Sheet Music

The Deschampsia cespitosa complex in central and northern Europe: a morphological analysis

Fifty-four exomorphological characters were recorded from 130 herbarium specimens belonging to Deschampsia cespitosa (L.) P. Beauv. subsp. cespitosa, D. alpina (L.) Roem. et Schult., D. bottnica (Wahlenb.) Trin., D. cespitosa subsp. glauca (Hartm.)
of 18
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
   Botanical Journal of the Linnean Society  (2000),  134:   495–512. With 4 figuresdoi:10.1006/bojl.2000.0343, available online at on The  Deschampsia cespitosa  complex in centraland northern Europe: a morphological analysis  J. CHIAPELLA ∗  Departamento de Bota´nica, Universidad Nacional del Comahue, Quintral 1250,8400 Bariloche, Argentina  Received September 1999; accepted for publication February 2000  Fifty-four exomorphological characters were recorded from 130 herbarium specimens be-longing to  Deschampsia cespitosa   (L.) P. Beauv. subsp.  cespitosa  ,  D. alpina   (L.) Roem. et Schult.,  D. bottnica   (Wahlenb.) Trin.,  D. cespitosa   subsp.  glauca   (Hartm.) C. Hartm.,  D. littoralis   (Gaudin)Reuter,  D. cespitosa   subsp.  parviflora   (Thuill.) K. Richter and  D. wibeliana   (Sond.) Parl. Thesewere treated as Operational Taxonomical Units (OTUs) and studied using cluster analysis,principal component analysis and theMann–Whitney U-test;boxplots of selectedquantitativecharacters were also made. To evaluate the relationship between environment and mor-phology, a redundancy analysis was carried out.Due to extensiveoverlapping in the diagnosticcharacters, all the taxa at specific level are reduced to subspecies of   Deschampsia cespitosa  .Geographical distribution is considered to be of considerable importance in the identificationof the taxa. © 2000 The Linnean Society of London  ADDITIONAL KEY WORDS:—morphology – infraspecific classification – pseudovivipary – species complex – variation gradient. CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . 495Material and methods . . . . . . . . . . . . . . . . . . . 496Morphological analysis . . . . . . . . . . . . . . . . . . 496Relationship between morphological variability and environment . . . . 497Results . . . . . . . . . . . . . . . . . . . . . . . . 497Discussion . . . . . . . . . . . . . . . . . . . . . . . 499 Acknowledgements . . . . . . . . . . . . . . . . . . . . 507References . . . . . . . . . . . . . . . . . . . . . . . 507 Appendix 1: taxonomic treatment . . . . . . . . . . . . . . . . 508 Appendix 2: material studied . . . . . . . . . . . . . . . . . 510 INTRODUCTION The genus  Des champsia   P. Beauv. comprises about 40 species of annual andperennial grasses (Mabberley, 1990) occurring in cold temperate regions of both ∗ Present address: Institut fu¨r Botanik, Universita¨t Innsbruck, Sternwartestrasse 15, 6020 Innsbruck, Austria. E-mail: jorge· 4950024–4074/00/120495 + 18 $35.00/0  © 2000 The Linnean Society of London   J. CHIAPELLA496 hemispheres. The main subject of this paper,  D. cespitosa   (L.) P. Beauv., is anextremely variable grass with a wide distribution. It is a mainly northern hemispherespecies with an almost complete holarctic circumpolar area (Hulte´n, 1962); itspresence in West Greenland has been attributed to human introduction during the1950s (Polunin, 1959). Forms referred to it have been described also for Argentina (Parodi, 1949; Nicora, 1978), Brazil and Chile (Parodi, 1949), Australia and New Zealand (Edgar, 1993) and Central Africa (Hedberg, 1995). The taxonomy of the group, particularly in Central Europe, is still puzzling, because of the extrememorphological variability and ambiguous boundaries between the taxa. Hulte ´n(1962), Hedberg (1986) and Conert (1987) cited about 40 subspecies or varieties, together with an extensive synonymy. The partial treatments existing for  D. cespitosa  and allies (Polunin, 1959; Hulte´n, 1962; Kawano, 1963, 1966; Hess, Landolt & Hirzel, 1967; Chrtek & Jirasek, 1965; Clark e, 1978; Lid, 1985; Conert, 1987; Schmeil, 1988; Adler, Oswald & Fischer, 1994; Lauber & Wagner, 1998) represent only restricted views of a taxonomical problem which clearly exceeds the extent of regional floras.ThepresentstudydealswiththevariantsfoundincentralEurope(asgeographicallydefined in Hegi’s  IIlustrierte Flora von Mitteleuropa   [Conert, 1987] and northern Europe(including Norway, Sweden and Finland). The main objectives of this work are toanalyse morphological variation in relation to taxon boundaries, to evaluate char-acters that have been used to delimit taxa and to relate patterns of morphological variation with environmental variables. MATERIAL AND METHODS  Morphological analysis  One hundred and thirty herbarium specimens previously determined as  D. cespitosa  (L.) P. Beauv., subsp.  cespitosa  ,  D. alpina   (L.) Roem. et Schult.,  D. bottnica   (Wahlenb.)Trin.,  D. cespitosa   (L.) P. Beauv., subsp.  glauca   (Hartm.) C. Hartm.,  D. littoralis   (Gaudin)Reuter,  D. cespitosa   sbsp.  parviflora   (Thuill.) K. Richter and  D. wibeliana   (Sond.) Parl.belonging to IB, GZU, M, WU and S (Holmgren, Holmgren & Barnet, 1990) werestudied, treating each specimen as an operational taxonomic unit (OTU). The statesof 55 exomorphological characters (variables) were recorded and coded. The characters analysed were selected after Kawano (1963, 1966), Chrtek & Jirasek  (1965) and Conert (1987). Kawano (1963, 1966) analysed the variability of   D.cespitosa  , measuring culm height and branching, size, branching and scabrousness of panicles, shape, size and hairiness of glumes and lemmas, length of the awns andtheir point of insertion in the lemmas, and occurrence of vivipary (pseudovivipary).Chrtek & Jirasek (1965) provided a comprehensive account of the morphologicalfeatures used in the classification of   D. cespitosa   and relatives. In addition to thecharacters used by Kawano, they included length and width of the leaf blade, sizeand colour of the spikelets, number of florets of the spikelet, type of awn, morphologyof the palea and length of the ligule. Besides morphological features, Conert (1987)added geographic distribution to separate  D. cespitosa   (depicted as a species broadlydistributed) from  D. wibeliana   (restricted to inundation zones of the inferior courseof the Elbe, Eider and Weser rivers in Germany) and  D. littoralis   (found in shorelines   DESCHAMPSIA CESPITOSA  IN CENTRAL AND NORTHERN EUROPE 497 of rivers and lakes of Switzerland, with pseudoviviparous spikelets in the Bodensee).The characters selected (Table 1) include variables of continuous variation(characters 1, 5, 6, 7, 8, 12, 13, 18, 19, 26, 27, 30, 31, 36, 37, 44, 49), two-statecharacters (2, 16, 20, 21, 25, 29, 35, 38, 46, 47, 48, 50, 54) and multistatediscontinuous characters (3, 4, 9, 10, 11, 14, 15, 17, 22, 23, 24, 28, 32, 33, 34, 39,40, 41, 42, 43, 45, 51, 52, 53). The following analyses ere carried out. Cluster analysis.  (CA), with similarity matrices calculated by applying Pearson’scorrelation coe ffi cient. A dendrogram was obtained with Euclidean distance andWard’s method, which uses an analysis of variance approach to evaluate the distancesbetween clusters. The distortion of this technique was evaluated by the copheneticcorrelation coe ffi cient. Principal component analysis.  (PCA), in which a basic correlation matrix was calculatedwith the Pearson product-moment correlation among pairs of characters. The firsttwo principal components were used to plot the distribution of the OTUs. Theanalysis was carried out on standardized variables.  Morphometric analysis.  A non-parametric Mann–Whitney  U  -test for pairs of samples(Sokal & Rohlf, 1981) was carried out between  D. cespitosa   and each of the consideredtaxa. Box plots featuring medians, first and third quartiles and range of selectedcharacters were drawn according to Sokal & Rohlf (1981: 152). Relationship between morphological variability and environment  To assess the relationship between morphological variation and the environment,a redundancy analysis (RDA) was carried out. The biplot produced representsenvironmental variables and the associated morphological variables. A subset of 70specimens was selected. Environmental variables included annual average rainfall,annual a verage temperature of meteorological stations nearest to the collection localities (Walle´n, 1970), soils (Ganssen & Ha¨drich, 1965) and altitude above sea level data extracted from the specimen labels.The following software was used: Statistica (StatSoft, 1995) for the CA, Mann– Whitney  U  -test and box plots, NTSYS-pc (Rohlf, 1986) for the PCA, and Canoco 4.0 (Ter Braak & Smilauer, 1998) for the RDA. RESULTS Cluster analysis.  The dendrogram (Fig. 1) showed overlapping among all the OTUs.  D. cespitosa   subsp.  cespitosa   was scattered through the whole cluster, while the OTUsof the other taxa had more defined positions (e.g.  D. alpina   and  D. glauca   ).  D. bottnica  ,the hybrid  bottnica  × cespitosa   and  D. parviflora   also occupied defined positions, althoughless consistent than those of   D. alpina   and  D. glauca  . Finally  D. littoralis   and  D. wibeliana  were the taxa which showed the highest dispersion through the dendrogram. Thecophenetic correlation coe ffi cient showed a low cluster distortion (  r  = 0.89). Principal component analysis.  All the characters were used. A relatively low percentageof the variability is explained by the first 10 principal components (Table 2). Thefirst two axes accounted for only 16.7% of the total variability. Successive axes   J. CHIAPELLA498 T    1. Morphological characters and environmental variables used in numerical analyses of the  Deschampsia cespitosa   complex in central and northern Europe Morphological character Character states1 Total height (mm)2 Tillering intravaginal (0), extravaginal (1)3 Nodes of the fertile shoot (n o  )4 Nodes of the panicle (n o  )5 Length of panicle (mm)6 Width of panicle (mm)7 Distance between 1st and 2nd panicle nodes (mm)8 Distance between 2nd and 3rd panicle nodes (mm)9 Ramifications of the 1st node of the panicle (n o  )10 Ramifications of the 2nd node of the panicle (n o  )11 Scabrousness of secondary ramifications glabrous (0), few (1), numerous (2)12 Width of penultimate leaf (mm)13 Length of blade of leaf (mm)14 Scabrousness of veins of leaf-abaxial glabrous (0), few (1), numerous (2)15 Scabrousness of veins of leaf-abaxial glabrous (0), few (1), numerous (2)16 Nature of margin of leaf membranous (0), scarious (1)17 Scabrousness of leaf sheaf glabrous (0), few (1), numerous (2)18 Height of leaf node (mm)19 Length of ligule leaf (mm)20 Nature of ligule membranous (0), scarious (1)21 Shape of ligule apex sharp (0), obtuse (1)22 Colour of spikelets purple-violet (0), greenish (1), golden-leather yellow (2)23 Florets per spikelet 1 (0), 2 (1), 3 or more (2)24 Hairiness of rachilla glabrous (0), few (1), numerous (2)25 Shape of glume 1 lanceolate (0), acute (1)26 Length of glume 1 (mm)27 Width of glume 1 (mm)28 Number of nerves of glume 1 (n o  )29 Shape of glume 2 lanceolate (0), acute (1)30 Length of glume 2 (mm)31 Width of glume 2 (mm)32 Number of nerves of glume 2 (n o  )33 Scabrousness of veins of glumes absent (0), only midvein (1), all veins (2)34 Scabrousness between veins absent (0), present (1), abundant (2)35 Nature of margin of glumes membranous (0), scarious (1)36 Length of lemma 1 (mm)37 Width of lemma 1 (mm)38 Nature of lemma membranous (0), scarious (1)39 Relevant size of apical teeth all similar (0), lateral larger (1), central larger (2)40 Number of apical teeth (n o  )41 Number of nerves of lemma 1 (n o  )42 Scabrousness of nerves of lemma 1 glabrous (0), few (1), numerous (2)43 Scabrousness between nerves glabrous (0), few (1), numerous (2)44 Length of awn (mm)45 Insertion of awn basal 1/3 (0), medium 1/3 (1), superior 1/3 (2)46 Nature of awn straight (0), bended (1)47 Nature of awn II not twisted (0), twisted (1)48 Scabrousness of awn absent (0), present (1)49 Length of palea 1 (mm)50 Nature of palea 1 membranous (0), hyaline (1)51 Shape of palea dorsum rounded (0), bi-keeled (1), flattened (2)52 Number of nerves of palea 1 (n o  )53 Scabrousness of nerves of palea 1 glabrous (0), few (1), numerous (2)54 Pseudovivipary absent (0), present (1) continued    DESCHAMPSIA CESPITOSA  IN CENTRAL AND NORTHERN EUROPE 499 T    1—  continued  Environmental variables Variable states1 Altitude above sea level (m)2 Annual mean temperature (  ° C)3 Annual mean precipitation (mm)4 Soil type eutrophic to oligotrophic (partially podolized) brown soils (1),podzolic (2), chernozems (3), rendzina (4), hydromorphic (5),complex of young and tundra soils in high alpine regions (6),light chestnut (7), grey brown podzolic soils (8) explained important amounts of variability, evidence of high among-charactercorrelation. The bidimensional plot of all the OTUs in the first two axes of thePCA (Fig. 2) indicates distinctive but loosely cohesive groups for each taxon,overlapping to di ff   erent degree with the common subsp.  cespitosa  .  Morphometric analysis.  The Mann–Whitney  U  -test detected several characters withsignificant di ff   erences (Table 3). The taxa with more di ff   erences at the  P  <0.005level were  D. alpina   (eight characters out of 17) and  D. bottnica   (nine characters). Alsoat the  P  <0.005 level,  D. littoralis   and  D. cespitosa   subsp.  parviflora   had five characterswith significant di ff   erences.  D. wibeliana   showed the lowest di ff   erentiation from  D.cespitosa  , with the highest number of non-significant di ff   erences (12 characters). Boxplots of selected morphological characters are presented in Figure 3. Moderatepositive correlation was found between plant height and size of panicle (  r  = 0.67)and length of leaves (  r  = 0.68), and a weak negative correlation between plant heightand the size of spikelets (estimated in base to the length of the glume) (  r  =− 0.27).From this it followed that taller plants showed a tendency to have generally largerpanicles and leaves and, to a lesser degree, smaller spikelets. Relation between morphology and environment.  The RDA relationships between a subsetof 70 specimens and environmental variables (Fig. 4) showed association betweenrainfall and plant height, and between colour of spikelets and altitude. Thus, tallerplants were related to higher rainfalls, and darker spikelets to increasing altitude. DISCUSSION The abundance of overlapping among the character values of the considered taxaprovides support for the hypothesis of a closely related group. The high phenotypicplasticity of   D. cespitosa  , and its ability to produce ecotypes, has been long ac-knowledged (Lawrence, 1945; Kawano, 1963, 1966; Ward, 1969; Pearcey & Ward, 1972). In North America, five morphologically di ff   erent forms of   D . cespitosa   werefound; however, they were cytologically uniform (Lawrence, 1945) and thus con-sidered as ecotypes. An opposite situation to North America can be described forcentral and northern European populations of   D. cespitosa  , where di ff   erent chro-mosome counts have been reported (Table 4) and the taxa showed a high degreeof morphological similarity. Adaptation to particular environments often results inthe formation of ecotypes and may represent the beginning of speciation (Bradshaw,1965), yet the di ff   erences between them may be mainly physiological and genetic,
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks