Phylogeneticpositions S.- C. Kim of Actites and Sonchushydrophi lus. (Sonchinae: Asteraceae) based on ITS and chloroplast non-coding DNA sequences

Transcription

1 etal. CSIRO PUBLISHING Australian Systematic Botany 17, Phylogenetic positions of Actites megalocarpa and Sonchus hydrophilus (Sonchinae: Asteraceae) based on ITS and chloroplast non-coding DNA sequences Seung-Chul Kim A,C, Christina T. Lu A and Brendan J. Lepschi B A Department of Botany and Plant Sciences, University of California, Riverside, CA 92521, USA. B Australian National Herbarium, Centre for Plant Biodiversity Research, GPO Box 1600, Canberra ACT 2601, Australia. C Corresponding author; sckim@citrus.ucr.edu Abstract. Phylogenetic positions of the Australian endemic taxa Actites megalocarpa and Sonchus hydrophilus within the subtribe Sonchinae were determined on the basis of ITS sequences of nuclear rdna and the psba trnh (GUG) intergenic spacer of chloroplast DNA. Both ITS and cpdna phylogenies suggest that the monotypic genus Actites is not closely related to the members of Sonchus section Asperi, as previously suggested. Rather, this study indicates that it is more closely related to the members of Sonchus sections Maritimi (S. maritimus) and Arvenses (S. arvensis). It also suggests that S. maritimus from section Maritimi is one of the closest relatives of Actites in Australia, although an alternative origin from section Arvenses is possible. Actites and Embergeria, once treated as congeneric taxa, appear to have originated independently in Australia and New Zealand, respectively. Sonchus hydrophilus is closely related to the S. asper complex, S. oleraceus and S. kirkii. This study suggests that S. kirkii may be involved in the origin of S. hydrophilus in Australia. SB03019 Phylogeneticpositions S.- C. Kim of Actites and Sonchushydrophi lus Introduction The subtribe Sonchinae (Bremer 1993, 1994) comprises 11 genera and approximately 130 species. It consists of two alliances: (1) Launaea Cass. and the related genera Aetheorhiza Cass. and Reichardia Roth, and (2) Sonchus L. and a group of closely related genera, i.e. Actites Lander, Babcockia Boulos, Embergeria Boulos, Kirkianella Allan, Lactucosonchus (Sch. Bip.) Svent, Sventenia Font Quer and Taeckholmia Boulos. Most species occur in the Mediterranean region, Africa, and the Canary Islands, while three genera, Actites, Embergeria and Kirkianella, occur in the Australasian region. The genus Actites contains a single species, A. megalocarpa (Hook. f.) Lander, a fleshy perennial herb endemic to coastal sand dunes and cliffs on the southern and eastern coasts of Australia (Lander 1976). Actites megalocarpa was originally described as Sonchus asper var. megalocarpa Hook. f., but was subsequently afforded specific status by Black (1929) as S. megalocarpus (Hook. f.) J. M. Black. Boulos (1965) transferred it to the genus Embergeria, as E. megalocarpa (Hook. f.) Boulos, apparently on the basis of leaf texture and cypsela size (see Lander 1976). Most recently, Lander (1976) erected the monotypic genus Actites to accommodate A. megalocarpa, separating his new genus from Sonchus and Embergeria on a combination of morphological (achene shape and pappus type) and palynological characters. Embergeria grandifolia (Kirk) Boulos resembles Actites megalocarpa in habit, and occupies a similar habitat (coastal sand dunes and cliffs on the Chatham Islands off New Zealand), but differs in several morphological characters (see above). It has been suggested that both Actites and Embergeria evolved from Sonchus, most likely from section Asperi Boulos, during the late Pliocene in Australia and New Zealand, respectively (Boulos 1967; Pons and Boulos 1972; Roux and Boulos 1972; Lander 1976). While there is support for the exclusion of A. megalocarpa from Embergeria (e.g. Cooke 1986), the segregation of this taxon into a monotypic genus has not been universally accepted, despite general recognition of Actites by most authors (e.g. Murray 1992a, 1992b; Jeanes 1999; Wheeler et al. 2002). Cooke (1986) and Sennikov and Illarionova (2001) suggest Actites should be included within Sonchus, arguing that differences in cypsela morphology CSIRO 9 March /SB /04/010073

2 74 Australian Systematic Botany S.-C. Kim et al. Table 1. Sources of plant materials for ITS and cpdna intergenic spacer sequences Voucher specimens are deposited in the Australian National Herbarium (CANB) and/or the Western Australian Herbarium (PERTH). All sequences are deposited in GenBank Taxon Voucher GenBank ITS psba trnh Actites A. megalocarpa Heyligers (eastern Australia) AY AY Lepschi (western Australia) AY AY Lepschi (eastern Australia) AY (eastern Australia) AY AY (eastern Australia) AY AY Sonchus S. hydrophilus Lepschi (western Australia) AY AY (western Australia) AY AY Lepschi & Lally 2349 (western Australia) AY AY Paget 2783 (south-eastern Australia) AY AY Stajsic 871 (south-eastern Australia) AY AY S. asper subsp. asper Adams 4153 (eastern Australia) AY AY Keighery (western Australia) AY S. asper subsp. glaucescens Lepschi (eastern Australia) AY AY S. oleraceus Adams 4154 (eastern Australia) AY Gibson & Brown 3853 (western Australia) AY Lepschi & Lally 4683 (eastern Australia) AY AY S. tenerrimus Lepschi 3070 (western Australia) AY AY between Actites and Sonchus s.l. are insufficient for generic delimitation. Sonchus hydrophilus Boulos is endemic to Australia. It occurs in a variety of habitats, but particularly in wet sites such as along watercourses and seepage lines (Boulos 1973; B. Lepschi unpublished data). Boulos (1973) placed it in Sonchus section Asperi and on the basis of morphological and palynological characters, suggested that S. hydrophilus was probably an autotetraploid that had originated from Sonchus asper subsp. glaucescens (Jordan) Ball. Boulos (1965) also records this species as occurring in New Zealand and New Guinea, although according to Garnock-Jones (1988), records of this species from New Zealand are erroneous, and refer to the endemic S. kirkii (T. Kirk.) Allan. Similarly, the status of this species in New Guinea requires confirmation. Koster (1976) cast doubt on the taxonomic validity of S. hydrophilus, reducing it to forma status within S. asper, but it is likely that the material examined by Koster represented S. asper rather than true S. hydrophilus. Recent work by Kim et al. (1996), using internal transcribed spacer (ITS) sequences of nuclear rdna (nrdna), has provided insights into the phylogenetic relationships among genera of the subtribe Sonchinae (sensu Bremer 1993). In particular, the ITS tree from this study did not support the hypothesis of the origin of Embergeria in New Zealand and suggested that Embergeria and Kirkianella, another New Zealand endemic, are most closely related to the members of Sonchus sections Arvensis (Kirp.) Boulos and Maritimi (Kirp.) Boulos. Further research using psba trnh (GUG) cpdna intergenic spacer sequence (Kim et al. 1999) also suggested similar relationships among members of Sonchinae, but with less resolution. These studies, however, did not include other important Australian endemics, such as Actites megalocarpa and Sonchus hydrophilus (Kim et al. 1996, 1999). Therefore, the phylogenetic positions of these two taxa within Sonchinae need to be determined to understand their origin and evolution in Australia. The purpose of this study is to determine the phylogenetic position of the genus Actites within the subtribe Sonchinae, and to determine the phylogenetic position and taxonomic status of Sonchus hydrophilus relative to S. asper. Materials and methods To determine the phylogenetic position of Actites megalocarpa, we sampled five individuals, four from eastern and one from western Australia (Table 1). We sampled five individuals of Sonchus hydrophilus, three from western and two from eastern Australia. We further sampled several introduced weedy Sonchus species from Australia to address the taxonomic status of S. hydrophilus, including S. asper subsp. asper, S. asper subsp. glaucescens, S. oleraceus and S. tenerrimus L. (Table 1). Total genomic DNA was isolated from 20 mg of silica-dried leaf material or herbarium specimens with DNeasy plant mini kits (Qiagen, Valencia, CA) and then quantified on a TKO-100 fluorometer (Hoefer Scientific Instruments, San Francisco, CA). The ITS of nrdna and psba trnh (GUG) intergenic spacer of cpdna were amplified as described previously (Kim et al. 1996, 1999). Reaction conditions were

3 Phylogenetic positions of Actites and Sonchus hydrophilus Australian Systematic Botany 75 an initial 1 min at 95 C followed by 34 cycles of 1 min at 94 C, 1 min at 50 C and 1 min at 72 C, with a final extension of 5 min at 72 C. PCR products of approximately 100 µl for each sample were purified with the QiaQuick PCR Purification kit (QIAGEN, Valencia, CA). Direct sequencing of PCR products was performed with ABI PRISM BigDye Terminator v3.0 Ready Reaction Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) and extension products were purified and subsequently separated on ABI377 automated sequencing machine (Applied Biosystems). In the case of the ITS of nrdna, two additional internal sequencing primers, ITS 2 and ITS 3, were used to sequence both strands of ITS 1 and ITS 2. Basecalling and sequence-editing were performed with Sequencher 4.1 (Genecodes, Inc., Ann Arbor, MI). The boundaries of the ITS and rdna coding regions were determined on the basis of the previous study (Kim et al. 1996). Both ITS 1 and ITS 2 and psba trnh (GUG) intergenic spacer sequences were manually entered and visually aligned to the previously existing MacClade (version 4; Maddison and Maddison 2000) data matrices for the subtribe Sonchinae (available upon request from the first author). Phylogenetic analyses of the aligned datasets using Fitch parsimony were performed with PAUP (version 4.0; Swofford 2001) using HEURISTIC search option with TBR branch swapping and MULPARS on. Based on the previous phylogenetic analysis, only Prenanthes purpurea L. was used as an outgroup (Kim et al. 1996). Gapped nucleotide positions in alignments were treated as missing values. Bootstrap analysis (Felsenstein 1985) using HEURISTIC search from a simple addition sequence with TBR branch swapping was performed with 100 replications to provide a measure of support for the clades. Phylogenetic analyses by PAUP ver 4.0 were also performed with neighbour-joining (Saito and Nei 1987) and UPGMA according to the Kimura 2-parameter method (Kimura 1980). Results Length variation and base composition of the ITS region and psba trnh spacer region The length of ITS 1 in Australian Sonchus species is 255 base pairs (bp), whereas Actites is 256 or 257 bp. The A+T content of ITS 1 ranges from 55.3 to 55.7% in Sonchus species, while Actites varies from 53.3 to 53.5%. In the case of the 5.8S coding region, all species have a total length of 163 bp. The length of ITS 2 varies from 222 to 225 bp in Sonchus species, while Actites is 223 bp. The A+T content in ITS 2 of Sonchus species ranges from 49.3 to 51.8%, whereas Actites is 53.8%. Among Australian Sonchus asper populations, there are two single-base pair changes and three indels in ITS 1 and ITS 2, respectively. Sonchus asper subsp. glaucescens has one polymorphic site (i.e. Y = C+T) when compared with two individuals of S. asper subsp. asper in ITS 1. All S. hydrophilus individuals have identical ITS sequences, including the 5.8S region. A total of three base pair changes (two transitions and one transversion) were found between S. asper and S. hydrophilus in both the ITS 1 and 2 regions. Furthermore, there was one non-synonymous change (A to G) between two species in the 5.8S coding region. In the case of Actites ITS sequences, one western individual (Lepschi ) was different from the eastern plants by a 1-bp deletion in ITS 1 and a 1-bp change (A to C) in ITS 2. No evidence of multiple rdna repeat types was observed in any of the Australian taxa analysed, even in putative polyploid species. In addition, polymorphism at individual nucleotide sites was not encountered, with the exception of one site in S. asper subsp. glaucescens. Therefore, ITS sequence data in this study provide no evidence for different ITS length variants or major sequence variants within individual samples. The length of the psba trnh intergenic spacer of cpdna varies from 357 [S. oleraceus L. (Lepschi & Lally 4683); S. asper subsp. asper (Adams 4153)] to 423 bp [Actites (all four accessions)]. Sonchus asper subsp. asper (Adams 4153) has a 43-bp deletion shared with S. oleraceus (Lepschi & Lally 4683). All S. hydrophilus accessions have a length of 400 bp and are identical, except for a 1-bp change (A to C) in an eastern individual (Stajsic 871). In Actites, one western individual (Lepschi ) differs in two base pairs (two transversions) from the rest of the eastern accessions. Unlike the ITS sequences, no fixed difference between S. asper and S. hydrophilus were found. The A+T content of Australian Sonchus species and Actites is in the range of %. Phylogenetic analyses A total of 443 characters was used for phylogenetic analyses of ITS sequences for 57 taxa including one outgroup. There are 249 variable sites (56%), and 163 characters (approximately 37%) are cladistically informative. The heuristic search found 6338 equally parsimonious trees with a length of 558, a consistency index (CI) of ( excluding cladistically uninformative changes), and a retention index (RI) of A strict consensus tree is shown in Fig. 1. The ITS phylogeny based on parsimony optimality criterion suggests that the two genera Reichardia and Launaea are basal in Sonchinae. Populations of Actites are monophyletic (100% bootstrap support) and the Sonchus maritimus and S. arvensis clade is sister to the Actites clade (99% bootstrap support). This clade in turn is sister to the two monotypic New Zealand endemics Embergeria and Kirkianella (91% bootstrap support; Fig. 1). Populations of Sonchus hydrophilus comprise a moderately supported clade (84% bootstrap support) that is part of a polytomy that includes other Sonchus species such as S. asper, S. oleraceus, and S. kirkii (100% bootstrap support). In the 50% majority-rule consensus tree (not shown), S. kirkii, endemic to New Zealand, appears to be sister to the S. hydrophilus clade (67%). The neighbour-joining tree (Fig. 2) based on ITS sequences shows similar relationships among major clades. The phylogenetic relationships of Actites and S. hydrophilus relative to other species in Sonchinae are the same as suggested in the parsimony tree. Several major incongruencies between parsimony and neighbour-joining trees were apparent: (1) the position of the African Sonchus clade [S. saudensis Boulos, S. schweinfurthii Olive et Hiern,

4 76 Australian Systematic Botany S.-C. Kim et al. Fig. 1. Strict consensus of 6338 shortest trees resulting from maximum parsimony analysis using the ITS sequences (163 informative characters; 558 steps; CI excluding cladistically uninformative changes = , RI = ). Bootstrap values (%) are shown above branches. The species sequenced in this study are in bold.

5 Phylogenetic positions of Actites and Sonchus hydrophilus Australian Systematic Botany 77 Fig. 2. Neighbour-joining tree of subtribe Sonchinae based on ITS sequences. Bootstrap values are above branches. The species sequenced in this study are in bold.

6 78 Australian Systematic Botany S.-C. Kim et al. Fig. 3. Strict consensus of equally parsimonious trees resulting from maximum parsimony analysis using the psba trnh (GUG) intergenic spacer (44 parsimony informative characters; 124 steps; CI excluding cladistically uninformative characters = ; RI = ). Bootstrap supports (%) are above branches. The species sequenced in this study are in bold.

7 Phylogenetic positions of Actites and Sonchus hydrophilus Australian Systematic Botany 79 and S. luxurians (R. E. Fries) C. Jeffrey]; (2) the position of the monotypic genus Aetheorhiza; and (3) the position of S. palustris L. All these positions were weakly supported (below 33% bootstrap values) by bootstrap analysis in parsimony optimality criterion. The neighbour-joining tree further suggested that S. kirkii is closely related to S. hydrophilus (Fig. 2). The UPGMA tree (not shown) is similar to the neighbour-joining tree and the phylogenetic positions of Actites and S. hydrophilus are the same as the neighbour-joining tree. In the case of cpdna psba trnh intergenic spacer sequences for 48 taxa, total aligned sequences were 417, with 83 variable sites (approximately 20%). Thirty-nine variable characters (approximately 9%) were uninformative, whereas 44 characters (approximately 10%) were cladistically informative. Parsimony analysis found equally parsimonious trees with a length of 124 (CI = , CI excluding cladistically uninformative characters = ; RI = ). A strict consensus tree (Fig. 3), like the one based on the ITS sequences, suggests that Actites is closely related to the two New Zealand endemic taxa Kirkianella and Embergeria, and one species of Sonchus, S. maritimus L. It also suggests that S. hydrophilus is closely related to other species of Sonchus, such as S. asper, S. oleraceus, and S. kirkii (Fig. 3). Overall, the cpdna phylogeny showed less resolution, with lower bootstrap support, compared to the ITS phylogeny. Both ITS and cpdna phylogeny suggested the same phylogenetic positions for Actites and S. hydrophilus within the subtribe Sonchinae. Discussion Phylogenetic position and evolution of the monotypic genus Actites As previously suggested by several authors (e.g. Boulos 1967; Pons and Boulos 1972; Lander 1976), it seems justified to conclude that the monotypic Australian endemic Actites is derived from subgenus Sonchus. This study, based on both ITS and cpdna non-coding sequences, strongly suggests that Actites is closely related to the members of subgenus Sonchus, especially sections Maritimi (S. maritimus) and Arvenses (S. arvensis L.). In the ITS phylogeny (Figs 1, 2), two monotypic genera from New Zealand, Kirkianella and Embergeria, are in turn sister to the clade of Actites and S. maritimus S. arvensis. These relationships were strongly supported by high bootstrap values (over 99%). In the cpdna phylogeny (Fig. 3), Actites is clearly closely related to S. maritimus, Kirkianella, and Embergeria. However, the relationships among them were poorly resolved, due in part to slowly evolving cpdna intergenic spacer compared to the ITS of nrdna. Nevertheless, these two independent phylogenies provide several insights into the origin of Actites. First, Actites may not have evolved from the members of section Asperi as suggested previously (see Lander 1976). All members of section Asperi included in this study (three of six species: S. asper, S. kirkii, and S. hydrophilus) were not closely related to Actites (Figs 1 3) and sections Asperi and Sonchus also appear not to be monophyletic. The three remaining species not included in this study, i.e. S. mauritanicus Boiss. & Reut., S. gigas Boulos ex Humbert, and S. macrocarpus Boulos & C.Jeffrey, are geographically restricted to Mauritius, Africa and Malagasy, and Egypt, respectively. Therefore, it seems unlikely that any of these species are responsible for the origin of Actites in Australia. Rather, both the ITS and cpdna phylogenies suggest that Actites may be derived from either Sonchus sections Maritimi or Arvenses. We were able to obtain DNA material from only three of the seven species of these two sections (S. maritimus and S. palustris for section Maritimi and S. arvensis for section Arvenses) and thus it is not possible to precisely determine the origin of Actites. Both phylogenies suggest that Sonchus palustris (section Maritimi) is not closely related to Actites (Figs 1 3) and thus, it is unlikely that this species is ancestral to Actites. Sonchus arvensis also may not be directly involved in the origin of Actites because this species appears to be a hexaploid (2n = 54; Boulos 1961; Löve and Löve 1961), whereas Actites may be a tetraploid (2n = 36; Pons and Boulos 1972). In addition, the cpdna haplotype of S. arvensis is very different from several of the Pacific endemics in this study (Fig. 3). It seems plausible that S. maritimus (section Maritimi), which occurs in southern and western Europe, the Mediterranean, and northern Africa, is one of the closest relatives of Actites. This species is not only closely aligned with Actites in both phylogenetic trees (Figs 1 3), but is also a diploid (2n = 18; Boulos 1960; Delay 1968) with a coastal distribution. Alternatively, it is possible that Actites may have evolved from several other diploid or tetraploid Sonchus species in section Arvenses, such as S. brachyotus DC. (2n = 18; Roux and Boulos 1970), S. arvensis subsp. uliginosus (M. Bieb.) Bég. (2n = 36; Shumovich and Montgomery 1955; Löve and Löve 1961), and S. wightianus DC. (n = 9; Mitra and Datta 1967). These taxa are widely distributed throughout Europe and Asia (S. arvensis subsp. uliginosus; also naturalised in North America), eastern Asia (S. brachyotus) and western to south-eastern Asia (S. wightianus). They are geographically close to Australia and cytological data suggest that some may be involved in the origin of Actites. A western population of Actites (Lepschi ) has different ITS sequences (1-bp deletion in ITS 1 and 1-bp substitution in ITS 2) compared to the eastern populations and is sister to the eastern populations in the phylogenetic tree (Figs 1, 2). Furthermore, the western Actites accession has a different cpdna haplotype compared with the eastern populations (see Fig. 3 and Results section). These findings suggest genetic differentiation among populations of Actites

8 80 Australian Systematic Botany S.-C. Kim et al. and suggest that it may have originated in western Australia. It is possible that members of section Maritimi or Arvenses migrated from south-east Asia to Australia and New Zealand during the late Pliocene (Raven 1972; Lander 1976), and Actites has arisen through either polyploidisation or cladogenesis/anagenesis. Our study supports Lander s taxonomic treatment (1976) of Actites and suggests that Actites and Embergeria originated independently in Australia and Chatham Islands, respectively, from members of subgenus Sonchus (Figs 1 3). Garnock-Jones (1988) suggested that Embergeria is closely related to S. kirkii, another New Zealand endemic. However, this study suggests that S. kirkii is not closely related to Embergeria, and that Kirkianella, another monotypic New Zealand endemic (Allan 1961), is sister to Embergeria (Figs 1 3). Kirkianella is a pentaploid (2n = 90) and heptaploid (2n = 129) (Beuzenberg and Hair 1984), which suggests both Embergeria and Kirkianella originated through polyploidisation, but progenitors or ancestors cannot be determined in this study. Our study suggests that, like other members of subgenus Sonchus, Actites retained a dimorphic pappus, whereas both Embergeria and Kirkianella lost glochidial pappus bristles after dispersal to New Zealand. A loss of certain traits after long-distance dispersal seems a common feature of island plants (Carlquist 1965) and the presence or absence of a pappus in tribe Lactuceae seems to be under simple genetic control (Vlot et al. 1992; Van Houten et al. 1994). Phylogenetic position and taxonomic status of Sonchus hydrophilus The ITS phylogeny indicates that S. hydrophilus in Australia is monophyletic (Figs 1, 2). This monophyly is moderately supported by bootstrap analysis (84% support). ITS sequences strongly suggest that S. hydrophilus is closely related to the S. asper complex, S. oleraceus, and S. kirkii. Two species of Sonchus, S. tenerrimus and S. bourgeaui Sch. Bip., in section Sonchus are in turn sister to the S. hydrophilus S. asper complex S. oleraceus S. kirkii clade. The phylogenetic relationships of S. hydrophilus relative to the other species in the same clade are poorly resolved. However, the ITS 50% majority-rule consensus tree (not shown) and the neighbour-joining tree (Fig. 2) suggest that S. kirkii, endemic to New Zealand, is basal to S. hydrophilus (67% in majority-rule consensus tree and 76% bootstrap support in the neighbour-joining tree). This suggests that S. hydrophilus is probably not closely related to the S. asper complex as suggested by Boulos (1973). Rather, S. kirkii from New Zealand is one of the closest relatives of S. hydrophilus and shares a common origin with it (Fig. 2). Palynological studies (Pons and Boulos 1972) have shown that S. hydrophilus has a mixture of tetracolporate and tricolporate pollen grains, suggesting a possible polyploid origin. Boulos (1973) suggested that S. hydrophilus is probably an autotetraploid, derived from S. asper subsp. glaucescens. The ITS phylogeny, however, does not support this view and suggests that S. kirkii is one of the closest relatives of S. hydrophilus (Figs 1, 2). The cpdna phylogeny does not provide much insight into the origin of S. hydrophilus owing to very low resolution (Fig. 3). However, all four individuals of S. hydrophilus have an identical cpdna haplotype to S. kirkii, S. asper and S. oleraceus (Adams 4154). This supports the suggestion that S. kirkii may be involved in the origin of S. hydrophilus in Australia. The chromosome number of S. hydrophilus has yet to be determined and thus, its ploidy level is unknown. There is no indication of different ITS length variants or repeat types in this species, which may suggest that this species is either a diploid or an autotetraploid if indeed polyploid. In conclusion, this study demonstrates that both Actites and Sonchus hydrophilus are monophyletic. The monotypic genus Actites appears not to be derived from Sonchus section Asperi as previously suggested. Rather, it appears to have arisen from members of Sonchus section Maritimi or possibly Arvenses. Our data suggest that Actites and Embergeria, considered congeneric by some authors, may have originated independently in Australia and New Zealand, respectively. This study also suggests that Sonchus hydrophilus is related to members of sections Asperi and Sonchus, with S. kirkii from New Zealand a likely close relative of S. hydrophilus. Acknowledgments We thank Dan Crawford for helpful comments on an early draft of this manuscript. This work was supported by new faculty set-up funds and Regents Faculty Fellowships at the University of California at Riverside to the first author. Laurie Adams, Kate Brown, Neil Gibson, Petrus Heyligers, Terena Lally and Neville Walsh are thanked for collection of plant materials used in this study. References Allan HH (1961) Flora of New Zealand. Vol I. (Ed. HH Allan) pp (Government Printer: Wellington) Beuzenberg EJ, Hair JB (1984) Contributions to a chromosome atlas of the New Zealand flora 27 Compositae. New Zealand Journal of Botany 22, Black JM (1929) Flora of South Australia. Part 4. (Government Printer: Adelaide) Boulos L (1960) Cytotaxonomic studies in the genus Sonchus. 2. The genus Sonchus, a general systematic treatment. Botaniska Notiser 113, Boulos L (1961) Cytotaxonomic studies in the genus Sonchus. 3. On the cytotaxonomy and distribution of Sonchus arvensis L. Botaniska Notiser 114, Boulos L (1965) Sonchus and Embergeria. In Supplement to Black s Flora of South Australia. (Ed. HJ Eichler) pp (Government Printer: Adelaide) Boulos L (1967) Taeckholmia, a new genus of Compositae from Canary Islands. Botaniska Notiser 120,

9 Phylogenetic positions of Actites and Sonchus hydrophilus Australian Systematic Botany 81 Boulos L (1973) Révision systématique du genre Sonchus L. s.l. IV. Soud-genre I. Sonchus. Botaniska Notiser 126, Bremer K (1993) New subtribes of the Lactuceae. Novon 3, Bremer K (1994) Asteraceae cladistics and classification. (Timber Press: Portland, OR) Carlquist S (1965) Island life. (The Natural History Press: New York) Cooke DA (1986) Asteraceae. In Flora of South Australia. Part 3. (Eds JP Jessop, HR Toelken) pp (Government Printer: Adelaide) Delay J (1968) Halophytes II. Informations Annuelles de Caryosystématique et Cytogénétique 2, Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, Garnock-Jones PJ (1988) Asteraceae Lactuceae. In Flora of New Zealand. Vol. IV. (Eds CJ Webb, WR Sykes, PJ Garnock-Jones) pp (Botany Division DSIR: Christchurch, NZ) Jeanes JA (1999) Actites. In Flora of Victoria. Vol. 3. (Eds NG Walsh, TJ Entwisle) p (Inkata Press: Melbourne) Kim S-C, Crawford DJ, Jansen RK (1996) Phylogenetic relationships among the genera of the subtribe Sonchinae (Asteraceae): evidence from ITS sequences. Systematic Botany 21, Kim S-C, Crawford DJ, Jansen RK, Santos-Guerra A (1999) The use of non-coding region of chloroplast DNA in phylogenetic studies of the subtribe Sonchinae (Asteraceae: Lactuceae). Plant Systematics and Evolution 215, Kimura M (1980) Simple method for estimating evolutionary rate of base substitution through comparative studies of nucleotide sequences. Journal of Molecular Evolution 16, Koster JT (1976) The Compositae of New Guinea. Blumea 23, Lander NS (1976) Actites, a new genus of Compositae from Australia. Telopea 1, Löve A, Löve D (1961) Chromosome numbers of central and northwest European plant species. Opera Botanica 5, Maddison WP, Maddison DR (2000) MacClade, version 4. (Sinauer: Sunderland, MA) Mitra K, Datta N (1967) IOBP chromosome number reports XIII. Taxon 16, Murray L (1992a) Actites. In Flora of New South Wales. Vol. 3. (Ed. GJ Harden) pp (University of NSW Press: Sydney) Murray L (1992b) Sonchus. In Flora of New South Wales. Vol. 3. (Ed. GJ Harden) pp (University of NSW Press: Sydney) Pons A, Boulos L (1972) Révision systématique du genre Sonchus L. s.l. III. Étude palynologique. Botaniska Notiser 125, Raven PH (1972) Evolution of subalpine and alpine plant groups in New Zealand. New Zealand Journal of Botany 11, Roux J, Boulos L (1970) IOBP chromosome number reports XXV. Taxon 19, Roux J, Boulos L (1972) Révision systématique du genre Sonchus L. s.l. II. Étude caryologique. Botaniska Notiser 125, Saito N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4, Sennikov AN, Illarionova ID (2001) Morphological and anatomical structure of achenes in the genus Sonchus s.l. (Asteraceae). Botanicheskii Zhurnal 86, Shumovich W, Montgomery FH (1955) The perennial sow thistles in northeastern North America. Canadian Journal of Agricultural Sciences 35, Swofford DL (2001) PAUP*: Phylogenetic analysis using parsimony (*and other methods) Version 4.0. (Sinauer: Sunderland, MA) Van Houten WL, Van Raamsdonk LWD, Bachmann K (1994) Interspecific evolution of Microseris pygmaea (Asteraceae, Lactuceae) analyzed by co-segregation of phenotypic characters (QTLs) and molecular markers (RAPDs). Plant Systematics and Evolution 190, Vlot EC, Van Houten WH, Mauthe S, Bachmann K (1992) Genetic and non-genetic factors influencing deviations from five pappus parts in a hybrid between Microseris douglasii and M. bigelovii (Asteraceae: Lactuceae). International Journal of Plant Sciences 153, doi: / Wheeler JR, Marchant NG, Lewington M (2002) Flora of the South West. Part 2. (Australian Biological Resources Study/ University of Western Australia Press: Perth) Manuscript received 17 July 2003, accepted 28 November