| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Department of Botany, The Natural History Museum, Cromwell Road, London SW7 5BD UK; 3Jardín de Aclimatación de La Orotava, Calle Retama Num. 2, E-38400, Puerto de la Cruz, Tenerife, Canary Islands, Spain; 4Department of Biological Sciences, Florida International University, University Park, Miami, Florida 33199 USA; Fairchild Tropical Garden, 11935 Old Cutler Road, Miami, Florida 33156 USA
Received for publication October 5, 2003. Accepted for publication March 11, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: biogeography • Cabo Espichel • Canary Islands • Convolvulaceae • Convolvulus • Jacquemontia • Madeira • Oceanic Islands
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
]; 40% in the Canary Islands [Santos Guerra, 1999
]) and a predominant woody habit among endemics (ca. 70% of Canary Island endemics are woody; Aldridge, 1979
). However, whilst the Macaronesian region is typical of oceanic island systems in several respects, the region in general, and the Canary Islands in particular, differ markedly from more isolated Pacific island groups such as Hawaii, in that they are close to potential continental source areas. The easternmost Canary Islands are presently 95 km from the coast of North Africa and at periods during the last 20 million years (my) they have probably been as close as 65 km (García-Talavera, 1997
, 1999
). There are also more than 20 large volcanic sea mounts located between the Canary Islands, Madeira, Selvages, and the continent, several of which are presently less than 100 m below sea level. Recent dating studies have shown that some of the sea mounts in this area are at least 68 my old (Geldmacher et al., 2001
), and it is conceivable that they may have facilitated dispersal between the continent and islands by serving as "stepping stones" during glacial periods when sea levels were lower (García-Talavera, 1997
).
|
) raises the possibility that genera could have undergone multiple independent colonizations into the region. Back-colonization to the continent from Macaronesia could also have occurred, with the islands serving as a source area for continental endemics (Santos-Guerra, 1999
; Park et al., 2001
; Fuertes-Aguilar et al., 2002
; Goertzen et al., 2002
). Such relationships between island and continental floras are more complex than those generally hypothesized for other island oceanic island systems such as the Hawaiian and Galapagos archipelagos.
Recent molecular phylogenetic analyses of Macaronesian plant groups have provided valuable insights into the relationships of the region's endemics. Whilst each group seems to have a unique pattern of diversification in the region (Francisco-Ortega et al., 2001b
), a number of more general patterns are starting to emerge. Multiple independent colonizations into the region may seem plausible but in the majority of genera so far investigated, the Macaronesian endemic taxa form a single monophyletic group. In most cases, the sister group of the Macaronesian clade is North African or western European and typically Mediterranean (e.g., Argyranthemum Webb ex Sch.Bip. (Asteraceae), Francisco-Ortega et al., 1997
; Isoplexis (Lindl.) Loud. (Scrophulariaceae), Carvalho and Culham, 1998
; Bencomia Webb & Berth. alliance (Rosaceae), Helfgott et al., 2000
; Sideritis Tourn.ex Linn. (Lamiaceae), Barber et al., 2002
; see also Santos-Guerra, 1999
), although sister group relationships with taxa distributed elsewhere, notably the New World (e.g., Madeiran species of Sedum L. (Crassulaceae), van Ham and t'Hart, 1998
; Pericallis D.Don (Asteraceae), Panero et al., 1999
; Bystropogon L'Hérit. (Lamiaceae), Trusty et al., in press), East Africa (e.g., Solanum L. (Solanaceae), Olmstead and Palmer, 1997
; Camptoloma Benth. (Scrophulariaceae), Kornhall et al., 2001
) and southern Africa (e.g., Phyllis L. (Rubiaceae), Andersson and Rova, 1999
) have also been revealed (Andrus et al., 2004
).
Multiple congeneric colonizations into the region have been demonstrated for a number of genera investigated so far (Table 1) but documented examples of back-colonization of the continent from the Macaronesian region remain limited. Molecular phylogenetic analyses of Euphorbia L. subsect. Pachyclade (Boiss.) A.DC. (Euphorbiaceae; Molero et al., 2002
) and Teline Medikus (Fabaceae; Percy and Cronk, 2002
) are consistent with back-colonization of the continent but are equivocal. Back-colonization of the continent from Macaronesia is the most plausible explanation for the pattern of relationships of continental and island taxa in Tolpis Adans. (Asteraceae, Park et al., 2001
; Moore et al., 2002
). In Aeonium Webb & Berth. (Crassulaceae), the African representatives of the genus are nested within an otherwise exclusively Macaronesian clade, a pattern consistent with dispersal from Macaronesia to Africa (Mort et al., 2002
).
|
). One genus for which both multiple colonizations into Macaronesia and back-colonization of the continent have been hypothesized is Convolvulus L. (Convolvulaceae). Ten species of Convolvulus are endemic to the Canary Islands and one is endemic to Madeira and Desertas (Table 2). Of the remaining species found in Macaronesia, C. arvensis L., C. althaeoides L., and C. farinosus L. are considered introduced (Kunkel, 1991
). The status of the weedy annual species C. siculus L., which is also widespread around the Mediterranean basin, is uncertain (Kunkel, 1991
).
|
). Among the island endemic taxa, however, there is considerable morphological variation and taxonomic treatments of Convolvulus have placed the Macaronesian endemic species in two (Webb and Berthelot, 1840
), three (Peter, 1891
), or four (Sa'ad, 1967
) different infrageneric groupings. Most recently, Mendoza-Heuer (1971
, 1983
) recognized two distinct groups of Macaronesian endemic Convolvulus, one comprising the endemic climbers and the other comprising the endemic non-climbers (Table 2). Furthermore, Mendoza-Heuer (1971
, 1983
) proposed that these two groups had resulted from two independent colonizations of the region.
Mendoza-Heuer (1971
, 1983
) was largely silent on the relationships of the endemic species of Convolvulus to other members of the genus but previous classifications have suggested a number of relationships that remain to be tested. These include a possible New World link for the Macaronesian endemic climbers (Choisy, 1845
; Peter, 1891
) and an eastern Mediterranean/Middle Eastern link for both C. scoparius L. (Sa'ad, 1967
) and C. floridus L. (Peter, 1891
). Additionally, Sa'ad (1967)
proposed a North African link between the spiny endemic species C. caput-medusae Lowe and the North African spiny subshrub C. trabutianus Schweinf. & Muschler. This proposal was, however, vigorously challenged by Mendoza-Heuer (1971
, 1983
) who maintained, on the basis of floral morphology, that all three non-climbing Macaronesian endemic species formed a single group (Table 2, Group I).
A close relationship has also been proposed between the C. canariensis L. group of endemics (Table 2, Group IIa) and the Portuguese endemic C. fernandesii Pinto da Silva and Teles, which is confined to coastal cliffs at Cabo Espichel, in Portugal (Fig. 1). The scrambling woody habit of C. fernandesii is unlike any other species of Convolvulus distributed in the western Mediterranean, and Pinto da Silva and Teles (1981)
hypothesized that C. fernandesii is a continental neo-endemic resulting from recent back-colonization of Cabo Espichel from the Macaronesian region. An alternative explanation, however, was provided by Bramwell and Bramwell (2001)
, who considered C. fernandesii and its putative relatives in the Macaronesian region to be relicts of a formerly widespread woody flora that covered the Mediterranean region during the Tertiary. As extinction occurred elsewhere in the Mediterranean following climate change in the Pleistocene, these woody taxa were able to survive in a few suitable refugial areas including Macaronesia and Cabo Espichel (e.g., Bramwell, 1972
; Bramwell and Richardson, 1973
; Aldridge, 1979
; Sunding, 1979
; Médail and Quezel, 1999
; Bramwell and Bramwell, 2001
).
There are therefore a number of competing hypotheses to explain the diversity and evolutionary history of Convolvulus in Macaronesia that remain to be tested. Potentially, Convolvulus demonstrates one of the most complex patterns of relationship in the region. There are very few genera for which multiple introductions into Macaronesia and back-colonization of the continent have been hypothesized, and Convolvulus therefore presents a rare example with which to explore these phenomena and gain new insights into the evolution and relationships of the Macaronesian flora. To date, hypotheses concerning the relationships and evolution of Macaronesian Convolvulus have been based exclusively on intuitive assessments of morphology. In many groups, similar hypotheses have been refuted by the application of more rigorous analytical techniques (e.g., Bremer and Humphries, 1993
) or new sources of data, notably DNA sequence data (e.g., Böhle et al., 1996
; Kim et al., 1996
; Francisco-Ortega et al., 1997
; Helfgott et al., 2000
; Mort et al., 2002
). In this paper, therefore, a molecular phylogenetic analysis of relationships of the Macaronesian endemic Convolvulus is presented using data from the internal transcribed spacer regions (ITS) of nrDNA. Three specific questions are addressed: (1) How many endemic groups of Convolvulus occur in the Macaronesian region and how many colonizations into the region are necessary to explain the diversity among the endemic species? Is there any support for the Mendoza-Heuer (1983)
hypothesis of two independent colonization events? (2) What are the closest continental relatives of the endemic species? We aim to test (a) the putative New World link between the island climbers (Table 2, Group II) and C. nodiflorus Desr. (Choisy, 1845
; Peter, 1891
); (b) the possible close relationship between the endemic spiny subshrub C. caput-medusae, C. trabutianus from North Africa and other spiny subshrubs (Sa'ad, 1967
; cf. Mendoza-Heuer, 1971
, 1983
); and (c) the hypothesized relationships between C. scoparius, C. floridus, and Eastern Mediterranean/Arabian species (Choisy, 1845
; Peter, 1891
; Sa'ad, 1967
) (3) What (and where) are the closest relatives of C. fernandesii? Is this species a neo-endemic and the result of speciation following back-colonization of Cabo Espichel from the Macaronesian region (Pinto da Silva and Teles, 1981
) or is it a Tertiary relict (Bramwell and Bramwell, 2001
)?
The results of the Convolvulus study are discussed in relation to other groups that demonstrate a complex colonization history in the region, to gain new insights into the processes underpinning the evolution of the Macaronesian endemic flora. The patterns of relationships of the Macaronesian and Mediterranean floras revealed by molecular data are also reviewed, and the implications of these data for our understanding of the evolution of the flora of these two regions are discussed.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Peter, 1891
; Sa'ad, 1967
). The remaining Convolvulus species included in the analysis represent a morphologically and geographically diverse sample of the genus from the Mediterranean and Western Asia.
Three species of Calystegia R.Br. were included in the study as this genus has been shown to be nested within a paraphyletic Convolvulus (Stefanovi
et al., 2002
). Seven other species, representing related genera of Convolvulaceae, were included as outgroups on the basis of higher-level analyses (Manos et al., 2001
; Stefanovi et al., 2002
).
Molecular methods
Total genomic DNA was extracted from 
0.1 g of dried leaf material (silica gel-dried or herbarium specimens) using a modified hexacyltrimethylammonium bromide (CTAB) micro-extraction protocol (Doyle and Doyle, 1987
). Total extracted DNA was further purified, without precipitation, using GFX purification kits (Amersham Biosciences, Little Chalfont, UK).
Standard polymerase chain reaction (PCR) procedures were applied to amplify the ITS region together with part of the 17S and 26S gene region. For most taxa, the region was amplified as a single fragment using primers 17SE and 26SE (Sun et al., 1994
). However, for problematic accessions, the region was amplified in two fragments using primers 17SE and ITS2 (White et al., 1990
) to amplify the ITS 1 region and ITS3 (White et al., 1990
) and 26SE to amplify the ITS 2 region. Betaine (1.2 mol/L) was added to prevent the formation of secondary structures in difficult templates and PCR reactions were carried out using a Techgene Thermal Cycler (Techne, Cambridge, UK) using 30 cycles of: 1 min of denaturation at 94°C, 3 min of annealing at 49°C, 1 min of extension at 72°C, and 8 min final extension at 72°C. Amplification products were purified using GFX purification kits following the manufacturer's protocols.
Dideoxy cycle sequencing (28 cycles: 30 of denaturation at 95°C, 15 of annealing at 50°C, 4 min of extension at 60°C) with big dye terminators was performed in 10 µL volumes using a Hybaid Omnigene Thermal Cycler. Excess dye-labelled nucleotides from the sequence reactions were removed by standard ethanol/sodium acetate precipitation. Sequence products were subsequently resuspended and run on an ABI 377 DNA sequencer (Applied Biosystems, Foster City, California, USA).
Sequence alignment and phylogenetic analyses
Sequence data were edited and assembled using Lasergene Navigator (DNAStar, Madison, Wisconsin, USA). Verified sequences were then aligned by eye in Se-Al (version 1.0al; Rambaut, 1996
) prior to phylogenetic analysis. Gaps were not coded as informative characters for analysis because many were overlapping and consequently problematic for coding.
Phylogenetic analyses of the data were performed using PAUP* Beta version 4.0b5 (Swofford, 1998). To determine the effect of outgroup alignment on the ingroup topology two analyses were performed: the first included all accessions whereas the second included only the ingroup taxa (Convolvulus + Calystegia). All parsimony analyses were simultaneous and unconstrained with character state changes unordered and weighted equally (Nixon and Carpenter, 1993
). An heuristic search comprising 1000 random replicates was performed with accelerated transformation (ACCTRAN), saving all minimal trees (MULPARS), and tree bissection and reconnection (TBR) options. Trees were rooted on the most distal outgroup (Jacquemontia mexicana) based on knowledge from previous phylogenetic analyses (Stefanovi et al., 2002
).
The ensemble consistency index (CI; Kluge and Farris, 1969
) and retention index (RI; Farris, 1989
) were calculated in PAUP*. The robustness of clades in the strict consensus tree was evaluated by nonparametric bootstrap analysis (Felsenstein, 1985
) and by computing decay values (Bremer, 1988
). Bootstrap values were determined from 100 bootstrap replicates each comprising 100 random stepwise addition heuristic searches with TBR branch swapping, saving two trees per replicate. Decay analysis was conducted using AutoDecay (Eriksson, 1998
) in conjunction with PAUP*. For each of the decay constraint trees, a heuristic analysis with 10 replicates and TBR swapping was performed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Parsimony analyses
Parsimony analysis of the complete data set (including outgroups) resulted in 672 most parsimonious trees (length = 1232 steps; CI [excluding uninformative characters] = 0.491; RI = 0.800). Analysis of ingroup taxa alone also resulted in 672 trees but of length 719 steps (CI [excluding uninformative characters] = 0.545; RI = 0.869). Both analyses resulted in identical sets of ingroup topologies. Bootstrap and decay values obtained for the ingroup-only analysis were also similar to those obtained from analysis of the complete data set, and as inclusion of outgroups did not affect either the ingroup topology or the general levels of support, in the remainder of this paper, we confine our discussion to the results from the analysis of the entire data set.
Figure 2 shows the strict consensus tree from the complete analysis with bootstrap and decay values indicated and species distributions provided. One phylogram is shown in Fig. 3. The analysis results in a strongly supported group comprising Convolvulus (excluding the New World C. nodiflorus) and Calystegia (hereafter the Convovlulus clade; Fig. 2, bootstrap = 99%, decay value = 17). Within this clade, two clades are resolved corresponding to a clade of perennial nonclimbing species, most of which are shrubs and subshrubs with sessile leaves (Fig. 2, clade A; bootstrap = 71%; decay value = 2) and a clade of annual or perennial herbs and suffrutescent plants typically with a climbing or trailing habit and petiolate leaves (Fig. 2, clade B; bootstrap = 96%; decay value = 8).
|
|
Section I (Table 2) and is resolved within the larger clade of nonclimbing species (clade A). Within clade I, C. floridus (the only species in this clade found on all of the Canary islands) is resolved as sister group to the other two species (bootstrap = 98%; decay value = 3). Clade II (Fig. 2, bootstrap = 90%, decay value = 4) comprises the endemic climbers (Table 2, Section II) together with the Portuguese endemic C. fernandesii. This clade is nested within clade B comprising Convolvulus species that typically possess a climbing or trailing habit. Phylogenetic structure within the Macaronesian clade II is weakly supported but C. volubilis Link. and C. glandulosus (Webb) Sa'ad are resolved as successive sister species to the rest of the clade (Fig. 2, bootstrap values = 79% and 65%, respectively; decay value of one in both cases). From Fig. 3, it is apparent that there are very short branch lengths in much of the Macaronesian clade II. A close relationship with Mediterranean species for both Macaronesian clades is supported in the analysis. Thus, clade I is resolved in a clade with C. dorycnium (E. Mediterranean), C. cneorum (E. Mediterranean), and C. lanuginosus Desr. (W. Mediterranean) (Fig. 2, bootstrap = 95%; decay value = 3), and clade II is resolved as sister group to a clade comprising C. gharbensis Batt. & Pit., C. supinus Coss. & Kral., C. sabatius Viv. (all western Mediterranean), and C. siculus L. (circum-Mediterranean) (Fig. 2, bootstrap = 100%; decay value = 16).
The spiny North African species C. trabutianus, which was considered closely related to the Macaronesian endemic C. caput-medusae by Sa'ad (1967)
, is resolved as sister group to a clade comprising the Western Asian species C. desertii Hochst. and C. prostratus Forsk. (Fig. 2, bootstrap = 100%; decay value = 13). The analysis also suggests that Convolvulus nodiflorus is erroneously placed in the genus as this species is placed in a basal trichotomy with the outgroup Jacquemontia mexicana (Fig. 2, bootstrap = 100%; decay value = 52).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
that the Macaronesian endemic species of Convolvulus are the result of two independent colonizations into the region. Macaronesian Clade I is nested in a derived position within the Convolvulus clade that comprises perennial nonclimbing species, most of which are shrubs and subshrubs with sessile leaves (Fig. 2, clade A). Macaronesian clade II is nested within the clade of annual or perennial herbs and suffrutescent plants typically with a climbing or trailing habit and petiolate leaves (Fig. 2, clade B). It is therefore apparent that the two colonizations of Macaronesia by Convolvulus were from distantly related lineages with distinct morphologies.
Several other plant genera have also undergone multiple independent colonizations into Macaronesia and information on these genera is summarized in Table 1. It should be noted that two of the genera included in Table 1, namely Pulicaria Gaertner and Teline Medikus, have been shown to be nonmonophyletic and that low bootstrap values (<75%) for relevant nodes provide only weak phylogenetic support for multiple introductions in several other genera (see Table 1). Of the remaining examples, however, several demonstrate a similar pattern to Convolvulus (Fig. 2), in which the Macaronesian endemic congeners are distantly related. In Plantago L., for example, the three colonizations occurred from within three different subgenera (Rønsted et al., 2002
), whilst the three island colonization groups of Euphorbia subsect Pachyclade (Molero et al., 2002
) and the two island colonization groups in Solanum L. (Bohs and Olmstead, 2001
) are, in each case, morphologically and phylogenetically distinct. Indeed, some authors have treated one of the Macaronesian colonization groups in Solanum as a separate genus (Normania Lowe) because of its morphological distinctiveness (Bohs and Olmstead, 2001
). In Lavatera L., the two introductions into Macaronesia are not only morphologically distinct but the phylogenetic position of the Macaronesian species suggests that there was one recent and one ancient colonization into the region (Fuertes-Aguilar et al., 2002
). The temporally distinct colonizations hypothesized for these taxa supports the more general conclusion that the Macaronesian region includes both endemics linked to an ancient Tertiary flora and others that have a more recent origin (e.g., Helfgott et al., 2000
; Francisco-Ortega et al., 2001b
). A similar scenario is also consistent with the molecular study of Ilex L. (Manen et al., 2002
) in which the two Macaronesian species may have been the result of colonization events at very different times although the results are weakly supported.
Olea L., Hedera L., and the Asteriscus Miller alliance are the only genera for which a pattern of multiple colonization into Macaronesia by distantly related congeners does not appear to apply. In Hedera, one Macaronesian colonization was from the polyploid clade whilst two other colonizations occurred from within the diploid clade. Within the latter, relationships remain poorly resolved, and it is conceivable that the two colonizations were from closely related lineages. It is notable, however, that ITS data suggest that the three colonizations of Macaronesia by Hedera occurred independently to the Canaries, Madeira, and the Azores. Similarly, in Olea, whilst Macaronesia was colonized independently on two occasions by the same species (O. europaea L.), the two colonizations were to the Canaries and Madeira independently. Thus, whilst the Macaronesian lineages in Olea and Hedera might not be distantly related, in each case they conform to a "checkerboard distribution" (Diamond, 1975
), i.e., they are geographically isolated within the region.
The Asteriscus alliance is the only example to date for which multiple colonization of the same archipelago by closely related congeners is a plausible interpretation of the results of molecular phylogenetic analyses. One introduction resulted in A. graveolens Less. subsp. stenophyllus (Link) Greuter, and the second introduction resulted in a clade comprising A. sericeus (L.fil.) DC. and A. intermedius (dc.) Pit. & Proust. However, some of the relationships of continental and island taxa in this complex remain weakly supported and further data are necessary to robustly establish the number of island colonizations in the group.
Further support for multiple introductions into Macaronesia for some genera comes from the patterns of interspecific hybridization between Macaronesian endemics. Interspecific hybrids are common within genera or groups of genera that have radiated in the islands following a single colonization event (Francisco-Ortega and Santos-Guerra, 2001
). For example, hybrids are known to occur easily among several species of the Aeonium alliance, Argyranthemum, the Bencomia alliance, the Gonospermum Less. alliance, Echium L., Pericallis, Sideritis L., and the woody Sonchus L. alliance (reviewed by Francisco-Ortega and Santos-Guerra, 2001
). Species of these genera form monophyletic groups in Macaronesia and hybrids within each of these assemblages have been detected both in the wild and in cultivation (Hansen and Sunding, 1993
; Francisco-Ortega and Santos-Guerra, 2001
). In contrast, there are no reports for hybrids between members of the two Convolvulus clades, between Ilex canariensis Poir. and I. perado Ait. or between Euphorbia balsamifera Ait. and any of the species of the E. dendroides L. clade even though congeneric species overlap in their distribution range and, in some instances, are cultivated in public gardens of the islands. It is unknown to what extent hybridization occurs between species from different colonization events of Asteriscus, Hedera, Lavatera, Olea, Plantago, Pulicaria, Teline, and Solanum.
A pattern of phylogenetic distinctiveness and/or a checkerboard distribution pattern appears to apply to most, if not all, examples of multiple congeneric endemic Macaronesian plant lineages examined to date. Other genera for which multiple introductions have been hypothesized to account for the Macaronesian endemics include Limonium Mill. (Plumbaginaceae), Salvia L. (Labiatae), and Viola L. (Violaceae) (Fuertes-Aguilar et al., 2002
), Dracaena L. (Agavaceae; Marrero et al., 1998
), and Hypericum L. (Guttiferae; Robson, 1977
). Artemisia L. (Asteraceae), Campanula L. (Campanulaceae), Helianthemum Miller (Cistaceae), and Senecio L. (Asteraceae) also probably conform to this pattern. Nearly all of these genera have a broad continental distribution, are species-rich and ecologically and morphologically diverse and it is conceivable that explicit phylogenetic analyses of many of these groups will demonstrate a similar pattern of phylogenetically distinct congeneric Macaronesian lineages.
Those genera listed in Table 1 that are monophyletic, together with Convolvulus and those genera for which multiple introductions are hypothesized but have yet to be rigorously examined, total approximately 20. This is significantly more than the number of multiple introductions demonstrated for other volcanic oceanic islands, notably Hawaii (one: Rubus L., Howarth et al., 1997
) and the Galapagos Islands (one: Gossypium L., Wendel and Percival, 1990
). However, bearing in mind the close proximity of the Macaronesian islands to the mainland, it represents a small proportion of the total number of genera with endemic Macaronesian representatives (approximately 232; Humphries, 1979
).
The limited number of introductions into island groups such as Hawaii has often been attributed to the extreme difficulty of dispersing to geographically isolated archipelagos during the short period of time such young islands have been in existence (e.g., Carlquist, 1995
). This explanation does not appear to be satisfactory for the Macaronesian region since several islands are of considerable age (e.g., Carracedo, 1994) and many Macaronesian endemic groups appear capable of dispersing over the distances necessary to colonize Macaronesia from the continent. In Convolvulus, for example, the Macaronesian climber clade (Clade II) is distributed in both the Canary Islands and Madeira (Fig. 2), suggesting dispersal over approximately 400 km following initial colonization of the region (Fig. 1). In other genera (e.g., Aichryson Webb & Berth., Euphorbia, Pericallis, and Tolpis that occur in both the Azores and Canaries; Echium and Sonchus that occur in the Cape Verdes and Canaries), dispersal over greater distances is consistent with their distribution in the region. Difficulty of dispersal to the Macaronesian region, particularly for groups with sister groups distributed in the near-continent may therefore be an inadequate explanation for the limited number of introductions to the region.
A more significant factor limiting the number of successful colonizations into the region is likely to be establishment following initial dispersal of propagules to the islands (Carlquist, 1974
) as competitors already in residence may effectively "close" an island system to future potential colonizers (MacArthur and Wilson, 1967
; Whittaker, 1998
). Silvertown (2004)
recently discussed dispersal and establishment barriers in relation to the colonization of Macaronesia. Using published molecular phylogenies of a number of plant groups, Silvertown (2004)
demonstrated that in genera that had undergone a single colonization of the region, the mean number of species per endemic group was seven times greater than that for genera that had undergone multiple colonizations into the region. He therefore hypothesized that groups with many species are more effective at excluding repeat colonizations by relatives than groups with fewer species since larger groups should preempt more of the niche or habitat space available for colonization.
The pattern of relationships of congeneric Macaronesian lineages discussed in this paper is also consistent with this hypothesis: the patterns of phylogenetic distinctiveness or checkerboard distributions of those genera that have undergone multiple introductions into the region support the hypothesis that niche preemption has played an important role in restricting the number of new colonizations into Macaronesia. Multiple congeneric colonization of Macaronesia appears only to have occurred when either (1) different archipelagos within the region have been colonized independently or (2) when the colonizing groups are sufficiently phylogenetically, morphologically, and, presumably, ecologically distinct that they are not in direct competition.
Whilst the sample of genera in Table 1 is limited, it is interesting to note that Macaronesian endemic animal groups are also consistent with this pattern. Phylogenetic analyses have shown that several animal genera have undergone multiple independent colonization into the Macaronesian region namely Calathus Bonelli (Coleoptera; Emerson et al., 1999
), Dysdera Latreille (Aracnida; Arnedo et al., 2001
), Tarentola Gray (Reptilia; Carranza et al., 2002
), and possibly Chalcides Laurenti (Reptilia; Brown and Pestano, 1998
) (although the precise relationships of the Macaronesian representatives of the latter remain to be robustly resolved). In all of these examples, the independent Macaronesian lineages demonstrate a checkerboard distribution and are confined either to different archipelagos or to different islands within an archipelago.
Back-colonization of the continent from Macaronesia
Results presented in this paper place C. fernandesii from Cabo Espichel in Portugal in an unresolved terminal polytomy within the Macaronesian climber group (Fig. 2, Clade II). This is consistent with the view expressed by Pinto de Silva and Teles (1981)
that C. fernandesii forms part of the Macaronesian climber clade. The placement of C. fernandesii in a late branching position in the Macaronesian climber clade and the extremely short branches within this clade (Fig. 3) furthermore support the hypothesis of Pinto de Silva and Teles (1981)
that C. fernandesii is a continental neo-endemic of recent origin, resulting from back-colonization of the continent from Macaronesia (a minimum distance of ca. 1200 km). The traditionally held view of this distribution as relictual (e.g., Bramwell and Bramwell, 2001
) is not supported.
The coastal area near Lisbon where C. fernandesii occurs is among the most floristically diverse areas of the Iberian peninsula (Lobo et al., 2001
). The Portuguese endemic Euphorbia pedroi Molero and Rovira, which is part of the Mediterranean-Macaronesian pachycaul group within Euphorbia (Molero et al., 2002
), also occurs in this area and whilst molecular phylogenetic data have been unable to robustly resolve the relationships within this clade (Molero et al., 2002
), an origin for this species by back-colonization from Macaronesia to Portugal remains a plausible hypothesis.
Recent back-colonization of Portugal from Macaronesia may have been facilitated by the volcanic seamounts located between the Macaronesian archipelagos and the continent. These could have served as stepping stones during glacial periods when sea levels were lower (García-Talavera, 1997
). In the case of the Hawaiian archipelago, palaeogeographic reconstructions have been integrated with phylogenetic hypotheses to gain new insights into the evolution of the Hawaiian biota (Price and Clague, 2002
) and a similar approach for the complex Macaronesian region may shed further light onto the possible role of sea mounts in the colonization history of the Cabo Espichel region.
Sister group relationships of Macaronesian Convolvulus and their systematic implications
A Mediterranean sister group relationship is suggested for both Macaronesian clades in the analysis (Fig. 2). Macaronesian clade I is placed in an unresolved polytomy with C. dorycnium, C. cneorum, and C. lanuginosus, which lends some support to the suggestions of Peter (1891)
and Sa'ad (1967)
that C. floridus and C. scoparius, respectively, are closely related to C. dorycnium. Convolvulus dorycnium is a shrub up to 1 m high, distributed in the eastern Mediterranean whilst C. cneorum is a suffrutescent plant up to 30 cm high, also distributed in the eastern Mediterranean, and C. lanuginosus is a subligneous plant to 30 cm high with a western Mediterranean distribution. The sister group of the Macaronesian erect endemic species is therefore predominantly, though not exclusively, central/eastern Mediterranean. Woodiness may be interpreted as a plesiomorphy for Macaronesian clade I, although C. floridus in particular shows a general increase in size and woodiness when compared with its close continental relatives.
Sa'ad (1967)
, in her revision of Convolvulus from the Canary Islands, Mediterranean, and Near and Middle East, treated C. trabutianus from Morocco as conspecific with the Canarian spiny subshrub C. caput-medusae. The expanded C. caput-medusae sensu lato (s.l.) was placed in section Acanthocladi that comprised all spinescent subshrubs. This suggestion was challenged by Mendoza-Heuer (1983)
, and the results presented in this paper demonstrate unequivocally that C. caput-medusae is more closely related to other Canarian endemics (C. floridus and C. scoparius) than it is to C. trabutianus from Morocco (Fig. 3). The sample of section Acanthocladi included in the analysis suggests that the group is highly polyphyletic (Fig. 3) and that the spines of species from the Canary Islands, the western Mediterranean, and the eastern Mediterranean are non-homologous. Whilst taxon sampling in the present analysis does not allow us to propose a new infrageneric classification of Convolvulus, it does highlight the need for both a careful reexamination of putative morphological homologies and for further work on the infrageneric classification of the group.
Macaronesian clade II, comprising the climbing species, is resolved as sister group to a clade of erect or trailing herbs that are distributed in the western Mediterranean (C. supinus, C. sabatius, and C. gharbensis) or more generally in the Mediterranean and adjacent regions (C. siculus) (Fig. 2). Whilst C. siculus and C. gharbensis are entirely herbaceous, C. supinus and C. sabatius are herbs that are slightly woody at the base. The prominently woody habit of the Macaronesian climbers however, is a synapomorphy for this clade and provides a further clear example of derived insular woodiness (Carlquist, 1974
). The short branch lengths within this clade (Fig. 3) suggest that the Macaronesian climbers may have undergone a recent and rapid radiation.
Choisy (1845)
and Peter (1891)
placed the Macaronesian climbers in section Strophoaulos Don. within an informal group comprising all climbers that were at least slightly woody at the base. The other representatives of this group were New World. Choisy's (1845)
and Peter's (1891)
circumscription of genera within Convolvulaceae differed significantly from currently accepted generic concepts within the family and several species included in section Strophoaulos were later transferred to Jacquemontia Choisy. The only non-Macaronesian species in section Strophoaulos currently retained within Convolvulus is C. nodiflorus (Stearn, 1969
), a robust climber reaching 6 m in height, distributed in tropical America. The results presented in this paper demonstrate that C. nodiflorus is not closely related to the Macaronesian climbers and does not appear to belong to Convolvulus. Indeed, its placement in an unresolved basal trichotomy with Jacquemontia mexicana, the outgroup taxon in the present analysis, suggests that the earlier placement of this species in Jacquemontia may be correct.
Stefanovi et al. (2002)
demonstrated that Convolvulus and Jacquemontia are distantly related in Convolvulaceae. This result contrasted markedly with earlier taxonomic treatments that had considered the two genera closely related in the tribe Convolvuleae. The two genera have been distinguished on the basis of pollen morphology (tricolpate in Convolvulus, pantocolpate in Jacquemontia) and stigma morphology (elongate in Convolvulus, flattened and tongue-like in Jacquemontia) although several species, including C. nodiflorus and C. rozynskii (Standley) Lewis and Oliver, have been placed at different times in either Convolvulus or Jacquemontia. Stearn (1969)
demonstrated that C. nodiflorus possesses tricolpate pollen and as a result, suggested that this species belonged in Convolvulus. However, the molecular data strongly disagree with this placement. Further work is necessary to clarify the circumscription and diagnosis of these two genera.
General patterns of sister-group relationships of the Macaronesian flora
The Macaronesian climber group of Convolvulus is one of a number of groups for which a close relationship with highly disjunct taxa has been hypothesized. Whilst molecular phlylogenetic analyses support several such sister group relationships (see below), others have been strongly refuted (Andrus et al., 2004
), notably the endemic genus Argyranthemum (Francisco-Ortega et al., 1996
), the Bencomia alliance (Helfgott et al., 2000
), the Gonospermum alliance (Francisco-Ortega et al., 2001a
), and Echium (Böhle et al., 1996
) in addition to the Macaronesian climber group of Convolvulus. All of these groups have been shown to have their closest relatives in the Mediterranean. Indeed, molecular phylogenetic analyses suggest that the sister groups of most Macaronesian endemic taxa examined to date have a predominantly Mediterranean distribution. Those genera for which the results of molecular cladistic analyses are consistent both with (1) a single colonization of Macaronesia and (2) a Mediterranean or predominantly Mediterranean sister group for the Macaronesian clade are summarized in Table 3. From this table it can be seen that there is at least some molecular support for a Mediterranean sister group for the endemic Macaronesian clades of 32 genera/alliances.
|
); Pericallis (Asteraceae; New World; 15 species; Panero et al., 1999
); Madeiran Sedum clade (New World; Crassulaceae; two species; van Ham and t'Hart, 1998
); Ocotea Aubl. (South Africa; Lauraceae; one species; Chanderbali et al., 2001
); Phyllis L. (South Africa; Rubiaceae; one species; Andersson and Rova, 1999
); Madeiran species of Saxifraga L. (Eurosiberia; two species; Vargas et al., 1999b
); Canarian species of Silene L. (Eurosiberia; seven species; Clement et al., 1997
). The Macaronesian endemic groups accounted for in Table 1 (25 groups) and Table 3 (32 groups), together with Convolvulus and the seven genera described above, represent 64 Macaronesian groups that have been subject to molecular phylogenetic analyses to date. Of these, however, eight have sister taxa that are either (1) imprecisely known because of conflicting phylogenies (Pinus canariensis), (2) inadequately resolved (Hedera [three clades]; Ranunculus L.) or (3) established on the basis of extremely limited sampling of continental species (Euphorbia [three clades]). Excluding these clades from consideration, the sister-group relationships of 56 endemic Macaronesian clades have been resolved using molecular data. The distributions of the sister taxa of these groups are summarized in Table 4.
|
; Brochmann et al., 1997
; Jardim and Francisco, 2000
; Santos-Guerra, 2001
; Schäffer, 2003
), we estimate that the total number of species endemic to Macaronesia is approximately 874, and the analyses summaried in Table 4 account for 276 of these, i.e., 32% of all Macaronesian endemic species. As can be seen from Table 4, 71% of Macaronesian clades examined have a Mediterranean, or predominantly Mediterranean, sister group, and these clades comprise 80% of the Macaronesian endemic species subject to molecular phylogenetic analysis. Thus, both in terms of number of clades and species diversity, the Mediterranean element of the Macaronesian flora is the dominant element and floristic links with other remote regions account for only a relatively minor component of the endemic plant biodiversity.
In this respect, Convolvulus conforms to the general pattern of relationships of the Macaronesian flora as both Macaronesian clades have Mediterranean sister groups. It is notable however, that the sister taxa of the two Macaronesian Convolvulus clades differ in their areas of distribution within the Mediterranean basin with the sister group of the climber clade (Fig. 2, Clade 2) predominantly western Mediterranean and the sister group of the erect clade (Fig. 2, Clade 1) predominantly eastern Mediterranean in distribution.
Of the clades considered in Tables 3 and 4, 18 (32%) have an exclusively or predominantly western Mediterranean distribution (i.e., Spain, Portugal, Morocco, Algeria, Tunisia, and South France) and 19 (34%) have a sister group distributed in both eastern and western Mediterranean basins, whereas only three (5% of all clades) have an exclusively or predominantly eastern Mediterranean distribution (Hypochaeris L., Vaccinium L., and the erect Convolvulus clade [Fig. 2, clade 1]). Relationships between Macaronesian taxa and those distributed in the western Mediterranean, i.e., in the near-continent, therefore account for 66% of the groups examined and are far more abundant than disjunctions between Macaronesia and the eastern Mediterranean.
The disjunct relationships hypothesized for some Macaronesian groups has been used to support the hypothesis that much of the Macaronesian flora is relictual in nature. However, this view has been challenged by many molecular phylogenetic studies during the last 10 years, and Table 4 illustrates that disjunctions to the Eastern Mediterranean, the New World, southern Africa, eastern Africa, or eastern Asia account for only 19% of Macaronesian groups studied in molecular analyses to date (see also Andrus et al., in press
).
It seems plausible that the Macaronesian flora does indeed comprise lineages that diverged from their closest continental relatives at different times and that some lineages may be ancient (e.g., Fuertes-Aguilar et al., 2002
). Macaronesian lineages with sister taxa distributed in different areas (such as the two Convolvulus lineages) may be the result of temporally distinct waves of colonization into Macaronesia that reflect the complex geological and climatic history of the region. The apparently congruent patterns of sister group relationships with taxa distributed in the near-continent also need close scrutiny as these may conceal "pseudo-congruent" patterns sensu Cunningham and Collins (1994
; see Donoghue and Moore, 2003
). Macaronesian clades with sister groups distributed in the near-continent include both groups that have radiated spectacularly in the islands and evolved distinctive island traits (e.g., Aeonium alliance, Argyranthemum, the Convolvulus climbers, Sideritis) and groups that show little or no evidence of island radiation (e.g., Avena L., Androcymbium Willd., Erucastrum C.Presl.). It is conceivable that the Macaronesia–Mediterranean area relationship specified by these groups may be pseudo-congruent in that they may have resulted from diversification from continental sister groups at different times and, presumably, with different underlying causes.
As Donoghue and Moore (2003)
note, knowledge of the timing of events distinguishes those groups that diversified during the same time period, and therefore might have experienced the same causal events, from those that diversified during different time periods and require different causal explanations. The application of biogeographic approaches that attempt, either explicitly or implicitly, to incorporate temporal information on the diversification of Macaronesian lineages from their continental relatives may provide new insights into the extremely complex patterns of relationships between the Macaronesian islands and continental floras that are well illustrated by groups such as Convolvulus.
|
|
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Aldridge A. E. 1979 Evolution within a single genus: Sonchus in Macaronesia. In D. Bramwell [ed.], Plants and islands, 279–293. Academic Press, London, UK
Alicchio R. L. Aranci L. Conte 1995 Restriction fragment length polymorphism based phylogenetic analysis of Avena L. Genome 38: 1279-1284[ISI]
Andersson L. J. H. E. Rova 1999 The rps16 intron and the phylogeny or the Rubiodeae (Rubiaceae). Plant Systematics and Evolution 214: 161-186[CrossRef][ISI]
Andrus N. J. Trusty A. Santos-Guerra R. K. Jansen J. Francisco-Ortega 2004 Using molecular phylogenies to test phytogeographical links between East/South Africa, Sourthern Arabia and the Macaronesian islands: a review, and the case of Vierea and Pulicaria section Vieraeopsis (Asteraceae). Taxon 53: 333-346[CrossRef][ISI]
Arnedo M. A. P. Oromi C. Ribera 2001 Radiation of the spider genus Dysdera (Araneae, Dysderidae) in the Canary Islands: cladistic assessment based on multiple data sets. Cladistics 17: 313-353[CrossRef][ISI]
Badr A. W. Martin U. Jensen 1994 Chloroplast DNA restriction site polymorphism in Genisteae (Leguminosae) suggests a common origin for European and American lupines. Plant Systematics and Evolution 193: 95-106[CrossRef][ISI]
Barber J. C. J. Francisco-Ortega A. Santos-Guerra K. G. Turner R. K. Jansen 2002 Origin of Macaronesian Sideritis L. (Lamioideae: Lamiaceae) inferred from nuclear and chloroplast sequence datasets. Molecular Phylogenetics and Evolution 23: 293-306[CrossRef][ISI][Medline]
Böhle U. R. H. H. Hilger W. F. Martin 1996 Island colonization and evolution of the insular woody habit in Echium L (Boraginaceae). Proceedings of the National Academy of Sciences, USA 93: 11740-11 745
Bohs L. R. G. Olmstead 2001 A reassessment of Normania and Triguera (Solanaceae). Plant Systematics and Evolution 228: 33-48[CrossRef][ISI][Medline]
Bramwell D. 1972 Endemism in the flora of the Canary Islands. In D. H. Valentine [ed.], Taxonomy phytogeography and evolution, 141–159. Academic Press, London, UK
Bramwell D. Z. I. Bramwell 2001 Wild flowers of the Canary Islands, 2nd ed. Editorial Rueda, Madrid, Spain
Bramwell D. I. B. K. Richardson 1973 Floristic connections between Macaronesia and the East Mediterranean region. Monographiae Biologicae Canariensis 4: 118-125
Bremer K. 1988 The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795-803[CrossRef][ISI]
Bremer K. C. J. Humphries 1993 Generic monograph of the Asteraceae: Anthemideae. Bulletin of the Natural History Museum (London), Botany Series 23: 71-177
Brochmann C. O. H. Rustan W. Lobin N. Kilian 1997 The endemic vascular plants of the Cape Verde Islands, W Africa. Sommerfeltia 24: 1-356
Brown R. P. J. Pestano 1998 Phylogeography of skinks (Chalcides) in the Canary Islands inferred from mitochondrial DNA sequences. Molecular Ecology 7: 1183-1191[CrossRef][Medline]
Carlquist S. J. 1974 Island biology. Columbia University Press, New York, New York, USA
Carlquist S. J. 1995 Introduction. In W. L. Wagner and V. A. Funk [eds.], Hawaiian biogeography: evolution of a hotspot archipelago, 1–13. Smithsonian Institution Press, Washington, D.C., USA
Carracedo J. C. 1994 The Canary Islands: an example of structural control on the growth of large oceanic-island volcanoes. Journal of Volcanology and Geothermal Research 60: 225-241
Carranza S. E. N. Arnold J. A. Mateo P. Geniez 2002 Relationships and evolution of the North African geckos, Geckonia and Tarentola (Reptilia: Gekkonidae), based on mitochondrial and nuclear DNA sequences. Molecular Phylogenetics and Evolution 23: 244-256[CrossRef][ISI][Medline]
Carvalho J. A. A. Culham 1998 Conservation status and preliminary results on the phylogenetics of Isoplexis (Lindl.) Benth. (Scrophulariaceae), an endemic Macaronesian genus. Boletim do Museo Municipal do Funchal Suplemento 5: 109-127
