneGlean (Bio-101, Galifomia), or were used direcdy for in-gel Ugations. Amplification reactions that yielded only a single band were purified using Prep-A-Gene (Biorad). Ligations were conducted overnight using the pGEM-T vector system (Promega). Following electroporation, recombinants were identified and plasmids were isolated using standard procedures.

Initially, five clones were screened and sequenced for each ligation reaction, and if the results indicated that the sequences were variable, an additional five clones were selected for further sequencing. Both strands were sequenced using standard dideoxy sequencing methods with [a-^^S], electrophoresis in Long Ranger (FMG) sequencing gels, and autoradiography.

Sequences were aligned using Glustal W (Thompson et al., 1994) and the resulting alignments were adjusted manually if necessary. For each sequence, the entire ITS region 95

Table 2. Primers (5' to 3') used for PCR amplification of the ITS region.

EMmer Nucleotide Sequence Primer Nucleotide Sequence ITS5 GGAAGTAAAAGTCGTAACAAGG N5.8S ATCGAGTCTTTGAACGCA rrssnew GGAAGGAGAAGTCGTAACAAGG ITS4 TCCTCCGCTTATrGATATCG rrsL TCGTAACAAGGTTTCCGTAGGTG C26A TTTCTTTTCCTCCGGT N18L18 AAGTCGTAACAAGGTTTC ITS4B CAGGAGACTTGTACACGGTCCAG ITS IF CTTGGTCATTTAGAGGAAGTAA rrsp^ AACTCAGCGGGTAGTCCC rrsw^ GTGACCCTGACCAAAACAGA rrsR'^ CCWCCWTGYGCTGTGC ITS2 GCTGCGTTCTTCATCGATGC ITS3 GCATCGATGAAGAACGCAGC ITS2C TGCGTTCAAAGACTCGAT C5.8S TGCGTTCAAAGACTCGAT

'a' indicates the primers designed in this study.

(ITS 1, 5.8S, and ITS2) and its three components were used as query sequences in GenBank and EMBL searches using the FASTA (Pearson and Lipman, 1988) and BLAST (Altschul et al., 1990) programs. 5.85 sequences of three angiosperms, including two grasses, and various fungi were downloaded from GenBank and EMBL, and included in the final data matrix. Maximum parsimony analysis was performed using PAUP version 3.1.1 (Swofford,

1993), with plants as the outgroup in the first round, and three ascomycetes as the outgroup for subsequent phylogenetic analyses of sequences from basidiomycetes. All characters were weighted equally, and only the heuristic option was used in the search for shortest trees because of the size of the matrix. A strict consensus tree was generated from the equally shortest trees for each round of analysis. Decay analysis was used to evaluate support for individual clades (Bremer, 1988).

Leaf material for scarming electron microscopy (SEM) was taken from herbarium specimens. Representative samples approximately 0.5 X 1 cm were excised from the middle third of mature foliage leaves, and were either mounted direcdy or sonicated in xylene for about

10 min to remove epicuticular wax. Specimens were mounted on brass discs with silver paste 96 or double-stick tape, coated with Au-Pd in a Polaron E5100 sputter-coater, then viewed at 15 kV in a JEOL JSM-35 scanning electron microscope. Photographs were taken using Polaroid

Type 665 positive-negative film.

RESULTS

ITS regions were so variable in their lengths (573 to 823 bp) and sequences that it was not possible to align them except for in the 5.8S region. When the ITS 1 and rrS2 components of each sequence were used as separate units in FAST A and BLAST searches of GenBank and

EMBL, only the begirming of the ITS 1 region and the end of the ITS2 region remraed matches, and these were due to overlap with the ribosomal small and large subunit sequences, respectively. Because the spacer sequences were so variable, only the 5.8S sequences were used for phylogeny reconstruction.

Alignment of the 5.8S gene sequences resulted in a data matrix of 160 nucleotide positions by 54 taxa, as shown in Figure 1. Inspection of this matrix reveals a number of base positions where different nucleotides are found in fiingi and plants, and others where the 5.8S genes of Ascomycota and Basidiomycota are readily distinguished. At positions 114 and 124, for example, each of the three major groups exhibits a different base, although two basidiomycetes share the characteristic adenine of ascomycetes at 114, and several basidiomycetes share the characteristic cytosine of ascomycetes at 124. For the most part, ascomycetes, except for Candida albicans, are distinguished from other taxa by nucleotides at positions 1, 33, and 153, and basidiomycetes are uniquely defined by the possession of a diymine at position 157. In addition to character-states at these latter four positions, these two major groups of fungi often display different nucleotides at positions 29, 121, 135, and 139.

Plants possess nucleotides that distinguish them from the fiingi at a total of thirteen positions

(marked by asterisks in Fig. I), which is 8.1% of all characters. Nucleotide substitutions at 20 30 40 SO 60 70 30 Thexmomyces la&uginosus ???AAACTTTCAACAATGGA TCT( \ TGAAGAACGCAGCGAAATGC GATAAGTAATGTGAATTGC Verticilliuxn tricorpxis TTAG T C... Spichloe typhina. TCA ACirSnionXUni nijUntm TCA ..C Trichoderma longibrachiatum TCA phialophora americnna GCA TCA Pilobeisidiella neoformans ATA Hecerobasidion annosimi ATAC Rhizocconia solani AACT..G.. Kyphodontia aspera ATAC Peniophora niida ATAC Puccinia sorghi AAAT .T... A r .TCA A T T Uromyces scill2u:\zm ATAT .T A r .TCA A T T Peridexmitim pini AAAT .T... r A.. -T T Cronarcium flaccidum AAAT .T... r A...T .... T VoXvariella volvacea ATAC Lentinula lateritia ATAC ...G. . .c r Arundinsiria gigantea I ATAC ..c r .TC Arundinaria gigantea 2 ATAC ..c r .TC Anmdinaria gigantea 3 ATAC . .c r .TC Arundixiaria gigantea 4 ATAC ..c r .TC Anzndinaria gigantea 5 ATAC . .c r .TC C... Bashania fargesii I ATAC ...G. ..c r .TC Bashania fargesii 2 ara ...G. . .c r .TC Bashania fargesii 3 ATAC .c r ,TC . ...G Basiumia feurgesii 5 ATAC .c r .TC Bashania fargesii 6 ATAC G. .c r Bashania fargesii 7 ATAC .TG.. .c .TC Chinonobanbtisa aannorea 1 ACA .c .TC ChimoaobamhiAsa aarmorea 2 ATAC .TG.. .c ... Chimonobamtnisa imunnorea 3 ATAC .TG.. .c .TCC Chifflonobanbusa maxioorea 5 ATA ...G. .c r .TC T. Phylloscachys ptibescens 1 ATAC .c r .TC Phyllostachys ptibescens 2 ATAC .c r .TC PhyXXostachys pubescens 3 ATAC .c r .TC PhyXXostachys pubescens 4 TTAC..T.. r .TC PhyXXostachys ptibescens 5 ATAC ...G. .c r .TC Pseudosasa japonica 1 ATAC .c r .TC Sasa variegata 1 ATA ...G. .c r .TC Sasa variegata 2 ATA .c r .TC T. Sasa veiriegata 5 ATAC Shibataea kuxnasaca 1 AAAC .c r .TC ..CC.CC Shibataea kisnasaca 3 ATA .c... c.. r .TC T. Yushania exiXis 1 ATA ...G. .c .TC . ...c Yushania exiXis 2 ATA ...G. .c r .TC ....c Yushania exiXis 4 ATA Vicia faba GAATG...C. .GG. .C... .A. .A. .c..TT T CT.GG Triticum vulgare ACACG...C .GG. .C... .A. .C. .c..TC T CC.GG Oryza sativa ACACG...C .GG. .C... .A. .C. .c. -TC T CC.GG Arundinaria gigantea ACACG..-C .GG. .C... .A. .C. .c. -TC T CC.GG Pseudosasa japonica ACACG..-C -GG. .C... .A. .C. .c..TC T CC.GG PhyXXostachys ptibescens ACACG...C .GG. .C... .A. -C. .c..TC T CC.GG Shibataea )ctixnasaca ACACG...C

FIG. 1. Aligned data matrix of 5.8S sequences for representative ascomycetes (A), basidiomycetes (B) and angiospermous plants (P). Although Candida albicans is classified as an ascomycete, we boxed it separately to emphasize that it shares only one of the three character-states that define the other ascomycetes. * denotes nucleotide positions diat distinguish most fimgal and plant sequences; + denotes nucelotide positions that distinguish Basidiomycota and Ascomycota; (denotes nucleotide positions that differ for each of the three groups. 98

90 1^0 no 120 no i40 iso AGAATTCCTGAATCATCGAA TCTTTGAACGCACATTGCGC CCTCTGGTATTCCGG-GGG- G-CATGCCTGTCCGAGCGTC

A—

. .TA. AC .A C— .C. . .T...... A...... C T.CT...... TT-. -A- .. . .TT.. -TA- . .A. . C T.CT...... A-. .A- ... .TT.. -T.. - .CA. .T.. . -A- .G. .CA. ...TT.. -T.. . .-T .CA. ....TT.. .T.. - .-T -CA. ....TT.. -T. . . C T. .T. --Cra.T.C.-AA.- AG. .T...... A...... C .. .A. T.CT...... A-. .A- .-T .. .TT.. -T.. - .A...... C. . . .A. T.CT... . - - -A-. A- .-T...... TT.. .T. - ...... A-. A- .-T...... TT.. -T.. - .A...... C - . .A. T.CT...... A-. .A- .-T...... -TT.. -T. . - -A- -T. . . .-T ...TT.. .T. . - A- .-T . .T .T. . . -T -A- ...TT.. -T.. . C. T. . .A-A-. CA. .. .TT.. .T.. . C. . C. .TT.. -T. . . .A...... C. . .G-. T.CT...... TT-. A- .. .TT.. -A. - - •A- .-T CAT .T.. . .A...... C. T.GT...... A-C A" C-G . . .TT.. -T. - - .. .TT. . -T. - . A- . -T .T.. . .A...... C. . .GC. . .CT. . .C. . .A-. C. . .. .TT. . -T. - - .A...... C. . -GC. . .CT. . -C. . .A-. c.. . . .TT.. .T.. . .A...... C. . .GC. . .CT...... -A-. C- . . .TT.. .T- - . ..-T.. .A...... C. . .GC. T.CT...... A-. A- ...TT.. -T. - . .A...... C. . . .A. T.CT...... A-. A' .-T ...TT.. .TC. . .T .A. . C. . . .A. T.CT...... A-. A- .-T ...TT.. .T. . . A' .-T ..T .T T A- TTC. .T A- ...... TT.. .T. . - C- A—G...... TT.. -T.. . T A- TTC. .T.. . c. . . .A. GA. . . - - -ATC- .-T .T.. . A- .T.. . GA...... -ATC- , -T T.. .T. - . C...... C.. G AG. . .GA. cc. A. .TT.AG .G. C.T. .-C.T.G .T-. - C cc. -c-TC.CC. AG G C, . .-C.T.G c cc..c- . . .CC AG G c. .-C.T.G c cccc.- . . .CC. AG ,G. r ..C.T.G c ccc .c- . . .CC. AG .G. c ..C.T.G c ccc -C-. . .CC AG .G. c --C.T.G c ccc.c- . . .CC. AG .G. r --C.T.G i- 4 99 positions 88 and 93 may also be usefiil in distinguishing plants from fungi but are not unique.

Phyiogenetic analysis of 42 aligned 5.8S sequences was conducted treating the four observed indels (positions 136,140, 142, and 148; Fig. 1) as missing data. A strict consensus of the 210 equally most parsimonious (L = 183) trees, rooted with plants, is shown in Fig. 2.

Each tree had a consistency index (CI) of 0.50 and a retention index (RI) of 0.80, indicating that most of the character-state change observed in the data matrix is actual synapomorphy. In the strict consensus tree, sequences from all grasses, including the four woody bamboos whose sequences were obtained after surface sterilization of fresh leaves, formed a monophyletic clade well separated from the fungal clade, while all sequences obtained from bamboos without surface sterilization grouped with fungi to form another monophyletic clade.

Within the fungal clade, a trichotomy was observed, consisting of (1) an ascomycete clade with

Candida albicans in the basal position; (2) a clade of basidiomycetes including all sequences from untreated bamboo leaves except Yushania exilis 4; and (3) a clade consisting solely of the sequence from this latter taxon. We note that although the Candida yeasts are classified as ascomycetes (Alexopoulos etaL, 1996) and associate closely with other ascomycetous yeasts in an analysis of 18S rRNA (Bams et ai, 1991), Candida albicans does not share the ascomycete synapomorphies identified in this data set.

Decay analysis showed that the whole fungal clade, as weU as the ascomycete clade, survived in strict consensus trees a minimum of two steps longer. Aldiough the basidiomycete clade itself was not well-supported in the decay analysis, the clear separation of the plants and ascomycetes in the tree demonstrates that the sequences from untreated bamboo leaves are not of plant or ascomycetous origin, but instead are related to basidiomycetes.

Based on the results from this initial phyiogenetic analysis, a second analysis was conducted wherein we included three ascomycetes and all available sequences from basidiomycetes in an effort to fiirther identify the sequences from untreated bamboo leaves.

The aligned matrix had 43 rows and 160 columns (two single-bp indels). Parsimony analysis FIG. 2. Strict consensus of the 15 shortest trees found in parsimony analysis of angiosperm, fungal, and "contaminating fungal" 5.8S sequences, rooted with angiosperms. Sequences generated in this study are indicated with shaded boxes; arable numerals following species names denote individual clones. Numbers above branches denote decay indices: 0 - clade exists in the most parsimonious trees only; 1 - clade survives in a strict consensus of trees one step longer; 2 - clade survives in the strict consensus of trees two or more steps longer. Thermomyces lanuginosus Verticillium tricorpus Epichloe typhina Acremonium uncinatum Ascomycota Trichoderma longibrachiatum Phialophora americana Candida albicans Heterobasidion annosum Rhizoctonia solani

GRmondbanbiisccmSmof^

Puccinia sorghi Uromyces scillanun Volvariella volvacea Sa^vdhegaUi^Iri BasKSS^fSfxeaiiZi

SSsa^Srie^Z:?, SISBataeSBimasaceii' Basidiomycota pmiosmfv^mBSSm^r

CtumonoBamBusSmarmbr^S

BasftmSTfiirgesiSSl FseiidosaMraodriZaT Lentinula lateritia BMftgiua'f^ge^4 Vicia faba Oryza sativa Triticum vulgare Plantae Ariu^SSna'^gcntea: PseudSsdM^fapdmcal Pft^UosiacW^uSScens^ 102 of the 11 known basidiomycetes and the 29 sequences we produced, plus three ascomycetes as the outgroup, generated 414 equally shortest trees of 136 steps, each with a CI of 0.49 and a

RI of 0.75. The strict consensus tree (Fig. 3) indicates that while the basidiomycetes are supported as monophyletic, the data, with one exception, do not provide good resolution. The exception is the rusts (Puccinia sorghi, Uromyces scillarum, Peridermium pini, and

Cronartium flaccidum), which form a well-supported clade. Although the position of Yushania exilis 4 is ambiguous in Fig. 2, it does not share any of the characters diagnostic of ascomycetes and does possess the thymine at position 157 that is characteristic of basidiomycetes. Not surprisingly, therefore, in the analysis in which only fungal taxa are included, Yushania exilis 4 appears embedded within the basidiomycete clade (Fig. 3).

We surmised that additional resolution within the basidiomycete clade might emerge from analyses of die spacer sequences (ITS 1 and ITS2), but these regions were highly variable and could not be aligned reliably even among the sequences we generated from the bamboo hosts. This fact, and the paucity of available 5.8S DNA sequences from basidiomycetes in sequence data bases precluded a more precise identification of the plant-associated fungi.

DISCUSSION

All 5.8S sequences isolated from bamboo leaves that were not surface sterilized prior to

DNA extraction were identified as having been derived from basidiomycetes (Figs. 1 - 3).

These sequences were obtained following PGR amplification using many different PGR protocols, a variety of different amplification primers (Table 2), and irrespective of the use of

PGR additives such as DMSO and glycerol. No adjustment to the PGR methology resulted in the amplification of angiosperm rather than fungal ITS sequences. Because the same primers amplify true bamboo ITS regions after fiingi are removed from the leaves by surface sterilization, we concluded that plant DNA preparations contaminated with fungal DNA caused the ampUfication problems rather than some feature of primer design or reaction conditions. FIG. 3. Strict consensus of 414 equally most-parsimonious trees found in parsimony analysis of 5.8S sequences from basiomycetes and putative basidiomycetous fiingi associated with bamboos, rooted with three representative ascomycetes. Sequences isolated from bamboos are indicated by plant species names in shaded boxes; arabic numerals denote individual clones.

Numbers above branches denote decay indices: 0 - clade exists in the most parsimonious trees only; 1 - clade survives in a strict consensus of trees one step longer; 2 - clade survives in the strict consensus of trees two or more steps longer. Filobasidiella neoformans Heterobasidion annosum Rhizoctonia solani

Hyphodentia aspera Peniophora nuda EhvUtfstacttwpubescens^Ie

Puccinia sorghi Uromyces scillarum Peridermium pini Cronartium flaccidum Volvariella volvacea Lentinula lateritia ScSSydnegdia^Il

YusHmua&lS2T SttsaiyimegateeSi' • m !•' 1mm '• .''/mn i i y,-: Basiumui^nestiu • Basidiomycota

PsewSMsSidpomcd 1'

5ga>atee

Chmionbbimbusa^ CWnSmtidmbuWiimnm

CKmumbtxiMmajtidrmbrmS Bashemm^^WS}

^Sx^^^îgêdS YusJîamUs:Ij- Y^funiaj^S^ BasfîS^rgai£6 Epichloe typhina Acremonium uncinatum Ascotnvcota Phialophora americana 105

The rrS primers that have become popular for phylogenetic purposes in plants

(Baldwin et al., 1995) were originally designed for use in fiingi (White et al., 1990). Thus, it

is not surprising that inadvertant isolation of fiingi from plants has previously been reported

(Bobola et al., 1992; Ritland and Straus, 1993; Ritland et al., 1993; Smith and Klein, 1994,

1996). To ameliorate this problem, new putatively angiosperm-specific primers have been designed, e.g., ITSSnew (Olmstead laboratory), N18L18, C26A, C5.8S and N5.8S (Zimmer laboratory), and ITSL (Hsiao laboratory). These primers are similar to and often overlap extensively with the original fungal primers, from which they typically differ by only several nucleotides. We designed three new primers, ITSW, ITSP and ITSR, which are more specific for grasses and bamboos, in an attempt to amplify true bamboo ITS DNA firom the species in this study. These primers also failed when used on the initial extractions, except for the combination of ITSW and ITSP for Arundinaria gigantea, which did amplify the true bamboo

ITS region. Given the success of surface sterilization, we suggest that this procedure be considered whenever fungal contamination might be a problem and primers are "universal", as they are for the ITS region.

In addition to documenting the presence of contaminating basidiomycetes, the sequence data and phylogenetic results clearly suggest the presence of a diversity of epiphyllous fiingi.

Sequences from some bamboo hosts, such as Arundinaria gigantea (clones 1-4) and

Pseudosasa japonica 1, formed a relatively well-supported clade (Fig. 3), perhaps indicating that a small number of closely related basidiomycete species or varieties are associated with these hosts. Other host bamboos, however, particularly Yushania exilis, Bashaniafargesii,

Sasa variegata, and Phyllostachys pubescens, are associated with fungal sequences scattered throughout the cladogram (Fig. 3). This result suggests that one bamboo species may host a number of different basidiomycete species. Phyllostachys pubescens appears to host at least four different basidiomycetes, whereas B. fargesii hosts six. 106

Although the 5.8S region in the fungal sequences revealed limited variation, the ITS I and ITS2 regions were highly variable (data not shown). Using this information in addition to the phylogenetic results, we infer conservatively that the fiingi isolated from untreated bamboo leaves represent at least 22 different species. Precise identification of these fiingi is not possible given present limitations, but some suggestions of relationships are evident. For example, clone 1 firom the bamboo host Shibataea kumasaca is sister to Rhizoctonia solani, a well known soil fungus that causes root rot and damping off. Two nearly identical sequences from Phyllostachys pubescens (clones 1,2) associated with Peniophora nuda, a wood-rotting fungus. It is unlikely that any of these fiingi normally would occupy the leaves of a host plant, but they could easily be present as contaminants from splashing during watering or from nearby dead culms, and both of these host plants are cultivated in greenhouses. A majority of the fiingal sequences from bamboo hosts, however, did not associate with any of the known basidiomycete sequences. Clearly though, the fiingi associated with any given host bamboo species are, for the most part, not closely related. It is noteworthy that none of our fungal sequences associated with the rust clade, the one group of basidiomycetes that could be expected to occur on bamboo cuhns or in bamboo leaves.

Based on molecular evidence to date, woody bamboos fall into two major clades, the temperate woody bamboos and the tropical woody bamboos (Clark et al., 1995; Kelchner and

Clark, in prep.; Zhang and Clark, in prep.). The eight species sampled in the present study all belong to the temperate clade. Within this clade, Arundinaria, Pseudosasa, Bashania, Sasa, and Yushania are considered to be closely related to each other, while Phyllostachys,

Shibataea, and Chimonobambusa form another group of related genera (Keng, 1957; Keng,

1987; Zhang, 1992). Although most fiingi bom Arundinaria gigantea clustered with the clone

1 from Pseudosasa japonica, a fungus from Phyllostachys pubescens (clone 5) is sister to that assemblage, and in another clade, a fungal species from Bashania fargesii (clone 7) is sister to two fiingi from Chimonobambusa marmorea (clones 2, 3) (Fig. 3). We conclude that there is 107 no evidence in our data that closely related basidiomycetes associate with closely related host bamboos, although this possibility clearly has not been ruled out.

It is well known that there are close associations between plants and fungi; according to some estimates, about 80 % of vascular plants host fungi, and coevolution between plants and fiingi has been suggested (Lyal, 1986; Thompson, 1994; Alexopoulos et al., 1996; Bacon and

Hill, 1996). Grasses, including bamboos, have been reported to host numerous fungi (Hino,

1961; White, 1987; Clay and Leuchtmann, 1989; White and Glenn, 1994; Rollo et al., 1995;

White and Morgan-Jones, 1996; Bacon and Hill, 1996; Stone et al., in prep), and coevolutionary relationships between ascomycetes and grasses have been proposed (Bacon and

Hill, 1996; White and Morgan-Jones, 1996). Hino (1961) indicated that both ascomycetes and basidiomycetes were parasitic on bamboos, with a considerable number of ascomycetes and

Fungi Imperfecti present. Basidiomycetes were relatively few in number (approximately 15 % of fungal diversity on the bamboos examined), and all were found on rotten culms. All endophytic fiingi documented from grasses are ascomycetes or Fungi Imperfecti presumed to be ascomycetous in origin (Clay, 1988, 1990; White and Morgan-Jones, 1996; Stone et al., in prep.).

All leaves we used for DNA extraction were asymptomatic and apparentiy healthy.

Under the scarming electron microscope, however, fungal hyphae can be seen clearly on the adaxial and abaxial leaf surfaces (Fig. 4) of both tropical and temperate bamboos. Examination of the leaf material (taken from herbarium specimens) revealed that at least two types of fiingi were present. The larger, pigmented hyphae (Fig. 4C) most likely belonged to a mildew

(ascomycete), perhaps the result of not drying the specimen quickly enough. A different type of hyphae, smaller and hyaline, was observed on other asymptomatic bamboo leaves (Fig. 4A-

B, D), and could not be positively identified as either ascomycetous or basidiomycetous (L.

Tiffany, pers. comm.). Figure 4D shows an unti-eated bamboo leaf with its epicuticular wax, but the fungal hyphae remained even after sonication in xylene for about 10 minutes (Fig. 4A- 108 109 HO

C). Species of the Balansieae, a group of ascomycetous fiingi endophytic in grasses, can utilize host waxes as energy sources during their epiphytic growth stage (White et ai, 1991;

White and Morgan-Jones, 1996), so epicuticular wax on bamboo leaves could provide a potential source of energy for epiphytic fungi.

The amplification of true bamboo DNA from several hosts after surface sterilization of the leaves, including four hosts from which only fiingai DNA was amplified initially, indicates that fungi exist as epiphytes in the bamboo phylloplane. The Shibataea kwnasaca I and

Phyllostachys pubescens (clones 1,2) fungal sequences may be attributed to contamination from non-epiphyllous sources, but this explanation is unsatisfactory for most other sequences, given that no work on fiingi is carried out in this laboratory, and other work with ITS sequencing has not produced similar results. In addition, the sheer diversity of fungal sequences from the bamboo hosts argues against laboratory contamination, as does the observation that no ascomycetous sequences were derived. There was no macroscopic evidence of fungi on the bamboo leaves used for DNA extraction, but our results show that basidiomycetes are present. This raises the intriguing possibility that at least some of these fungi form a component of a natural bamboo phylloplane mycoflora. We note, however, that the eight species of woody bamboos we used were collected from non-native settings in greenhouses or botanical gardens. Fungi are present on asymptomatic bamboo leaves collected from their native habitats (Fig. 4), suggesting the possiblity that a variety of basidiomycetous fiingi form part of the natural mycoflora of bamboo leaves. Further investigation of this interesting, potentially natural ecological association is warranted.

Amplification from contaminating DNAs is a potential problem wherever different kinds of organisms share a close association, and examples of this are numerous (e.g., Boboia et al., 1992; Smith and Klein, 1994, 1996). The present study, in addition to highlighting the potential for errors in phylogeny reconstruction based on rDNA sequences, also serves as a reminder to interpret unexpected phylogenetic results cautiously. Contamination and Ill counterfeit PGR amplification, for example, might lead to phylogenetic phenotypes that mimic bona fide genetic mechanisms such as horizontal interspecific transfer (Nishida and Sugiyama,

1995).

ACKNOWLEDGEMENTS

We thank A. Liston, T. Harrington, D. Nickrent, C. Hsiao, and M. Hershkovitz for comments and suggestions; J. Stone, G. Bills and J. White for their endophyte fiingi manuscript; L. Tiffany for examination of fungal hyphae on bamboo leaves and for helpful discussion; T. Harrington, C. Hsiao, and E. Zimmer for ITS primers; T. Hsiau and T.

Harrington for use of their unpublished basidiomycete 5.8S sequences; R. Small and R. Cronn for pGEM-T cloning assistance; and T. Seelanan for computer assistance. Scanning electron microscopy was carried out in the Bessey Microscopy Facility of Iowa State University. We also acknowledge the financial support provided by National Science Foundation Grant DEB-

9218657.

LITERATURE CITED

Alexopoulos, C. J., Mims, C. W., and Blackwell, M. (1996). "Introductory mycology,"

John Wiley & Sons, Inc., New York.

Altschul, S. F., Gish, W., Miller, W., Mayers, E. W., and Lipman, D. J. (1990). Basic

local alignment search tool. J. Mol. Biol. 215: 403-410.

Bacon, C. W. and Hill, N. S. 1996. Symptomless grass endophytes: Products of

coevolutionary symbioses and their role in the ecological adaptations of grasses. In

"Endophytic Fungi in Grasses and Woody Plants" (S. C. Redlin and L. M. Cards,

Eds.), pp. 155-178, APS Press, St. Paul, Minnesota. 112

Baldwin, B. G., Campbell, C. S., Porter, J. M., Sanderson, M. J., Wojciechowski, M. F.,

and Donoghue, M. J. (1995). The ITS region of nuclear ribosomal DNA: A valuable

source of evidence on angiosperm phylogeny. Ann. Mo. Bot. Gard. 82: 247-277.

Bams, S. M., Lane, D. J., Sogin, M. L., Bibeau, C., and Weisburg, W. G. (1991).

Evolutionary relationships among pathogenic Candida species and relatives. J. Bact.

173: 2250-2255.

Bentham, G., and Hooker, J. D. (1883). "Genera Plantarum 3(2)," L. Reeve & Company,

London.

Blaney, C. (1995). Fungi in the family tree. Bioscience 45: 746.

Bobola, M. S., Smith, D. E., and BClein, A. S. (1992). Five major ribosomal repeats

represent a large and variable fraction of the total genomic DNA of Picea rubens and P.

mariana. Mol. Biol. Evol. 9: 125-137.

Bremer, K. (1988). The limits of amino acid sequence data in angiosperm phylogenetic

reconstruction. Evolution 42: 795-803.

Cayouette, M., Gackstetter, T., and Moores, J. (1996). The effects of various additives on

experimentally determined PCR annealing temperatures. Strategies 9: 33-35.

Clark, L. G., Zhang, W., and Wendel, J. F. (1995). The phylogeny of the grass family

(Poaceae) based on ndhF sequence data. Syst. Bot. 20: 436-460.

Clay, K. (1988). Fungal endophytes of grasses: A defensive mutualism between plants and

fiingL Ecology €9: 10-16.

Clay, K. (1990). Fungal endophytes of grass. Annu. Rev. Ecol. Syst. 21: 275-295.

Clay, K., and Leuchtmarm, A. (1989). Infection of woodland grasses by fungal endophytes.

Mycologia 81:805-811.

Clayton, W. D., and Renvoize, S. A. (1986). "Genera graminum," Her Majesty's Stationery

Office, London. 113

Downie, S. R., and Katz-Downie, D.S. (1996). A molecular phylogeny of Apiaceae

subfamily Apioideae: Evidence from nuclear ribosomal DNA internal transcribed

spacer sequence. Am. J. Bot. 83: 234-251.

Frothingham, R., Blitchington, R. B., Lee, D. H., Greene, R. C., and Wilson, K. H.

(1992). UV absorption complicates PGR decontamination. BioTechniques 13:20S-

210.

Hino, 1. (1961). "Icones Fungorum Bambusicolorum Japonicorum," The Fuji Bamboo

Garden, Japan.

Holttum, R. E. 1956. The classification of bamboos. Phytomorph. 6: 73-90.

Hsiao, C., Chatterton, N. J., Asay, K. H., and Jensen, K. B. (1993). Phylogenetic

relationships of 10 grass species: An assessment of phylogenetic utility of the internal

transcribed spacer region in nuclear ribosomal DNA in monocots. Genome 37: 112-

120.

Hsiao, G., Chatterton, N. J., Asay, K. H., and Jensen, K. B. (1994). Phylogenetic

relationships of the monogenomic species of the wheat tribe, Triticeae (Poaceae),

inferred from nuclear rDNA (internal transcribed spacer) sequences. Genome 38: 211-

223.

Hsiao, C., Chatterton, N. J., Asay, K. H., and Jensen, K. B. (1995). Molecular phylogeny

of the Pooideae (Poaceae) based on nuclear rDNA (ITS) sequences. Theor. Appl.

Genet. 90: 389-398.

Hunziker, J. H., and Stebbins, G. L. (1987). Chromosomal evolution in the Gramineae. In

"Grass Systematica and Evolution" (T. R. Soderstrom, K. W. Hilu, C. S. Campbell,

and M. E. Barkworth, Eds.), pp. 179-187, Smithsonian Institution Press, Washington

D.C.

Kellogg, E. A. and Watson, L. (1993). Phylogenetic studies of a large data set. I.

Bambusoideae, Andropogonodae, and Pooideae (Gramineae). Bot. Rev. 59: 273-343. 114

Keng, P. C. (1987). A systematic key to the tribes and genera of subfam. Bambusoideae

(Gramineae) occurring m China and its neighborhoods. J. Bamboo Res. 6: 13-28.

Keng, Y. L. (1959). "Flora lUustralis Plantarum Sinicarum - Gramineae," Peking.

Kim, K.-J., and Jansen, R. K. (1994). Comparisons of phylogenetic hypotheses among

different data sets in dwarf dandelion {Krigia, Asteraceae): Additional information

from intemai transcribed spacer sequences of nuclear ribosomal DNA. Plant Syst.

Evol. 190: 157-185.

Liston, A., and Alvarez-Buylla, E. (1995). Intemai transcribed spacer sequences of conifers:

"There is a fiingus among us." Inoculum 46: 26. (Abstract.)

Liston, A., and Alvarez-Buylla, E., Robinson, W. A., and Oliphant, J. M. (1995). Length

variation in the nuclear ribosomal DNA ITS of Coniferales, Ginkgoales, and Gnetales.

Am. J. Bot. 82 (Suppl.): 146. (Abstract.)

Lu, Y. H., and Negre, S. (1993). Use of glycerol for enhanced efficiency and specificity of

PCR amplification. Trends Genet. 9: 297.

Lyal, C. H. C. (1986). Coevolutionary relationships of lice and dieir hosts: A test of

Farenholtz's Rule. In "Coevolution and Systematics" (A. R. Stone and D. L.

Hawksworth, Eds.), pp. 77-91, Clarendon Press, Oxford.

Munro, W. (1868). A monograph of the Bambusaceae, including descriptions of all the

species. Trans. Linn. Sac. London 26: 1-157.

Nishida, H., and Sugiyama, J. (1995). A common group I intron between a plant parasitic

fungus and its host. Mol. Biol. Evol. 12: 883-886.

Olmstead, R. G., and Palmer, J. D. (1994). Chloroplast DNA systematics: A review of

methods and data analysis. Am. J. Bot. 81: 1205-1224.

Paterson, A. H., Brubaker, C. L., and Wendel, J. F. (1993). A rapid method for extraction

of cotton (Gossypium spp.) genomic DNA suitable for RFLP and PCR analysis. Plant

Mol. Bio. Rep. 11: 122-127. 115

Pearson, W. P., and Lipman, D. J. (1988). Improved tools for biological sequence

comparison. Proc. Natl. Acad. ScL, USA 85: 2444-2448.

Rao, V. B. (1994). Strategies for direct sequencing of PCR-amplified DNA. PCR Methods

Applic. 4: 15-23.

Ridand, C. E. and Straus, N. A. 1993. High evolutionary divergence of the 5.8S ribosomal

DNA in Mimulus glaucescens (Scrophulariaceae). Plant Mol. Biol. 22: 691-699.

Ritland, C. E., Ritland, K. and Straus, N. A. 1993. Variation in the ribosomal internal

transcribed spacers (ITS1 and ITS2) among eight taxa of the Mimulus guttatus species

complex. Mol. Biol. Evol. 10: 1273-1288.

Roberts, L. (1991). Finding DNA sequencing errors. Science 252: 1255-1256.

RoUo, F., Sassaroli, S., and Ubaldi, M. (1995). Molecular phylogeny of the fungi of the

iceman's grass clothing. Curr. Genet. 28: 289-297.

Roux, K. H. (1995). Optimization and troubleshooting in PCR. PCR Methods Applic. 4:

185-194.

Sarkar, G., and Sommer, S. S. (1990). Shedding light on PCR contamination. Nature 343:

27.

Schultz, B., Wanke, U., Draeger, S., and Aust, H.-J. (1993). Endophytes from herbaceous

plants and shrubs: Effectiveness of surface-sterilization methods. Mycol. Res. 97:

1447-1450.

Smith, D. E., and Klein, A. S. (1994). Phylogenetic inferences on the relationship of North

American and European Picea species based on nuclear ribosomal 18S sequences and

the internal transcribed spacer 1 region. Mol. Phylogenet. Evol. 3: 17-26.

Smith, D. E., and Klein, A. S. (1996). Erratum. Mol. Phylogenet. Evol. 5: 286-287.

Soderstrom, T. R. (1981). Some evolutionary trends in the Bambusoideae (Poaceae). Ann.

Mo. Bot. Gard. 68: 15-47. 116

Soderstrom, T. R., and Ellis, R. (1987). The position of bamboo genera and allies in a

system of grass classification. In "Grass systematics and evolution" (T. R.

Soderstrom, K. W. Hilu, C. S. Campbell, and M. E. Barkworth, Eds.), pp. 225-238,

Smithsonian Institution Press, Washington, D.C.

Swofford, D. L. (1993). "PAUP: Phylogenetic analysis using parsimony. Version 3.11,"

Champaign, Illinois Natural History Survey.

Thompson, J. (1994). "The coevolutionary process," Chicago University Press, Chicago.

Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994). Clustal W: Improving the

sensitivity of progressive multiple sequence alignment through sequence weighting,

position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:

4673-4680.

White, J. F. (1987). The widespread distribution of endophytes in the Poaceae. Plant

Disease 71: 340-342.

White, J. F., and Glenn, A. E. (1994). A study of two fungal epibionts of grasses:

Structural features, host relationships, and classification in the genus Myriogenospora

(Clavicipitales). Am. J. Bot. 81: 216-223.

White, J. F. and Morgan-Jones, G. (1996). Morphological and physiological adaptations of

Balansieae and trends in the evolution of grass endophytes. In "Endophytic Fungi in

Grasses and Woody Plants" (S. C. Redlin and L. M. Carris, Eds.), pp. 133-154, APS

Press, St. Paul, Minnesota.

White, J. F., Breen, J. P., and Morgan-Jones, G. (1991). Substrate utilization in selected

Acremonium, Atkinsonella, and Balansia species. Mycologia 83: 601-610.

White, T. J., Bums, T., Lee, S., and Taylor, J. (1990). AmpUfication and direct sequencing

of fungal ribosomal RNA genes for phylogenetics. In "PCR Protocols: A Guide to

Metiiods and Application" (M. A. Innis, D. H. Gelfand, J. J. Smnsky, and T. J.

White, Eds.), pp. 315-322, Academic Press, San Diego. 117

Zhang, W. (1992). The classification of Bambuseae (Poaceae) in China. J. Am. Bamboo

Sac. 9: 25-42. 118

CHAPTER 4. GENERAL CONCLUSION

General Discussion

Previous molecular and morphological studies provided differing resolutions regarding the phylogeny of bambusoids, due to either limited sampling of the bambusoid representives with a focus on other grass groups (Esen and Hilu 1989; Davis and Soreng 1993; Kellogg and

Watson 1993; Watanabe et al. 1993; Nadot et al. 1994; Barker and Linder 1995; Duvall and

Morton 1996; Liang and Hilu 1996), or insufficient phylogenetic information (Hamby and

Zimmer 1988; Doebley et al. 1990). However, the present and some other studies (Clark et al.

1995; Clark and Kelchner, in prep.; Zhang et al., unpubl.) consistently resolve a monophyletic bambusoid subfamily consisting of only two tribes: Bambuseae, the woody bamboos, and

Olyreae (including Parianeae and Buergersiochloeae), the herbaceous olyroid bamboos; all other traditional bambusoids either associate with other major groups of grasses, or represent their own distinct lineages.

The herbaceous olyroid bamboos were recognized as monophyletic in most recent morphological and anatomical studies, because of several synapomorphies, such as the unisexual spikelets, crenate silica bodies, and other olyroid anatomical characters (Soderstrom and Ellis 1987), but their relationship with Buergersiochloa. and the recognition of the

Parianeae, were always problematic. Both molecular and morphological/anatomical analyses of this study, however, clearly illustrated the robust resolution of the olyroid group, and its relationships with Buergersiochloa and Pariana/Eremitis. The woody bamboos also were always regarded as a monophyletic group morphologically, but the phylogeny within the group was a puzzle. The lack of sufficient morphological and anatomical character information. 119 especially for inflorescence characters, and the subjective selection of the key diagnostic characters, left this particular group a long way from being well understood. Although our analyses strongly supported the monophyletic woody bamboos, they did not provide robust resolution below the tribal level.

Of particular significance was the resolution of Puelia at the base of the higher grass clade in the parsimony analysis. This strengthens the conclusion that characters such as the forest habitat, broad, pseudopetiolate leaf blades, arm and fiisoid cells in the chlorenchyma, and trimerous flower parts, all used previously to define the Bambusoideae s. 1., are in fact symplesiomorphies for the whole family. Because of the presence of several florets per spikelet in t^ielia. we conclude that the transition from one floret per spikelet to two or more occurred at the base of the higher grasses. Biogeographically, Puelia is African and almost certainly of Gondwanan origin, which is consistent with Clark et al.'s (1995) suggestion that early diversification within the grass family took place in the southern hemisphere.

The position of Streptogvna was not so strongly resolved in this study, but we demonstrated that Streptogvna associated with the oryzoid clade, albeit weakly.

Morphologically and anatomically, Streptogvna is of uncertain affinities; it did not belong to the

'core bambusoids' (Soderstrom and Ellis 1987), nor was it related to the Streptochaeteae

(Kellogg and Watson 1993). Our molecular analysis showed that Streptogvna was closer to the oryzoids, rather than to the bambusoids by its resolution at the base of the oryzoid clade to form a monophyletic unit, which is separate from the bambusoid clade, while morphological and anatomical data also indicated that Streptogvna was further distant from the bambusoids than from the Oryzeae. The results imply that Streptogvna may represent a remote group in the oryzoid lineage, which should probably be recognized as its own subfamily.

The molecular analyses provided powerful resolution for the deeper branches of the bambusoid subfamily, but it also reminded us of new potential problems. Our DNA sequence data were generated by using the polymerase chain reaction (PGR) technique, as is done in 120 most other molecular studies; however, PGR does not always amplify target regions specifically (Olmstead and Palmer 1994). The close association between bamboos and fiingi resulted in the non-specific PGR amplification of fiingal DNA, and provided false phylogenetic resolution. Ascomycetes were reported associated with many grasses and bamboos (Hino

1961; White 1987), but our analysis indicated all the fungi associated with the sampled bamboos belonged to the basidiomycetes, and none of them was an ascomycete. The association between the host bamboos and fiingi did not show the coevolutionary pattern suggested by some studies G-yal 1986; Thompson 1994). The selective environment of sampling (our host plants were growing in a greenhouse or botanical garden) may have contributed to the different conclusion in our study, but the possibility of a natural basidiomycetous mycoflora inhabiting the leaves of bamboos deserves further investigation.

Recommendations for Further Research

Phylogenetic study of the bambusoids still has not attracted the same level of attention as the other grasses, and more research is needed to describe basic diversity and fiarther explore phylogenetic relationships within this economically and ecologically important subfamily. This study is the first inclusive smdy of bambusoid phylogeny, but it is clear that we need more both molecular and morphological data to establish a reUable phylogeny. With regard to molecular data, although similar resolution was observed in other smdies (Clark and Kelchner, in prep.; Zhang et al. unpubl.), all smdies used plastid DNA sequence data. More cpDNA analyses may provide additional resolution, but because of the uniparental inheritance of plastids, it is even more inportant to use nuclear data in the future to estimate bambusoid phylogeny. Morphologically and anatomically, extensive character and character state selection and wider sampling are necessary in order to get rigorous resolution. Intensive and thorough 121 studies of individual features, such as inflorescences and rhizomes, will be extremely helpful for understanding character evolution in the bamboos, and therefore, provide more reliable information for phylogenetic reconstruction.

Two African genera, Puelia and Guaduella. have not been well studied so far either morphological or from molecular perspectives. It will be very interesting to include these two groups in fiiture phylogenetic analyses of the bambusoids and higher grasses. The ndhF sequence data showed a very intriguing resolution for Puelia. but the study included only one species. The inclusion of more representives from the group will allow a more robust resolution. Guaduella is one of two groups from which we failed to get reliable ndhF sequence data in the study, but well perserved herbarium or fresh leaf materials will provide a chance to isolate high quality DNA for friture molecular studies. Detailed morphological and anatomical studies will help us to understand and revaluate the molecular phylogeny for both genera as well as for the bambusoids.

Monophyly of the redefined Bambusoideae was well resolved in this study; however, the relationships within this clade, especially those of the woody bamboos, remain obscure.

Insufficient information from both morphological/anatomical and molecular data is the main reason for the lack of resolution. More rigorous evaluation of both inflorescence and vegetative characters, with particular attention to branching, is needed but will require some time and dedicated effort. The long generation time in the woody bamboos may be directly related to the relatively low levels of sequence evolution seen in the ndhF gene and the rpil6 intron (Clark et al. 1995; Clark and Kelchner, in prep.; Zhang et al., unpubl.). However, for molecular studies, we can use some alternative molecular techniques, such as microsatellites and AFLPs, which are usually used to resolve inter- or intraspecific phylogenetic relationships, to collect enough variable information, or choose fast-evolving gene(s) or other DNA regions such as ITS and IGS to obtain more informative sequence data. When we generate nuclear 122

DNA sequence data for plant phylogenetic analyses, we should be aware of the potential for errors caused by fungal contamination, which has already been demonstrated in this study.

Literature Cited

BARKER, N. P., H. P. LINDER, and E.H.HARLEY. 1995. Polyphyly of Arundinoideae

(Poaceae): Evidence from rbcL sequence data. Systematic Botany 20:423-435.

CT.ARK, L. G., W. ZHANG, and J. F. WENDEL. 1995. A phylogeny of the grass family

(Poaceae) based on ndhF sequence data. Systematic Botany 20; 436-460.

DAVIS, J. I and R. J. SORENG. 1993. Phylogenetic strucmre in the grass family (Poaceae)

as inferred from chloroplast DNA restriction site variation. American Journal of

Botany 80: 1444-1454.

DOEBLEY, J., M. DURBIN, E. M. GOLENBERG, M. T. CLEGG, and D. P. MA. 1990.

Evolutionary analysis of the large subunit of carboxylase (rbcL) nucleotide sequence

data among the grasses (Poaceae). Evolution 44: 1097-1108.

DUVALL, M. R. and B. R. MORTON. 1996. Molecular phylogenetics of Poaceae: an

expanded analysis of rbcL sequence data. Molecular Phylogenetics and Evolution 5:

352-358.

ESEN, A., and K. W. HILU. 1989. Immunological affinities among subfamilies of the

Poaceae. American Journal of Botany 76: 196-203.

HAMBY, T. K. and E. A. ZIMMER. 1988. Elibosomal RNA sequences for inferring

phylogeny within the grass family (Poaceae). Plant Systematics and Evolution 160:

29-37.

HINO, I. 1961. Icones Fungorum Bambusicolorum Japonicorum. The Fuji Bamboo

Garden, Japan. 123

KELLOGG, E. A., and L. WATSON. 1993. Phylogenetic studies of a large data set. L

Bambusoideae, Andropogonodae, and Pooideae (Gramineae). Botanical Review 59:

273-343.

LIANG, H., and K. W. HILU. 1996. Application of the matK gene sequences to grass

systematics. Canadian Journal of Botany 74: 125-134.

LYAL, C. H. C. 1986. Coevolutionary relationships of lice and their hosts: A test of

Farenholtz's Rule. Pp. 77-91 in Coevolution and Svstematics . eds. A. R. Stone and

D. L. Hawksworth. Oxford : Clarendon Press.

NADOT, S., R. BAJON, and B. LEJEUME. 1994. The chloroplast gene rps4 as a tool for

the study of Poaceae phylogeny. Plant Systematics and Evolution 191: 27-38.

OLMSTEAD, R. G., and J. D. PALMER. 1994. Chloroplast DNA systematics: A review of

methods and data analysis. American Journal of Botany 81: 1205-1224.

SODERSTRM, T. R., and R. ELLIS. 1987. The position of bamboo genera and allies in a

system of grass classification. Pp. 225 - 238 in Grass svstematics and evolution, eds.

T. R. Soderstrom, K. W. Hilu, C. S. Campbell, and M. E. Barkworth. Washington,

D. C.: Smithsonian Instimtion Press.

WATANABE, M., M. ITO, and S. KURITA. 1994. Chloroplast DNA phylogeny of Asian

bamboos (Bambusoideae: Poaceae) and its systematic implication. Jouma of Plant

Research 107: 253-261.

WHITE, J. F. 1987. The widespread distribution of endophytes in the Poaceae. Plant

Disease 71: 340-342.

THOMPSON, J. 1994. The coevolutionarv process. Chicago: Chicago University Press. 124

APPENDIX. COMPLETE NDHF GENE SEQUENCE DATA MATRIX

10 20 30 40 SO 60 Oryza sativa TGGGTAATCCCTCl'rCTXJaJACrreCAGTTATTATGTCAATGUGU'i'i'i'GGALTI'i'ri'C'fT Leersia virginica T C MaXtebrunia peciolaca T Ehrharta calycina C A Streptogyna americana 1 A Streptogyna americana 2 A Diarrhena obovaca c...A Brachyelytrum erecmni A A Poa pratensis T C T.. Avena sativa A C T.. Phaenospenoa globosma ? C...A Joinvillea ascendens ...A A...T. -..T C...C A..A C.C... screpcochaeca angiistifolia A...T....T G A...G C Anomochloa marancoidea A..CT....T G A C Pharus latifolius A...T T A Pharus lappulaceus A...T A Leptaspis banksii T A...C Puelia olyriformis Buergersiociiloa bambusoides T A Pariana radiciflora A Eremitis sp. nov. A Olyra latifolia A .A T.I faumilis A Lithachnfi pauciflora ? A Sucrea maculata A Raddia distichopfaylla T A Anindinaria gigeintea A Pseudosasa japonica A Ampelocalamus scandens A Bashania fargesii A Shibataea kumasaca A rh mannorea A Phyllostachys pxibescens A Phyllostachys bambusoides A Sasa variegata A Fargesia robusta A Yushania exilis A Glaziophyton mirabile A Alvimia gracilis A Chiisquea lacifolia CA C Chusquea circinaca C Neurolepis aperta T G A Arthrostylidiuitt ecuadorense A Apoclada simplex A Rhipidocladum pictieri A Ocatea acuminaca A Guadua paniculaca A Melocanna baccifera A Baiobiisa aff. banibos A Bambusa scenostachya A Cephalostachyma pergracile A Schizostachyum luzcnicum A Racemobambos microplylla A HicJcelia madagascariensis A A Nastiis elatus A Sporobolus indicus Zoysia matrella

Phragmites aus&ralis Molinia caerulea 125

70 80 90 100 LIO 120

Oryza saciva Leersia virginica Kaltebrunia peciolata Ehrfaarta calycina Streptogyna americana 1 Screptogyzia americana 2 Oiarrhena obovaca Brachyelytrum erectum Poa pracexisis Avena saciva Phaenospexsaa globosisa Joinvillea ascendens Screpcochaeca asgustifolia Anomnrhloa oarancoidea Pharus latifolliis PhcLTus lappulaceus Leptaspis banJcsii Puelia olyriformis Buergersiochloa bambusoides Pariana radiciflora gremitis sp. nov. Olyra latifolia Lithachne huznilis A G G.TC. Lithachne pauciflora .G.TC. Sucrea maculaca Raddia discichophylla Anmdinaria gigantea Psetidosasa japonica Ampelocalamus scandens Bashania fargesii Shibataea kumasaca Chimonobambusa mamorea Phylloscacfays pubescens Pbylloscachys bambusoides Sasa variegata Fargesia robusta Yushania exilis Glaziophyton mirabile Alvimia gracilis Chusquea lacifolia Chusquea circinata Meurolepis aperca Arthroscylidium ecuadorense Apoclada simplex Rhipidocladum pictieri Otatea acuminata Guadua paniculaca Helocanna baccifera Bambusa a£f. bambos Bambusa stenoscachya Cephaloscachyum pergracile Schizostachyum luzonicum Racemobambos microphylla Hic)celia madagascariensis NasCus elacus Sporobolus indicus Zoysia macrella Zea mays Sorghum bicolor Phragmices auscralis Holinia caerulea 126

130 140 150 160 170 180

Oryza sativa AGTATAGCTATCCTATTCTCAGTPCACCroTCTATTCAACAAATaAATGGAAGTTCTATC Leers ia vixginica cc Halcebrunia peciolata cc Khrharta calycina .. .C C T Streptogyna americana I Strepcogyna americana 2 Diarrhena obovaca Brachyelytrum erectun n A .c Poa pracensis . . . .A, . c... Avena saciva Phaenospenna globosun Jolnvillea ascendens r a ...A..T Str^cochaeta angustifolia Anomochloa marantoidea .C T Pharus lacifolius A Pharus Xappulaceus Leptaspis banksii PueXia olyrifonnis Suergersiochloa bambusoides c. Pariana radiciflora c. Eremitis sp. nov. c. Olyra latifolia c. Lithachne humilis c. Litfaachne pauciflora Sucrea naculata Raddia disticbophylla Arundinaria gigantea A Pseudosasa japonica A Ampelocalamus scandens A Bashania fargesii A Shibataea bjmasaca A Chimonobambusa marmorea A Phyllcscachys pubescens A Phyllostacfays banbusoides Sasa variegata A Fargesia robusca A Yushania exilis _ a Glaziophyton mirabile . . C.. T A .G Alvimia gracilis T A , .G Chusquea Iaci£olia T A Chusc[uea circinaca Neurolepis aperta .. .T A Arthroscylidium ecuadorense T A Apoclada simplex T A Rhipidocladum pitcieri T A .G Ocatea acuminaca T A Guadua paniculata T a Helocanna baccifera A Bambusa a£f. bambos A Bambusa stenoscachya a C- Cephaloscachyum pergracile A Schizoscachyum liizonicim A Racemobambos oicrophylla A Hickelia madagascariensis Nascus elatus . a Sporobolus ixidicus c A. .G Zoysia oacrella n r A- .G Zea mays c Sorghum bicolor Phragmices auscralis htolinia caeriilea 127

190 200 210 220 230 240

HaXtebrunia peciolaca A....C Ehrharta calycina C SCreptogyna americana 1 G Streptogyna americana 2 G Diarrhena obovaca A G Brachyelytrunx erectun G Poa pratensis G Avena sativa T G Phaenosperma globosina G Jolnvillea ascendens A G G Strepcochaeta anguscifolia A C G G Anomochloa marantoidea TA C....G G Pharus latifolius A C G G Phanis lappulaceus A C....G G Leptaspis banksii A C G G Puelia olyriformis A G Buergersiochloa bambusoides G Pariana radiciflora T..G Eremicis sp. nov. T T..G Olyra lacifolia T..G Lithachne humilis T TC T..G Lithachne jjauciflora T TC T..G Sucrea maculata T T..G Raddia distichophylla T..G Arundinaria gigantea G Pseudoscisa japonica G Ampelora 1 aimjs scandens G Bashania fargesii G Shibacaea Joanasaca G Chlmonobambusa tnarmorea G Phylloscachys pubescens G Phylloscachys bambusoides G Sasa variegata G Fargesia robtisca G Yushania exilis G Glaziophycon mirabiie C G Alvimia gracilis A G Chi;squea latifolia G Chusquea circinaca G Neurolepis aperta G Arthroscylidium ecuadorense A G Apoclada simplex G Rhipidocladum pitcieri G Otacea acuminata G Guadua panictilaca G Melocanna bacci£era G Bambusa aff. bambos G Baabusa scenoscachya G Cephalostachyum pergracile G Schizoscachyum luzonicum G Racemobambos microphylla G HicJcelia madagascariensis G Nascxis elacus G Sporobolus indicus A G Zoysia matrella A G 2ea mays A G Sorghum bicolor A G Phragmites australis A G Molinia caerulea A G 128

250 260 270 280 290 300

Oryza sativa ATCGACCCGCTTACVirim'lATGTTAATACTftATTACTACTGTAGGAATCCTGG'i'I'C'rr Leersia virginica Maltebrunia petiolaca Ehrharta calycina ..T T Streptogyna americana 1 Strepcogyna americana 2 Diarrhena obcvata ..T Brachyelytrrm erectum ..T Poa pracensis ..T G Avena sativa ..T Phaenospezna gloisosxm ..T Joinvillea ascendens T G. C C.A Streptocbaeca angustifolia TA Anomochloa marantoidea TT Phanis laCifolius G A Pharus lappulaceus G A Leptaspis banksii G A Puelia olyrifcrmis Buergersiochloa bambusoides C Pariana radiciflora T Eremitis sp. nov. T Olyra latifolia G....C Lithachne humilis C Lithachne pauciflora C Sucrea maculata C Raddia distichcphylla C Arundinaria gigeincea C Pseudosasa japonica C An5)elccalamis scandens c Bashania fargesii C Shibataea kmnasaca C rh->mnr>nhamh»ga mamorea C Phyllostachys pubescens C Phyllostachys bambusoides C Sasa variegaca C Fargesia rofausca C Yushania scilis C Glaziophyton mirabile T Alvimia gracilis T Chusquea latifolia C Chusquea circinata C C Neurolepis aperta C Arthroscylidium ecuadorense C T Apoclada simplex T Rhipidocladum piccieri T Otatea acuminata T Guadua paniculata T Melocanna baccifera C Bambusa aff. bambos C Bambusa stenostachya A c cephaloscachyum pergracile C Schizostachyum luzonicum C Racemobambos microphylla A C Hickelia madagascariensis C Nastus elatiis C Sporobolus indictis T..C T Zoysia matrella C T Zea mays G C G. T.A Sorghum bicolor G C G T.A Phragmices auscralis C T Molinia caerulea C T 129

310 320 330 340 350 260

Oryza sativa ATTTATAGTGATGATTAIATGTCrCACGATGAGGGATATTroAGATTTTrTGTTTATATA Leersia virginica Malcebrunia peciolata r G ...A Ehrfaarta calycina r .T.. c Strepcogyna americana 1 r Strepcogyna americana 2 r ...A Diarrhena obovaca r n ...A Brachyelytrum ereccum r r; ...A T Poa pratensis r G ...A C C Avena saciva r G ...A. .G Phaenospezna globosum r a ...A Joinvillea ascendens r .C...... A S&reptochaeta angusti£olia r ...A Anomochloa marantoidea ..c.. r T ...A Pharos latifolius r .. .A Pharus lappulaceus r A Lepcaspis banksii r . - .A Puelia olyrifonnis A... Buergersiochloa bambusoides - A Pariana radici£lora r ...T, .. .A Eremicis sp. nov. r ...T, Olyra laci£oXia r r .C...... A Lithachne faumilis .CA.C..c .. .A Lichacfane pauci£lora ra.r..c...... A Sucrea maculaca r r .c ...A Raddia discichopbylla r GT ...A Arundinaria gigantea r ...A Pseudosasa japonica r ...A Anipelacalainus scandens r ...A Bashania £argesii r ...A Shibataea kumasaca r ...A Chinonobambusa mannorea r ...A Phyllostacfays pubescens r .. .A Phylloscacfays bambusoides Sasa variegaca r .. .A Fargesia robusta r ...A Yushania exilis r .. .A Glaziophyton mirabile r .. .A Alvimia gracilis r .. .A Chusquea Iati£oIia r .. .A Chusquea circinaca r ...A Keurclepis aperta r .. .A Arthrostylidiimi ecxiadorense r .. .A Apoclada simplex r .. .A Rhipidocladum pitcieri .. .A A Ocacea acuminata r .. .A Guadua paniculata r .. .A Helocanna baccifera r .. .A Bafflbusa a££. bambos r ...A Banbusa stenostachya ..C. r ...A Cephaloscachyum pergracile r ...A Schizostachyum luzonicum .. .A Racemobambos microphylla ...A Hickelia oadagascariensis r ...A Nascxis elacus r .. .A Sporobolus indicus C Zoysia matrella . c. r c Zea mays .. .A Soirghiim bicolor .. .A Phragmites australis .. .A Koliiiia caenilea 130

370 380 390 400 410 420

Oryza saciva j^:;i'iUTiUiaA!r!ACTTCCaTGTTGGGATrGGTTACTAGTTCCAATTTGATACAAATTTAT Leersia vlrgixiica Maltebnmia petiolata Ehrbarta calycina C Streptogyna americana 1 Streptogyna americana 2 Diarrhena obovaca Brachyelytrum erectum T Poa pratensis c T.. c Avena saciva ..C.A... .c Phaenospenna globosizm c Joinvillea ascendens ...T.. ...A C Streptochaeta angusti£oIia Anomochloa marantoidea Pharus Xacifolius Pharus lappulaceus T Leptaspis banksii Puelia olyriforsiis Buergersiochloa bambusoides Pariana radiciflora TC. T.. , Eremitis sp. nov. c. ...T Olyra lacifolia ..T T.. Lichachne humilis ..T T...c Lifhachne pauciflora Sucrea maculaca Raddia discicbopbylla Arundinaria gigaxicea Pseudosasa japonica Ampelocalamus scanripns Bashania fargesii Shibataea kumasaca T Chinonobainbusa naraorea Phyllostachys pubescens Phyllostachys bambusoides C Sasa variegaca Fargesia robusta Yushania exilis Glaziophyton nirabile Alvimia gracilis Chusquea latifolia Chusquea circinaca Neurolepis aperta Arthrostylidium eciiadorense Apoclada simplex T Phipidocladum pittieri Otatea acuminata T Guadua paniculata T Helocanna baccifera Bambusa aff. bambos Bambusa stenostachya Cephalostachyum pergracile Schizostachyum luzonicum Racemobambos microphylla Kickelia oadagascariensis Nastus elatus Sporobolus indicus Zoysia matrella Zea mays Sorghum bicolor Phragmices australis Molinia caerulea 131

430 440 450 460 470 480 oryza sativa TTTTTTTGGGAGCTTGTGGGAATGTGTTCCTATTTATTGATAGGCTTTTGGTTTACACGG Leers ia virginica Haltebrunia petiolata Ehrfaarta calycina A Streptogyna americana 1 A Strepcogyna americana 2 A Diarrhena obovata A..C A Bracfayelytrum erectum A A Poa pratensis A. .c A Avena sativa A..C..A A Phaenospezma giobosum A Joinvillea ascendens A..G A A C Streptochaeca angustifolia A T A C Anqnochloa marantoidea A A T A C Pharos latifolius A A T C Pharus lappulaceus A A T C Leptaspis banksii A. Puelxa olyriformis A Buergersiochloa bambusoides A Pariana radiciflora A C Ereadtis sp. tiov. A c Olyra latifolia A Lithachne humilis A Lichachne pauciflora A Sucrea maculaca A Raddia distichophylla A Arurdinaria gigantea A Pseudosasa japonica A Agpelocalamus scandens A Bashania fargesii A Shibataea kumasaca A ry,i Tunnnhamhi i ca fiwnncrea A Phylloscachys pubescens A Phyllostachys bambusoides A Sasa variegaca A Fargesia robusta A Yushania exilis A Glaziophycon mirabile A C Alvimia gracilis A Cbusquea lacifolia A T Chusquea circinaca A T Neurolepis aperta A Art±rostylidium ecuadorense A Apoclada simplex A Rhipidocladum pittieri A Otacea acuminata A A Guadua paniculata A Kelocanna baccifera A aff. bambos A Bafflbusa scenoscachya A Cephalostachyum pergracile A SchizosCachyum luzonicum A Racemobajnbos microphylla A Hickelia madagascariensis A Nastus elacus A Sporobolus indicus A A Zoysia macrella A A Zea mays A c G... Sorghum bicolor A c G... Phragmites australis A Molinia caerulea A 132

490 500 510 520 530 540

Oryza sativa CCAATTGCAGCGACjTGCITGTCAAAAAGC^'i'iTOrAACTAATCGTGTAGGGGATTTTGGT Leersia virginica A Haltebrunia petiolaca Ehrharta calycina Streptogyna americana 1 Strepcogyna americana 2 Diarrhena obovaca A G Brachyelytrum erectum ..G Poa praCensis C Avena saciva C Phaoiospezma globosum ..T..C C JoinxrIXlea ascendens G Streptocbaeca angustifolia C Anamochloa oarancoidea Pharus Iati£olius ...C C Phams lappulaceus ...C C Leptaspis banksii Puelia olyri£ormis Buergersiochloa bambusoides Pariana radiciflora A Cremitis sp. nov. Olyra latifolia Lichachne humilis G Lithachne pauci£lora G Sucrea maculata A Raddia discichophylla A Arundinaria gigancea Pseudosasa japonica Ampelocalainus scandens Bashania fargesii Shibataea kuznasaca Chimonobafflbusa marmorea. Phylloscachys ptibescens Phyllostachys banbusoides Sasa variegata Fargesia robusca YUshania exilis Glaziophyton loirabiXe •A. Alvimia gracilis Chusquea latifolia Chusquea circinata Neurolepis aperta Arthroscylidium ecuadorense Apoclada simplex Rhipidocladua piccieri T Otatea acuminata Guadua paniculata Melocanna bacci£era Bambusa a££. bambcs Banbusa stenostachya Cephalostachyum pergracile Schizostachyum luzonicum A Elacemobambos microphylla Hickelia madagascariensis Nastus elatus Sporobolus indicus ..C T. Zoysia matrella Zea mays Sorghum bicolor Phragmites australis .G A Holinia caerulea .G A 133

550 560 570 580 590

Oryza sativa Leersia virginica KalCebrunia peciolata Ehrharta calycina Streptogyna aisericaiia 1 Strepcogyna americana 2 Diarrhena obovata Bracfayelytrum erecttim Poa pratensis Avena sativa Phaenosperma globosum Joinvillea ascendens Streptochaeta anguscifolia Ananochloa marancoidea Pharus latifolius Pharus lappulaceus Leptaspis banksii Puelia olyriformis Buergersiochloa bambusoides Pariana radici flora Eremitis sp. nov. Olyra latifolia Lithachne humilis Lithachne pauciflora Sucrea maculaca Raddia distichopbylla Arundinarla gigancea Pseudoscisa japonica Ampeloc2Llamus scandens Bashania fargesii Shibataea kumasaca rh i mnnnhamhi i ca marmorea Phylloscachys pubescens Phylloscachys bambusoides Sasa variegaca Pargesia robusca Yushania exilis Glaziophyton mirabile Alvimia gracilis Chusquea latifolia Chusquea circixiata Meurolepis aperta Arthrostylidium ecuadorense Apoclada simplex Rhipidocladum pittieri Otatea acisninata Guadua paniculata Kelocanna baccifera Bambusa aff, bambos Bambusa stenostachya Cephalostachyum pergracile Schizostachyum luzonicum Racemobambos microphylla Hickelia tnadagascariensls Nastus elatus SporoboXus indicus Zoysia matrella Zea mays Sorghum bicolor Phragmites australis Holinia caenilea 134

610 620 630 640 650 660

Oryza sativa TTTAAAATAGCTAATAACrGGATTCCTAATAATGAGATTAACTCCTrGCTTACrATTTTG Leersia virginica ...... (a...... Kaltebruxiia pe&iolaca .G.. . ;; Ebrharta calycina ..c... .G... .-T. .A. .... -C. Streptogyna americana 1 .G... ..T. .A...... C. screpcogyna americana 2 ..c... .G... ..T. .A. .... Oiarrhena obovata .G... ..T. -A..A...... c. Brachyelytnim ereccuai ..c... .C .G... ..T. .A. .c. Poa pratensis ..c... .T .GA.. . -...T. .A...... c. Avena sativa .GA.. ..T. .T. .A...... c. Phaenospexina globosum ..c... .G... ..T. .TG.A...... c. Joinvillea ascendens ..c... .G...C. .T. .A. C.. . .c. Screptochaeca axigustifolia ..cc.. A. -G... ..T. .AA .... .c. Anomochloa marantoidea ..cc.. ...AG. ..A.. ..T. .A...... c. Pbanis lacifolius C. G..G. .. ..T. .A. ..T. .c. Pharus lappulaceus ..c... c. G..G.. . ..T. ..T. .c. Leptaspis banksii ..c... c. G..G... ..T. .A. ..T. .c. Puelia olyriformis ..c... .G... ..T. .AT .... .c. Buergersiochloa bambusoides ..c... C .G... ..T...... c. .A Pariana radiciflora C .G... ..T. .A. .... c. .A Eremicis sp. nov. r Olyra lacifolia c .GA.. ..T. .A...... c. .A Lithachne humilis T... G c .G... .-T. .A...... c. Lithachne pauciflora ....T... G .G ..T. A r Sucrea maculaCa c .G... .GT. .A. .c. .A Raddia discichophylla c .G... ..T. .T. .A. .. .A c. .A Arundinaria gigantea ..c... .G... ..T. ... .A. c. Pseudosasa japonica ..c... .G... ..T. .A. c. Ampelocalannis scandens ..c... .G... ..T. .A. c. Bashania fargesii ..c... .G... ..T. .A. c. Shibataea kumasaca ..c... .G... ..T. .A. c. Qumonobaxnbusa marmorea ..c... -G... ..T. .A. c. Phyllostacbys piibescens ..c... .G... ..T. .A. c. Phyllostachys bambusoides ..c... .G... ..T. .A. .... c. Sasa variegaca ..c... .G... ..T. .A. c. Pargesia robusca c G.. . ..T. A r Yushania ^cilis ..c... .G... ..T. .A. c. Glaziophyton mirabile ..c... .G... ..T. .A. c. Alvimia gracilis .G ..T. A r Chusquea latifolia ..c... .G.. . ..T. .A. c. Chusquea circinata c... .G ..T. r Keurolepis aperta ..c... .G.. . a..T. .A. c. Arthrostylidium ecuadorense ..c... .G... ..T. .A. .... c. Apoclada simplex G. .G.. . ..T. .A. .... c. Rhipidocladum pittieri ..c... .G... ..T. .A. .. .A c. Otatea acumixmta G..G. .. ..T. .A. .... c. Guadua paniculata ..c... G..G. . . ..T. .A. .... c. Melocaxma baccifera ..c... .G... ..T. .A. .... c. Bambusa aff. bambos ..c... .G... ..T. .A. .... c. Bambusa stenostachya ..c... .G... ..T. .A. c. Cephalostachyum pergracile ..c... .G... ..T. .A. c. Schizostachyum luzonicum ..c... .G... ..T. .A. c. Racemobambos micropbylla .G... ..T. .A. c. Hickelia madagascariensis ..c... .G... ..T. .A. c. Nastus elatxis ..c... .G... ..T. .A. c. Sporobolus indicus ..c... .G.. . ..T. -A..A. c. Zoysia matrella ..c... .G.. . . -T..A. .A. .... c. Zea mays ..c... .G.. . ..T. .A. c. Sorghum bicolor ..c... .G. . . ..T. .A. .... c. Phragmites australis ..c... .G.. . ..T. .A. c. Molinia caerulea ..c... .G.. . ..T. .A. c. 135

670 680 690 700 710 720

Oryza saciva TGTGC'fmTl'ATTMTCLTiUilVjaXJrrGCGAAATCGGCACAA'rTCCCTCTTCACgrA Leersia virginica Haltebrunia peciolaca A - Ehrharca calycina ..T. . A.... .A. - c. Strepcogyna americana 1 A.. T. Strepcogyna americana 2 A.. ...T. Diarrhena obovaca A.... .A.. T. Brachyelycrum erect:mn A.. T. G A.. Poa pracensis A.. .. .A.. .. .C. r? Avena sativa A.. c. Fhaenospexnia globostm A.. T. Joinvillea ciscendens .C. ... .G. . A.. T.. T. r c Streptochaeca anguscifolia .C... A.... .A.. .. .C. Anomochloa marancoidea .A. .. ..T.. A.. A.. T. .G Pharus latifolius .c. A.. T. Fharus lappulaceus -C... A.. .. .T. Lepcaspis banksii .c... A.. ...T. Puelia olyriformis .c... A.. T. Buergersiociiloa banbusoides A.. .. .T. Pariana radiciflora A.. ...T. G Eremitis sp. nov. A.. .. .A. Olyra lacifolia A.. .. .T. Lithachne humilis A.. .. .T. A.. Lichachne pauciflora A.. .. .T. A.. Sucrea maculaca ..C.. A.. .. .C. Baddia distichophylla ..CA. A.. ...c. Arundfnaria gigancea A.. .. .T. Pseudosasa japonica A.. .. .T. Ampelocalanus scandens A.. .. .T. Bashania fargesii A.. .. .T. Shibataea kumasaca A.. .. .T. Chimonobambusa marmorea A.. .. .T. Phylloscachys pubescens A.. .. .T. Pbylloscachys bambusoides A.. .. .T. Sasa variegaca A.. .. .T. Fargesia robusta A.. , .. -T. Yushania exilis A.. , .. .T. Glaziophycon mirabile Alvimia gracilis Chusquea lacifolia A...... T. Chusquea circinaca Heurolepis aperca A.. . . .A...... T. Arthrostylidium ecuadorense A.. , .. .C. Apoclada simplex A...... T. Rhipidocladum piccieri A...... T. Ocacea acuminaca A.. . T Guadua paniculata A...... T. Helocamia bacci£era A...... T. Banbusa af£. bambos A...... A...... T. Bambusa scenoscachya A...... A...... T. CephalosCachyum pergracile A...... T. Schizoscachyum luzonicum A.. . ..T. Hacemobambcs microphylla A...... A... . .T. Hickelia madagascariensis A.. . ..C. Nascus elacus A.. . ..T. Spoi^holus indicus ..T A...... A... . .T. .G. Zoysia macrella ..T . A...... T.... .T. G Zea mays A.. ,. ..A.... .T. T Sorghum bicolor ... .A. .. , ..A...... T. ....T Phragmices auscralis ..T A... . .A... . .T. Holinia caerulea 136

730 740 750 760 770 780

Oryza sativa TGGTTACCCGATGCTATGGAAGGACCCaCCCCCATTTXaGCTCTTATACflaSCAATTTTT Leersia virginica Haltebruoia peciolaca Ehrharta calycina Streptogyna americana 1 G Strepcogvna americana 2 ?G Dlarrhena obovaca Brachyelytrum erectum Poa pracensis Avena sativa Fbaenospenna globostm Joinvillea ascendens Strepcochaeca anguscifolia Anomochloa marancoidea Pharus laci£olius Pharus lapptilaceus Leptaspis banksii Puelia olyriforais Buergersiochloa bambusoides Pariana radiciflora Eranicis sp. nov. G T Olyra latifolia Lichachne humilis Licfaachne pauciflora Sucrea oaculata Raddia distichopbylla Arundinaria gigancea Pseudosasa japonica Ampelocalamus scandens Basbania fcirgesii Shibacaea kumasaca Chifflonobanbusa namiorea Phylloscachys pubescens Phylloscachys bambusoides Sasa variegaca Fargesia robusca Yushania exilis Glaziophyton mirabile T G AG Alvimia gracilis AG Chusquea lacifolia Chusquea circinaca Neurolepis aperca Archroscylidium ecuadorense T AG Apoclada simplex T.... G AG Rhipidocladum piccieri A Ocacea actmninaca Guadua paniculaca T.... G Helocanzia baccifera Bambusa a££. bambos T G Banbusa scenoscachya cephaloscachyum pergracile T G Schizoscachyum luzonicum T G Racemobambos nicropfaylla T G Hickelia madagascariensis A Nascus elanis Sporobolus indicus Zoysia matrella Zea mays Sorghum bicolor Phragmices auscralis Holinia caerulea 137

790 800 aio 320 330 340

Oryza saciva CITftraGCraaCTTCITCCrCTTTTCATATCCCTACCTTTGATAATGaGTTTCATTTCT Leersia virginica AT Halcebrunia peciolaca AT Ehrbarta calycina Screpcogyxia americaca 1 ..c.., ..G.., Str^cogyna araericana 2 Diarrhena obovaca T . Brachyelytrum ereccum Poa pracensis ..c... G .T...... TA. ...C.,. .A Avena saciva Pbaenospexna globosum . c.., ...T., ...A. . Joinvillea ascendens . G...... TA. Screpcochaeca angustifolia AnoQochloa marancoidea A. ., .T Pharus lacifolius pfaarus lappulaceus ..c.., .T.. . Lepcaspis banksii c. ...AT. ,.G Puelia olyrifonnis . Buergersiochloa bambusoides ..c.., •T... .C r .A ..T Pariana radiciflora ...c... .T.. .T... .G. .A ..T Eremitis sp. nov. .A ..T Olyra latifolia .T... .A ..TG Lithacfane humilis ...c... .T...... G. .c .c..A . .TG Lichacfane pauciflora ...c...... », ... . .T...... G. .c .c..A ..IG Sucrea maculata . c... .T... G .A ..TG Raddia discicbophylla . c... .T.. . .CA. .. .A ..TG Arundinaria gigancea Pseidnsasa japonica . c... .A Anpelocalamus scandens . c... .A Bashania Cargesli .A Shihataea Icumasaca Chinonobambusa marmorea .A PhyllosCachys pubescens ..c... .A Phylloscacbys banbusoldes . c... .A Sasa variegata ..c... .A Fargesia robusca . c... .A Yushania exilis . c.. .A Glaziophyton mirabile . c... A .A Alvimia gracilis Chusquea lacifolia . c... .A Chusquea circinaca . c... .A Neurolepis aperta c. . .A Arthrostylidium ecuadorense Apoclada simplex .A Rhipidocladum piccieri ..T...... A. .A Ocacea aciimiiiaca . c... .A Guadua paniciilaca ..c... .A Helocaima bacci£era . c... .A Banbusa a££. bajnbos c. . .A Bambusa scenostachya . c... r .A Cephalostachyum pergracile c... .A Schizoscacbyum luzonicum ..c... .A Racemcbambos microphylla c.. . .C a .A Kickelia madagascariensis .-C... .A Nascus elatus ..c...... A. T .A Sporobolus indicus G, G .T . .T Zoysia matrella G. .T Zea mays ..c.... .A. r Sorghum bicolor . r _ r Phragmices australis , c Holinia caerulea 138

850 860 870 880 890 900

Oryza saciva TTAATAGGTACACTAACACri'nVi'i'AGGAGCCACTrTAGCTCiTGCrCAGAGAGATATT Leersia vixginica Haltebnixiia peciolata Ehrharta calycina Streptogyna americana I Strepcogyzsa americana 2 .G ..A... c .T Diarrhena obovata aracbyelytnim erectua .. .G ..A...... c .T G Poa pracensis Avena saCiva Phaenosperna globosxim ...G ..A...... c G..T C Joinvillea ascendens .. .G ..A...... c G..T G... Streptochaeta angustifolia ...G ..A...... C..T. G..T Anomochloa marantoidea .. .G .... .A...... C..T. G..T Pharos laclfolius .. .G .... .A...... C .T T Phams lappulaceus ...G ..A...... c .T T Leptaspis banksii .. .G ..A...... c .T T Puelia olyrifonnis Buergersiochloa bambusoides Parlana radici flora .. .G G. ..A...... c .T Cremitis sp. xiov. ...G G...... A...... c .T Olyra laCi£olia ...G .C.. .A...... c .T Lithacfane humilis .. .G .c.. .A...... G .T Lithacfane pauciflora ...G .c. ..A...... G .T Sucrea maculata Raddia disticbopfaylla Arundinaria gigantea Pseudosasa japonica AmpelocaXaimis scandens Bashaxiia fargesii Shibacaea kumasaca Chimonobambusa marmorea PhylXostachys pvibescens Phyllostachys bambusoides Sasa variegata Fargesia robusca Vushania exilis Glaziophycon mirabile ...G ..A...... c G. .T C Alvifflia gracilis .. .G ..A...... c G..T Chusquea lacifolia Chusquea circinata ...G G. ..A...... c G..T Meurolepls aperta Archrostylidium ecuadorense ...G ..A... G..C G..T Apoclada simplex ...G ..A...... c G..T C Rhipidocladum pictieri ...G ..A...... c G..T OtaCea acuminata ...G .A...... c G..T Guadua paniculaca Helocanna baccifera Bambusa aff. bambos Bambusa stenostachya Cephalostachyum pergracile .G .A...... c .T Schizostachyum luzonicum Racemobambos microphylla .G .A...... c .T Hickelia madagascariensis .G .A...... c .GC Nastus elatus Sporobolus indicus .G .A.. . .A Zoysia matrella .. .G .A Zea mays .A...... c Sorghum bicolor Phragmites australis .. .G .A...... c .T Molinia caerulea 139

910 920 930 940 950 960

Oryza saciva AAAAGAAGCTTaGCCTATTCTACAATGTCraAlTCGGTTATATGATGTTAGCTCTAGGT Leers ia virginica T Kaltebruxiia petiolata T Ehrharta calycina T A Streptogyna americana 1 Streptogyna americana 2 Diarrhena obcvata Bracfayelytrua erectm Poa pratensis Avena saciva T Phaenospenna globosina Joinvillea ascendens G G— Strepcochaeca angustifolia G Anomocbloa marantoidea Phams lacifolixjs Pharus lappxilaceus Lepcaspis baxiksii Puelia olyriformis Buergersiochloa bambusoides Pariana radiciflora Eremicis sp. nov. Olyra lacifolia Lithachne humilis Lichachne pauci flora Sucrea maculata Raddia distichophylla C Arundinaria gigancea Pseudosasa japonica Anqpelocalaxnus scandens Bashania fargesii Shibataea kiimasaca Chimonobambusa marmorea Phyllostachys pxibescens Phyllostachys bambusoides Sasa variegata Fcirgesia robusca Yusheinia exilis Glaziophyton mirabile Alvimia gracilis Chusquea latifolia Chusquea circinata Neurolepis aperta Arthrostylidium ecuadorense Apoclada simplex Rhipidocladum pittieri Otacea acuminata Guadua paniculaca Melocanna baccifera Bambusa aff. bambos Bambusa stenostachya Cephalostachyum pergracile Schizostachyum luzonicum Racemobambos microphylla HicJcelia madagascariensis A G... Nastus elacus Sporobolus indicus T Zoysia matrella T Zea mays Sorghum bicolor Phragmites australis Molinia caerulea 140

970 980 990 1000 10X0 X020

Oryza saciva ATALiUrrclTATCAAGCTtXrmaTTCCftrrTGATCaCTCaTGCTTATTCGAAftGLlTlA Leers ia virglnica Haltebrunia pecioXata Ehrharta calycina G T Strepcogyna anericana 1 Strepcogyna americana 2 Oiarrhena obovata T T BracbyeXytruxti erection A A Poa pracensis T C A Avena sativa T A phaenosperma globosxim A Joinvillea ascendens C. Streptorhaeta angustifolia A Anomachloa marantoidea Pharos latifolius A Pharos lappulaceus A Leptaspis banJcsii A Poelia olyri£oniiis T A BoergersiochXoa baafausoides .C C Pariana radiciflora .C Eresoitis sp. nov. .C Olyra latifoXia ,C Lithachne humilis Lithacfane pauci flora Socrea macolata c. Raddia disticbophylla .c c. Arondinaria gigantea Pseudosasa japonica Ampelocalaxnos scandens Bashania fargesii Shibataea kumasaca Chimonobambosa marmorea PhyllosCachys ptifoescens Phyllostachys bambusoides Sasa ^/ariegata Fargesia robusta Yoshania exilis Glaziophyton mirabile Alvimia gracilis Chusquea latifolia .A, Chosquea circinaca A, Nexirolepis aperta G •A, Arthrostylidlum ecoadorense Apoclada sisiplex Rhipidocladuzn pittieri Otatea acuzoinaca Guadoa paniculata Helocarma baccifera Bambosa a££. bambos Banbosa stenostachya Cephalostachyom pergracile Schizostachyum luzonictsa Racanobanbos microphyXla HickeXia madagascariensis Nast\is eXatos SporobcXos indicus Zoysia matreXXa Zea mays sorghum bicoXor Phragmites austraXis HoXinia caeruXea 141

1030 1040 1050 1060 1070 1080

Oryza saCiva TTATiVnaXaTCCGGftTCl^riarTCATTCflAIGGAACCTCrTGTTGGATArrCflCCft Leersia virginica ..G Maltebrunia petiolata ..G Ehrharca calycina ..G T Screpcogyna americana 1 .-G C Strepcogyna americana 2 ..G •iarrhena obovaca ..G T C Brachyelytrum ereccum ..G C C Poa pracensis .-G C Avena saciva - .G AA Phaenosperxna globose ..G c Join^rillea ascendens ..G A CA T.CG screpcochaeta angustifolia ..G A A C Anonochloa marancoidea A A Pharus laCi£ollus ..G A A C pharus lappulaceus ..G A A C Leptaspis barJcsii ..G A A C Puelia olyriformis ..G G C C Buergersiochloa banbusoides ..G C Pariana radiciflora ..G T C C C Eremi tis sp. nov. ..G C C C Olyra latifolia ..G C Lithachne humilis ..G CA G Lichachne pauciflora ..G A G Sucrea maculaca ..G CC Raddia distichophylla ..G CC Anindinaria giganCea ..G CA Pseudosasa japonica .-G CA Ainpelocalainus scandftns . -G CA Bashania feurgesii ..G CA Shibacaea kumasaca ..G CA Chimonobambusa oannorea .-G CA Phyllostachys pubescens .-G CA Phyllostachys bambusoides .-G CA Sasa variegata ..G CA. Fargesia robusca .-G CA. Yushania exilis Glaziophyton mirabile Alvimia gracilis CA Chusquea lacifolia Chusquea circinaca CA Neurolepis aperta Artiiroscylidium ecuadorense . CA Apoclada simplex Rhipidocladum pictieri . ra Ocacea acuminata Guadua paniculaca Melocanna baccifera Bambusa a£f. bambos Bambusa scenoscachya cephaloscachyum pergracile Schizoscachyum luzonicum Racsnobambos microphylla Hickelia madagascariensis Nascus elaCus Sporobolus indicus Zoysia macrella Zea mays Sorghum bicolor Phragmices auscralis Molinia caemlea Oryza saciva Leers ia virginica Haltebrunia petiolata Ehrharta calycina screpcogyna americana I Strepcogyna americana 2 Diarrhena obovata Brachyelytrum ereccum Poa pracensis Avena saciva Phaenospezma glcbosuxn Joinvillea ascendens Strepcocfaaeca aziguscifolia AnoDOchloa oarancoidea Pharus latifolius Pharos lappulaceus Lepcaspis banksii Puelia olyrifonnis Buergersiochloa bambusoides Pariana radiciflora Eremicis sp. nov. Olyra latifolia Lithachne hxjmilis Lichacbne pauciflora Sucrea maculaca Etaddia distichopfaylla Aruzodinaria gigancea Pseudosasa japonica Ampelocalamus scandens Bashania fargesii Shibataea kumasaca Chimonobafflbusa mamorea PhylXoscachys pubescens Phyllostachys bambusoides Sasa variegaca Fargesia robusta Vushania ^cilis Glaziophyton mirabile Alvimia gracilis Chusquea latifolia Chusquea circinata Neurolepis aperta Arthrostylidiim ecuadorense Apoclada simplex Rhipidocladum pitcieri Otacea acuminata Guadua paniculata Melocaiina baccifera Bambusa aff. bambos Bambusa scenoscacbya Cephalostachyum pergracile Schizostachyum luzonicum Racemobambos microphylla Hickelia madagascariensis Nascus elacus Sporobolus indicus Zoysia macrella Zea mays Sorghum bicolor Phragmites auscralis Molinia caerulea 143

1150 1160 1170 liao 1190 1200

Oryza saciva ALTiUmUTi'iVI^XSGGTACCLTi'i'CICrnVJlUiTfllTCCSCCTC'n'UCi'lU.'nx.'i'GG Leersia virginica dalcebrunia petiolata Ehrbarca calycina A C Screptagyna anericana 1 AC A T — Str^cogyna americana 2 ...AC A T — Diarrhena obovata ...AC A Bracbyelytrum erectum ..CAC A Poa pratensis ...AC A C T— Avena sacxva T.CGC A G T... Phaenospezna globosum AC A Joinvillea ascenriens ..GAC T A T— Streptorhaeta angusci£olia AC T A Anomochloa marantoidea —AC C A T — Pharus latifolius ..CAC A Pharus lappulaceus ..CAC A Leptaspis banksii ..CAC A Puelia olyriEormis ...AC C..A G Buergersiochloa bambusoides ..CAC T G Pariana radiciflora ..CAC A C Eremicis sp. nov. ..CAC A Olyra lacifolia ..CAC G C tiichachne humilis ..CAC G C Licbacfane pauciflora ..CAC A C Sucrea maculaCa ..CAC A C Raddia distichophylla ..CAC A G C Amndinaria gigancea AC G Pseudosasa japonica ...AC G Ampelocalamus scandens ..CAC G Basbania fargesii —AC G Shibacaea kumasaca ...AC G Chimonobambusa mannorea ...AC G Phylloscachys pubescens —AC G Phylloscachys bambusoides AC G Sasa variegaca AC G Fargesia robusta ...AC G Yushania exilis ...AC G Glaziophyton mirabile .G.AC G Alvimia gracilis .G.AC G Chusquea lacifolia ...AC G Chusquea circinaca —AC A Neurolepis aperca ..CAC G T... Arthrostylidium ecuadorense .G.AC G Apoclada simplex .G.AC G Rhipidocladum pitcieri .G.AC G Otatea acuminata .G.AC G Guadi.ia paniculata .G.AC G Helocanna baccifera ...AC G Bambusa aff. bambos ...AC G Bambusa stenostachya ..GAC G Cephalostachyum pergracile ...AC G Schizostachyum luzonicum ...AC G RacCTiobamhos microphylla ...AC G Hickelia madagascariensis ...AC G Nastus eiatus ..CAC G C Sporobolus indicus ...AC T A Zoysia matrella ACC T A Zea mays ....«: T A Sorghum bicolor ...AC T A C Phragmites australis ...AC A Holinia caerulea ...AC A 144

1210 1220 1230 1240 1250 1260

Leersia virginica Maltebrunia petiolata Ehrfaarta calycina G Strepcogyna aaericana 1 G C Streptogytia americana 2 G C Diarrhena obovata G c Brachyelytrum ereccma G G C Poa pratensis G..T G C Avena sativa G G Phaenospenna globosum G G G..G Joinvillea ascexxdens T G AAC...G.C Streptochaeta angustifolia G T Anomochloa narancoidea T G GACA. C Phanis latifolius ..A T G G..TA C Pharus lappulaceus ..A T G TA C Leptaspis banksii T G AA C Puelia olyriformis C....G C Bueirgersiochloa bambusoides G C Pariana radiciflora G C Eremitis sp. nov. G c Olyra latifolia G C Lichacbne humilxs A C Lithachne pauciflora A C Sucrea maculata T G C Haddia distichophylla T G C Arundineiria gigantea G C Pseudosasa japonica G C Ampelocalanais scandens G c Bashania fargesii G C Shibataea kumasaca G T C Chifflonobambusa manoorea G C Phyllostachys pubescens G C Phyllostachys bambusoides G C Sasa variegata G C Fargesia robusta G C Vushania exilis G C Glaziophyton mirabile G C Alvimia gracilis G C Chusquea latifolia G C Chusquea circinaca G C Neurolepis aperta G A G C Archrostylidium eciiadorense G C ^oclada simplex C G C Rhipidocladum pittieri G C Otatea acuminata C G C Guadua paniculata C G C Melocanna baccifera G C Bambusa aff. bambos G C Bambusa stenostacfaya G C Cephalostachyum pergracile G C Schizostachyum luzonicum G C Racemobambos microphylla G C Hickelia madagascariensis G C Nastus elatus G C Sporobclus indicus G T Zoysia matrella C G T 2ea mays G C Sorghum bicolor A G C Phragmites australis G C Molinia caerulea G C 145

1270 1280 1290 1300 1310 1320

Oryza saciva TCrnU'ACTCCAGGATTftACTGaJfVrTATftU.Vm'axaTATATTTaCTTAt-rm'GAT Leersia virginica Halcebrunia peciolaca G Ehrbarta calycina Screpcogyna anericana 1 screpcogyna americana 2 Diarrhena obovaca ..c Brachyelytrvan ereccum G Poa pracensls c. .A T Avena saciva Pbaenosperma giobosum ..c ..c .A Joinvillea ascendens .G.. ,c G C. .. Strepcochaeca anguscifolia G c... flnomochloa narancoidea ,c G c... r Phaxus lacifolius .c G c... Pharus lappulaceus c G c... Lepcaspis banksii r . n c. _ . Puelia olyrifonais G BuergersiochXoa bambusoides Pariana radiciflora a Eremicis sp. nov. r Olyra lacifolia Lichachne bumilis c... Lichachne pauciflora c... Sucrea maculaca Raddia discichophylla Arundinaria gigancea Pseudnsasa japotiica Arapelocalamus scandens Bashania fargesii ShibaCaea kunasaca Chimonobacibusa naxmorea PhyllosCachys pubescens PhyllosCachys bambtisoides Sasa variegaca Fargesia robusca Yusbania exilis Glaziophycon mirabile •A Alvimia gracilis • A, Chusquea lacifolia Chusquea circinata Neurolepis aperta Archroscylidium ecuadorense •A, Apoclada simplex •A, Rhipidocladui piccieri .A, Ocacea acuminaca ,A, Guadua paniculaca Melocaima baccifera Bambusa aff. banbos Bambusa scenoscachya Cephaloscachyum pergracile SchizosCachyum luzonicum Hacemobanbos nicrophylla Hiclcelia nadagascariensis Nascus elaCus Sporobolus indicus • A A. Zoysia macrella • A A. Zea nays • A Sorghum bicolor • A Phragmices auscralis -A Molinia caerulea -A 146

1330 1350 1360 1370 1380

Oryza sativa GGGlAri'iXXGTCTTCAiiVl'i'CAAAATTACAgrAGTACTAAAGAGGA- - -TTCGTTGTAT Leers ia virgixiica .A. Malcebrunia peciolata .A. Ehrharta calycina StrepCogyna americana I Streptogyna americana 2 Dlarrhena obovata c. G .., .C. Bracfayelytrum ereccim ...c c. AG . Poa pratensis ...cc. G .., .C. Avena sativa ..c T. c. G .C.C., Phaenosperzna globosxim A.... c. G . Joinvillea ascendens c. r TAGTTC. .c Streptochaeta angustifolia c. AG , .C... Anomochloa marantoidea G c. AG . .C... Pharus latifolius c. AG ... .A. .C. .. Pharos lappulaceus c. AG . .C... Leptaspis hanksii c. AG Puelia oXyriformis c... c. G Buergersiochloa banbusoides c.. A ...AC. G Pariana radiciflora c.. T....AC. .... G-— Eremicis sp. nov. c.. ...AC. G , Olyra latifoXia cc.. .G ...AC. .c.. G , Lithacbne humilis ..A. .CC...... G ...AC. .A.. G , Lithachne pauci flora ..A..CC.. .G ...AC..A. . G . Sucrea macuXaCa CC.. .G ...AC. ..C. .AG , Raddia distichopbyXXa CC.. .C.AC. AAG , Anindinaria gigantea CC.. r G . PseT.idosasa japonica .. ..cc.. r G . AznpeXocaXaznus scandens ....C. G . Bashcuiia fargesii CC.. C- G . Shibataea kumasaca cc.. c. G . Chimonobainbusa marxnorea cc.. c. G . PhyXXcscachys pubescens cc.. c. G . PhylXostachys bambusoides cc...... c. .... C.G . Sasa variegaca .cc. G .. Fargesia robusta .cc. G .. Yushania exilis .cc. G .. Glaziophyton nirabiXe .cc. G .. AXvimia graciXis G .. Chusquea latifoXia cc.. c G1 .. Chusquea circinaca cc.. c 'G .. NeuroXepis aperta cc... .c AC Gt C. ArthrostyXidium ecuadorense cc.. c 'G .. ApocXada simpXex G .. RhipidocXadum pictieri .cc. G .. Ocatea acuminata .cc. G .. Guadua panicuXata .cc. G .. HeXocanna baccifera .cc. G .. Banbusa a£f. bambos .cc. G C. Banbusa stenostachya .cc. G C. CephaXostachyum pergraciXe .cc. G .. Schizostachyum luzonicuxn .cc. G .. Hacemobaobos nicrophylla .cc. G—C. HickeXia madagascariensis .cc. G—.. Nastus eXatus ..c. G—.. SporoboXus indicus C.A.G .. Zoysia macreXXa ..A.G .. Zea mays ..A.G .. Sorghum bicoXor ..A.G .. Phragmites austraXis ..A.G .. MoXinia caeruXea ..A.G .. 147

1390 1400 1410 1420 1430 1440

Oryza sativa Leersia virginica ..u ...... A HalCebrunia pecialata ..c .A Ehrharca calycina ..T ...... C .c. .c A.A... .T Strepcogyna americana 1 ..C .c .c... ..G...... SCreptogyna americana 2 ..c .c.c... ..G... Diarrhena obovaca ..c .c .c... ..G... .G... Brachyelytrum erectum ..c ...... c .A. ..iV.GT.G...... T Poa pracensis ..c .A... .c... .G... . .T Avena sativa ..C..G... .c .c... ..G... .G.G. Pbaenospenna globosum ..c G.....c .c... ..G... .G... Joinvillea ascendens ..TC ..A.. .c .c... .G... ..C Screptocbaeca angustifolia ..T ..A...G .c...... G. Anonachloa narancoidea ..T ..A.. .c...... C.. ...G...... Fharus lacifolius ..C ...... C .c... A .G... Pharus lappulaceus ..C ...... C .c... A .G... Lepcaspis banksii ..c ...... c .c... A ...... G.. . Puelia olyriformis ..c ...... c .c...... Buergersiochloa bambusoides ..c ...G. .c .c... ..G... ..A Pariana radiciflora ..c .c .cc.. A ..G... ..A EremiEis sp. nov. ..c .c .cc.. A ..G...... A. .G Olyra lacifolia ..c G.... .c C.G. ..G... ..A LiChacfane humilis ..c ...c...... c... A ..A... .G... .TAG lathachne pauciflora ..c ...c. c... A ..A.. . .G... .TAC Sucrea maculaca ...... c .c c... ..G.. . .G... ..A Raddia discicbophylla ...... c ...... c c... ..G... .G... ..A Arundinaria gigancea ..c ...... c c... ..G...... A Pseudosasa japonica ...... c .c c... ..G...... A Ampelocalaimis scandens ..c .c c... ..G... ..A Bashania fargesii ..c .c c... ..G...... A Shibacaea tounasaca ...... c .c c... ..G...... A Chimocobambusa marmorea ...... c .c c... ..G...... A Phyllostacbys pubescens ..c c c... ..G... ..A Phylloscacbys bambusoides ...... c .c c... ..G.. . ..A Sasa variegata ..c c c... ..G...... A Fargesia robusta ..c ...... c c... ..G.. . ..A ¥usbania exilis ...... c .c c... ..G...... A Glazicphyton mirabile ..c .c c... ..G... ..A Alvimia gracilis ...... c .c c... ..G...... A Chusquea lacifolia ..c .c c... ..G... ..A Oiusquea circinaca ..c ...... c c... ..G...... A Neurolepis aperta ..c ...... c c... ..G... ..A.. ..A Arthrostylidium ecuadorense ..c .c c... ..G... ..A Apoclada simplex ..c c c... ..G... ..A Hbipidocladum pitcieri ..c A c ..G... ..A Ocacea acuminata ..c c c... ..A... ..A Guadua peiniculata ..c ..... c c... ..G...... A Melocanna baccifera ..c c c... ..G...... A Bambusa aff. bambos ..c c c... ..G...... A Bambusa stenostacbya ..c c c... ..G...... A Cephalostacbyum pergracile ..c c C. ../ ..G...... A Scbizostacfayum luzonicum ..c c c... ..G... ..A Bacemobambos microphylla ..c c c... ..G... ..A Hickelia nadagascariensis ..c c c... ..G...... A Hastus elatus ..c c c.. . ..G... ..A Sporobolus indicus ...c.. . .c ..C.. c CG...... T Zoysia matrella ...c... .c ..C. . c c... .T Zea mays ..c ..T.. c... .G .T Sorghum bicolor ..c ..T. . c... .C... .T Phragmites australis ..c ..C.. c c... .G... .T Molinia caerulea 148

1450 1460 1470 1480 1490 1500

Oryza sativa ^C^CGAAGAGTG<^'-rriYi-i-rr^ .A Leers ia virgiziica Kaltebrunia petiolata .. .A A Ehrheirta calycina .. .A.C A A ..C.AG. Streptogyxia americasa I G..A A G. Streptogyna americazia 2 G..A A G. Oiarrbena obovaca ...A A .. .A.G. Brachyelytnun erectum ...A C A G. Poa pracensis .G.A.A A AG.C.AC ...A.G. Avena sativa ...A.A A T A..C ...A.G. Phaenospenna globosum ...A A AA G. JoinvilXea ascendens .. .AT.. .T.A C..C A AC G...... C.A. Strepcochaeta angustifolia ...A T AC G... .A.G.C. Anomochloa marancoidea .. .A T A A GA GG G..A -GG...A. Phariis lacifoliiis G..T T AC G... C G. Pharus lappulaceus G..T T AC G... C G. Leptaspis banks ii G..T T AC G... C G. Puelia oXyriformis ...A A G. Buergersiochloa bambusoides ...A.A T A G. Pariana iradiciflora ...A G A A G..A.G. Erenicis sp. nov. ...A G C.A A G..A.G. Olyra latifolia ...A A .. .G.G. Lithachne humilis ...C A .A.G.G. Lithachne pauciflora ...C A .A.G.G. Sucrea maculaca ...A A ...G.G. Haddia distichopbylla ...A A A ..GG.G. Arundinaria gigantea ...A A G. Pseudosasa japonica ...A A G. Ampelocalanais scandens .. .A AG G. Bashania fargesii .. .A AG G. Shibataea kumasaca ...A A G. rh i mnnobambusa mannorea .. .A AG G. Phylloscachys pubescens ...A AG G. PhylloscaclQrs bambusoides ...A AG G. Sasa variegata ...A A G. Fargesia robusca ...A AG G. Yushania exilis .. .A AG G. Glaziopbyton mirabile ...A.A A G. Alvimia gracilis ...A.A A G. Chusquea lacifolia .. .A G.TC ..C. .G. Chusquea circinaca .. .A A A ..C. .G. Neurol^is aperta .. .A A G. Arthroscylidium ecuadorense .. .A.A A .. .C.G. Apoclada simplex .. .A C A G. Rhipidocladum pittieri ...A.A A G. Otatea acuminata .. .A C A G. Guadua paniculata .. .A C A G. Helocanna baccifera .. .A A G. Bambusa aff. bambos ...A A G. Bambusa stenostacbya ...A A G. cephaloscachyxim pergracile .. .A A G. Schizostachyimi luzonicum .. .A A G. Racemobainbos microphylla .. .A A G. Hickelia madagascariensis .. .A.A A G. Nastus elatus .. .A A G. Sporobolus indicus T.G.A T A..C TTCTGG.AC..G. Zoysia matrella .. .AT C A..C. .C .TC..G. Zea mays .. .AT A..C.A ..C. .C. Sorghum bicolor ...AT A..C ..C..C. Phragmites australis .. .AT A..C.GC ..C.CG. Holinia caerulea .. .AT A..C.GC ..C. .G. 149

1510 1520 1540 1550 1560

Oryza sativa AATACAGGAAATAGGATAGCXTrcCTTTAGTACTTCATTGGGGACTAAAAACACmTGTC Leersia vlrginica .G.. ..C .AG Malcebninia peciolata ..C .AG Ehrharta calycina ..A... ..G Strepcogyna americana 1 ..A... ..C..T.. .G Strepcogyna axoericana 2 ..A. .. T.. .G Diarrhena obovata ..A... ..C..T.. .G Brachyelycrxan erecCimi ..A. .. C.C..T..AG Poa pratensis ..A... A..C..T.. .G Avena sativa ..AC.. G..C..T.. .G Phaenosperma globosum ..A...... C..T.. .G Joizivillea ascendens .GT.A... ..AT..T..T Streptochaeca angustifolia .C..A...... C..T..AGT.. . AnomochJoa caarancoidea .G..AT.. ...C..T..CGT.. . Pharus lacifolius A... ..AC..T..AG Pharus lappulaceus ..AC..T..AG Leptaspis banlcsli A.. . ..AC..T..AG Puelia olyriformis A...... C..T.. .G Buergerslochloa bambusoides A...... C..T..AG Pariana radiciflora A...... C..T..AG Eremicis sp. nov. A...... C..T..A Olyra lacifolia A... T...C..T..AG Lithachne bumilis A.. . T...CC.T..AG Litfaachne pauciflora A.. . T...CC.T..AG Sucrea maculata A.. . T...C..T..AG Raddia distichophylla A.. . T...CC.T..AG Arundinaria gigancea A...... C..T.. .G Pseudosasa japanica A...... C..T. Ampelocalamus scandens A.. . .C..T.. .G Bashania fargesii A...... C..T.. .G Shibataea kumasaca A...... C..T.. .G rVi i mnnobflTnh*«pa marTDorea A...... C..T.. .G Phyllostachys pubescens A...... C..T.. .G Phyllostachys bambusoides A...... C..T.. .G Sasa variegaca A...... C..T.. .G Pargesia robusta A...... C..T.. .G Yushania exilis A...... C..T.. .G Glaziophyton mirabile A...... C..T.. .G Alvimia gracilis AA.. ...C..T.. .G Chusquea lacifolia A...... C..T.. .G Chusquea circinata A...... C..T.. .G Neurolepis aperta A...... C..T. Arthrostylidium ecuadorense AA.. ...C..T.. .G Apoclada simplex A...... C..T.. .G Rhipidocladum pitcieri A...... C..T.. .G Otatea acuminaca A...... C..T.. .G Guadua paniciilata A...... C..T.. .G Melocanna bacciCera A...... C..T...GT... Bambusa a£f. bambos A...... C..T...GT... Bambusa scenoscachya A...... C..T...GT... Cephalostachyum pergracile A...... C..A. ..GT... Schizoscacbyum luzonicum A...... C..T...GT... Racemobambcs microphylla A...... C..T...GT... Hiclcelia madagascariensis A...... C..T...GT... Nastus elatus A...... C..T...GT... Sporobolus indicus A.. . .C.CC.T..AG Zoysia oacrella A.. . .C.C..T. .AG Zea mays A.. . .C.C..T. . .G Sorghum bicolor A.. . .C.C..T.. .G Phragmites auscralis A... .C.C..T.. .G Holinia caerulea A.. . .C.C..T...G 150

1570 1580 1590 1600 1610 1620

Oryza sativa TATCCTCATCJ^^CCGGGAAATaCTATGCTATTTCCTCTIXrrTAaaTTAC'iX^-iUU'GTACT Leersia virginica A Halcebrunia peciolata A G... Ehrharta calycina A T Streptogyna anericana 1 A T Strepcogyna americana 2 A T Dlarrhena obovata G A T T Brachyelytrum erectum A C T Poa pracensis G A T T...7 Avena sativa G A...A T T Fhaenosperma giobosuia G A T Joizxvillea ascendens G...A C G.C.. .G.A. .A.T Streptochaeca angiisci£oIia G...A AT...T Ancmochloa marancoidea A A C C CT...T Fharus lacifolius A C T.C T Pharus lappulaceus A C T.C T Lepcaspis banksii A ????...? Puelia olyriformis A C T Buergersiochloa bambusoides A C T CT Pariana radici flora A.A C T T Eremitis sp. nov. A.A T T T Olyra latifolia A C T CT Lichachne hiimi lis A...A. .G C..C.. .T GCT..T. Litbachne pauciflora A...A. .G C..C.. .T GCT..T. Sucrea maciilata A C T CT Raddia distichopbylla A C T CT..T. Amndinaria gigancea A C T T Pseudosasa japonica A C T T Ampelocalamus scandens A C T T Bashania fargesii A C T T Shibacaea kumasaca A C T T Chimonobainbusa manoorea A C T T Phyllostachys pxibescens A C T T Phyllostachys bambusoides A C T T Sasa variegata A C T T Fargesia robusca A C T T Yushania exilis A C T T Glaziophyton mirabile A T C T T Alviiaia gracilis A T C T T Chusquea latifolia A C..C.. .T T Chusquea circinata A C..C.. .T T Meiurolepis aperta A C..C.. .T T Arthrostylidium ecuadorense A.C T C T T Apoclada simplex A T C T T Hhipidocladum pittieri A T C T T Otatea acuminata A T C T T Guadua paniculata A T C T T Melocanna baccifera A C T T Bambusa aff. bambos A C T T Banbusa stenostachya A C T T CephalostachyuzR pergracile A. CT T T.. . . Schizostachyiim luzonicum A C T T Racemohamhos microphylla A C T T Hickelia madagascariensis A C T T Mastus elatus A C T T Sporobolus indicus GC.. .A T Zoysia matrella A T Zea mays A A T Sorghum bicolor A. A T Phragmites australis A T Itolinia caerulea A T 151

1630 1640 1650 1660 1670 1680

Oryza saciva TTGTTTATTGGMCTATRCGflATCCATTTT GATAATGAA Leersia virginica MalCeiirunia peciolaca Ehrfaarca calycina c... c GATAAT- GGAtSTAATT...... G SCrepCogyna americana I c... c GAIAAT- GGAGTAATG...... G Screpcogyna americana 2 c . . .GAIAAT- GGAGTAATG .. .G Diarrhena obovaca c... TC .TC GATAAT- GGAGCAATG...... G Brachyelytnan erectum c GATAAT- GGAGTAATG. .. . .CG Poa pracensis ..A. .C... T .TC GATAAT- GGAGCAATG...... G Avena sativa c... TC .TC GATAAT- GGAGCGACG...... G Pbaenaspezna globosum c. TC ..C... .GATAAT- GGAGCAATG...... G Joinvillea ascendens ...C.CG.. A T ...G screptochaeca anguscifolia Anomochloa marancoidea Phaxus lacifolitis Pharus lappulaceus Lepcaspis banksii Puelia olyriformis c... C GATAAT- GGAGTAATG...... G Buergersiochloa bambusoides c... C GATGAT- -GTAGTAATG...... G Pariana radiciflora c... c c GATAAT- GGAGTAATG...... G Eremicis sp. nov. c... c GATAAT- GGAGTAATG...... G Olyra lacifclia c GATAAT- GGAGTAATG... Lichacfane humilis ..T c GATAAT- GGAGTAATT.. . Lichacfane pauciflora c... raVTAAT- «3AGTAATT. .. Sucrea maculaca .C. .. GATAATAATAATGGAC3TAATG.. . Raddia discichophylla c... nATaATAATAATGGAGTAATG... Arundinaria gigancea c... c GATAAT- GGAGTAATG...... G Pseudosasa japonica c... c GATAAT- GGAGTAATG...... G Anpelocalanus scandens c... c GATAAT- GGAATAATG...... G Basfaania fargesii c... c GATAAT- GGAGTAATG...... G Shibataea kunasaca c... c GATAAT- GGAGTAATG...... G Chinonohamhusa mannorea c... c GATAAT- GGAGTAATG...... G Phyllostachys pxjbescens c GATAAT- GGAGTAATG...... G PbyllosCachys bambusoides GATAAT- GGAGTAATG...... G Sasa variegaca c GAIAAT- GGAGTAATG...... G Fargesia robusCa c GATAAT- GGAGTAATG...... G YUshania exilis .GATAAT- GGAGTAATG. .. Glaziophyton mirabile .GATAAT- GGAGTAATGA. . .. -G Alvimia gracilis c... c GATAAT- GGAGTAATGA.. .. .G Chusquea lacifolia . .A. .C... c GATAAT- GGAGTAATG...... G Chusquea circinaca ..A..C... c GATAAT- GGAGTAATG...... G Neurolepis aperta c GATAAT- GGAGTAATG.G. .. .G Archroscylidium ecuadorense . ...TCTAAT- GGAGTAATGA. . .. .G Apoclada simplex C. .. AC GATAAT- GGAGTAATG... .A.G RMpidocladum piccieri c... c A GATAAT- GGAGTAATGA. . .. .G Ocacea agiminata c... c GATAAT- GGAGTAATG... .A.G Guadua paniculaca c... c GATAAT- GGAGTAATG... .A.G Kelocanna baccifera c GATAAT- GGAGTAATG...... G Bambusa aff. bambos . ...GATAAT- GGAGTAATG...... G Bambusa stenoscachya c... c GATAAT- GGAGTAATG...... G Cephaloscachyum pergracile c... c GATAAT- GGAGTAATG...... G Schizoscacfayum luzonicum c... c GATAAT- GGAGTAATG...... G Racemobambos microphylla c... c GATAAT- GGAGTAATG...... G Hickelia oadagascariensis c... c ..T GATAAT- GGAGTAATG...... G Nascus elacus c c.... GATAAT- GGAGTAATG... Q Sporobolus indicus c... c GATAAT- GGAGTAATA...... G Zoysia macrella . GATAAT- GGAGTAATA...... G Zea mays GATAAT- GGAGTAAAG...... G Sorghum bicolor GGAGTAAAA Phragmices auscralis c... c GATAAT- GGAGCAATG...... G Molinia caerulea c... c GATAAT- GGAGCAATG...... G 152

1690 1700 1710 1720 1730

Oryza sativa ATAGGGGAATTRACCATATTATCAAAGTtXXrTAACTCCCTCAATCAAC'I'ri'rit.CAAGAA Leersia virginica Malcebrunia peciolaca Ehrharta calyci na Strepcogyna axoericana 1 Streptogyna americana 2 .G... .G.. Diarrhena obovata .C. , .G...... T A AC...G.. Brachyelytrum ereccum .C. , .G...... T A G.. Poa pratensis .CA. ...T A GACT..G.. Avena saciva ...T A AC...G.. Phaenospezna globosxm .c.. Joinvillea ascendens G..AT.. .G...C A A Streptochaeta angusti£oXia G..ATA. .G...T A...T A Anomochloa marantoidea ...AT. . .G... A...T AG A Phaxus latifolius G..AT. . .G... A A Pharus lappulaceus G..AT.. A A Lepcaspis banJcsii G..AT.. A ..A Puelia olyriformis G...T.. ..A Buergersiocbloa bambusoides ...C.. ..A Pariana radiciflora ...CA. ..A Erendtis sp. nov. ...CA. ..A Olyra latifolia ...C.. ..A Lithachne huxnilis ...C.. ..A Lichachne pauciflora ..A Sucrea maculaca ...CA. .A Raddia distichophyXla ...CA. .A Aruzidinaria gigantea ...C.. .A Pseudosasa japoxiica ...C.. .A..T. Afflpelocalamus scandens ...C.. .A Bashania fargesii ...C.. .A Shibacaea kumasaca ...CA. Chimonobanibusa oazinorea .A Phyllostachys pubescens ...C.. .A Phyllostachys bambusoides ...C.. .A Sasa variegata ..AC.. .A..T. Fargesia robusca ...C.. .A Yusbania exilis ...c.. .A Gleiziopfayton mirabile ...CA. .A Alvimia gracilis ...C.. .A Chusquea latifolia ...C.. .A..T. Chusquea circinaca ...c.. .A..T. Neurolepis aperca ...c.. .A..T. Arthroscylidium ecuadorense ...c.. .A Apoclada simplex ...C.. .A Rhipidocladum pictieri ...c.. .A Otacea acuminata ...c.. .A Guadua paniculaca ...c.. .A..T. Helocanna baccifera ...c.. .A.G.. Baabusa aff. banbos ...c.. .A.G.. Bambusa stenostachya ...c.. .A.G.. cephaloscachyum pergracile ...c.. .A.G.. SchizostaciQaim luzonicuzn ...c.. .A.G.. Racemobanbos microphylla ...c.. .A.T.. Hickelia xnadagascariensis ...c.. .A.G.. Nastus elacus ...c.. .A.G.. Sporobolus indicus C.TC.. .A.G.. Zoysia matrella C.TC.. .A.GT. Zea mays ..TT.. .A sorghum bicolor ..TT.. .A Phragmites australis ..TC.. .A Molinia caerulea ..TC.. .A 153

175Q 1760 1770 1780 1790 180

Oryza saciva AGTTCTAATTCGTCCATAAATrCATATGAATTTAICACTAATGCAATTTCTTCrGTAAGT Leersia virginica T Malcebrvmia petiolata T Ehrharta calyciua C..T C Streptogvna americaua 1 T C Streptogyna americana 2 T C Diarrhena obovaca T..T T BrachyelyCrum ereccum .C T..T Poa pracensis T.TT T G Avena saCiva T T.T A Phaenosijenna globosum C TT-.T C Joinvillea ascendent ...C.G T.-T.G G T.A..C G C.-C Screptochaeca angustifolia TA T G T....C G C Anomochloa marantoidea TG T G T.A..C G A—CA Pharus lacifolius ....TG T G T....C G C..C Pharus lappulaceus T T G T C G C LepCaspis banksii ....TG T G T....C GG C Puelia olyriformis T C—C Buergersiochloa bambusoides T T G Pariana radiciflora T C Eremitis sp. nov. T C Olyra laciEolia T C Uchachne hiimilia .A—G T C T Licbachne pauci£lora .A...G T C T Sucrea niaculaca G T T...T Raddia discichophylla T T—T Arundinaria gigancea T Pseudoscisa japonica T Anpelocalamus scandens T C Bashania fargesii T ShibaCaea kumasaca T Chimonobamhusa nwnnorea T Phyllostachys pubescens T Phyllostachys bambusoides T Sasa variegata T Fargesia robusca T Yusbania exilis T Glaziophyton mirabile T Alvimia gracilis T TC Chusquea lacifolia G—T Chusquea circinaca G—T Neurolepis aperta ..C T A Archroscylidium ecuadorense T T T A G Apoclada simplex T Rhipidocladum pictieri T C..G Ocacea acuminata T Guadua paniculaca T Melocaima baccifera T Bambusa af£. bamJaos T Bambusa scenoscachya T CephalosCachyum pergracile T Schizostachyum luzonicum T C Racemobambos microphylla T Hiclcelia madagascariensis T G Nascus elanis T Sporobolus indicus G T Zoysia macrella T Zea mays .A T C Sorghum bicolor .A T C Phragmices auscralis T..T C Molinia caerulea T..T C 154

laio 1820 1830 1840 1850 1860

Oryza sativa CTAGCTAri'in'iUy'iCrATTCRTAGCATATA'l'U'i'i'i''i'ATGGA'lViljCri'ACTC'rTTi'ri'I' Leersia virginica .G. ...C MaXtebrunia peciolata ...C Ehrharta calycina ..cc..c...... T... Streptogyna americana 1 ...T... Streptogyna americana 2 ...T... Oiarrhena obovaca .. .C. .C ...T... BracfayBlytrum erectura ...G...... T... Poa pratensis .. .CA.C ...C..A...... T... Avena saciva ...C..C ...T... Pbaenospezsia globosum ...T... Joinvillea ascendens . .G ...T..A...... TG.. Streptochaeta angustifolia . .G ...T..A Anomochloa marancoidea . .G ...T. .A Pharus Xacifolius . ..T. .A Pharus lappulaceus ...T..C... ,..T..A Leptaspis baxiJcsii ,..T..A Puelia olyrifonais ...c..c...... T... Buergersiochloa bambusoides ...G.. ...T..C Pariana radici flora ...G.. ...c..c... ,..T..C Eremicis sp. nov. ...G.. ...T..C Olyra latifolia ...G.. ,..T..C Lithachne huxnilis ...A. .G ,..T.TC Lithachne pauci£Iora .. .A. .G ...T.TC Sucrea maculaca ...G... C Raddia discicbophylla ...G...... T..C Amndinaria gigancea ...G...... T..C Pseudosasa japonica ...G... ..T..C Ampelocalamus scandens ...G... ..T..C Bashaxiia £argesii ...G... ..T..C Shibataea ktrmasaca ...G... ..T..C Chimonobambusa marmorea ...G... ..T..C Phyiloscacbys pubescens ..T..C Phyllostachys bambusoides ...G... ..T. .C Sasa vsuriegaca ...G... ..T..C Fargesia robusta ..CG... ..T. .C Yushania exilis ...G... . .T. X Glaziophyton mirabile ...G... ..T..C Alvimia gracilis ...G... ..T..C Chusquea latifolia ...A. .. ..T..C Chusquea circinata .. .A. .. ..T..C Neurolepis aperca ...G... ..T..C Arthroscylidium ecuadorense ... G. .. ..T. .C Apoclada simplex ...G... ..T..C Rhipidocladum pictieri ...G... ..T..C Otatea acuminaca ...G... ..T..C Guadua paniculata ...G... ..T..C Melocanna baccifera ...G... ..T..C Bambusa aff. bambos ...G... .-T... Bambusa scenoscachya ...G... ..T... Cephalostachyum pergracile ...G... ..T. .C Schizoscachyum luzonicuni ...G... ..T..C Racemnhamhos microphylla ...G... ..T... Hickelia madagascariensis ...G... ..T..C Nastiis elatus ...G... ..T..C Spcrobolus icdicus C.A.GC... ..T. . Zoysia natrella ...C... C.A.GC... ..T. . Zea mays A ..c ..T. . Sorghum bicolor A ..c ..T. . Phragmices aiistralis G . .T. . Holinia caerulea G ..T. . 155

1870 1880 1890 1900

Oryza saciva CAGAATTTGGA- -AAAGGGGGTCCGAAA Leersia virginica ..C. Maltebnmia peciolata ..C. .. .A.. . Ehrharta calycina ...A.A. Screpcogyna axoericana 1 RAA, Strepcogyna americana 2 HAA Diarrhena obovata AA. Brachyelytrum erecaan ...C A Poa pracensis .TAC- .G.AAA... Avena saciva .TAC- .G.AAA... Phaenospema globosisn A... Jolnvillea ascendens .C..C. .GC. ..AA.A... Strepcochaeca anguscifolia ..c. ..AAAA... Ancnochloa oarancoldea ..AA.AA.. Pharus lacifolius AA... Pharus lappulaceus AA.. Lepcaspis banlcsii AAA.. Puelia olyriformis .TTTAATAAA. .T.C. AA.. Buergersiochloa bambusoides ..TC. ..AA.A... Pariana radiciflora .AT.- ..TC. ..AA.A... Eremitis sp. tiov. .AT.— ..TC. ..AA.A. .. Olyra laci£olla ..TC. ..AA.A. .. Lithachne humilis .TTC. . .AA.AA.. Lithacfane pauciflora .TTC. ..AA.AA.. Sucrea maculata .TTC. ..AA.A... Raddia discichophylla ..TC. ..AA.A... Arundinaria gigancea ..TC. ...A.A... Pseudosasa japonica ..TC. ...A.A... Ampelocalamus scauidens ..TC. ...A Bashania fargesii ..TC. ...A Shibataea kumasaca ..TC. .. .A.A... rh mnni^hamVii i marffiorea ..TC. .. .A Phylloscachys piibescens .TC. .. .A Phylloscachys bambusoides .TC. .. .A Sasa variegaca .TC. .. .A.A. .. Fargesia robusca .TC. .. .A Yushaiiia exilis .TC. .. .A Glaziopfayton mirabile .A. . ..C ...A.A... Alvimia gracilis .A. . ..C ...A.A... Chusquea lacifolia .A. . . . .A. ..C A...T Chusquea circinaca .A. . . . .A. ..c ...A.A...T Neurolepis aperca .TC ...A.A... Archroscylidium ecuadorense .A. . ...A.A... Apoclada simplex .A. . .T.C. ...AAA... Rhipidocladum pittieri .A. . ...C. ...A.A... Ocacea acuminata .A. . .T.C. ...A.A. .. Guadua paniculaca .A. . .T.C. ...AAA... Melocanna baccifera .CT. .. .A.A. .. Bambusa aff. bambos .AT. .. .A. . .C. ...A.A... Bambusa stenoscachya .AT. .T.A. . .C. .. .A.A. .. Cephalostachyum pergracile .CT. . .C. .. .A.A. .. Schizoscachyum luzonicum .AT. ..C. .. .A.A. .. Racemobambos microphylla .AT. .T.A. ..C. ...A.A... Hickelia madagascariensis .AT. ...A.A... Nascus elacus .AT. ..C. ...A.A. .. Sporobolus indicus . .A. .. .A. .TC. .T. •C.GT. .AA.. CC Zoysia matrella ..A. .C.A. ..C. .T.C- C.G. . .AA.. Zea mays .TCA. .. .AA.AA.. Sorghum bicolor .A. .TCA. .. .C. .. .AA.AA.. Phragmices auscralis .A. .TT. . . ..C. C. .- AA. . btolinia caerulea .A. .TT. . . ..C. C..- . ..AA. .AA. 156

1930 1940 1950 1960 1970 1980

Oryza saciva AAGTArrTTTTCCATCAAC TAAAAAAAAACMaTATAGTTGGTCATATAAICGC Leersia virginica G A T KalCebrunia peciolaca G T Ehrharta calycina .G CGG..AC.G Strepcogyna americana 1 ..C CAG G T Strepcogyna americana 2 ..C C3U3 G T Diarrhena abovaca ..CC TG G C T Brachyelytrun ereccun ..CC G G .G C T Foa pracensis .. .C G G .T C T Avena saCiva TC...C TG..A..A T C Fhaenospema globosum .G.GTCC ...TG. .G . C T Jolnvillea ascendens ..AAC.. C..AG. .G . C...G..C Streptochaeta anguatifolia ..AAG.. ..CGG. .G . G.G C Anonochloa narantoidea ..AAT.. ..CAG. .G . G C T Fharus latifolius ..AACC. ...GG. .G . G C Fharus lappulaceus ..AACC. ...GG. .G . C Lepraspis banksii .. .A.C. ...GG. .G . Puelia olyrifonnis CC. ...GG. .G . -C T Buergersiochloa banbusoides c.. ...GG. .G . -C T Fariana radiciflora ...AC.. ...GG. .G . -C T Eremitis sp. nov. ..TAC.. ...GG. .G . • C T Olyra latifolia ..AAC.. GG. .GTACAAG. • C T Lichacbne humilis ..AAC.. ...C.G. .G . T Lithachne pauci£lora ..AAC.. ..C.G. .G . T Sucrea naculata ..AAC.. ..CGG. GG . -C T Baddia distichopfaylla ...T.. GG. , .C T Arundinaria gigantea .T.. ..CGG. , -C T Pseiidnsasa japonica . .T.... .CGG. . -C T Ampelocalamus scandens ....T.. ...CGG. , • C T Bashania fargesii ...T.. ...CGG. , .0 T Shibataea kunasaca T.. ..CGG. .G . -C T Chinonobanbusa mannorea .T.. ..CGG. .G , -C T Phyllostachys pubescens T.....CGG. , .C T Phyllostachys banbusoides T.. ..CGG. .G . -C T Sasa variegata ...T.. ..CGG. , .C T Fargesia robusca .. T.... .CGG. .G— . -C T Yusfaama exilis T.... .CGG. , .C T Glaziophyton mirabile ..CGG. , .C T Alvimia gracilis ...GG. .G . -C T Chusquea latifolia ..A. -C. ..CGG. .G . • C Chusquea circinata TAC.... .CGG. ^ -C T Neurol^is aperta G. , •T Arthrostylidium ecuadorense ...GG. .G . .C C. -T Apoclada simplex ...GG. .G . • C T Piiipidocladum pittieri ...GG. .G . .C T Otatea acuminata C.. ...GG. .G . • C T Guadua paniculata C.. ...GG. .G . -C T Helocanna bacci£era C.. ...GG. .G . -C T Banbusa a£f. bambos C.. ...GG. .G . • C Banbusa stenostacfaya C.. ...GG. .G . • C T Cephalostachyum pergracile C.. ...GG. .G . .C T Schizostacfayum luzonicum C.. .. .GG. , • C T Racemobanbos microp^lla c.. ...GG. .G . -C T Hickelia madagascariensis ...GG. , -C T Nastus elatus c.. GG. .G , C T Sporobolus indicus ...A.C. G.. .G . Zoysia natrella .. .A.C. ...G...G .. . G G C..C Zea mays ...AGC. ...GG. .G .G . C. -C T Sorghum bicolor ...AGC. .. .GG..G .G . C. -C T Fhragmices australis ...A.C. GG. .A . C C. -C T Holinia caerulea ...A.C. ...GG. .A . C C.AC T 157

1990 2000 2010 2020 2030 2040

Oryza sativa GGTTATATAGATATTTTCTATACTAGGACCTrTACCTTGGGTAIAAGAGGATTAACCGAA. Leersia virginica A C A.. MalCebrunia peciolata A A.. Ehrharta calycina TT CA A A.. Strepcogyna americana 1 GT C A.. Screpcogyna americana 2 GT C A.. Diarrhena obovaca GT C.C Brachyelytrum ereccum G G...GT C G Poa pracensis GTA C.C Avena sativa GT C C Pbaenospema globosum T A—G. C.C T T... Joinvillea ascendens C GTT C..C GC Streptochaeta angustifolia A—GT Anomochloa marantoidea .C.G T ....GT T Pharus lacifolius T T.C G Pharus lappulaceus T T.C G Lepraspis banJcsii C ^ T.C G Puelia olyriformis GT c Buergersiorh1oa baobusoides c Pariana radiciflora c Eremicis sp. nov. ...c Olyra latifolia TTG...... c A.. Lithacfane humilis TTG ...c A.. .T... Lithachne pauciflora TTG...... c A.. .T... Sucrea naculaca TTG...... c A.. Raddia discichophylla TTG c A.. Arundinaria gigancea GT c .T... Pseudosasa japonica GT c .T. .. Ampelocalamus scandens GT ,,.c .T... Bashania fargesii GT ...c .T... Shibacaea kumasaca GT .. .c .T... Chiaonobainbusa marmorea GT c .T. .. Phyllostachys pubescens GT c .T... Phylloscachys banbusoides GT ...c .T... Sasa variegata GT ...c Fcirgesia robusca GT ...c .T. .. Yushania exilis GT - ..c .T.. - Glaziopbycon nirabile GT Alvimia gracilis RT Chusquea lacifolia GT c Chusquea circinata GT c Keurolepis aperta . A. _. GT .. .R Arthroscylidium eoiadorense GT Apoclada simplex GT Rhipidocladum pictieri GT Otatea acuminata A.GT Guadua paniculata GT Helocanna baccifera GTT c Bambusa aff. bambos GT c Bambusa stenostachya GT c Cephalostachyum pergracile GT c Schizostachyum luzonicum GT c Racemobambos microphylla G GT ...c Hickelia madagascariensis GT .T.C Nastus elatus GT.... C.A Sporobolus Indicus GT .T.C A.. . -TG. Zoysia matrella G.. GT .T.C ..TG. Zea mays T.. GTT... .T.C.A A.. . . .G. Sorghum bicolor .T.C.A A.. . . .G, .T. .. Phragmites australis G A.GT T.C A G Holinia caerulea G A.GT T.C A G 158

2050 2060 2070 2080 2090 2100 Oryza sativa CTAACGCAGri'iU"iO:AiaAGGGTGTCftTTGATCGAATTACCAATGGAGTAGGTCl'l'UCT Leersia virginica Halcebrunia peciolaca Ehrharta calycina T Screpcogyna americana 1 ...T.. Strepcogyna americana 2 ...T.. Diarrhena obovaca ...T.. Brachyelytrum erectum T..T.. Poa pratensis ...T.. Avena saciva T..T.. Fhaenospenna globosum ...T.. Joinvillea ascendens T Screpcochaeca angustifolia T..T.. Anomochloa marancaidea T..T.. Pharus laCi£olius A..T.. Pharus lapptilaceus A..T.. Lepcaspis banksii T..T.. Puelia olyriforais T..T.. Buergersiochloa bambusoides ...T.. Pariana radici£lora ...T.. Eremitis sp. nov. ...T.. Olyra lacifolia T..T.. LiChacfane hunilis T..T.. Lichachne pauciClora T..T.. Sucrea maculaca T..T.. Raddia discichophylla T..T.. Arundinaria gigancea ...T.. Pseudosasa japonica ...T.. Aapelocalarmis scandens ...T.. Basbania fargesii ...T.. Shibacaea Icmnasaca ...T.. mannorea ...T.. Phyllostachys pubescens ...T.. Phyllostaci^s bambusoides ...T.. Sasa variegata Fargesia robusta ...T.. Yushania exilis ...T.. Glaziophyton mirabile ...T.. Alvimia gracilis ...T.. Chusquea latifolia ...T.. Chusquea circinaca ...T.. Neurolepis aperta ...T.. Arthrostylidium ecuadorense ...T.. Apoclada simplex ...T.. Rhipidocladum pittieri ...T.. Otatea acuminata ...T.. Guadua paniculata ...T.. Melocanna baccifera ...T.. Bambusa a£f. bambos ...T.. Bambusa stenostachya ...T.. Cephalostachyum pergracile ...T.. Schizostachyum luzonicum Racemobambos microphylla ...T.. HicJcelia madagascariensis ...T.. Nastus elacus ...T.. Sporobolus indicus T..T..A Zoysia matrella T..T..A Zea mays T..T..A Sorghtim bicolor T..T..A Phragmites australis ...T..A Molinia caerulea ...T..A 159

2110 Oryza saciva AGTTTTTGTATRGGAGAA Leersia virginica Halcebnmia peciolaca Ehrharta calycina Screpcogyna americana 1 G, Screpcogyna americana 2 G, Diarrhena obovata Brachyelytrum ereccum Poa pracensis Avena saciva Phaenosperma globostm Joinvillea ascendens screpcochaeca anguscifolia Anomochloa marantoidea Pharus lacifallus Pharus lappulaceus Lepcaspis banksii Puelia olyriformis Buergersiochloa bambusoides Pariana radiciflora Eremitis sp. nov. Olyra lacifolia Lichacbne humilis LiChacfane pauci£lora Sucrea maculaca Raddia discichophyila Arundinaria gigancea Pseudosasa japonica Ampelocalamus scandens Bashania fargesii Shibacaea kumasaca Chimonobambusa mazmorea Phylloscachys pubescens Phyllostachys bambusoides Sasa variegata Fargesia robusta Yushania exilis Glaziophyton mirabile Alvimia gracilis Chusquea lacifolia Chusquea circinaca Neurolepis aperca Archrostylidium ecuadorense Apoclada simplex Rhipidocladum piccieri Otacea aeimi nata Guadua paniculaca Helocanna baccifera Bambusa a££. bambos Bambusa scenoscachya Cephaloscachyum pergracile Schizoscachyum luzonicum Racemobambos micrcphylla Hickelia madagascariensis Nastus elatus Sporobolus indicus Zoysia macrella Zea mays G, Sorghum bicolor G, Phragraites auscralis G, Holinia caerulea G, 160

ACKNOWLEDGMENTS

I am grateful to those who have contributed to this study in many ways. My major professors, Lynn Clark and Jonathan Wendel, and my committee members, Nels Lersten,

Donald Farrar, and Elwood Hart, have ail been exemplary teachers who have helped me broaden my knowledge of plant systematics and evolution, as well as other biological fields. I thank Lyim Clark for recruiting me from China, the Graduate School and Botany Department for providing me the assistantships during my whole smdy program, Jonathan Wendel for allowing me use of his molecular lab facility, and the crew in Jonathan Wendel's lab and other botany systematics graduates for their help and their giving me the priority to use the lab facility and computer at the late stage of my study. The advice and suggestions from Aaron Liston

(Oregan State University), Tom Harrington (Plant Pathology Department, Iowa State

University), Cathy Hsiao (USDA and Utah State University), Daniel Nickrent (Southern

Illinois University), and Mark Hershkovitz (Smithsonian Institution) lead me to escape the fungal trap.

My two Chinese colleagues and friends, Tongpei Yi (Forestry School of Sichuan

Province) and Hsueh Jiarong (Southwestern Forestry College) accompanied me on the trips for collecting specimens and leaf materials in Yurman and Sichuan. Gerald Guala (University of

Florida) assisted me to collect Arundinaria gigantea and Pseudosasa japonica in Florida. Nigel

Barker (University of Cape Town, South Africa), G. Davidse (Missouri Botanical Garden), S.

Dransfield and S. Renvoize (Royal Botanic Garden, Kew), J. Davis (Cornell University), E.

Kellogg (Harvard University), W. Chu (Taiwan Forestry Research Institution), die Botany

Department of Smithsonian Institution, and the Herbarium of Missouri Botanical Garden sent me leaf materials and herbarium specimens. R. Ohnstead (then of University of Colorado), and E. Zimmer (Smithsonian Institution) provided primers for DNA sequencing. T. 161

Harrington and T. Hsiau allowed me to use their unpublished fungal sequence data. Also, D.

Lewis and S. Mahoney made life easier in the Ada Hayden Herbarium (ISC) and Richard Pohl

Conservatory.

I would also like to thank my wife, Ying Fang, a graduate student in the Entomology

Department, for the tremendous help, for her taking care of our lovely but active boy for most of the time, and allowed me spend much more time on my study program.