Patterns of Genetic Variability in Populations of Adenostyles Cass

Delpinoa, n.s. 44: 103-114. 2002

Patterns of genetic variability in populations of Adenostyles Cass. complex (Asteraceae) along the Apennine chain

1 1 1 ANNA MARIA PALERMO ,GIUSEPPE PELLEGRINO ,MARIA ELENA NOCE , 2 1 LILIANA BERNARDO ,ALDO MUSACCHIO

1Dipartimento di Ecologia, 2Orto Botanico, Università della Calabria, I-87036 Arcavacata di Rende, CS, Italy.

Riassunto. In popolazioni disgiunte di Abstract. Allozymes and RAPDs markers were Adenostyles presenti lungo l'Appennino, catena tested in disjunct populations of Adenostyles com- montuosa che rappresenta un buon modello di plex along the Apennine chain, which is a good gradiente eco-geografico, sono stati testati allo- eco-geographic north-south gradient model. In our zimi e RAPD. I RAPD sono risultati più variabili survey, RAPD was higher than allozyme variability, rispetto agli allozimi, giacché ogni individuo ha with each individual showing a different RAPD mostrato un genotipo differente dagli altri. genotype. Both kinds of markers pointed out that Entrambi i marcatori hanno dimostrato che la majority of total genetic diversity resides within diversità genetica maggiore è presente all'interno populations, while geographic groups are scarcely delle popolazioni, mentre i gruppi geografici divergent and no clear geographic pattern was sono scarsamente divergenti e non è stato riscon- detected. Indeed, the dendrograms indicate that trato alcun gradiente geografico. La struttura populations group together without congruence genica spaziale non sarebbe da attribuire al flus- with their regional location. Data tend to exclude so genico, ma piuttosto rappresenterebbe la con- that this spatial genetic structure may be attributed seguenza di contatti secondari. Questi eventi pro- to a constantly maintained gene flow, but rather babilmente avvennero durante i cambiamenti cli- could be explained as a consequence of secondary matici del Quaternario, mentre la separazione fra contact events. These events probably occurred le popolazioni esaminate sarebbe avvenuta di during the climatic changes of Quaternary, and the recente. In aggiunta, sono stati discussi gli aspet- separation of the examined populations seems very ti conservazionistici, considerando sia la bassa recent. A conservational point of view is discussed, consistenza delle popolazioni sia la riduzione dei considering the low number of individuals in these loro ambienti. populations and the scarcity of their habitats.

Key words: Adenostyles, allozymes, Apennine chain, Genetic variability, RAPDs, Spatial genetic structure

INTRODUCTION vivor populations on the Alps, few studies have been conducted on the behaviour of In these last years, with the develop- herbaceous species during glaciations in ment of molecular markers, greater atten- the Mediterranean basin (COMES & tion has been placed to phylogeographic KADEREIT, 1998), which has already pro- studies on plants. In particularly, refugia ven to be an important refugium area for and recolonization of arctic and alpine taxa many tree species (COMES &KADEREIT, have been investigated (i.e., BAUERT et al., 1998; WILLIS, 1996). In this context the 1998; STEHLIK et al., 2001), with the aim Apennine chain, which has been viewed as of substantiating or disproving the nuna- an useful model of eco-geographic north- takker theory (STEHLIK, 2000). south gradient, represents a potential refu- Independently from the occurrence of sur- gium for alpine species and, eventually, a secondary colonisation route. In fact, its ties of an unique species, namely A. alpina orophilous floristic component includes, (FIORI, 1925-29), whereas in more recent among others, several Alpine species, time they have been treated discordantly as capable of reaching different latitudes and different subspecies or species (PIGNATTI, that have spread along the chain presu- 1982; WAGENITZ, 1983). In particular for mably during the Quaternary climatic these territories A. glabra (Mill.) DC., A. change. Due both to the peculiar orogra- australis (Ten.) Nyman and A. orientalis phic conformation and to the anthropic Boiss. have been reported, although their pressure, the populations of these species, differentiation has been based on few living in disjunction with respect to their highly variable and quantitative characters. main areas of distribution, are often Infact, these species are notoriously diffi- strongly isolated and progressively rare- cult to determine and some reports of their fied (PIGNATTI, 1980; PASSALACQUA & presence are questioned (PIGNATTI, 1982). BERNARDO, 1997). It has been shown that populations from The study of spatial structure of these central and northern zones of Apennines species with genetic markers may contri- were erroneously referred to A. glabra, bute to elucidate their evolutionary history whereas they belong to A. australis. On the and to shed light on the ecological and contrary, the alpine populations are refer- genetical influences of the colonisation red to A. glabra and not to A. australis. and isolation events experienced by these (PIGNATTI, 1982). In any case, regardless of taxa (THOMPSON, 1999; FREVILLE et al., taxonomic implications Apenninic 2001). Moreover, recent occurrence of Adenostyles complex may represent a good bottleneck or expansion in range may have model for exploring the pattern of genetic significant effects on actual population variation within and among populations of genetic structure (HANNAN &ORICK, an Alpine element, progressively further 2000). Unfortunately, there is a lack of from its main distribution area. As a matter interest towards the genetic structure of of fact, its effectively linear scattering natural plant populations along the transet along a north-south environmental gra- of this region (ZHANG et al., 2001; dient may reveal patterns of gene flow or GRÜNANGER et al., 1998; FRIZZI & phylogeografic routes beside their taxono- TAMMARO, 1991). mic evaluation. Consequently, in this paper The genus Adenostyles Cass. is preva- we chose to treat all examined populations lently distributed throughout the Alpine as belonging to Adenostyles complex. territories, but its range extends also to In addition, because no population Northern Spain, Corsica (France), Greece genetic information exists on these spe- and along the entire Apennine chain in cies, genetic variability analysis may Italy (including mountains of Calabria and encourage future conservation measures. Sicily) (WAGENITZ, 1983; TUTIN, 1976). Knowledge of population genetic structure The Apenninic entities of Adenostyles, provides an historical perspective of evolu- which occupy wet and shady habitats in tionary changes that characterise a species beech forests above 1400 m a.s.l., repre- and allows to predict how populations will sent a complex of strictly related taxa, respond to future natural events similar to each other from both a morpho- (VRIJENHOEK, 1987; HUENNEKE, 1991; logical and ecological points of view. In LUTZ et al., 2000). Consequently, the suc- the past, they have been regarded as varie- cessful preservation of populations of rare,

104 endangered or fragmented taxa may greatly benefit from similar genetic variation results. In this paper we investigate the allozymic and RAPD variability of populations of the Apenninic Adenostyles entities. These markers have been widely applied to describe the genetic structure and divergence existing within and among populations (HAMRICK &GODT, 1989; EDWARDS, 1998), with the aim of understanding the evolutionary factors responsible for their variation (SUN et al., 1999). RAPDs allow to emphasise variability in populations of endemic or rare plant species, even when allozymes failed in this goal (AYRES & RYAN, 1999). In brief, the major objectives of the present study are a) to evaluate relative levels Fig. 1 - Geographic distribution of populations of of genetic variation within Apenninic Adenostyles complex along the Apennine populations, b) to quantify differentiation chain among populations and gene flow, c) to scattered populations and the relative scar- assess genetic distances between popula- city of their habitats (CORBETTA et al., tions in relation to their geographic distan- 1998). ces, d) to indicate the potential consequences of disjunction and the causes responsible for the present Adenostyles distribu- MATERIALS AND METHODS tion, e) to determine the possible conservational implications, since Adenostyles may Plant material be included at least among the rare species, Samples were collected from two popu- considering the low consistency of their lations in alpine localities and from eight

Tab. 1 - Sampled populations of Adenostyles

Sites Region Abbr. N. of sampled analyzed Allozyme sRAPDs

Cozzo del Pellegrino Pollino massif (northern Calabria) PO-1 15 8 Tavolara Pollino massif (northern Calabria) PO-2 15 8 River Pollino massif (northern Calabria) PO-3 15 8 Botte Donato Sila (central Calabria) SI-1 11 8 Fontana Colonnello Sila (central Calabria) SI-2 66 Cervati Cilento (Campania) CI-1 15 8 Sicignano Cilento (Campania) CI-2 15 8 Mount Velino Marsica mountains (Abruzzo) AB-1 11 8 Magasa Eastern Alps (Lombardia) AL-1 15 8 Trento Eastern Alps (Trentino) AL-2 44

105 populations along the Apennines chain mic number of all Adenostyles is constant (Fig. 1). All populations were constituted throughout the genus (2n=38) (FEDEROV, by a few specimens, but in all cases the 1974). numbers of studied samples were represen- tative (Tab. 1). One or two leaves were col- RAPDs lected from each of individuals per popula- Stored leaf tissue (approx. 0.3 g) of each tion and stored on ice in individually label- sample was frozen in liquid nitrogen and led plastic bags until they were transferred ground into a fine powder. Total DNAs to a -80°C freezer, where they were kept were then extracted according to DOYLE & until processed. We conducted RAPD DOYLE (1987). assays on leaf tissue from a subset of the RAPD markers were generated in 25 l same individuals used for the isozyme containing 0.05 mmol of each of deoxynu- analyses. cleotide, 2 mmol MgCl2, 1x reaction buffer, 0.5 U Taq polymerase (Bioline USA, Allozymes Inc), 0.04 mmol decamer primer and 10 ng Enzymes were extracted by grinding sample DNA. PCR reactions were carried leaf tissue with pH 7.5 Tris-HCl extraction out, using primers supplied by Operon buffer (SOLTIS et al., 1983), modified Technologies Inc. (Alameda, CA), in a adding 1 mM PMSF. Crude extracts were 2400 Perkin-Elmer thermocycler, program- absorbed on Whatmann 3MM paper wicks med for 40 cycles of 30 sec at 94 °C, 40 sec and stored at -80°C. Out of 20 enzyme at 35 °C and 1 min at 72 °C. This was fol- systems tested, five were resolved by 10% lowed by 5 min at 72 °C. Amplification horizontal starch gel electrophoresis, with products were separated and visualised at the following buffer systems: Continuous 80 V for 2 h in 1,3% TBE agarose gels con- Tris-Citrate (SHAW &PRASAD, 1970) for taining 0.5 g/ml ethidium bromide. menadione reductase (MNR, EC 1.6.99.2) In a preliminary analysis 20 decamer and Leucine aminopeptidase (LAP, EC primers were screened for suitability. 3.4.11.1), Tris-Versene-Borate (BREWER, Seven (OPD-05, OPD-06, OPD-07, OPD- 1970) for Acid phosphatase (ACP, EC 08, OPD-11, OPD-12, OPD-18) were 3.1.3.2), Phosphate-Citrate (HARRIS, 1966) selected on the basis of best repeatability, for Malate dehydrogenase (MDH, EC clearness of the amplification pattern and 1.1.1.37) and Lithium Versene-Borate their ability to reveal levels of diversity in (SHAW &PRASAD, 1970) for Glutamate Adenostyles specimens. Each DNA sample oxaloacetate transaminase (GOT, EC was amplified at least twice with all pri- 2.6.1.1). We selected these enzyme systems mers. Polymorphic bands were scored as because they stained with sufficient inten- present or absent. sity and resolution to be scored with confi- dence. Enzyme activity staining was carried out according to WENDEL &WEEDEN DATA ANALYSIS (1990) and BREWER &SING (1970). Genetic interpretation of band patterns Allozymes followed standard principles (WEEDEN & The standard genetic parameters were WENDEL, 1990; WENDEL &WEEDEN, calculated using TFPGA software 1990), taking in account that the chromoso- (MILLER, 1997) to estimate allelic frequen-

106 cies, percent polymorphic loci at 95% cri- tances, was carried out with AMOVA 1.55 terion (P95), number of alleles per locus software to partition the total variance into (A), expected and observed heterozygosity covariance components, and to find their (He and Ho respectively) and departure significance levels, according to inter-indi- from Hardy-Weinberg equilibrium for vidual and inter-population differences each populations. (EXCOFFIER et al., 1992) at the level of geo- Population substructure was estimated graphic region. with Wright's F statistics (WRIGHT, 1965), Shannon Information Index (I) to esti- FIT, FIS and FST, according to WEIR & mate variability at population and species COCKERHAM (1984), using the computer level and Nei's (NEI, 1987) genetic identi- software FSTAT (GOUDET, 2000). These ties were calculated using POPGENE 32 analyses indicate reductions in heterozygo- software (YEH et al., 1997). sity expected under random mating in indi- Nei's genetic distances, calculated with viduals relative to subpopulations (FIS), or the same software for all pairwise popula- the total variation in the species (FIT), and tion comparisons, were used to construct the amount of variation distributed among the unweighted pair group method with populations (FST). Use of this method arithmetic averages (UPGMA). allowed us to test whether these F values In addition, significant differences and were significantly different from zero. correlation between the two molecular data Nei's (NEI, 1987) genetic identities, use- set markers were evaluated using the two- ful to evaluate the interpopulation differen- tailed t-test and the chi-square test and the tiation, were obtained with the same pro- relationship between genetic distance and gram. A phenogram based on Nei's genetic geographic distance was explored using distances was constructed using the the Mantel test (SOKAL &ROHLF, 1995). unweighted pair group method with arithmetic averages (UPGMA) of SNEATH & SOKAL (1973) using TFPGA software. RESULTS

RAPDs Allozymes Due to the dominant nature of RAPD From the evaluation of 12 enzymatic markers, allele frequencies must be deter- systems, 5 have furnished reliable and mined under the assumptions that only two repeatable bands. Five putative loci have alleles exist at a locus and that the popula- been identified, each with 2 alleles. tions are in Hardy-Weinberg equilibrium Monomorphic loci have not been detected, unless estimates of inbreeding are availa- although a fixed allele of MNR enzyme ble from other data. Estimates of popula- has been observed in the Sila and Alps tion differentiation GST were calculated populations. under the assumptions of population equi- The number of alleles per locus, the librium using POPGENE 32 software percentage of polymorphic loci are repor- (YEH et al., 1997). For two alleles at a ted in Tab. 2. Only three populations (PO- locus, as applicable in RAPD analysis, 3, AB-1 and AL-1) have heterozytosity GST is identical to Wright's FST. values that meet expected values under An analysis of molecular variance Hardy-Weinberg equilibrium, while two (AMOVA), using squared Euclidean dis- populations (SI-1 and SI-2) show negative

107 Tab. 2 - Summary of allozyme variation within populations of Adenostyles: percentage of polymorphic loci (P95), mean number of alleles per locus (A), observed heterozygosity (Ho), expected heterozygosity (He), standard error (SE).

Site P95 A Ho (SE) He (SE)

Pollino massif (northern Calabria) PO-1 80 1.8 0.250 (0.287) 0.329 (0.246) PO-2 100 2.0 0.274 (0.180) 0.430 (0.105) PO-3 60 1.6 0.257 (0.329) 0.276 (0.256) Population mean 80 1.8 0.259 (0.265) 0.345 (0.202)

Sila (central Calabria) SI-1 80 1.8 0.415 (0.325) 0.384 (0.219) SI-2 20 1.2 0.120 (0.268) 0.093 (0.209) Population mean 50 1.5 0.267 (0.296) 0.238 (0.214)

Central Apennines CI-1 80 1.8 0.121 (0.146) 0.279 (0.213) CI-2 80 1.8 0.254 (0.293) 0.393 (0.222) AB-1 100 2.0 0.431 (0.126) 0.458 (0.037) Population mean 87 1.9 0.269 (0.188) 0.376 (0.157)

Eastern Alps AL-1 80 1.8 0.355 (0.342) 0.397 (0.225) AL-2 60 1.6 0.200 (0.274) 0.257 (0.235) Population mean 70 1.7 0.277 (0.308) 0.327 (0.230) fixation indexes, indicative of an excess of the different loci, with all values always heterozygotes. significantly different from zero Wright F statistics are variable among (p<0.001). The partitioning of variability

a b

Fig. 2 - Allozyme (a) and RAPD (b) UPGMA dendrogram of Nei's genetic distance between populations of Adenostyles

108 within and among populations was also (Fst) at geographic level is higher (0.338) shown to be significantly different from than that of narrowly distributed species zero (p<0.001). The mean Fst for polymor- with a mixed or outcrossing breeding phic loci is 0.082. system (0.2-0.28). All the pairwise values The Nei's (NEI, 1972; 1987) genetic dis- between populations are significantly dif- tances (ranging from 0.04 to 0.38) have ferent from zero, aside from AB-1 vs CI-1 been used to elaborate the similarity den- (P<0.05) and PO-1 vs AB-1, which are drogram (Fig. 2a), which does not indicate negative. geographic structuring among the popula- Nei's genetic diversity values of populations. tion and geographic levels are indicative of a good variability, in accord to those obtai- RAPDs ned with various Nei's formulations in The seven primers used in the study multilocus estimates, which may be biased allowed the scoring of a total of 45 loci, by the dominant character of the RAPDs with an average of 6.4 loci per primer. marker (NYBOM &BARTISH, 2000). Number of bands per primer ranges from 5 The RAPD dendrogram (Fig. 2b) does (OPD-6, OPD-18) to 11 (OPD-12) and not resolve populations and geographic their length from 370 to 1050 bp. We found groups, indicating that intrapopulation a high amount of RAPD variation in the polymorphisms exceeds interpopulation examined populations of Adenostyles, ones. Corroborating this, the Mantel's test revealed by the unique RAPD multiband indicated no correlation between genetic phenotype. distance and geographic distance among Nested AMOVA evaluations at geogra- samples. Otherwise, the Shannon Index phic levels revealed a significant (0.55), which is less influenced by the (P<0.001) genetic diversity within dominance of the markers, is indicative of similar variability patterns. Tab. 3- Summary of analysis of molecular variance (AMOVA) The Gst value (0.40) is higher than that relative to perennial herbaceous plant spe- Variance component cies with a narrow (0.22) and regional Level of variation Df Absolute %P(0.35) distribution (NYBOM &BARTISH, Among geographic 4 -0.283 -3.62 0.7353 2000). However, the correlation between regions the RAPD-based Gst and the allozyme- Among populations 5 2.864 36.60 <0.001 derived Fst, is not significant (paired T-test within geographic = - 1.104; 0.1

109 evident pattern of spatial genetic structure lity in Adenostyles. Indeed, the relative has been detected and the geographic dendrograms point out that populations groups are scarcely divergent (HAMRICK & group without congruence with their regio- GODT, 1989; NYBOM &BARTISH, 2000). nal location (Fig 2a, b). A lack of con- The Apennine populations of Adeno- gruence between genetic differentiaton styles used in the present study show a among populations, which is low, and geo- higher allozyme diversity than that usually graphic pattern of population groups, has found in rare species or in isolated popula- been reported for Saxifraga oppositifolia, tions at the border of the distribution area an alpine taxon living in habitats with a (ESSELMAN et al., 2000; GITZENDANNER & small-scale heterogeneity. This finding has SOLTIS, 2000). At the same time, allozymic been attributed to a rare gene flow among genotypes of the examined populations of relatively distant examined populations Adenostyles appear to be rather homoge- (GUGERLI et al., 1999). neous. They differ only for the relative fre- In this respect, the high genetic diver- quencies of common alleles and, excluding sity and variation found among the the unique allele of MNR fixed in the indi- Adenostyles populations, does not suggest viduals collected from Sila mountains and that gene flow is still occurring, but rather eastern Alps, suggest that all populations points out a common genetic exchange are part of a larger genetic pool. only (if at all) recently interrupted The same populations exhibit greater (SCHAAL et al., 1998). The discrepancy in variation at RAPD loci than at allozyme levels and structuring of genetic diversity, loci. RAPD variation is most apparent in as measured by allozymes and RAPD, different band frequencies among popula- undoubtedly, is the result of recent and tions, and every individuals exhibited a historical genetic and ecological factors. unique multilocus phenotype. Usually, Moreover, the lack of molecular gradients, PCR-based DNA markers reveal equal or being not correlated to geographic distri- higher levels of genetic variability (WOLFE bution, tends to exclude a history of migra- &MORGAN-RICHARDS, 1998; ESSELMAN et tion and recolonization of the Apennine al., 2000). Greater diversity in DNA mar- chain. Additionally, other episodes, like kers such as RAPD is attributed to the fact long distance dispersal or founder events, that many loci are amplified, providing a that usually imply a reduction of the gene- genome-wide survey (WILLIAMS et al., tic variability may account for this pheno- 1990). In contrast to allozyme loci, non- menon (COMES &KADEREIT, 1998; coding region, such as those that may be THOMPSON, 1999). The UPGMA pheno- amplified with random primers, are unli- gram does not show a pattern of sequential kely to be under heavy selective con- relationships among populations from straints. Consequently, allozymes are con- north to south, so excluding a scenario of sidered more sensitive to selective pressu- unidirectional range expansion, but rather res, while the RAPD variability depends display a distribution of Adenostyles primarily on evolutionary mechanisms, occurring via recent habitat fragmentation like mutations and drift (SUN et al., 1999). of a single and wide-spread ancestral entity However, in this case both allozyme distributed along whole Apennine chain. A and RAPD data have failed to find clear plausible explanation could be that during geographic patterns of the genetic variabi- the Quaternary Adenostyles, being part of

110 the vegetation that grew at the boundaries nomic proposal. It is more probable that of glacial zones along Apennine chain, these taxa have experienced some situa- finally managed to establish contact with tions of parapatry and/or hybridization lea- the wet habitats where they are at presently ding to genetic as well as to morphological largely located (WILLIS, 1996). In this homogenization. manner, these taxa underwent periodic Finally, one of the purposes of this work expansion and retreats of their vegetational was to exploit molecular data in a biologi- areas with connected episodes of secon- cal conservation framework, an appropria- dary contacts, which could explain their te objective if one considers the low consi- present genetic similarity and the absence stency and the disjunction of the examined of any geographic gradient within populations (BARRET &KHON, 1991). Apennine populations (Fig. 1) (THOMPSON, Units appropriate for conservation and 1999). As a consequence, in the postglacial management should be established on the period these populations became progres- basis of their evolutionary potential. In this sively isolated with the appearance of regard, variation at allozyme and RAPD dryer conditions. In these isolated popula- loci in Apennine populations of tions, the alteration of their pollinator set Adenostyles complex may be interpreted as (CONTI et al., 2000) and/or the insurgence an indication of genetic potential, showing of autogamy (CRAWLEY, 1977) may have that Apennines populations of Adenostyles caused the little morphological diversifica- complex still have the possibility to persist tion (i.e number of florets, WAGENITZ, and sustain future evolutionary change. 1983). Such a scenario appears to be compati- Acknowledgements This paper was sup- ble with the aforementioned difficulty in ported from a MURST PRIN (2000) grant. discriminating taxa of Adenostyles, independently from each of the particular taxo-

LITERATURE CITED AYRES D. R. & RYAN F. J. 1999. Genetic No genetic variation detected within isolat- diversity and structure of the narrow ed relict populations of Saxifraga cernua endemic Wyethia reticulata and its con- in the Alps using RAPD markers. gener W. bolanderi (Asteraceae) using Molecular Ecology, 7: 1519-1527. RAPD and allozyme techniques. American BREWER G. J. & SING C. F. 1970. An Journal of Botany, 86 (3): 344-353. introduction to isozyme techniques. BARRET S. C. & KHON J. R. 1991. Academic Press, New York & London. Genetic and evolutionary consequences of COMES H. P. & KADEREIT J. W. 1998. small population size in plants: implica- The effects of Quaternary climatic changes tions for conservation. In: Falk D.A., on plant distribution and evolution. Trends Holsinger K.E. (Eds.). Genetics and con- in Plant Science, 3 (11): 432-438. servation of rare plants. Pag. 3-30. Oxford CONTI E.,SURING E.,BOYD D., University Press, New York. JORGENSEN J., GRANT J. & KELSO S. 2000. BAUERT M.R.,KAELIN M., Phylogenetic relationships and character BALTISBERGER M. & EDWARDS P. J. 1998. evolution in Primula L.: the usefulness of

111 ITS sequence data. Plant Biosystems, 134 FRIZZI G.&TAMMARO F. 1991. (3): 385-392. Electrophoretic study and genetic affinity CORBETTA F., ABBATE G., FRATTAROLI A. in the Campanula elatines and C. fragilis R. & PIRONE G. F. 1998. S.O.S. VERDE! (Campanulaceae) rock-plants group from Vegetazioni e specie da conservare. Italy and W. Jugoslavia. Plant Systematic Edagricole, Bologna. and Evolution, 174: 67-73. CRAWLEY M. J. 1977. Sex - Plant ecolo- GITZENDANNER M. A. & SOLTIS P. S. gy. Pag. 156-213. Blackwell Science, 2000. Patterns of genetic variation in rare Oxford. and widespread plant congeners. American DOYLE J. J. & DOYLE J. L. 1987. A rapid Journal of Botany, 87 (6): 783-792. DNA isolation procedure for small quanti- GOUDET J. 2000. FSTAT, a program to ties of fresh leaf tissue. Phytochemistry estimate and test gene diversities and fixa- Bullettin, 19 (1): 11-15. tion indices (version 2.9.1). EDWARDS K. J. 1998. Randomly ampli- GRÜNANGER P., CAPORALI E., MARZIANI fied polymorphic DNAs (RAPDs). In: G., MENGUZZATO E. & SERVETTAZ O. 1998. Karp A., Isaac P.G., Ingram D.S., Molecular (RAPD) analysis on Italian taxa Molecular tools for screening biodiversity. of the Ophrys bertolonii aggregate (Eds.). Plants and animals. Chapman & (Orchidaceae). Plant Systematic and Hall London. Evolution, 212: 177-184. ESSELMANN E.J.,CRAWFORD D.J., GUGERLI F., EICHENBERGER K.& BRAUNER S., STUESSY T. F., ANDERSON G. SCHNELLER J. J. 1999. Promiscuity in pop- J. & SILVA O. M. 2000. RAPD markers ulations of the cushion plant Saxifraga diversity within divergence among species oppositifolia in the Swiss Alps as inferred of Dendroseris (Asteraceae: from random amplified polymorphic DNA Lactucaceaea). American Journal of (RAPD). Molecular Ecology, 8: 453-461. Botany, 87 (4): 591-596. HAMRICK J. L. & GODT M. J. W. 1989. EXCOFFIER L., SMOUSE P. E. & QUATTRO Allozyme diversity in plant species. In: J. M. 1992. Analysis of molecular variance Brown A. H. D., Clegg M. T., Kahler A. L., inferred from metric distances among Weir B. S. (Eds.). Plant population genet- DNA haplotypes: application to human ics, breeding and genetic resources. Pag. mitochondrial DNA restriction data. 43-63. Sinauer, Sunderland, Mas- Genetics, 131: 479-491. sachusetts, USA. FEDEROV F. 1974. Chromosome num- HANNAN G. L. & ORICK M. W. 2000. bers of flowering plants. Otto Koeltz Isozyme diversity in Iris cristata and the Science Publishers. threatened glacial endemic I. lacustris FIORI A. 1925-29. Flora Analitica (Iridaceae). American Journal of Botany, d’Italia. Vol 2: 582-583. Edagricole, 87 (3): 293-301. Bologna. HARRIS H. 1966. Enzyme polymor- FREVILLE H., JUSTY F. & OLIVIERI I. phism in man. Proceedings of the Royal 2001. Comparative allozyme and Society, London B, 169: 298-310. microsatellite population structure in a nar- HUENNEKE L. F. 1991. Ecological impli- row endemic plant species, Centaurea cations of genetic variation in plant popu- corymbosa Pourret (Asteraceae). lations. In: D. A. Falk & K. E. Holsinger. Molecular Ecology, 10: 879-889. (Eds.). Genetics and Conservation of Rare

112 Plants. Pag. 31-44. Oxford University SHAW C. R. & PRASAD R. 1970. Starch Press, New York. gel electrophoresis of enzymes - a compi- LUTZ E.,SCHNELLER J.J.& lation of recipes. Biochemical Genetics, 4: HOLDEREGGER R. 2000. Understanding 297-320. population history for conservation pur- SNEATH P. H. A. & SOKAL R. R. 1973. poses: population genetics of Saxifraga Numerical taxonomy. Freeman W. H. San aizoides (Saxifragaceae) in the lowlands Francisco. and lower mountains north of the Alps. SOKAL R.R.&ROHLF F. J. 1995. American Journal of Botany, 87 (4): 583- Biometry. W. H. Freeman, New York. 590. SOLTIS D. E., HAUFLER C. H., DARROW MILLER M. P. 1997. Tools for popula- D.C. & GASTONY G. J. 1983. Starch gel tion genetic analyses (TFPGA) 1.3: A win- electrophoresis of ferns: a compilation of dows program for the analysis of allozyme grinding buffers, gel and electrode buffers and molecular population genetic data. and staining schedules. American Fern Computer software distributed by author. Journal, 73 (1): 9-27. NEI M. 1972. Genetic distance between STEHLIK I. 2000. Nunataks and periph- populations. American Naturalist, 106: eral refugia for alpine plants during quater- 283–292. nary glaciation in the middle part of the NEI M. 1987. Molecular evolutionary Alps. Botanica Helvetica, 110: 25-30. genetics. Columbia University Press, New STEHLIK I.,SCHNELLER J.J.& York. BECHMANN K. 2001. Resistance or emigra- NYBOM H.&BARTISH I. V. 2000. tion: response of the high-alpine plant Effects of life history traits and sampling Eritrichium nanum (L.) Gaudin to the ice strategies on genetic diversity estimates age within the Central Alps. Molecular obtaneid with RAPD markers in plants. Ecology, 10: 357-370. Perspectives in Plant Ecology, Evolution SUN G.L.,SALOMON B.& VON and Systematic, 3/2: 93-114. BOTHMER R. 1999. Genetic diversity in PASSALACQUA N. G. & BERNARDO L. Elymus caninus as revealed by isozyme, 1997. Flora relitta di altitudine RAPD and microsatellites markers. dell’Appennino meridionale: quale ori- Genome, 42: 1-12. gine? Biogeographia, 19: 105-117. THOMPSON J. D. 1999. Population dif- PIGNATTI E., PIGNATTI S., NIMIS P. & ferentation in Mediterranean plants: AVANZINI A. 1980. La vegetazione ad insights into colonization history and evo- arbusti spinosi emisferici. Contributo alla lution and conservation of endemic interpretazione delle fasce di vegetazione species. Heredity, 82: 229-236. delle alte montagne dell’Italia mediter- TUTIN T. G. 1976. Adenostyles Cass. In: ranea. C.N.R. AQ/1/79, Roma. Tutin, T. G., Heywood, V. H., Burges, N. PIGNATTI S. 1982. Flora d’Italia. Vol. 3: A., Moore, D. M., Valentine, D. H., 14-16. Edagricole, Bologna. Walters, S. M. & Webb, D. A. (Eds.). Flora SCHAAL B. A., HAYWORTH D. A., OLSEN Europea. Pag 189. Cambridge University K. M., RAUSCER J. T. & SMITH W. A. 1998. Press, Cambridge. Phylogeographic studies in plants: prob- VRIJENHOEK R. C. 1987. Population lems and perspectives. Molecular Ecology, genetics and conservation. In: D. Western 7: 465-474. & M. C. Pearl. (Eds.). Conservation for the

113 21st Century. Pag. 89-98. Oxford markers. Nucleic Acids Research, 18: University Press, New York. 6531-6535. WAGENITZ G. 1983. Die Gattung WILLIS K. J. 1996. Where did all the Adenostyles Cass. (Compositae- flowers go? The fate of temperate Senecioneae). Phyton (Austria), 23 (1): European flora during glacial periods. 141-159. Endeavour, 20 (3): 110-114. WEEDEN N. F. & WENDEL J. F. 1990. WOLFE K. & MORGAN-RICHARDS M. Genetics of plant isozymes. In: Soltis D. 1998. PCR markers distinguish Plantago E., Soltis P. S. (Eds.). Isozymes in plant major subspecies. Theoretical and Applied biology. Dudley T. R., Ph. D., General Genetics, 96: 282-286. Editor. Pag. 46-72. Chapman and Hall, WRIGHT G. 1965. The interpretation of London. population structure by F-statistics with WEIR B. & COCKERHAM C. C. 1984. special regard to systems of mating. Estimating F-statistics for the analysis of Evolution, 19: 395-420. population structure. Evolution, 38: 1358- YEH F. C., YANG R. C., BOYLE T. B. J., 1370. YE Z. H. & MAO J. X. 1997. POPGENE, WENDEL J. F. & WEEDEN N. F. 1990. the user-friendly shareware for population Visualization and interpretation of plant genetic analysis. Molecular Biology and isozymes. In: Soltis D. E., Soltis P. S. Biotechnology Centre, University of (Eds.). Isozymes in plant biology. Dudley Alberta, Edmonton, Alberta, Canada. T. R., Ph. D., General Editor. Pag. 5-45. ZHANG L., COMES H. P. & KADEREIT J. Chapman and Hall, London. 2001. Phylogeny and quaternary history of WILLIAMS J. G. K., KUBELIK A. R., the European montane/alpine endemic LIVAK K. J., RAFALSKI J. A. & TINGEY S. V. Soldanella (Primulaceae) based on ITS 1990. DNA polymorphisms amplified by and AFLP variation. American Journal of arbitrary primers are useful as genetic Botany, 88 (12): 2331-2345.

Finito di stampare nell’ottobre 2003

114