A Phylogenetic Analysis of Diurideae (Orchidaceae) Based on Plastid DNA Sequence Data

American Journal of Botany 88(10): 1903–1914. 2001.

A PHYLOGENETIC ANALYSIS OF DIURIDEAE (ORCHIDACEAE) BASED ON PLASTID DNA SEQUENCE DATA1

PAUL J. KORES,2 MIA MOLVRAY,3,9 PETER H. WESTON,4 STEPHEN D. HOPPER,5 ANDREW P. BROWN,6 KENNETH M. CAMERON,7 AND MARK W. CHASE8 Natural Heritage Inventory, Oklahoma Biological Survey, University of Oklahoma, 111 E. Chesapeake St., Norman, Oklahoma 73019 USA; 3Department of Botany-Microbiology, University of Oklahoma, 223 Cross Hall, Norman, Oklahoma 73019 USA; 4 Royal Botanic Garden, Mrs. Macquaries Road, Sydney, NSW 2000, Australia; 5Kings Park and Botanical Garden, Perth, Western Australia, Australia; 6Department of Conservation and Land Management, Perth, Western Australia, Australia; 7Cullman Molecular Systematics Laboratory, New York Botanical Gardens, Bronx, New York, USA; and 8Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK 2

DNA sequence data from plastid matK and trnL-F regions were used in phylogenetic analyses of Diurideae, which indicate that Diurideae are not monophyletic as currently delimited. However, if Chloraeinae and Pterostylidinae are excluded from Diurideae, the remaining subtribes form a well-supported, monophyletic group that is sister to a ‘‘spiranthid’’ clade. Chloraea, Gavilea, and Megastylis pro parte (Chloraeinae) are all placed among the spiranthid orchids and form a grade with Pterostylis leading to a monophyletic Cranichideae. Codonorchis, previously included among Chloraeinae, is sister to Orchideae. Within the more narrowly delimited Diurideae two major lineages are apparent. One includes Diuridinae, Cryptostylidinae, Thelymitrinae, and an expanded Drakaeinae; the other includes Caladeniinae s.s., Prasophyllinae, and Acianthinae. The achlorophyllous subtribe Rhizanthellinae is a member of Diurideae, but its placement is otherwise uncertain. The sequence-based trees indicate that some morphological characters used in previous classifications, such as subterranean storage organs, anther position, growth habit, fungal symbionts, and pollination syndromes have more complex evolutionary histories than previously hypothesized. Treatments based upon these characters have produced conflicting classifications, and molecular data offer a tool for reevaluating these phylogenetic hypotheses. Key words: Acianthinae; Chloraeinae; Codonorchis; Diurideae; DNA sequences; matK; Megastylis; molecular systematics; monocots; Orchidaceae; phylogenetic relationships; trnL-F.

Diurideae are a Southern Hemisphere tribe of terrestrial orchids found primarily in Australia, New Zealand, and New Caledonia, but with outlying representatives in Malesia, portions of Oceania, Japan, eastern Asia, and South America. The tribe was established by Endlicher in 1842 to accommodate five Australian genera of orchids. Variously circumscribed, Diurideae have been recognized in many major treatments of the family (Schlechter, 1926; Mansfeld, 1937, 1955; Lavarack, 1971, 1976; Dressler, 1981, 1993; Rasmussen, 1982, 1985; 1 Manuscript received 12 September 2000; revision accepted 15 March 2001. This study would not have been possible without the kind assistance of many people. The authors would like to thank J. M. Veillon at the Herbarium of ORSTOM, J.-J. Villegente, C. Laudereau, and Mr. Guerassimoff in New Caledonia for essential help with collections and information. Mr. T. Chavreau, Service de l’Environment, provided much help with the permitting process. In Western Australia we are deeply indebted to Garry Brockman, Greg Bussell, and Allan Tinker who helped us find critical species. Andrew Perkins, studying with Peter Weston, provided much assistance in New South Wales. Ruth Rudkin kindly provided both specimens and contact with the late Karl Robatsch, who was generous with plants samples from his extensive private collection. Paula Rudall, Head, Anatomical Section, Jodrell Laboratory, Royal Botanic Gardens Kew, and Tom Eggeling, Environmental Planning Officer for the Falkland Islands, provided critical specimens. We also wish to thank the following individuals and institutions for help in obtaining the numerous collections used as sources of DNA for this study: the staff of the Molecular Systematics Section, Jodrell Laboratory, Royal Botanic Gardens Kew; Kingsley Dixon, Kings Park Botanic Garden, Perth Australia; and the members of the Native Orchid Society of Western Australia. This study was funded by NSF Postdoctoral fellowship for Paul Kores, BIR-9508358. 9 Author for correspondence (e-mail: [email protected]).

Burns-Balogh and Funk, 1986; Clements, 1995; Szlachetko, 1995). As delimited in the most widely accepted account (Dressler, 1993), Diurideae include ten subtribes with ;43 genera containing .900 species. Nine of these subtribes (Acianthinae, Caladeniinae, Cryptostylidinae, Diuridinae, Drakaeinae, Prasophyllinae, Pterostylidinae, Rhizanthellinae, and Thelymitrinae) are predominantly from Australia (or Australia, New Zealand, and New Caledonia), whereas the tenth (Chloraeinae) is from South America and New Caledonia. As characterized by Dressler (1993), Diurideae have prominent tuberoids, nonarticulate, convolute, chartaceous foliage, and acrotonic or pleurotonic anthers. However, none of these characters is unique to them, and some representatives of the tribe lack tuberoids or have scale-like leaves. As a result, Diurideae remain difficult to characterize morphologically, and their circumscription and possible affinities within the family have been problematic. Historically, floral characters have been of the highest importance in orchids. Much of the work has focused on the column, specifically on the placement of the pollinia and associated structures with respect to the rostellum and stigma. Most orchid treatments make a distinction between acrotonic and basitonic anthers, but various authors have defined these terms differently. Dressler (1981, 1993) stressed the spatial relationship between the anther and rostellum: acrotonic anthers were characterized as having the rostellum or viscidium associated with the apex of the anther, whereas basitonic anthers have the rostellum or viscidium associated with the base of the anther. Rasmussen (1982), on the other hand, used the

1903

1904

AMERICAN JOURNAL

OF

BOTANY

[Vol. 88

TABLE 1. Specimens used. Names are followed by Chase (MWC) number, deposited at Kew, or Kores & Molvray (KM) number, at OKL; trnLF sequences are represented by the first EMBL accession number, matK sequences are the second accession number. GenBank accession no.a Species

Orchidoideae Diurideae Acianthinae Acianthus caudatus Acianthus cf. elegans Acianthus confusus Acianthus cymbalariifolius Acianthus exsertus Acianthus fornicatus Corybas diemenicus Corybas neocaledonica Cyrtostylis huegelii Stigmatodactylus sikokianus Caladeniinae Adenochilus nortonii Caladenia cairnsiana Caladenia catenata Caladenia falcata Caladenia latifolia Cyanicula gemmata Cyanicula sericea Caladenia barbarossa Elythranthera emarginata Eriochilus cucullatus Glossodia major Leptoceras menziesii Praecoxanthus aphyllus Cryptostylidinae Coilochilus neocaledonicus Cryptostylis cf. subulata Cryptostylis erecta Cryptostylis ovata Diuridinae Diuris laxiflora Diuris sulphurea Orthoceras strictum Drakaeinae Caleana major Chiloglottis trapeziformis Drakaea concolor Leporella fimbriata Lyperanthus serratus Lyperanthus suaveolens Megastylis rara Paracaleana nigrita Pyrorchis nigricans Rimacola elliptica Spiculaea ciliata Prasophyllinae Genoplesium fimbriatum Genoplesium sp. indet. Microtis parviflora Microtis unifolia Prasophyllum brevilabrum Prasophyllum cyphochilum Prasophyllum ringens Thelymitrinae Calochilus campestris Thelymitra carnea Thelymitra ixioides Codonorchideae Codonorchidinae Codonorchis lessonii Codonorchis lessonii Orchideae Coryciinae Corycium carnosum

Voucher number

trnL-F

matK

262KM 322KM 330KM 329KM 565MWC 263KM 564MWC 316KM 247KM 328KM

GBAN-AJ409369 GBAN-AJ409372 GBAN-AJ409370 GBAN-AJ409371 GBAN-AJ409373 GBAN-AJ409374 GBAN-AJ409389 GBAN-AJ409390 GBAN-AJ409399 GBAN-AJ409453

GBAN-AJ309990 GBAN-AJ309991 GBAN-AJ309999 GBAN-AJ309992 GBAN-AJ309993 GBAN-AJ309994 GBAN-AJ310010 GBAN-AJ310011 GBAN-AJ310019 GBAN-AJ310075

265KM 227KM 321KM 305KM 235KM 308KM 307KM 260KM 309KM 566MWC 568MWC 306KM 314KM

GBAN-AJ409375 not yet dep. GBAN-AJ409376 GBAN-AJ409377 GBAN-AJ409378 GBAN-AJ409397 GBAN-AJ409398 GBAN-AJ409406 GBAN-AJ409407 GBAN-AJ409410 GBAN-AJ409415 GBAN-AJ409421 GBAN-AJ409440

GBAN-AJ309995 GBAN-AJ309996 GBAN-AJ309997 GBAN-AJ309998 GBAN-AJ310000 GBAN-AJ310017 GBAN-AJ310018 GBAN-AJ310026 GBAN-AJ310027 GBAN-AJ310028 GBAN-AJ310033 GBAN-AJ310039 GBAN-AJ310057

338KM 332MWC 337KM 336KM

GBAN-AJ409388 GBAN-AJ409395 GBAN-AJ409393 GBAN-AJ409394

GBAN-AJ310009 GBAN-AJ310015 GBAN-AJ310014 GBAN-AJ310016

209KM 273KM 571MWC

GBAN-AJ409403 GBAN-AJ409404 GBAN-AJ409433

GBAN-AJ310023 GBAN-AJ310024 GBAN-AJ310050

266KM 569MWC 834MWC 259KM 312KM 552MWC 326KM 313KM 311KM 268KM 239KM

GBAN-AJ409379 GBAN-AJ409382 GBAN-AJ409405 GBAN-AJ409420 GBAN-AJ409423 GBAN-AJ409424 GBAN-AJ409427 GBAN-AJ409436 GBAN-AJ409448 GBAN-AJ409449 GBAN-AJ409452

GBAN-AJ310001 GBAN-AJ310003 GBAN-AJ310025 GBAN-AJ310038 GBAN-AJ310040 GBAN-AJ310041 GBAN-AJ310044 GBAN-AJ310053 GBAN-AJ310065 GBAN-AJ310066 GBAN-AJ310072

226KM 270KM 553MWC 328KM 272KM 261KM 310KM

GBAN-AJ409413 GBAN-AJ409414 GBAN-AJ409428 GBAN-AJ409429 GBAN-AJ409441 GBAN-AJ409442 GBAN-AJ409443

GBAN-AJ310031 GBAN-AJ310032 GBAN-AJ310045 GBAN-AJ310046 GBAN-AJ310058 GBAN-AJ310059 GBAN-AJ310060

267KM 271KM 269KM

GBAN-AJ409380 GBAN-AJ409454 GBAN-AJ409455

GBAN-AJ310002 GBAN-AJ310076 GBAN-AJ310077

332KM 333KM

GBAN-AJ409386 GBAN-AJ409387

GBAN-AJ310008 GBAN-AJ310007

1132MWC

GBAN-AJ409391

GBAN-AJ310012

October 2001] TABLE 1.

KORES

ET AL.—DIURIDEAE

(ORCHIDACEAE)

1905

Continued. GenBank accession no.a Species

Disinae Disa glandulosa Disperis capensis Monadenia bracteata Pterygodium catholicum Habenariinae Habenaria repens Peristylus ngoyensis Orchidinae Holothrix sp. Ophrys apifera Satyriinae Satyrium nepalense Spirantheae Chloraeinae Chloraea gaudichaudii Chloraea indet. Gavilea australis Gavilea sp. Cranichidinae Cranichis fertilis Ponthieva racemosa Goodyerinae Gonatostylis vieillardii Goodyera viridiflora Pristiglottis montana Zeuxine vieillardii Zeuxine strateumatica Megastylidinae Megastylis gigas Megastylis glandulosa Pachyplectroninae Pachyplectron ariflolium Pterostylidinae Pterostylis longifolia Pterostylis picta Spiranthinae Aa paleacea Sacoila lanceolata Sarcoglottis acaulis Epidendroideae Acineta superba Bulbophyllum longiscapum Dendrobium prasinum Nervilia cf. aragoana Spathoglottis vieillardii Tropidia effusa Eriinae Eria bulbophylloides Neottieae Cephalanthera damasoniana Epipactis helleborine Listera smallii Palmorchideae Palmorchis trilobata Vanilloideae Cleistes rosea Pogonia ophioglossoides Vanilla planifolia

Voucher number

trnL-F

matK

678MWC 1203MWC 216KM Kurzweil 1882 NBG

GBAN-AJ409401 GBAN-AJ409402 GBAN-AJ409430 GBAN-AJ409447

GBAN-AJ310021 GBAN-AJ310022 GBAN-AJ310047 GBAN-AJ310064

381MWC 324KM

GBAN-AJ409418 GBAN-AJ409437

GBAN-AJ310036 GBAN-AJ310054

675MWC 536MWC

GBAN-AJ409419 GBAN-AJ409432

GBAN-AJ310037 GBAN-AJ310049

539MWC

GBAN-AJ409450

GBAN-AJ310070

331KM 551 339KM 340KM

GBAN-AJ409383 GBAN-AJ409384 GBAN-AJ409411 GBAN-AJ409412

GBAN-AJ310004 GBAN-AJ310005 GBAN-AJ310029 GBAN-AJ310030

401MWC 398MWC

GBAN-AJ409392 GBAN-AJ409439

GBAN-AJ310013 GBAN-AJ310056

323KM 327KM 320KM 315KM

GBAN-AJ409416 GBAN-AJ409417 GBAN-AJ409444 GBAN-AJ409459 GBAN-AJ409458

GBAN-AJ310034 GBAN-AJ310035 GBAN-AJ310061 GBAN-AJ310081 GBAN-AJ310080

318KM 317KM

GBAN-AJ409425 GBAN-AJ409426

GBAN-AJ310042 GBAN-AJ310043

529MWC

GBAN-AJ409434

GBAN-AJ310051

264KM 240KM

GBAN-AJ409445 GBAN-AJ409446

GBAN-AJ310062 GBAN-AJ310063

535MWC 342KM 585MWC

GBAN-AJ409368 not yet dep. not yet dep.

GBAN-AJ309989 GBAN-AJ310067 GBAN-AJ310068

106MWC 302KM 304KM 319KM 325KM 301KM

(Whitten et al.) not yet dep. GBAN-AJ409400 GBAN-AJ409431 GBAN-AJ409451 GBAN-AJ409456

not yet dep. not yet dep. GBAN-AJ310020 GBAN-AJ310048 GBAN-AJ310071 GBAN-AJ310078

303KM

GBAN-AJ409409

GBAN-AJ310069

575MWC 199MWC 486MWC

GBAN-AJ409381 GBAN-AJ409408 GBAN-AJ409422

not yet dep. not yet dep. not yet dep.

Dressler s.n. FLAS

GBAN-AJ409435

GBAN-AJ310052

335KM 100KM 334KM

GBAN-AJ409385 GBAN-AJ409438 GBAN-AJ409457

GBAN-AJ310006 GBAN-AJ310055 GBAN-AJ310079

a The prefix GBAN-has been added to link the online version of American Journal of Botany to GenBank but it is not part of the actual accession number.

point of attachment between the viscidium and pollinia: acrotonic anthers have the viscidium attached near the apex of the pollinia and basitonic anthers near the base. Orchids that lack a detachable viscidium cannot be categorized as either

acrotonic or basitonic using the latter definition. Further confusing the issue, some authors recognize an intermediate condition termed mesotonic or pleurotonic (Mansfield, 1955; Dressler, 1993) in which the viscidium (or viscidia) are asso-

1906

AMERICAN JOURNAL

ciated with the middle of the pollinia. In yet another definition of acrotony and basitony, Clements (1995) emphasized the attachment of the anther to the column. According to Clements, the anther is attached to the column via a rudimentary anther filament in the acrotonic condition, whereas in the basitonic condition the individual locules of the anther are attached directly to the column for their entire length. Given the different definitions of acrotony and basitony, characterizing anther types within Diurideae has been problematic. Dressler (1981, 1993) viewed the tribe as having both acrotonic and pleurotonic anthers. Rasmussen (1982) categorized the genera within the tribe with distinct viscidia as having either acrotonic or basitonic anthers, but Arthrochilus, Caladenia, Caleana, Chiloglottis, Cyanicula, Drakaea, Elythranthera, Glossodia, Leporella, Leptoceras, Lyperanthus, Paracaleana, Praecoxanthus, Pyrorchis, Rimacola, and Spiculea all lack a detachable viscidium and could not be categorized. Clements (1995) considered all Diurideae to possess acrotonic anthers. In part because of these differing interpretations, acrotonic, pleurotonic, and basitonic anthers have been reported within Diurideae, leading different authors to interpret this condition as indicative of possible affinities to Spiranthoideae (sensu Dressler, 1981), Neottioideae (sensu Rasmussen, 1982) or both the latter and Epidendroideae (sensu BurnsBalogh and Funk, 1986). More recently, increasing emphasis has been placed on the use of vegetative and embryological characters in higher classification of the orchids (Dressler, 1981; Clements, 1995; Freudenstein and Rasmussen, 1999). Unlike Spiranthoideae and Epidendroideae, but similar to terrestrial, basitonic orchidoid orchids, Diurideae have tubers. The similarity led Dressler (1981) to group Diurideae, Orchideae, and Diseae in the subfamily Orchidoideae. Subsequent work (Kores, 1995; Pridgeon and Chase, 1995), however, has thrown enough doubt on the homology of the tubers in different lineages that these structures are now referred to as root-stem tuberoids (Dressler, 1993). In addition, molecular studies (Cameron et al., 1999; Kores et al., 2000; Kores, in press) have identified lineages basal to Orchidoideae with tuber-forming representatives, raising the possibility that their presence in Diurideae and Orchideae is plesiomorphic or convergent. Reliance on vegetative similarity links the acrotonic (or pleurotonic) Diurideae with the basitonic Orchideae rather than the acrotonic Spiranthoideae (Dressler, 1981, 1993). The mixed affinities of major floral and vegetative characters found in Diurideae led to a number of different placements of the tribe and subtribes. Lavarack (1971, 1976), Garay (1972), and Rasmussen (1982, 1985, 1986) assigned Diurideae to Neottioideae. Dressler (1981, 1993) included their constituents within Orchidoideae. Burns-Balogh and Funk (1986) split Diurideae into three lineages that were dispersed among Spiranthoideae, Neottioideae, and Epidendroideae. Szlachetko (1991, 1995) also divided the tribe, but he included the resulting groups within Orchidoideae and a new subfamily, Thelymitroideae. Clements (1995) excluded two of the subtribes from Diurideae and placed the remainder of the tribe in Orchidoideae. Chloraeinae and Pterostylidinae, the two subtribes he excluded, were assigned to Spiranthoideae (sensu Dressler, 1993). In addition, Clements also suggested that Diurideae and spiranthoid orchids were sister groups, which agreed with the findings of Kores et al. (1997) and Cameron et al. (1999) based on rbcL sequence data. Recent work has used plastid matK sequences to increase

OF

BOTANY

[Vol. 88

the number of phylogenetically informative characters in an effort to better support the relationships of Diurideae (Kores et al., 2000; Kores, in press). This work has confirmed the sister relationship between Diurideae and Cranichideae and hence the need to subsume Spiranthoideae under Orchidoideae. It has also confirmed that the ten former subtribes do not all belong together. The members of Chloraeinae sampled at the time showed affinities to Cranichideae, not Diurideae, as did Pterostylidinae. Within Diurideae, a clade composed of Acianthinae and Prasophyllinae was sister to a lineage composed of two clades. One, the ‘‘core Caladeniinae,’’ was sister to a weakly supported clade composed of a basal lineage of Diuris, Orthoceras, and Cryptostylis followed by members of Thelymitrinae, Drakaeinae, and a few genera formerly placed within Caladeniinae (sensu Dressler, 1993). The classification of the large Australasian genus Caladenia has been problematic. The formal classification by Bentham (1873) remained the only comprehensive treatment until very recent research. Hopper and Brown in Hoffman and Brown (1992, 1998) proposed the reinstatement of Leptoceras (R.Br.) Lindl. and the erection of three new genera: Cyanicula, Praecoxanthus, and Drakonorchis. Our data are discussed at greater length later in this paper and clearly support these proposals except for the status of Drakonorchis, which is embedded within Caladenia. Hopper and Brown (2000) have recently formalized the description of Cyanicula and Praecoxanthus and erected a new subgeneric classification within Caladenia that includes the subgenera Caladenia, Calonema, Drakonorchis, Elevata, and Phlebochila. In earlier work (Kores et al., 2000), the strict consensus tree showed considerable resolution but bootstrap support for some lineages was low, especially so for relationships among major clades of Diurideae. As had happened repeatedly, this difficult tribe required more data to provide a well-supported hypothesis of generic relationships. Two regions, matK and trnL-F, were selected based on our previous work, as well as that of Jarrell and Clegg (1995) for matK and Geilly and Taberlet (1994), Sheahan and Chase (2000), and Chase et al. (2000) for trnL-F. The latter region encompasses the small 39 trnL exon (the only coding portion), the trnL intron, and the trnLF intergenic spacer. The current work reports on results from multiple-gene matrices with increased taxon sampling. METHODS Plants sampled for this study included 95 taxa, representing 53 Diurideae in 31 genera, ten genera of Cranichideae, nine species and four genera of Chloraeinae, ten Orchideae-Diseae, ten Epidendroideae, and three genera of Vanilloideae (Table 1). Total genomic DNA was extracted from fresh or silica-dried plant material using a standard 2X CTAB protocol (Doyle and Doyle, 1987) but purified by ultracentrifugation using a CsCl2-ethidium-bromide density gradient (1.55 g/ mL). Amplification was carried out in a Perkin-Elmer thermal cycler using 100-mL polymerase chain reactions (PCR). For matK, the protocol used 2.5 units of Taq polymerase (Promega, Madison, Wisconsin, USA), 2 mL 4% bovine serum albumin, 2.8 mmol/L MgCl2, and 100 ng of the two PCR primers, matK-19F (Molvray, Kores, and Chase, 2000) and trnK2R (Johnson and Soltis, 1994, 1995). The PCR profile included an initial premelt of 2 min 30 sec at 948C, followed by 28 cycles of 1 min denaturation (948C), 1 min annealing (528C), and 2 min 30 sec elongation (728C) with 8 sec added per cycle, followed by a 7 min final extension at 728C. The two PCR primers for trnL-F were c and f (Geilly and Taberlet, 1994). The PCR protocol differed from that of matK in that magnesium concentration was reduced to 2.5 mmol/ L. The trnL-F PCR profile included an initial premelt of 2 min at 948C,

October 2001]

Fig. 1.

KORES

ET AL.—DIURIDEAE

(ORCHIDACEAE)

1907

Bootstrap consensus tree based on matK sequence data.

followed by 30 cycles of 1 min denaturation (948C), 30 sec annealing (508C), and 1 min elongation (728C); and ending with a 7 min final extension at 728C. The PCR products of both regions were purified using the Wizard PCR purification columns (Promega) following the manufacturer’s protocols. Cycle sequencing was carried out directly on the purified PCR product using the ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems, Foster City, California, USA), with 10 ng of primer, 3 mL of sequence dilution buffer, and 2 mL of cycle sequence mix in a 20mL reaction volume. For trnL-F, primers c and f (Geilly and Taberlet, 1994) were used for all taxa, supplemented by the internal forward and reverse primers (e and d) in templates with long homopolymer A or T regions. Six primers were used to sequence the complete matK–trnK region. These included the two PCR primers, three internal forward primers, and one internal reverse primer. The primer sequences are listed in Molvray, Kores, and Chase (2000). Cycle sequencing conditions for both regions were as follows: 26 cycles of 15 sec denaturation (968C), 1 sec annealing (508C), and 4 min elongation (608C) using a Perkin-Elmer 9700 thermal cycler. Sequencing reactions were purified by ethanol precipitation and run on an ABI Prism 377 automated sequencer (Applied Biosystems). Electropherograms were assembled and edited with Sequencher 3.1 software (GeneCodes, Ann Arbor, Michigan, USA). Sequences were aligned manually. Two regions in noncoding trnL-F and two in the matK–trnK intergenic spacer have sufficiently high levels of variability between major clades that homology of the sequences at those positions cannot be assumed with confidence. The regions were excluded a priori from analysis. They were found in the aligned matrix at positions 1852–1895 and 2080–2115 in matK-trnK and 128–206, and 1478– 1564 in trnL-F, for a total of 245 characters or 6.3% of the 3881 character data set. (Inclusion of these positions turns out not to affect topology, but does reduce bootstrap support slightly on some clades.) In addition, the trnLF intron has an indel of highly variable length, which is absent in many Diurideae, but can be as long as 292 bp, for instance in Dendrobium. Alignment indicative of homology could not be established between subfamilies or orchidoid tribes. The indel was placed between positions 574 and 874, and the whole region was excluded both from analysis and totals of numbers of characters.

Fig. 2.

Bootstrap consensus tree based on trnL-F sequence data.

Phylogenetic analyses were performed on the two data matrices separately and in combination using PAUP* 4.02 (Swofford, 1998). Starting trees were obtained using random sequence addition, searched using equally weighted maximum parsimony (Fitch, 1971) with tree bisection-reconnection (TBR) branch swapping and MULPARS and Steepest Descent in effect. Tree limit was set at 5000. Positions where the gaps occurred were treated as missing data (Swofford, 1993). Internal support was assessed by bootstrapping (Felsenstein, 1985), using equally weighted characters. Bootstrap percentages (BP) for each node were computed after resamplings followed by a maximum parsimony (MP) reconstruction (bootstrap option in PAUP* 4.02b, with 500 replicates of heuristic search, one random sequence addition per replicate, TBR branch swapping and maxtrees 5 100). The matK and trnL-F trees bootstrap consensus trees (Figs. 1 and 2) did not exhibit hard incongruence (Seelanan, Schnabel, and Wendel, 1997), so the data sets were used in combined analyses. Increased resolution coupled with higher levels of bootstrap support were taken as another indication that phylogenetic signal was compatible between data sets and that direct combination was valid (Olmstead and Sweere, 1994). Three vanilloid genera were designated as outgroups for all analyses, based on past work (Kores et al., 1997, 2000; Cameron et al., 1999).

RESULTS The two regions of the genome used in this study appear to have similar properties (Table 2). Both are characterized by

1908 TABLE 2.

AMERICAN JOURNAL

OF

BOTANY

[Vol. 88

Tree and variability statistics associated with the data matrices used in the present study. Sequence characteristic

Combined matK-trnL-F

matK

trnL-F

Number of characters Invariant characters Uninformative variable characters Parsimony-informative variable characters Consistency index Retention index Maximum parsimony tree length Number of trees Pairwise sequence divergence (range in %)

3636 1483 730 1423 0.48 0.65 6982 8 2–14

2097 830 422 845 0.45 0.65 4303 8 1–15

1539 653 308 578 0.56 0.70 2610 .5000 3–18

numerous indels, have several short homopolymer A and T regions, and display a similar range of pairwise sequence divergence values for the taxa sampled here. This similarity could be due to the likelihood that matK is a pseudogene in Orchidaceae (Kores et al., 2000). The matK region has a similar level of substitution for all three codon positions, stop codons are found within it, and indels occur that create reading frame shifts. Dissimilarities are that the trnL-F region appears to have more indels than matK, and some indels are much larger than in matK. The matK–trnK region sequenced included 2097 characters after alignment. Of these, 1267 were variable and 845 were potentially parsimony informative. The analysis yielded 8 equally parsimonious trees with a length of 4303 steps, a consistency index (CI) of 0.45, and a retention index (RI) of 0.65. The topology obtained using matK (Fig. 1) was similar to that obtained in a combined matK–trnL-F analysis (Figs. 3 and 4), but had less resolution and lower bootstrap support among some of the lineages.

The trnL-F region sequenced included 1539 characters after alignment, of which 886 were variable and 578 were potentially parsimony informative. Analyses based on trnL-F alone produce .5000 most parsimonious trees with a length of 2610 steps, a CI of 0.56, and an RI of 0.70. These trees are characterized by a large polytomy (Fig. 2). Despite the lack of resolution within monandrous orchids based on trnL-F, the region does provide additional support for certain clades when combined with matK, most importantly from our perspective in Diurideae. The combined matK–trnL-F data matrix comprised 3636 characters, of which 1423 were parsimony informative. Uncorrected pairwise distances between taxa from different genera based on the combined matrix ranged from a minimum of 2% to a maximum of 14%, with most of the values falling within a 7–12% range. Analysis of the combined matrix using Fitch parsimony yielded eight equally most parsimonious trees (Figs. 3 and 4). These trees had a length of 6982 steps, a CI of 0.48, and an RI of 0.65.

Fig. 3. Tree 1 of 8 based on combined matK and trnL-F sequences, nondiurid taxa. Nodes that collapse in a strict consensus are indicated with arrows. Branch lengths are shown above the branches; bootstrap support percentages below the branches. Percentages $83% are shown in boldface.

October 2001]

KORES

ET AL.—DIURIDEAE

(ORCHIDACEAE)

1909

Fig. 4. Combined matK and trnL-F tree continued, diurid taxa. Nodes that collapse in a strict consensus indicated with arrows. Branch lengths are shown above the branches; bootstrap support percentages are shown below the branches. Percentages $83% are shown in boldface.

The topology based on the combined matK and the trnL-F regions bears similarity to trees in Cameron et al. (1999) based on the plastid gene rbcL and in Kores et al. (2000) and Kores (in press) based on matK. Subfamily Orchidoideae sensu lato forms a strongly supported clade (BP 100%) sister to a moderately supported Epidendroideae (BP 78%) (Fig. 3). Within Orchidoideae, three major lineages are evident. The first lineage includes Codonorchis, formerly considered a member of Chloraeinae and Diurideae, and its sister group, the basitonic orchids (BP 87%). The basitonic Orchideae form a strongly supported clade (BP 99%) that includes Diseae sensu Dressler (1993). The spiranthid orchids form the second major lineage (BP 99%), which includes Cranichideae, Megastylidinae, Pterostylidinae, and core Chloraeinae. The third major clade is composed of core Diurideae (BP 100%). Diurideae are sister to the spiranthid clade (BP 99%), and that combined clade is sister to Codonorchis–Orchideae. Focusing on the spiranthid lineage (Fig. 3), three strongly supported suprageneric subclades are apparent. One of these includes the genera Pachyplectron through Gonatostylis and encompasses Goodyerinae and Pachyplectroninae (BP 100%). The second includes the genera Sacoila through Cranichis and

represents Cranichidinae, Prescottiinae, and Spiranthinae (BP 100%). These two lineages, which include all subtribes of Cranichideae (except the unsampled Manniellinae), are sister to each other, although bootstrap support is currently lacking. Two other genera, Pterostylis and Megastylis pro parte, form a strongly supported group with the Cranichideae (BP 98%), but their precise relationship to each other or to the other members of Cranichideae is uncertain. Clements (1995) suggested Pterostylis should be included among spiranthid orchids based on embryological characters, but otherwise both Pterostylis and Megastylis were formerly regarded as members of Diurideae. In our analysis, Megastylis pro parte is sister to Cranichideae, with Pterostylis sister to the combined cranichid–Megastylis clade, but there is no bootstrap support for either of these relationships. The third subclade within the spiranthid lineage, which is sister to all the other spiranthid orchids, includes the genera Chloraea and Gavilea (BP 100%), both South American representatives of the subtribe Chloraeinae, formerly in Diurideae. In the core diurid lineage (Fig. 4), two subclades are apparent: one includes the genera from Adenochilus to Acianthus and represents the subtribes Acianthinae, Prasophyllinae, and

1910

AMERICAN JOURNAL

core Caladeniinae (BP 80%); the other includes the genera from Orthoceras to Caleana representing the subtribes Drakaeinae, Thelymitrinae, Cryptostylidinae, Diuridinae, and portions of Chloraeinae (Megastylis rara) and Caladeniinae (Rimacola, Lyperanthus, Pyrorchis, Leporella) (BP 87%). Each of these subclades is considered in more detail below. The Acianthinae–Prasophyllinae–core Caladeniinae clade can be further resolved into a moderately supported Acianthinae–Prasophyllinae lineage (BP 75%) that is sister to the core Caladeniinae clade. Acianthinae are strongly supported (BP 100%), but the genus Acianthus appears to be polyphyletic. In this treatment, the three species from Australia, Acianthus exsertus, A. fornicatus (the type species), and A. caudatus, form a strongly supported monophyletic group (BP 100%), whereas the three taxa from New Caledonia, A. confusus, A. elegans, and A. cymbalariifolius, form a grade within the strongly supported clade that includes Stigmatodactylus sikokianus (BP 100%). The latter clade is sister to a moderately supported lineage composed of Corybas and Cyrtostylis (BP 74%). The sister group to Acianthinae, Prasophyllinae, is also strongly supported (BP 97%) and can be further resolved into two lineages, Microtis (BP 100%) and a lineage made up of Genoplesium and Prasophyllum (BP 100%). Core Caladeniinae, sister to the Acianthinae–Prasophyllinae clade, lack bootstrap support and the most basal node collapses in a strict consensus of the eight trees obtained with equally weighted characters. However, the node above Adenochilus is strongly supported (BP 97%). Parenthetically, it is worth noting that Adenochilus is strongly supported (BP 88%) as sister to the Eriochilus–Caladenia clade in internal transcribed spacer (ITS) analysis of Caladeniinae alone (P. J. Kores and M. Molvray, unpublished data), and that this relationship is likely to be supported in further work. Within core Caladeniinae, Eriochilus branches off after Adenochilus, followed by Leptoceras and Praecoxanthus, the last of which is sister to a strongly supported clade comprising Caladenia, Cyanicula, Elythranthera, and Glossodia (BP 100%). Cyanicula, Elythranthera, and Glossodia form a weakly supported clade (BP 63%) that is sister to Caladenia (BP 100%). The other major subclade within the core diurids includes representatives from five subtribes. Four of these subtribes constitute well-supported, monophyletic lineages: Diuridinae (Diuris and Orthoceras; BP 100%), Cryptostylidinae (Cryptostylis and Coilochilus; BP 100%), Thelymitrinae (Thelymitra and Calochilus; BP 100%), and Drakaeinae (Spiculaea to Caleana; BP 90%). Diuridinae are sister to a clade comprising Cryptostylidinae, Thelymitrinae, Drakaeinae, and several genera previously assigned to Caladeniinae. Within that clade, Cryptostylidinae are sister to the combined Thelymitrinae– Drakaeinae–Caladeniinae (pro parte) clade, but the relationship is weakly supported (BP 57%). The Thelymitrinae–Drakaeinae–Caladeniinae (pro parte) clade is well supported (BP 100%). Within this subclade, Thelymitrinae are sister to a well-supported lineage (BP 88%) made up of the genera Leporella, Pyrorchis, Lyperanthus, Rimacola, Megastylis (pro parte), and the representatives of subtribe Drakaeinae (Spiculaea to Caleana). The segregate genera Pyrorchis and Lyperanthus are sister to each other (BP 96%), as are the genera Paracaleana and Caleana (BP 100%). Also of interest is the placement of Megastylis rara near Drakaeinae and Rimacola elliptica with strong support (BP 93%). The position of Rhizanthella could not be resolved in our studies. Unlike some other achlorophyllous orchids (M. Mol-

OF

BOTANY

[Vol. 88

vray, unpublished data), the plastid genome appears to be highly aberrant, and all attempts to amplify the plastid genes used in this study have failed. Recent work by P. J. Kores and M. Molvray (unpublished data) using nuclear 26S and 18S rDNA sequences place Rhizanthella in Diurideae (bootstrap support 93%), but its position within the tribe has not yet been resolved. DISCUSSION Expanded sampling within Diurideae and the addition of trnL-F sequences has clarified relationships in Diurideae and provided strong support for many of the major lineages placed in this problematic tribe. The current study has also clarified relationships in Orchidoideae as a whole and identified a number of polyphyletic groups recognized in previous treatments based on morphological characters. Many of the previous findings based on single plastid genes (Cameron et al., 1999; Kores et al., 2000; Kores, in press) are strongly supported in the current treatment, and some new relationships are reported here for the first time. Our study confirms that Diurideae are not monophyletic unless all of the representatives from the subtribes Chloraeinae (with the exception of Megastylis rara) and Pterostylidinae are excluded. The molecular data strongly support the transfer of the core Chloraeinae (Bipinnula, Chloraea, Gavilea, and Geoblasta) and Pterostylidinae to the spiranthid lineage, a placement supported by a number of morphological characters. All of the Diurideae sensu stricto, with the exception of Cryptostylis, Rimacola, Megastylis rara, Adenochilus, and Rhizanthella slateri, have tuberoids, but these structures are absent in all genera of core Chloraeinae except Geoblasta. In addition, most Diurideae have only one leaf per shoot, but core Chloraeinae, Megastylis, and Codonorchis have shoots with multiple leaves. Pterostylis also has multiple leaves per shoot, although, like Diurideae, the genus has prominent storage organs. The seed coat morphology in Pterostylis is typically spiranthid (Molvray and Kores, 1995). Further, both Pterostylis and the representatives of the core Chloraeinae have a spiranthid embryo pattern (Clements, 1995, 1999). Another strongly supported finding is that subfamily Orchidoideae consists of three lineages, although the composition of these lineages differs somewhat from previous accounts based on single gene regions. The lineages are referred to in this discussion as the Codonorchis–Orchideae clade, the spiranthid clade, and the diurid clade. Only the last of these lineages closely conforms to the currently accepted circumscription for an existing tribe. Codonorchis has already been afforded subtribal status by Brieger (1974–1975), and in view of the molecular data Cribb and Kores (2000) have proposed its elevation to the rank of tribe. However, we feel it would be premature to alter the delimitation of existing tribes within spiranthids or to propose new ones at this time because of our limited sampling within this clade. The placement of Codonorchis sister to Orchideae has not been suggested in any previous classifications for the family. The genus includes from one (e.g., Hawkes, 1965) to three (e.g., Mabberley, 1997) species and is widely distributed in the temperate and subtropical regions of South America. The two specimens used in this study were obtained from Chile and the Falkland Islands, representing the eastern and westernmost limits of its range. Codonorchis has stalked subterranean storage organs similar to those found in some members of Orchideae and like most Orchideae it has several leaves per

October 2001]

KORES

ET AL.—DIURIDEAE

shoot. However, its floral structure is much more similar to the typical diurid rather than the basitonic orchidoid taxa. It has a long slender column without auricles, the anther is clearly acrotonic, the pollinia are not sectile, and only a single viscidium is present. Despite this obvious similarity in floral structure, embryological development in Codonorchis follows a pattern aberrant for a diurid (Clements, 1995, 1999). Its placement sister to Orchideae is moderately well supported in the current treatment, which has a number of interesting implications. The position of Codonorchis sister to Orchideae offers further evidence that the basitonic anther is a derived condition, possibly a rapidly derived condition. Similarly, the presence of tuberoids in Codonorchis, Orchideae, and most Diurideae suggests their absence in most spiranthid orchids is a derived condition. The other two lineages in Orchidoideae are Diurideae and the spiranthid orchids. The sister relationship between them, previously reported by Clements (1995), Kores et al. (1997), Cameron et al. (1999), Kores et al. (2000), and Kores (in press), is strongly confirmed by our molecular data, although there are few morphological characters that form unambiguous synapomorphies for this clade. Seed coats in both lineages have at least a few, sometimes many, intercellular gaps (Molvray and Kores, 1995), but this character can be hard to interpret in some cases. Further, developmental studies have not yet been done, so there is no guarantee that gaps in different lineages are homologous. Clements (1995, 1999) noted that the embryos of Orchideae have a multicelled suspensor, those of Cranichideae lack a suspensor, and those of Diurideae have a reduced, one- or two-celled suspensor. If reduction in the number of suspensor cells is considered a derived condition, then this multistate character could be ordered to constitute a synapomorphy for the diurid–spiranthid clade. Despite the paucity of synapomorphies, the sister relationship between Diurideae and spiranthids has important implications for character phylogeny of underground storage organs and column morphology. The position of Chloraea and Gavilea, which lack tuberoids, as sister to other spiranthids, implies that the underground storage organs found in Pterostylis are either plesiomorphic or derived independently of those in core diurids. Further, the presence of tuberoids in both Orchideae and core diurids indicates that these organs are plesiomorphic and therefore unlikely to be phylogenetically informative until their homologies are subjected to detailed developmental studies. Acrotony has been used to define relationships within the Orchidaceae, despite the fact that there are problems defining the term. If we accept Dressler’s or Rasmussen’s definition of acrotony, in which the viscdium or rostellum is associated with the apex of the anther or its pollinia, then the distribution of the character on the tree indicates symplesiomorphy or convergence. Acrotonic anthers are present in some epidendroids, the spiranthids, most diurids, and Codonorchis, whereas basitonic anthers, sensu Dressler, are restricted to core Orchideae or to core Orchideae and Diurideae sensu Rasmussen. However, floral developmental studies by Kurzweil (1987a, b, 1988) show that all monandrous orchids are similar in early floral organogeny and that acrotony subsequently develops in a variety of different ways. In more-derived epidendroid orchids, acrotony is achieved by inward bending of the anther brought about by differential growth at its base. However, in the spiranthids and less-derived epidendroids, it is achieved by late growth of the rostellum. Within Diurideae, even among

(ORCHIDACEAE)

1911

sister genera such as Thelymitra and Calochilus, acrotony is achieved via distinctly different developmental pathways (Andrew Perkins, Royal Botanic Gardens, Sydney, Australia, personal communication). In Thelymitra, the rostellum is pushed near the tip of the anther by the growth of the stigma as a whole, whereas in Calochilus it results from elongation of the central sublobe of the median stigma lobe between the anther loculi. Other Diurideae, notably Acianthinae, are considered to have an intermediate, pleurotonic anther. However, Acianthinae are sister to Prasophyllinae that are characterized by a typical acrotonic anther. Given the close and well-supported relationship between the two subtribes, the transition between the two states may not be as significant as previously thought. Our results are consistent with Rasmussen’s (1982) hypothesis that pleurotony results from a more or less symmetrical increase in the size of the pollinia in both the upper and lower halves of the locule during maturation. The sister relationship between acrotonic Codonorchis and basitonic Orchideae provides more evidence that column morphology may be more labile than previously hypothesized. The lack of unambiguous morphological synapomorphies is also evident in the tribe Diurideae. All, except the suspensorless Townsonia, have a one- or two-celled suspensor, and the embryo develops entirely within the integuments. This type of development has been termed the diurid embryo pattern (Clements, 1995, 1999). However, Megastylis ‘‘glandulifera’’ (fide Clements, 1995) has a diurid embryo pattern, but in our analysis, Megastylis glandulosa is unequivocally a member of the spiranthid clade, making yet another exception to a diurid synapomorphy. In contrast, Megastylis rara is a true diurid, but its embryology has not been investigated. Most diurids have a one-leafed shoot, but there are some reversals (Diuris, Orthoceras, Arthrochilus, Chiloglottis, Aporostylis, Leporella, and Rimacola). Cryptostylis, Townsonia, Adenochilus, and Megastylis rara have one-leafed shoots connected by a rhizome, such that even though the whole plant has more than one leaf, the individual shoots are still one-leafed. Morphological synapomorphies for the major lineages within Diurideae are similarly difficult to identify. One of the most striking aspects of Diurideae are their variously appendaged columns. They may be partially free from the column and variously shaped, or entirely united with the column and winglike. Several authors have used these differences as characters (Dressler, 1993; Clements, 1995; Szlachetko, 1995). However, there is little agreement between the plastid tree topology and what might be expected based on column appendages. Diuridinae, Thelymitra, and Prasophyllinae have partially free, vascularized, obliquely falcate column appendages, yet the latter is in one major clade of the tribe, the former two in another. Given that these processes often help to orient potential pollinators (Jones, 1970, 1974a, b, 1981) and the widespread evidence of great morphological plasticity in response to pollinator selection, it appears likely that column appendages are yet another feature subject to considerable convergence. At the subtribal level, morphological synapomorphies are sometimes present. Acianthinae have a broad, reniform leaf, a petiole fused with the peduncle, and lack secondary roots. Prasophyllinae have a cylindrical leaf, sectile pollen, and a hamulus. The latter two characters also occur outside Diurideae (Freudenstein and Rasmussen, 1997, 1999). Caladeniinae have a recurved, often three-lobed labellum, clavate appendages on the labellum, and a distally winged column; the plants are almost always pubescent. Thelymitrinae have a column with

1912

AMERICAN JOURNAL

prominent, vascularized column appendages united with a median filament forming a hood around the style. Multiple leaves per shoot, which may be a symplesiomorphy, are characteristic of Diuridinae, as well as a column with prominently vascularized column appendages equal to or surpassing the anther. Cryptostylidinae have rhizomes and nonresupinate flowers with a simple lip. Considering relationships between subtribes within the Diurideae, obvious morphological synapomorphies appear rare. In the present analysis, Prasophyllinae are sister to Acianthinae, but the only character they appear to have in common is secretory tissue at the base of the labellum. Australian species of Acianthus and Cyrtostylis are reported to secrete nectar from paired glands at the base of the labellum (Jones, 1988). Secretory tissue associated with the base of the labellum has also been reported within New Caledonian Acianthus by Kores (1995) and for species of Genoplesium and Prasophyllum by Jones et al. (1999). Nectar production in Microtis has been suggested by Peakall and Beattie (1989). In their study of pollination in Microtis parviflora, they noted that ants forage persistently, visiting individual flowers and inflorescences repeatedly for nectar. They also observed that ants (Iridomyrmex gracilis) visited only newly opened flowers and that pollinia attachment and pollen transfer to the stigma occurred while they probed the base of the labellum. There are no reports of nectar production in Townsonia or Corybas, but the former is reported to be autogamic while the latter relies on pollination by deceit (Jones et al., 1999). As a consequence, the secretory function of the labellum may have been secondarily lost in these genera. Other higher level relationships within Diurideae, such as the Diuridinae–Drakaeinae clade are not supported by any apparent nonmolecular synapomorphies. Some authors have suggested that the mycorrhizal associations of terrestrial orchids may have phylogenetic significance (Warcup, 1981; Clements, 1988; Dressler, 1993). The patterns of fungal associations are complex, judging by the trees presented here. Some well-supported subtribes are uniform with respect to symbionts. For instance, the Diuridinae–Drakaeinae clade has a preponderance of taxa that share association with Tulasnella. Burnettia will likely be a member of the Lyperanthus–Pyrorchis clade and also has the same fungus. Exceptions are Megastylis rara, Rimacola, Leporella, and Cryptostylis. All members of Caladeniinae clade share Sebacina, but other clades vary in their associations. Within Acianthinae, Acianthus and Corybas associate with Tulasnella, whereas Cyrtostylis has Sebacina. In Prasophyllinae, Microtis has Sebacina, but Genoplesium and Prasophyllum share Ceratobasidium. Repeated switches appear to have occurred among fungal symbionts and their hosts. Although it may be difficult to extract phylogenetic information from the patterns at higher taxonomic levels, fungal associations certainly bear examination and comparison based on a robust phylogenetic tree. Our results indicate that Cryptostylidinae belong with other diurids, not with the spiranthoid orchids as postulated earlier (Dressler, 1981; Rasmussen, 1985; Burns-Balogh and Funk, 1986). Stern et al. (1993), in their study of Spiranthoideae sensu Dressler (1981, 1993), noted that Cryptostylis does not share the anatomical synapomorphies he identified for the subfamily. In our analysis, Cryptostylidinae are embedded within a well-supported clade made up of four monophyletic lineages. Within this larger clade the subtribe is sister to the combined Thelymitrinae–Drakaeinae subclade, but this relationship has only weak bootstrap support. The molecular data do not sup-

OF

BOTANY

[Vol. 88

port the elevation of Cryptostylidinae to tribal rank as proposed by Szlachetko (1995) and Clements (1995). The trees based on the molecular data presented here also indicate that the Chloraeinae are polyphyletic. One species, Megastylis rara, is placed with representatives of Drakaeinae (Diurideae), whereas the other two species, Megastylis gigas and M. glandulosa, are sister to Cranichideae in the spiranthid lineage. Gavilea and Chloraea (the core Chloraeinae) are also placed in the spiranthid lineage, but they are not sister to that part of Megastylis. As previously mentioned, Codonorchis, the other representative of Chloraeinae, is sister to Orchideae. In light of the molecular data, we have a possible explanation for the unusually high morphological diversity that has been noted within the subtribe by some authors (Ackermann and Williams, 1981; Clements, 1995; Molvray and Kores, 1995). Once Codonorchis and Megastylis are removed from Chloraeinae, the remaining genera, which are all South American, form a well-supported group likely to remain monophyletic with increased sampling. Polyphyly and paraphyly are also evident at the generic level. Megastylis has been mentioned above because it crosses tribal lines. Our trees indicate that Genoplesium is a grade leading to Prasophyllum, and its status as a segregate genus is doubtful. Acianthus is polyphyletic. The Australian clade, which contains the type species of the genus, is sister to a clade composed of one lineage of Corybas and Cyrtostylis and another lineage of Stigmatodactylus and New Caledonian Acianthus. Further sampling is needed within the latter lineage to resolve relationships between the two taxa. Townsonia is also a member of Acianthinae, based on our unpublished ITS data. The DNA trees indicate that morphological plasticity in response to selective pressure is generally high in Diurideae. Kores et al. (2000) have noted, for instance, that the rhizomatous habit has arisen at least four times in four separate lineages. Three of these rhizomatous genera lack root-stem tuberoids (Adenochilus, Cryptostylis, and Rimacola), and the fourth (Townsonia) has very reduced tuberoids. Most diurids form well-developed root-stem tuberoids, which give rise to deciduous leafy shoots, but the chlorophyllous, rhizomatous species are all evergreen. These evergreen diurids (with the exception of Australian Cryptostylis) live in habitats with abundant moisture throughout the year, hence it seems reasonable to speculate that the evergreen rhizomatous habit lacking tuberoids has evolved repeatedly in different lineages in response to increased water availability. In the case of Cryptostylis, many of the non-Australian species seem to grow in mesic habitats, but the Australian species are often found in drier areas as well. The genus is characterized by moderately fleshy roots, which some authors have interpreted as subterranean storage organs (Clements, 1995), and these structures may facilitate survival in more xeric locations. Pollination syndromes involving sexual deception of male hymenopterans, with all their highly derived floral modifications, have arisen at least six times: Cryptostylis–Ichneumonidae, Calochilus–Scolidae, Leporella–Formicidae, Spiculaea– Paracaleana–Tiphiidae. It is noteworthy that two of these independent innovations involve the same hymenopteran subfamily, Thynnidae. The current study provides one more example in the close relationship between the recently segregated Pyrorchis and Lyperanthus. The apparently large difference between Lyperanthus and Pyrorchis in leaf morphology, the latter having a fleshy, prostrate leaf, may be less important

October 2001]

KORES

ET AL.—DIURIDEAE

than it appears. Pyrorchis tend to prefer relatively exposed, sandy substrates, whereas Lyperanthus are found in more sheltered locations. Adaptations that reduce water loss may account for the difference in leaf morphology. The morphological diversity of Diurideae and their lack of easily identifiable morphological synapomorphies have made them difficult to synthesize into an ordered framework that can be used to focus other studies. However, trees based on plastid DNA sequences are beginning to show the levels of support that promise to make them useful in this process. Molecular data have given us useful tools for studying phylogenies, but perhaps most interestingly, they facilitate the detection of phylogenetically informative morphological characters and suggest which characters may be most in need of further study to understand developmental processes and homologies. LITERATURE CITED ACKERMAN, J. D., AND N. H. WILLIAMS. 1981. Pollen morphology of the Chloraeinae (Orchidaceae: Diurideae) and related subtribes. American Journal of Botany 68: 1392–1420. BENTHAM, G. 1873. Orchidaceae. Flora Australiensis, vol. 6, 267–396. L. Reeve and Co., London, UK. BRIEGER, F. G. 1974–1975. Unterfamilie: Neottioideae. In F. G. Brieger, R. Maatsch, and K. H. Senghas [eds.], Rudolph Schlechter. Die Orchideen. Bd. 1, Teil A, (Fasc. 5 and 6), 284–358. Verlag Paul Parey, Berlin, Germany. BURNS-BALOGH, P., AND V. A. FUNK. 1986. A phylogenetic analysis of the Orchidaceae. Smithsonian Contributions to Botany 6: 1–79. CAMERON, K. M., M. W. CHASE, M. W. WHITTEN, P. J. KORES, D. C. JARRELL, V. A. ALBERT, T. YUKAWA, H. G. HILLS, AND D. H. GOLDMAN. 1999. A phylogenetic analysis of Orchidaceae: evidence from rbcL nucleotide sequences. American Journal of Botany 86: 208–224. CHASE, M. W., A. Y. DE BRUIJN, G. REEVES, A. V. COX, P. J. RUDALL, M. A. T. JOHNSON, AND L. E. EGUIARTE. 2000. Phylogenetics of Asphodelaceae (Asparagales): an analysis of plastid rbcL and trnL-F DNA sequences. Annals of Botany 86: 935–951. CLEMENTS, M. A. 1988. Orchid mycorrhizal associations. Lindleyana 3: 73– 86. CLEMENTS, M. A. 1995. Reproductive biology in relation to phylogeny of the Orchidaceae especially the tribe Diurideae. Ph.D. dissertation, Australian National University, Canberra, Australia. CLEMENTS, M. A. 1999. Embryology, vol. 1. In A. M. Pridgeon, P. J. Cribb, M. W. Chase, and F. N. Rasmussen [eds.], Genera Orchidacearum Oxford University Press, Oxford, UK. CRIBB, P. J., AND P. J. KORES. 2000. The systematic position of Codonorchis (Orchidaceae: Orchidoideae). Lindleyana 15: 169–170. DOYLE, J. J., AND J. L. DOYLE. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemistry Bulletin 19: 11–15. DRESSLER, R. 1981. The orchids: natural history and classification. Harvard University Press, Cambridge, Massachusetts, USA. DRESSLER, R. 1993. Phylogeny and classification of the orchid family. Dioscorides Press, Portland, Oregon, USA. ENDLICHER, S. L. 1842. Mantissa Botanica sistens Generum Plantarum supplementum secundum. Wien, Austria. FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. FITCH, W. M. 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Systematic Zoology 20: 406–416. FREUDENSTEIN, J. V., AND F. N. RASMUSSEN. 1997. Sectile pollinia and relationships in the Orchidaceae. Plant Systematics and Evolution 205: 125–146. FREUDENSTEIN, J. V., AND F. N. RASMUSSEN. 1999. What does morphology tell us about orchid relationships? A cladistic analysis. American Journal of Botany 86: 225–248. GARAY, L. A. 1972. On the origin of Orchidaceae, II. Journal of the Arnold Arboretum 53: 202–215. GEILLY, L., AND P. TABERLET. 1994. The use of chloroplast DNA to resolve plant phylogenies: noncoding versus rbcL sequences. Molecular Biology and Evolution 11: 769–777.

(ORCHIDACEAE)

1913

HAWKES, A. D. 1965. Encyclopedia of cultivated orchids. Faber and Faber, London, UK. HOFFMAN, N., AND A. BROWN. 1992. Orchids of south-west Australia, 2nd ed. University of Western Australia Press, Nedlands, Australia. HOFFMAN, N., AND A. BROWN. 1998. Orchids of south-west Australia, 2nd ed. with supplement. University of Western Australia Press, Nedlands, Australia. HOPPER, S. D., AND A. P. BROWN. 2000. New genera, subgenera, combinations, and species in the Caladenia alliance (Orchidaceae: Diurideae). Lindleyana 15: 120–126. JARRELL, D. C., AND M. T. CLEGG. 1995. Systematic implications of the chloroplast-encoded matK gene on the tribe Vandeae (Orchidaceae). American Journal Botany 82(suppl.): 137. JOHNSON, L. A., AND D. E. SOLTIS. 1994. matK DNA sequences and phylogenetic reconstruction in Saxifragaceae s. str. Systematic Botany 19: 143–156. JOHNSON, L. A., AND D. E. SOLTIS. 1995. Phylogenetic inference in Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK sequences. Annals of the Missouri Botanical Garden 82: 149–175. JONES, D. L. 1970. The pollination of Corybas diemenicus H. M. R. Rupp and W. H. Nicholls ex H. M. R. Rupp. Victorian Naturalist 87: 372– 374. JONES, D. L. 1974a. The pollination of Acianthus caudatus R. Br. Victorian Naturalist 91: 272–274. JONES, D. L. 1974b. The pollination of Calochilus holtzei F. Muell. American Orchid Society Bulletin 43: 604–606. JONES, D. L. 1981. The pollination of selected Australian orchids. In L. Lawler and R. D. Kerr [eds.], Proceedings of the Orchid Symposium, 13th International Botanical Congress, Sydney, Australia 1981, 40–43. Orchid Society of New South Wales, Sydney, Australia. JONES, D. L. 1988. Native orchids of Australia. Reed Books, Frenchs Forest, Australia. JONES, D. L., H. WAPSTRA, P. TONELLI, AND S. HARRIS. 1999. The orchids of Tasmania. Melbourne University Press, Carlton South, Australia. KORES, P. J. 1995. A systematic study of the genus Acianthus (Orchidaceae: Diurideae). Allertonia 7(3): 1–220. KORES, P. J. In press. Circumscription and phylogeny of the spiranthoid orchids and implications for character evolution. In J. Clark, G. Tingley, and J. Biro [eds.], Proceedings of the 16th World Orchid Conference. Harbour Publishing, Vancouver, British Columbia, Canada. KORES, P. J., K. M. CAMERON, M. MOLVRAY, AND M. W. CHASE. 1997. The phylogenetic relationships of Orchidoideae and Spiranthoideae (Orchidaceae) as inferred from rbcL plastid sequences. Lindleyana 12: 1–11. KORES, P. J., P. WESTON, M. MOLVRAY, AND M. W. CHASE. 2000. Phylogenetic relationships within Diurideae: inferences from plastid matK DNA sequences. In K. L. Wilson and D. A Morrison [eds.], Monocots: systematics and evolution. CSIRO Publishing, Collingwood, Victoria, Australia. KURZWEIL, H. 1987a. Developmental studies in orchid flowers I: epidendroid and vandoid species. Nordic Journal of Botany 7: 427–442. KURZWEIL, H. 1987b. Developmental studies in orchid flowers II: orchidoid species. Nordic Journal of Botany 7: 443–451. KURZWEIL, H. 1988. Developmental studies in orchid flowers III: neottioid species. Nordic Journal of Botany 8: 271–282. LAVARACK, P. 1971. The taxonomic affinities of the Neottioideae. Ph.D. dissertation, University of Queensland, Brisbane, Queensland, Australia. LAVARACK, P. 1976. The taxonomic affinities of the Australian Neottioideae. Taxon 25: 289–296. MABBERLEY, D. J. 1997. The plant book. Cambridge University Press, Cambridge, UK. MANSFELD, R. 1937. Uber das System der Orchidaceae-Monandrae. Notizblatt des Koniglichen Botanishen Gartens und Museums zu Berlin-Dahlem 13: 666–676. ¨ ber die Verteilung der Merkmale innerhalb der OrMANSFELD, R. 1955. U chidaceae-Monandrae. Flora 142: 65–80. MOLVRAY, M., AND P. KORES. 1995. Character analysis of the seed coat in the Spiranthoideae and Orchidoideae, with special reference to the Diurideae (Orchidaceae). American Journal of Botany 82: 1443–1454. MOLVRAY, M., P. J. KORES, AND M. W. CHASE. 2000. Polyphyly of mycoheterotrophic orchids and functional influences on floral and molecular characters. In K. L. Wilson and D. A Morrison [eds.], Monocots: systematics and evolution. CSIRO Publishing, Collingwood, Victoria, Australia.

1914

AMERICAN JOURNAL

OLMSTEAD, R. G., 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. PEAKALL, R., AND A. J. BEATTIE. 1989. Pollination of the orchid Microtis parviflora R. Br. by flightless worker ants. Functional Ecology 3: 515– 522. PRIDGEON, A. M., AND M. W. CHASE. 1995. Subterranean axes in tribe Diurideae (Orchidaceae): morphology, anatomy, and systematic significance. American Journal of Botany 82: 1473–1495. RASMUSSEN, F. N. 1982. The gynostemium of the neottioid orchids. Opera Botanica 65: 1–96. RASMUSSEN, F. N. 1985. Orchids. In R. M. Dahlgren, H. T. Clifford, and P. F. Yeo [eds.], The families of the monocotyledons, 249–274. SpringerVerlag, Berlin, Germany. RASMUSSEN, F. N. 1986. Ontogeny and phylogeny in Orchidaceae. Lindleyana 1: 114–124. SCHLECHTER, R. 1926. Das System der Orchidaceen. Notizblatt des Botanischen Gartens und Museums zu Berlin-Dahlem 9: 563–591.

OF

BOTANY

[Vol. 88

SEELANAN, T., A. SCHNABEL, AND J. WENDEL. 1997. Congruence and consensus in the cotton tribe (Malvaceae). Systematic Botany 22: 259–290. SHEAHAN, M. C., AND M. W. CHASE. 2000. Phylogenetic relationships within Zygophyllaceae based on DNA sequences of three plastid regions, with special emphasis on Zygophylloideae. Systematic Botany 25: 371–384. STERN, W. L., M. W. MORRIS, W. S. JUDD, A. M. PRIDGEON, AND R. L. DRESSLER. 1993. Comparative vegetative anatomy and systematics of Spiranthoideae (Orchidaceae). Protoplasma 172: 49–55. SWOFFORD, D. L. 1993. PAUP: phylogenetic analysis using parsimony, version 3.1. Formerly distributed by Illinois Natural History Survey, Champaign, Illinois, USA. SWOFFORD, D. L. 1998. PAUP* 4.0b: phylogenetic analysis using parsimony (and other methods). Sinauer, Sunderland, Massachusetts, USA. SZLACHETKO, D. L. 1991. Thelymitroideae, a new subfamily within Orchidaceae. Fragmenta Floristica et Geobotanica 36: 33–49. SZLACHETKO, D. L. 1995. Systema Orchidalium. Fragmenta Floristica et Geobotanica 3: 1–152. WARCUP, J. H. 1981. The mycorrhizal relationships of Australian orchids. New Phytologist 87: 371–381.