| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2 Botanical Garden and Museum, University of Oslo, Trondheimsveien 23 B, N-0562 Oslo, Norway; and 3 The University Courses on Svalbard, P. O. Box 156, N-9170 Longyearbyen, Svalbard, Norway
Received for publication July 1, 1999. Accepted for publication December 16, 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Arctic • genotypic diversity • morphology • Potentilla • RAPDs • Rosaceae • taxonomy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, and their "the more the better" principle)? Multivariate analyses have traditionally been applied to determine whether the variation in "overall" morphology is discontinuous, but the inclusion of characters that vary at random across taxa will necessarily tend to conceal the information provided by taxonomically significant characters. An additional problem is that the genetic component of the morphological variation observed in field-collected material is unknown.
Use of molecular markers that regularly identify many genetic polymorphisms at low taxonomic levels, such as random amplified polymorphic DNAs (RAPDs; Williams et al., 1990
), provides a possibility to address the relationship between variation in individual morphological characters and genotypic variation. Some tens of randomly selected markers, derived from all over the target genomes, may provide a sufficient basis for grouping of populations. After an initial molecular analysis, morphological characters can be selected and analyzed with the specific aim to test the hypothesis that the groups of multilocus genotypes observed are morphologically distinguishable and should be considered different taxa. Alternatively, hypotheses forwarded on the basis of variation in specified morphological characters can be tested with molecular data.
In this study, we address the usefulness of a combined molecular and morphological approach to test some previously proposed hypotheses on taxonomic delimitations and relationships in Potentilla, a genus well known for its taxonomic complexity in the Arctic (e.g., Hultén, 1945
; Hiitonen, 1949
; Rønning, 1961
; Soják, 1985, 1986, 1989
; Elven and Elvebakk, 1996
; Eriksen, 1996, 1997
). Extensive reticulate evolution via hybridization and polyploidy, combined with facultative, pseudogamous agamospermy, have probably caused many of the taxonomic problems in this genus (Müntzing, 1928
; Smith, 1963
; Acharaya Goswami and Matfield, 1974
; Soják 1985, 1986
; Eriksen, 1996
).
A total of 11 taxa from three sections of Potentilla have been reported throughout the years from the arctic Norwegian archipelago of Svalbard, and six species were accepted in a recent checklist by Elven and Elvebakk (1996)
. Their tentative conclusions regarding four of these species were further examined in the present study. They hypothesized that the P. nivea L. complex consists of three species in Svalbard, viz. P. chamissonis Hultén, P. insularis Soják, and P. nivea L. ssp. subquinata (Lange) Hultén. Within P. pulchella R. Br., three "eco-morphotypes" were described (Elven and Elvebakk, 1996
), and herein we test whether these "eco-morphotypes" should be considered different intraspecific taxa. Because of the long-term confusion regarding the taxonomy of Potentilla in Svalbard, we found that the background for forwarding these two hypotheses and a third hypothesis, concerning hybrid origin of P. insularis, needed to be further clarified. Below, we therefore present an updated survey of the relevant taxonomic literature as well as results from an initial re-examination of herbarium material of Potentilla from Svalbard.
The Potentilla nivea complex and P. pulchella in Svalbard
The taxonomy of the Potentilla nivea complex has been controversial. A proposal for a solution to the long-lasting controversy regarding nomenclature and typification in this complex was recently presented by Eriksen, Jonsell, and Nilsson (1999)
, and these issues will not be further addressed here.
Hultén (1945)
determined the distribution of P. nivea ssp. subquinata to arctic America, Greenland, Svalbard, and northern Fennoscandia. He also described a new species, P. chamissonis, and reported it to have an amphi-Atlantic distribution with its main concentration on the western side (for further details, see Goworuchin, 1932
; Juzepchuk, 1941
; Hiitonen, 1949
; Jurtzev, 1984
). Potentilla rubricaulis Lehm. of the P. nivea complex was described from Canada by Lehmann (1830), and it was reported from Svalbard by Rønning (1961
; repeated by Ball, Pawlowski, and Walters, 1968
, and Hultén and Fries, 1986
). Potentilla pedersenii (Rydb.) Ostenf. was treated as a synonym for P. rubricaulis by Hultén (1945)
, whereas Soják (1986)
recognized P. pedersenii as a distinct species distributed along the arctic coasts of Canada and in Greenland, but not in Svalbard. At the same time, Soják (1986)
described a new species from Svalbard, P. insularis. This species was typified on a specimen collected at Hyperitthatten (28 August 1908, leg. H. Resvoll-Dieset, O!) and did in reality replace P. rubricaulis and P. pedersenii in a Svalbard context.
Elven and Elvebakk (1996)
questioned a report of P. lyngei Jurtz. and Soják from Svalbard, which was based on a single specimen collected at Gipshuken. We re-examined this specimen (19 July 1908, leg. H. Resvoll-Dieset, O!), which was determined by Soják, and found it to be a mixture of two taxa (stems from P. insularis and leaves from P. pulchella; K. T. Hansen and R. Elven, unpublished data). Thus, convincing evidence for the presence of P. lyngei in Svalbard is lacking, and this particular problem is not addressed further in this study.
The taxonomy of P. pulchella has also been controversial. Brown (1824) described P. pulchella based on material from Melville Island in arctic Canada. Sommerfelt (1833) described P. keilhaui Sommerf. from Svalbard, but later it was shown to be conspecific with P. pulchella (see Elven and Elvebakk, 1996
). Elven (1994)
and Elven and Elvebakk (1996)
reported that conspicuous morphological variation within P. pulchella had caused some of the taxonomic confusion in Potentilla in Svalbard. Some large plants of P. pulchella growing in bird cliffs were, for example, often referred to the P. nivea complex. In addition, Potentilla multifida L. was reported from Svalbard by Nathorst (1883) based on plants collected at Kapp Thordsen, Dickson Land, and other authors accepted this taxon for Svalbard (Resvoll-Holmsen, 1927
; Ball, Pawlowski, and Walters, 1968
; Hultén and Fries, 1986
). However, Rønning (1961)
considered these plants to represent a form of P. pulchella. We re-examined the material from the area for the present study and concluded that the plants previously referred to P. multifida represent part of the variation within P. pulchella, in agreement with Rønning (1961)
. Potentilla multifida is, therefore, considered to be absent from the flora of Svalbard and is not further addressed.
Presentation of the taxa and the hybrid hypothesis for P. insularis
The taxa and names used by Hultén (1945)
and Elven and Elvebakk (1996)
for the Svalbard material of Potentilla were tentatively accepted as an initial framework for the present study. Because our final conclusions also were in agreement with this treatment, the names P. chamissonis, P. insularis, P. nivea, and P. pulchella will be used in the following.
All taxa of the P. nivea complex occur in south-facing scree slopes and in manured bird cliff meadows in Svalbard. According to Elven and Elvebakk (1996)
, the taxa are most easily distinguished by petiole indument (curly hairs in P. nivea vs. straight hairs in P. chamissonis and P. insularis) and number of leaflets per leaf (many in P. insularis, few in P. nivea and P. chamissonis). No data are available on the reproductive biology of the Svalbard populations.
Potentilla nivea has a circumpolar distribution and comprises a variety of different races that are more or less separated morphologically and geographically, and P. nivea ssp. subquinata has been reported as amphi-Atlantic (e.g., Hultén, 1945
; Soják, 1986
). Potentilla nivea varies from low-ploid (diploid?) to decaploid (2n = 14?, Krogulevich, 1978
; 2n = 28?, Belaeva and Siplivinsky, 1976
; Zhukova and Petrovsky, 1976
; 2n = 49, Belaeva and Siplivinsky, 1976
; 2n = 54–56, Engelskjøn, 1979
; 2n = 56, Dansereau and Steiner, 1956
, Knaben and Engelskjøn, 1967
; 2n = 63, Böcher and Larsen, 1950
; Engelskjøn, 1979
; 2n = 70, Zhukova and Petrovsky, 1980
). Chromosome numbers have not yet been reported for the Svalbard populations of the species, but the mainland Norwegian populations are high-ploid (2n = 54–63; Knaben and Engelskjøn, 1967
; Engelskjøn, 1979
).
Potentilla chamissonis has been reported to have a western amphi-Atlantic distribution (i.e., northeastern America, Greenland, Svalbard, and Fennoscandia; Hultén, 1945
; Hultén and Fries, 1986
). The chromosome number has not been determined in Norwegian populations, but the species varies from heptaploid to possibly 11-ploid in other geographic areas (2n = 49, Dansereau and Steiner, 1956
; 2n = 56, Böcher and Larsen, 1950
; 2n = 77?, Müntzing in Hultén, 1945
).
When Potentilla insularis was described from Svalbard, it was also reported from eastern Greenland by Soják (1986)
. He suggested that this species, which has an unknown chromosome number, originated as a hybrid between P. lyngei (sect. Multifidae Rydb.) and P. chamissonis (sect. Niveae Rydb.). His main argument was that plants with digitate or somewhat pinnate leaves probably originated as hybrids between sections Multifidae (pinnate leaves) and Niveae (ternate leaves). He also argued that hybrid taxa that exclusively have straight hairs on the petiole, such as P. insularis, must have been derived from taxa in section Niveae that exclusively have straight hairs. He therefore concluded that P. chamissonis must be one of the progenitors of P. insularis. Soják (1986)
also believed that P. insularis had originated independently from the same parental species in Greenland and Svalbard. However, we showed above that P. lyngei does not occur in Svalbard, and another taxon of section Multifidae must therefore be considered as a possible progenitor of P. insularis. Potentilla pulchella is the only one present in Svalbard, and in the present study, we therefore forward the following modification of Soják's (1986)
original hybrid hypothesis: P. insularis originated as a hybrid between P. pulchella as one parent and one of the taxa in the P. nivea complex as the other parent.
Potentilla pulchella has a circumpolar, but scattered distribution with large gaps in arctic Asia (e.g., Porsild and Cody, 1980
; Hultén and Fries, 1986
). It is consistently tetraploid with 2n = 28, also in Svalbard (Flovik, 1940
; Holmen, 1952
; Dansereau and Steiner, 1956
; Jørgensen, Sørensen, and Westergaard, 1958
). The three "eco-morphotypes" in P. pulchella are (1) the "normal morphotype," consisting of large, hairy plants growing in cliffs, in manured bird cliff meadows, on ridges, and in scree slopes; (2) the "Sassen morphotype," consisting of small, glabrescent plants growing on gravel shore terraces; and (3) the "beach morphotype," consisting of small, hairy plants growing on silt shore terraces (Elven and Elvebakk, 1996
).
Synopsis of the aims of this study
In particular, we wanted to explore the usefulness of a detailed, combined molecular (RAPDs) and morphological approach for unraveling complex variation at low taxonomic levels and to use this approach to test the following hypotheses forwarded for some arctic members of Potentilla: (1) the P. nivea complex consists of three taxa in Svalbard (P. chamissonis, P. insularis, and P. nivea; Elven and Elvebakk, 1996
), (2) the three "eco-morphotypes" within P. pulchella described by Elven and Elvebakk (1996)
are genetically differentiated and should be considered different intraspecific taxa, and (3) P. insularis originated as a hybrid between P. pulchella and one of the taxa in the P. nivea complex (modified from Soják, 1986
, cf. above).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
; Smith, 1963
; Acharaya Goswami and Matfield, 1974
; Asker, 1977
; Asker and Jerling, 1992
; Eriksen, 1996
).
RAPD analysis
DNA was isolated from 10 to 30 mg of silica-dried, field-collected leaves or from 50 to 100 mg of fresh leaves from cultivated plants using the CTAB method as described by Gabrielsen et al. (1997)
, except that a purifying step with RNAse (10 µg/mL DNA extract) was added. For DNA quantification, 5 µL of the DNA stock and 1 µL 6x loading buffer were run on a 0.7% agarose gel stained with ethidium bromide. The intensity of the DNA bands was compared with a known concentration of a 
marker cut with EcoR1 and HindIII. The RAPD-PCR (polymerase chain reaction) protocol followed Gabrielsen et al. (1997)
. One nanogram template DNA was used in each PCR reaction, determined after an initial concentration test. The PCR products were separated on 1.4% agarose gels and visualized by ethidium bromide staining.
A total of 60 primers from kits A, C, and D (Operon Technologies, Alameda, California, USA) were used for preanalyzing RAPD variation among four plants, one of each tentative taxon. Thirty-three primers that produced smeary bands or no bands at all were excluded. The remaining 27 primers were further tested using 12 plants. Twelve primers (A01, A04, A05, A10, A14, A15, C01, C02, C08, C13, C18, and D08) that produced the most distinct, reproducible, and polymorphic bands were selected for full RAPD analysis of all 136 plants. The gels were scored conservatively, i.e., only the most reliable bands were scored (as 1, present, or 0, absent). The reproducibility of all initially scored bands was rechecked by comparing the profiles of a number of individual plants that were run 2–4 times: (1) in the first primer test, (2) in the second primer test, (3) in the main analysis, and/or (4) in final reruns to verify marker alignment across gels.
Morphological analysis
We initially explored the variation in a number of characters that previously had been used to discriminate among the hypothesized taxa (e.g., Hultén, 1945
; Eriksen, 1997
) and other characters that potentially could discriminate among the groups of multilocus phenotypes identified in the analyses of the RAPD data. A final set of 48 primary morphological characters was selected for full analysis. In addition, 16 derived characters were computed based on these primary characters, resulting in a total set of 64 morphological characters (see Table 4). The hair density on the petiole was recorded separately for two types of hairs, curly and straight (cf. Hultén, 1945
). The morphological analysis was mainly performed on field-collected material. However, because some plants did not flower at the time of collection, but later flowered in the phytotron, it was necessary to include some cultivated material for measurement of floral characters. To test for possible differences in floral characters measured in cultivated vs. field-collected material, both types of material were measured for 12 plants. Four characters (petal insertion depth, stamen length, and anther length and width) were excluded after this test; for other floral characters, the two types of material gave consistent results.
|
), resulting in a set of 37 characters (see Table 4). The morphometric data and the RAPD data were subjected to principal coordinate analyses (PCO; Gower, 1966, 1967
), minimum spanning tree analyses (MST), and UPGMA cluster analyses using NTSYS-pc (Rohlf, 1998
). The morphometric matrix was first standardized by ranging, i.e., the variation in each character was scaled between 0 and 1. A distance matrix was computed using average Manhattan distance (city block), given by 1/n
k |xki - xkj|, where n is the number of plants, k is the character, and i and j are a pair of plants. Spearman's rank correlation coefficients (r) were calculated between all morphological characters and between the PCO axes and the morphological characters using SPSS for Windows (Noru
is, 1993
). The RAPD matrix was analyzed using three different similarity or distance coefficients—Dice, Simple matching, and Euclidean distance (Rohlf, 1998
). These coefficients gave similar results, and only the analyses based on Dice's similarity are shown. This coefficient is defined as 2a/(2a + b + c), where a is the number of shared bands, and b and c are the number of bands present in one sample but absent in the other sample. Thus, this coefficient does not take shared absence of bands into account.
Analyses of Molecular Variance (AMOVA; Excoffier, Smouse, and Quattro, 1992
) were carried out on the RAPD data using the WINAMOVA 1.55 program provided by L. Excoffier (http://anthropologie.unige.ch/ftp/comp). Amatrix of Euclidean squared distances between RAPD phenotypes was used, and the AMOVA variance components were used as estimates of molecular diversity at each hierarchic level. The variance among species, the variance among populations within species, and the variance among individuals within populations were calculated from the complete RAPD matrix. The variance among populations and the variance among individuals within populations were also calculated in separate AMOVA analyses for each taxon. Significance levels of the variances were obtained with tests including 100 permutations for each analysis. The AMOVA analyses were carried out on five plants from each population, which were selected at random from larger populations.
Three estimates of intrapopulational genotypic diversity were calculated following Ellstrand and Roose (1987)
. The proportion of distinguishable genets was calculated as G/N, where G is the number of RAPD phenotypes and N is the number of plants analyzed. Genotypic diversity (D), the complement of the Simpson index corrected for finite sample size (Pielou, 1969
), was calculated as D = 1 - {[ ni(ni - 1)]/[N(N - 1)]}, where ni is the number of plants with RAPD phenotype i, and N is the number of plants analyzed. D equals 1 if every plant in a population represents a distinct RAPD phenotype, and D equals 0 if all plants in a population have identical RAPD phenotypes. An estimate of genotypic evenness (E; Fager, 1972
) was calculated as E = (Dobs - Dmin)/(Dmax - Dmin), where Dmin = [(G - 1)(2N - G)]/[N(N - 1)] and Dmax = [N(G - 1)]/[G(N - 1)], where Dobs is the complement of Simpson's index given above, G is the number of RAPD phenotypes, and N is the number of plants analyzed. The higher the value of E obtained, the more evenly the RAPD phenotypes are distributed in the population.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
In a PCO analysis of the total RAPD data set (Fig. 3), the three first axes extracted 85.1% of the variation. The first axis (54.1%) clearly separated P. pulchella from the P. nivea complex. The second axis (22.8%) separated P. nivea from P. insularis and P. chamissonis. In an attempt to further clarify the relationships within the P. nivea complex, a separate PCO analysis was performed on a reduced RAPD data set, from which P. pulchella was excluded (Fig. 4). In this analysis, P. nivea was clearly separated from P. chamissonis and P. insularis along axis 1 (54.4%), and P. chamissonis was fairly well separated from P. insularis along axis 2 (18.5%) and axis 3 (7.9%; not shown). The RAPD phenotype observed in four of the plants from the mixed population was intermediate between P. insularis and P. nivea, and the other phenotypes of this population grouped within P. insularis or P. nivea. The minimum spanning tree that was superimposed on this PCO analysis connected P. nivea to P. insularis via the phenotype from the mixed population, and P. insularis to P. chamissonis.
|
|
|
Morphological analysis
Several individual morphological characters discriminated fairly well among the four groups of RAPD phenotypes that were identified in the analyses of the molecular data, revealing that these RAPD groups corresponded to the four hypothesized taxa (Fig. 2, Table 4). The variation in some of these morphological characters is illustrated in Figs. 2 and 5–6. Although there was a more or less continuous variation in the total material in many individual quantitative characters, the combination of such characters with some qualitative characters allowed for unambiguous classification of each plant into the group that corresponded with its RAPD phenotype.
Qualitative style characters and other floral characters discriminated P. pulchella most clearly from the other taxa (Figs. 2, 5). Potentilla pulchella invariably had a distinctly bottle-shaped style with papillae at the base, whereas the members of the P. nivea complex invariably had tube-shaped styles with nonpapillous bases. Potentilla pulchella was also distinguished from the P. nivea complex by the distance from the end leaflet to the first leaflet pair, petal length, petal width, petal length:width, sepal length, and episepal length (characters 5, 29, 30, 31, 60, 33, and 35, respectively; Table 4). Potentilla nivea was the only taxon observed with curly hairs on the petiole, and it had a lower end leaflet lobe length:width ratio (character 54) and a lower petal length:width ratio (character 60) than P. insularis and P. chamissonis (Figs. 2, 6, Table 4). Potentilla insularis was most easily distinguished from P. chamissonis by its higher number of leaflets per leaf (Fig. 2), by its higher number of lobes on the end leaflet, by its higher first leaflet length:width ratio, by its somewhat higher end leaflet length:width ratio, and by its somewhat higher lobe length:width ratio on the end leaflet (characters 4, 9, 55, 51, and 54, respectively; Table 4).
The PCO and UPGMA analyses of the morphological data revealed a very similar structure in the data set, and only the PCO analyses are therefore reported (Figs. 7–8). In a PCO analysis of the total morphological data set (Fig. 7), the first three PCO axes extracted 64.7% of the variation. The overall variation pattern was more or less continuous in this analysis, but the hypothesized taxa corresponded roughly to different parts of this variation pattern. Potentilla pulchella was most distinct; axis 1 (39.7%) separated most plants of P. pulchella from the taxa of the P. nivea complex. Along axis 2 (17.9%), P. insularis was fairly well separated from P. nivea, but these two taxa were connected via P. chamissonis.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Further analyses to determine whether these multilocus phenotypes could be classified into groups were performed without any a priori assumption of their taxonomic identity (except, of course, that all plants had to be recognized in the field as belonging to the group under study). The taxonomic hypothesis on the number of taxa in the Potentilla nivea complex was tested by subsequently determining whether it was possible to identify any morphological characters that discriminated between the groups identified in the molecular analysis, and, in this case, where a taxonomic hypothesis already had been proposed, to determine whether the molecular groups correspond to the previously hypothesized taxa. In Potentilla pulchella, on the other hand, no groups of RAPD phenotypes were observed in spite of its conspicuous morphological variation, thus providing support for the alternative hypotheses that the variation in this species is caused by plastic responses to different local environments or by taxonomically insignificant genetic variation at a small number of loci.
The more or less continuous variation revealed in the multivariate analyses of the morphological characters (Figs. 7, 8) also reflects a commonly encountered problem when attempting to interpret so-called "overall morphological variation" (cf. Sneath and Sokal, 1973
). In spite of this continuous, "overall" variation, we identified several individual morphological characters that could be used to classify each plant correctly to the particular group that corresponded to its RAPD phenotype. This result demonstrates that quantitative morphological characters showing complex variation across taxa can conceal the information provided by the taxonomically more significant characters. More or less random variation in quantitative characters among closely related species is probably a common pattern, because the phenotypic expression of such characters usually is determined by many genes, and diverging populations may maintain similar polymorphisms at many loci for a long time.
The number of taxa in the Potentilla nivea complex in Svalbard
The molecular data and the variation in several individual morphological characters strongly support the hypothesis of Elven and Elvebakk (1996)
that the P. nivea complex consists of three taxa in Svalbard. The AMOVA analysis of the RAPD data for the P. nivea complex showed that the three postulated taxa were much more clearly differentiated (77.6% of the total variation) than the populations within these taxa (20.4%; Table 2). The most reliable characters were the occurrence of curly hairs in P. nivea, which clearly separated it from P. chamissonis and P. insularis, and the number of leaflets, end leaflet length:width ratio, first leaflet length:width ratio, and lobe length:width ratio on the end leaflet, which separated P. insularis from P. chamissonis. Several of these characters were also emphasized by Elven and Elvebakk (1996)
.
Potentilla nivea is the most distinct taxon within the complex in Svalbard, genetically as well as morphologically, whereas P. insularis and P. chamissonis are somewhat more similar to each other (cf. Figs. 2, 4, 8). It is nevertheless possible that it is most reasonable to recognize the three taxa at the same taxonomic level, or, alternatively, that P. nivea is recognized as one species and P. insularis and P. chamissonis as different subspecies of another species. However, it is not certain that similar patterns of variation exist in other geographic areas, and the results of this study should therefore be compared with studies of this widespread complex on a broader geographic scale before any final taxonomic conclusions are drawn.
Evidence for recent hybridization within the Potentilla nivea complex was only obtained from a single site in Svalbard. The "mixed population" (number 4) was collected at Hyperitthatten, the only site where all three taxa of the complex were found in close proximity. One plant analyzed from this "population" clearly belonged to P. nivea and another plant belonged to P. insularis, based on the genetic as well as the morphological data. In the PCO analysis of the morphological data, a third plant from this population was intermediate between P. chamissonis and P. nivea, but in the PCO analysis of the RAPD data, this plant grouped with P. nivea. Four additional plants from the mixed population had identical RAPD phenotypes that were intermediate between P. insularis and P. nivea, but these plants grouped with P. chamissonis in the morphological analysis. Thus, the situation at the Hyperitthatten site appears to be very complex, and it is possible that hybridization occurs between all three taxa at this site.
Potentilla pulchella: adaptively plastic in the heterogeneous arctic environment?
The hypothesis that the three "eco-morphotypes" within P. pulchella in Svalbard represent different intraspecific taxa was not supported by the molecular data. The species contained one dominant multilocus RAPD phenotype (37 of 45 plants), and this phenotype occurred in the populations of all three "eco-morphotypes." The "eco-morphotypes" could only be recognized by their conspicuous difference in plant size and degree of hairiness in our morphological analysis, and by their occurrence in different local habitats. The morphological variation in P. pulchella seems to be associated with different abiotic conditions. The "Sassen morphotype" as well as the "beach morphotype" are characterized by dwarfish growth in strongly windexposed habitats, but they occur on different types of soil (coarse, "Sassen morphotype;" fine-grained, "beach morphotype"). The larger plants of the "normal morphotype" occur in less exposed sites, such as manured bird-cliff meadows, where the soil is rich in nutrients. The morphological forms within P. pulchella may thus represent different phenotypically plastic expressions of the same genotype, serving to maximize its fitness in different environments (cf. Bradshaw, 1965
; MacDonald and Chinnappa, 1988
). It is possible, however, that the "eco-morphotypes" are genetically different in spite of their similarity at the high number of randomly distributed RAPD loci examined. They may differ at a small number of loci controlling the particular morphological traits observed, but the RAPD data may still yield an accurate estimate of their overall genetic similarity.
The hypothesis of hybrid origin of Potentilla insularis
The hypothesis that P. insularis originated as a hybrid between P. pulchella and another taxon of the P. nivea complex (i.e., P. nivea or P. chamissonis; modified from Soják, 1986
; cf. above) is not supported by the RAPD data. Potentilla insularis was not intermediate between the proposed parental taxa in the PCO analysis of the RAPD data, and it showed no clear additivity for individual RAPD markers. The data rather suggest that P. pulchella is genetically distinct from the entire P. nivea complex in Svalbard. Although P. insularis was somewhat intermediate between P. pulchella and P. chamissonis/nivea in the PCO analysis of the morphological data (Fig. 7), it was intermediate only in a few individual morphological characters (petal width, leaf hairiness, sepal hairiness, and lobe shape characters). The most important character influencing the position of P. insularis in the PCO analysis was probably the number of leaflets (usually, five vs. three in P. chamissonis and P. nivea; cf. Table 4). However, occasional intermediacy in morphological characters does not necessarily result from hybridization (e.g., Stace, 1989
; Rieseberg, 1995
).
Intrapopulational variation
Very low levels of RAPD variation were observed within the populations of Potentilla in Svalbard. The intrapopulational proportion of the total variation was 6.4–16.7% in the separate AMOVA analyses for each taxon (Table 2), and all populations but one contained only one or two multilocus RAPD phenotypes. It is possible that the Svalbard taxa of the P. nivea complex analyzed herein are more or less agamospermous. There are so far no data available on their reproductive biology, but several taxa of Potentilla from other geographic areas are facultatively agamospermous (Müntzing, 1928
; Smith, 1963
; Acharaya Goswami and Matfield, 1974
; Asker, 1977
; Asker and Jerling, 1992
; Eriksen, 1996
; Eriksen and Fredrikson, 2000
). The Svalbard taxa are large-flowered, and visiting insects have been observed many times (personal observations). The plants may be pseudogamous, i.e., they must be pollinated for induction of asexual seed development (cf. Müntzing, 1928
; Smith, 1963
). In the mainly allogamous Saxifraga oppositifolia L., which also has been analyzed in Svalbard, most of the RAPD variation was found within populations (64.8%; Gabrielsen et al., 1997
). In a RAPD analysis of the autogamous Saxifraga cespitosa L. in Svalbard, most of the variation was found among populations (59.3%), but there was still a considerable amount of intrapopulational variation (Tollefsrud et al., 1998
).
The low levels of variation in the populations analyzed of Potentilla were also demonstrated by their low proportions of distinguishable genets (mean G/N = 0.22) and, in particular, by their low levels of genotypic diversity (D; mean 0.20, range 0.00–0.60; Table 3). The genotypic diversity in the populations of Potentilla is thus considerably lower than that observed in the clonal species Saxifraga cernua L. in Svalbard (mean 0.52, range 0.10–0.81; Gabrielsen and Brochmann, 1998
), which is almost as diverse as clonal plants in general (mean D = 0.62, Ellstrand and Roose, 1987
; mean D = 0.75, Widén, Cronberg, and Widén, 1994
). The mean genetic evenness in the populations of Potentilla was also relatively low (E = 0.36; Table 3). This result demonstrates that the populations of Potentilla often have a single dominating RAPD phenotype, whereas other RAPD phenotypes, if occurring at all, are rare. This pattern is in accordance with the structure observed in Saxifraga cernua in Svalbard, where a few RAPD phenotypes dominated in each population (mean E = 0.42; Gabrielsen and Brochmann, 1998
).
|
|
|
|
|
FOOTNOTES |
|---|
4 Author for correspondence (e-mail: christian.brochmann{at}toyen.uio.no
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Asker, S. 1977 Pseudogamy, hybridization and evolution in Potentilla. Hereditas 87: 179–184
———, and L. Jerling. 1992 Apomixis in plants. CRC Press, Boca Raton, Florida, USA
Ball, P. W., B. Pawlowski, and S. M. Walters. 1968 Potentilla L. In T. G. Tutin, V. H. Heywood, N. A. Burges, D. M. Moore, D. H. Valentine, S. M. Walters, and D. A. Webb [eds.], Flora Europaea 2, 36–47. Cambridge University Press, Cambridge, UK
Belaeva, V. A., and V. N. Siplivinsky. 1976 Chromosome numbers and taxonomy of species of Baikal flora. II. Botanicheskii Zhurnal SSSR 61: 873–880 (in Russian)
Böcher, T. W., and K. Larsen. 1950 Chromosome numbers of some arctic or boreal flowering plants. Meddelelser om Grønland 147(6): 1–32
Bradshaw, A. D. 1965 Evolutionary significance of phenotypic plasticity in plants. Advances in Genetics 13: 115–155
Brown, R. 1824 Botany. A list of plants collected in Melville Island. Supplement Appendix XI. In W. E. Parry, Journal of a voyage for the discovery of a north-west passage. Murray, London, UK
Dansereau, P., and E. E. Steiner. 1956 Studies in Potentillae of high latitudes and altitudes. II. Central Baffin Island populations. Bulletin of the Torrey Botanical Club 83: 113–135[CrossRef]
Ellstrand, N. C., and M. L. Roose. 1987 Patterns of genotypic diversity in clonal plant species. American Journal of Botany 74: 123–131[CrossRef][ISI]
Elven, R. 1994 6th edition of J. Lid and D. T. Lid's Norsk Flora. Det Norske Samlaget, Oslo, Norway (in Norwegian)
———, and A. Elvebakk. 1996 Part 1: vascular plants. In A. Elvebakk and P. Prestrud [eds.], A catalogue of Svalbard plants, fungi, algae, and cyanobacteria, Norsk Polarinstitutt Skrifter 198, 9–55. Norsk Polarinstitutt, Oslo, Norway
Engelskjøn, T. 1979 Chromosome numbers from Norway, including Svalbard. Opera Botanica 52: 1–38
Eriksen, B. 1996 Mating systems in two species of Potentilla from Alaska. Folia Geobotanica et Phytotaxonomica 31: 333–344
———. 1997 Morphometric analysis of Alaskan members of the genus Potentilla sect. Niveae (Rosaceae). Nordic Journal of Botany 17: 621–630[ISI]
———, and M. Fredrikson. 2000 Megagametophyte development in Potentilla nivea L. (Rosaceae) from northern Swedish Lapland. American Journal of Botany 87: 642–651
———, B. Jonsell, and Ö. Nilsson. 1999 Proposal to conserve the name Potentilla nivea (Rosaceae) with a conserved type. Taxon 48: 165–166[CrossRef]
Excoffier, L., P. E. Smouse, and J. M. Quattro. 1992 Analysis of molecular variance inferred from metric distances among DNA haplotypes: applications to human mitochondrial DNA restriction data. Genetics 131: 479–491[Abstract]
Fager, E. W. 1972 Diversity: a sampling study. American Naturalist 106: 293–310[CrossRef][ISI]
Flovik, K. 1940 Chromosome numbers and polyploidy within the flora of Spitzbergen. Hereditas 26: 430–440[ISI]
Gabrielsen, T. M., K. Bachmann, K. S. Jakobsen, and C. Brochmann. 1997 Glacial survival does not matter: RAPD phylogeography of Nordic Saxifraga oppositifolia. Molecular Ecology 6: 831–842
———, and C. Brochmann. 1998 Sex after all: high levels of diversity detected in the arctic clonal plant Saxifraga cernua using RAPD markers. Molecular Ecology 7: 1701–1708[CrossRef]
Gower, J. C. 1966 Some distance properties of latent root and vector methods used in multivariate analysis. Biometrica 53: 325–338
———. 1967 Multivariate analysis and multidimensional geometry. Statistician 17: 13–28[CrossRef]
Goworuchin, V. S. 1932 Flora of Ural. Academiae Scientiarum URSS, Moscow, USSR (in Russian)
Hiitonen, I. 1949 Über die ostfennoskandischen Formen und Bastarde der Kollektivart Potentilla nivea L. nebst Erörterung einiger anderen Arten der Niveae Gruppe. Archivum Societatis Zoologicae Botanicae Fennicae "Vanamo" 2: 23–33
Holmen, K. 1952 Cytological studies in the flora of Peary Land, North Greenland. Meddelelser om Grønland 128(5): 1–40
Hultén, E. 1945 Studies in the Potentilla nivea group. Botaniska Notiser 1945: 127–148
———, and M. Fries. 1986 Atlas of North European vascular plants north of the Tropic of Cancer. Koeltz Scientific Books, Königstein, Germany
Jurtzev, B. A. 1984 Potentilla L. In B. A. Jurtzev [ed.], Flora arctica URSS. IX, 1. Droseraceae-Rosaceae, 137–234. Nauka, Leningrad, USSR (in Russian)
Juzepchuk, S. 1941 Potentilla L. In V. L. Komarov [ed.], Flora of USSR. X, 78–223. Academiae Scientiarum URSS, Moscow, USSR (in Russian)
Jørgensen, C. A., T. Sørensen, and M. Westergaard. 1958 The flowering plants of Greenland. A taxonomical and cytological survey. Biologiske Skrifter Danske Videnskabernes Selskab 9(4): 1–172
Knaben, G., and T. Engelskjøn. 1967 Chromosome numbers of Scandinavian arctic-alpine plant species. II. Acta Borealia A. Scientia 21: 1–57
Krogulevich, R. E. 1978 Karyological analysis of the species of the flora of eastern Sayana. In L. I. Malyshev and G. A. Peshlcova [eds.], Flora of the Prebaikal, 19–48. Novosibirsk, USSR (in Russian)
Lehmann, J. G. C. 1830 Novarum et minus cognitarum stirpium pugilli 2(11). Hamburg, Germany
MacDonald, S. E., and C. C. Chinnappa. 1988 Patterns of variation in the Stellaria longipes complex: effects of polyploidy and natural selection. American Journal of Botany 75: 1191–1200[CrossRef][ISI]
Müntzing, A. 1928 Pseudogamie der Gattung Potentilla. Hereditas 11: 267–283
Nathorst, A. G. 1883 Nya bidrag til kännedomen om Spetsbergens kärlväxter, och dess växtgeografiska förhållanden. Kungliga Svenska Vetenskaps-Akademiens Handlingar 20(6): 1–88 (in Swedish)
Noruis, M. J. 1993 SPSS for Windows, Release 6.0. SPSS Inc., Chicago, Illinois, USA
Pielou, E. C. 1969 An introduction to mathematical ecology. Wiley-Interscience, New York, New York, USA
Porsild, A. E., and W. J. Cody. 1980 Vascular plant of continental Northwest Territories. National Museum of Natural Sciences, Ottawa, Ontario, Canada
Resvoll-Holmsen, H. 1927 Svalbards flora med en del om dens plantevekst i nutid og fortid. J. W. Cappelens Forlag, Oslo, Norway (in Norwegian)
Rieseberg, L. H. 1995 The role of hybridization in evolution: old wine in new skins. American Journal of Botany 82: 944–953[CrossRef][ISI]
Rohlf, F. J. 1998 NTSYS-pc. Numerical taxonomy and multivariate analysis system, version 2.0. Exeter Software, New York, New York, USA
Rønning, O. I. 1961 Some new contributions to the flora of Svalbard. Norsk Polarinstitutts Skrifter 124: 1–20
Smith, G. L. 1963 Studies in Potentilla L. Embryological investigations into the mechanism of agamospermy in British P. tabernaemontani Aschers. New Phytologist 62: 264–282[CrossRef]
Sneath, P. H. A., and R. R. Sokal. 1973 Numerical taxonomy: the principles and practice of numerical classification. Freeman, San Francisco, California, USA
Soják, J. 1985 Some new northern hybrids in Potentilla L. Preslia 57: 263–266
———. 1986 Notes on Potentilla. I. Hybridogenous species derived from intersectional hybrids of sect. Niveae x sect. Multifideae. Botanische Jahrbücher für Systematik 106: 145–210
———. 1989 Notes on Potentilla (Rosaceae). VIII. P. nivea L. agg. Candollea 44: 741–762
Sommerfelt, C. 1833 Bidrag til Spitsbergens og Bereen-Eilands flora, efter herbarier medbragte af M. Keilhau. Magazin for Naturvidenskaberne 11: 232–252 (in Norwegian)
