| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Structure and Development |
School of Botany, University of Melbourne, Parkville, Melbourne, Victoria 3052, Australia
Received for publication March 7, 2003. Accepted for publication October 30, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Drimys • floral development • molecular phylogeny • parametric bootstrapping • SH test • Tasmannia • Winteraceae
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Vink, 1970
). The genus has often been linked with the New World genus Drimys J.R. & G. Forst., and has been considered as a section within Drimys by some authors (Smith, 1943b
; Vink, 1970
, 1988
). This is because both genera have a monopodial growth habit and a flower with a protective calyptra that is shed just prior to anthesis.
Taxonomic controversy surrounding Tasmannia and Drimys has been active for over 50 years (Smith, 1943a
, b
, 1969
; Ehrendorfer et al., 1968
; Vink, 1970
, 1988
, 1993
; Sampson et al., 1988
; Doust, 2000
; Endress et al., 2000
; Karol et al., 2000
). Drimys, as delimited by Smith (1943b)
, was composed of two sections: section Drimys in South America and section Tasmannia in Australia, New Guinea, Malesia, and the Philippines. However, Smith later changed his opinion and proposed elevating the sections to generic status (Smith, 1969
). He was influenced in this decision by the cytological studies of Ehrendorfer et al. (1968)
, who showed that chromosome numbers in Tasmannia are n = 13, but in Drimys sensu stricto and the other genera they are n = 43. Smith (1969)
also pointed out differences in anatomy, morphology, and chemical composition as well as the unisexual and dioecious flowers of Tasmannia. Vink (1970)
disagreed with the elevation of the sections to generic status and refused to recognize the two genera proposed by Smith (1969)
, citing the monopodial construction of the inflorescence of both Drimys and Tasmannia, the common presence of a calyptra that encloses the bud until anthesis, and the fact that some flowers on male plants of Tasmannia show evidence of bisexuality by occasionally forming fruit.
An initial phylogeny of the family using internal transcribed spacer (ITS) sequences of nuclear ribosomal DNA data indicated that Tasmannia and Drimys do not form a monophyletic group (Suh et al., 1993
). Later molecular analysis using ITS nuclear and trnL chloroplast sequence data and a greater number of outgroup taxa showed that Takhtajania Baranova & J. F. Leroy was basal in the family and that Tasmannia was sister to the group of Drimys, Pseudowintera Dandy and Zygogynum Baill. sensu lato (Karol et al., 2000
) (Zygogynum s.l. = Zygogynum Baill. + Bubbia Tiegh. + Exospermum Tiegh. [Vink, 1977
, 1985
]). The clades of both Drimys and Tasmannia had strong bootstrap support, and there were no molecular data to suggest that the two genera could form a monophyletic group. However, a combined analysis of ITS and morphological data suggested that Tasmannia and Drimys formed a monophyletic group (Linder and Crisp, 1995
). A recent morphological analysis also found the two genera formed a single clade and that Takhtajania was part of a clade including Pseudowintera and Zygogynum s.l. (Endress et al., 2000
).
The phylogenetic analyses that group Tasmannia and Drimys together do so because of the shared characters of a monopodial growth habit and a persistent calyptra. Other genera in the family have a sympodial growth habit and floral calyptras that stop growing and are shed or burst early in the development of the floral bud (Doust, 2000
). The shared monopodial growth habit does not extend to shared inflorescence architecture: Drimys bears anthotelic racemose florescences in the axils of the inflorescence bracts whereas Tasmannia bears either single flowers or multiple flowers in ranks borne on a common mound of tissue (Vink, 1970
; Doust, 2000
). The shared persistent calyptra is also more complex than at first glance; developmental studies by Tucker and Gifford (1966)
, Vink (1970
, 1988
), and Sampson et al. (1988)
show that Tasmannia lanceolata and Drimys winteri have lateral sepals, whereas T. xerophila and all other Tasmannia species that have been observed have only medial sepals (Vink, 1970
; Doust, 2000
). Tasmannia lanceolata also has adaxial and abaxial lobes that have been interpreted as either an inner pair of sepals or as variously lobed regions of the calycine calyptra (Tucker and Gifford, 1966
; Vink, 1970
; Doust, 2000
). Tasmannia differs from Drimys, and the other genera in the family, because it is the only genus where flowers are unisexual and dioecious (Vink, 1970
).
The present study focuses on floral development in Tasmannia in relationship to that of Drimys, within an expanded phylogeny of the family. Further species of Drimys and Tasmannia were added to the phylogenetic reconstruction in order to verify that Drimys species form a monophyletic group and to ascertain the placement of Tasmannia lanceolata within the Tasmannia clade. The molecular phylogenies produced were used to test the monophyly of Tasmannia and Drimys (Vink, 1988
), as well as the hypothesis that Tasmannia and Drimys form one monophyletic clade and that Takhtajania, Pseudowintera, and Zygogynum s.l. form another (Endress et al., 2000
). A critical reevaluation of floral development in the light of the present molecular analysis may resolve conflicts between the phylogenies suggested by previously published molecular and morphological data sets and indicate whether Tasmannia should be regarded as being part of Drimys or recognized as a separate genus.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
. The ITS genomic region was amplified using the polymerase chain reaction (PCR), using the leu1 and 4R primers (White et al., 1990
; Malcomber, 2002
). The PCR products of ITS sequences were cloned using the pGEM-T easy vector system (Promega, Madison, Wisconsin, USA) as described in Malcomber (2002)
. Four clones of each ITS DNA accession were amplified using M13 forward and reverse primers. The PCR products for the trnL intron were PCR amplified using primers C and F from Taberlet et al. (1991)
and cleaned by gel purification using the QIAquick PCR purification kit (Qiagen, Valencia, California, USA). Cleaned PCR products from both DNA regions were quantified by comparison with DNA of a known concentration (pGEM 10 and 25 ng; Applied Biosystems, Foster City, California, USA) and fluorescence-labelled using the Big Dye (Applied Biosystems) cycle sequencing protocol. Four primers were used for ITS sequencing, the two plasmid primers t7 and sp6, and two ITS specific internal primers, ITS2 (White et al., 1990
) and ITS3B (Baum et al., 1994
). The trnL region was sequenced using primers C, D, E, and F (Taberlet et al., 1991
). Both forward and reverse DNA strands were sequenced on an ABI 377 automated sequencer (Applied Biosystems). Contig assembly and editing of sequences used Sequencher, version 3.1 (Gene Codes Corporation, Ann Arbor, Michigan, USA). Sequences were a minimum of 85% double-stranded. Sequences were aligned by eye with the aligned data set of Winteraceae trnL and ITS sequences of Karol et al. (2000)
, using MacClade 4.0 (Maddison and Maddison, 2000
). Accession numbers are listed along with specimen voucher information in the supplementary data accompanying the online version of this article. Selection of outgroup taxa was based on previous analyses by Suh et al. (1993)
and Karol et al. (2000)
.
Phylogenetic analyses were performed using both maximum parsimony (MP) and maximum likelihood (ML) algorithms, treating character states as unordered. The best model for each of the maximum likelihood analyses was calculated by Mrmodeltest version 1.1b (Nylander, 2002
), using a neighbor-joining tree. Analyses were conducted using PAUP* version 4.0b10 (Swofford, 1999
), with heuristic searches, tree-bisection-reconnection (TBR) branch swapping, 100 (MP) or 10 (ML) random addition sequence replicates, and gaps treated as missing data. Full heuristic bootstrap analyses (Felsenstein, 1985
) for the parsimony analyses were conducted using 100 (MP) or 1 (ML) random addition sequence replicates and 1000 bootstrapped data sets.
Combinability of the two molecular data partitions was analyzed using ML topologies and the PAUP* implementation of the Shimodaira-Hasegawa (SH) test, with resampling estimated by log-likelihood (RELL) optimization and 1000 bootstrap replicates (Goldman et al., 2000
). Two SH tests were performed, one in each data partition (with likelihood parameters being estimated from the data in that partition). Each test compared the likelihood of (a) the best tree topology for that partition (e.g., for the ITS partition, the best ML ITS tree), (b) the best topology for the other partition (e.g., the best ML trnL topology from the trnL partition), and (c) the best topology for that partition constrained by the topology of the best tree from the other partition (e.g., the best ML ITS tree for the ITS partition, constrained by the trnL tree).
A comparison of trees (a) and (b) is a crude test of congruence of tree topologies generated from different data partitions, and can be analyzed either via an exact likelihood ratio test (Huelsenbeck and Bull, 1996
) or by a resampling-based approach (SH test; Goldman et al., 2000
). However, the results of the comparison of topologies (a) and (b) in the two data partitions via an SH test appeared to be sensitive to differences in the resolution of the trees being compared (see Results), so a comparison of constrained and unconstrained trees calculated from the same data partition (trees [a] and [c]) was also used. This type of comparison has been more commonly used to test how well different topological hypotheses explain the data (Fishbein et al., 2001
; Zanis et al., 2002
), rather than as an explicit test of combinability, although the logic is similar. Such a comparison, based on the same data partition, circumvents the problem of finding significant differences between trees with similar topologies but very different levels of resolution.
An SH test was also used to examine two main morphological hypotheses of evolutionary relationships in Winteraceae. Topologies encoding (a) Drimys and Tasmannia as a monophyletic group (Vink, 1988
); and (b) Drimys and Tasmannia as a monophyletic group and Takhtajania, Pseudowintera, and Zygogynum s.l. as a separate monophyletic group (Endress et al., 2000
) were used to produce constrained ML molecular phylogenies. These were tested against the unconstrained ML phylogeny.
Parametric bootstrapping was employed to further test the specific hypothesis that Drimys and Tasmannia should be considered as separate genera (Huelsenbeck et al., 1996
; Swofford et al., 1996
; SOWH test: Goldman et al., 2000
; Fishbein et al., 2001
; Buckley, 2002
; Zanis et al., 2002
). A fully resolved ML tree, calculated under the constraint of Drimys and Tasmannia being a monophyletic group, was used as the basis for estimating branch lengths under maximum likelihood in PAUP*. The best model for the ML analysis was calculated by Mrmodeltest version 1.1b (Nylander, 2002
), using a neighbor-joining tree. Analyses were conducted using PAUP* version 4.0b10 (Swofford, 1999
), with heuristic searches, TBR branch swapping, 10 random addition sequence replicates, and gaps treated as missing data. Seq-Gen version 1.2.6 (Rambaut and Grassly, 1997
) was then used to simulate 100 data sets using the topology and branch lengths from the constrained tree. Each of these data sets was analyzed in PAUP* without constraint and with the constraint of the tree from which the data sets were simulated. The difference in tree length between constrained and unconstrained trees for each simulated data set was calculated and used to form a null probability distribution. The difference in tree length between trees with Drimys and Tasmannia as polyphyletic or as a monophyletic group was calculated from the original data set and compared against the null probability distribution.
Developmental characters, defined by comparison of developmental sequences (see below), were optimized on each of the most parsimonious trees using MacClade 4.0 (Maddison and Maddison, 2000
). In cases where character states were equivocal at a node, all equally parsimonious reconstructions were examined.
Morphological and developmental analysis
Previous work has shown that there are two main morphological groupings in Tasmannia (Smith, 1943a
; Vink, 1970
; Sampson et al., 1988
). One of these consists only of Tasmannia lanceolata, whose first pair of sepals is arranged laterally, while the other group comprises all other species examined and has two sepals arranged medially (adaxially and abaxially). Material at all stages of development of Tasmannia lanceolata (Poir.) A.C. Smith and one species from the other group, T. xerophila (P. Parm.) Gray were collected for this study from populations at around 800–1000 m a.s.l. in the Great Dividing Range northeast of Melbourne, Australia. Voucher specimens were collected from each plant and are stored in the herbarium at the University of Melbourne (MELU). Fixation, dissection, and scanning electron microscope (SEM) analysis are as detailed in Doust (2000
, 2001
).
Following the arguments of Hufford (1995)
we regard the development of morphological structures as only observable through a succession of instantaneous morphologies (ontogenetic states), such as those provided by SEM micrographs. The description and comparison of ontogenetic states provides an alternative to the view of ontogeny as passing through a prescribed and rigid series of developmental stages. This is necessary because developmental events are often disassociated during morphological diversification of related taxa (heterobathmy; Takhtajan, 1991
), so that recognizable and homologous ontogenetic states can potentially arise at spatially or temporally different locations in the developmental process (Hufford, 1995
; Doust and Kellogg, 2002b
). The molecular phylogeny can be used as a guide to selecting phylogenetically relevant comparisons of ontogenetic states and in inferring direction of character state evolution (Doust and Kellogg, 2002a
, b
).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
). Maximum likelihood analyses gave two trees with the same likelihood value (to five decimal places). The topologies of the two trees were identical apart from small reciprocal changes in branch lengths. One of the maximum likelihood trees is illustrated in Fig. 1. The relationships suggested by the maximum likelihood trees are the same as for the maximum parsimony trees. Bootstrap values from both maximum parsimony and maximum likelihood analyses were similar and are noted below.
|
, with Takhtajania sister to the other members of the family. The next clade to diverge is formed by the six species of Tasmannia, which are sister to a clade comprising Drimys, Pseudowintera, and the species of Zygogynum, Exospermum, and Bubbia (Zygogynum s.l., Vink, 1985
). The clade of Tasmannia species has high bootstrap support (M 100%, ML 98%), as does the sister clade of Drimys, Pseudowintera, and Zygogynum s.l. (MP 98%, ML 89%). The two Drimys accessions form a well-supported clade (MP 100%, ML 98%), as do those of Pseudowintera (MP 99%, ML 99%) and Zygogynum s.l. (MP 100%, ML 100%).
Testing alternative hypotheses of relationships
A variety of methods were used to investigate alternative hypotheses of relationships that had been suggested from morphological analyses (Vink, 1988
; Endress et al., 2000
). In particular, the relationship of Drimys and Tasmannia in the molecular trees is very different to that found in morphological trees, where the monopodial growth habit and the persistence of the calyptra up until anthesis have been used to unite the two genera. We looked at the trees calculated from the bootstrapped data sets and found that only five out of 11 763 trees in the MP analysis (0.043%) and 0 out of 1173 trees in the ML analysis (0%) had Drimys and Tasmannia as a monophyletic group. Thus there is little support from the bootstrapped data sets for Drimys and Tasmannia forming a clade.
To further test the placement of Drimys and Tasmannia, the SH test was used to examine the hypotheses: (a) that Drimys and Tasmannia should form a monophyletic group (Vink, 1988
); and (b) that Drimys and Tasmannia form one monophyletic group and Takhtajania, Pseudowintera, and Zygogynum s.l. form a second monophyletic group (Endress et al., 2000
). These hypotheses were used as topological constraints to construct constrained ML trees, which were compared to the unconstrained ML trees. Hypothesis (b) was significantly different from the unconstrained tree and therefore can be rejected as an adequate description of the data (Table 2). However, the probability for hypothesis (a) being significantly different was P = 0.070, a value that is notable but does not allow us to confidently reject the null hypothesis (Table 2).
|
). The difference in log-likelihood between constrained and unconstrained trees calculated from the original data matrix was –lnL = 8.029 84, which falls outside the probability distribution calculated from the resampled data sets (largest difference –lnL = 7.505 33). The difference between the constrained and unconstrained trees is therefore highly significant (P < 0.01). However, this result needs to be taken with some care as Buckley (2002)
has shown that parametric bootstrapping can produce a high Type I error (probability of false positive results) if the model of sequence evolution is incorrect. However, the results of the three methods of analysis make it unlikely that Drimys and Tasmannia form a monophyletic group.
Morphology and development
All species of Tasmannia have monopodial shoot systems that produce leaves with axillary vegetative buds during the vegetative phase and bracts with axillary floral buds in the reproductive phase. The inflorescence is a raceme terminated by a vegetative bud, and after the production of the flowers, production of leaves recommences (Fig. 2A). In common with most other genera in the family, species of Tasmannia have flowers whose free organs show considerable variation in arrangement (Vink, 1970
; Doust, 2000
). All species are dioecious, with unisexual flowers, although bisexual flowers have rarely been observed (as evidenced by the production of fruit in otherwise staminate flowers; Vink, 1970
). The species included in this study usually bear only one flower per floral bract, but other species of Tasmannia may have up to 11 flowers per floral bract [various forms of Tasmannia (Drimys) piperita from Papua New Guinea; Vink, 1970
].
|
). An alternative interpretation would be to interpret the first and second pair of organs in T. lanceolata and the first pair of organs in T. xerophila as prophylls, and the inner petaloid organs as tepals. In other members of the Winteraceae the inner series of organs are more tepaloid, as they can form a gradation from thick outer to thin inner petaloid structures. However, in Drimys and Tasmannia, all of the inner series of organs are alike and petaloid. It is not possible to definitely differentiate between sepals and petals vs. prophylls and tepals for flowers in Winteraceae, but the particular interpretation adopted does not affect the comparisons of floral development detailed below. It may be that patterns of gene expression will be able to differentiate between prophylls and sepals, but in the absence of such evidence we have labelled the first initiated organs as sepals rather than prophylls and the inner series of organs as petals.
Tasmannia xerophila
During development the floral buds are protected by a number of tightly enclosing straw-colored floral bracts, with the most basal bracts being the largest. These bracts abscise just before anthesis. In more distal positions on the inflorescence all of the flowers are solitary but towards the base of the inflorescence some of the axillary floral buds produce a bracteole or two on their pedicels (Fig. 3A). In the very lowermost positions these bracteoles may themselves have axillary buds, at least some of which are floral buds. Staminate flowers produce stamens and a nonfunctional carpel; pistillate flowers produce a number of lateral carpels. The processes of sepal and petal initiation are similar in both staminate and pistillate flowers.
|
Soon after the initiation of the petals the sepals enclose the floral apex, the adaxial sepal overlapping the abaxial sepal (Fig. 3E). The extent of this overlap can be seen when the adaxial sepal has been removed (Fig. 3F). The sepal tips are then borne aloft by the action of an intercalary meristem, which forms the tube of the calyptra. Growth of the calycine calyptra keeps pace with the development and growth of the developing bud. At anthesis the calyptra splits into two halves, the splits appearing between the adaxial and abaxial sepal positions. The two halves of the calyptra then reflex to a horizontal position (Fig. 2B).
Petals are initiated soon after the sepals, before the floral apex is covered by the growth of the calyptra. Two petals are initiated, one at each pole of the elliptic floral apex, lateral to the placement of the sepals and to the subtending floral bract (Figs. 3C, D, 4A). The floral apex is at this stage approximately 50 µm high and has a projected surface area of approximately 2.0 x 104 µm2. Further petals may be initiated, appearing in either an abaxial or adaxial position (Fig. 4H).
|
In pistillate flowers there are generally no stamens or staminodes, and carpel initiation directly follows on from the initiation of the petals. There are between three and 10 carpels laterally initiated on the floral apex; the first four are initiated on either side of the petal primordia, towards the poles of the elliptic apex (Fig. 4E, F) and further carpels are initiated to fill in the remaining space on the meristem. The placement of the carpels can vary in both height and transverse position on the apex (Fig. 4G). In some flowers, primordia are differentiated into petals in some of the positions where carpels would otherwise differentiate (Fig. 4H). The carpel develops as an open structure until fusion of the edges occurs (Fig. 4I). Later in development a stipe starts to grow beneath the region of the locule, elevating it above the receptacle.
Tasmannia lanceolata
During development, the floral buds are tightly enclosed by red floral bracts; basal bracts are shorter than more distal ones. Floral buds are usually single in the axil of each bract. As in T. xerophila, staminate flowers produce stamens and a nonfunctional carpel; however, pistillate flowers generally produce only a single terminal carpel. The processes of sepal and petal initiation are similar in both staminate and pistillate flowers.
The first two sepals initiated are at the poles of the elliptic floral apex and, unlike T. xerophila, are oriented laterally to the subtending bract (Fig. 5A, B). The floral primordium is at this stage 100–130 µm high and has a projected surface area between 1.1 x 104 and 2.0 x 104 µm2. Two more sepals are initiated adaxially and abaxially soon after the first two sepals (Fig. 5C). The adaxial sepal is the more prominent of this second pair of sepals and has a more or less triangular shape; the abaxial sepal is often no more than a line of tissue with no free portion, although it occasionally is more pronounced (Fig. 5F). Both lateral and medial sepals are the result of unequal growth of the calyx rim. The continued growth of the calyx occurs via an intercalary meristem, as in T. xerophila, so that a calycine calyptra is formed, bearing the sepal tips at its apex. Some further growth of the two medial sepal tips may also occur (Fig. 5F). In most flowers, the presence of the adaxial and abaxial sepal tips continues to separate the lateral sepal tips throughout the growth of the bud. At anthesis, the calyx calyptra splits in two and the halves reflex. The dehiscence line of the calyptra is between the adaxial and abaxial sepals and bisects the lateral sepals.
|
Stamens are generally found only in the staminate flowers, although two examples were found of a single stamen being initiated in a pistillate flower. Stamen initiation commences before the calyptra has enclosed the floral apex (Fig. 6A), and stamen primordia appear first towards the lateral poles of the more or less elliptic floral apex although they are very soon initiated on the abaxial and adaxial sides of the apex as well (Fig. 6B). The stamens can be positioned in whorls or spirals (Fig. 6C, D).
|
In pistillate flowers the floral apex continues to grow after petal formation and may become more circular in cross-section (Fig. 6E). The initiation of the single terminal carpel uses up the entirety of the floral apex. During differentiation of the carpel, two parallel ridges appear, oriented in a roughly medial plane, creating a furrow or cleft at the apex of the carpel (Fig. 6F). The ridges continue growth, but the dorsal ends of the ridges together with the dorsal bridge between them grows at a faster rate than the ventral ends of the ridges (Fig. 6G). For most of its development the carpel grows as an open U-shaped structure, which finally fuses by the intertwining of the papillae at the margins. However, there is a short section towards the base of the carpels where the carpel grows as a completely closed ring. This ring of tissue can be seen as a zone between where the carpel starts to swell from the top of the floral meristem and where the lower edge of the cleft begins (Fig. 6H, I). At maturity the carpel has a stigmatic crest running almost to the base of the carpel on the ventral side (Figs. 2C, 6H), and the distinction between the stipe and the cross-zone is obscured (Fig. 7I).
|
Evolution of floral development
Figure 7 illustrates the developmental sequence of the two Tasmannia species and of terminal and lateral flowers of Drimys winteri. Developmental time is relative, being confounded with size, so that increased size may be due to longer or faster periods of growth. In this diagram size is used as a proxy for time and the lengths of the lines connecting the various defined developmental stages signify the amount of time/size change that occurs between stages. Wavy lines show where time/size has increased relative to other taxa. The approximate height of the floral primordium at sepal initiation and the half-ellipsoid surface area (area = basal perimeter length multiplied by height) of the floral meristem at petal initiation stages are indicated. There are a number of points throughout the developmental trajectory of these species where differences arise. An early difference is that T. xerophila achieves a larger floral primordium size and is twice the height of the other taxa before sepals are initiated and that the sepals of T. xerophila are initiated medially rather than laterally. The increased floral primordium height in T. xerophila may have facilitated the novel substitution of medial for lateral sepals. Another difference at sepal initiation is that the terminal flowers of D. winteri inflorescences initiate a ring meristem whereas lateral flowers, that have elliptic floral meristems, initiate two lateral sepals (Doust, 2000
, 2001
). This suggests that formation of lateral sepals may be linked to shape of the meristem. The lateral sepals in T. lanceolata are very similar to those in the lateral flowers of D. winteri, and an occasional lobe which resembles a medial sepal can also be found in D. winteri. Thus T. lanceolata and D. winteri are developmentally similar in these early ontogenetic stages.
Many of the developmental changes mentioned above are autapomorphies, yet the extended growth of the calyptra can be optimized onto the phylogeny either as a gain each in the lineages leading to Tasmannia and Drimys or a gain on the branch between Takhtajania and Tasmannia, followed by a loss on the branch between Drimys and Pseudowintera (Fig. 8).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
[ITS]; Karol et al., 2000
[ITS + trnL]). Additional taxa of Tasmannia and Drimys do not change the basic phylogenetic relationships that had been described by those studies.
These molecular phylogenies conflict with hypotheses of relationships between Drimys and Tasmannia based on morphological data (Vink, 1993
; Endress et al., 2000
). The SH test indicated the most constrained hypothesis, that of a tree with Drimys + Tasmannia as one monophyletic group and Takhtajania + Pseudowintera + Zygogynum s.l. as another, is significantly less likely than the unconstrained molecular tree. The less constrained hypothesis, that Drimys + Tasmannia form a monophyletic group, was notably less likely than the unconstrained tree on the SH test (P = 0.070), but not significant at the P < 0.05 level. However, less than one-tenth of one-percent of MP trees and no ML trees from the pseudoreplicated bootstrap data sets had Drimys and Tasmannia grouped together. As well, the results of the parametric bootstrap test showed that topologies where the two genera were constrained to be monophyletic were significantly less likely than topologies where the two genera were separate. These methods offer different insights into the topologies suggested by the data, but support the topologies that were consistently produced in the phylogenetic analysis, and the finding that Tasmannia and Drimys should be regarded as separate genera rather than a monophyletic group. It is possible that the marginal significance for the SH test for this hypothesis may reflect a lack of power due to too few informative characters, a deficiency that could be resolved by adding more molecular data.
Development
The patterns of development show that there is a mixture of similarities and differences between Tasmannia lanceolata, T. xerophila, and Drimys winteri (Fig. 7). In fact, the only consistent similarity between the two genera is the extended growth of the calyptra, which has two equally parsimonious optimizations onto the molecular topology (Fig. 8). Consistent developmental differences between the two genera include the delay in petal initiation in Drimys winteri and the divergence of development into unisexual male and female flowers in Tasmannia. However, there are also marked differences between the two species of Tasmannia, both in the size of the floral primordium when sepals initiate, the position where sepals first initiate, the number of sepals initiated, and the number and placement of carpels (several and lateral in T. xerophila, single and apparently terminal in T. lanceolata). Sepal initiation in T. lanceolata is more similar to the lateral flowers of Drimys winteri, yet this pattern is found throughout the family and does not provide evidence for monophyly for the two species (Doust, 2000
).
Tucker and Gifford (1966)
state that two lateral sepals are initiated in T. lanceolata, and that the margins of the two sepals eventually become appressed and fused edge to edge, so that the lower portions grow as a cylinder. No evidence was found in this study for sepal fusion, and the growth of the calyptra as a collar, bearing the free sepals aloft, appears to better describe the development of the calyptra. Vink (1970)
considered that only two lateral sepals are initiated but that it was a matter of opinion whether the adaxial and abaxial lobules were interpreted as a second pair of sepals or a result of the unequal growth of the ridges connecting the lateral sepals. However, as the lateral sepals in all genera are themselves the result of unequal growth this does not preclude the medial sepals likewise being so. The developmental analysis shows that adaxial and abaxial structures are more appropriately interpreted as sepals because the placement of the first whorl of petals alternates with the positions of both sepal pairs.
The shared position of the medial sepals in T. lanceolata and T. xerophila might be considered as evidence for their homology, although their appearance and manner of initiation are quite different. In T. lanceolata the medial sepals are the second set of primordia initiated and are relatively narrow and at approximately equal heights on the floral apex, whereas in T. xerophila they are the first set of primordia initiated and are wide and at different heights on the floral meristem. Thus it seems unlikely that the two forms of medial sepals are homologous. More sampling of the morphological diversity within Tasmannia is needed to shed light on the possible homology of the two forms of medial sepals.
In both species of Tasmannia the position of initiation of further petals, stamens, and carpels depends on the position of those already present. In a number of cases, more petals are differentiated than is the norm and are found where carpels or stamens would normally have been differentiated. This indicates that organ identity may be somewhat flexible in the flower but that organ position is less so. This is also evidenced by the conserved position yet changed organ identity of stamens and carpels in male and female flowers of Tasmannia xerophila. In later flower development in both species the initially regular whorled patterning evident in early flower development may become more irregular. This is most noticeable in male flowers, because more and smaller organs (stamens) are initiated in male as opposed to female flowers. An explanation of the irregular arrangement of floral organs in Tasmannia was first attempted by Vink (1970)
, who attributed the lack of regularity to the influence of an uneven base to the flower. Vink also noted that sepals and petals in Tasmannia were more regularly arranged and considered it likely that sepal and petal arrangement were under separate genetic control from that of stamens and/or carpels. However, there is no need to posit two sets of control factors for organ arrangement, because the smaller size of the stamen and carpel primordia makes it easier for disturbances in arrangement to emerge in that part of the flower. It is likely that irregularities in organ arrangement are due to irregularities in floral meristem shape, as shown for Drimys winteri (Doust, 2001
).
The differences in sepal initiation and floral organ arrangement between Drimys and most species of Tasmannia do not support grouping of the two into a single genus. The initiation of lateral sepals in T. lanceolata is similar to D. winteri, yet later patterns of floral organ initiation are more similar to other species of Tasmannia. The production of a calyptra tube by an intercalary meristem occurs throughout the family, and it is only its continued growth that provides a morphological link between Tasmannia and Drimys. The developmental evidence presented above makes it likely that the continued growth of the calyptra has been separately derived in each of the two lineages. Developmental analysis reveals little evidence for uniting Drimys and Tasmannia, but rather supports the molecular phylogenetic findings that they are better regarded as separate monophyletic genera.
|
FOOTNOTES |
|---|
2 Current address: Department of Biology, University of Missouri—St Louis, 8001 Natural Bridge Road, St Louis, Missouri 63121 USA. adoust{at}umsl.edu
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Buckley T. R. 2002 Model misspecification and probabilistic tests of topology: evidence from empirical data sets. Systematic Biology 51: 509-523[CrossRef][ISI][Medline]
Doust A. N. 2000 Comparative floral ontogeny in Winteraceae. Annals of the Missouri Botanical Garden 87: 366-379[CrossRef][ISI]
Doust A. N. 2001 The developmental basis of floral variation in Drimys winteri (Winteraceae). International Journal of Plant Sciences 162: 697-717[CrossRef]
Doust A. N. E. A. Kellogg 2002a Inflorescence diversification in the panicoid "bristle grass" clade (Paniceae, Poaceae): evidence from molecular phylogenies and developmental morphology. American Journal of Botany 89: 1203-1222
Doust A. N. E. A. Kellogg 2002b Integrating phylogeny, developmental morphology and genetics: a case study of inflorescence evolution in the ‘bristle grass' clade (Panicoideae, Poaceae). In Q. Cronk, R. Bateman, and J. Hawkins [eds.], Developmental genetics and plant evolution, 298–314.Taylor & Francis, London, UK
Ehrendorfer F. F. Krendl E. Habeler W. Sauer 1968 Chromosome numbers and evolution in primitive angiosperms. Taxon 17: 337-468[CrossRef]
Endress P. K. A. Igersheim F. B. Sampson G. E. Schatz 2000 Floral structure of Takhtajania and its systematic position in Winteraceae. Annals of the Missouri Botanical Garden 87: 347-365[CrossRef][ISI]
Felsenstein J. 1985 Confidence limits of phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Fishbein M. C. Hibsch-Jetter D. E. Soltis L. Hufford 2001 Phylogeny of Saxifragales (Angiosperms, Eudicots): analysis of a rapid, ancient radiation. Systematic Biology 50: 817-847[CrossRef][ISI][Medline]
Huelsenbeck J. P. J. J. Bull 1996 A likelihood ratio test to detect conflicting phylogenetic signal. Systematic Biology 45: 92-98[CrossRef][ISI]
Huelsenbeck J. P. D. M. Hillis R. Neilsen 1996 A likelihood-ratio test of monophyly. Systematic Biology 45: 546-558[CrossRef]
Hufford L. 1995 Patterns of ontogenetic evolution in perianth diversification of Besseya (Scrophulariaceae). American Journal of Botany 82: 655-680[CrossRef][ISI]
Giussani L. M. J. H. Cota-Sanchez F. O. Zuloaga E. A. Kellogg 2001 A molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis. American Journal of Botany 88: 1993-2012
Goldman N. J. P. Anderson A. G. Rodrigo 2000 Likelihood-based tests of topologies in phylogenies. Systematic Biology 49: 652-670[CrossRef][ISI][Medline]
Karol K. G. Y. Suh G. E. Schatz E. A. Zimmer 2000 Molecular evidence for the phylogenetic position of Takhtajania in the Winteraceae: inference from nuclear ribosomal and chloroplast gene space sequences. Annals of the Missouri Botanical Garden 87: 414-432[CrossRef][ISI]
Linder H. P. M. D. Crisp 1995 Nothofagus and Pacific biogeography. Cladistics 11: 5-32[CrossRef][ISI]
Maddison D. R. W. P. Maddison 2000 MacClade 4: analysis of phylogeny and character evolution, version 4.0. Sinauer and Associates, Sunderland, Massachusetts, USA
Malcomber S. T. 2002 Phylogeny of Gaertnera Lam. (Rubiaceae) based on multiple DNA markers: evidence of a rapid radiation in a widespread, morphologically diverse genus. Evolution 56: 42-57[ISI][Medline]
Nylander J. A. 2002 Mrmodeltest version 1.1b. Program distributed by the author. Department of Systematic Zoology, EBC, Uppsala University, Sweden, http://www.ebc.uu.se/systzoo/staff/nylander.html
Rambaut A. N. C. Grassly 1997 Seq-Gen: an application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees. Computational Applied Biosciences 13: 235-238
Sampson F. B. J. B. Williams P. S. Woodland 1988 The morphology and taxonomic position of Tasmannia glaucifolia (Winteraceae), a new Australian species. Australian Journal of Botany 36: 395-413[CrossRef]
Smith A. C. 1943a Taxonomic notes on the old world species of Winteraceae. Journal of the Arnold Arboretum 24: 119-164
Smith A. C. 1943b The American species of Drimys. Journal of the Arnold Arboretum 24: 1-33
Smith A. C. 1969 A reconsideration of the genus Tasmannia (Winteraceae). Taxon 18: 286-290[CrossRef]
Suh Y. L. B. Thien H. E. Reeve E. A. Zimmer 1993 Molecular evolution and phylogenetic implications of internal transcribed spacer sequences of ribosomal DNA in Winteraceae. American Journal of Botany 80: 1042-1055[CrossRef][ISI]
Swofford D. L. 1999 PAUP*. Phylogenetic analysis using parsimony and other methods, version 4. Sinauer Associates, Sunderland, Massachusetts, USA
Swofford D. L. G. J. Olsen P. J. Waddell D. M. Hillis 1996 Phylogenetic inference. In D. M. Hillis, C. Moritz, and B. K. Mable [eds.], Molecular systematics, 407–514. Sinauer, Sunderland, Massachusetts, USA
Taberlet P. L. L. Gielly G. Pautou J. Bouvet 1991 Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105-1109[CrossRef][ISI][Medline]
Takhtajan A. 1991 Evolutionary trends in florwering plants. Columbia University Press, New York, New York, USA
Tucker S. C. E. M. Gifford 1966 Carpel development in Drimys lanceolata. American Journal of Botany 53: 671-678[CrossRef][ISI]
Vink W. 1970 The Winteraceae of the Old World I. Pseudowintera and Drimys—morphology and taxonomy. Blumea 18: 225-354
Vink W. 1977 The Winteraceae of the Old World II. Zygogynum—morphology and taxonomy. Blumea 23: 219-250
Vink W. 1985 The Winteraceae of the Old World V. Exospermum links Bubbia to Zygogynum. Blumea 31: 39-55
Vink W. 1988 Taxonomy in Winteraceae. Taxon 37: 691-698[CrossRef][ISI]
Vink W. 1993 Winteraceae. In K. Kubitzki, J. G. Rohwer, and V. Bittrich [eds.], The families and genera of vascular plants, vol. II, Flowering plants—Dicotyledons, 630–638. Springer-Verlag, Berlin, Germany
White T. J. T. Birns S. Lee J. Taylor 1990 Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In M. Innis, D. Gelfand, J. Sninsky, and T. White [eds.], PCR protocols: a guide to methods and applications, 315–322. Academic Press, San Diego, California, USA
Zanis M. J. D. E. Soltis P. E. Soltis S. Mathews M. J. Donoghue 2002 The root of the angiosperms revisited. Proceedings of the National Academy of Sciences, USA 99: 6848-6853
Zharkikh A. 1994 Estimation of evolutionary distances between nucleotide sequences. Journal of Molecular Evolution 39: 315-329[CrossRef][ISI][Medline]
| ||
