SYSTEMATICS OF NOTHOFAGUS (NOTHOFAGACEAE)

Transcription

1 American Journal of Botany 84(9): SYSTEMATICS OF NOTHOFAGUS (NOTHOFAGACEAE) BASED ON RDNA SPACER SEQUENCES (ITS): TAXONOMIC CONGRUENCE WITH MORPHOLOGY AND PLASTID SEQUENCES PAUL S. MANOS Department of Botany, Box 90338, Duke University, Durham, North Carolina Phylogenetic relationships were examined within the southern beech family Nothofagaceae using 22 species representing the four currently recognized subgenera and related outgroups. Nuclear ribosomal DNA sequences encoding the 5.8s rrna and two flanking internal transcribed spacers (ITS) provided 95 phylogenetically informative nucleotide sites from a single alignment of 588 bases per species. Parsimony analysis of this variation produced two equally parsimonious trees supporting four monophyletic groups, which correspond to groups designated by pollen type. These topologies were compared to trees from reanalyses of previously reported rbcl sequences and a modified morphological data set. Results from parsimony analysis of the three data sets were highly congruent, with topological differences restricted to the placement of a few terminal taxa. Combined analysis of molecular and morphological data produced six equally parsimonious trees. The consensus of these trees suggests two basal clades within Nothofagus. Within the larger of the two clades, tropical Nothofagus (subgenus Brassospora) of New Guinea and New Caledonia are strongly supported as sister to cool-temperate species of South America (subgenus Nothofagus). Most of the morphological apomorphies of the cupule, fruit, and pollen of Nothofagus are distributed within this larger clade. An area cladogram based on the consensus of combined data supports three trans- Antarctic relationships, two within pollen groups and one between pollen groups. Fossil data support continuous ancestral distributions for all four pollen groups prior to continental drift; therefore, vicariance adequately explains two of these disjunctions. Extinction of trans-antarctic sister taxa within formerly widespread pollen groups explains the third disjunction; this results in a biogeographic pattern indicative of phylogenetic relationship not vicariance. For the biogeographically informative vicariant clades, area relationships based on total evidence support the recently advanced hypothesis that New Zealand and Australia share a unique common ancestry. Contrary to previous thought, the distribution of extant Nothofagus is informative on the area relationships of the Southern Hemisphere, once precise phylogenetic relationships are placed in the context of fossil data. Key words: biogeography; ITS sequences; Nothofagus; Nothofagaceae; rdna; taxonomic congruence. One of the most noteworthy taxonomic elements of the relictual Southern Hemisphere temperate flora is the widespread genus Nothofagus (Nothofagaceae). These wind-pollinated evergreen and deciduous trees known as southern beeches have figured prominently in discussions of the biogeographical history of the land masses inhabited by living and extinct members of the genus (see Humphries, 98a, b; Tanai, 986 for review). Nothofagus is a particularly appealing genus in this regard because its distribution is trans-antarctic (Fig. ) and characterized by well-established areas of endemism. Historical explanations put forth to account for the dis- Manuscript recieved 6 August 996; revision accepted 7 January 997. The author thanks Bruce Baldwin, David Baum, Peter Fritsch, M. Alejandra Gandolfo, Bill Hahn, Peter Linder, Jean-Marc Moncalvo, Rytas Vilgalys, and Rick Zechman for helpful comments on earlier versions of the manuscript. Plant material was graciously provided by the Wau Ecology Institute, Mount Tombah Botanic Garden, Christchurch Botanic Gardens, University of California Botanical Garden, Andrew Drinnan, Robert Hill, Jenny Read, Andrea Premoli, Rytas Vilgalys, Kevin Nixon, M. Alejandra Gandolfo, and Steve and Barbara Manos. I am especially grateful to Bruce Baldwin, Suzanne Fredericq, Joey Spatafora, and Miguel Volovsek for expert technical advice. Additional thanks to Kevin Nixon for assistance in the field. This work was supported by the A.W. Mellon Foundation. 37 junct distribution of the 35 species of southern beech have been diverse and largely dependent on the prevailing theories of earth history (e.g., stable vs. mobile continents). Hooker (853) offered one of the first and simplest explanations for the numerous plant species distributed among New Zealand, Australia, and Tasmania, and cold-temperate South America by hypothesizing the existence of a formerly continuous flora fragmented by geological and climatic events. Although this hypothesis would eventually gain further credence, later biogeographers instead chose to present complicated narratives to explain the history of terrestrial life in the Southern Hemisphere, citing Nothofagus as a prime example. Darlington (965), motivated by the idea that continental drift occurred too early to affect Nothofagus, argued for a Northern Hemisphere origin of the genus with subsequent migration from the Indo-Australian Archipelago southward. He inferred several long-distance dispersals on the basis of austral fossil pollen deposits that showed compositional changes through time as represented by the distinct pollen types found among extant species. Critical to this argument for a northern origin is the assumption of a close phylogenetic relationship between Nothofagus and the genus Fagus (the true beech), both traditionally recognized as members of the predominantly Northern

2 38 AMERICAN JOURNAL OF BOTANY [Vol. 84 Fig.. The general distribution of Nothofagus in the Southern Hemisphere modified from Humphries (98a, b). Hemisphere family Fagaceae. Within a similar stabilist context, van Steenis (962) accepted this broad concept of Fagaceae with Nothofagus originating north of Australia, but explained the arrival of the genus to austral areas and South America by way of land bridges. Other contemporaries, most notably Cranwell (963) and Moore (97), needed only to evaluate the occurrence of the oldest Nothofagus microfossils known at that time to suggest a southern origin for the genus. This view was couched in the mobilist perspective with dispersal restricted to overland routes. Increasing fossil evidence has demonstrated that southern Gondwana served as a center for the diversification of various biota, including Nothofagus (see Crame, 989 and Drinnan and Crane, 989 for review). Brundin (965) was one of the first to suggest that clues to the biogeographical history of trans-antarctic relationships should be sought primarily within a phylogenetic context and that the role of fossils would be enhanced in such a context. With more complete documentation of the fossil pollen record for Nothofagus, the age and origin of the genus are now known with greater certainty. The microfossil record for Nothofagus currently includes eight distinct pollen types, four of which occur among living species (Dettman et al., 990). These four types have been recovered from the Early Campanian of the Late Cretaceous ( 70 Mya) of Western Antarctica and South America, and although the sequence of appearance of the four types is not apparent, all four types occur in slightly younger deposits of Australia and New Zealand (Dettman et al., 990). The distribution and diversity of these pollen fossils and presence of several additional older pollen types ( 83 Mya) assignable to Nothofagus strongly suggest that the ancestors of modern-day Nothofagus species had evolved and radiated in the Southern Hemisphere. Given the low probability of over-water or bird dispersal for the genus (Preest, 963; Burrows and Lord, 993), numerous attempts have been made to reconstruct the phylogeny of Nothofagus and to test the vicariant nature of its pattern of distribution (Melville, 973; Cracraft, 975; Humphries, 98a; Humphries, Cox, and Neilsen, 986; Linder and Crisp, 995). The analyses of Melville (973) and Cracraft (975) suffered from dependence on a classification erected using misleading morphological characters (van Steenis, 953, 97), whereas Humphries (98a; Humphries, Cox, and Neilsen, 986) analyses, although based on Hennigian concepts, produced poorly resolved trees and provided no support for pollen type as a grouping criterion within the genus. Based on these findings, Humphries (98a) ech-

3 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 39 TABLE. Most recent classification (Hill and Read, 99) of the Nothofagus species sampled in this study. Classification reflects pollen types designated by Dettman et al. (990). Distribution of taxa indicated with the following abbreviations: Aust. Australia, N.C. New Caledonia, N.G. New Guinea, N.Z. New Zealand, S.A. South America, Tas. Tasmania. Subgenus (pollen type) Species Distribution No. of samples Nothofagus (fusca A) Fuscospora (fusca B) Lophozonia (menziesii) Brassospora (brassii) N. antarctica (Forst) Oerst. N. betuloides (Mirb.) Oerst. N. dombeyi (Mirb.) Oerst. N. pumilio (Poeppl. & Endl.) Krasser. N. nitida (Phil.) Krasser. N. alessandri Espin. N. gunnii (Hook.) Oerst. N. truncata (Col.) Ckn. N. fusca (Hook) Oerst. N. solandri (Hook.) Oerst. N. alpina (Poeppl. & Endl.) Oerst. N. glauca (Phil.) Krasser. N. obliqua (Mirb.) Oerst. N. menziesii (Hook.) Oerst. N. cunninghamii (Hook.) Oerst. N. moorei (Muell.) Krasser. N. aequilateralis (Baill.) Steen. N. balansae (Baill.) Steen. N. perryi Steen. N. resinosa Steen. N. brassii Steen. N. grandis Steen. S.A. S.A. S.A. S.A. S.A. S.A. Tas. N.Z. N.Z. N.Z. S.A. S.A. S.A. N.Z. Aust. Tas. Aust. N.C. N.C. N.G. N.G. N.G. N.G oed Patterson s (98, p.483) claim that Nothofagus is uninformative on the interrelationships of southern hemisphere areas and suggested that poor resolution in the area cladogram was due to the evolution of several species groups before the breakup of Gondwana. Using Nothofagus as an example, Humphries (98a, b) convincingly argued that narrative theories of biogeography are often based on unreasonable assumptions and incomplete knowledge of phylogenetic relationships; however, his own conclusions on the historical biogeography of Nothofagus remain unsatisfactory. Recent phylogenetic studies based on morphology (Hill and Jordan, 993) and DNA sequences from the chloroplast rbcl gene (Martin and Dowd, 993) have greatly increased our understanding of Nothofagus phylogeny, but they are not strongly supported and disagree about relationships among major subgeneric groups. However, both studies provide evidence for the phylogenetic importance of pollen type. The general morphology of these pollen types distinguishes four monophyletic groups within the genus. This is consistent with the most recent treatment of the genus (Hill and Read, 99; see Table ), wherein the four groups are considered subgenera. If pollen type is indicative of monophyly within extant Nothofagus, then according to the distribution of taxa within subgenera the minimum number of trans-antarctic relationships within the genus is two (Table ). Nonetheless, increased phylogenetic resolution within and among monophyletic groups could potentially infer additional trans-antarctic and other putative vicariant relationships within Nothofagus. A recent biogeographical analysis of Nothofagus based on combined parsimony analysis of the two data sets discussed above suggested that extinction and sympatric speciation have shaped the modern distribution of Nothofagus (Linder and Crisp, 995). However, the phylogenetic hypothesis produced in that study was weakly supported, mostly due to conflicting phylogenetic signal between data sets. At present, no worker has critically examined these primary data sources and presented additional data to rigorously address both the monophyly of the groups recognized by pollen morphology and their higher order relationships. In particular, the broader relationships of the tropical, morphologically divergent taxa of New Guinea and New Caledonia (subgenus Brassospora) have added to the complexity of Nothofagus systematics since their discovery in the late 940s. Its morphologically distinctive characteristics have prompted several workers to recognize this subgenus at the generic level (Baumann-Bodenheim, 983) or to suggest an early divergence from cool-temperate species (Hjelmqvist, 963). In light of temporal data drawn from the fossil record, it is clear that a fully resolved, well-supported phylogenetic hypothesis for Nothofagus is needed to evaluate the nature of the processes that may have produced the trans- Antarctic relationships. In this vein, the question of whether all trans-antarctic patterns of relationship within Nothofagus are best explained by vicariance, extinction, or secondary dispersals is key to understanding the importance of Nothofagus in Southern Hemisphere biogeography. To further explore the systematics and historical biogeography of Nothofagus, I present a phylogenetic analysis of nucleotide sequences from 22 species of Nothofagus and two outgroups using the ITS region of nuclear ribosomal DNA (nrdna). To independently evaluate the phylogenetic hypothesis obtained from nrdna, the same two previously published data sets bearing on the relationships of Nothofagus are reanalyzed and compared for taxonomic congruence, including an assessment of the appropriateness of combining the three data sets. Lastly, morphological character evolution within the genus is examined within a phylogenetic context.

4 40 AMERICAN JOURNAL OF BOTANY [Vol. 84 MATERIALS AND METHODS Nuclear ribosomal ITS sequencesplant material was collected from naturally occurring populations or cultivated plantings. Vouchers are deposited at the Duke University Herbarium (DUKE). All species representing the cool-temperate element of the genus (South America, New Zealand, and Australia) were sampled (Table ). For most of these taxa, more than one individual was sampled for intraspecific sequence variation. Because of difficulties in obtaining tissue from all species representing the tropical to subtropical distribution of the genus, exemplars of subgenus Brassospora were sampled: four species were sampled from New Guinea and two from New Caledonia. New Guinean Nothofagus sampled (graciously provided by Peter Martin) are the same as those used by Martin and Dowd (993) to generate cpdna rbcl sequences (see below). Total genomic DNA was isolated from fresh, silica-gel dried, and/or air-dried leaf material following Manos, Nixon, and Doyle (993). Sequence was determined for the complete ITS and most of the 5.8s gene of nrdna. Primer pairs ITS-5/ITS-2 and ITS-3/ITS-4 (sequences given in Baldwin, 992) were used for single-stranded amplifications following Baldwin (992; see his fig. ). Double-stranded amplifications were obtained using the same methods, but modified to account for equal volumes of each 0 mol/l primer. All sequences were obtained using two methods of manual sequencing. For most DNAs, single-stranded products were sequenced following Baldwin (992). For several DNAs, double-stranded products were sequenced using the specifications suggested in the Stratagene Cyclist Exo Pfu DNA sequencing Kit. For each DNA, both strands were sequenced. Two DNAs were sequenced automatically to verify manually generated sequences using dye terminator reactions and the protocols provided in the supporting manual of the Applied Biosystems 373 DNA Sequencer. All ITS DNA sequences were aligned manually by first comparing sequences from species and species groups of Nothofagus reported to be closely related on the basis of morphological evidence. Once these alignments were determined, more divergent groups of sequences were compared until all Nothofagus sequences and comparable outgroup sequences were aligned. The ITS sequences also were evaluated for several standard descriptive parameters, including size, percentage G C content, percentage pairwise divergence, percentage of aligned sites with gaps, and percentage of phylogenetically informative sites with gaps. Ribosomal DNA data analysisfrom sequence alignments, a final data matrix was constructed for ITS and 5.8s nrdna sequences (Appendix ). Indels were treated for phylogenetic analysis in three ways: () deletions were treated as missing data, (2) characters corresponding to indel positions were not included, (3) indels were coded as presence/ absence characters in a supplemental matrix. For the few overlapping indels (e.g., positions and ) where homology assessment was impossible, deletions were coded as missing in the corresponding cells of the supplemental matrix. Sequence data were analyzed using the maximum parsimony algorithms implemented with PAUP version 3. (Swofford, 993) and Hennig86 (Farris, 988). Preliminary results of a broadscale phylogenetic analysis of taxa related to Nothofagus ( higher hamamelids) using coding regions of rdna and ITS sequences indicated that two genera of ceae ( and ) are appropriate outgroups on the basis of sequence alignment (Manos and Steele, 996). Data were analyzed with PAUP using 00 randomorder entry replicate searches and the branch-and-bound search options. The shortest trees recovered by PAUP were verified by using Hennig86 and the command IE or implicit enumeration. Relative support for clades recovered by parsimony was assessed with 00 bootstrap replicates. Morphological data analysisseveral morphological phylogenetic analyses of Nothofagus have been published (Melville, 973; Humphries, 98; Humphries, Cox, and Neilsen, 986; Philipson and Philipson, 988); however, only Hill and Jordan s (993) analysis is based on critical evaluation of morphological variation (see Hill and Read, 99). For this study, Hill and Jordan s (993) study was used as a guide for outlining patterns of morphological variation. Though most characters from their study also were used here, a number of characters and character states were reinterpreted for this study (see Appendix ). A revised data set was constructed for analysis using unweighted parsimony analysis (Appendices 2 and 3). Only taxa included in the ITS and rbcl studies were considered; thus, exemplar taxa were used to represent New Guinean and New Caledonian species. Specimens were surveyed from four herbaria (DUKE, CU, US, GH) to verify the taxonomic distribution of many of the macroscopic character state observations made by Hill and Jordan (993) for the 22 species of Nothofagus included in the present analysis. The distribution of microscopic pollen characters also was reviewed in Hanks and Fairbrothers (976), Praglowski (982), and Dettman et al. (990). The revised matrix consists of 23 characters; 6 are coded as binary and seven are coded as nonadditive multistate characters (Appendices 2 and 3). Possible outgroups for Nothofagus include taxa of the higher Hamamelididae, a clade of predominantly wind-pollinated families. Historically, Nothofagus has been considered the sister genus to Fagus, with both comprising subfamily Fagoideae of Fagaceae (see Kubitzki, 993 for a modern version of this traditional treatment). However, recent phylogenetic analyses of morphology (Nixon, 989), cpdna restriction sites (Manos, Nixon, and Doyle, 993) and cpdna sequences (Soltis et al., 995; Manos and Steele, 996, in press) support the exclusion of Nothofagus from Fagaceae and recognition of the genus at the family level. Hill and Jordan (993) concluded that even though evidence exists to question a close relationship of Fagus and Nothofagus, the morphologically divergent family ceae, which Nixon (989) suggested to be sister to Nothofagus, provided little relevance for obtaining character state polarities, and thus they employed Fagus as outgroup. Fagus also was used here to root trees and polarize characters because it shares several vegetative and reproductive features with the ingroup. CpDNA rbcl data analysismartin and Dowd (993) presented a parsimony-based phylogenetic analysis of cpdna sequences of the rbcl gene obtained from 23 species of Nothofagus and two outgroup taxa, Liquidambar (Hamamelidaceae) and Fagus. For this study, Nothofagus sequences were obtained from Genbank (L , L3350-L3363) and reanalyzed using Fagus (L3340), (L2634), and (X5669) as outgroups. Alignments were unambiguous with no missing data. Parsimony analysis with equal character weighting was conducted on a matrix of 345 bp, including sequences from the same taxa sampled in the ITS study. RESULTS Nuclear ribosomal DNA datathe boundaries of the internal transcribed spacers (ITS, ITS 2) and nrdna coding regions in the 24 taxa included here were determined by comparison to several published sequences obtained from a range of angiosperms (Yokota et al., 989; Venkateswarlu and Nazar, 99; Baldwin, 992; Wojciechowski et al., 993). No differences were observed between manually generated sequences and automated sequences. Comparative descriptions and analyses of sequences for all taxa are based on 588 bp of the ITS/5.8s region (Appendix ). Sequence data for the 5.8s rdna gene were incomplete for most taxa; however, the same region of partial sequence obtained for all taxa was included in the final analyses (33 bp of a total of 64 bp). The length of ITS varied from 28 to 228 bp, whereas that of ITS 2 varied from 204 to 227 bp. Percentage G

5 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 4 TABLE 2. Pairwise distances between selected species of Nothofagus and two outgroups. Absolute distances are shown below the diagonal and mean distances (adjusted for missing data) are shown above the diagonal. Species N. antarctica 4. N. betuloides 5. N. alessandri 6. N. gunnii 7. N. solandri 8. N. grandis 9. N. balansae 0. N. alpina. N. cunninghamii 2. N. menziesii 3. N. moorei Fig. 2. One of two most-parsimonious trees for Nothofagus based on ITS nrdna sequences. Length 246 steps, CI 0.76, RI 0.88; branch lengths are drawn according to scale. Clade designations follow the subgeneric classification of Hill and Read (99; see Table ). Percentage of 00 bootstrap replications is given for nodes with bootstrap values 50%. C content across the entire sequence varied from 62.7% in to 53.4% in Nothofagus pumilio. The highest percentage G C within Nothofagus was observed for N. alpina (58.3%). The ITS lengths and general pattern of nucleotide composition in the sequences obtained from Nothofagus and outgroup taxa are similar to those reported for a broad sample of angiosperms (Baldwin et al., 995). Alignment of the final matrix required the introduction of 2 small - or 2-bp indels (insertion or deletion mutations) distributed throughout ITS and 2, and several larger indels as shown in Apppendix. In some instances, small indels overlap with larger indels and the two largest indels overlap for much of their span. Twenty-three of 228 positions (0.%) within ITS had at least one sample with a gap. Sixteen of these were required to align Nothofagus and outgroup sequences. Similarly, 36 of 227 positions within ITS 2 (6.3%) had at least one gap, seven of which were required to align Nothofagus and outgroup sequences. For the entire sequence, values of pairwise percentage sequence divergence ranged from.2 to.% among Nothofagus species, and from 9.3 to 24.8% between Nothofagus and outgroup taxa (Table 2). Intraspecific sampling of morphologically defined species resulted in only one case of ITS sequence variation. For the three individuals of Nothofagus dombeyi examined, two possessed ITS sequences that closely matched sequences from individuals of N. pumilio, whereas one was more similar to the sequences obtained from the three individuals of N. antarctica. Any resolution among these species is not considered adequate for inferring organismal relationships and further sampling within these species is necessary to establish the extent of intraspecific ITS variation within these species. For convenience, one sequence from N. dombeyi was included in the analysis. Within Nothofagus, a total of 92 positions were phylogenetically informative (Appendix ): 43 in ITS, six in 5.8s, and 43 in ITS 2. For ITS, three out of 43 phylogenetically informative sites included gaps; for ITS 2, nine out of 43 sites included gaps. Therefore, 2 (3%) of the phylogenetically informative positions of the final matrix include at least one species with gaps. The level of phylogenetically informative nrdna ITS sequence variation detected within Nothofagus is approximately five times greater than that from a comparable length of cpdna rbcl sequence in a similar sample of taxa (Martin and Dowd, 993). When sequences from all taxa were considered and gaps treated as missing data, parsimony analysis recovered two equally most-parsimonious trees of 246 steps. One of these trees is shown with branch lengths drawn to indicate the relative number of changes per branch under ACCTRAN optimization (Fig. 2). The other tree differs only in the relative placement of N. fusca and N. truncata, which are resolved as sister taxa. Upon removal of the 2 phylogenetically informative sites containing

6 42 AMERICAN JOURNAL OF BOTANY [Vol. 84 Fig. 3. Strict consensus tree of eight equally parsimonious trees for Nothofagus based on morphological characters of Hill and Jordan (993) modified and reanalyzed (Appendices 2 and 3). Length 46 steps, CI 0.70, RI Clade designations follow the subgeneric classification of Hill and Read (99; see Table ). Percentage of 00 bootstrap replications is given for nodes with bootstrap values 50%. gaps (Appendix ; sites 03, 70, 383, 387, 39, 393, 399, 404, 539, 543, 575, and 576) parsimony analysis recovered 2 equally most-parsimonious trees (CI 0.72; RI 0.88). The strict consensus (not shown) is similar to the trees based on all data, except for decreased resolution within subgenus Nothofagus. A third parsimony analysis conducted on an expanded matrix with indels coded as presence/absence characters reinforces the topology shown in Fig. 2 by increasing the bootstrap support for most of the groups recovered from the analysis with complete sequences. Bootstrap support for all higher order structure and several other clades is moderate to high (Fig. 2). Outgroup rooting suggests two major lineages within Nothofagus, one comprising a nested set of three of Hill and Read s (99) subgenera ((Brassospora- Nothofagus)- Fuscospora), the other represented solely by subgenus Lophozonia. Within subgenus Brassospora there is good support for two monophyletic groups: the New Guinean species and the New Caledonian species. Irrespective of the ITS polymorphism detected among individuals of N. dombeyi of the South American subgenus Nothofagus, the other evergreen species N. nitida and N. betuloides are supported as sister taxa, whereas the two deciduous species, N. pumilio and N. antarctica, are not. Within Fuscospora, there is moderate support for a sister relationship between the deciduous species N. alessandri and N. gunnii. Basal to this grouping is a grade of evergreen New Zealand species, N. fusca, N. truncata, and N. solandri. Within Lophozonia, two clades are well supported, one formed by the three deciduous South American species, the other by the three evergreen species. In the latter group, there is strong support for a close relationship between N. cunninghamii from southeastern Australia and Tasmania and N. moorei from eastern Australia, relative to the relationship of either taxon to the New Zealand species, N. menziesii. Morphological dataphylogenetic analysis of Nothofagus using Fagus as outgroup resulted in eight equally most-parsimonious trees of 46 steps. In all trees, three of the four subgenera of Nothofagus form monophyletic groups, the exception being subgenus Fuscospora. A strict consensus (Fig. 3) summarizes the branching pattern common to the four trees; topology varies within Fuscospora, whereas the polytomies within subgenus Lophozonia and Brassospora are due to lack of resolution in all eight trees. In the consensus, two higher order monophyletic groups of Nothofagus are supported in the same hierarchical arrangement as the ITS tree: (Lophozonia- (Fuscospora- (NothofagusBrassospora))). The larger clade includes three of the four subgenera and the smaller clade comprises subgenus Lophozonia. Hill and Jordan s (993) phylogenetic hypothesis differs mainly in the higher order branching relationships among monophyletic groups, which they resolved in the following arrangement: (Fuscospora- (Nothofagus- (Lophozonia Brassospora))). Given that the morphological data set contains only 23 characters it is not surprising that several of the branches are weakly supported. However, within each of the cool-temperate subgenera (see Fig. 3), there is moderate support for the following groups of evergreen species: (N. dombeyi- (N. beluloidesn. nitida)), (N. fusca -N. truncata), and (N. cunninghamii- (N. menziesii N. moorei)). CpDNA datafor the cpdna rbcl sequences published by Martin and Dowd (993), 43 of a total of 345 sites were phylogenetically informative within Nothofagus. Reanalysis of these sites using parsimony and Fagus,, and as outgroups produced 53 equally most-parsimonious trees of 38 steps. One of the mostparsimonious trees is presented to indicate branch lengths (Fig. 4). This tree is equivalent to the strict consensus, except for the positions of N. cunninghamii and N. moorei, which were otherwise unresolved. The overall tree topology is similar to the trees obtained for ITS data. The genus is divided into two clades with strong to moderate bootstrap support for the monophyly of the recognized subgenera. Support for the initial divergence within the genus and monophyly of Lophozonia is strongest; however, several nodes providing structure within the larger clade are weakly supported, most notably the branch subtending the sister group relationship of the subgenera Brassospora and Nothofagus. Differences relative to the ITS and morphology trees include the following: () for subgenus Nothofagus, strong support in the rbcl trees for the deciduous species N. antarctica and N. pumilio as sister taxa, (2) for Fuscospora, moderate support in the rbcl trees for a clade of N. fusca, N. solandri, and N. truncata and no support for a sister species relationship between N. gunnii and N.

7 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 43 Fig. 4. One of 53 most-parsimonious trees for Nothofagus based on chloroplast DNA rbcl sequences published by Martin and Dowd (993). Length 38 steps, CI 0.67, RI 0.82; branch lengths are drawn according to scale. Clade designations follow the subgeneric classification of Hill and Read (99; see Table ). Percentage of 00 bootstrap replications is given for nodes with bootstrap values 50%. alessandri, and (3) for Lophozonia, low resolution in the rbcl trees overall, including no support for a clade uniting N. glauca, N. alpina, and N. obliqua. Most internal branch lengths appear to be uniformly short as indicated by the distribution of site changes on several of the minimum length trees. Martin and Dowd (993) reported rate homogeneity in these same rbcl sequences, but they also pointed out that a large degree of uncertainty surrounds this result, given the relatively small number of changes per branch. Combined datacombined analysis of the three data sets (ITS, cpdna, and morphology) is justified because the higher order structure of the trees produced by each independent analysis was identical. Within the monophyletic groups (subgenera sensu Hill and Read, 99) supported by each analysis, several relationships of terminal taxa well-nested within monophyletic groups conflict across data sets; however, these incongruities often are not strongly supported. Combined analysis could potentially resolve these species-level relationships, especially those supported weakly or unresolved in each independent analysis. From the combined total of 6 phylogenetically informative characters, parsimony analysis recovered six equally most-parsimonious trees of 443 steps. The difference between the number of extra steps from the combined data and the sum of the steps for the three individual data sets, adjusted to include the same outgroup taxa (47 steps), was 26 steps. Incongruence between data sets was low as determined by the Mickevich and Farris Fig. 5. Strict consensus tree of six equally parsimonious trees for Nothofagus based on combined analysis of rdna ITS sequences, chloroplast DNA rbcl sequences, and morphological characters. Length 443 steps, CI 0.70, RI Clade designations follow the subgeneric classification of Hill and Read (99; see Table ). Percentage of 00 bootstrap replications is given for nodes with bootstrap values 50%. (98) index (I MF 0.06). The strict consensus (Fig. 5) shows two collapsed nodes that are resolved in various combinations among the six minimum length trees. The combined data consensus tree provides strong support for the four subgenera of Hill and Read (99) (Table ) and their interrelationships. Combined data reinforce relationships within the following monophyletic groups: () strong support for two clades within subgenus Brassospora, (2) resolution within subgenus Fuscospora, including strong support for a clade comprising the New Zealand species, N. fusca, N. solandri, and N. truncata relative to the successively basal N. gunnii and N. alessandri, and (3) complete resolution and strong to moderate support within subgenus Lophozonia. One of the most-parsimonious trees is presented as a basis for discussing patterns of morphological character state evolution (Fig. 6; Appendices 2 and 3). The distribution of morphological synapomorphies in the combined analysis is generally similar to the pattern produced in the morphological phylogenetic analysis because Lophozonia and Fagus, the outgroup for the morphological analysis, share numerous putatively plesiomorphic character states. This is not the case for characters 7 and 22, which have a different distribution in the combined analysis as Fagus and Lophozonia have different character states. Subgenus Lophozonia is monophyletic and supported by cupules with simple trichomes (character 3 character state ), an open staminate perianth bearing

8 44 AMERICAN JOURNAL OF BOTANY [Vol. 84 of the distribution of phylogenetically informative characters within Brassospora based on more complete sampling, see Hill (994). DISCUSSION Fig. 6. One of six most-parsimonious trees for Nothofagus based on combined analysis of nrdna ITS sequences, chloroplast DNA rbcl sequences, and morphological characters. Length 443 steps, CI 0.70, RI The distribution of morphological characters is shown with boxes; nonhomoplasious characters indicated with solid boxes; homoplasious characters indicated with open boxes. Character number indicated above boxes and character state indicated below box (see Appendices 2 and 3). Deciduous taxa are shown in boldface. Figure abbreviations: AUST Australia, NC New Caldedonia, NG New Guinea, NZ New Zealand, SA South America, Tas Tasmania. 20 stamens (6 ), and pollen with V-shaped colpus ends (0 ). The larger monophyletic group containing the remaining subgenera is supported by cupules with membranous lamellae (2 2), pollen with short colpi (8 ) that are internally thickened (9 ), and leaves with solitary type C unicellular trichomes (8 ). Subgenus Fuscospora is monophyletic and supported by the presence of unicellular trichome type A (7 0) and absence of conical unicellular trichomes (9 ). Within Fuscospora, a monophyletic subclade comprising N. solandri, N. fusca, and N. truncata is supported by revolute leaf vernation (3 2), and N. fusca and N. truncata are further supported by leaves with a complete fimbrial vein (5 ). The sister-group relationship of Brassospora and Nothofagus is supported by peltate stipules (2 ) and evensized stomata (23 0). Subgenus Nothofagus is monophyletic and characterized by pollen colpi with external thickenings (8 2), leaves without T-pieces of cutin at stomatal poles (20 ), and with stomata oriented parallel to the long axis of the leaf (2 ). Subgenus Brassospora is monophyletic and supported by cupules with two symmetrical valves ( ), dimerous pistillate flowers (5 2), pollen with constricted polygonal amb (7 ), wood without tracheids ( ), conduplicate leaf vernation (3 3), and giant stomata (22 0). Within Brassospora, the monophyly of N. balansae and N. aequilateralis is supported by spiral leaf arrangement (4 ). For further discussion Phylogeny and taxonomic congruencephylogenetic analysis of ITS sequences supports previous phylogenetic hypotheses based on rbcl sequences (Martin and Dowd, 993) and morphology (Hill and Jordan, 993), and the derivative combined analysis of Linder and Crisp (995) in recognizing four monophyletic groups within Nothofagus, in agreement with those groupings indicated by pollen morphology (Dettman et al., 990). The analysis of rbcl sequences by Martin and Dowd (993) was the first to suggest the presence of two basal clades within Nothofagus. Analysis of ITS sequences (Fig. 2) and reanalysis of morphology (Fig. 3) in this study independently support this divergence. ITS sequence provides the strongest support for the hierachical arrangement of the four monophyletic groups of Nothofagus, as exemplified by the higher boostrap support for the sister-group relationship of the subgenera Brassospora and Nothofagus (Fig. 2). Because taxonomic congruence (sensu Mickevich, 978; Kluge, 989) was observed across the data sets, parsimony analysis of the combined data offers the best available estimate of the phylogeny of Nothofagus (Fig. 5). The phylogeny based on all three data sets (Fig. 5) strongly supports the sister-group relationship between the tropical subgenus Brassospora (New Guinea and New Caledonia) and the cool-temperate subgenus Nothofagus (South America). Although earlier classifications and phylogenetic hypotheses of Nothofagus consistently recognized subgenus Brassospora on the basis of its distinctive morphology (e.g., van Steenis, 953, 97; Humphries, 98a; Philipson and Philipson, 988), its relationship to cool-temperate taxa remained obscure. Phylogenetic analyses of morphology and combined data suggest that many of the distinctive morphological character states of subgenus Brassospora are apomorphic within Nothofagus (see Fig. 6). Taxonomic and inferred phylogenetic relationships among the 6 cool-temperate species prior to the classification of Hill and Read (99; Table ) had been unstable and based on few characters. Hill and Read (99) attempted to use cupule and leaf morphology to form phenetic groups of species, but their recent classification is largely based on pollen morphology. Phylogenetic data presented here support that classification by resolving three monophyletic groups of cool-temperate species, but indicate that the ten cool-temperate species of Nothofagus and Fuscospora are more closely related to the tropical species of Brassospora than to cool-temperate species of Lophozonia. Interestingly, species of subgenus Nothofagus were often grouped with species of subgenus Fuscospora on the basis of pollen similarity, (fusca-type pollen of Cookson and Pike, 955), leaf morphology, and habit. These and other morphological similarities between the two groups represent likely plesiomorphies or convergences within Nothofagus. As demonstrated by morphological phylogenetic analysis presented here, Fuscospora in particular has retained numerous plesiomorphic characters and thus its

9 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 45 monophyly is not supported in all morphological trees (Fig. 3). The arrangement of the four groups resolved in each analysis is counter to that produced in Hill and Jordan s (993) recent cladistic analysis based on morphology. The difference in the resolution obtained in that study, (Fuscospora- (Nothofagus- (LophozoniaBrassospora))), relative to the one produced here (Fig. 3), may be due to several factors. For example, three- and four-state transformation series are proposed in their study for four pollen characters, many of which are not supported in the pollen literature (Hanks and Fairbrothers, 976; Praglowski, 982; Dettman et al., 990). By replacing the four pollen characters in Hill and Jordan s original data with the three pollen characters designated here (see discussion below), a resolution similar to the one shown in Fig. 3 is obtained. Morphological implicationsearlier morphological investigations into the pollen of Nothofagus (e.g., Cranwell, 939; Cookson and Pike, 955; Hanks and Fairbrothers, 976; Praglowski, 982) recognized only three pollen types within Nothofagus, named after characteristic species representing each of the types (fusca, menziesii, and brassii types; see Table ). Quantitative studies of the three pollen types suggested that the menzesii pollen type (subgenus Lophozonia) is distinct from the more similar fusca and brassii pollen types (Hanks and Fairbrothers, 976; Praglowski, 982). Additional morphological characters and numerous molecular characters also support gross pollen morphology by resolving two clades of Nothofagus. By recognizing a fourth pollen type from within the fusca type, Dettman et al. (990) established two phenetic groupings that are strongly supported as clades (subgenera Nothofagus and Fuscospora) by phylogenetic analysis of ITS and cpdna sequences (Martin and Dowd, 993) and to a lesser extent, morphological characters other than those from pollen. Although grouping Nothofagus species according to generalized pollen morphology apparently has phylogenetic significance, it has been difficult to determine apomorphic pollen character states. Hill and Jordan (993) were the first to incorporate hypothesized pollen transformation series into phylogenetic analysis with other morphological characters; however, their delimitation of character states from continuous size and shape variables reported by other workers appears unjustified in most cases. The systematic approach taken in this study attempted to maximize the information content of pollen morphological differences by using readily observable character states, most of which are polarized. In the current phylogenetic context, several evolutionary changes in pollen morphology are noteworthy (Fig. 6); pollen synapomorphies for the large clade of Nothofagus include consistently shorter colpi (8 ) with internal thickenings (9 ). Reduction in colpus length also is synapomorphic in related clades of higher hamamelids on the basis of morphological (Nixon, 989) and molecular inferences (Manos, Nixon, and Doyle, 993; Manos and Steele, in press). With the exception of subgenus Fuscospora, which could not be defined by a single pollen apomorphy, the remaining three pollen types possess derived features to support monophyletic groups of Nothofagus. Within the large clade of Nothofagus, constricted-polygonal amb in polar view (Fig. 6, 7 ) is synapomorphic for the species of subgenus Brassospora, and conspicuous outward colpus thickenings (Fig. 6, 9 2) is synapomorphic for species of subgenus Nothofagus. The only possible pollen synapomorphy for subgenus Lophozonia is the morphology of the internal end of the colpus, which is V-shaped instead of the U-shaped type observed in all other Nothofagus. However, the polarity of this condition remains equivocal. Morphological variation in cupule and pistillate flowers were key features in previous classifications and evolutionary interpretations of species relationships of Nothofagus (e.g., van Steenis, 953). In considering the patterns of variation in cupules and flowers in the current phylogenetic context, it is apparent that the large clade of Nothofagus possesses the bulk of within-group character-state change, whereas the species of Lophozonia are invariant for the characters scored (Appendices 2 and 3). Hill and Jordan (993) considered the presence of glandular lamellae on the cupules of species of subgenus Lophozonia to be synapomorphic. Because the polarity of this character is uncertain, the loss of glands and transformation to the membranous condition could be interpreted as a synapomorphy for the large clade of Nothofagus (Fig. 6, 2 2). Support for this interpretation is based on the hypothesis that the cupule of Nothofagus is stipular in origin (van Steenis, 953), suggesting that colleters found on the stipules of Nothofagus species may be homologous with the glands of individual cupule lamellae. Thus their presence may be the ancestral condition within the genus. Given this interpretation, the branching aspect of this multicellular structure might then be considered synapomorphic for the six species of subgenus Lophozonia. The stabilized arrangement of cupules and flowers found in most cool-temperate Nothofagus is a symmetrical four-valved cupule that subtends a dichasium of three flowers, two tricarpellate and lateral and one bicarpellate and central. Within the large clade of Nothofagus, morphological changes include two independent reductions in the number of cupule valves and at least one occurrence of complete loss of the cupule in N. resinosa (Fig. 6, 3). From the plesiomorphic condition of four-valved cupules, two distinct derivations of two-valved cupules are supported in all trees (Fig. 6). In Brassospora, two-valved cupules subtend 3 bicarpellate flowers (depending on the species), with a symmetric orientation of valves suggesting fusion of adjacent cupule valves ( ). In N. pumilio, a member of subgenus Nothofagus, two-valved asymmetric cupules subtend a single tricarpellate flower suggesting the loss of the opposing pair of cupule valves ( 2). This latter interpretation requires the concommitant loss of adjacent flowers of the dichasium, leading to the asymmetrical cupule/fruit arrangement. Developmental studies are needed to confirm this. Reduction in cupule valve number also may be seen in N. solandri of subgenus Fuscospora. This species often bears three-valved cupules and one of the three valves is thicker than the other two, suggesting fusion between adjacent valves. This condition is similar to that hypothesized to account for each valve of the two-valved cupules in Brassospora; however, the species does not constantly express the three-valved condition (Hill and Read, 99).

10 46 AMERICAN JOURNAL OF BOTANY [Vol. 84 Flower number and type also have phylogenetic significance within Nothofagus. Higher numbers of flowers per cupule are known to occur sporadically on individuals of certain species (Hill and Read, 99; Herendeen, Crane, and Drinnan, 995), but the nearly constant expression of a seven-flowered cupule in N. alessandri prompted earlier workers to hypothesize an ancestral position for the species (van Steenis, 953; Melville, 973). Although the condition of a more complicated inflorescence might be indicative of the ancestral state in Nothofagus, the seven-flowered cupule is phylogenetically uninformative in the context provided here. Of greater phylogenetic significance is the distribution of apomorphies pertaining to reduction in flower and carpel number from the more generalized ancestral condition described above. Subgenus Brassospora, for example, is defined by a change to strictly bicarpellate flowers (Fig. 6, 5 2). Reduction to a single bicarpellate flower per cupule (4 2) is restricted to species of Brassospora, where the condition apparently has arisen multiple times (Hill and Jordan, 993; Hill, 994). The single-flowered condition also is found in N. pumilio of subgenus Nothofagus; however, because it is tricarpellate, its origin is probably different from the single-bicarpellate-flowered condition observed in Brassospora. The structure of the staminate inflorescence has provided subtle characters to aid in the identification of certain species of Nothofagus, but the general morphology at higher taxonomic levels is phylogenetically valuable. The shape of the perianth in most species and the outgroup Fagus is essentially tubular with 20 stamens per flower; however, the condition for most species of Lophozonia is that of a reduced, flat, or open perianth bearing 20 stamens per flower (Fig. 6, 6 ). Recent ontogenetic studies suggest that this condition is the result of the fusion of several flowers within dichasia (Rozefelds and Drinnan, 994), and therefore is derived within the genus, further supporting the monophyly of subgenus Lophozonia. Biogeographic implicationsphylogenetic analysis of combined data has produced a well-supported, fully resolved phylogeny to serve as a framework for the interpretation of the historical biogeography of Nothofagus (Figs. 5, 6). This phylogeny supports four monophyletic groups: two groups with a trans-antarctic distribution of taxa (subgenera Fuscospora and Lophozonia) and two groups with relatively localized distributions in South America (subgenus Nothofagus) and New Guinea and New Caledonia (subgenus Brassospora). Although these distribution patterns were predicted by Hill and Read s (99) classification, data from ITS and rbcl (Martin and Dowd, 993) have contributed greatly to reconstructing hierarchical relationships among and within monophyletic groups, including the detection of a third trans-antarctic relationship for sister subgenera Nothofagus and Brassospora (Figs. 3, 4). A reduced area cladogram illustrates the three trans-antarctic relationships within Nothofagus (Fig. 7). To adequately test the vicariance model for Nothofagus, knowledge of precise interrelationships must be reconciled with the past and present distribution of the genus. Data on fossil pollen assignable to modern pollen types support a continuous ancestral distribution for the four modern Fig. 7. Area cladogram for Nothofagus based on the tree shown in Fig. 6. Figure abbreviations: AUST Australia NC New Caledonia, NG New Guinea, NZ New Zealand, and SA South America. groups, spanning from South America to Australia by at least the Eocene (Dettman et al., 990). Vicariance is generally supported in the area cladogram showing trans-antarctic relationships within two of the four groups of Nothofagus (Figs. 6, 7). For the subgenera Brassospora and Nothofagus, groups now geographically restricted, fossil pollen bearing apomorphies found in extant species groups indicate that extinction events in the Tertiary eliminated other likely within-section trans-antarctic relationships, resulting in a biogeographic pattern between subgenera that superficially resembles vicariance (Fig. 8A). Leaf and cupule fossils similar to N. betuloides and N. dombeyi (subgenus Nothofagus) recovered from Oligocene deposits in Tasmania (Hill, 99) further support the presence and later extinction of this group from at least one austral area. There are no macrofossils to support the presence of subgenus Brassospora in South America; however, cupules bearing apomorphies (e.g., one fruit per cupule; Hill, 994) observed in living species are known from Oligocene deposits in Tasmania and leaves resembling living New Caledonian species are known from the Miocene of New Zealand (Campbell, 985). The occurrence of Brassospora in New Caledonia typically has been explained by vicariance events within a larger landmass that may have included either New Zealand, Australia, or both in the Late Cretaceous to Early Tertiary (Crook and Belbin, 978; Stevens, 989). The totality of fossil evidence suggests widespread extinction of subgenera Brassospora and Nothofagus, the former from South America, Antarctica, New Zealand, and Australia, and the latter from Antarctica, Tasmania, and New Zealand. Because widespread extinction appears likely in two of the four monophyletic groups of Nothofagus, the contribution of extant Nothofagus to understanding historical relationships among the landmasses of the Southern Hemisphere requires a shift in focus. Only Fuscospora and Lophozonia include living species with distributions indicative of vicariance. On the basis of combined parsimony analysis, each of these clades supports a more recent relationship of New Zealand to Australia (including Tasmania) relative to South America (Fig. 8B, C). Most reconstructions of Gondwana hypothesize an early divergence ( Mya) of New Zealand from a continuous landmass including South America, Antarctica, and Australia (Wilford and Brown, 994). Because the relationship of (SA (AUST, NZ)) is inconsistent with both the accepted history for the break-up of Gondwana (NZ (SA-AUST)), and the biogeographical pattern of (SA, NZ) AUST)) observed for trans-antarctic insect groups

11 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 47 Fig. 8. Biogeographic patterns for subgenera of Nothofagus. (A) Trans-Antarctic relationship between Nothofagus and Brassospora; hypothesized extinctions indicated with dashed lines and parentheses. (B) Trans-Antarctic relationship within Fuscospora. (C) Trans-Antarctic relationship within Lophozonia. (e.g., Brundin, 965; Cranston and Naumann, 99), the pattern within Nothofagus potentially could be explained by dispersal. Indeed, the presence of Nothofagus in the Tertiary of New Zealand often has been attributed to colonizations via long-distance dispersal from Australia (Mildenhall, 980; Truswell, Kershaw, and Sluiter, 987; Hill, 994; Macphail et al., 994). Using the microfossil record for New Zealand, these authors pointed to the Early Tertiary absence of certain pollen types and later appearance to suggest that Nothofagus crossed the Tasman Sea several times in the Early to Mid Tertiary, even though the Tasman Sea is believed to have reached its present width of 850 km by the Early Tertiary, 60 Mya (Stock and Molnar, 982) and the farthest distance recorded for airborne transport of Nothofagus seed is no more than 5 km (Burrows and Lord, 993). For the two extant groups in question, fossil pollen classified as fusca A type first occurs in New Zealand at the Cretaceous Tertiary boundary, 65 Mya, whereas the menziesii type first occurs there in the Eocene, 5 million years later (Dettman et al., 990). Linder and Crisp (995) rejected dispersalist explanations on the basis of congruent phylogenies supporting the (SA(AUST, NZ) pattern of area relationships in several other trans-antarctic plant groups including Nothofagus (based on the combined data sets of Hill and Jordan, 993 and Martin and Dowd, 993), and genera of Winteraceae, Iridaceae, and Elaeocarpaceae. Concerning New Zealand Nothofagus, they argued against the use of absence of fossil pollen to indicate ancestral absence of the genus. To explain the (SA(AUST, NZ) pattern found for plant groups, the vicariance view considers it more parsimonious to invoke ancestral presence and subsequent extinction from a previously continuous distribution rather than multiple dispersal events across taxonomic groups. If this plant biogeographic pattern accurately reflects a shared history and not concordant dispersal events, the accepted history of the break-up of Eastern Gondwana and biogeographical inferences based on insect phylogenies may require revision (Linder and Crisp, 995). In considering the clades supporting the Australia New Zealand relationship, the occurrence of fossil equivalents for several key taxa provides additional distributional data for inferring historical biogeography. Oligocene fossil equivalents of N. gunnii (Hill, 984, 99) and N. fusca (Hill, 984) of subgenus Fuscospora indicate that each taxon had a broader ancestral distribution, the former from Antarctica and the latter from Tasmania. Similarly, Eocene and Oligocene fossil equivalents of N. moorei and N. cunninghamii have been recovered from Tasmania and New Zealand (see Hill, 99 for review). These fossils suggest that extinction may be confounding patterns of endemism that might otherwise be interpreted as the product of strict vicariance or secondary dispersal. Although the processes of vicariance and extinction appear to have created patterns of endemism, continuous ancestral distribution of taxa in Australia (including Tasmania), New Zealand, and Antarctica supports a single common biogeographical history. Discriminating between hypotheses in support of the ancestral presence of taxa or secondary dispersal as the cause of disjunctions within Fuscospora and Lophozonia with the molecular data at hand is an attractive prospect, but not free of problems. Preliminary data on the relative substitution rate of ITS sequences (Manos and Uyenoyama, unpublished data) suggest that the ITS of subgenus Fuscospora is evolving significantly slower than other Nothofagus lineages and rate heterogeneity exists within subgenus Lophozonia (see Fig. 2). Additional problems concerning the use of ITS sequences for biogeographic hypothesis testing include relatively low numbers of substitutions between trans-antarctic species (see Table 2), a feature shared with rbcl sequences in comparisons of the same taxa (Martin and Dowd, 993).

12 48 AMERICAN JOURNAL OF BOTANY [Vol. 84 At present, faster evolving chloroplast genes are being investigated to increase the number of substitutions between taxa exhibiting the (SA(NZ, AUST) pattern. A combined gene approach using multiple substitution rates and various calibration points taken from the fossil and geologic record may be able to statistically support or refute long-distance dispersal as a biologically realistic event in the biogeographic history of Nothofagus in the Southern Hemisphere. CONCLUSIONS The extensive fossil history of Nothofagus indicates that diversification of the genus was a significant vegetational event in the Late Cretaceous of Gondwana. Four of the eight pollen types described in the fossil literature define the four modern clades independently supported by molecular data. Therefore, the presence of additional pollen types in the fossil record of the genus most likely represents the evolution and subsequent extinction of a considerable portion of this diversification. The explosive radiation of Nothofagus as marked by pollen diversity occurred within a narrow time frame, estimated to be between 83 and 70 Mya (Dettman et al., 990). Molecular data have captured the remaining hierarchical order of cladogenesis from that period, whereas macrofossil data provide additional temporal context by indicating that key evolutionary novelties within the genus had developed by the Oligocene, 40 Mya (Hill, 99). Reconciling the well-established fossil record of Nothofagus with a robust hypothesis of phylogeny suggests that extinction and vicariance events remain the most viable historical explanation for the modern distribution of the genus. For the subgenera Brassospora and Nothofagus, extinction appears to have limited the biogeographical informativeness of living taxa for understanding the biogeography of the Southern Hemisphere. Brassospora is the most species-rich of all subgenera. Although this group remains poorly known at the species level, its modern distribution and basic morphology represent key evolutionary issues within the genus. Morphologically, the group is defined by many apomorphies, including distinctive wood anatomy, conduplicate leaf vernation, and unique pistillate inflorescence structure. That the two major clades are disjunct and only occur at the northern (New Guinea) and eastern (New Caledonia) limits of the range of the genus suggests a more widespread distribution in the past. Indeed, micro- and macrofossils suggest widespread extinction throughout Gondwana, with overland migration to New Guinea during the Miocene (Khan, 974). The sister group to Brassospora is subgenus Nothofagus, a group of five species restricted to the western limit of the range of the genus. Species of this cool-temperate clade often occur in sympatry and several of them superficially resemble species of Fuscospora and Lophozonia. Phylogenetic relationships within the subgenus are inconsistent across the three data sets and the resolution based on combined data is unsatisfactory. Complete resolution at the species level is ambiguous due to intraspecific ITS polymorphism within N. dombeyi. The extent of conflict introduced by sampling error involving cpdna is unknown, as Martin and Dowd (993) included only one sample per species. Hybridization also may be a factor, as sympatric species belonging to the same pollen group are known to be interfertile (e.g., Cockayne and Atkinson, 926). Putative hybrids have been reported between certain combinations of South American species of subgenus Nothofagus (e.g., Donoso and Atienza, 984; Premoli, 996b). The trans-antarctic relationship of this group to subgenus Brassospora probably predates most of the major geological events of the Southern Hemisphere, and therefore may not be indicative of vicariance. The presence of austral micro- and macrofossils unequivocally assigned to this clade suggests widespread extinction leading to isolation of this subgenus in South America. The basal group within the large clade of Nothofagus is the trans-antarctic subgenus Fuscospora. The only South American species is the rare Chilean endemic N. alessandri, which occupies the basalmost position in the clade. Earlier claims of the primitive nature of this deciduous species appear to be justified on the basis of explicit phylogenetic analysis. The sister group to N. alessandri is composed of two basic morphological groups, an evergreen species complex in New Zealand and a single deciduous species, N. gunnii, endemic to Tasmania. The evergreen species are defined by several vegetative features, whereas N. gunnii is largely plesiomorphic for morphological characters. The broader distribution of Oligocene fossil equivalents of several living taxa indicates that current patterns of endemism were produced by extinction and that speciation events may have predated the formation of geologic barriers. Nevertheless, phylogenetic relationships in this section support an area relationship of (SA(AUST, NZ)). Phylogenetic analyses of separate and combined data sets support the recognition of two basal clades within Nothofagus, and resolution of subgenus Lophozonia as sister to the morphologically diverse subgenera of Nothofagus. Given the uncertainty surrounding appropriate outgroups for Nothofagus, the well-supported hierarchical structure within the genus has identified Lophozonia as a functional outgroup for assigning polarity to phylogenetically informative morphological character states within the large clade of Nothofagus. With the exception of morphological characters associated with habit and leaf morphology, Lophozonia is morphologically uniform; however, ITS data strongly support a trans-antarctic relationship indicative of vicariance. Phylogenetic relationships within this section also support the biogeographic pattern of (SA(AUST, NZ)), although the process of extinction appears to have produced current distributions that have been variously interpreted. LITERATURE CITED BALDWIN, B. G Phylogenetic utility of the internal transcribed spacers of nuclear ribosomal DNA in plants: an example from the Compositae. Molecular Phylogenetics and Evolution : 3 6.,M.J.SANDERSON, J.M.PORTER, M.F.WOJCIECHOWSKI, C.S. CAMPBELL, AND M. J. DONOGHUE The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82: BAUMANN-BODENHEIM, M. G Morphologie und klassifikation neu-caledonischer arten der fagaceen-gattung Trisyngyne. Botanica Helvetica 93: BRUNDIN, L On the real nature of transantarctic relationships. Evolution 9: BURROWS, C. J., AND J. M. LORD Recent colonisation by Noth-

13 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 49 ofagus fusca at Cass, Canterbury. New Zealand Journal of Botany 3: CAMPBELL, J. D Casuarinaceae, Fagaceae, and other plant megafossils from Kaikorai Leaf Beds (Miocene) Kaikorai Valley, Dunedin, New Zealand. New Zealand Journal of Botany 23: COCKAYNE, L., AND E. ATKINSON On the New Zealand wild hybrids of Nothofagus. Genetica 8: 43. COOKSON, I. C., AND K. M. PIKE The pollen morphology of Nothofagus Bl. subsection Bipartitae Steen. Australian Journal of Botany 3: CRACRAFT, J Historical biogeography and earth history: Perspectives for a future synthesis. Annals of the Missouri Botanical Garden 62: CRAME, J. A Origins and evolution of the antarctic biota. Geological Society, London. CRANSTON, P. S., AND I. D. NAUMANN. 99. Biogeography. In I. D. Nauman [ed.], The insects of Australia, 2d ed., vol., Melbourne University Press, Melbourne. CRANWELL, L. M Southern beech pollen. Records of the Auckland (N. Z.) Institute and Museum 2: Nothofagus: living and fossil. In J. L. Gressitt [ed.], Pacific basin biogeography, Bishop Museum Press, Honolulu, HI. CROOK, K. A. W., AND L. BELBIN The southwest Pacific area during the last 90 million years. Journal of the Geological Society of Australia 25: DARLINGTON, P. J iogeography of the southern end of the world. Harvard University Press, Cambridge. DETTMANN, M. E., D. T. POCKNALL, E.J.ROMERO, AND M. C. ZAMALOA Nothofagidites Erdtman ex Potonie, 960: a catalogue of species with notes on the paleogeographic distribution of Nothofagus Bl. (Southern Beech). New Zealand Geological Survey, Lower Hutt, New Zealand. DONOSO, C., AND J. ATIENZA Hibradacion natural entre Nothofagus betuloides (Mirb.) Oerst. y Nothofagus nitida (Phil.) Krasser. Medio Ambiente 7: 9 6. DRINNAN, A. N., AND P. R. CRANE Cretaceous paleobotany and its bearing on the biogeography of austral angiosperms. In T. N. Taylor and E. L. Taylor [eds.], Antarctic paleobiology, Springer-Verlag, New York, NY. FARRIS, S Hennig 86, ver..5. Stony Brook, New York, NY. GANDOLFO, M. A., AND E. J. ROMERO Leaf morphology and a key to species of Nothofagus Bl. Bulletin of the Torrey Botanical Club 9: HANKS, S. L., AND D. E. FAIRBROTHERS Palynotaxonomic investigation of Fagus L. and Nothofagus Bl.: light microscopy, scanning electron microscopy, and computer analysis. In V. H. Heywood [ed.], Botanical systematics : 42. Academic Press, London. HERENDEEN, P. S., P. R. CRANE, AND A. N. DRINNAN Fagaceous flowers, fruits, and cupules from the Campanian (Late Cretaceous) of central Georgia, USA. International Journal of Plant Sciences 56: HILL, R. S Tertiary Nothofagus macrofossils from Cethana, Tasmania. Alcheringa 3: Tertiary Nothofagus (Fagaceae) macrofassils from Tasmania and Antarctica and their bearing on the evolution of the genus. Botanical Journal of the Linnean Society 05: Nothofagus smithtonensis (Nothofagaceae), a new macrofossil species from Oligocene sediments in northwest Tasmania, Australia, and its phylogenetic significance. Review of Paleobotany and Palynology 80: 5 2., AND G. J. JORDAN The evolutionary history of Nothofagus (Nothofagaceae). Australian Systematic Botany 6: 26., AND J. READ. 99. A revised infrageneric classification of Nothofagus (Fagaceae). Botanical Journal of the Linnean Society 05: HJELMQVIST, H Some notes on Nothofagus from New Guinea and New Caledonia. Botaniska Notiser 6: HOOKER, J. D Botany of the antarctic voyage of H. M. Discovery ships Erebus and Terror in the years In Flora Novae-Zelandiae, vol. 2, pt.., introductory essay. Reeve, London. HUMPHRIES, C. J. 98a. Biogeographical methods and the southern beeches (Fagaceae: Nothofagus). In V. A. Funk and D. R. Brooks [eds.], Advances in cladistics: proceedings of the first meeting of the Hennigian Society, New York Botanical Garden, Bronx, NY.. 98b. Biogeographic methods and the southern beeches. In P.L. Forey [ed.], The evolving biosphere, Cambridge University Press, Cambridge., J. M. COX, AND E. S. NEILSEN Nothofagus and its parasites: a cladistic approach to coevolution. In A. R. Stone and D. L. Hawksworth [eds.], Coevolution and systematics, Clarendon Press, Oxford. KHAN, A. M Palynology of Neogene sediments from Papua (New Guinea) stratigraphic boundaries. Pollen et Spores 6: KLUGE, A. G A concern for evidence and the phylogenetic hypothesis of relationship among Epicrates (Boidae, Serpentes). Systematic Zoology 38: KUBITZKI, K Fagaceae. In K. Kubitzki, J. G. Rohwer, and V. Bittrich [eds.], The families and genera of vascular plants, vol. 2, Springer-Verlag, Berlin. LINDER, H. P., AND M. D. CRISP Nothofagus and Pacific biogeography. Cladistics : MACPHAIL, M. K., N. F. ALLEY, E.M.TRUSWELL, AND I. R. K. SLUITER Early Tertiary vegetation: evidence from spores and pollen. In R. S. Hill [ed.], History of the Australian vegetation: Cretaceous to recent, Cambridge University Press, Cambridge. MANOS, P. S., K. C. NIXON, AND J. J. DOYLE Cladistic analysis of restriction site variation within the chloroplast DNA inverted repeat region of selected Hamamelididae. Systematic Botany 8: , AND K. P. STEELE A phylogenetic analysis of higher Hamamelididae based on morphology and sequences from plastid and nuclear DNA. American Journal of Botany 82: 49 (Abstract)., AND. In press. Phylogenetic analyses of higher Hamamelididae based on plastid sequence data. American Journal of Botany MARTIN, P. G., AND J. M. DOWD Using sequences of rbcl to study phylogeny and biogeography of Nothofagus species. Australian Systematic Botany 6: MELVILLE, R Continental drift and plant distribution. In D. H. Tarling and S. K. Runcorn [eds.], Implications of continental drift to the earth sciences, : Academic Press, London. MICKEVICH, M. F Taxonomic congruence. Systematic Zoology 27:43 58., AND J. S. FARROS. 98. The implication of congruence in Menidia. Systematic Zoology 30: MILDENHALL, D. C New Zealand Late Cretaceous and Cenozoic plant biogeography: a contribution. Palaeogeography, Palaeoclimatology, Palaeoecology 3: MOORE, D. M. 97. Connections between cool temperate floras with particular reference to southern South America. In D. H. Valentine [ed.], Taxonomy, phytogeography, and evolution, Academic Press, London. NIXON, K. C The origin of Fagaceae: evolution, systematics, and fossil history of the Hamamelidae. In P. R. Crane and S. Blackmore [eds.], Systematic Association Special Vol. No. 40B, Clarendon Press, Oxford. PATTERSON, C. 98. Methods of paleobiogeography. In G. Nelson and D. E. Rosen [eds.], Vicariance biogeography: a critique, Columbia University Press, New York, NY. PHILIPSON, W. R., AND M. N. PHILIPSON A classification of the genus Nothofagus (Fagaceae). Botanical Journal of the Linnean Society 98: PRAGLOWSKI, J Fagaceae L. (Fagoideae). In S. Nilsson [ed.], World pollen and spore flora II, 28. Almqvist and Wiksell, Stockholm. PREEST, D. S A note on the dispersal characteristics in the seed of the New Zealand podocarps and beeches. In J. L. Grassitt [ed.], Pacific Basin Biogeography, Bishop Museum Press, Honolulu, HI. PREMOLI, A. C. 996a. Leaf architecture of South American Nothofagus (Nothofagaceae) using traditional and new methods in morphometrics. Botanical Journal of the Linnean Society 2:

14 50 AMERICAN JOURNAL OF BOTANY [Vol b. Allozyme polymorphisms, outcrossing rates, and hybridization of South American Nothofagus. Genetica 97: ROZEFELDS, A. C., AND A. N. DRINNAN Ontogeny and floral diversity of staminate flowers in the the southern beeches (Nothofagus). Australian Systematic Botany (Abstract) SOLTIS, D. E., P. S. SOLTIS, D. R. MORGAN, S. M. SWENSON, B. C. MULLIN, J. M. DOWD, AND P. G. MARTIN Chloroplast gene sequence data suggest a single origin of the predisposition for symbiotic nitrogen fixation in angiosperms. Proceedings of the National Academy of Sciences, USA 92: STEVENS, G. R The nature and timing of biotic links between New Zealand and Antarctica in Mesozoic and Early Cretaceous times. In J. A. Crame [ed.], Origins and evolution of the Antarctic biota, The Geological Society, London. STOCK, J., AND P. MOLNAR Uncertainties in the relative positions of the Australia, Antarctica, Lord Howe, and Pacific plates since the Late Cretaceous. Journal of Geophysical Research 87: SWOFFORD, D. L PAUP: phylogenetic analysis using parsimony, version 3.. Illinois Natural History Survey, Champaign, IL. TANAI, T Phytogeographic and phylogenetic history of the genus Nothofagus Bl. (Fagaceae) in the southern hemisphere. Journal of the Faculty of Science, Hokkaido University, series 4, 2: TRUSWELL, E. M., A. P. KERSHAW, AND I. R. SLUITER The Australian-south-east Asian connection: evidence from the paleobotanical record. In T. C. Whitmore [ed.], Biogeographical evolution of the Malay Archipelago, Clarendon Press, Oxford. VAN STEENIS, C. G. G. J Results of the Archbold expeditions: Papuan Nothofagus. Journal of the Arnold Arboretum 34: The land-bridge theory in botany with particular reference to tropical plants. Blumea : Nothofagus, key genus of plant geography, in time and space, living and fossil, ecology and phylogeny. Blumea 9: VENKATESWARLU, K., AND R. NAZAR. 99. A conserved core structure in the 8 25S rrna intergenic region from tobacco, Nicotiana rustica. Plant Molecular Biology 7: WILFORD, G. E., AND P. J. B ROWN Maps of late Mesozoic- Cenozoic Gondwana break-up: some paleogeographical implications. In R. S. Hill [ed.], History of the Australian vegetation: Cretaceous to recent, 5 3. Cambridge University Press, Cambridge. WOJCIECHOWSKI, M. F., M. H. SANDERSON, B. G. BALDWIN, AND M. J. DONOGHUE Monophyly of aneuploid Astragalus (Fabaceae): evidence from nuclear ribosomal DNA internal transcribed spacer sequences. American Journal of Botany 80: YOKOTA, Y., T. KAWATA, Y. IIDA, A. KATO, AND S. TANIFUJI Nucleotide sequences of the 5.8S rrna genes for phylogenetics. Journal of Molecular Evolution 29: APPENDIX 2. Characters and character states for Nothofagus. The characters listed below were extracted from the phylogenetic analysis of Hill and Jordan (993) or are represented in modified form according to personal interpretations, observations, and data presented in Hanks and Fairbrothers (976), Praglowski (982), Dettman et al. (990), and Hill and Read (99). A brief discussion of modified character state codings follows certain character descriptions. Phylogenetic hypotheses based on morphological data were derived from the matrix provided in Appendix 3. All characters were treated as nonadditive in the parsimony analysis. Reproductive Characters. Cupule valve number Cupules with four valves (0), cupules with two symetrical valves (), cupules with two asymmetrical valves (2), cupules absent (3), cupules polymorphic (?). Hill and Jordan (993) used two states: cupules with four valves or two to three asymmetrical valves and cupules with two bilaterally symmetrical valves. The two-valved cupule of N. pumilio is constant, whereas N. solandri is polymorphic for three- and four-valved cupules (Hill and Read, 99). Cupules are absent from N. resinosa. The revised coding accounts for the range of cupule valve morphology and does not combine potentially informative states. 2. Cupule appendage typecupules with scales (0), cupules with glandular lamellae (), cupules with membranous lamellae (2), cupules polymorphic within a species (?). Hill and Jordan (993) used the following transformation series: nonglandular appendages present species polymorphic glandular appendages present. Although the cupules of certain individuals of species of subgenus Nothofagus produce glands, this apparent polymorphism should not be treated as the intermediate state of an ordered character. 3. Cupule vestiturevestiture absent (0), vestiture of simple trichome (). This character is described by Hill and Read (99), but it was not included in Hill and Jordan s (993) analysis. 4. Fruit number per cupulethree or more (0), 2 (), (2). Hill and Jordan (993) attempted to unite this character with carpel number per fruit per cupule in an ordered transformation series. The two characters are independent. Fruit number is stabilized at three for many Nothofagus species, but cupules can be found with more than three fruits (Hill and Read, 99; Herendeen, Crane, and Drinnan, 995). 5. Carpels per flower per cupuledimerous and trimerous (0), trimerous (), dimerous (2). The mixture of dimerous and trimerous fruits within a single cupule is common in the cool-temperate species of Nothofagus. The change to a single trimerous fruit is an autapomorphy for N. pumilio, whereas the change to bicarpellate fruits is a synapomorphy for species of subgenus Brassospora. 6. Staminate perianthperianth tubular with 20 stamens, centripetal development (0), perianth broadly campanulate with 20 stamens, centrifugal development (), perianth campanulate with 20 stamens (2). Hill and Jordan (993) coded the morphology of the staminate perianth as a transformation series separate from the number of stamens per flower. These characters are correlated because developmental studies have shown that the flowers bearing numerous stamens are the result of the fusion of dichasia (Rozefelds and Drinnan, 994). 7. Pollen shapeamb circular (0), amb constricted polygonal (). Hill and Jordan (993) coded pollen outline or amb as a transformation series of three states: amb peritreme amb perigoniotreme amb goniotreme. The distinction between pollen peritreme and perigoniotreme is not consistent in the primary literature. 8. Pollen aperture lengthaperture 5.0 um (0), aperture 4 um (). Hill and Jordan (993) used the ratio of colpus length to equatorial diameter to characterize a three-state transformation series. The qualitative differences between these states are not consistent in the primary literature. Aperture length alone establishes two qualitatively distinct states. 9. Pollen aperture wallthin (0), internally thickened (), externally thickened (2). Hill and Jordan (993) recognized a four-state transformation series based on pollen aperture thickening. The four states are not supported by the primary literature. Instead, three states are coded here to describe the variation present. 0. Shape of aperture ends U-shaped (0), V-shaped (), unknown (?). Although this character was not used by Hill and Jordan (993), the two states appear to consistently define two groups within Nothofagus. Vegetative Characters. Wood anatomytracheids present (0), tracheids absent () 2. Stipule attachmentnot peltate (0), peltate (). 3. Leaf vernationplicate (0), planar (), revolute (2), conduplicate (3). 4. PhyllotaxyDistichous (0), spiral (). 5. Fimbrial veinincomplete or absent (0), complete (). Contrary to Hill and Jordan (993), recent morphometric and taxonomic studies of leaf variation do not support the presence of a complete fimbrial vein in N. nitida (Premoli, 996a) and N. solandri (Gandolfo and Romero, 992). 6. Glandular trichomespresent (0), absent (). 7. Solitary unicellular trichome type APresent (0), absent (). 8. Solitary unicellular trichome type CPresent (0), absent (). 9. Conical unicellular trichomepresent (0), absent (). 20. T-pieces of cutin at stomatal polespresent (0), absent (). 2. Stomatal orientationrandom (0), parallel to long axis of leaf (). 22. Giant stomatapresent (0), absent (). 23. Stomatal sizeeven (0), uneven ().

15 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 5 APPENDIX. The ITS sequences for the 22 Nothofagus and two outgroup taxa examined in this study. The symbol ^ indicates those sites that are phylogenetically informative within Nothofagus. *ITS TCGAAACCTGCCCAGCAGAACGACCCGCGAACTTGTATAAACAACCGGGGGCGGGGGGCG TCGAAACCTGCCCAGCAGAACGACCCGTGAACCTCTTGAAACAACTCGGGGTGCCGTGCG N. antarctica TCGAAACCTGCCTAGCAGAACGACCCGCGAACTAGTTTATAAATTTGGGGGAACGTGGGC N. nitida TCGAAACCTGCCTAGCAGAATGACCCGCGAACTAGTTTATAAATTTGGGGGAACGTGGGC N. pumilio TCGAAACCTGCCTAGCAGAATGACCCGCGAACTAGTTTATAAATTTGGGGGAACGTGGGC N. betuloides TCGAAACCTGCCTAGCAGAATGACCCGCGAACCAGTTTATAAATTTGGGGGAACGTGGGC N. dombeyi TCGAAACCTGCCTAGCAGAATGACCCGCGAACTAGTTTATAAATTTGGGGGAACGTGGGT N. alessandri TCGAAACCTGCCTAGCAGAACGACCCGTGAACTAGTTTCAAAAATTGGGGGCACGTGGCT N. gunnii TCGAAACCTGCCTAGCAGAATGACCCGTGAACTAGTTTCAAAAATTGGGGGCACGTGGCT N. solandri TCGAAACCTGCCTAGCAGAATGACCCGCGAACTAGTTTCAAAATTTGGGGGCACGCGGCT N. fusca TCGAAACCTGCCTAGCAGAATGACCCGCGAACTAGTTTCAAAATTTGGGGGCACGCGGCT N. truncata TCGAAACCTGCCTAGCAGAATGACCCGCGAACTAGTTTCAAAATTTGGGGGCACGCGGCT N. grandis TCGAAACCTGCCTAGCAGAACGACCCGTGAACAAGTTTCAAAATTCGGGGGAACCTGGGC N. brassii TCGAAACCTGCCTAGCAGAACGACCCGTGAACAAGTTTCAAAATTCGGGGGAACCTGGGC N. perryi TCGAAACCTGCCTAGCAGAACGACCCGTGAACAAGTTTCAAAATTTGGGGGAACCTGGGC N. resinosa TCGAAACCTGCCTAGCAGAACGACCCGTGAACAAGTTTCAAAATTTGGGGGAACCTGGGC N. balansae TCGAAACCTGCCTAGCAGAACGACCCGTGAACAAGTTTCAAAATTTGGGGGAACATGGGC N. aequilateralis TCGAAACCTGCCTAGCAGAACGACCCGTGAACAAGTTTCAAAATTTGGGGGAACATGGGC N. alpina TCGAAACCTGCCTAGCAGAACGACCCGCGAACTAGTTTCAAATTCTGGGGGCATGCGGCG N. glauca TCGAAACCTGCCTAGCAGAACGACCCGCGAACTAGTTTCAAATTCTGGGGGCATGCGGCG N. obliqua TCGAAACCTGCCTAGCAGAACGACCCGCGAACTAGTTTCAAATTCTGGGGGCACGCGGCG N. cunninghamii TCGCAACCTGCCTAGCAGAACGACCCGCGAACTGGTTTCAAATTTTGGGGGCATGCGGTG N. menziesii TCGAAACCTGCCTAGCAGAACGACCCGCGAACTTGTTTCAAATTTTGGGGGCACGCGGTG N. moorei TCGAAACCTGCCCAGCAGAACGACCCGCGAACTGGTTTCAAATTTTGGGGGCATGCGGTG ^ ^ ^^ ^^ ^^^^ ^ ^^^ ^^ TTCTCGCCCCGTTCCCCCGAACGGCGGGG AGACACTCGTGCCTTCTTGCC G ATCTCGCCTT TCCCCCGAACGGTAGGG AGACACTTGTTCATCCCTGCC G N. antarctica ATCACGCCTTGTTCACCC-TACGATCGGG AGACA-GTGCTGGCCATTGTCTACCCCGA N. nitida ATCATGCCTTGTTCACCC-TACGATCGGG AGACA-GTGCTGGCCATTGTCTACCCCGA N. pumilio ATCATGCCTTGTTCACCC-TACGATCGGGGGAGACA-GTGCTGGCCATTGTCTACCCCGA N. betuloides ATCATGCCTTGTTCACCC-TACGATCGGGGGAGACA-GTGCTGGCCATTGTCTACCCCGA N. dombeyi ATCATGCCTCGTTCACCC-TACGATCGGG AGACA-GTGCTGGCCATTGTCTACCCCGA N. alessandri ATCGTGCCTTGTTCCCCC-TACGATCGGG AGGCG-GTGCTAGCCATTGTCTACCCCGA N. gunnii ATCATGCCTTGTTCCCCC-TACGATCGGG AGGCG-GTGCCAGCCATTGCCTGCCCCGA N. solandri ATCATGCCTCGTTCCCCC-TACGATCGGG AGGCA-GTGCTAGCCATTGTCTGCCCCGA N. fusca ATCATGCCTCGTTCCCCC-TACGATCGGG AGGCAAGTGCTAGCCATTGTCTGCCCCGA N. truncata ATCATGCCTCGTTCCCCC-TACGATCGGG AGGCAAGTGCTAGCCATTGTCTGCCCCGA N. grandis ATCATGCCTTGTTCCCCC-TACGTTCGGG AGACA-GTGCTAGCCATTGTCTACCCCGA N. brassii ATCATGCCTTGTTCCCCC-TACGTTCGGG AGACA-GTGCTAGCCATTGTCTACCCCGA N. perryi ATCATGCCCTGTTCCCCC-TACGTTCGGG AGACA-GTGCTAGCCATTGTCTACCCCGA N. resinosa ATCATGCCCTGTTCCCCC-TACGTTCGGG AGACA-GTGCTAGCCATTGTCTACCCCGA N. balansae ATCATGCCTTGTTCCCCC-TACGTTCGGG AGACA-GTGCTAGCCATTGTCTATCTCGA N. aequilateralis ATCATGCCTTGTTCCCCC-TACGTTCGGG AGACA-GTGCTAGCCATTGTCTATCTCGA N. alpina ATCGTGCCTCGTGCCCC TACGGTCGGG-CAGACA-G -CACTGTCTACCCCGA N. glauca ATCGCGCCTCGTGCCCC TACGGTCGGG-CAGACA-G -CACTGTCTACCCCGA N. obliqua ATCGTGCCTCGTGCCCC TACGGTCGGG-CAGACA-G -CACTGTCTACCCCGA N. cunninghamii ATCGTGCCTCGTGCCCC TATGGTCGGG-TAGACG-GTGCTGG-CACTGTCTACCCCGA N. menziesii ATCGTGCCTCGTGCCCC TACGGTCGGG-TAGACG-GTGCTAG-CACCATCTACCCCGA N. moorei ATCGTGCCTCGTGCCCC TAYGGTCGGG-TAGACG-GTGCTAG-CACTGTCTACCCCGA ^^ ^^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^^ ^ AACAA CGAACCCCGGCGCGGTCTGCCCAAGGAACTTCAA-CTAAAGAGTGCC-TCCC AACAA CGAACCCCGGCGCGGTCTGCCCAAGGAACTTTAA-CGAAAGAATTCC-TCCC N. antarctica TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. nitida TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. pumilio TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. betuloides TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. dombeyi TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. alessandri TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. gunnii TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAATTATAA-CAAAAGAGCGTCATCTT N. solandri TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. fusca TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT

16 52 AMERICAN JOURNAL OF BOTANY [Vol. 84 APPENDIX. Continued. ^^ ^^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^^ ^ N. truncata TACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCATCTT N. grandis AACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAG-GCGTCATCTT N. brassii AACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAG-GCGTCATCTT N. perryi AACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAG-GCGTCATCTT N. resinosa AACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAG-GCGTCATCTT N. balansae AACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGGGCGTCATCTT N. aequilateralis AACAACAACGAACCCCGGCGCGGGCTGCGCAAGGAACTATAA-CAAAAGGGCGTCATCTT N. alpina CACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAAGCAAAAGAGCGTCGTCTT N. glauca CACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAAGCAAAAGAGCGTCGTCTT N. obliqua CACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAAGCAAAAGAGCGTCGTCTT N. cunninghamii CACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCGTCTT N. menziesii CACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCATCGTCTT N. moorei CACAACAACGAACCCCGGCGCGGACTGCGCAAGGAACTATAA-CAAAAGAGCGTCGTCTT ^ ^ ^ * GTCGCCTCGGAAACGGCGTGCGTGCGGGGGG-GA-ATCTTGTGCAAAACCATAACGACTC GTCGCCTCGGAAACGGTGTGT-TGCGGA-AATGA-ATCTTGTCTAGAACCATAACGACTC N. antarctica GTCGTCCCGGTAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. nitida GTCGTCCCGGTAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. pumilio GTCGTCCCGGTAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. betuloides GTCGTCCCGGTAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. dombeyi GTCGTCCCGGTAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCAATAAACAAAACGACTC N. alessandri GTCGTCCCGGTAACGGTGTGCGCGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. gunnii GTCGTCCCGGTAACGGTGTGCGCGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. solandri GTCGTCCCGGTAACGGTGTGCGCGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. fusca GTCGTCCCGGTAACGGTGTGCGCGCAAGTG-CGATGTCTTGTCTATAAACAAAACGACTC N. truncata GTCGTCCCGGTAACGGTGTGCGCGCAAGCG-CGATGTCTTGTCTATAAACAAAACGACTC N. grandis GTCGCCCCGGTAACGGTGTGCGTGCAAGTGGCGATGTCTTGTATATTAACAAAACGACTC N. brassii GTCGCCCCGGTAACGGTGTGCGTGCAAGTGGCGATGTCTTGTATATTAACAAAACGACTC N. perryi GTCGCCCCGGTAACGGTGTGCGTGCAAGTGGCGATGTCTTGTATATTAACAAAACGACTC N. resinosa GTCGCCCCGGTAACGGTGTGCGTGCAAGTGGCGATGTCTTGTATATTAACAAAACGACTC N. balansae GTCGCCCCGGTAACGGTGTGCGTGCAAGTG-CAATGTCTTGTATATTAACAAAACGACTC N. aequilateralis GTCGCCCCGGTAACGGTGTGCGTGCAAGTG-CAATGTCTTGTATATTAACAAAACGACTC N. alpina GTCGTCCCGGGAACGGTGTGCGTGCAAGCG-CGATGTCTTGTCTAGAAACAAAACGACTC N. glauca GTCGTCCCGGGAACGGTGTGCGTGCAAGCG-CGATGTCTTGTCTAGAAACAAAACGACTC N. obliqua GTCGTCCCGGGAACGGTGTGCGTGCAAGCG-CGATGTCTTGTCTAGAAACAAAACGACTC N. cunninghamii GTCGTCCCGGGAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCTAGAAACAAAACGACTC N. menziesii GTCGTCCCGGGAACGGTGTGCGCGCAAGTG-CGATGTCTTGTCTAGAAACAAAACGACTC N. moorei GTCGTCCCGGGAACGGTGTGCGTGCAAGTG-CGATGTCTTGTCTAGAAACAAAACGACTC ^ ^ ^ ^ ^ ^ ^^ TCGGCAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGCGAATCATCGAG TCGGCAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGCGAATCATCGAG N. antarctica TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. nitida TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. pumilio TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. betuloides TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. dombeyi TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. alessandri TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. gunnii TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. solandri TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. fusca TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. truncata TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. grandis TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. brassii TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. perryi TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. resinosa TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. balansae TCGACAACGGATATCTCGGCTCTCGCATCATCGAATTGCAGAATCCCGTGAATCATCGAG N. aequilateralis TCGACAACGGATATCTCGGCTCTCGCATCATCGAATTGCAGAATCCCGTGAATCATCGAG N. alpina TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. glauca TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. obliqua TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. cunninghamii TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG

17 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 53 APPENDIX. Continued. ^ ^ ^ ^ ^ ^ ^^ N. menziesii TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG N. moorei TCGACAACGGATATCTCGGCTCTCGCATCGATGAATTGCAGAATCCCGTGAATCATCGAG ^^^ TCTTTGAACGCAAGTTGCGCCCGAAGCCATCTGGTCGAGGGCACGTCTGCCTGGGTGTCA TCTTTGAACGCAAGTTGCGCCCGAAGCCACCTGGCCGAGGGCACGTCTGCCTGGGTGTCA N. antarctica TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. nitida TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. pumilio TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. betuloides TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. dombeyi TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. alessandri TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. gunnii TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. solandri TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. fusca TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. truncata TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. grandis TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. brassii TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. perryi TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. resinosa TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. balansae TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. aequilateralis TCTTTGAACGCAAGTTGCGCCCGAAGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. alpina TCTTTGAACGCAAGTTGCGCCCGATGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. glauca TCTTTGAACGCAAGTTGCGCCCGATGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. obliqua TCTTTGAACGCAAGTTGCGCCCGATGCCTTTTGGTTGAGGGCACGTCTGCCTGGGCGTCA N. cunninghamii TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGCTGAGG ACGTCTGCCTGGGCGTCA N. menziesii TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGCTGAGG ACGTCTGCCTGGGCGTCA N. moorei TCTTTGAACGCAAGTTGCGCCCGAAGCCTTCTGGCTGAGG ACGTCTGCCTGGGCGTCA ^ ^ ^ *ITS CGCATCGTTGCCCCCAACCCCATCGCCTCTCCAAGAGACGAGGGCGGTCTGC-GGGGCGG CGCATCGTTGCCCCCAACCCCATCTCCTTGTAAAGGGACGAGGGGGCCTGT GGGGCAG N. antarctica CGCATCGTTGCCCCCAAAACCACAACCCC-AAAAGGGATAAAGGTGGTGTTGCGGGGCGG N. nitida CGCATCGTTGCCCCCAAAACCATAACCCC-AAAAGGGATAAAGGTGGTGTTGCGGGGCGG N. pumilio CGCATCGTTGCCCCCAAAACCATAACCCC-GAAAGGGATAAAGGTGGTGTTGCGGGGCGG N. betuloides CGCATCGTTGCCCCCAAAACCATAACCCC-AAAAGGGATAAAGGTGGTGTTGCGGGGCGG N. dombeyi CGCATCGTTGCCCCCAAAACCATAACCCC-GAAAGGGATAAAGGTGGTGTTGCGGGGCGG N. alessandri CGCATCGTTGCCCCCAAAACCATAACCC -TGGTGTTGCGGGGCGG N. gunnii CGCATCGTTGCCCCCAAAACCATAACCC -TGGTGTTGCGGGGCGG N. solandri CGCATCGTTGCCCCCAAAACCACAACCC -TGGTGTTGCGGGGCGG N. fusca CGCATCGTTGCCCCCAAAACCATAACCC -TGGTGTTGCGGGGCGG N. truncata CGCATCGTTGCCCCCAAAACCATAACCC -TGGTGTTGCGGGGCGG N. grandis CGCATCGTTGCCCCCAAAACCACAACCCCCAAGAGGGAT-AAGACGGTGTTGCGGGGCGA N. brassii CGCATCGTTGCCCCCAAAACCACAACCCCCAAGAGGGAT-AAGACGGTGTTGCGGGGCGA N. perryi CGCATCGTTGCCCCCAAAACCACAACCCCCAAGAGGGAT-AAGACGGTGTTGCGGGGCGA N. resinosa CGCATCGTTGCCCCCAAAACCACAACCCCCAAGAGGGAT-AAGACGGTGTTGCGGGGCGA N. balansae CGCATCGTTGCCCCCAAAACCACAACCCCCAAGAGGGAT-AAGATGGTGTTGCGGGGCGA N. aequilateralis CGCATCGTTGCCCCCAAAACCACAACCCCCAAGAGGGAT-AAGATGGTGTTGCGGGGCGA N. alpina CGCATCGTTGCCCCAAAAAC AC-AAGGTGGTGTTGCGGGGCGG N. glauca CGCATCGTTGCCCCAAAAAC AC-AAGGTGGTGTTGCGGGGCGG N. obliqua CGCATCGTTGCCCCAAAAAC AC-AAGGTGGTGTTGCGGGGCGG N. cunninghamii CGCATCGTTGCCCCAAAAACCACAACACCCGAGAGG-AT-AAGGCGGTGTTGCGGGGCGG N. menziesii CGCATCGTTGCCCCAAAAACCACAACACCCGAGAGG-AC-AAGGTGGTGTTGCGGGGCGG N. moorei CGCATCGTTGCCCCAAAAACCACAACACCCGAGAGG-AT-AAGGTGGTGTTGCGGGGCGG ^ ^ ^ ^ ^ ^ ^^ ^ ACATTGGCCTCCCGTGA-CTCTCGCTCGCGGCTGGCCTAAAAGCGAGTCCTCGGCGACGA AAATTGGCCTCCCGTGAGCTCATGCATGCGGTTGGCCTAAAAGCGAGTCCTCGGCGACGC N. antarctica ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCAACAA N. nitida ATGTTGGTCTCCCGTGTGCTGATGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAG N. pumilio ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. betuloides ATGTTGGCCTCCCGTGTGCTGATGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAG

18 54 AMERICAN JOURNAL OF BOTANY [Vol. 84 APPENDIX. Continued. ^ ^ ^ ^ ^ ^ ^^ ^ N. dombeyi ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. alessandri ATGATGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGATAA N. gunnii ATGATGGTCTCCCGTGTGCTATTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. solandri ATGATGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. fusca ATGATGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. truncata ATGATGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. grandis ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTTCTTGGCGACAA N. brassii ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTTCTTGGCGACAA N. perryi ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTTCTTGGCGACAA N. resinosa ATGTTGGTCTCCCGTGTGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTTCTTGGCGACAA N. balansae ATGTTGGTCTCCCGTGT-CTGTTGCTCGCGGTTGGCCAAAAAGCGAGTTCTTGGCGACAA N. aequilateralis ATGTTGGTCTCCCGTGT-CTGTTGCTCGCGGTTGGCCAAAAAGCGAGTTCTTGGCGACAA N. alpina ATGTTGGCCTCCCGTGCGCTGTTGCTCGCGGTTGGCCAAAAAGCGAGTCCTCGGCGACAA N. glauca ATGTTGGCCTCCCGTGCGCTATTGCTAGCGGTTGGCCAAAAAGCGAGTCCTCGGCGACAA N. obliqua ATGTTGGCCTCCCGTGCGCTATTGCTCGCGGTTGGCCAAAAAGCGAGTCCTCGGCGACAA N. cunninghamii ATGTTGGCCTCCCGTGCGCAATCGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA N. menziesii ATGTTGGCCTCCCGTGCGCAATCGCTCGCGGTYGGCCAAAAAGCGAGTCCTCGGAGACAA N. moorei ACGTTGGCCTCCCGTGCGCAATCGCTCGCGGTTGGCCAAAAAGCGAGTCCTTGGCGACAA ^ ^ ^ ^^^^ ^ ^ ^ GCGCCACGACAATCGGTGGT-TGACAAACCCTCGTGTCCC-GTCGTGCGCGGCT-CGTCG GCGCCACGACAATCGGTGGT-TGTCAAACCCTCGTGTCCC-GTCGTGCGTGACTGCGTCG N. antarctica GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTCGTGCGTCCGCT-GTTG N. nitida TTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGCGTCCGCT-GTTG N. pumilio GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGCGTACACT-GTTG N. betuloides TTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTCGTGCGTCCGCT-GTTG N. dombeyi GTGCCACGACAATCTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGCGTACACT-GTTG N. alessandri GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGTGTTCTTT-GTTG N. gunnii GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGTGTTCTTT-GTTG N. solandri GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGTGTTCTTT-GTTG N. fusca GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGTGTTCTTC-GTTG N. truncata GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTTCCCTTGTTGTGTGTTCTTT-GTTG N. grandis GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTCCCCTTGTTGTGTGTTCTTT-GTTG N. brassii GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTCCCCTTGTTGTGTGTTCTTT-GTTG N. perryi GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTCCCCTTGTTGTGTGTTCTTT-GTTG N. resinosa GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTCCCCTTGTTGTGTGTTCTTT-GTTG N. balansae GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTCCCCTTGTTGTGTGTTCTTT-GTTG N. aequilateralis GTGCCACGACAATTTGTGGTATGAGAAACCCTCGTCCCCTTGTTGTGTGTTCTTT-GTTG N. alpina GTGCCACGACAATCTGCGGTATGAGAAACCCTCGTTCCCTTGTCGTGCGTTCTTT-GT-G N. glauca GTGCCACGACAATCTGCGGTATGAGAAACCCTCGTTCCCTTGTCGTGCGTTCTTT-GT-G N. obliqua GTGCCACGACAATCTGCGGTATGAGAAACCCTCGTTCCCTTGTCGTGCGTTCTTT-GT-G N. cunninghamii GTGCCACGACAATCTGTGGCATGAGAAACCCTCGTTCCCTTGTCGTGCGTTCTTT-GTCG N. menziesii GTGCCACGACAATCTGTGGCATGAGAAACCCTCGTTCCCTTCTCGTGCGTTCTTCTGTCG N. moorei GTGCCGCGACAATCTGTGGCATGAGAAACCCTCGTTCCCTTGTCGTGCGTTCTTCTGTCG ^ ^ ^ ^ ^ ^ ^ ^ ^^^ ^ CTCGTCTTGTGCTCTGTGACCCTGTAGCGTCGCGCTAGCGAGGCTCTT CTCATCGTGTGCTCCTTGACCCTGCTGTGTCGCGCTAGCGAGGCTTCC N. antarctica CCCGTAGAGCTCTCTGCGACCCTTTTGCATTGC GATGCTTTC N. nitida CCTGTAGAGCTCTCTGCGACCCTTTTGCATTGC GATGCTTTC N. pumilio CCCGTAGAGCTCTCTGTGACCCTTTTGCATTGC-ATTGCGATGCTTTC N. betuloides GCCGTAGAGCTCTCTGCGACCCTTTTGGCTTGC GATGCTTTC N. dombeyi CCCGTAGAGCTCTCTGTGACCCTTTTGCATTGC-ATTGCGATGCTTTC N. alessandri GCCGTAGAGCTCTGTGCGACCCTCTTGCATTGC GATGCTTTC N. gunnii CCCGTAGAGCTCTCTGTGACCCTCTTGCATTGC GATGCTTTC N. solandri CCCGTAGAGCTCTGTGTGACCCTCTTGCATTGC GATGCTTTC N. fusca CCCGTAGAGCTCTGTGTGACCCTCTTGCATTGC GATGCTTTC N. truncata CCCGTAGAGCTCTGTGTGACCCTCTTGCATTGC GATGCTTTC N. grandis CCTCTTGAGCTCTATGTGACCCTTTTGCATTGC GATGCTTTC N. brassii CCTCTTGAGCTCTATGTGACCCTTTTGCATTGC GATGCTTTC N. perryi CCTCTTGAGCTCTATGTGACCCTTTTGCATTGC GATGCTTTC N. resinosa CCTCTTGAGCTCTATGTGACCCTTTTGCATTGC GATGCTTTC N. balansae CCTCTTGAGCTCTATGTGACCCTTTTGCATTGC GATGCTTTC N. aequilateralis CCTCTTGAGCTCTATGTGACCCTTTTGCATTGC GATGCTTTC N. alpina CC-GCGGGGCTCTCCGTGACCCTCTTGC GACTGC-GCGATGCTTTC

19 August 997] MANOSSYSTEMATICS OF NOTHOFAGUS 55 APPENDIX. Continued. ^ ^ ^ ^ ^ ^ ^ ^ ^^^ ^ N. glauca CC-GCGGGGCTCTCCGTGACCCTCTTA -GCTGC-GCGATGCTTTC N. obliqua CC-GTGGGGCTCTCCGTGACCCTCTTGC-TCGCTGCTGCGATGCTTTC N. cunninghamii CCCGTTGGGCTCTCCGTGACCCTCTTGCATTGCTGTTGCGATGCTTTC N. menziesii CCCGTTGGGCTCTCCGTGACCCTCTTGCATTGCTGTTGCGATGCTTTC N. moorei ACCGTTGGGCTCTCCGTGACCCTCTTGCATTGCTGTTGCGATGCTTTC ^ ^^^^ ^ ^^ ^ ^ ^^ APPENDIX 3. Morphological data matrix for 23 species of Nothofagus and the outgroup Fagus. Taxon Characters Fagus ? N. antarctica N. nitida N. pumilio 2? N. betuloides 0? N. dombeyi N. alessandri N. gunnii N. solandri? N. fusca N. truncata N. grandis N. brassii N. perryi N. resinosa 3? N. balansae N. aequilateralis N. alpina N. glauca N. obliqua N. cunninghamii N. menziesii N. moorei