Many taxonomic distinctions are made or refined on the basis of herbarium material that is either dried or preserved in spirit medium. Hitherto, examination of internal structure has only been possible by the destructive sectioning of the preserved material. In this paper, the use of nuclear magnetic resonance (NMR) imaging for the non-destructive, non-invasive, complete three-dimensional structural examination of herbarium material is demonstrated for the first time. The experimental materials were the fruiting structures of two species of Southern Hemisphere Podocarpaceae: Acmopyle pancheri and Podocarpus nivalis. Material dried in accordance with standard herbarium techniques was used, as well as material preserved in spirit and freshly gathered fruits. The dried material was subsequently rehydrated using standard techniques, and protocols established for the specimens. Appropriate selection of NMR imaging parameters allowed a variety of anatomical features to be highlighted on a single specimen. Fresh specimens from living material gave the best NMR signals. Dry specimens gave no signal except from the lipid in the seed, but when rehydrated the images yielded almost as much information about internal structure as did a fresh specimen of the same taxon. Thus, NMR imaging has great potential value as a non-invasive method for obtaining details of the internal structure of fruits and seeds and is particularly useful when, as in the case of Acmopyle, the sclerotesta of the seed is too lignified for sectioning by conventional methods.

Recent trends in compilation of world consensus family classifications from existing floristic and monographic data, and demands for alpha-taxonomic and other traditional phenetic data for analysis with phylogenetic reconstructions derived from DNA sequences are discussed. Obstacles hindering the production of a meaningful, comprehensive data set for Apiaceae include: (1) the lack of comparable non-molecular phenetic data; (2) incomplete coverage of family accounts in recent Floras, particularly in the southern hemisphere; (3) large, artificial genera awaiting monographic treatment; and (4) the lack of database systems that handle differences in taxonomic opinion (alternative classifications). The use of electronic communication, particularly the Internet, can help to accelerate progress in these areas through promoting collaboration and information exchange. The contribution of the Apiales Resource Centre website (especially the umbellifer areas: www.umbellifers.com) is highlighted.

The family Apiaceae (Umbelliferae) can be credited with two major landmarks in botanical history: the first systematic monographic treatment of any plant group (Morison, 1672), and the first international symposium dedicated to systematic research on a plant family (Heywood, 1971). The 1970 symposium on the Biology and Chemistry of the Umbelliferae held at the University of Reading, UK, resulted from the large body of research interest in the family around the world at that time, and helped to stimulate further work on the Apiaceae. It also provided a model for similar symposia on major plant groups in the years to follow, including Asteraceae (Heywood et al., 1977), Brassicaceae (Vaughan et al., 1976), Lamiaceae (Harley & Reynolds, 1992), Solanaceae (Hawkes et al., 1979), and Fabaceae (Summerfield & Bunting, 1980; Polhill & Raven, 1981). Growing interest in umbellifers soon resulted in a second international symposium on the family held at the Centre Universitaire de Perpignan, France, in 1977 (Cauwet-Marc & Carbonnier, 1982). Although a large role of this second symposium was to review progress on a major co-operative research programme focused mainly on the tribe Caucalideae, participants with other interests were also involved, and wider developments in the systematics of the family were discussed.

The taxonomy and nomenclature of the Rubus species (sub)endemic to Madeira and the Azores are treated. Only three species names have to be retained: R. bollei Focke (synonyms: R. canariensis Focke, R. concolor Lowe, R. suspiciosus Menezes, R. vahlii Frid.), R. hochstetterorum Seub., and R. serrae Soldano (R. grandifolius Lowe). The last two species are described and illustrated. Three names are lectotypified. A key to Rubus of the Macaronesian region (five native species, one hybrid, one neophyte) is provided.

Three rare alpine, short-stemmed or stemless species of Saussurea DC. subgen. Saussurea are treated: S. ovata Bentham (W Kunlun Shan), S. elliptica C. B. Clarke (S. kuschakewiczii C. Winkler, syn. nov.; Tian Shan, Pamir, Karakorum, Zanskar) and S. atkinsonii C. B. Clarke (W Himalaya). The species are not closely interrelated and morphologically rather isolated. They are confined to different chains or mountain sections of Western High Asia, with their possible close relatives found in geographically far distant areas. Putative relationships of the taxa are discussed, revealing conflicts with the infrageneric classification proposed by Lipschitz (1979). Saussurea candolleana (DC.) Schultz Bip. (S. clarkei Hook. f., syn. nov.) is briefly discussed.

This study is a contribution to the further understanding of cytological patterns in Aeschynanthus(Gesneriacaeae). Chromosome numbers are reported for 12 species from six sections; nine of these are new counts. Two basic numbers, x=16 and x=15, are generally encountered. Aeschynanthus gracilis proved to be of exceptional interest, as its rare somatic number, 2n=28, confirms the occurrence of a third basic number, x=14, in the genus. Variation in chromosome number in relation to seed morphology is examined.

Phylogenetic relationships involving the angiosperm order Apiales (Apiaceae and Araliaceae) are troublesome at nearly every taxonomic level and have eluded several generations of botanists. Because of difficulties in interpreting and polarizing morphological character states at deeper phylogenetic levels, most studies in Apiales have focused on relationships between the two families and among/within the apialean genera. In the present study, however, recent contributions from molecular analyses are reviewed and combined using a ‘supertree’ approach to test traditional hypotheses of relationships involving Apiales, and to re-evaluate assumptions of character-state evolution in the order. Results from this study confirm that Apiales form a monophyletic group with Pittosporaceae (along with Griselinia G. Forst., Melanophylla Baker, Torricellia DC. and Aralidium Miq.), and should be transferred out of subclass Rosidae (away from both Cornales and Sapindales) to the Asteridae (in a position close to Asterales and Dipsacales). These findings are also supported by several lines of morphological, anatomical, and phytochemical evidence, and provide a more satisfactory framework for interpreting relationships and character-state evolution within the major clades of Apiales.

A study of the fruit structure of Annesorhiza filicaulis Eckl. & Zeyh. has shown that only the four commissural ribs are expanded, resulting in a fruit wing configuration characteristic of the genus Peucedanum. Hysteranthous or seasonal leaves, as present in A. filicaulis, are not unique to Annesorhiza but are also encountered in Chamarea Eckl. & Zeyh. and in Peucedanum sulcatum Sond. and related species. The correct generic delimitation of Peucedanum (one or several genera) is one of the last challenges in the alpha taxonomy of southern African Apiaceae, and A. filicaulis complicates the matter even further. Nevertheless, A. filicaulis is clearly a misfit in Annesorhiza and should be included in future studies of Peucedanum sensu lato, even though the latter may prove not to be monophyletic. Annesorhiza filicaulis is formally transferred to Peucedanum, and the correct nomenclature, typification and known geographical distribution of this poorly known species is presented.

Wood anatomical data confirm the close relationships of most Araliaceae to Apiaceae, but do not indicate any intermediate groups between the two families. Heteromorpha Cham. & Schltdl., Bupleurum L. and Melanoselinum Hoffm. form a well-delimited group distinguished from other woody Apiaceae by helical thickenings on their vessel walls, septate fibres, and mostly homogeneous rays. The woodiness in Nirarathamnos Balf.f. and Myrrhidendron J. M. Coult. & Rose is likely to be of secondary origin.

A review of the taxonomy and distribution of the European species of Damasonium Mill. has been carried out to provide background information for conservation. Damasonium alisma Mill., D. bourgaei Coss. and D. polyspermum Coss. are recognized as distinct species, and descriptions and a key are presented. Lectotypes are designated for D. bourgaei and D. polyspermum. Damasonium alisma is the most northerly taxon, predominantly occurring in France and England, but extending eastwards to Russia. Damasonium bourgaei is widespread around the Mediterranean eastwards at least to India. Damasonium polyspermum occurs mainly in the western Mediterranean.

Examination of morphology, ploidy and interfertility in the two subspecies of the Old World Senecio flavus (Decne.) Sch. Bip. (Compositae) and the closely related New World S. mohavensis A. Gray does not support the subspecific taxonomy of S. flavus. On the basis of our results S. flavus subsp. breviflorus Kadereit is transferred to S. mohavensis as a new subspecies: S. mohavensis subsp. breviflorus (Kadereit) M. Coleman comb. nov. The new subspecies has a distribution that includes Arabia, the Middle East, Sinai, Iran, Afghanistan, Djibouti, and the Thar Desert of Pakistan. The type subspecies of S. mohavensis occurs in the Mojave and Sonoran deserts of North America, providing an unusual disjunct distribution at the species level. Separation from S. flavus is based upon differences in morphology and chromosome number. Senecio flavus is diploid (2n = 20), while both subspecies of S. mohavensis are tetraploid (2n = 40). Further support for the new taxonomic treatment is provided by the results of controlled crosses. No artificial hybrids have been generated from crosses made between the previously recognized subspecies of S. flavus, while crosses between the newly recognized subspecies of S. mohavensis have produced fertile hybrids. The fertility of the hybrids is significantly lower than the parental taxa (P<0.001), indicating partial genetic divergence since isolation. Previous studies of isozyme and cpDNA variation in all three taxa also support the new treatment. The similarity of the S. mohavensis subspecies suggests a relatively recent separation, although the amount of genetic divergence does not support a post-Colombian introduction. Given that land bridges to North America via Beringia and the North Atlantic last existed in the Oligocene, long-distance dispersal seems the most likely explanation. Natural dispersal to rather than from the New World is supported, but whether this took place in an easterly or westerly direction is unclear. The evolution of S. mohavensis remains equivocal.

The stem and leaf anatomy of the ten species of Jasione L. (Campanulaceae) in the Iberian Peninsula were investigated; their infra-specific taxa were also studied. The species differ from each other anatomically and can be identified by their anatomical characters. The anatomical evidence supports the taxonomic treatment that will be published in the forthcoming Flora iberica Volume 14. The possible relations between the anatomy and the ecology of these plants are discussed. Specialized small multicellular structures (trichoids) present on the leaf surface, whose interest has not previously been recognized, are described and their possible function discussed.

Recent molecular studies indicate that the araliaceous tribes Myodocarpeae R. Vig. (Delarbrea Vieill., Pseudosciadium Baill. and Myodocarpus Brongn. & Gris.) and Mackinlayeae R. Vig. (Apiopetalum Baill., Mackinlaya F. Muell. and several genera of Hydrocotyloideae Link (Apiaceae)) comprise basally branching lineages within Apiales, an interpretation consistent with data from morphology and wood anatomy. Comparison of selected features in these genera, and in close relatives of Apiales, suggests that ancestral character states for the order may include: simple leaves, inflorescences in panicles of umbellules, flowers with articulated pedicels and a bicarpellate gynoecium, an andromonoecious, duodichogamous sexual system, septate fibres, the absence of radial canals, and the presence of paratracheal axial parenchyma in the wood.

Problems of infrageneric classification in Pedicularis L. (Scrophulariaceae), and the phytogeography of the genus in the Flora of Bhutan area, are discussed. Six new species and one new variety of Pedicularis L. are described. Pedicularis garckeana Prain ex Maxim. is separated from P. ser. Robustae Prain as the new, monospecific series Garckeanae R. R. Mill (Sikkim and adjacent Xizang). Six of these new taxa are endemic to Bhutan (P. dhurensis R. R. Mill sp. nov., P. longipedicellata P. C. Tsoong var. lanocalyx R. R. Mill var. nov., P. melalimne R. R. Mill sp. nov., P. microloba R. R. Mill sp. nov., P. sanguilimbata R. R. Mill sp. nov. and P. woodii R. R. Mill sp. nov.) and one (P. yarilaica R. R. Mill sp. nov.) occurs in N Bhutan and S Xizang. The type material of Phtheirospermum auratum Bonati (=Pedicularis aurata (Bonati) H. L. Li) is shown to be a mixed gathering; the name is lectotypified and an emended description of P. aurata given. Several other names are lectotypified. Three sectional names originally proposed by Yamazaki (each ‘based’ on invalid ‘names’ of groups earlier recognized as subgreges by Tsoong) are validated as P. sect. Asthenocaulus [P. C. Tsoong ex] R. R. Mill, P. sect. Nothosigmantha T. Yamaz. ex R. R. Mill, and P. sect. Rhizophyllastrum T. Yamaz. ex R. R. Mill, while a fourth invalid section (‘P. sect. Brachystachys (P. C. Tsoong) T. Yamaz.’) is here recognized and described as P. sect. Brevispica R. R. Mill, sect. nov. Pedicularis sect. Cryptorhynchus T. Yamaz. and P. sect. Elephanticeps T. Yamaz. are provided with emended descriptions. The nomenclature of P. pennelliana P. C. Tsoong and related taxa of the P. gracilis Wall. ex Benth. complex is discussed.

The terrestrial orchid genus Bhutanthera Renz (subfamily Orchidoideae, tribe Orchideae), comprising five species, three newly described here, is established. A new species, Herminium pygmaeum Renz, and three new varieties, Bulbophyllum cauliflorum Hook.f. var. sikkimense N. Pearce & P. J. Cribb, Cephalanthera erecta (Thunb.) Bl. var. oblanceolata N. Pearce & P. J. Cribb, and Gymnadenia orchidis Lindl. var. pantlingii Renz, are here newly described. Five new combinations and one new name are proposed. Four new sections of Bulbophyllum are also established and their relationships discussed.

A preliminary analysis of New Caledonian Araliaceae has shown varying degrees of support for previous hypotheses on sexual system evolution in the family. Andromonoecy, rather than hermaphroditism, is strongly supported as the ancestral sexual system for the Araliaceae. Derivation of hermaphroditism from andromonoecy in Delarbrea harmsii R. Vig. and Polyscias bracteata (R. Vig.) Lowry, ined. is also strongly supported. Support for the derivation of dioecy from andromonoecy in several New Caledonian species of Polyscias J. R. Forst. & G. Forst. depends in part on an assumption that andromonoecy cannot be secondarily derived from dioecy. The derivation of hermaphroditism from andromonoecy in Schefflera candelabra Baill. and S. pseudocandelabra R. Vig. is equivocal. The hypothesis that the immediate ancestors of Arthrophyllum Blume were andromonoecious is not supported.

A new species of Agapetes (Ericaceae) from Thailand is described and illustrated.

Systematic relationships within the Hawaiian group Tetraplasandra A. Gray, Munroidendron Sherff and Reynoldsia sandwicensis A. Gray are investigated here based on parsimony analysis of nucleotide sequences from the internal transcribed spacer (ITS 1 and 2) regions and the non-transcribed spacer of 5S (5S-NTS) nuclear ribosomal DNA. Results indicate that Tetraplasandra, Munroidendron and Reynoldsia sandwicensis form a single monophyletic group, and that Reynoldsia A. Gray is paraphyletic. Both data sets support combining Reynoldsia sandwicensis with Munroidendron and treating it as distinct from Reynoldsia species outside Hawaii. Both data sets support two major clades: a Munroidendron/Reynoldsia sandwicensis clade and a Tetraplasandra clade, but they are discordant in relation to the monophyly of Tetraplasandra. 5S-NTS results support treating Tetraplasandra, Munroidendron and Reynoldsia sandwicensis as a single genus, whereas ITS results suggest that Tetraplasandra represents a distinct genus from Reynoldsia sandwicensis and Munroidendron. Tetraplasandra gymnocarpa (Hillebr.) Sherff is the only member of the Araliaceae with a completely superior ovary, and in T. kavaiensis (H. Mann) Sherff the ovary is partially superior. Together, they form a well-supported clade within the Hawaiian complex, confirming the hypothesis that in Tetraplasandra superior ovaries evolved in Hawaii from an epigynous ancestor.

The occurrence of a monodominant tropical forest dominated by the tree species Brosimum rubescens Taub. (Moraceae) in the transition zone between the cerrado region and the Amazonian rain forests along the Araguaia valley in the state of Mato Grosso in Brazil is reported. A 6000m2 (40×150m) area located in the central portion of a forest (14°50′47″S, 52°08′37″W) on the Eldorado Farm in Nova Xavantina was sampled to determine the structure, phytosociology and soil properties. A total of 44 tree species was found in the forest, with a Shannon diversity index (H′) of 2.37 and the Pielou evenness index (J′) of 0.63. These values were similar to those reported by other workers in an Amazonian monodominant forest dominated by Peltogyne gracilipes. Average tree density equal or greater than (≥) 5cm dbh was 1066ha−1 and basal area 37.46m2 ha−1. Dead standing individuals represented 7.03% of the total density and lianas 8.12%. The main families were Moraceae, Rubiaceae, Burseraceae, Hippocrateaceae and Leguminosae–Caesalpinioideae. The species with the highest importance value indices were B. rubescens (124.55), Amaioua intermedia var. brasiliana (32.97), Cheiloclinium cognatum (22.41), Tetragastris balsamifera (19.58) and Protium pilosissimum (16.64). Brosimum rubescens alone accounted for 85.5% of the individuals ≥30cm dbh, 80.35% of those ≥20cm dbh, 57.33% of those ≥10cm dbh, and 35.85% of those ≥5cm dbh. The forest was composed of upper, middle and lower storeys. Diameter distribution was unbalanced for the community and for the main species, suggesting that establishment has been episodic. The surface soil was acid, with high exchangeable Al levels and low fertility, as is the case for most cerrados and Amazonian forests. The availability of K was in the range of adequacy, as compared with most native soils. However, concentrations of available Mg in the soil was higher than that of available Ca, with Mg/Ca ratios above three, unlike most cerrados and Amazonian forests, but similar to the soils of many monodominant tropical forests reported in the literature. Soils were characterized by high concentrations of available Fe. Besides the high Mg/Ca ratios, a high proportion of K as compared with the bivalent cations Ca and Mg, and toxicities of Fe and Mn due to strong acidity could be factors influencing the occurrence of species in this forest.