| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721
Received for publication February 17, 1998. Accepted for publication December 8, 1998.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Boswellia • Bursera • Burseraceae • burseras • Commiphora • copal • cuajiotes • evolutionary trends • ITS sequences • phylogenetic analysis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
100 species distributed from the southwestern United States to Peru. It reaches its maximum diversity in the Pacific slopes of Mexico where 80 species occur and 70 are endemic (Rzedowski and Kruse, 1979
; Daly, 1993
).
Burseras are typically low or medium-size trees, and some are also small shrubs or large trees. Many species have succulent trunks colored blue, yellow, green, red, or purple. The outer bark of some species exfoliates in papery flakes or sheets of bright colors often different from that of the trunk. The genus is notable for its terpenoid secretions and exudates (Mooney and Emboden, 1968
; Becerra and Venable, 1990
; Becerra, 1994,
1997
). These secretions arise from a system of resin canals found in the cortex of the trunk and in the leaves (Guillaumin, 1909
) and are known to provide chemical defense against specialized herbivores (Becerra and Venable, 1990
). The shape, size, and morphology of the leaves are highly variable among species. They are alternate and usually once- or twice-pinnate, but some are unifoliolate or trifoliolate. Leaves range from <1 to 50 cm long. Leaflet size and shape are also highly variable. All species are drought deciduous and usually flower at the end of the dry season. Most burseras are dioecious or polygamo-dioecious, although a few are apomictic or hermaphrodite. The actinomorphic flowers are usually less than a centimetre long and are arranged in inflorescences that may include as few as one or up to 70 or more flowers. The flower color is usually pale yellow, although in some species it is wine-red. Most species are generalists in terms of pollinators, being visited by bees, flies, and small wasps. The predominantly bird-dispersed fruits are small, single-seeded dehiscent drupes with the seed partially or completely covered by a cream to red pseudoaril.
In Mexico, Bursera is one of the most important physiognomic components of the tropical dry forests. In many places, like the depression of the Balsas River in Michoacán and Guerrero or the dry canyons of the Tehuantepec River in Oaxaca, the genus becomes the dominant or codominant woody taxon, surpassing the legumes in diversity and abundance (Miranda, 1947
; Rzedowski, 1978
). They are also conspicuous constituents of habitats like desertscrub and thornscrub of the Mexican central and northern deserts and occur to a lesser extent in lowland tropical rain forests and higher altitude woodland. As is true of other Burseraceae, Bursera is a good indicator of relatively intact natural vegetation, because few species occur in secondary growth (Rzedowski and Kruse, 1979
; Daly, 1992
).
Because of its importance in the Mexican vegetation, the genus has been the subject of taxonomic studies continuously since the end of last century. Nevertheless, its classification is still problematic and its phylogenetic relationships are not well understood. Part of the problem is the scarcity of herbarium specimens with flowers. Since the small, superficially uniform flowers appear in the spring before the macroscopically more variable leaves, it is unusual to have matching specimens of flowers and leaves. As a result, the taxonomy of the group has been largely inferred from leaf morphology and bark characteristics, and to a lesser extent by fruit morphology, the morphology of the cotyledons and resin chemical profiles. However, in other groups of plants, bark and, in particular, leaf characteristics have proven to be very labile and often unreliable for clarifying phylogenetic affiliations (Stebbins, 1950
; Donoghue and Sanderson, 1992
). Also, some species like B. schlechtendalii, B. bipinnata, B. paradoxa, and B. tecomaca have very distinctive leaf characteristics making it difficult to determine their affinities with other species using leaves. Becerra (1997)
has also shown that there is considerable convergence in the presence and absence of particular terpenes. Thus, this is a group for which molecular data might prove especially useful for determining phylogenetic affiliations.
The purpose of this investigation was to infer the evolutionary relationships within Bursera and between Bursera and related genera. To do this, we have conducted a phylogenetic analysis of nucleotide sequences of the internal transcribed spacer (ITS) region of the nuclear ribosomal DNA from 57 species and varieties of Mexican Bursera and five outgroup taxa.
Systematics of Bursera
The Burseraceae includes 20 genera and >600 species of trees and shrubs from the subtropical and tropical regions of Africa, Asia, North America, and South America. Having resin ducts containing aromatic terpenes and essential oils is characteristic of the family (Standley, 1923
; Gillett, 1991
; Rzedowski and Guevara-Féfer, 1992
). Recent studies using molecular and anatomical data suggest that the sister family to the Burseraceae is the Anacardiaceae (Terrazas-Salgado, 1994
). Bursera, together with the tropical genera Triomma, Aucoumea, Boswellia, Commiphora, and Beiselia form the tribe Bursereae or Boswelliae (Engler, 1931
; Lam, 1932
; Rzedowski and Kruse, 1979
; Forman et al., 1991
), which is considered to be an advanced group within the Burseraceae. Bursera is distinguished from the rest of the family by having a valved, one-seeded, dehiscent drupe and a calyx that opens in the bud (Gillett, 1980
, 1991
).
The genus has been divided into two subgenera or sections (McVaugh and Rzedowski, 1965
; Rzedowski, 1968
; Becerra, 1997
). Section Bullockia includes species with a trivalvate fruit, four merous flowers, branch rosettes with cataphylls, and grayish-red or grey rough bark. The outer layers of the bark are, in most cases, rough and do not separate in sheets. Toledo (1982)
further divided this section into two groups: the species whose fruits are completely or almost completely covered by the pseudoaril and the species in which the pseudoaril covers the fruit only partially. In species of section Bursera, flowers are three, four, and five merous (often five merous in the male and three merous in the female), fruits are bivalvate, leaves have no cataphylls, and the colorful bark peels off in papery sheets, giving rise to the common name "cuajiote," which means "leprous tree" in Nahuátl. There are three distinctive groups of species in this section according to Toledo (1982)
. In the first group (the mulatos), the cotyledons are trilobate as in the species of section Bullockia and trees have red-exfoliating bark. The second group, the red-barked cuajiotes, includes species with multilobate cotyledons and red-exfoliating bark. In the last group, the yellow-barked cuajiotes, the cotyledons are also multilobate, but the bark exfoliates in yellow sheets.
Since the early studies, evidence has suggested that within the Burseraceae Bursera is very closely related to the genus Commiphora and also to Boswellia, but the relationships among these three genera are not clear. Boswellia includes 30 species distributed in the drier parts of tropical Africa, Arabia, Pakistan, Iran, Ceylon, and India. They are trees with some traits shared with section Bursera like exfoliating bark, five merous flowers, and a usually trilocular ovary. Their main difference from the genus Bursera is that in Boswellia the locules of the ovary are all fertile, with more than one seed (Engler, 1931
; Rzedowski and Kruse, 1979
). Commiphora includes 200 species, and their distribution is similar to that of Boswellia. They are trees or shrubs sharing some similarities with species of section Bullockia. They have a nonexfoliating bark, four-merous flowers, and bilocular ovaries. Their main difference with the genus Bursera is their calyx, which is closed in the bud (Gillett, 1980
). Because of the morphological affinities between section Bursera and Boswellia and between section Bullockia and Commiphora, Rzedowski and Kruse (1979)
hypothesized that Bursera might be diphyletic. Other authors have suggested that Bursera is not monophyletic because it includes species that should belong to Commiphora. Based on pollen morphology Rzedowski and Palacios (1985)
and Palacios (1984)
suggested that B. tecomaca and B. sarcopoda were closer to Commiphora than to Bursera.
Some systematists have also proposed that Commiphora and Bursera should be combined into a single genus (Gillett, 1980
). Others consider that Bursera is closer to Boswellia and have combined them with Triomma into the subtribe Burserinae, leaving Commiphora in the other subtribe, the Commiphorinae (Lam, 1932
).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). We included the three varieties in the analyses. We also included the species Terebinthus acuminata and B. diversifolia. McVaugh and Rzedowski (1965)
recognized that T. acuminata should belong to Bursera. However, they deferred a formal name change until it could be determined whether it was distinct or conspecific to Bursera attenuata or B. simaruba. Bursera diversifolia has been recognized as a group of forms that resulted from hybridization, between B. bipinnata and usually B. copalifera (Toledo, 1982
). Our collection would correspond to the putative B. bipinnata x B. copalifera hybrid.
|
Most Bursera species were collected in the field, whereas the outgroups were obtained from private collections and botanical gardens (Table 1). Small branches of the plants were collected to make vouchers specimens (Table 1) and a few leaves were stored in plastic bags, which contained 20–30 g of silica gel for quick drying to preserve DNA. One set of voucher specimens was deposited at the Herbarium of the University of Arizona and another set was deposited at the National Mexican Herbarium (National Autonomous University of Mexico, Mexico City).
DNA extraction, amplification, and sequencing
Total genomic DNA was isolated from 0.1 g of the dry leaf materials following the protocol described by Doyle and Doyle (1987)
. Because of the high amounts of resins present in the leaves, we substituted 4XCTAB (hexadecyltrimethylammonium bromide) for 2XCTAB. DNA was subsequently purified with the S&S Elu-Quick DNA purification kit according to the instructions of the manufacturer (Scheiler & Schuell, Keene, New Hampshire; order number 74450).
The internal transcribed spacers (ITS1 and ITS2) and the 5.8S coding region were amplified from total genomic DNA by the polymerase chain reaction (PCR). The amplifications were performed in 50-µL reactions using a Perkin-Elmer Cetus GeneAmp reagents kit (Cetus Corporation, Norwalk, Connecticut; part number N808-0009). The reactions included 27.6 µL of sterile water, 3.0 µL of glycerol, 5.0 µL of 10x PCR buffer, 5.0 µL of 200 µmol/L dNTPs in an equimolar ratio, 4.0 µL of 25 mmol/L MgCl2 solution, 0.2 µL of AmpliTaq DNA Polymerase, 0.2 µL of pg32 (T4 Gene 32 protein from Ambion, catalog number 2424), 2 µL of ITS4 (TCCTCCGCTTATTGATATGC) primer, 2 µL of ITS5 (GCAAGTAAAAGTCGTAACAAGG) primer, and 1 µL of template. Reaction samples included positive and negative controls. The cycling amplification profile followed the one described by Baldwin (1992)
. PCR products were purified using Microcon 100 microconcentrators (W. R. Grace and Co., Beverly, Massachusetts; product number 42413), and then sequenced with the automated sequencer ABI-373 (Applied Biosystems, Foster City, California).
To obtain reliable sequences, all the species included in the study were sequenced at least once with primers ITS5, ITS4, and also with internal primers ITS2 (GCTGCGTTCTTCATCGATGC) and ITS3 (GCATCGATGAAGAACGCAGC). The ITS region in Bursera comprises 700 base pairs, and each reading from the sequencer included 420 perfectly readable base pairs, for most species. Thus, the central section of the ITS region (including the 5.8 cistron) was read four times (twice in each direction), and the distal sections of the region were read twice (once in each direction).
Alignment of sequences and phylogenetic analyses
Sequences were aligned with the Genetics Computer Group software package (GCG; Madison, Wisconsin), and subsequently further aligned with the program Sequencher (1995)
. Finally, correction of the last small misalignments was performed manually with Sequencher. Since none of the sequences were identical, we included all taxa in the analysis. Each gap position was treated as missing data to retain information about nucleotide substitutions in taxa with the insertion (Wojciechowski et al., 1993
). Also, entire gaps that were unambiguous and potentially informative (shared by two or more species) were scored and entered as separate characters. All sequences were submitted to GenBank (accession numbers GBAN-AF080003 to GBAN-AF080064; the prefix GBAN- has been added for linking the on-line version of American Journal of Botany to GenBank but is not part of the actual accession number).
Phylogenies were inferred from the sequence matrix using parsimony analysis with PAUP 3.1.1 (Swofford, 1993
) under the Fitch criterion. We performed heuristic searches because of the large number of species included. To maximize the probability of finding the shortest possible trees our phylogenetic analyses implemented several combinations of sequence-addition and branch swapping. We did (1) SIMPLE addition and tree bisection reconnection (TBR) branch swapping; (2) CLOSEST addition and TBR branch swapping; (3) 20 replicates of RANDOM addition and nearest-neighbor interchange (NNI) branch swapping; and (4) 400 replicates of RANDOM addition and TBR branch swapping. Character state changes were weighted equally. The amount of phylogenetic information in the analysis was estimated using the consistency index (Kluge and Farris, 1969
), the retention index (Farris, 1989
) and the g1 statistic (from 10 000 random trees; Huelsenbeck, 1992
).
To estimate the relative robustness of individual clades, we implemented a bootstrap analysis, which involved 500 bootstrap searches with 40 replicates of RANDOM addition and TBR branch swapping.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The searches generated four maximally parsimonious trees that involved 1326 evolutionary steps. All the random-addition replicates converged on the same four trees, suggesting that we probably did find the optimal trees (Swofford, 1993
). Their consistency index was 0.57 and their retention index was 0.74. The g1 statistic was -0.22, suggesting significant phylogenetic signal (Huelsenbeck, 1991
). Only the search that involved NNI branch swapping converged on longer trees (six trees of 1327 steps).
The strict consensus tree is highly resolved (Fig. 1), which suggests that ITS was a good choice for this molecular study. The four most parsimonious trees differed topologically in the resolution of the position of B. heteresthes and B. mirandae from section Bullockia. In two of these maximally parsimonious trees, the root of the clade with these two species is unresolved, whereas in the other two trees their root is resolved and situated as the sister clade of the clade that includes B. palmeri, B. infiernidialis, B. penicillata, B. stenophylla, and B. hindsiana. Also, B. palmeri appears in two of the trees as the sister species of the four other species in its clade, whereas in other two trees it is the sister species of only B. penicillata and B. infiernidialis.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Rather, our results are consistent with previous arguments that Commiphora and Bursera are closely related (Rzedowski and Kruse, 1979
; Gillett, 1980
). The members of Commiphora and Bursera sampled resolve as a monophyletic sister groups, with considerable ITS sequence differentiation between them and 100% bootstrap support for each. Even B. tecomaca, which has been repeatedly suggested to belong to Commiphora (Palacios, 1984
; Rzedowski and Palacios, 1985
), clearly falls within the clade of Bursera section Bullockia with 100% bootstrap support. Although our sampling of Bursera was extensive, our analysis includes only three of the 200 species of Commiphora. Thus, while these Bursera species are clearly more related to each other then to the three species of Commiphora, a broader taxonomic sample of Commiphora is desirable to confirm separate monophyly. However, further sampling is unlikely to place Bursera section Bursera closer to Boswellia than to Commiphora.
The division of the genus into two sections is strongly supported by our DNA phylogeny. McVaugh and Rzedowski (1965)
stated that these sections form good natural groups. In our phylogeny they are each monophyletic with 100% bootstrap values. Traditionally, the morphological trait most frequently used to distinguish the two sections is the presence or absence of the exfoliating bark (McVaugh and Rzedowski, 1965
). Yet, B. paradoxa and B. mirandae have anomalous traits. The overall morphology, including a trivalvate drupe and trilocular ovary of B. paradoxa, indicates its affiliation with section Bursera (Guevara-Féfer and Rzedowski, 1980
), but its nonpeeling gray bark is like that of section Bullockia. The opposite occurs with B. mirandae, which has the general morphology of a Bullockia species but also has a peeling bark (Toledo, 1984
). Our phylogeny confirms that these species should be classified according to their general morphology, not their bark characteristics. Closer examination of the traits of these species reveals that the exfoliation in B. mirandae is atypical in its thickness as compared to the usually papery thinness of the exfoliation of species of section Bursera (Toledo, 1984
). Likewise, seedlings and young juveniles of B. paradoxa retain their sectional traits of papery exfoliating bark (personal observation).
Subdivisions of the sections into monophyletic groups
McVaugh and Rzedowski (1965)
and Rzedowski and Kruse (1979)
recognized the mulato group within section Bursera as a natural, well-defined group (Table 1). These species are distinguished from the rest of the genus by their red exfoliating bark, leaflets with entire margins and apices mostly acuminate, and trilobate cotyledons. Monophyly is strongly supported by the molecular phylogeny (Fig. 1). Terebinthus acuminata falls into this group, confirming the widely held belief that it is indeed a Bursera. Its position in the phylogeny suggests that it may be more closely related to B. attenuata than to B. simaruba. Further population studies are required to determine whether these two species should be combined.
Our consensus tree contains three other recognizable groups in section Bursera: the fragilis group, the fagaroides group, and the microphylla group. Until more reproductive characteristics are gathered, the most conspicuous difference between these three groups and the mulatos is that the mulatos have trilobate seedlings and the leaflets with margin entire, whereas the other three groups have multilobate seedlings. The fragilis group is distinguished by having generally serrate (or crenate) leaflet margins, whereas the fagaroides and microphylla groups have generally entire margins. The microphylla group has distinctive small linear leaflets and red trunks.
Although bootstrap support is not high for any of these groups, except the mulato group, the general outlines of these groups are consistent with previous systematic treatment. The microphylla group has been consistently recognized as a set of related species (Bullock, 1936
; McVaugh and Rzedowski, 1965
; Toledo, 1984
), although reservations regarding monophyly have sometimes been expressed based on the possibility of convergence in leaf morphology (e.g., Toledo, 1982
). The monophyly of this group in our strict consensus tree suggests that leaf morphology of this group is a good synapomorphy. The fagaroides and fragilis groups have also been consistently recognized as natural groupings, though B. paradoxa, B. trifoliolata, and the closely related pair B. chemapodicta and B. schlectendalii are distinctive and have always been difficult to place (McVaugh and Rzedowski, 1965
; Guevara-Féfer and Rzedowski, 1980
). Nonetheless, the remaining species have been recognized as having clear affinities with one of these two groups. Thus, our strict consensus tree confirms the idea that these are likely to be good groupings with the possible exception of the four "difficult" species. All four of these difficult species have attached to the fagaroides clade in the molecular phylogeny. It is perhaps indicative of their distinctive status that this clade has the lowest bootstrap value of the four. Indeed, when these four species are excluded from analysis, the bootstrap value for the fagaroides group increases to 59%. Of the four, B. trifoliolata is most convincingly placed with strong bootstrap support for its close relationship to the yellow-barked B. aptera and B. discolor.
The division of the nonmulato species of section Bursera into red-bark cuajiotes and yellow-bark cuajiotes (Table 1; Toledo, 1982
) is not supported by the molecular phylogeny. The red-barked microphylla group comes out closer to the yellow-barked fagaroides group than to the red-barked fragilis group. Also, the position of the red-barked species, B. schlechtendalii, B. trifoliolata, and B. chemapodicta within the fagaroides group, further detracts from the idea that red and yellow bark are good synapomorphies.
Several smaller clades within these groups have some support. Bursera chemapodicta was originally described as a local endemic derived from B. schlectendalii. Their close relationship is strongly supported (Fig. 1). In the fragilis group, B. denticulata, B. crenata, B. kerberi, B. lancifolia, and B. trimera appear to be closely related with two sister-species pairs involving reductions in leaflet number: the multifoliate B. lancifolia to tri- or unifoliate B. trimera and trifoliate B. kerberi to unifoliate B. crenata.
Further divisions of the Bullockia section based on the results of our phylogenetic analysis are less reliable because bootstrap values are low. This is somewhat expected from the lower level of DNA divergence at the lower branches of the tree, which result in lower bootstrap values even in the absence of contradictory characters (Sanderson, 1989
). Nevertheless, some patterns can still be discerned. Toledo (1982)
divided section Bullockia into two groups. The first one included species whose seeds are completely or almost completely (two-thirds or more) covered by the pseudoaril. The second group comprised species only partially covered (less than two-thirds) by the pseudoaril. The strict-consensus tree contains a clade that consists exclusively of members of the covered-pseudoaril group, which we have labeled as the copalifera group (Fig. 1). Although all of the species in this clade are from the covered-pseudoaril group, several species with seeds completely or almost completely covered by the pseudoaril are not part of this clade (i.e., B. cuneata, B. coyucensis, B. palmeri, and B. diversifolia [which is probably a hybrid species]). That this clade exclusively contains plants previously posited to have close systematic affinity suggests that it may approximate a valid, monophyletic group. The rest of the section contains few monophyletic groups, and the characteristic of having a partially covered seed is not a good synapomorphy uniting them.
Several smaller clades have some support. Bursera biflora, B. glabrifolia, and B. xochipalensis form a monophyletic group in the strict-consensus tree, albeit with only 67% bootstrap support. Rzedowski suggests in his original description of B. xochipalensis (cited in Toledo, 1982
) that it is closely related to B. glabrifolia. Likewise B. biflora has been posited to be a unifoliate derivative of B. gabrifolia (Bullock, 1936
). Thus, agreement with previous systematic hypotheses suggests some validity for this clade. It is interesting that B. diversifolia, a putative hybrid between B. bipinnata and B. copalifera of the covering-aril clade attaches to the B. biflora clade. Toledo (1982)
suggested that B. cuneata, B. copalifera, and B. velutina are closely related species that replace each other ecologically along an altitudinal gradient. Our study confirms that B. copalifera and B. velutina are sister taxa but suggests that B. cuneata is not closely related. Bursera cuneata's positioning with B. aloexylon and B. coyucensis is supported by an unusual bifurcation of the style of females (the style divides below the stigma) in the specimens we examined of these three species (personal observation). There is strong bootstrap support for a close relationship between B. mirandae and B. heteresthes. This positioning seems reasonable on morphological grounds and Toledo (1982
, 1984
) previously suggested that B. mirandae may be close to "species such as B. laxiflora, B. bonetti, or B. heteresthes." Bursera penicillata and B. infiernidialis form a strongly supported species pair in the molecular phylogeny. Although reasonable on morphological grounds, their relationship has not been previously suggested.
Other details of the molecular analysis shed light on the taxonomy of the species. For example, B. fragilis, a species of northwestern Mexico, has been regarded as synonymous with B. lancifolia of south central Mexico (Bullock, 1936
; Johnson, 1992
). Although they have disjunct distributions, these two species exhibit considerable overlap in leaf and leaflet size and morphology, and in fruit characteristics. In our phylogeny, however, they are located rather far apart from each other, casting doubt on their conspecificity (Fig. 1).
The same is true of B. laxiflora and B. filicifolia, which, because of their similar morphological features, are sometimes considered synonymous (Felger and Lowe, 1970
; Turner, Bowers, and Burgess, 1995
). The former occurs in most of the state of Sonora, whereas the latter occurs only in southern Baja California. In our phylogeny they are quite distantly related for conspecifics.
According to the molecular phylogeny, two varieties of B. fagaroides, var. fagaroides and var. purpusii, form one clade, and var. elongata is in another more ancestral clade. This could be explained by a hybrid origin of var. elongata, between var. fagaroides and another species from the fragilis group (McVaugh and Rzedowski, 1965
). But it is also possible that var. elongata is a completely different species from var. fagaroides. This study does not provide the best kind of evidence to resolve the placement of these three varieties, but, given their genetic differences in the ITS region (Fig. 2), perhaps they should be tentatively regarded as three separate species.
Evolutionary tendencies within the genus
There has been considerable discussion about the evolutionary tendencies of different characteristics of species of Bursera (Rzedowski and Kruse, 1979
; Toledo, 1982
). With this DNA phylogeny we can begin to make more definitive statements about evolutionary trends.
Colorful, peeling bark is probably one of the most striking features of this genus. This exfoliating bark may serve several functions (Rzedowski and Kruse, 1979
). Peeling may provide light access to the photosynthetic layers of the trunk by occurring when the bark becomes opaque. It may discourage the establishment of lichens, which block light. It may impede the establishment of epiphytic seedlings, like those of Ficus, and it may attract fruit dispersers from a long distance. According to our phylogeny, the presence of an exfoliating bark is a primitive trait in section Bursera, with a subsequent independent evolution of a smooth nonpeeling bark in B. paradoxa. Red-peeling bark is primitive and yellow-peeling bark is derived, though there are some reversions as mentioned above. The opposite occurs in section Bullockia where smooth bark is ancestral, with subsequent evolution of peeling bark in B. mirandae.
Another interesting feature of burseras is the large variation in the size of leaves and the size, shape, and number of leaflets. Leaf size and shape are considered to be of adaptive significance in several ecological contexts (Givnish, 1979
). Small, long-lived sclerophyllous leaves are considered important adaptations to nutrient stress, yet large systematic variation in nutrient availability is not readily apparent across Bursera habitats. Furthermore, most burseras occur in seasonally dry habitats and they produce one cohort of leaves that usually lasts the full wet season. This provides little opportunity for specialization with regard to longevity, with the possible exception of B. simaruba, which lives in evergreen rainforest, and B. microphylla and B. hindsiana, which occur in deserts where rainfall is more sporadic and somewhat less seasonal. However, there is a strong gradient in temperature and aridity across the habitats in which Bursera is found. Thus, leaf adaptations to mesic vs. arid gradients are expected. Leaf size and subdivision have important impacts on leaf energy budgets. Small leaves or leaflets can regulate their temperature to be near ambient, whereas the temperature of large leaves tends to increase above ambient upon exposure to high levels of solar radiation. The most important impact of high leaf temperature in arid habitats is greater transpirational loss of hard-to-obtain water. Reduction of leaf area or subdivision of leaves is thus favored in arid habitats.
The evolution of leaves in the section Bursera is instructive in this regard. The gross macroevolutionary pattern is one of reduction of leaf size as this section evolved into more arid habitats. First, the large-leafed simaruba clade separated from the rest (Table 2). Then, the smaller leaved fragilis clade separated. Finally, the even smaller leafed fagaroides clade separated from the smallest leaved microphylla clade. Species in these groups tend to live in progressively drier habitats along this evolutionary sequence (E. Huerta, J. X. Becerra, and D. L. Venable, unpublished data). Within each of these clades there was subsequent evolution of larger and smaller leaves in moister and drier habitats, respectively (e.g., B. morelensis vs. B. arida or B. ariensis vs. B. schlechtendalii).
|
; Toledo, 1982
). Bursera instabilis tends to have 1–3 leaflets, which is a reduction in number in the simaruba clade. In the fragilis clade B. trimera is, in our phylogeny, a trifoliate sister species to B. lancifolia, which frequently has 5–7 leaflets. Bursera crenata is a unifoliate sister species to the usually trifoliate B. kerberii. In the fagaroides clade B. chemapodicta and B. schlechtendalii are unifoliate derived species in contrast to B. paradoxa, which has evolved reduced leaflet size in the form of unusual narrowly linear leaflets. In all of these cases the derived species occur in hotter drier habitats. Leaf area of the whole microphylla clade is low. This condition has been achieved by the production of many small strap-shaped leaflets. Bursera arida represents a derived reduction in leaflet number in response to aridity within this clade. Similar trends occur in section Bullockia. Bursera biflora is the often unifoliate sister species of pinnate B. glabrifolia and B. xochipalensis in our phylogeny. Bursera heteresthes is the trifoliate sister species of pinnately compound B. mirandae. Bursera infiernidialis is a trifoliate sister species to pinnate B. penicillata. Bipinnate leaves with very small leaflets have evolved at least twice (B. bipinnata and B. stenophylla) from pinnate ancestors. The independent status of the often-bipinnate B. laxiflora and B. filicifolia from each other and from B. stenophylla is suggested by our most parsimonious trees. However, this part of the tree has short branch lengths and low bootstrap values, making the conclusion of independence tentative.
Basal breeding systems appear to be dioecious, though B. simaruba has been reported to be dioecious or polygamodioecious. At the base of section Bullockia, B. tecomaca and the clade involving B. bonetti are all dioecious. Although most species and populations of Bursera are dioecious, considerable variation in the breeding system occurs, mostly at the within-species level. In B. fagaroides var. fagaroides, we have observed all-female, presumably agamospermous populations, populations with males, females, and andromonoecious individuals (inconstant males with varying proportions of hermaphroditic flowers), and at least one population with male, female, andromonoecious inconstant males, and hermaphroditic individuals (E. Huerta and D. L. Venable, personal observations). We have observed dioecious and hermaphroditic populations of B. fagaroides var. elongata. Bursera medranoana (not in our phylogeny, but putatively a hybrid between B. morelensis and B. schelectendalii; Rzedowsky and Ortiz, 1988
) appears to also be agamospermous, consisting only of female individuals (Rzedowski and Ortiz, 1988
; and E. Huerta and D. L. Venable, personal observations). In most populations of B. galeottiana the males appear to be inconstant, producing a few hermaphroditic flowers and fruits (E. Huerta and D. L. Venable, personal observation). Bursera schlechtendalii produces males, females, and inconstant males with hermaphroditic flowers and fruits, at least in some populations. Since Bursera flowers are small and often emerge before the species-identifying foliage, it is hard to gather information on the intraspecific distribution of sexual systems for the whole genus. Nonetheless, among-population variation in asexual vs. sexual or dioecious vs. hermaphrodite breeding systems present a rare opportunity for comparing these systems which typically occur in different genera or families of plants (Barrett, Harder, and Warley, 1996
) and we are currently investigating these systems.
Likewise, much remains to be learned about the fruits of Bursera. The color and degree to which the pseudoaril covers the usually black seed are well documented (Rzedowski and Kruse, 1979
; Toledo, 1982
) and there are clear taxonomic patterns. All of section Bursera tends to have the pseudoaril completely covering the seed, and the color of the aril tends to be cream or yellow with some tendencies toward red, for example, in B. schlechtendalii and B. crenata. The pseudoaril is more variable in section Bullockia, with the B. copalifera clade having completely or almost completely covering, usually orange to red pseudoarils. Most of the remaining species have the aril covering half or less of the seed. Bursera infiernidialis is unusual in having the pseudoaril covering half of one face of the seed, but most of the other. These seeds are attractive to birds and secondarily consumed by rodents once on the ground (Rzedowski and Kruse, 1979
).
There is considerable variation in ecologically relevant fruit traits that are not well understood for the genus as a whole. Pseudoarils range from dry and papery to thick and fleshy with different water and sugar contents (Bates, 1992
). Also seeds are matured either simultaneously or over a prolonged period (Rzedowski and Kruse, 1979
). Bates (1992)
studied B. hindsiana and B. microphylla on the coast of Sonora, Mexico, and found that fruits of B. hindsiana were used by a variety of generalist birds. It had higher dry mass of aril and higher sugar content per seed than B. microphylla, and fruits were matured simultaneously and eaten in the autumn. In contrast, fruits of B. microphylla were used by overwintering gray vireos who set up territories around them and used them almost exclusively. They matured a few at a time throughout the winter and had lower sugar content and volume. Similar patterns have been seen in the northeastern United States deciduous forest where some plants produce high-quality fruits that mature synchronously during autumn bird migration, whereas others produce lower quality fruits more gradually, which are used by nonmigrants through the winter (Stiles, 1989
). Greenberg (1993)
has shown that white-eyed vireos overwintering in the Yucatan Peninsula feed heavily (96% of frugivory observations) on B. simaruba throughout the winter. These birds have territories that include B. simaruba trees, and they defend their trees against conspecifics as well as other species of vireos. Bursera simaruba's fruits ripen slowly for a 7–8 mo period coinciding with the white-eyed vireo overwintering (September–early April). Rzedowski and Kruse (1979)
list B. fagaroides, B. longipes, and B. morelensis (representing three of the four section Bursera clades) as other species that slowly ripen fruits over many months and B. bipinnata and B. copalifera (both members of the covering-pseudoaril clade of section Bullockia) as examples of simultaneous maturing fruits. Yet, the general distribution of these and other frugivore-related traits such as quantity and quality of fleshy pseudoaril, as well as the quantity and quality of dispersal provided by different dispersal agents, is not yet known.
Thus, Bursera exhibits high lability in many ecologically interesting traits and provides a good opportunity to test hypotheses on their evolutionary tendencies. Such tests using this phylogeny will also help in understanding the evolutionary success of this highly diverse and abundant group of the dry tropics of Mexico.
|
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Barrett, S. C. H., L. D. Harder, and A. C. Worley. 1996 The comparative biology of pollination in flowering plants. Philosophycal Transactions of the Royal Society of London, Series B 351: 1271–1280.
Bates, J. M. 1992 Frugivory on Bursera microphylla (Burseraceae) by wintering gray vireos (Vireo vicinor, Vireonidae) in the coastal deserts of Sonora, Mexico. Southwestern Naturalist 37: 252–258.[CrossRef][ISI]
Becerra, J. X. 1994 Squirt-gun defense in Bursera and the chrysomelid counterploy. Ecology 75: 1991–1994.[CrossRef][ISI]
———. 1997 Insects on plants: macroevolutionary chemical trends in host use. Science 276: 253–256.
———, and D. L. Venable. 1990 Rapid terpene-bath and "squirt-gun" defense in Bursera schlechtendalii and the counterploy of chrysomelid beetles. Biotropica 22: 320–323.[CrossRef][ISI]
Bullock, L. L. 1936 Notes on the Mexican species of the genus Bursera. Bulletin of Miscellaneous Information, Kew 1936: 347–387.
Daly, D. C. 1992 The genus Tetragastris and the forests of eastern Brazil. Kew Bulletin 45: 179–194.[CrossRef]
———. 1993 Notes on Bursera in South America, including a new species. Studies in Neotropical Burseraceae VII. Brittonia 45: 240–246.[CrossRef][ISI]
Donoghue, M. J., and M. J. Sanderson. 1992 The suitability of molecular and morphological evidence in reconstructing plant phylogeny. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants. Chapman & Hall, New York, NY.
Doyle, J. J., and J. L. Doyle. 1987 A rapid DNA isolation method for small quantities of fresh tissues. Phytochemical Bulletin 19: 11–15.
Engler, A. 1931 Burseraceae in Engler & Prantl, Die Natürllichen Pflanzenfamilien. Ed. 2, 19A: 405–456.
Farris, J. S. 1989 The retention index and homoplasy excess. Systematic Zoology 38: 406–407.[CrossRef]
Felger, R. S., and C. H. Lowe. 1970 New combinations for plant taxa in northwestern Mexico and southwestern United States. Journal of the Arizona Academy of Science 6: 82–84.
Forman, L. L., P. E. Brandham, M. M. Harley, and T. J. Lawrence. 1991 Beiselia mexicana (Burseraceae) and its affinities. Kew Bulletin 44: 1–31.
Gillett, J. B. 1980 Commiphora (Burseraceae) in South America and its relationship to Bursera. Kew Bulletin 34: 569–589.
———. 1991 Burseraceae. In R. M. Polhill [ed.], Flora of East Africa. A. A. Balkema, Rotterdam.
Givnish, T. 1979 On the adaptive significance of leaf form. In O. T. Solbrig, S. Jain, G. B. Johnson, and P. H. Raven [eds.], Topics in plant population biology, 375–407. Columbia University Press, New York, NY.
Greenberg, R. S. 1993 Forest migrants in non-forest habitats in the Yucatán peninsula. In J. M. Hagan and B. W. Johnston [eds.], Ecology and conservation of neotropical migrant land birds. Smithsonian Institution Press, Washington, DC.
Guevara-Féfer, F., and J. Rzedowski. 1980 Notas sobre el género Bursera (Burseraceae) en Michoacán (Mexico). I. Tres especies nuevas de los alrededores de la presa del Infiernillo, con algunos datos relativos a la región. Boletín de la Sociedad Botánica de México 39: 63–81.
Guillaumin, A. 1909 Recherches sur la structure et le development de Burseracées: application a la systématique. Annales des Sciences Naturelles, Botanique Serie 9, 10: 201–301.
Huelsenbeck, J. P. 1991 Tree-length distribution skewness: an indicator of phylogenetic information. Systematic Zoology 40: 257–270.[CrossRef]
Johnson, M. B. 1992 The genus Bursera (Burseraceae) in Sonora, Mexico and Arizona, U.S.A. Desert Plants 10: 126–143.
Kluge, A. G., and J. S. Farris. 1969 Quantitative phyletics and the evolution of the anurans. Systematic Zoology 18: 1–32.
Lam, H. J. 1932 The Burseraceae of the Malay Archipelago and peninsula. Bulletin Jardin Botanique de Buitenzorg Serie 3, 12: 296–297.
McVaugh, R., and J. Rzedowski. 1965 Sinopsis of the genus Bursera L. in western Mexico, with notes on the material of Bursera collected by Sessé & Mociño. Kew Bulletin 18: 317–382.[CrossRef]
Miranda, F. 1947 Estudios sobre la vegetación de México—V. Rasgos de la vegetación de la Cuenca del Rio Balsas. Revista de la Sociedad Mexicana de Historia Natural 8: 95–114.
Mooney, H. A., and W. A. Emboden, Jr. 1968 The relationship of terpene composition, morphology and distribution of populations of Bursera microphylla (Burseraceae). Brittonia 20: 44–51.[CrossRef][ISI]
Palacios, R. 1984 La morfología de los granos de polen de las especies mexicanas del género Bursera. Biotica 9: 153–182.
Rzedowski, J. 1968 Notas sobre el género Bursera (Burseraceae) en el Estado de Guerrero (México). Anales de la Escuela Nacional de Ciencias Biológicas de México 17: 17–36.
———. 1978 Vegetación de México. Limusa. México DF.
———, and F. Guevara-Féfer. 1992 Burseraceae. In J. Rzedowski and G. Calderón de Rzedowski [eds.], Flora del Bajío y de regiones adyacentes. Instituto de Ecología A. C. Centro Regional del Bajío, Pátzcuaro, Michoacán, México.
———, and H. Kruse. 1979 Algunas tendencias evolutivas en Bursera (Burseraceae). Taxon 28: 103–116.[CrossRef][ISI]
———, and E. Ortíz. 1988 Estudios quimiotaxonómicos de Bursera (Burseraceae). II. Una especie de origen híbrido de la Barranca de Tolantongo, Estado de Hidalgo. Acta Botanica Mexicana 1: 11–20.
———, and R. Palacios. 1985 La presencia de Commiphora (Burseraceae) en México. Taxon 34: 207–210.[CrossRef][ISI]
Sanderson, M. J. 1989 Confidence limits on phylogenies: the bootstrap revisited. Cladistics 5: 113–129.[ISI]
Sequencher. 1995 3.0 user manual. Gene Codes Corp., Ann Arbor, MI.
Standley, P. C. 1923 In Trees and shrubs of Mexico. Contributions of the United States National Herbarium 23: 542–552.
Stebbins, L. 1950 Variation and evolution in plants. Columbia University Press, New York, NY.
Stiles, E. W. 1989 Fruits seeds and dispersal agents. In W. G. Abrahamson [ed.], Plant-animal interactions, 100–112. McGraw Hill, New York, NY.
Swofford, D. L. 1993 Phylogenetic analysis using parsimony (PAUP) version 3.1.1. Smithsonian Institution, Washington, DC.
Terrazas-Salgado, T. 1994 Wood anatomy of the Anacardiaceae: ecological and phylogenetic interpretation. Ph.D. dissertation, University of North Carolina, Chapel Hill, NC.
Toledo, C. A. 1982 El género Bursera en el estado de Guerrero. Tesis profesional. Facultad de Ciencias UNAM, México D.F.
———. 1984 Contribuciones a la flora de Guerrero. Tres especies nuevas del género Bursera (Burseraceae). Biotica 9: 441–449.
Turner, R. M., J. E. Bowers, and T. L. Burgess. 1995 Sonoran desert plants: an ecological atlas. University of Arizona Press, Tucson, AZ.
Wojciechowski, M. F., M. J. Sanderson, B. G. Baldwin, and M. J. Donoghue. 1993 Monophyly of aneuploid Astragalus (Fabaceae)—evidence from nuclear ribosomal DNA internal transcribed spacer sequences. American Journal of Botany 80: 711–722.[CrossRef][ISI]
This article has been cited by other articles:
|
|
 |
|
  B. K. Dunphy and J. L. Hamrick Estimation of gene flow into fragmented populations of Bursera simaruba (Burseraceae) in the dry-forest life zone of Puerto Rico Am. J. Botany, November 1, 2007; 94(11): 1786 - 1794. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Sotuyo, A. Delgado-Salinas, M. W. Chase, G. P. Lewis, and K. Oyama Cryptic Speciation in the Caesalpinia hintonii Complex (Leguminosae: Caesalpinioideae) in a Seasonally Dry Mexican Forest Ann. Bot., November 1, 2007; 100(6): 1307 - 1314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. X. Becerra The impact of herbivore-plant coevolution on plant community structure PNAS, May 1, 2007; 104(18): 7483 - 7488. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. X. Becerra From The Cover: Timing the origin and expansion of the Mexican tropical dry forest PNAS, August 2, 2005; 102(31): 10919 - 10923. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. B. NAVARRO, V. N. SUAREZ-SANTIAGO, and G. BLANCA A New Species of Haplophyllum A. Juss. (Rutaceae) from the Iberian Peninsula: Evidence from Morphological, Karyological and Molecular Analyses Ann. Bot., October 1, 2004; 94(4): 571 - 582. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hamzeh and S. Dayanandan Phylogeny of Populus (Salicaceae) based on nucleotide sequences of chloroplast TRNT-TRNF region and nuclear rDNA Am. J. Botany, September 1, 2004; 91(9): 1398 - 1408. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Weeks and B. B. Simpson Molecular genetic evidence for interspecific hybridization among endemic Hispaniolan Bursera (Burseraceae) Am. J. Botany, June 1, 2004; 91(6): 976 - 984. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. X. Becerra Synchronous coadaptation in an ancient case of herbivory PNAS, October 28, 2003; 100(22): 12804 - 12807. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. X. Becerra, D. L. Venable, P. H. Evans, and W. S. Bowers Interactions Between Chemical and Mechanical Defenses in the Plant Genus Bursera and Their Implications for Herbivores Integr. Comp. Biol., August 1, 2001; 41(4): 865 - 876. [Full Text] [PDF]
|
|
|
|
|
|
J. X. Becerra and D. L. Venable Macroevolution of insect-plant associations: The relevance of host biogeography to host affiliation PNAS, October 26, 1999; 96(22): 12626 - 12631. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||
