| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Department of Systematic Botany, Lund University, Östra Vallgatan18–20, SE-223 61 Lund, Sweden; 3Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3DS, UK
Received for publication December 21, 2000. Accepted for publication March 16, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: AFLP • Dactylorhiza • Orchidaceae • phylogeny • polyploid evolution • systematics
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Soltis and Soltis, 1993
). For instance, it appears to be a general pattern that polyploid taxa have evolved repeatedly from progenitors of lower ploidy, which has resulted in higher levels of genetic diversity at the polyploid level than would otherwise be expected. Some studies also indicated that gene flow and introgression may occur across ploidy levels (Lord and Richards, 1977
; Brochmann, Stedje, and Borgen, 1992
; Menken, Smit, and Den Nijs, 1995
). Furthermore, restructuring of the hybrid genomes in allopolyploids may have resulted in new adaptively valuable combinations of parental characters; thus, polyploids are not evolutionary dead ends (Soltis and Soltis, 1993
) and may evolve independently of their parental taxa. Finally, the polyploid genome may later be diploidized, and over a longer timescale the whole cycle could be repeated (Stebbins, 1971
; Grant, 1981
; Leitch and Bennett, 1997
; Soltis and Soltis, 1999
).
In the orchid genus Dactylorhiza Nevski, allotetraploids have evolved on several occasions due to repeated hybridization (Hedrén, 1996a
) between two main groups of parental taxa, one consisting of the diploid marsh orchids, D. incarnata sensu lato (s.l.), and the other consisting of the diploid D. maculata subsp. fuchsii and the autotetraploid D. maculata subsp. maculata. Based on morphological patterns, the allotetraploid derivatives have been described as a large number of taxonomic species, but it is not known to what degree the various regional and local populations of these species are indeed of common ancestry or if similar morphological types have evolved on several occasions in separate areas.
These problems have profound implications for the formulation of conservation strategies that should be applied to the complex. We know that many constituent members of the complex, particularly allotetraploids, decreased in numbers during the latter part of the 20th century due to destruction of habitat. Most of these taxa grow in calcareous fens, which are often drained, both intentionally and accidentally.
If gene flow from the diploid to the tetraploid level is extensive, due to allotetraploids evolving repeatedly from the parental groups, then the allotetraploid taxa recognized on a morphological basis may contain unrelated populations with local or regional distributions. In that case, conservation efforts could be concentrated on the parental groups because they may be regarded as the basis for further evolution in the complex; from them, allotetraploids, if lost, could be easily regenerated, although the existence of certain local allotetraploids with adaptations to specific habitats and unique combinations of characters should not be disregarded. On the other hand, if allotetraploids evolve on rare occasions from the parental groups, then allotetraploid taxa present today may be coherent evolutionary units over large areas, and the morphologically defined taxonomic units are likely to reflect the evolutionary pattern of the complex. In that case, more conservation effort may be expended on the tetraploids.
One of us has previously used allozymes to describe evolutionary patterns in the complex (Hedrén, 1996a, b, c, d
). However, although allozymes may give valuable insights into the general patterns of evolution, allozyme loci are not variable enough to reveal many patterns of detailed relationships that are of interest both in evolutionary and conservation contexts. In this paper, we investigate the degree to which the recently developed amplified fragment length polymorphisms technique (AFLP; Vos et al., 1995
) can provide such information. We include some more distantly related members of Dactylorhiza, as well as members of related genera, to understand if AFLP data can be used as well to study such patterns.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
. The diploid taxa D. incarnata vars. incarnata, ochroleuca, cruenta, borealis, and subspp. coccinea and pulchella have been regarded as members of D. incarnata s.l. The diploid D. maculata subsp. fuchsii, the autotetraploid D. maculata subsp. maculata, and the diploid D. foliosa (Sundermann and Watke, 1973
) were treated as the D. maculata group. Allozyme data indicated that the following taxa are allotetraploids with origins in taxa with a high degree of similarity to members of D. incarnata s.l. and D. maculata/fuchsii: D. majalis subspp. majalis, lapponica, traunsteineri (Hedrén, 1996a
), praetermissa (Hedrén, 1996b
), purpurella (Hedrén, 1996c
), alpestris (Hedrén, unpublished data), D. sphagnicola (Hedrén, 1996a
), and D. elata (Hedrén, unpublished data). In addition, based on chromosome numbers and morphology (Bateman and Denholm, 1983
; Delforge, 1995
), the following taxa are also regarded as allotetraploids with a similar origin: D. majalis subspp. scotica, traunsteineroides, cordigera, and cambrensis. We also examined one sample each of D. iberica, D. romana, and D. sambucina, all diploid taxa. These species are not generally thought to be parental taxa in the polyploid complex described above, although Bateman, Pridgeon, and Chase (1997)
noted the possibility that D. sambucina is an alternative parent to at least some of the allotetraploids and so should be more carefully examined. Finally, we also included one sample each of three genera closely related to Dactylorhiza: Gymnadenia conopsea, Pseudorchis albida, and Coeloglossum viride, the latter included in Dactylorhiza by Bateman, Pridgeon, and Chase (1997)
.
Most of the material analyzed in this study was collected in Sweden. Some additional samples were included from other areas in Europe, particularly from the UK. A large portion of the British material was studied by Pridgeon et al. (1997)
, and DNA extracts from these plants were taken from the DNA Bank at the Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, UK. Some samples were collected by colleagues in other parts of Europe. The geographic origin of the material studied is given in Table 1, which also includes authors of scientific names.
|
. When available, we collected flowers. Orchid flowers have a thinner cuticle than the vegetative parts, thereby permitting more rapid desiccation. For Dactylorhiza taxa growing in mixed populations, we avoided flowers that had been pollinated and thus potentially contaminated with foreign DNA (i.e., pollen from other taxa).
DNA extraction
DNA was extracted from 
10–50 mg dry mass (1–5 flowers), either purified by using a cesium chloride-ethidium bromide gradient or QIAquick columns according to the manufacturer's protocols (QIAGEN, Crawley, West Sussex, UK). The gradient-purified samples are included in the general collection of DNA samples kept at the Jodrell Laboratory, as indicated in Table 1.
AFLP
We followed the general protocol described by Applied Biosystems (Warrington, Cheshire, UK) that takes advantage of an automated sequencer and computer analysis of fragment length variation. Sample DNA was restricted with the endonucleases EcoRI and MseI and ligated to appropriate double-stranded adapters according to the manufacturer's protocols. Two steps of amplification followed, a preselective amplification in which we used primers with a 1 base pair (bp) extension, and a second amplification in which primers with 3 bp extensions were used, thereby further reducing the number of fragments. For the second amplification, we initially tried nine different primer combinations, including combinations that were found to give good amplification and suitable numbers of fragments (typically 50–150 per accession) in Orchis simia (Qamaruz-Zaman et al., 1998
). From these, we selected three combinations, -ACT/-CTT, -AGG/-CAA, and -ACC/-CAC, for the extensions to the EcoRI and MseI sites, respectively.
Data analysis
The fragment data generated by the automated sequencer were analyzed by the computer programs GeneScan and Genotyper 2.0.1 (Applied Biosystems, Warrington, Cheshire, UK). In the latter program the band patterns were visualized as fingerprint traces that could be further inspected by eye. We used fragments in the range of 50–500 bp, and the computer-based system was set to consistently discard bands with a weak signal less than a threshold value recommended by the manufacturer. The data were extracted as a table in presence/absence format and subsequently carefully compared to the table generated automatically by the Genotyper program, whereby certain corrections were made. The use of internal size standards in each lane permitted exact calibration of different individuals against each other and made possible separation of nonhomologous fragments that were nearly equal in length. We scored some additional bands in samples that were not automatically scored by Genotyper in which the presence of fragments was obvious as distinct shoulders of more intense bands of an adjoining size class. We also recognized some additional bands in individual samples with more generally weak signals.
The presence/absence data were subjected to parsimony analysis and various phenetic analyses. The parsimony analysis was performed to investigate higher level patterns of relationship (if present) and included representatives of Pseudorchis, Gymnadenia, Coeloglossum, and some members of Dactylorhiza that were thought not to be components of the polyploid complex (i.e., D. sambucina, D. romana, and D. iberica) but excluded allopolyploid taxa as we knew a priori that they are the result of reticulate evolution. The computer-program PAUP 3.1.1 (Swofford, 1993
) was used to produce Wagner trees based on presence/absence data. A heuristic search strategy was implemented with TBR (tree bisection/reconnection) branch swapping. The "steepest descent" option was applied with MULPARS on (saving multiple parsimonious trees at each step). Group support was evaluated by means of bootstrap analysis (Felsenstein, 1985
). We used 5000 replicates and assigned equal weights to the characters. Gymnadenia and Pseudorchis were used as outgroups. Pridgeon et al. (1997)
found that these taxa were related to the Dactylorhiza clade (including Coeloglossum) in analyses of ITS rDNA data.
Phenetic analyses were performed on the members of Dactylorhiza that form the polyploid complex. Pairwise comparisons of all possible pairs of individuals by means of Jaccard coefficients (Jaccard, 1908
) generated triangular similarity matrices that were used for principal coordinates analyses (PCO; Gower, 1966
) and cluster analyses (unweighted pair-group method using arithmetic averages [UPGMA]; Sneath and Sokal, 1973
). Separate analyses were made for the whole polyploid complex, D. incarnata s.l. and the group of allotetraploid taxa. For the polyploid complex, we also calculated separate similarity matrices for the three different AFLP primer data sets analyzed separately and then compared pairwise by Mantel tests. All phenetic analyses were performed in NTSYS-pc 1.80 (Rohlf, 1994
).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
|
|
|
The same general pattern expressed in the phenogram was also revealed by PCO (Fig. 4) with the slight difference that the two specimens of D. majalis subsp. majalis here appeared to be more similar to each other and, likewise, that the two samples of D. majalis subsp. cordigera appeared closer to each other than indicated by the phenogram.
|
The D. maculata group
The lower part of the phenogram (Fig. 3) contains three subgroups, which corresponded to the three taxa included in this group (cf. the PCO; Fig. 2). The D. maculata subsp. fuchsii subgroup joined the D. maculata s.s. subgroup at a similarity level of 0.4 in the phenogram. The D. foliosa subgroup was more dissimilar and joined the other two subgroups at a similarity level of 0.3.
Mean similarity coefficients within D. maculata subsp. fuchsii and D. maculata s.s. were relatively similar, 0.57 and 0.52, respectively, whereas comparisons between the two taxa gave a mean similarity of 0.40; D. foliosa was somewhat more different from both.
Differentiation within D. incarnata s.l
Within the D. incarnata s.l. group shown in the phenogram (Fig. 3), the samples of D. incarnata s.s. were spread out over the subtree. Most samples of the characteristic color variety ochroleuca were concentrated on a shorter branch within the incarnata cluster, and most samples of var. cruenta grouped in a cluster that excluded the majority of ochroleuca samples, but were more interspersed with incarnata s.s. The same structure was evident from the separate PCO ordination including D. incarnata s.l. only; in the resulting plot of this ordination (Fig. 5), var. cruenta was concentrated to the lower left, var. ochroleuca was concentrated to the upper left, and D. incarnata s.s. was more evenly spread. In the phenogram, there was also some structure due to geographic origin of the samples. For instance, most specimens of D. incarnata s.s. from the province of Skåne (inc003–inc052) formed a cluster distinct from that containing the majority of D. incarnata s.s. from Gotland (inc103–inc142), but the differentiation was far from perfect, with several specimens of D. incarnata s.s. from these provinces falling in other parts of the D. incarnata s.l. cluster.
|
Comparisons of samples within D. incarnata vars. cruenta and ochroleuca resulted in slightly higher similarity coefficients (0.87 and 0.79, respectively; Table 3) than comparisons between these taxa and other members of D. incarnata s.l. (range 0.63–0.75). Comparisons between D. incarnata s.s. and other members of D. incarnata s.l. sometimes resulted in mean similarity coefficients equally as high as comparisons within D. incarnata s.s. All comparisons between members of D. incarnata s.l. gave higher similarity coefficients than comparisons between other taxa.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The AFLP data will be most useful in studying the first of these questions; AFLP fragments have been shown to be spread over the entire nuclear genome (e.g., Nilsson et al., 1999
), as expected from the distribution of sites for these restriction enzymes. Different primer combinations reveal independent subsets of this variation. Provided that the numbers of fragments studied are large enough, different primers should give similar patterns of differentiation among the samples studied. We used three different primer combinations, and a comparison of the pairwise similarity coefficients among individuals by means of Mantel tests (Table 2) revealed little difference between the primer data sets. Thus, the AFLP method generates data sets that reliably reflect differentiation between entire individual genomes.
Patterns of hybridization and polyploid speciation are usually described by means of phenetic methods such as the ones used here. For analysis of small data sets, e.g., data sets generated by allozyme studies in which the variation at each locus could be interpreted independently (Brochmann, Soltis, and Soltis, 1992
; Hedrén, 1996a
), most-parsimonious interpretations of the data may be possible by comparing relatively few alternative hypotheses. For large data sets generated by fingerprint methods (which would allow for a higher degree of resolution), computer-based methods must be used, but parsimony programs designed to study reticulate evolution are not available.
Higher relationships
Under the assumption that the fragments amplified by the AFLP method are distributed randomly and evenly over the size range investigated, the number of bands shared by two individuals by pure coincidence would approximately be given by a = n2/S, where n is the number of bands of different size classes found in each of the two individuals and S is the maximum possible number of size classes that could be distinguished. As already mentioned, the computer-based analysis of the output from an automated sequencer allows for a separation of fragments that are equal in length but differ in sequence by a (pseudo)difference of 0.5 bp or less. The range of fragments here studied was 50–500 bp, and a conservative estimate of the number of size classes would be 900 over this range relative to the degree of resolution discussed above. If two completely unrelated samples each have 70 recognized bands for one primer combination (a fairly typical number), then 5–6 bands would nevertheless be shared, equalling a Jaccard similarity coefficient of J = 0.037–0.045. Furthermore, the great majority of bands are concentrated in the lower size classes (50–250 bp). Two unrelated samples with 50 bands each in this range would accordingly share 6–7 bands for nonhomologous reasons (a Jaccard coefficient of J = 0.064–0.075), and the total number of shared bands over the entire range would be somewhat higher still. The number of bands shared by coincidence would be higher if the fragments are unequally distributed over the range studied.
We find in our data that comparisons between genera indeed give similarity coefficients close to these values (Table 3), and it seems possible that the regions of the genome analyzed by the AFLP procedure are not conserved enough to reveal relationships among these genera. However, it is still possible that a fraction of variation seen in AFLP is phylogenetically accurate, and that this information would be possible to extract using parsimony methods. It also seems likely that AFLP data represent coding as well as noncoding regions that would thus evolve at different rates. Although a large portion of the shared bands found in comparisons of taxa belonging to different genera are possibly false homologies stemming from rapidly evolving parts of the genome, some bands may still be true homologies that could be used in reconstructing phylogenies. Parsimony methods should be able to differentiate between these types of data if true homologies outnumber randomly coincident bands. Obviously, one way to increase the probability of finding accurate trees would be to enlarge the total number of characters used by parsimony by increasing the numbers of primers used to generate the AFLP data set.
Our tree contained less resolution than those presented by Pridgeon et al. (1997)
, which were based on ITS sequences. They also found that D. foliosa and D. maculata subsp. maculata formed a clade with subsp. fuchsii as sister. Other findings that are not supported, but which are also not contradicted by our results, include a close relationship of D. iberica to D. maculata s.s. and D. foliosa, a distant position of D. incarnata s.l. from D. maculata s.l. within the genus, and Coeloglossum embedded in Dactylorhiza. Dactylorhiza sambucina was not included in the ITS tree, but the closely related D. romana was more related to D. maculata s.l. than to D. incarnata s.l.
The polyploid complex and the allotetraploids
In the PCO (Fig. 2), the group of allotetraploids has a more or less intermediate position between the incarnata s.l. cluster and the maculata cluster, which is expected if the allotetraploids originated as hybrids between members of these groups. However, the allotetraploid group is somewhat displaced towards the back (i.e., they have higher values than expected for the second principal coordinate). This displacement may be an indication that the allotetraploids either have an origin in taxa slightly different from the present-day representatives of the putative parental group or that the tetraploid genomes have evolved further after the origin of the various allotetraploid taxa, perhaps by recombination between parental genomes. Both these possibilities indicate that the allotetraploids are relatively old and/or that they have their origin in areas where the parental groups have different genotypes from those in northwestern Europe. This same displacement for allopolyploid taxa is observed in Calopogon (Goldman, M. W. Chase, and M. F. Fay, unpublished data).
The allotetraploids are apparently more similar to the incarnata s.l. group than to the maculata/fuchsii group (Table 3), which is why they also cluster together with incarnata s.l. in the phenogram (Fig. 3). However, it is also evident from Table 3 that similarities within taxa in incarnata s.l. are much higher than in fuchsii or in maculata. It is likely that this difference in similarity coefficients is due to a higher degree of homozygosity in members of incarnata s.l. than in fuchsii or maculata, a difference that is clearly seen in single-locus data provided by allozyme markers (Hedrén, 1996a
). Thus, every comparison between an allotetraploid and a sample of incarnata s.l. will show a higher degree of similarity than comparisons with maculata/fuchsii, because the incarnata sample is likely to share more alleles with the incarnata s.l. parent that gave rise to the allotetraploid than is the maculata/fuchsii sample with the maculata/fuchsii parent.
Because the allotetraploid samples from the British Isles form a rather coherent group in the phenogram (Fig. 3), it may be speculated that these allotetraploids would be more closely related to samples of incarnata s.l. from the British Isles than to incarnata from other areas. We find no support for this hypothesis from our similarity data (Table 3). However, we studied only two British samples of incarnata s.l., and because these samples yielded slightly fewer bands than did incarnata s.l. on the average, we cannot reject the hypothesis of a close relationship.
The Dactylorhiza maculata group
Allozyme data indicate that D. maculata subsp. maculata is an autotetraploid (Hedrén, 1996a
) and that there is a close correspondence of D. maculata subsp. maculata with D. maculata subsp. fuchsii in both allele composition and allele frequencies. Accordingly, it may be hypothesized that D. maculata subsp. maculata has an origin in a diploid taxon closely related to present-day D. maculata subsp. fuchsii and that gene transfer from diploids to tetraploids may still be occurring. Therefore, it is surprising to find that D. maculata subsp. fuchsii and D. maculata s.s. come out as distinct groups in the analyses presented here. There is a possibility that some of the differences found are due to the fact that tetraploids are compared to diploids (as discussed above), but from inspection of the electropherograms it appears that a high proportion of the differences found are indeed due to presence of distinct fragments with good amplification, and that the separation of the two taxa is real. The clear separation of D. maculata s.s. and D. maculata subsp. fuchsii in Sweden could indicate that Swedish D. maculata s.s. arose from diploid stocks elsewhere (perhaps now extinct) and migrated independently of D. maculata subsp. fuchsii after the last glaciation.
The maculata s.s. cluster, as well as the majority of allotetraploids investigated, attain high values along the third (vertical) principal coordinate (Fig. 2). Assuming intermediacy of the allotetraploids between the parental groups, it seems likely that material similar to maculata s.s. rather than fuchsii gave rise to most allotetraploids. Considering the model for stepwise evolution of allopolyploids by means of unreduced gametes (Müntzing, 1930a, b
; deWet, 1980
; Ramsey and Schemske, 1998
), it is possible that allotetraploid Dactylorhiza taxa may have evolved by hybridization between D. maculata s.s. (4x, genome FFFF) and D. incarnata (2x, genome II). First, a triploid hybrid with genome constitution FFI would be formed, followed by amalgamation of an unreduced gamete from this hybrid with a normal gamete from D. incarnata, resulting in an allotetraploid plant (FFII; Hedrén, 1996a
). However, it is not necessary to postulate that the tetraploid maculata gave rise to the allopolyploids directly; it is also possible that the allotetraploids originated from the same group of diploids that gave rise to the autotetraploid maculata. As indicated by the plot, and as evident from the phylogenetic trees (Fig. 1), the Madeiran endemic D. foliosa is such a diploid species, although its limited occurrence today indicates that it may not have been directly involved. It is clear that diploids from the entire distribution area need to be investigated. In contrast to the other allotetraploids, the two specimens of the undescribed allotetraploid from Gotland (ssp098 and ssp105) have low values for the third principal coordinate, and it could be speculated that the non-incarnata parent was indeed the diploid fuchsii. This assumption also agrees with the fact that fuchsii is much more abundant on Gotland than maculata s.s. and that fuchsii is now found close to these allotetraploid populations today. Thus, these allotetraploids may have had a recent and local origin.
The PCO (Fig. 2) indicates that foliosa is more similar to maculata s.s. than to fuchsii. However, the phenogram (Fig. 3) reveals that maculata s.s. and fuchsii are more similar to each other than to foliosa. Apparently, foliosa differs from the two other taxa by other characters than those expressed along the first three axes in the PCO ordination. These three axes summarize the main differentiation pattern in the entire polyploid complex within which incarnata s.l. and maculata/fuchsii are the most differentiated groups.
Differentiation within Dactylorhiza incarnata s.l
Within the incarnata s.l. group the varieties cruenta, ochroleuca, and incarnata s.s. showed a complex pattern of relationships in which the variation was partly correlated with geographic origin. The single borealis sample appeared to be slightly more different. However, as this sample came from a more northern location in Sweden, genomic differentiation correlated with geographic origin and correlated to morphological characters cannot be separated. The western European subspecies coccinea and pulchella came out together and were somewhat more different from the remainder of incarnata s.l., but again the effect of geographic origin and general differentiation affecting morphology cannot be distinguished.
The relatively low degree of differentiation within incarnata s.l. as compared to D. maculata/fuchsii agrees with findings from allozyme analyses. Whereas Swedish samples of D. incarnata s.l. were fixed for single alleles at seven loci used to describe the general pattern of relationships within the allopolyploid complex, D. maculata subsp. fuchsii and D. maculata subsp. maculata each showed variation at all of these loci (Hedrén, 1996a
).
Taxonomic conclusions
We included several samples of D. incarnata var. incarnata, var. cruenta, and var. ochroleuca in our study. These taxa may be strikingly different from each other in corolla color, leaf spotting or leaf curvature. Because the taxa appear distinct, they are often treated as subspecies or even species in orchid floras (e.g., Delforge, 1995
). The three taxa also differ in ecology; whereas var. incarnata grows in a variety of rich fens, var. cruenta and var. ochroleuca are restricted to calcareous fens in which the latter, on average, seems to prefer slightly wetter and shadier subsites.
In our analyses, the three taxa are interdigitated, and the differences in external morphology are apparently not correlated with general differentiation of the genomes (Figs. 1, 3, 5). The separating characters may thus be due to just a few genes, but the taxa are probably fixed for different alleles at these loci. Intermediate plants are sometimes found in mixed populations. Accordingly, crosses between the three taxa may take place, and it seems likely that plants belonging to a given taxon could give rise to variable offspring approaching the other taxa.
An explanation for this type of differentiation pattern may be given by ecotype formation (Turesson, 1922
). We hypothesize that the forms recognized as ochroleuca and cruenta may occur in the progeny of typical incarnata plants and are adapted to particular habitats. If similar selection pressures occur at different places, each of these forms may be independently derived from incarnata on more than one occasion. Such a scenario is consistent with our data. The value of naming these forms may be questioned because they contain populations that are not more related to each other than to populations of the other taxa. Still, there may be a need for ecologists and conservationists to separate the various forms because the names bear information on the habitat in which the plants were found. We suggest that these forms are best treated as varieties (implying that the concept of subspecies should be restricted to situations in which the taxa are more coherent, with constituent populations being more closely related to each other than to populations of other subspecies; Jonsell, 2000
).
We analyzed only a few specimens of D. incarnata s.l. from the British Isles. However, extended sampling from this area may show that they are more genetically distinct and that it would be appropriate to treat them as subspecies (cf. Bateman and Denholm, 1985
).
In the D. maculata group, the northern European representatives D. maculata subsp. maculata and subsp. fuchsii were clearly differentiated from each other, and they can be viewed as monophyletic sister taxa. However, they are morphologically ill-defined and not always distinguishable, as pointed out by several continental authors (e.g., Hylander, 1966
; cf. also discussion in Bateman and Denholm, 1988
), for which reason we treat them as subspecies rather than species. The exact position of D. foliosa remains unclear. It appears to be morphologically distinct, and in the absence of any conclusive evidence regarding its phylogenetic position, we continue to treat it as a separate species.
The allotetraploid members of Dactylorhiza have apparently evolved on repeated occasions from the same set of broadly defined parental species. Although the allotetraploids have similar origins, they may be more or less differentiated in morphology and habitat preferences, and they are often found in geographically separate areas. It is implied that each recognized allotetraploid either has a single origin or several origins from the same subset of the parental groups; in any case, they each contain a unique combination of characters from the parental genomes and a subset of the variation found in each parental group. Although not monophyletic in a strict sense (as they have originated by reticulate evolution), they have unique origins and could be viewed as evolutionary units evolving relatively independently of other such units. On the other hand, there are a fairly large number of cases in which different allotetraploids are difficult to separate from each other. First, allotetraploids in separate geographic areas may approach each other in morphological characters, although they have different origins; examples of this are D. majalis subsp. alpestris in the Alps and in the Pyrennees, subsp. cordigera on the Balkans, subsp. majalis in north-central Europe, and subsp. cambrensis in the British Isles. Secondly, at sites where they coincide, different, otherwise morphologically well-defined allotetraploids may be connected to each other by intermediate plants, indicating that hybridization and backcrossing occur, leading to gene flow between the allotetraploids. Examples of this may be D. majalis subsp. traunsteineri and D. sphagnicola at several Swedish sites. We propose that the various allotetraploids should be recognized as subspecies of one species separate from the parental species, which is in accordance with the treatment presented by Bateman and Denholm (1983)
. This solution would take into account that the various subspecies are connected to each other by intermediate forms and may be difficult to separate from each other. At the same time we recognize that they are still relatively independent from each other and have unique origins. Treating the allotetraploids as subspecies would decrease the naming of new taxa and would make the taxonomic treatment of Dactylorhiza comparable to that used in most other plant groups. However, subspecies should be delimited as carefully as possible to reflect the evolutionary patterns in the complex, and much work remains to investigate the subspecies that have already been described. Hybrids between the allotetraploids and the diploid parents have lower fertility (Stace, 1975
); thus, we are motivated to treat the allotetraploid lineages as a separate species.
Taxonomic note
Many allotetraploid taxa have already been treated as subspecies of D. majalis and we follow this practice here if names at subspecies level are available. The southwestern European D. elata is also an allotetraploid, and an amalgamation of this species with D. majalis should be considered as well. The name D. elata would have priority over D. majalis, which would require numerous combinations to be made. However, we require better knowledge of the pattern of variation in D. elata before we can eventually decide upon the taxonomy of this complex, and we refrain from proposing new combinations until more data are available. We also treat D. sphagnicola as a separate species until this question has been settled.
Conservation
Conservation strategies for polyploid complexes containing rare species should consider among other factors the rate at which new polyploid derivatives are formed from the stock of parental taxa, the relationships of different polyploid derivatives, the area of origin of polyploids, and the size of their distribution areas. They should also evaluate the status of the parental species to ensure that future evolution of the complexes are not restricted by human activity. These factors are rarely well known for polyploid complexes, although the use of molecular markers has contributed considerably to the knowledge of polyploid evolution in recent years (e.g., Cook et al., 1998
).
Our data indicate that new polyploid derivatives arise at a moderate rate in Dactylorhiza. The allotetraploids occurring in the British Isles cluster together in our analyses, indicating that these allotetraploids have a regional origin different from those occurring in Scandinavia. This finding contrasts with some taxonomic treatments of certain allopolyploids, in which both British and Scandinavian populations are accommodated in D. majalis subsp. lapponica, subsp. majalis, or subsp. traunsteineri. On the other hand, there are indications that local populations in larger regions have the same origins (e.g., D. sphagnicola in southern Sweden, or D. majalis subsp. purpurella in northwestern Europe), which indicates that some allotetraploids may have spread over fairly extensive areas after formation.
It appears that Dactylorhiza contains a moderately large number of independently evolved allotetraploids. We propose that each of the allotetraploids should be given equal value in a conservation program. For effective conservation, we need a taxonomy that reflects as closely as possible their independent origins. In the circumscription of these allotetraploids, morphological data alone does not appear to be sufficient (which is also why we prefer to treat them as subspecies rather than as species) and should be supplemented by genetic, ecological, and geographical data.
Our data also indicate that present-day gene transfer from the diploid D. maculata subsp. fuchsii to the autotetraploid subsp. maculata is restricted and that new autotetraploid populations do not evolve often. It is suggested that in a conservation context these two subspecies should also be considered as independent entities of equal importance.
|
|
|
FOOTNOTES |
|---|
4 Author for reprint requests (e-mail: Mikael.Hedren{at}sysbot.lu.se
; Fax +46 46 222 42 34).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Bateman R. M. I. Denholm 1985 A reappraisal of the British and Irish dactylorchids. 2. The diploid marsh-orchids. Watsonia 15: 321-355
Bateman R. M. I. Denholm 1988 A reappraisal of the British and Irish dactylorchids. 3. The spotted orchids. Watsonia 17: 319-349
Bateman R. M. A. M. Pridgeon M. W. Chase 1997 Phylogenetics of subtribe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 2. Infrageneric relationships and reclassification to achieve monophyly of Orchis sensu lato. Lindleyana 12: 113-141
Brochmann C. P. S. Soltis D. E. Soltis 1992 Recurrent formation and polyphyly of Nordic polyploids in Draba (Brassicaceae). American Journal of Botany 79: 673-688[CrossRef][ISI]
Brochmann C. B. Stedje L. Borgen 1992 Gene flow across ploidal levels in Draba (Brassicaceae). Evolutionary Trends in Plants 6: 125-134
Chase M. W. H. G. Hills 1991 Silica gel: an ideal material for field preservation of leaf samples for DNA studies. Taxon 40: 215-220[CrossRef][ISI]
Cook L. M. P. S. Soltis S. J. Brunsfeld D. E. Soltis 1998 Multiple independent formations of Tragopogon tetraploids (Asteraceae): evidence from RAPD markers. Molecular Ecology 7: 1293-1302[CrossRef]
Delforge P. 1995 Orchids of Britain and Europe. Harper Collins, London, UK
deWet J. M. J. 1980 Origins of polyploids. In W. H. Lewis [ed.], Polyploidy: biological relevance, 3–15. Plenum Press, New York, New York, USA.
Felsenstein J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Gower J. C. 1966 Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53: 325-338
Grant V. 1981 Plant speciation. Columbia University Press, New York, New York, USA
Hedrén M. 1996a Genetic differentiation, polypoidization and hybridization in northern European Dactylorhiza (Orchidaceae): evidence from allozyme markers. Plant Systematics and Evolution 201: 31-55[CrossRef][ISI]
Hedrén M. 1996b The allotetraploid nature of Dactylorhiza praetermissa (Orchidaceae) confirmed. Watsonia 21: 113-118
Hedrén M. 1996c Electrophoretic evidence for allotetraploid origin of Dactylorhiza purpurella (Orchidaceae). Nordic Journal of Botany 16: 127-134
Hedrén M. 1996d Notes on the esterase variation in Swedish Dactylorhiza incarnata s. l. (Orchidaceae). Nordic Journal of Botany 16: 253-256[ISI]
Hylander N. 1966 Nordisk kärlväxtflora II. Almquist & Wiksell, Stockholm, Sweden
Jaccard P. 1908 Nouvelles recherches sur la distribution florale. Bulletin de la Société Vaudoise de Science Naturelle 44: 223-70
Jonsell B. [ed.] 2000 Flora Nordica 1. The Bergius Foundation, Stockholm, Sweden
Leitch I. J. M. D. Bennett 1997 Polyploidy in angiosperms. Trends in Plant Science 2: 470-476[CrossRef][ISI]
Lord R. M. A. J. Richards 1977 A hybrid swarm between the diploid Dactylorhiza fuchsii (Druce) Soó and the tetraploid D. purpurella (T. & T. A. Steph.) Soó in Durham. Watsonia 11: 205-210
Menken S. B. J. E. Smit H. J. C. M. Den Nijs 1995 Genetical population structure in plants: gene flow between diploid sexual and triploid asexual dandelions (Taraxacum section Ruderalia). Evolution 49: 1108-1118[CrossRef][ISI]
Müntzing A. 1930a Outlines to a genetic monograph to the genus Galeopsis with special reference to the nature and inheritance of partial sterility. Heriditas 13: 185-341[ISI]
Müntzing A. 1930b Über Chromosomenvermerhrung in Galeopsis-Kreutzungen und ihre phylogenetische Bedeutung. Heriditas 14: 153-172[ISI]
Nilsson N.-O. M. Hansen A. H. Panagopoulos S. Tuvesson M. Ehlde M. Christiansson I. M. Rading M. Riessler T. Kraft 1999 QTL analysis of Cercospora leaf spot resistance in sugar beat. Plant Breeding 118: 327-334[CrossRef][ISI]
Pridgeon A. M. R. M. Bateman A. C. Cox J. H. Hapeman M. W. Chase 1997 Phylogenetics of subtribe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 1. Intergeneric relationships and polyphyly of Orchis sensu lato. Lindleyana 12: 89-109
Qamaruz-Zaman F. M. F. Fay J. S. Parker M. W. Chase 1998 The use of AFLP fingerprinting in conservation genetics: a case study of Orchis simia (Orchidaceae). Lindleyana 13: 125-133
Ramsey J. D. W. Schemske 1998 Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annual Reviews in Ecology and Systematics 29: 467-501[CrossRef]
Rohlf F. J. 1994 NTSYS-pc. Exeter Software, Setauket, New York, USA
Sneath P. H. A. R. R. Sokal 1973 Numerical taxonomy. Freeman, San Francisco, California, USA
Soltis D. E. P. S. Soltis 1993 Molecular data and the dynamic nature of polyploidy. Critical Reviews in Plant Sciences 12: 243-273
Soltis D. E. P. S. Soltis 1999 Polyploidy: recurrent formation and genome evolution. Trends in Ecology and Evolution 14: 348-352
Stace C. A. 1975 Hybridization and the flora of the British Isles. Academic Press, London, UK
Stebbins G. L. 1971 Chromosomal evolution in higher plants. Edward Arnold, London, UK
Sundermann H. R. Wattke 1973 Neue Beiträge zur Zytotaxonomie der Erdorchideen (Dactylorhiza, Orchis und Serapias). Orchidee (Hamburg) 24: 116-118
Swofford D. L. 1993 PAUP: phylogenetic analyses using parsimony, version 3.1.1. Computer program distributed by the Illinois Natural History Survey, Champaign, Illinois, USA
Thompson J. R. Lumaret 1992 The evolutionary dynamics of polyploid plants: origins, establishment and persistence. Trends in Ecology and Evolution 7: 302-307[CrossRef]
Turesson G. 1922 The genotypical response of the plant species to the habitat. Heriditas 3: 211-350[ISI]
Vos P. R. Hogers M. Bleeker M. Reijans T. van de Lee M. Hornes A. Freijters J. Pot J. Peleman M. Kuiper M. Zabeau 1995 AFLP, a new technique for DNA fingerprinting. Nucleic Acids Research 23: 4407-4414
This article has been cited by other articles:
|
|
 |
|
  M. Hedren, S. Nordstrom, H. A. P. Hovmalm, H. A. Pedersen, and S. Hansson Patterns of polyploid evolution in Greek marsh orchids (Dactylorhiza; Orchidaceae) as revealed by allozymes, AFLPs, and plastid DNA data Am. J. Botany, July 1, 2007; 94(7): 1205 - 1218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. BATEMAN and P. J. RUDALL Evolutionary and Morphometric Implications of Morphological Variation Among Flowers Within an Inflorescence: A Case-Study Using European Orchids Ann. Bot., November 1, 2006; 98(5): 975 - 993. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. PEREZ-COLLAZOS and P. CATALAN Palaeopolyploidy, Spatial Structure and Conservation Genetics of the Narrow Steppe Plant Vella pseudocytisus subsp. paui (Vellinae, Cruciferae) Ann. Bot., April 1, 2006; 97(4): 635 - 647. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. F. FAY, R. S. COWAN, and I. J. LEITCH The Effects of Nuclear DNA Content (C-value) on the Quality and Utility of AFLP Fingerprints Ann. Bot., January 1, 2005; 95(1): 237 - 246. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Pellegrino, A. Musacchio, M. E. Noce, A. M. Palermo, and A. Widmer Reproductive Versus Floral Isolation Among Morphologically Similar Serapias L. Species (Orchidaceae) J. Hered., January 1, 2005; 96(1): 15 - 23. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Kellogg and J. L. Bennetzen The evolution of nuclear genome structure in seed plants Am. J. Botany, October 1, 2004; 91(10): 1709 - 1725. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. B. Shipunov, M. F. Fay, Y. Pillon, R. M. Bateman, and M. W. Chase Dactylorhiza (Orchidaceae) in European Russia: combined molecular and morphological analysis Am. J. Botany, September 1, 2004; 91(9): 1419 - 1426. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. H. Goldman, R. K. Jansen, C. van den Berg, I. J. Leitch, M. F. Fay, and M. W. Chase Molecular and cytological examination of Calopogon (Orchidaceae, Epidendroideae): circumscription, phylogeny, polyploidy, and possible hybrid speciation Am. J. Botany, May 1, 2004; 91(5): 707 - 723. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
