genus Fascicularia Mez

Literature references:

Smith & Downs 1979: 1711

Comments:

• Revision of the genus Fascicularia Mez (Bromeliaceae)
  GEORG ZIZKA, RALF HORRES, E. CHARLES NELSON AND KURT WEISING Botanical Journal of the Linnean Society (1999), 129: 315-332. With 4 figures
  .
  
  INTRODUCTION
  
  Vegetatively, Fascicularia is similar to Ochagavia Phil. Both genera are endemic to central and southern Chile and comprise a group of terrestrial or epiphytic bromeliad species with more or less succulent leaves.
  Following Ruiz & Pavon, Molina (1810) and Gay (1853) recognized Bromelia bicolor as the only species of this group. A few years later Philippi (1856) described the genus Ochagavia, at first including only one species endemic to Isla Robinson Crusoe, Juan Fernandez Islands (Ochagavia elegans Phil.). Shortly afterwards, Philippi (1857-1858) defined another genus, Rhodostachys, with a single species, R. andina Phil. Delimitation of the genera was not clear and this led to a number of new combinations by Bentham (in Bentham & Hooker, 1883) and Baker (1889). The taxonomy and nomenclature of this group of genera were further complicated by new names and combinations published in the late 19th century based on cultivated material from European gardens without preservation of type specimens.
  In a footnote to his treatment of the Bromeliaceae for Martius' Flora Brasiliensis, Mez (1894) created the new genus Fascicularia, including three species formerly placed in Rhodostachys (F. bicolor, F. litoralis, F. pitcairniifolia) and separated them from the remaining species of Rhodostachys on the basis of petal appendages and pollen morphology. Mez (1896) realized that Ochagavia elegans and the species of Rhodostachys were closely related and therefore united the two genera, at first erroneously selecting the younger name Rhodostachys. This was corrected by Mez (1934) and Smith & Looser (1934), who published the presently accepted concept for the species group, comprising the genera Fascicularia and Ochagavia.
  Ochagavia can best be distinguished from Fascicularia by the relative lengths of the style and the stamens, by sepal shape, petal appendages, petal colour, and pollen morphology (see also below). Although flowering plants of Fascicularia and Ochagavia display striking differences, the relevant characters are difficult to study in herbarium specimens. Moreover, only some of the available specimens include floral parts (in the present study 39/89); even then, because of the mode of anthesis in bromeliads, only a single or a few fully developed flowers may be available for investigation. These genera therefore exemplify a major problem in bromeliad taxonomy which is that for most genera and species, the number of herbarium specimens available is comparatively small (due to difficulties in accessing, drying and handling of these plants) and thus our knowledge of variability of characters regarded of taxonomic importance is insufficient. Modern revisions including characters beyond morphology are scarce. Recently, comparative studies of morphology and micromorphology (e.g. Bohme, 1988; Gross, 1988; Schill, Dannenbaum & Jentzsch, 1988; Brown & Gilmartin, 1989; Brown & Terry, 1992), palynology (Halbritter, 1992), anatomy (Horres & Zizka, 1995) and DNA sequences (Clark et al., 1993; Ranker et al., 1990; Linder & Kellogg, 1995; Terry & Brown, 1996; Terry, Brown & Olmstead, 1997) have improved our knowledge about suprageneric grouping and relationships of the family.
  RESULTS
  Morphology
  For the delimitation of species, Mez (1896, 1934), as well as Smith & Downs (1979) who adopted Mez' classification, relied principally on the following morphological
  characters: length of floral bracts relative to length of flower, flower length, as well as shape and apex of the sepals. Our morphological studies concentrated on the significant characters of leaves, inflorescence and flowers. For the 89 available herbarium specimens (including duplicates; see also Appendix), the sizes (length/ width) of leaf sheath and leaf blade were measured, and from 39 specimens measurements of floral parts were obtained.
  Our investigations revealed that flower and bract length are highly correlated, the bracts being always shorter than the developed flower. The absolute length of the flower appears to be highly variable (with a tendency toward longer flowers in ssp. bicolor), but does not seem to be of value in separating any groups within the genus. Although considerable variation in shape and indumentum occurs, all sepals were observed to be ovate to oblong, apically retuse or praemorse, with a short inconspicuous mucro. In our opinion, none of these characters serves to distinguish or characterize groups within Fascicularia.
  Differences in shape and morphology of the leaf blade led us to define two taxa within Fascicularia (see also below), although these cannot be clearly separated using other morphological characters (Nelson & Zizka, 1997). The width of the leaf blade is the only, easily accessible morphological character that can be obtained from herbarium specimens which correlates fairly well with the different leaf blade types (Fig. 1). Discrimination of the two taxa was also supported by the results of factor analysis and discriminant analysis (performed with software SPSS) based on 13 morphological characters (length/width of leaf sheath, leaf blade, inflorescence, floral bracts, sepals, petals as well as flower length) measured from 39 herbarium specimens.
  
  Anatomy
  The anatomy of the leaf-blade in Fascicularia bicolor is similar to other investigated Bromelioideae. The species has a one-celled sclerotic hypodermal cell-layer. The inner hypodermal layers are sharply differentiated from the (sclerotic) outer layer and the inner, assimilating tissue (chlorenchyma), and form characteristic water storage tissue. In Fascicularia, as in most other Bromelioideae but not in Tillandsioideae, water storage tissue is well developed only adaxially, and is usually lacking abaxially. Nevertheless, the amount of water storage tissue in the leaf-blade (measured as percentage of area of transverse section) in F. bicolor ssp. bicolor and Ochagavia carnea reaches the highest values among the investigated Bromelioideae, similar to Pitcairnioideae like Puya roezlii or Hechtia glabra (Horres & Zizka, 1995).
  The two subspecies differ in leaf-shape and amount of adaxial water storagetissue. In ssp. bicolor, the adaxial surface of the leaf blade is not channelled longitudinally but is flat to slightly convex. In transverse section this can be easily seen, as well as the fact that the water storage tissue is better developed than in ssp. canaliculata. In the latter subspecies, the leaves are conspicuously channelled longitudinally. The leaf blade in ssp. bicolor is considerably thicker than in ssp. canaliculata and usually wider (see fig. 1 in Nelson & Zizka, 1997). In herbarium material, determining the subspecies is difficult. Drying and pressing causes the water storage tissue to collapse and it cannot be fully restored. Comparison with herbarium leaves from cultivated material leads to only slightly different leaf-shape of the two subspecies when the leaves are dried.
  Keeping the principal habitats of the two subspecies in mind (saxicolous near coast, epiphytic in forest), the differences in leaf anatomy might be regarded as mere variations corresponding to different ecological conditions without taxonomic value. However, the leaf characters proved to be constant in cultivation for a period of several years and correspond with two different banding patterns of RAPDs found in the cultivated material. Although not clearly separable by other morphological, especially floral, characters, the two groups are considered to be taxonomically distinct.
  
  Geographical distribution and ecology
  The specimens of Fascicularia bicolor we investigated were collected in Chile from app. 34°S to 42°24'S (Fig. 2). Smith & Downs (1979) cite a specimen from 45° 17'S (Puerto Lagunas, Fricke s.n.; CM; not seen), which extends the distribution considerably further south, making Fascicularia the southernmost distributed member of Bromeliaceae. This area comprises mediterranean vegetation types in the north (`bosque esclerofilo'; Grau, 1995), temperate broadleaved forest (`bosque caducifolio templado') and the region of the Valdivian rain forest in the south. The area of the bromeliad genus Greigia extends nearly as far south, reaching Cucao on Chiloe (42°42'S, 74°05'W; Bettina Will, pers. comm.).
  Although there is no clear separation between the distributions of the subspecies, the natural range of ssp. bicolor extends farther north into areas with mediterranean climate but including also the two above described more southern zones. Nevertheless, the ssp. bicolor is principally saxicolous, occurring in rocky habitats near the coast. On the other hand, F. bicolor ssp. canaliculata appears to have its centre of distribution in the area of temperate forests - the Valdivian rain forest and the broadleaved temperate forest - and usually grows as an epiphyte. These two forest types intergrade to some extent and share a considerable number of evergreen tree species (Schmithusen, 1956).
  Because of the difficulties in determining some herbarium specimens, a few specimens could not be assigned without doubt to one of the subspecies, and have been included in the map (Fig. 2) with a separate symbol. Further investigation of Fascicularia populations in Chile, especially careful description of leaf morphology when collecting herbarium specimens, will allow more precise information about the geographical distribution of the subspecies.
  Outside its natural distribution, Fascicularia bicolor has been introduced into gardens, and has become naturalized in western Europe. In favourable habitats, for example in the Isles of Scilly, western Ireland and the Channel Islands, Fascicularia colonies have spread, or been spread, to cover substantial areas of sand dunes and oak woodland.
  
  DNA analysis
  In recent years, the comparative analysis of DNA sequence variation has become increasingly important as a complementary tool to infer genetic relatedness among taxa (Bachmann, 1992; Hillis, Moritz & Mable, 1996). The random amplified polymorphic DNA (RAPD) marker technique described by Williams et al. (1990) is one of the most frequently used approaches of molecular taxonomy, mainly because of its technical simplicity. RAPD analysis makes use of the polymerase chain reaction (PCR; Mullis, Ferre & Gibbs, 1994) to generate polymorphic amplification products from anonymous regions of genomic DNA. Electrophoretic separation of RAPD products on agarose gels typically results in fingerprint-like banding patterns. Pairwise genetic distance matrices can be calculated from the extent of band sharing, and the results analysed phenetically. In plants, genetic relationships based on RAPD data generally proved to correlate well with expectations from morphological, ecological or phenological data, especially if very closely related taxa were analysed (see e.g. Adams & Demeke, 1993; Ramser et al., 1996,1997; Furman et al., 1997).
  In the present work, we used the RAPD technique to assess intraspecific relationships among 12 Fascicularia bicolor accessions (eight of ssp. bicolor, and four of ssp. canaliculata). To analyse the delimitation of the genus Fascicularia, we also included six accessions belonging to the related genera Ochagavia and Greigia (i.e. four accessions of Ochagavia carnea, and one accession each of 0. elegans and Greigia sphacelata). All plant material originated from the living collection of the Palmengarten (Frankfurt/ Main, Germany). Genomic DNA was isolated from freshly harvested leaves by the method of Doyle & Doyle (1987), treated with RNAse, and used for PCR without further purification. RAPD reactions were carried out with arbitrarily chosen GC-rich decamer primers (Operon, Alameda, FA), using the conditions described by Ramser et al. (1996, 1997). Of 14 primers initially tested, eight produced distinct polymorphic fingerprints and were used for further study. All assays were repeated independently, and only strong, reproducible bands were included in the final analysis. Banding patterns derived from primers OPG-7 and OPG-9 are exemplarily shown in Figure 3. Scoring for presence/ absence of a band at a particular position resulted in 98 polymorphic characters across all primers (Table 1). This binary character matrix was transformed into a pairwise distance matrix using the Jaccard coefficient of similarity ( Jaccard, 1908), and analysed by the unweighted pair group method using arithmetic averages (UPGMA; Sneath & Sokal, 1973). The resulting dendrogram is shown in Figure 4. The 12 Fascicularia accessions are separated into two clusters which coincide well with the subspecies designation recognized on the basis of leaf blade anatomy and morphology. Thus, the RAPD data fully support the division of Fascicularia bicolor into two subspecies. Our DNA analysis also supports a clear-cut delimitation of Fascicularia from Ochagavia and Greigia. Somewhat surprisingly, however, the latter two genera are not well distinguished from each other. Whereas the four 0. carnea accessions form a separate cluster, the RAPD patterns suggest a higher similarity of 0. elegans (which is endemic to the Juan Fernandez Islands) to Greigia sphacelata than to the other Ochagavia accessions analysed. This finding is in some contradiction with morphological data, since Ochagavia species are morphologically very similar to each other, but clearly distinct from all Greigia species. The genera Ochagavia and Greigia are currently under revision.
  The few DNA studies performed on Bromeliaceae so far (see Terry et al., 1997, and references cited therein) are mainly concerned with the still uncertain
  suprageneric phylogeny of the family. Our RAPD results underline that there is also a demand for a revision of the species and genus concept in the Bromeliaceae, taking into account morphological and anatomical as well as molecular data. In the course of such a revision, the taxonomic relevance of commonly used morphological descriptors may have to be re-evaluated.
  
  Descriptions and key
  Fascicularia Mez in Mart.
  Fascicularia Mez in Mart., Fl. Bras. 3,3: 627 (1894); Mez in DC., A. & C., Monogr. Phan. 9: 7 (1896); Harms in Engler & Prantl, Nat. Pflanzenfam, ed. 2, 15a: 132 (1930); Mez in Engler, Pflanzenr. 4,32: 5 (1934); Smith & Looser, Rev. Univ. 18,8: 1076 (1934); Munoz, Fl. silvestres Chile: 151 (1966); McClintock, Watsonia 10: 289 (1974-75); Smith & Downs, Fl. Neotrop. 14,3: 1711 (1979); Weber, J. Brom. Soc. 34,6: 265 (1984); Marticorena & Quezada, Gayana 42: 81 (1985); Benoit, Red List Chilean Terrestrial Plants: 119 (1989); Rauh & Gro13, Bromelien: 381 (1990); Nelson & Zizka, New Plantsman 4,4: 233 (1997).
  Bromelia auct. non L., p.p.; Ruiz & Pavon, Fl. Peruv. 3: 33 (1802); Molina, Sagg. Chil. ed. 2: 284 (1810), Gay, Hist. bot. Chil. 4: 9, t. 68 (1853); Morren, Belgique Hort. 23: 229, t. 14 (1873); Morren, Belgique Hort. 26: 161-162, pl. 10-11 (1876); Witte, Sempervirens 47: 553-557 (1889); Wittmack, Gartenflora 39: 345 (1890) Billbergia Schult. f. non Thunb., p.p. in Roem. & Schult., Syst. 7,2: 1255 (1830) Hechtia Riviere non Klotzsch, p.p.; Journ. Soc. Centr. Hort. France: 361 (1871); Morren, Belgique Hort. 26: 161 (1876); Wittmack, Gartenflora 39: 345 (1890) Rhodostachys auct. non Phil., p.p.; Bentham in Bentham & Hooker, Gen. Plant. 3,2; 662 (1883); Wittmack in Engler & Prantl, Nat. Pflanzenfam. 2,4: 45 (1888); Baker, Handb. Brom.: 27 (1889).
  
  Terrestrial or epiphytic herbs, stemless or with short, stout stem.
  Chromosome number. No information traced.
  
  Comments. The genus is best distinguished from the vegetatively similar genus Ochagavia by the relative length of style and stamina (Fasciculata: included in the flower; Ochagavia: emerging), shape of sepals (Fascicularia: retuse or praemorse, with short inconspicuous mucro; Ochagavia: acute with pungent apex), appendages and colour of petals (Fascicularia: appendages present; petals blue to violet; Ochagavia: appendages absent; petals rose), and pollen characters (Fascicularia: irregularly monocolpate; Ochagavia: monocolpate).
  The septal nectaries of F. bicolor are of special interest: Although the flowers are clearly epigynous as characteristic for Bromelioideae, the morphology of the septal nectaries approaches that of genera from Pitcairnioideae (Dyckia, Deuterocohnia, Abromeitiella). This might point toward Fascicularia representing a more primitive genus among the Bromelioideae (Bohme, 1988).
  
  Nomina excludenda
  1. Fascicularia pitcairniifolia (hort. ex B. Verl.) Mez in DC., A. & C., Monogr. Phan. 9: 10 (1896); Harms in Engler & Prantl, Nat. Pflanzenfam., ed. 2, 15a: 132 (1930); Mez in Engler, Pflanzenr. 4,32: 5 (1934); Smith & Downs, Fl. Neotrop. 14,3: 1711 (1979); McClintock, Watsonia 10: 289, 290 (1974-75); Rauh, J. Brom. Soc. 35,4: 213 (1984); Marticorena & Quezada, Gayana 42: 81 (1985); Rauh & GroI3, Bromelien: 382 (1990); Benoit, Red List Chilean Terrestrial Flora: 119 (1989). Hechtia pitcairniifolia hort. ex B. Verl., Revue Hort. 40: 211-212, pl. 1-2 (1868); Morren, Belgique Hort. 26: 161 (1876) ("pitcairniaefolia"); Wittmack, Gartenflora 39: 345 (1890) ("pitcairniaefolia"). Type: Plate 1 and 2, accompanying description pp. 211-212, Rev. Hort. (1868) (lecto).
  Rhodostachys pitcairniifolia (hort. ex B. Verl.) Benth. in Benth. & Hook. f, Gen. Plant. 3,2: 662 (1883); Baker, Handb. Brom.: 28 (1889) ("pitcairniaefolia"); C.H. Wright, Bot. Mag.: tab. 8087 (1906).
  Bromelia pitcairniifolia (hort. ex B. Verl.) K. Koch, Gartn. Wochensehr. 11: 325 (1868) ("pitcairnifolia"); Morren, Belgique Hort. 26: 161 (1876) ("pitcairniaefolia"); Wittmark, Gartenflora 39: 345 (1890) ("pitcairniaefolia").
  The illustration in the protologue, designated as the lectotype by Smith & Downs (1979), depicts a plant that obviously belongs to the genus Ochagavia. Its flowers have stamens clearly exceeding the petals, this being characteristic for species of Ochagavia, not occurring in the vegetatively very similar genus Fascicularia (stamens included).
  For detailed discussion see Nelson & Zizka (1997). In the protologue (and by several later authors), the epithet is spelled "picairniaefolia", the name referring to the genus Pitcairnia. This has been treated as orthographic error by subsequent authors according to ICBN (art. 60.8 and rec. 60G.1) and changed to "pitcairniifolia".
  
  Nomina dubia
  1. Billbergia joinvillei Van Houtte Hortus, Cat. 138: 92 (1871), nom. nud.; Morren, Belgique Hort. 26: 161 (1876); Wittmack, Gartenflora 39: 345 (1890), pro syn. Being first published as a nomen nudum, the latter two authors (Morren, 1876; Wittmack, 1890) only mention the name or cite name and place of publication among other synonyms. No type specimen, plate or description exist.
  Morren (1876) and Wittmack (1890) combine in their list of synonyms for a described taxon (Morren: Bromelia joinvillei, Wittmack: Rhodostachys pitcairniaefolia var. kirchhoffiana) names, which are in part regarded here as synonyms of Fascicularia bicolor, in part as synonyms of a species of the genus Ochagavia (e.g. Hechtia pitcairniifolia hort. ex B. Verl.). Though it is quite probable that the name belongs to the synonymy of Fascicularia bicolor or Ochagavia, it cannot be assigned with certainty and therefore is regarded as doubtful.
  The same applies also to the following names cited below (3. Hechtia carnea, 5. Hechtia joinvillei, 6. Pourretia africana, 7. Pourretia flexilis, 8. Pourretia mexicana) which seem to have been in use in European gardens in the 19th century.
  
  2. Fascicularia litoralis (Phil.) Mez in DC., A. & C., Monogr. Phan.: 9: 8 (1896); Mez in Engler, Pflanzenr. 4,32: 7 (1934); Smith & Downs, Fl. Neotrop. 14,3: 1714; McClintock, Watsonia 10: 289 (1974-75); Marticorena & CZuezada, Gayana 42: 81 (1985); Rauh, J. Brom. Soc. 35, 4: 213 (1984); Benoit, Red List Chilean Terrestrial Flora: 119 (1989).
  Rhodostachys litoralis Phil., Linnaea 30: 201 (1859) and Linnaea 33: 246 (1864-1865). Type: Chile, in regione litorali provinciae Colchagua, aliarumque ad rupes maritimas; leg. H. Volckmann s.n. (SGO?, not seen, probably destroyed); Baker, Handb. Brom.: 29 (1889) ("littoralis").
  Philippi's (1859: 202) description in the protologue is most probably based on a plant of the genus Ochagavia: ". . . petala ... rosea ... Stamina petala aequantia. Stylus staminibus paullo longior". The type specimen (leg. H. Volckmann) could not be traced by subsequent authors (Baker, 1889; Mez, 1896, 1934; Smith & Downs, 1979) nor by us. Therefore, the name is regarded as doubtful.
  
  3. Hechtia carnea hort. ex Wittm., Gartenflora 39: 346 (1890), nom. nud. For discussion see above (1. Bromelia joinvillei).
  
  4. Hechtia gracilis hort. ex Haage & Sehmidt, Cat.: 209 (1911), nom. nud.; Cat.: 210 (1912); Cat.: 211 (1913); Cat.: 188 (1916); Smith & Downs, Fl. Neotrop. 14,3: 1714 (1979), pro syn. No additional information about this name could be traced.
  
  5. Hechtia joinvillei hort. ex E. Morren, Belgique Hort. 26: 161 (1876), nom. nud. For discussion see above (l. Bromelia joinvillei).
  
  6. Pourretia africana hort. ex Wittm., Gartenflora 39: 345 (1890), pro syn. For discussion see above (1. Bromelia joinvillei).
  
  7. Pourretia flexilis hort. ex E. Morren, Belgique Hort. 26: 161 (1876), nom. nud.; Wittmack, Gartenflora 39: 345 (1890), pro syn. For discussion see above (1. Bromelia joinvillei).
  
  8. Pourretia mexicana hort. ex E. Morren, Belgique Hort. 26: 161 (1876), nom. nud.; Wittmack, Gartenflora 39: 345 (1890), pro syn. For discussion see above (1. Bromelia joinvillei).
  One species, endemic to Central and Southern Chile.