Genetic Relationships Within and Among Iberian Fescues (Festuca L.) Based on PCR-Amplified Markers

1170 Genetic relationships within and among Iberian fescues (Festuca L.) based on PCR-amplified markers

Pedro J.G. de Nova, Marcelino de la Cruz, Juan V. Monte, and Consuelo Soler

Abstract: The genus Festuca comprises approximately 450 species and is widely distributed around the world. The Iberian Penninsula, with more than 100 taxa colonizing very diverse habitats, is one of its main centers of diversification. This study was conducted to assess molecular genetic variation and genetic relatedness among 91 populations of 31 taxa of Iberian fescues, based on several molecular markers (random amplified polymorphic DNA, amplified fragment length polymorphisms, and trnL sequences). The analyses showed the paraphyletic origin of the broad-leaved (subgenus Festuca, sections Scariosae and Subbulbosae, and subgenus Schedonorus) and the fine-leaved fescues (subgenus Festuca, sections Aulaxyper, Eskia, and Festuca). Schedonorus showed a weak relationship with Lolium rigidum and appeared to be the most recent of the broad-leaved clade. Section Eskia was the most ancient and Festuca the most recent of the fine-leaved clade. Festuca and Aulaxyper were the most related sections, in concordance with their taxonomic affinities. All taxa grouped into their sections, except F. ampla and F. capillifolia (section Festuca), which appeared to be more closely related to Aulaxyper and to a new independent section, respectively. Most populations clustered at the species level, but some subspecies and varieties mixed their populations. This study demonstrated the value in combining different molecular markers to uncover hidden genetic relationships between populations of Festuca. Key words: Festuca, AFLPs, RAPDs, trnL. Re´sume´ : Le genre Festuca comprend environ 450 espe`ces et il est largement re´pandu dans le monde. L’Ibe´rie, avec plus de 100 taxons occupant des habitats tre`s diffe´rents, constitue un des principaux centres de diversification. Cette e´tude a e´te´ reáliseé pour e´valuer la variation geńe´tique molećulaire et le degre´ de parente´ geńe´tique entre 91 populations appartenant a` 31 taxons de fe´tuques ibe´riennes a` l’aide de plusieurs types de marqueurs molećulaires (RAPD, AFLP et des se´quences de trnL). Ces e´tudes ont re´ve´le´ une origine paraphyle´tique des fe´tuques « a` feuilles larges » (sous-genre Festuca, sections Scariosae et Subbulbosae ainsi que le sous-genre Schedonorus) des fe´tuques « a` feuilles e´troites » (sous-genre Festuca, sections Aulaxyper, Eskia et Festuca). Le sous-genre Schedonorus montre une relation e´troite avec le Lolium rigidum et semble le plus rećent au sein du clade « a` feuilles larges ». La section Eskia est la plus ancienne et la section Festuca la plus rećente au sein du clade « a` feuilles e´troites ». Les sections Festuca et Aulaxyper sont les plus apparenteés conforme´- ment a` leurs affinite´s taxonomiques. Tous les taxons ont e´te´ classe´s au sein de leurs sections a` l’exception du F. ampla et du F. capillifolia (section Festuca) qui ont semble´ plus proches, respectivement, de la section Aulaxyper et d’une nouvelle section inde´pendante. La plupart des populations e´taient groupeés ensemble au niveau de l’espe`ce, mais certaines sous-espe`ces et varie´te´se´taient entremeˆleés. Cette e´tude de´montre l’utilite´ de combiner diffe´rents types de marqueurs pour e´lucider les relations geńe´tiques cacheés entre les populations de Festuca. Mots cle´s : Festuca, AFLP, RAPD, trnL. [Traduit par la Re´daction]

Introduction tribe Poeae) is comprised mainly of rhizomatous and cespi- tose perennial (sometime annual) allogamous grasses, with a Genus Festuca L. (family Poaceae, subfamily Pooideae, genome size ranging from diploid (2n = 14) to dodecaploid Received 2 March 2006. Accepted 14 June 2006. Published on (2n = 84) (Kergue´len and Plonka 1989). The number of spe- the NRC Research Press Web site at http://genome.nrc.ca on cies varies, according to different authors, between approxi- 8 November 2006. mately 300 and 450 (Tzvelev 1976; Clayton and Renvoize 1986; Watson and Dallwitz 1992), and increases periodically Corresponding Editor: J.P. Gustafson. as new species are described (Kergue´len et al. 1993; de la P.J. de Nova, J.V. Monte, and C. Soler.1 Departamento de Fuente and Ortuń˜ez 1994a, 1994b; Chusovljanov 1998; Biotecnologıá, INIA Apdo, 1045-28800, Alcala´ de Henares, Gonza´lez-Ledesma et al. 1998; Kulikov 1998, 2000; de la Spain. Fuente et al. 1999a, 1999b; Fletcher and Stace 2000). The M. de la Cruz. Departamento de Biologıá Vegetal, E.T.S.I. genus has a broad geographic distribution in both hemi- Agrońomos, Universidad Politećnica de Madrid, 28040 Madrid, spheres, and the area of distribution of its taxa varies from Spain. narrow endemic to subcosmopolitan. Some species of Fes- 1Corresponding author (e-mail: [email protected]). tuca are valued as fodder, as ornamental plants, or as grass

Genome 49: 1170–1183 (2006) doi:10.1139/G06-077 # 2006 NRC Canada de Nova et al. 1171 coverage for eroded soils. However, both the potential ge- ing unit (La Canaleja, Madrid, Spain). The taxa, netic resources and the natural genetic variability of most distribution of the populations in Spain, and ploidy level species remain unexplored. are reflected in Table 1 and Fig. 1. A population named The Iberian Peninsula is considered one of the main cen- calcicolous-F. indigesta was included in the section Festuca, ters of diversification of the genus (Saint-Yves 1930), according to its morphoanatomical characteristics, but it could where up to 109 taxa have been described (78 species and not be assigned to any known taxa (T.E. Dıáz-Gonza´lez, per- 31 subspecies and varieties), and it is recognised that the sonal communication, 2002). list is still incomplete (Cebolla and Rivas 2003). Current intrageneric taxonomy distinguishes 9 subgenera within DNA extraction and RAPD amplification Festuca (Clayton and Renvoize 1986), 3 of which appear Total DNA was extracted from the young green leaves of in the Iberian Peninsula: Festuca, Drymanthele V. Krecz 30 individual specimens of each population, in accordance and Bobrov, and Schedonorus (P. Beauv.) Peterm (Kergue´len with the method of Dellaporta et al. (1983). RAPDs were and Plonka 1989). Most Iberian species are classified within performed in a reaction volume of 25 mL, containing subgenus Festuca (in sections Aulaxyper, Eskia, Festuca, 12.5 ng total DNA, 2.5 mL10Â reaction buffer, 7.5 ng pri- Scariosae and Subbulbosae); only 6 taxa belong to subgenus mer (Kits M, R, S, and T, Operon Technologies, Alameda, Schedonorus (sections Plantynia and Schedonorus) and 2 to Calif.), 100 mmol dNTP/L, and 1.25 U Taq DNA polymer- subgenus Drymanthele (Clayton and Renvoize 1986, Cebolla ase (Sigma, St. Louis, Mo.). Amplifications were carried out and Rivas 2003). in a PTC-100 thermocycler (MJ Research Inc.). PCR fol- Within the Iberian Peninsula, genus Festuca has a broad lowed the method of Williams et al. (1990), with minor ecological range, and its species appear from sea level to modifications: 40 cycles of 1 min at 94 8C, 1 min at 36 8C, the highest summits of Sierra Nevada and the Pyrenees, and 2 min at 72 8C. Amplification products were then re- growing in a variety of habitats, from damp to dry soil or solved by electrophoresis in 1.8% agarose gel, using TAE rock crevices and from meadows to scrub and forest. Fes- buffer, and visualized with ethidium bromide staining. Each tuca frequently plays a role in the structure of plant com- reaction and corresponding electrophoresis was repeated at munities. least twice. The great morphologic similarity between different species, the existence of numerous ecotypes, and the range of intrapopulational differentiation of taxa make the study and AFLP procedures classification of the genus a difficult task. To overcome AFLP analysis was performed in accordance with the problems such as the interference of the environment over methods described by Vos et al. (1995). Genomic DNA the morphoanatomical characteristics usually used in classi- (350 ng) was double-digested with 2 U each of EcoRI and fication, several molecular tools have been recently em- MseI (Roche Diagnostics GmbH, Mannheim, Germany), in ployed: restriction fragment length polymorphism (Xu and a final volume of 35 mL1Â buffer A, and incubated for 3 h Sleper 1994), random amplified polymorphic DNA at 37 8C. Adaptors EcoRI and MseI (5 and 50 pmols, re- (RAPD) (Stammers et al. 1995; Siffelova et al. 1997), inter- spectively) were ligated to the resulting fragments (30 mL nal transcribed spacer (Charmet et al. 1997; Gaut et al. of the digestion mix), using 1 U T4 DNA ligase (Roche Di- 2000; Torrecilla and Catalań 2002), and chloroplast DNA agnostics GmbH) in a final volume of 50 mL1Â buffer li- variability analysis (Lehva¨slaiho et al. 1987; Darbyshire gase, and incubated for 2 h at 25 8C. The ligation mix was and Warwick 1992; Catalań et al. 2004; Torrecilla et al. diluted 1/10, and 5 mL were added to the preamplification, 2004). containing 1Â reaction buffer, 30 ng each of primers In this paper, we try to shed light on the genetic variabil- EcoRI+1 and MseI+1, 0.25 mmol each dNTP/L, and 1 U of ity and the relations and differences among populations of Taq DNA polymerase (Sigma) in a final volume of 20 mL. Iberian wild fescues, using the molecular random PCR am- Preamplification was performed in a PTC-100 thermocycler plification of RAPD, amplified fragment length polymor- as follows: 2 min at 95 8C; 20 cycles of 30 s at 94 8C, 1 min phisms (AFLPs) (Williams et al. 1990; Vos et al. 1995), at 56 8C, 1 min at 72 8C; and 5 min at 72 8C. The pream- and the sequencing of the intron of chloroplast gene trnL plification mix was diluted 1/50, and 5 mL were added to the (Taberlet et al. 1991). selective amplification, containing 1Â reaction buffer, 30 ng each of primers EcoRI+3 and MseI+3, 0.25 mmol each Materials and methods dNTP/L, and 1 U of Taq DNA polymerase in a final volume of 20 mL. Selective amplification was performed as follows: Plant material 13 touchdown cycles (–0.7 8C per cycle) of 30 s at 94 8C, A total of 91 populations, belonging to 2 subgenera and 6 30 s at 65 8C, 1 min at 72 8C; 23 cycles of 30 s at 94 8C, sections of the genus Festuca L. and growing in the Iberian 30 s at 56 8C, 1 min at 72 8C; and 10 min at 72 8C. Five Peninsula were selected. Samples of Lolium rigidum (Gen- combinations of primer+3 were used with either the Bank accession No. DQ376043), Agrostis castellana, and EcoRI+3 or MseI+3 primer end-labelled with 6-FAM Puccinellia fasciculata (acc. No. DQ376076) were added to (Table 2). The PCR products were separated by electropho- the study as controls. To conserve the original genetic struc- resis in a 6% polyacrylamide gel and analysed in an ABI ture of the natural populations, specimens were collected us- PRISM 310 Genetic Analyser (Applied Biosystems). The ing the method of Hawkes (1980). This plant material reproducibility of the AFLP results was tested by replicat- belongs to the collection of wild Poaceae obtained from nat- ing the complete procedure in 5 DNA samples for all the ural environments and maintained at the INIA plant breed- primer+3 combinations.

# 2006 NRC Canada 1172 Genome Vol. 49, 2006

Table 1. Taxa and populations included in the study.

Taxa and population ID Origin Latitude Longitude Chromosome No. GenBank acc. No. Subgen. Festuca Sect. Festuca F. ampla Hackel 2n =28 AMP1 Ca´diz 36845’49@N5823’00@W DQ376048 AMP2 Ciudad Real 38830’56@N3837’14@W AMP3 Jaeń38803’45@N4800’28@W AMP4 Jaeń38817’12@N2835’20@W F. arvernensis Auquier, 2n =28 Kergueleń & Markgr.- Dannenb. subsp. costei (St.-Yves) Auquier & Kergueleń ´ ARVc Barcelona 41828’27@N1848’14É DQ376070 F. capillifolia Dufour in 2n =14 Roemer & Schultes CAP1 Almerıá37817’22@N2830’55@W DQ376072 CAP2 Almerıá34842’26@N2813’17@W CAP3 Almerıá37815’48@N2828’15@W CAP4 Cuenca 39857’23@N1815’12@W F. clementei Boiss 2n =14 CLE Granada 37803’49@N3822’12@W DQ376073 F. curvifolia Lag. ex 2n =42 Lange CUR1 Madrid 40850’13@N3849’50@W CUR2 Madrid 40849’21@N3856’54@W DQ376058 CUR3 Madrid 40847’31@N4800’15@W DQ376059 CUR4 Madrid 40842’26@N4808’55@W F. gracilior (Hackel) 2n =14 Markgr.-Dannenb. GRA1 Cuenca 39857’23@N1815’12@W GRA2 Teruel 40810’10@N1818’25@W DQ376064 GRA3 Teruel 40812’14@N1821’08@W F. hystrix Boiss 2n =14 HYS1 Jaeń38807’05@N2834’10@W DQ376068 HYS2 Albacete 38835’26@N2825’38@W HYS3 Guadalajara 40850’40@N1853’02@W HYS4 Guadalajara 40841’55@N2813’15@W F. indigesta Boiss. 2n =42 IND1 Granada 37803’49@N3822’12@W DQ376071 IND2 Granada 37806’48@N3825’03@W Calcicolous F. indigesta 2n =42 INDc Oviedo 42859’55@N5855’52@W DQ376065 F. lemanii Bastard, Ess. 2n =28 LEM Gerona 42815’53@N3817’14@E DQ376060 F longiauriculata Fuente, 2n =14 Ortuń˜ez & Ferrero LON1 Granada 37807’34@N3802’07@W DQ376067 LON2 Almerıá37813’25@N2832’01@W F. marginata (Hackel) K. 2n =14 Richter subsp. alopecuroides (Hackel) K. Richter MARa1 Cuenca 40827’35@N2802’08@W DQ376063 MARa2 Teruel 40829’16@N1841’15@W MARa3 Teruel 40830’54@N1839’44@W MARa4 Teruel 40830’02@N1840’31@W

# 2006 NRC Canada de Nova et al. 1173

Table 1 (continued).

Taxa and population ID Origin Latitude Longitude Chromosome No. GenBank acc. No. F. rivas-martinezii 2n =14 Fuente & Ortuń˜ez subsp. rivas-martinezii RIVr1 Madrid 41803’32@N3828’11@W DQ376062 RIVr2 Madrid 41805’22@N3829’34@W RIVr3 Madrid 40858’35@N3849’18@W RIVr4 Madrid 40852’51@N3846’19@W RIVr5 A´ vila 40840’55@N4841’03@W F. rivas-martinezii 2n =28 Fuente & Ortuń˜ez subsp. rectifolia Fuente & Ortuń˜ez RIVre Teruel 40812’14@N1821’08@W DQ376061 (Tormoń) F. summilusitana Franco 2n =42 & Rocha Afonso SUM1 Palencia 43802’04@N4844’32@W DQ376069 SUM2 Palencia 43800’47@N4843’41@W SUM3 Leoń42854’50@N5859’15@W SUM4 Leoń42853’31@N5858’32@W SUM5 Leoń42852’43@N5857’48@W F. valentina (St-Yves) 2n =14 Markgr.-Dannenb. VAL1 Jaeń38806’51@N2833’43@W DQ376066 VAL2 Jaeń38806’30@N2834’46@W VAL3 Jaeń38807’05@N2834’10@W VAL4 Jaeń38807’28@N2833’03@W VAL5 Jaeń38806’44@N2832’50@W Sect. Aulaxyper F. heterophylla Lam. 2n =28 subsp. heterophylla HETh Barcelona 41828’27@N1848’14@E DQ376050 F. iberica (Hack.) K. 2n =28 Richter IBE1 Granada 37806’48@N3825’03@W IBE2 Madrid 41805’22@N3829’34@W IBE3 Madrid 40852’51@N3846’19@W IBE4 Guadalajara 41811’42@N3825’38@W DQ376046 F. nevadensis (Hackel) 2n =14 K. Richter var. gaetula (Maire ex St. Yves) Al-Bermani & Stace NEVg1 Jaeń37852’47@N2852’39@W DQ376049 NEVg2 Jaeń37853’02@N2853’00@W NEVg3 Jaeń37854’24@N2854’28@W F. nevadensis (Hackel) 2n =28 K. Richter var. nevadensis NEVn1 Granada 37858’16@N2827’06@W DQ376051 NEVn2 Jaeń37849’12@N2858’33@W NEVn3 Jaeń38807’11@N2840’21@W F. nigrescens Lam. 2n =42 subsp. nigrescens NIGn1 Palencia 43800’47@N4843’41@W NIGn2 Leoń42853’31@N5858’32@W DQ376055 NIGn3 Oviedo 43802’04@N5842’56@W F. nigrescens Lam. 2n =14 subsp. nigrescens NIGn4 Leoń42842’16@N6855’15@W

# 2006 NRC Canada 1174 Genome Vol. 49, 2006

Table 1 (continued).

Taxa and population ID Origin Latitude Longitude Chromosome No. GenBank acc. No. NIGn5 Lugo 42853’04@N7804’42@W NIGn6 Leoń42853’32@N6848’25@W NIGn7 Lugo 42855’19@N6851’56@W DQ376056 NIGn8 Lugo 42845’30@N7801’00@W NIGn9 Lugo 42844’07@N6859’55@W NIGn10 Lugo 42843’35@N6858’47@W F. rothmaleri (Litard.) 2n =28 Markgr.-Dannenb. ROT Madrid 41804’53@N3827’09@W DQ376047 F. rubra L. subsp. 2n =42 pruinosa (Hackel) Piper RUBp Oviedo 43839’20@N5850’56@W DQ376057 F. trichophylla (Ducros 2n =28 ex Gaudin) K. Richter subsp. asperifolia (St.-Yves) Al-Bermani TRIa Guadalajara 40840’02@N2817’13@W DQ376052 F. trichophylla (Ducros 2n =14 ex Gaudin) K. Richter subsp. scabrescens (Hackel ex Trabut) Catalań & Stace TRIs1 Guadalajara 40856’41@N2851’52@W TRIs2 Guadalajara 40855’34@N2855’21@W TRIs3 Cuenca 40836’56@N2816’40@W TRIs4 Teruel 40828’05@N1836’33@W TRIs5 Cuenca 40825’52@N1853’31@W TRIs6 Teruel 40821’09@N1844’32@W TRIs7 Teruel 40830’54@N1839’44@W DQ376053 TRIs8 Teruel 40830’05@N1839’01@W F. trichophylla (Ducros 2n =28 ex Gaudin) K. Richter subsp. trichophylla TRIt1 Palencia 42848’46@N4843’35@W TRIt2 Soria 41808’00@N2826’48@W DQ376054 TRIt3 Teruel 40821’09@N1844’32@W TRIt4 Guadalajara 40841’46@N1854’14@W Sect. Eskia F. burnatii St-Yves 2n =14 BUR1 Leoń42853’31@N5858’32@W DQ376075 BUR2 Oviedo 42859’08@N5853’27@W F. elegans subsp. merinoi 2n =28 (Pau) Fuente & Ortuń˜ez ELEm Leoń42840’19@N6847’29@W DQ376074 Sect. Scariosae F. scariosa (Lag.) 2n =14 Ascherson & Graebner SCA1 Almerıá37802’04@N2857’03@W DQ376045 SCA2 Almerıá37818’53@N2830’26@W SCA3 Almerıá37816’07@N2827’27@W Sect. Subbulbosae F. paniculata (L.) Schinz 2n =14 & Tell. subsp. paui Cebolla & Rivas Ponce PANpa Cuenca 40808’49@N2841’28@W DQ376044

# 2006 NRC Canada de Nova et al. 1175

Table 1 (concluded).

Taxa and population ID Origin Latitude Longitude Chromosome No. GenBank acc. No. Subgen. Schedonorus Sect. Schedonorus F. arundinacea Schreb. 2n =28 subsp. mediterranea (Hackel) K. Richter ARUm Ma´laga 36856’34@N4827’55@W DQ376042

Fig. 1. Map of the Iberian Penninsula with the localization of the studied sections: !, section Festuca; ~, section Aulaxyper; *, section Eskia; &, section Scariosae; ^, section Subbulbosae; +, section Schedonorus. Coordinates of individual taxa can be found in Table 1.

trnL intron amplification and sequencing amplified bands were discarded and fragments of identical The chloroplast noncoding region of the trnL (UAA) in- size were considered to be the same. Presence (1) or absence tron was used for phylogenetic analysis. Amplification was (0) were used as character states. The data matrices were an- performed using 100 ng genomic DNA, 1Â reaction buffer, alysed with the NTSYS computer package, version 1.60 0.2 mmol each dNTP/L, 100 ng of each primer (5’-CGA- (Rolhlf 1989). The Jaccard similarity coefficient (Jaccard AATCGGTAGACGCTACG-3’ and 5’-GGGGATAGAGGG- 1908) and the unweighted pair-group method, arithmetic ACTTGAAC-3’) (Taberlet et al. 1991), and 2 U of Taq averages (UPGMA) (Sneath and Sokal 1973) were com- DNA polymerase in a final volume of 100 mL. PCR con- bined to construct dendrograms. ditions were programmed as follows: 16 touchdown cycles (–0.5 8C per cycle) of 1 min at 94 8C, 1 min at 62 8C, trnL alignment and sequence analysis 2 min at 72 8C; 29 cycles of 1 min at 94 8C, 1 min at First, sequences were aligned using the CLUSTAL W 54 8C, 2 min at 72 8C; and 10 min at 72 8C. Amplifica- (1.82) program with default parameters (Thompson et al. tion product was extracted with QIAquick Gel Extraction 1994). After alignment, phylogenetic analysis was conducted Kit (Qiagen GmbH, Hilden, Germany), after 1% agarose with the PAUP 4.0 beta 8 program (Swofford 1998), using gel electrophoresis in TAE buffer. For each sample, direct the parsimony criterion (Swofford and Berlocher 1987), a and reverse sequences were obtained using the 2 amplifi- heuristic search procedure, and the MULPARS, TBR branch cation primers. swapping, and CLOSEST addition options. Bootstrap support was estimated with 1000 replicates (Sanderson 1989). RAPD and AFLP statistical analysis A sample of the species Puccinellia fasciculata was used in the study as the outgroup. The resulting tree was constructed RAPD and AFLP data were scored on 2 matrices. Faint using TreeView (Page 1996).

# 2006 NRC Canada 1176 Genome Vol. 49, 2006

Table 2. Sequences of the adaptors and primers used in amplified lata subsp. paui, the only representative taxon of section fragment length polymorphisms (AFLPs). Subbulbosae, appeared close to section Eskia. The clusters of sections Festuca and Aulaxyper grouped first at the high- Type Sequences est similarity, and the clusters of sections Subbulbosae, Es- Adaptor EcoRI+ 5’-CTCGTAGACTGCGTACC-3’ kia, Schedonorus, and Scariosae were added sequentially Adaptor EcoRI– 3’-CATCTGACGCATGGTTAA-5’ with progressive dissimilarity. The taxonomic markers Adaptor MseI+ 5’-GACGATGAGTCCTGAG-3’ Agrostis and Puccinellia appeared in the outermost posi- Adaptor MseI– 3’-TACTCAGGACTCAT-5’ tions of the dendrogram; Lolium joined section Scariosae. Primer EcoRI+1 5’-GACTGCGTACCAATTCA-3’ Within section Festuca, F. valentina, F. gracillior, and Primer MseI+1 5’-GATGAGTCCTGAGTAAC–3’ F. rivas-martinezii subsp. rectifolia grouped together in the Primer EcoRI+3 6-FAM-5’-GACTGCGTACCAATTCACC-3’ innermost cluster, showing the greatest genetic relatedness Primer EcoRI+3 6-FAM-5’-GACTGCGTACCAATTCACG-3’ (similarity over 0.488); F. summilusitana, F. indigesta, and Primer EcoRI+3 5’-GACTGCGTACCAATTCAGC-3’ calcicolous-F. indigesta grouped in another cluster (simi- Primer EcoRI+3 6-FAM-5’-GACTGCGTACCAATTCAGG-3’ larity over 0.462), and were joined by F. curvifolia Primer MseI+3 6-FAM-5’-GATGAGTCCTGAGTAACAG-3’ (with similarity of 0.315). Another cluster was formed Primer MseI+3 6-FAM-5’-GATGAGTCCTGAGTAACTG-3’ by F. arvenensis, F. marginata, and F. rivas-martinezii Primer MseI+3 5’-GATGAGTCCTGAGTAACAA-3’ subsp. rivas-martinezii (similarity over 0.255), and a 4th one by F. hystrix and F. longiauriculata (similarity over Results 0.211). F. lemanii and F. clementei remained as the outermost species, joining the rest of the section with simi- RAPDs larity indices of 0.156 and 0.102, respectively. PCR amplification with the 30 primers generated 1389 Within section Aulaxyper, F. nigrescens populations ap- fragments. Of these, 724 (52.12%) fragments were present peared in 2 distinct groups related to ploidy level (3 popula- only within populations, 986 (70.99%) were present only tions with 2n = 42 and 7 with 2n = 14). They were joined within species, subspecies and varieties, and 1248 (89.85%) by F. trichophylla populations (similarity of 0.263) and, in were present only within sections. None of the fragments turn, were joined by the group of F. nevadensis (with simi- was present in all populations analysed. larity of 0.236). The group of F. rothmaleri and F. hetero- Within populations, F. curvifolia had the maximum Jac- phylla joined with less similarity, and were subsequently card similarity coefficient (average 0.780) and F. iberica joined by the populations of F. ampla (with similarity of had the lowest (average 0.293). F. gracilior and F. rivas- 0.116). martinezii subsp. rectifolia were the most similar taxa (average 0.574), and the minimum similarity was between AFLPs F. burnatii and both F. capillifolia and F. arundinacea Five primer combinations generated 1347 fragments be- subsp. mediterranea (0.000). The 3 minima of similarity between 70 and 488 bp. A high percentage of fragments tween F. capillifolia and sections Aulaxyper, Eskia, and (52.26%) were under 200 bp, and only 5.50% exceeded their own section Festuca was remarkable. Festuca ampla 400 bp. No fragment was shared by all populations; 419 was much more similar to almost all populations of section fragments (31.11%) were present only within populations; Aulaxyper than to any population of its own section Festuca. 681 (50.56%) within species, subspecies, or varieties; and The average similarity of genus Festuca to the controls 1011 fragments (75.06%) were present only in sections. L. rigidum, A. castellana, and P. fasciculata were 0.008, Similarity values were, in general, higher than in RAPD 0.016, and 0.00001, respectively. analysis. Festuca longiauriculata had the highest average The highest similarity values within sections (average similarity within species (0.829); F. trichophylla subsp. similarity of all populations) were in Scariosae and Eskia scabrescens and F. iberica had the lowest (both 0.388). (0.316 and 0.331, respectively), and the lowest were in Festuca gracilior and F. rivas-martinezii subsp. rectifolia Aulaxyper and Festuca (0.235 and 0.231, respectively). showed the highest similarity between taxa (0.676), whereas The highest similarities between sections were between the lowest was found between F. burnatii and F. scariosa Aulaxyper and Festuca (0.081) and between Subbulbosae (0.034). Average similarity between genus Festuca and and Eskia (0.078); the least related sections were Schedonorus controls were 0.083, 0.031, and 0.042 for Lolium, Agrostis, and Eskia (0.003). and Puccinellia, respectively. Within sections, again Scar- The dendrogram clearly discriminated all taxa, showing iosae and Festuca had the highest and lowest values, re- high levels of variability even inside individual species spectively (0.743 and 0.351). Between sections, Festuca (Fig. 2). In general, populations of the same species clus- and Aulaxyper had the highest average similarity (0.169), tered together, but some populations of F. trichophylla, and Eskia and Scariosae had the lowest (0.047). F. iberica, F. marginata, and F. rivas-martinezii joined clus- The AFLP dendrogram also showed high variability within ters of other taxa. In the case of F. rivas-martinezii, the pop- sections and species, although with higher similarities than in ulations that did not cluster were from different subspecies. the RAPD case (Fig. 3). Again, Festuca and Aulaxyper ap- Species from the same section appeared together in the same peared more related to themselves than to the remaining sec- clusters, except in the case of F. capillifolia and F. ampla, tions (0.177). Section Schedonorus was the most similar to both of section Festuca. Festuca ampla clearly joined the both (0.098), followed by Scariosae (0.075) and Eskia- Aulaxyper cluster and F. capillifolia joined F. arundinacea Subbulbosae (0.071). The group of Agrostis and Puccinellia subsp. mediterranea (section Schedonorus). Festuca panicu- controls appeared most isolated, sharing a 0.037 similarity

# 2006 NRC Canada de Nova et al. 1177

Fig. 2. Unweighted pair-group method with arithmetic averages (UPGMA) dendrogram of Festuca populations, using the Jaccard similarity coefficient, based on 1389 random amplified polymorphic DNA (RAPD) markers.

# 2006 NRC Canada 1178 Genome Vol. 49, 2006

Fig. 3. UPGMA dendrogram of Festuca populations, using the Jaccard similarity coefficient based on 1347 amplified fragment length polymorphism (AFLP) markers.

# 2006 NRC Canada de Nova et al. 1179 with the main group of fescues. F. paniculata subsp. paui subsp. mediterranea (Schedonorus) appeared closely related (sect. Subbulbosae) joined section Eskia, as it did in the to L. rigidum, both with a common and recent origin. RAPD dendrogram. F. ampla appeared again inside the In the other branch, the Eskia group showed their ancient main group of section Aulaxyper, instead of within the Fes- origin, whereas sections Aulaxyper and Festuca diverged tuca group. F. capillifolia also quit section Festuca. Lolium much more recently. All taxa of section Aulaxyper formed a rigidum appeared closer to F. arundinacea subsp. mediter- paraphyletic grouping, except F. rubra subsp. pruinosa and ranea (similarity of 0.113), instead of joining F. scariosa, F. nigrescens subsp. nigrescens, which shared a mono- as it did in the RAPD dendrogram. In general, and as in phyletic origin. Section Festuca diverged even more re- the RAPD dendrogram, populations of the same taxa cently, and all their taxa except F. ampla, F. capillifolia, grouped together, except again in the case of F. tricho- and F. clementei formed a paraphyletic group. F. ampla phylla, F. iberica, F. rivas-martinezii, F. marginata, and joined the paraphyletic group of Aulaxyper taxa, whereas F. nigrescens. The genetic distance between the 2 ploidy F. capillifolia and F. clementei diverged much earlier and groups of F. nigrescens was even greater than in RAPDs, formed a paraphyletic grouping with F. elegans subsp. mer- and the separation of F. rivas-martinezii populations re- inoi of section Eskia. Both populations of F. curvifolia flected its intraspecific taxonomy. showed a monophyletic origin with a high bootstrap value. As in the RAPD dendrogram, the innermost cluster within The 2 populations of F. nigrescens subsp. nigrescens also section Festuca was formed by F. gracilior, F. valentina, showed a monophyletic origin but, in this case, with little and F. rivas-martinezii subsp. rectifolia (similarity of support of bootstrap, probably because of their different 0.622); F. indigesta, calcicolous-F. indigesta, and F. summi- ploidy level. lusitana were grouped in another cluster, with similarity over 0.494, and F. arvernensis, F. marginata, and F. rivas- Discussion martinezii clustered with a similarity of 0.439. In contrast, F. hystrix and F. clementei were the outermost species and The high level of polymorphism detected in this study re- joined the main group with similarities under 0.300. flects the high genetic diversity of genus Festuca. The de- gree of polymorphism was lower in AFLPs, probably Within section Aulaxyper, F. heterophylla and F. rothma- because most fragments (52.26%) were under 200 bp; these leri had the strongest interspecific relationship; these were smaller fragments are less likely to contain polymorphisms joined by F. ampla (from section Festuca) at a similarity because they have fewer base pairs than the bigger RAPD level of 0.371. As in the RAPD dendrogram, F. nigrescens fragments. However, with both RAPDs and AFLPs, all pop- subsp. nigrescens populations appeared in 2 groups but, in ulations were clearly differentiated and, for almost every this case, with an even weaker relationship between them- taxon, the geographic and the genetic distances between selves; the hexaploid populations grouped in the outermost their populations were highly correlated (data not shown). cluster within Aulaxyper. Festuca valentina and F. scariosa In addition, the high values of consistence and retention in- populations had the strongest intraspecific relationship (they dices of the phylogenetic tree, based on the trnL, together grouped at similarity levels of 0.751 and 0.703, respec- with the lower homoplasy index, showed the optimization F. capillifolia F. nevadensis tively); and had the lowest of the cladistic analysis. (under 0.643 and 0.702, respectively). As in the RAPD case, relationships between intraspecific taxa, such as sub- The results of the analyses clarified the relationships be- species and varieties, were not the strongest in the dendro- tween Festuca populations at different scales. At the section gram. level, both RAPDs and AFLPs, together with the trnL sequencing, highlighted the close relationship between sections Aulaxyper and Festuca, consistent with the similar patterns trnL sequences of chromosome C-banding (Dawe 1989) and their morphoa- The PCR amplification and sequencing of the homologue natomical similarity (on this basis, both were included in regions of the intron of the chloroplast trnL gene resulted in section Ovinae by Hackel 1882). Although Darbyshire and a product of unique sequences, ranging from 496 bp in Warwick (1992), based on chloroplast DNA restriction F. summilusitana to 516 bp in F. burnatii (among the popu- site variation, proposed a paraphyletic origin, the mono- lations of fescues), and to 519 bp in L. rigidum. After the phyletic origin of both sections was shown by Gaut et al. alignment and coding of gaps, the dataset comprised 530 (2000), Torrecilla and Catalań (2002), and Catalań et al. characters. A remarkable gap was a 9 bp deletion shared by (2004), using internal transcribed spacer sequences. Some all populations of section Festuca and Aulaxyper except authors (Tzvelev 1983; Holub 1984) proposed that the F. capillifolia and F. clementei. The relative frequencies of broad-leaved fescues (sections Scariosae, Schedonorus, bases were A, 0.372; T, 0.303; C, 0.142; and G, 0.183. and Subbulbosae) were ancestral to the ‘‘fine-leaved’’ fes- Transversions were more abundant than transitions (ratio, cues (sections Aulaxyper, Eskia, and Festuca). Our trnL 0.635). Phylogenetic analysis detected 50 variable charac- phylogram showed the paraphyletic origin of both groups, ters, of which 24 were parsimony informative. The most and a similar divergence time, corroborating the results of parsimonious tree (Fig. 4) had a total length of 66 (consis- Torrecilla and Catalań (2002). tency index, 0.833; retention index, 0.903; homoplasy index, Section Eskia can be considered an intermediate group 0.167), and the bootstrap support of the nodes ranged be- between broad- and fine-leaved fescues because of its ana- tween 51% and 98%. The tree showed 2 diverging branches; tomical similarities with section Scariosae (Torrecilla et al. in one of them, sections Scariosae, Subbulbosae, and Sched- 2004). In our study, section Subbulbosae joins section Eskia onorus shared a common ancestor, and F. arundinacea in both RAPDs and AFLPs, a fact that can be explained by

# 2006 NRC Canada 1180 Genome Vol. 49, 2006

Fig. 4. Strict consensus of the most parsimonious trees obtained from a phylogenetic analysis of trnL intron sequences. Numbers are bootstrap values of the nodes.

the existence of the homoplasy phenomena; however, both RAPD analysis. The relationship between Lolium and sections appear in separate branches of the phylogram and Schedonorus has been repeatedly described (Holub 1984; diverge at different times. Lehva¨slaiho et al. 1987; Morgan et al. 1988; Darbyshire Chandrasekharan et al. (1972) reported a high genetic af- 1993; Stammers et al. 1995; Pasakinskiene´ et al. 1998). Cy- finity between section Scariosae and section Schedonorus tological studies have noted the relationship between Lolium but, in our study, this relationship appears only in the mono- and ancestors of Festuca (Jenkin 1933; Malik and Thomas phyletic origin of both within the broad-leaved group. A re- 1966), whereas other studies, based on comparative anat- cent botanical study (de la Fuente and Ortu´n˜ez 2001) omy, have suggested a relationship between the reduced in- proposed the inclusion of F. scariosa in section Eskia, but florescence of Lolium and some ancestor of F. pratensis all of our analyses suggest that it should remain in its own (Essad 1962). On the basis of these relationships, some au- section. thors have concluded that Schedonorus should be treated as The species included as controls were only weakly associ- a subgenus of Lolium (Peto 1933; Crowder 1953; Stebbins ated with Festuca taxa, except in the case of L. rigidum, 1956; Terrell 1966; Darbyshire 1993). In contrast, the mor- which was associated with section Schedonorus in AFLPs phology of the spikelet, the existence of only diploid ge- and in the phylogenetic tree and to section Scariosae in the nomes (proof of its more recent origin), and the differences

# 2006 NRC Canada de Nova et al. 1181 in the proteins of the endosperm have been stated as reasons high genetic polymorphism and have been grouped with to keep Lolium and Schedonorus as independent genera populations of other subspecies of F. trichophylla, making (Aiken et al. 1997, 1998; Holub 1998). Our study supports their intraspecific taxonomic structure unrecognisable. The the idea of including Lolium as a subgroup of Schedonorus, taxonomic relatedness of F. trichophylla subsp. scabrescens as has been proposed by Terrell (1968). and F. iberica remains controversial. Al-Bermani et al. At the specific or intraspecific level, some interesting re- (1992) proposed that they were the same taxon, whereas lationships between species have been revealed. For inde la Fuente et al. (1997) reaffirmed the independence of stance, Bulinska-Radomska and Lester (1986) presented both taxa based, among other reasons, on their different ge- both F. heterophylla and F. rubra as transitional between nome size. The grouping patterns of both taxa did not con- sections Schedonorus and Ovinae (as the summation of firm either hypotheses because, although some populations Aulaxyper and Festuca), but our study shows that they are were grouped together, others remained entirely independ- fully Aulaxyper taxa. ent. The separation of F. nevadensis populations did not re- F. ampla has long been classified within section Festuca flect their varietal status, although, in this case, the (Hackel, 1882) although it shows some of the anatomical molecular techniques could not discriminate among vari- features that Stace et al. (1992) consider characteristic of eties because their differential characteristics appear mixed the ‘‘F. rubra aggregate’’: the absence of auricles and adax- in the populations (Kergue´len and Plonka 1989). In con- ial schlerenchyma strands in their leaf blades. Also, from an trast, the 10 populations of F. nigrescens subsp. nigrescens ecological point of view, the requirement of deep soil by formed 2 distinct groups, 1 for each ploidy level in the F. ampla (de la Fuente and Ortuń˜ez 1995) more closely re- RAPD and AFLP dendrograms, suggesting that they could sembles the ecological affinities of most Aulaxyper taxa than be considered different subspecies or even species. In addi- those of section Festuca, which are better adapted to shal- tion, it is noteworthy that F. nigrescens associates with low soils. All our analyses confirm these relationships and F. rubra subsp. pruinosa in the trnL phylogram, supporting suggest that F. ampla should be placed in section Aulaxyper. the monophyletic origin of the F. rubra group, which Al- F. clementei and F. capillifolia, both classified within sec- Bermani et al. (1992) discriminated from the F. tricho- tion Festuca (and both with the adaxial schlerenchyma char- phylla group on the basis of some morphoanatomical characteristic), appeared in an ancestral position on the branch of acteristics. fine-leaved fescues in the phylogram (together with Eskia taxa), and quite separate from the rest of section Festuca in Acknowledgements the RAPD and AFLP dendrograms. Saint-Yves (1922) ques- tioned the taxonomic position of F. capillifolia and proposed The authors gratefully acknowledge Dr. T.E. Dıáz Gonza´lez the creation of a subsection (Exaratae) to distinguish it from (Universidad de Oviedo, Spain) for his advice on calcicolous the rest of Festuca (a proposal that did not succeed in the F. indigesta, Dr. Pilar Catalań (Universidad de Zaragoza, taxonomic literature); our results agree with that proposition Spain) for her helpful remarks about this study, and Mer- but suggest that the rank of the taxonomic category should cedes Martıńez for her linguistic assistance. We also ac- be greater than a subsection (see also Torrecilla and Catalań knowledge the financial support given by INIA (project 2002; Catalań et al. 2004). SC97-065 and RTA02-038). Within section Festuca, the grouping of F. arvernensis subsp. costei, F. rivas-martinezii, and F. marginata subsp. References alopecuroides could be related to their anatomical similar- Aiken, S.G., Dallwitz, M.J., McJannet, C.L., and Consaul, L.L. ities (all of them bear discontinuous abaxial sclerenchyma 1997. Biodiversity among Festuca (Poaceae) in North America: strands in their leaf blades, a differential characteristic in diagnostic evidence from DELTA and clustering programs, and Festuca taxonomy), although 1 population of F. marginata an INTKEY package for interactive, illustrated identification subsp. alopecuroides and 1 of F. rivas-martinezii subsp. and information retrieval. Can. J. Bot. 75: 1527–1555. rectifolia did not join the group. The separation of the sub- Aiken, S.G., Gardiner, S.E., Bassett, H.C.M., Wilson, B.L., and species of F. rivas-martinezii into different groups supports Consaul, L.L. 1998. Implications from SDS-PAGE analyses of the idea of Cebolla and Rivas (2003) of 2 different species: seed proteins in the classification of taxa of Festuca and Lolium F. longifolia subsp. rivas-martinezii (Fuente & Ortuń˜ez) (Poaceae). Biochem. Syst. Ecol. 26: 511–533. Cebolla et al. and F. rectifolia (Fuente et al.) Cebolla & Al-Bermani, A.K.K.A., Catalań, P., and Stace, C.A. 1992. A new Rivas Ponce (Cebolla and Rivas 2003). circumscription of Festuca trichophylla (Gaudin) K. Richter The genetic similarity of F. valentina and F. gracilior (Gramineae). An. Jard. Bot. Madr. 50: 209–220. (both group in the AFLP and RAPD dendrogram) is consis- Bulinska-Radomska, Z., and Lester, R.N. 1986. Phylogeny of chro- tent with their morphoanatomical similarity; some authors mosome races of Festuca arundinacea and F. mairei (Poaceae) (de la Fuente and Ortuń˜ez 1998) treat them as the same spe- as indicated by seed protein electrophoresis. Plant Syst. Evol. 152: 153–166. cies, whereas others consider them different species (Cebolla Catalań, P., Torrecilla, P., Lo´pez Rodrı´guez, J.A., and Olm- and Rivas 2001, 2003). The absence of any special relation- stead, R.G. 2004. Phylogeny of the festucoid grasses of ship in the trnL sequence prevents further conclusions about subtribe Loliinae and allies (Poeae, Pooideae) inferred their status. from ITS and trnL-F sequences. Mol. Phylogenet. Evol. The calcicolous F. indigesta population was genetically 31: 517–541. PMID:15062792. similar to genuine F. indigesta, both in RAPDs and AFLPs, Cebolla, C., and Rivas, M.A. 2001. Festuca michaelis (Poaceae), which is consistent with their morphological similarities. une nouvelle espe`ce pour la Peńinsule Ibe´rique. Flora Mediterr. F. trichophylla subsp. scabrescens populations showed a 11: 363–371.

# 2006 NRC Canada 1182 Genome Vol. 49, 2006

Cebolla, C., and Rivas, M.A. 2003. Cata´logo del geńero Festuca L. Hackel, E. 1882. Monographia festucarum Europaearum. Kassel (Poaceae) en la Penıńsula Ibe´rica. Candollea, 58: 189–213. und Berlin. p. 216. Chandrasekharan, P., Lewis, E.J., and Borrill, M. 1972. Studies in Hawkes, J.G. 1980. Crop genetic resources. Field collection manual Festuca. 2. Fertility, relationships between species of the sec- for seed crops, root and tuber crops, tree fruit crops and related tions Bovinae and Scariosae, and their affinities with Lolium. wild species. International Board for Plant Genetic Resources Genetica, 43: 375–386. doi:10.1007/BF00156133. (I.B.P.G.R.) and the European Association for Research on Plant Charmet, G., Ravel, C., and Balfourier, F. 1997. Phylogenetic ana- Breeding (EUCARPIA), Birmingham, UK. lysis in the Festuca-Lolium complex using molecular markers Holub, J. 1984. New genera in Phanerogamae. Folia Geobot. Phy- and ITS rDNA. Theor. Appl. Genet. 94: 1038–1046. doi:10. totax. 19: 95–99. 1007/s001220050512. Holub, J. 1998. Reclassifications and new names in vascular plants Chusovljanov, D.V. 1998. A new species of the genus Festuca 1. Preslia, 70: 97–122. (Poaceae) from Altai. Bot. Z. 83: 113–115. Jaccard, P. 1908. Nouvelles recherches sur la distribution florale. Clayton, W.D., and Renvoize, S.A. 1986. Genera graminum. Bull. Soc. Vaudoise Sci. Nat. 44: 223–270. Grasses of the world. Kew Publishing, London, UK. Jenkin, T.J. 1933. Interspecific and intergeneric hybrids in herbage Crowder, L.V. 1953. Interspecific and intergeneric hybrids of Fes- grasses. Initial crosses. J. Genet. 28: 205–264. tuca and Lolium. J. Hered. 44: 195–203. Kergue´len, M., and Plonka, F. 1989. Les Festuca de la Flore de Darbyshire, S.J. 1993. Realignment of Festuca Subgenus Schedo- France (Corse comprise). Bull. Soc. Bot. Centre-Ouest, 10: 1–368. norus with the Genus Lolium (Poaceae). Novon, 3: 239–243. Kergue´len, M., Plonka, F., and Chas, E´ . 1993. Nouvelle contribu- doi:10.2307/3391460. tion aux Festuca (Poaceae) de France. Lejeunia, 142: 1–42. Darbyshire, S.J., and Warwick, S.I. 1992. Phylogeny of North Kulikov, P.V. 1998. A new species of the genus Festuca (Poaceae) American Festuca (Poaceae) and related genera using chlor- from the southern Urals. Bot. Z. 83: 110–115. oplast DNA restriction site variation. Can. J. Bot. 70: Kulikov, P.V. 2000. A new species of the genus Festuca (Poaceae) 2415–2429. from north-western Caucasus. Bot. Z. 85: 124–127. Dawe, J.C. 1989. Sectional survey of Giemsa C-banded karyotypes Lehva¨slaiho, H., Saura, A., and Lokki, J. 1987. Chloroplast DNA in Festuca L. (Poaceae: Festuceae): systematic and evolutionary variation in the grass tribe Festuceae. Theor. Appl. Genet. 74: implications. Ph.D. thesis, Viena University. 298–302. Dellaporta, S.L., Wood, J., and Hicks, J.B. 1983. A plant DNA Malik, C.P., and Thomas, P.T. 1966. Karyotypic studies in some minipreparation: version II. Plant Mol. Biol. Rep. 1: 19–21. Lolium and Festuca species. Caryologia, 19: 167–196. Essad, S. 1962. E´ tude geńe´tique et cytogeńe´tique des espe`ces Lo- Morgan, W.G., Thomas, H., and Lewis, E.J. 1988. Cytogenetic Stu- lium perenne L., Festuca pratensis Huds. et leurs hybrides. dies of Hybrids between Festuca gigantea Vill., and Lolium Ann. Inst. Nat. Rech. Agron. Ser. B. No. 8. multiflorum Lam. Plant Breed. 101: 335–343. doi:10.1111/j. Fletcher, R., and Stace, C.A. 2000. A new section and species of Fes- 1439-0523.1988.tb00306.x. tuca (Poaceae) naturalized in England. Watsonia, 23: 173–177. Page, R.D.M. 1996. TREEVIEW: An application to display phylo- de la Fuente, V., and Ortuń˜ez, E. 1994a. A new species of Festuca genetic trees on personal computers. Comput. Appl. Biosci. 12: L. from Spain. Bot. J. Linn. Soc. 114: 23–30. doi:10.1006/bojl. 357–358. PMID:8902363. 1994.1002. Pasakinskiene´, I., Anamthawat-Jońsson, K., Humpreys, M.W., Pa- de la Fuente, V., and Ortuń˜ez, E. 1994b. A new species of Festuca plauskiene, V., and Jones, R.N. 1998. New molecular evidence L. (Poaceae) from Pyrenees (Spain). Fontqueria, 40: 35–42. on genome relationships and chromosome identification in fes- de la Fuente, V., and Ortuń˜ez, E. 1995. Festuca henriquesii Hackel y cue (Festuca) and ryegrass (Lolium). Heredity, 81: 659–665. F. ampla Hackel en la Penıńsula Ibe´rica. Stud. Bot. 14: 129–141. doi:10.1046/j.1365-2540.1998.00446.x. de la Fuente, V., and Ortuń˜ez, E. 1998. Biosistema´tica de la sec- Peto, F.H. 1933. The cytology of certain intergeneric hybrids be- cioń Festuca del geńero Festuca L. (Poaceae) en la Penıńsula tween Festuca and Lolium. J. Genet. 28: 113–156. Ibe´rica. Ediciones de la Universidad Autońoma de Madrid, Can- Rolhlf, J.F. 1989. NTSYS.pc: numerical taxonomy and multivariate toblanco, Madrid. analysis system for the IBM pc microcomputer (and compati- de la Fuente, V., and Ortuń˜ez, E. 2001. Festuca sect. Eskia (Poa- bles). Applied Biostatistics INC, Setavek, New York. ceae) in the Iberian Peninsula. Folia Geobot. 36: 385–421. Saint-Yves, A. 1922. Les Festuca (Eu-Festuca)deL´ Afrique du de la Fuente, V., Ortuń˜ez, E., and Ferrero, L.M. 1997. Contribucioń Nord et des Iles Atlantiques. Candollea, 1: 1–63. al conocimiento del geńero Festuca L. (Poaceae) en el Paı´s Saint-Yves, A. 1930. Aperçu sur la distribution geógraphique des Vasco y Sistema Ibe´rico septentrional (Penıńsula Ibe´rica). Itin- Festuca (Subgen. Eu-Festuca). Candollea, 4: 146–165. era Geobot. 10: 317–351. Sanderson, M.J. 1989. Confidence limits on phylogenies: the boot- de la Fuente, V., Ortuń˜ez, E., and Ferrero, L.M. 1999a. A new spe- strap revisited. Cladistics, 5: 113–129. cies of Festuca L. (Poaceae) from the Iberian Peninsula. Parla- Siffelova, G., Pavelkova, M., Klabouchova, A., Wiesner, I., and Na- torea, 3: 65–77. sinec, V. 1997. Computer-aided RAPDs fingerprinting of acces- de la Fuente, V., Ortuń˜ez, E., and Ferrero, L.M. 1999b. New spe- sions from ryegrass-fescue complex. J. Agric. Sci. 129: 257–265. cies of Festuca L. section Festuca (Poaceae) in the Iberian Pe- Sneath, P.H.A., and Sokal, R.R. 1973. Numerical taxonomy. Free- ninsula. Lazaroa, 20: 3–9. man, San Francisco. Gaut, B.S., Tredway, L.P., Kubik, C., Gaut, R.L., and Meyer, W. Stace, C.A., Al-Bermani, A.K.K.A., and Wilkinson, M.J. 1992. The 2000. Phylogenetic relationships and genetic diversity among distinction between the Festuca ovina L., and Festuca rubra L. members of the Festuca-Lolium complex (Poaceae) based on aggregates in the British Isles. Watsonia, 19: 107–112. ITS sequence data. Plant Syst. Evol. 224: 33–53. doi:10.1007/ Stammers, M., Harris, J., Evans, G.M., Hayward, M.D., and For- BF00985265. ster, J.W. 1995. Use of random PCR (RAPD) technology to ana- Gonza´lez-Ledesma, M., Koch, S.D., and Go´mez, M. 1998. Two lyse phylogenetic reltionships in the Lolium/Festuca complex. new species of Festuca (Gramineae: Pooideae) from Mexico. Heredity, 74: 19–27. PMID:7852097. Novon, 8: 147–151. doi:10.2307/3391986. Stebbins, G.L. 1956. Taxonomy and evolution of genera with spe-

# 2006 NRC Canada de Nova et al. 1183

cial reference to family Gramineae. Evolution, 10: 235–245. Torrecilla, P., Lo´pez Rodrı´guez, J.A., and Catala´n, P. 2004. Phylo- doi:10.2307/2406009. genetic relationships of Vulpia and related genera (Poeae, Poa- Swofford, D.L. 1998. PAUP*. Phylogenetic analysis using parsi- ceae) based on analysis of ITS and trnL-F sequences. Ann. Mo. mony (*and other methods). Version 4. Sinauer Associates, Sun- Bot. Gard. 91: 124–158. derland, Massachusetts. Tzvelev, N.N. 1976. Grasses of the Soviet Union. Zlaki SSSR. Swofford, D.L., and Berlocher, S.H. 1987. Inferring evolutionary Translation for the Smithsonian Institution (1984), Amerind trees from gene frequency data under the principle of maximum Publishing Co., New Delhi. parsimony. Syst. Zool. 36: 293–325. doi:10.2307/2413068. Tzvelev, N.N. 1983. Grasses of the Soviet Union. Part II. Oxonian Taberlet, P., Gielly, L., Pauton, G., and Bouvet, J. 1991. Universal Press, New Delhi. primers for amplification of three non-coding regions of chloro- Vos, P., Hogers, R., Bleeker, M., Rijans, M., van de Lee, plast DNA. Plant Mol. Biol. 17: 1105–1109. doi:10.1007/ T., Hornes, M., et al. 1995. AFLP: a new technique for BF00037152. PMID:1932684. DNA fingerprinting. Nucleic Acids Res. 23: 4407–4414. Terrell, E.E. 1966. Taxonomic implications of genetics in ryegrass PMID:7501463. (Lolium). Bot. Rev. 32: 138–164. Watson, L., and Dallwitz, M.J. 1992. The grass genera of the Terrell, E.E. 1968. A taxonomic revision of the genus Lolium. world. C.A.B. International, Wallingford. pp. 1038. USDA-ARS Tech. Bull. 1392: 1–65. Williams, J.G.K., Kubelik, A.R., Livak, K.J., Rafalski, J.A., and Thompson, J.D., Higgins, D.G., and Gibson, T.J. 1994. CLUSTAL W: Tingey, S.V. 1990. DNA polymorphisms amplified by arbitrary improving the sensitivity of progressive multiple sequence align- primers are useful as genetic markers. Nucleic Acids Res. 18: ment through sequence weighting, position-specific gap penalties 6531–6535. PMID:1979162. and weight matrix choice. Nucleic Acids Res. 22: 4673–4680. Xu, W.W., and Sleper, D.A. 1994. Phylogeny of tall fescue and re- PMID:7984417. lated species using RFLPs. Theor. Appl. Genet. 88: 685–690. Torrecilla, P., and Catala´n, P. 2002. Phylogeny of Broad-leaved doi:10.1007/BF01253971. and Fine-leaved Festuca Lineages (Poaceae) based on Nuclear ITS Sequences. Syst. Bot. 27: 241–251.

# 2006 NRC Canada