Understanding how biodiversity is shaped through time is a fundamental question in biology. Even though tropical rain forests (TRF) represent the most diverse terrestrial biomes on the planet, the timing, location and mechanisms of their diversification remain poorly understood. Molecular phylogenies are valuable tools for exploring these issues, but to date most studies have focused only on recent time scales, which minimises their explanatory potential. In order to provide a long-term view of TRF diversification, we constructed the first complete genus-level dated phylogeny of a largely TRF-restricted plant family with a known history dating back to the Cretaceous. Palms (Arecaceae/Palmae) are one of the most characteristic and ecologically important components of TRF worldwide, and represent a model group for the investigation of TRF evolution.
We provide evidence that diversification of extant lineages of palms started during the mid-Cretaceous period about 100 million years ago. Ancestral biome and area reconstructions for the whole family strongly support the hypothesis that palms diversified in a TRF-like environment at northern latitudes. Finally, our results suggest that palms conform to a constant diversification model (the 'museum' model or Yule process), at least until the Neogene, with no evidence for any change in diversification rates even through the Cretaceous/Paleogene mass extinction event.
Because palms are restricted to TRF and assuming biome conservatism over time, our results suggest the presence of a TRF-like biome in the mid-Cretaceous period of Laurasia, consistent with controversial fossil evidence of the earliest TRF. Throughout its history, the TRF biome is thought to have been highly dynamic and to have fluctuated greatly in extent, but it has persisted even during climatically unfavourable periods. This may have allowed old lineages to survive and contribute to the steady accumulation of diversity over time. In contrast to other plant studies, our results suggest that ancient and steady evolutionary processes dating back to the mid-Cretaceous period can contribute, at least in part, to present day species richness in TRF.
Tropical rain forests (TRF) are the most biodiverse terrestrial ecosystems on the planet . They are characterised by a closed, multilayered canopy dominated by flowering plants (angiosperms ) and occur only in frost-free areas with high mean monthly temperatures and precipitation, and low seasonality . Today, TRF covers just 7% of the Earth's surface  in equatorial zones of the Americas, Africa and the Indo-Pacific, and is highly threatened by human activity . The origin and evolution of species-rich biomes raises fundamental questions in evolutionary biology  and, as such, the diversification of TRF has been much debated [3, 6].
Even though it is generally agreed that TRF is a relatively old biome, the location and timing of its origin remain uncertain mainly because the fossil record for tropical regions is highly incomplete, especially during the Cretaceous [7, 8]. Some direct  and indirect  evidence suggests that TRF was present in the mid-Cretaceous period (100 million years ago (Ma)) at middle paleolatitudes (for example, Laurasia) while other studies indicate that the first paleoflora attributable to modern day TRF are found in the Early Paleocene of North America  and Late Paleocene of South America [7, 8, 12] and Africa [13–15].
Whereas previous views suggested that the TRF biome has been ecologically stable over long periods of time , more recent data indicate that it is highly dynamic  having fluctuated both in extent [14, 18] and in the diversity of plants that it sustains [7, 8]. These views have led to three general evolutionary hypotheses that explain the high levels of present day species diversity found within TRF: (i) early, rapid speciation in response to favourable climatic conditions followed by a deceleration of diversification rates due to global cooling  and contraction of TRF (referred to here as the 'ancient cradle model', see [10, 20, 21]); (ii) constant diversification rates coupled with low extinction rates leading to a gradual accumulation of lineages in response to a long-lasting and stable tropical ecosystem (the 'museum model', see [16, 22]); and (iii) an increase in diversification rates towards the present in response to climatic, tectonic or biotic changes (the 'recent cradle model', see [23–25]). These hypothetical processes result in alternative lineage accumulation through time and thus different patterns of inferred branch length distributions (Figure 1). Several phylogenetic studies of TRF plant groups have provided evidence in support of the recent cradle model of diversification [25–27]. However, these studies were restricted to low taxonomic levels (species) and thus do not enhance our understanding of how these hypotheses might apply throughout the entire history of TRF, for example, on the long-term diversification dynamics of TRF. In contrast, studies of the early diversification of TRF plant lineages that permit tests of the above hypotheses are rare.
In the absence of a complete fossil record for tropical rain forests, family-level diversification analyses of large pantropical angiosperm groups that are ecologically characteristic of TRF can provide important insights into the historical construction of the biome . In this respect, the pantropical palm family (Arecaceae/Palmae) presents an ideal study group. First, palms are among the most important and characteristic components of TRF ecosystems worldwide in terms of species diversity (approximately 2,400 species), abundance of individuals and impact on the environment [28–30]. Based on the excellent taxonomic knowledge for this family , we calculated that over 90% of its species diversity is restricted to TRF (Figure 2). Water and energy-related variables are strong determinants of palm diversity [31, 32] and fundamental anatomical constraints inhibit palms from colonising cold environments [33, 34]. Second, the known history of palms extends far back into the Cretaceous although the details of the spatiotemporal origin of the family remain controversial . Direct evidence from unambiguous fossils associated with palms suggest that the family was already present during the Turonian (89-93.5 Ma, [35–37]) while more doubtful fossils have been recorded since the Aptian (112 Ma, ). More recently, several molecular clock estimates based on angiosperm wide phylogenies suggested a stem age for the family ranging from 91 to 120 Ma [38–41]. These studies were based on a very limited sampling within the family and thus do not provide reliable approximations for the crown node age and early diversification history. To date, most estimates of palm ages have focused on subfamily [42, 43] or tribal levels [44–46].
Here we investigate the origin and diversification of palms in space and time using the first complete generic-level sampling for any important TRF-restricted plant family . We estimated speciation events under a Bayesian framework using a relaxed molecular clock approach (BEAST, ), while the spatial origin of the family was inferred under a maximum likelihood method that implements the dispersal-extinction-cladogenesis model [49, 50]. Finally, we use palms as a model to explore the evolution of TRF biodiversity by testing which of the three TRF diversification hypotheses outlined above corresponds to the diversification history of the family.
Results and discussion
Evolutionary origin of palms
The fossil-calibrated molecular dating of a complete genus-level palm supertree  provides for the first time minimum age estimations for all major groups of the palm family (Table 1). The resulting chronogram (Figure 3b) suggests that the diversification of extant lineages of palms started in the mid-Cretaceous period at the Albian-Cenomanian boundary (crown node: 100 Ma, 95% highest posterior density (HPD) 108-92 Ma). The Cretaceous represents an important period for plant evolution as it witnessed the rise and diversification of flowering plants [51, 52]. The fossil record indicates that from the Albian to the middle Cenomanian angiosperms diversified extensively, becoming more abundant relative to other plants and establishing themselves as a major part of paleofloras by the end of that period . Our maximum likelihood analysis of geographic range evolution indicates that the most likely distribution of the most recent common ancestor of palms was centred on present day Central/North America and Eurasia, which corresponds to the Laurasian landmass at that time (Figure 3a). Notably, the oldest reliable palm fossils (Turonian to Campanian) have all been discovered in Europe and North America . A Laurasian origin for palms was previously suggested by Uhl and Dransfield  based on the prevalence of putatively primitive lineages in the northern hemisphere. From this ancestral area several subsequent dispersal events are inferred into the equatorial regions of South America, Africa and South East Asia, the present day distribution of palms. Finally, our ancestral biome analysis suggests that the earliest palm lineages were restricted to TRF (P = 0.984; Figure 3b and Table 1), a result that was further supported by a test of phylogenetic signal of the biome. In fact, adaptation to non-TRF biomes did not arise until the Paleocene within the fan palm subfamily Coryphoideae. Thus, our results support the notion that palms originated in a TRF-like biome and started to diversify during the mid-Cretaceous period in Laurasia.
Mean estimated ages, 95% confidence intervals and ancestral ecologies for the family and subfamilies
Age in Ma
Proportional likelihoods of ancestral ecology of branch leading to nodea
0 = 0.984/1 = 0.0/2 = 0.0
0 = 0.999/1 = 0.0/2 = 0.0
0 = 0.982/1 = 0.0/2 = 0.0
0 = 0.716/1 = 0.28/2 = 0.0
0 = 0.979/1 = 0.0/2 = 0.0
0 = 0.999/1 = 0.0/2 = 0.0
Bold entries indicate values significantly different from other states.
HPD = highest posterior density; Ma = millions of years.
Origin of tropical rain forests
The presence of TRF during the mid-Cretaceous period is controversial because pre-Cenozoic fossils associated with TRF are notoriously sparse [8, 11, 55, 56] in contrast to the relatively well documented fossil evidence of TRF during the early Cenozoic from equatorial to relatively high palaeolatitudes [11–14, 20, 57]. The earliest fossil flora interpreted as characteristic of TRF was found in the early Cenomanian from several formations in North America, for example, the Dakota formation of Kansas (99 Ma, [9, 53, 58]), which is consistent with our results. However, these conclusions, which were based on the physiognomy of leaf characters, such as shape and size, that are generally associated with megathermal vegetation [20, 53], have been questioned by some authors (see personal communication from Johnson in Morley ). Several studies have suggested that during the Cretaceous plant biodiversity was highest at mid paleolatitudes where the climate was more favourable while equatorial latitudes were exposed to a drier and hotter climate unlikely to have supported TRFs [20, 59]. In addition, simulations of major vegetation distributions during the mid-Cretaceous period (100 Ma) indicate that the presence of TRF in the Cenomanian of North America and Eurasia as well as other parts of the world is plausible [14, 60]. Finally, indirect evidence is provided by a diversification study of Malpighiales, a large plant order mainly restricted to TRFs. Using a molecular clock approach, it was estimated that the origin of this order dates to around 114 Ma with subsequent diversification during the Cenomanian , implying the presence of TRF at that time.
Even though molecular dating methods have been criticised in relation to interpretations of TRF origins  and are not assumption free, such approaches have played important roles in understanding the construction of other species rich biomes (for example, the Cape flora [61, 62]) and provide important insights when the fossil record is sparse or incomplete. Molecular dating of palms, one of the most characteristic TRF plant families, provides additional evidence that modern TRF might have already been in place 100 Ma, significantly earlier than suggested by unequivocal fossil evidence [13, 57]. It is most likely that formation of the TRF biome was a gradual process and thus the precise time at which modern TRFs can be recognised may be impossible to pinpoint. However, our results and other evidence discussed above [9, 10, 53] imply that the assembly of the TRF biome had already started during the mid-Cretaceous period and was not just a strictly Cenozoic process. It is puzzling that, to our knowledge, no macrofossil of palms has been recovered in Cenomanian deposits of North America . However, our results suggest that palms were just starting to diversify at this time and may not have been widespread, thus reducing the frequency of fossilization and probability of later discovery. Interestingly, studies of other species-rich TRF plant families yield timings for the earliest extant lineage diversification events (that is, crown node estimates) that largely post-date palms, for example Leguminosae (59 Ma, ), Annonaceae (89 Ma, ) and Rubiaceae (86 Ma, ). This would imply that palms represent one of the first extant plant families to have diversified within TRF since its origin. Thus, palms not only play a major role in present day TRF [28–30], but also appear to have been a key component in the assembly and diversity of this biome since the earliest stages of its evolution.
Early diversification of tropical rain forests
To depict global diversification at the family level and test the hypotheses of TRF diversification, we generated a semilogarithmic lineage-through-time (LTT) plot as well as 95% confidence intervals based on 1,000 randomly selected posterior trees from the BEAST output (Figure 3c). LTT plots are widely used to characterise the diversification of clades as a function of time [66–68], but they are sensitive to incomplete taxon sampling [69, 70], as is the case here (183 species sampled out of circa 2,400, approximately 7.6%, see Additional file 1). However, our sampling is phylogenetically representative and non-random as we included 100% of all described palm genera. Such a sampling strategy has the advantage of representing all the deeper nodes of the phylogeny and can provide a good estimation of diversification history up to a point in time after which undersampling at shallower nodes biases the inference [71, 72]. In order to restrict the interpretation of our LTT plot to the more accurately estimated parts we used a novel approach to identify a threshold after which the potential impact of incomplete taxon sampling becomes too important to permit accurate analyses of diversification rates (see Methods). Based on the chronogram, the percentage of total missing palm species increased sharply from 12% to 28% after circa 24 Ma (Figure 3c). Interpretation of the LTT plot was thus restricted to the period prior to that point in time (from 100 Ma to 24 Ma) and all nodes occurring after the 24 Ma threshold were excluded from subsequent analyses.
Between 100 Ma and 24 Ma, the LTT plot forms an almost straight line, which suggests that the palm family underwent a constant rate of diversification without any major shifts or radiations [67, 68]. This is also statistically supported by the better fit of the pure-birth model of diversification (constant diversification rates with no extinction) on the LTT plot than any other model tested (ΔAICRC = -1.421 between the pure birth model and the second best fitting model (the density-dependent or 'DDX' model); AIC = Akaike Information Criterion). Moreover, the ΔAICRC was not significantly different under the null hypothesis of diversification rate consistency when calculated from 10,000 phylogenies simulated under the pure-birth process (P value = 0.739). Finally, the γ statistic  also supported a constant rate hypothesis as it was not significantly different from zero (γ = -0.318, P value = 0.379). These results lend support to the museum model (Figure 1) in which diversification rates remain constant and extinction rates are low [16, 22]. This hypothesis was thought to be the consequence of the old and ecologically stable conditions of TRF over millions of years. Even though such views have now been replaced by the notion of greater dynamism in TRF evolution [14, 17, 73], the biome itself has never completely disappeared [14, 20] and has persisted in refuges during unfavourable climatic times. The existence of such refuges may have allowed comparatively old lineages to persist and contribute to present day species diversity. For example, lineages of another diverse TRF plant family, Annonaceae, were shown to have persisted in possible TRF refuges of East and West Africa for over 30 million years  even during climatically unfavourable times (for example, the global cooling of the Eocene/Oligocene boundary). We suggest here that TRF refugia may have played a similar role throughout the history of the palm family and, as a result, global palm species diversity is at least partly the result of a gradual accumulation of ancestral lineages through time, and cannot be attributed to ancient or recent speciation bursts alone.
At finer time scales, palm species diversity and diversification rates most likely fluctuated with extinction rates possibly increasing and decreasing at specific time frames in the past, perhaps in relation to climatic and geological changes. For example, studies of the palynoflora through the Late Paleocene-Eocene Thermal Maximum in Colombia indicate an increase in palm morphospecies after this time (56.3 Ma, ). Thus, it is probable that different clades within palms underwent alternative diversification scenarios leading to a heterogeneous pattern of diversification among the lineages within the family. However, our study implies that over a larger time scale these changes did not influence the overall global pattern of diversification in palms, at least until 24 Ma. This is also apparent during the major extinction episode at the Cretaceous/Paleogene boundary (K/Pg; ), which had no statistically significant effect on diversification rates (Figure 3c).
To date, relatively few family-level studies have provided evidence for the museum model of tropical plant diversification (Annonaceae , liverwort family Lejeunaceae ). This pattern contrasts with the study of Malpighiales evolution, which indicated that all major lineages originated within a short timeframe suggesting an early rapid speciation of the order, although no detailed diversification analysis was undertaken for this group . Interestingly, meta-analyses based on a large number of species-level dated molecular phylogenies of a range of plants and animals have also underlined the importance of the constant rate diversification model [77, 78]. Indeed a large number of phylogenies fitted the simplest model of diversification. For example, Morlon et al. , using a novel coalescent-based approach, found that 87 out of 289 phylogenies studied (30%) better fitted the Yule process (time constant rates with no extinction) than alternative models. Even though these results were obtained from a wide range of organisms occurring in different ecosystems, it nevertheless underlines the importance of such a simple process for explaining present day diversity.
Given the threshold of 24 Ma imposed on our LTT plot (Figure 3c), there is little indication about recent (Neogene) diversification patterns. However, it is clear that in order to attain present day species diversity (Figure 3c), rates must have increased, which suggests a shift in diversification that occurred after 24 Ma. This could have been achieved either by accelerating diversification rates, consistent with the 'recent cradle' model, or simply by higher overall constant rates. Recent high speciation rates within other TRF plant genera have been documented [13–15] and it seems likely that rapid speciation occurred within some species-rich palm genera given the very young stem node age estimations we inferred for them (for example, Pinanga, approximately 130 species, stem node approximately 12 Ma; Dypsis, approximately 140 species, stem node approximately 13 Ma). In fact, the 'museum' and 'recent cradle' hypotheses are not mutually exclusive; both mechanisms could be at work, possibly within different palm lineages and different time frames. For example, both the 'ancient' and 'recent' cradle models of diversification have been identified within the TRF-restricted leaf beetle family .
Our results from one of the most important TRF plant families suggests that present day TRF biodiversity can at least in part be explained by a steady accumulation of lineages dating back to the mid-Cretaceous period and is not just the result of rapid radiations, either recent  or ancient . The analysis of additional family-level diversification patterns of other TRF restricted plant groups will undoubtedly shed more light into the evolutionary forces that have led to the immense diversity of species found within modern TRF today.
This study builds upon the complete generic-level supertree analyses of palms by Baker et al. , the most extensive phylogenetic study of the family published to date. Here, the sampling is updated to be consistent with the latest family-wide monograph [28, 80], in which 183 genera are accepted (see Additional file 1 for the list of genera used). This was performed by repeating the supertree analyses of Baker et al.  with the addition of published plastid DNA sequence data for the recently described genus Tahina  (see below). Three commelinid monocot outgroup taxa (Costus, Dasypogon, Zea) were selected from the sampling of Baker et al. .
Palms have a rich fossil record dating from the Late Cretaceous onwards. Although the record is unusually rich among angiosperms, only a small fraction of palm fossils can be identified to specific taxonomic groups with confidence. Drawing on recent surveys of the palm fossil record [28, 82, 83], we selected the most reliable fossils (Table 2), judged on the basis of the credibility of their purported taxonomic affinities and reported ages. Nevertheless, none of these fossils is sufficiently informative to justify allocation to crown nodes ; they are thus applied conservatively to stem nodes throughout. Where authors provide a range of age estimates, we have used the most recent date. Where a geological time period alone is specified, we have used the date of the upper end of that period .
Names of fossils used to calibrate the tree, with the respective exponential prior parameters used
Hard lower bound (Ma)
Soft upper bound 95% (Ma)
Exponential mean (uncertainty)
Ma = millions of years.
The four selected fossils are widely distributed across the family and are located in three out of the five subfamilies. The earliest fossils that can be assigned unequivocally to a taxonomic group within palms are Late Cretaceous records of palmate leaves, the earliest of which is Sabalites carolinensis from the late Coniacian of South Carolina . Although an affinity with Sabal is implied by the genus name, the fossil could be linked with many coryphoid groups and its age is therefore used conservatively here as a calibration point for the stem node of subfamily Coryphoideae as a whole with an age of 85.8 Ma. Hyphaene kapelmanii, a fossil discovered at a late Oligocene site in Ethiopia , provides a further calibration point within the Coryphoideae. This fossil consists of a petiole fragment with a close resemblance to the modern genus Hyphaene due to the morphology of its wide, recurved marginal spines. We use this fossil as a constraint for the stem node of subtribe Hyphaeninae with an age of 27 Ma.
In subfamily Calamoideae, the unique structure of the pollen of subtribe Mauritiinae corresponds closely to fossil pollen in the genus Mauritiidites, specifically the clavate monosulcate grains with each spine inset and a swollen foot layer below. Mauritiidites has been recorded as early as the Maastrichtian of Africa  with numerous records soon after in the Palaeocene onwards of South America . We use it here as a calibration for the stem node of the Mauritiinae with an age of 65 Ma.
Fossil records of the coconut tribe Cocoseae, particularly of fossilised endocarps, are numerous . Until recently, well documented records appeared from the Middle Eocene onwards (for example, [90, 91]), but new research in the middle to late Palaeocene of Colombia has revealed compression fossils of large fruits that closely resemble the modern coconut, Cocos nucifera, both in size and surface morphology . In the absence of further substantiating evidence, we allocate this fossil to the stem node of the Attaleinae, the subtribe of tribe Cocoseae to which Cocos belongs, with an age of 54.8 Ma.
Finally, a number of other reliable fossils could not be used because they are made redundant by older fossils assigned to more distal nodes. Nypa is most notable here, given its outstanding macrofossil and microfossil records dating back to the Maastrichtian [28, 93]. Also significant are the distinctive diaperturate fossil pollen grains, usually referred to the form genus Dicolpopollis, which can be assigned with confidence to tribe Calameae of the Calamoideae [82, 83, 94–96]. The earliest records of this fossil palynomorph are from the Maastrichtian and its boundary with the early Palaeocene of Somalia and Borneo [88, 96].
For each fossil we applied an exponential prior, the parameters of which are given in Table 2. Finally, the stem node of palms was constrained by a uniform prior ranging from 110 to 120 Ma. This corresponds to the earliest monocot fossil . By doing this we imply that the stem of palms cannot be older than the oldest monocot fossil.
Molecular dating was carried out using BEAST 1.5.3 [48, 98]. For this analysis the 'most congruent supertree', based on the method and data of Baker et al. , was used as a topological constraint. This topology was based on extensive data sampling, including DNA sequence data, restriction fragment length polymorphisms (RFLP) and morphology, and represents the best family-wide estimation of phylogenetic relationships between palm genera to date. To update the taxonomic sampling of the supertree, we repeated the supertree analysis of Baker et al. . This study used a matrix representation with parsimony (MRP) analysis based on input trees generated from individual partitions and combinations of partitions with matrix elements weighted in proportion to bootstrap values of corresponding input tree clades. The strict consensus tree of this analysis was highly resolved with minor ambiguity in tribe Trachycarpeae and parts of tribe Areceae only. One most-parsimonious tree was selected at random and used as a constraint. This tree was pruned to include only the 183 genera accepted by Dransfield et al. . All DNA sequence datasets utilised by Baker et al.  were included in our molecular dating analysis (plastid DNA regions: atpB, matK, ndhF, rbcL, rps16 intron, trnD-trnT, trnL-trnF, trnQ-rps16; nuclear DNA regions: 18S, ITS, ms, prk, rpb2). The completeness of taxonomic sampling for each of these regions varies from 12% to 100%, with an average of 48%. Moreover, sampling for the chloroplast markers was much more complete than for the nuclear markers. Morphological and RFLP datasets used by Baker et al.  were excluded. The supertree topology was used as constraint by deleting in the XLM BEAST input file the following commands: subtreeSlide; narrowExchange; wideExchange; wilsonBalding. Each marker was individually partitioned in BEAUTi 1.5.3 http://beast.bio.ed.ac.uk/BEAUti and assigned the General Time Reversible model (GTR) with gamma-distributed rate variation (G). model of sequence evolution. Prior to our full analysis, we investigated the effect of missing data on the estimation of ages. We undertook a preliminary analysis on 2 datasets: 1 with all the 13 markers (missing data present) and 1 where the 5 nuclear markers were removed (missing data of 10%). A regression analysis between the ages obtained for all nodes was highly significantly positive (R = 0.88; t test: P < 0.001) indicating that missing data in our dataset are likely to have little influence on age estimations in molecular dating. We then undertook the full-scale analysis on the full 13-marker dataset.
In total, 8 individual analyses were carried out, 5 with 20 million generations and 2 with 30 million generations, resulting in a total of 160 million generations, and sampling every 1,000th generation. Individual analyses were performed in order to test for convergence of the results. Tracer 1.4  was used to check for convergence of the model likelihood and parameters between each run. Results were considered reliable once the effective sampling sizes (ESS) of all parameters exceeded 200. The resulting independent log and tree files were then combined using LogCombiner discarding 10% of generations as burn in per independent run.
Finally, deviation from a strict molecular clock was tested by running the analysis a second time with the strict clock enforced. We used the Bayes Factor as implemented in Tracer 1.4  to select the best-fitting model under the smoothed marginal likelihood estimate and with 100 bootstrap replicates . This test strongly supported the data as being non-clock like (ln BF = 1,158.2 ± 3.2 in favour of relaxed clock hypothesis), and thus the results under the relaxed clock are presented here.
To explore diversification rates in the family, we generated a semilogarithmic LTT plot. The mean LTT plot as well as the 95% confidence intervals was generated from a random selection of 1,000 posterior trees resulting from the BEAST analysis. LTT plots are sensitive to incomplete taxon sampling . However, the full generic sampling of palm genera (100% of genera included) means that sampling of extant lineages is complete in the older parts of the phylogeny, becoming progressively more incomplete towards the present [71, 101]. In order to avoid misinterpretation of the plot we restrict our analyses to the accurately estimated part of the LTT, which will be least influenced by the missing taxa. We do this by finding the point at which incomplete taxon sampling will likely begin to have a significant effect on the LTT plot. Under the assumption that all genera are monophyletic, speciation events within each genus will always be found after the stem node of that genus. Given the known age of each stem node for each genus, we calculated a cumulative total for the number of missing species as a function of time. By doing this we generated a time-dependent curve representing the increase of missing taxa from the origin of the family until the present that quantified the amount of uncertainty each part of the LTT plot contains (Figure 3c). In this study, the proportion of missing taxa was less than 12% from 100 Ma to 24 Ma, at which time a dramatic increase occurred with missing species rising to 24% and on to > 92% at the present time. This point represents the stem age of the genus Calamus, the most species-rich genus in palms . All nodes occurring after this 24 Ma threshold were deleted from subsequent diversification analyses.
Two different approaches were used to test for significant changes in diversification rates. First, we used a maximum likelihood method for fitting alternative diversification models to the LTT plot  using the R package LASER 2.2 . The test statistic for diversification rate constancy ΔAICRC is calculated as ΔAICRC = AICRC - AICRv, where AICRc is the AIC score for the best fitting rate-constant diversification model, and AICRv is the AIC for the best fitting variable-rate diversification model. A negative value for ΔAICRC indicates that the data is best approximated by a rate-constancy model. We fitted five different diversification models: (1) the constant-rate birth model (the Yule process; ) with the speciation rate (λ) being constant and the extinction (μ) set to zero; (2) the constant-rate birth-death model with two parameters, speciation (λ) and extinction (μ); (3) a pure birth rate-variable model where the speciation rate λ1 shifts to rate λ2 at time ts, with three parameters (λ1, λ2, ts); (4) an exponential density-dependent speciation rate 'DDX' model; and (5) a logistic density-dependent speciation rate 'DDL' model. The significance of the observed ΔAICRC was evaluated by simulating 10,000 trees under a pure birth constant diversification rate.
Second, we calculated the γ statistic of Pybus and Harvey , which provides a summary of the distribution of nodes in the phylogeny: if the internal nodes are closer to the root then γ < 0; if they are closer to the tips then γ > 0; if the nodes are equally spread out then γ = 0. The observed γ statistic was compared with the distribution of the γ statistic of 1,000 simulated phylogenies under a pure-birth model using the LASER package 2.2.
A presence-absence matrix was built representing the global distribution of palm genera (see Additional file 1 for the original data used to perform this analysis). We defined seven, non-overlapping major palm areas that reflect the distribution and endemism of genera as well as broad scale geological units, as follows: (A) South America, (B) North America (including Central American and the Caribbean), (C) Africa (including Arabia); (D) Indian Ocean (Madagascar, Mascarenes, Comoros and Seychelles), (E) India (including Sri Lanka), (F) Eurasia (including west Malesia to the west of Wallace's Line) and (G) Pacific Ocean (including east Malesia to the east of Wallace's Line, Australia and the Pacific Islands). Each genus was assigned to one or more of the major palm areas based on its known current distribution [28, 105].
Ancestral areas (AA) were reconstructed using a maximum likelihood method under the dispersal-extinction-cladogenesis model [49, 50] as implemented in the software Lagrange build 20091004 . We tested our ancestral area reconstruction under two different biogeographic models (see Additional file 1 for the parameters used to perform these analyses). The first model (M0) was unconstrained and we assigned an equal probability (P = 1.00) of dispersal between all areas during the whole time period considered. This model assumes that spatial relationships among areas have no effect on biogeographical patterns. For the second model (M1), we applied a more complex biogeographic scenario incorporating prior information on range evolution as well as dispersal probabilities between areas given discrete periods of time. This model was based on past climatic data, tectonic history and presence/absence of postulated land bridges [14, 19, 20, 106–108]. Five time frames were delimited and dispersal probabilities were assigned between all adjacent areas (see Additional file 1). Dispersal probabilities were set as following: low or no dispersal = 0.01; low dispersal = 0.25; medium dispersal = 0.5; high dispersal = 0.75; areas adjacent or very close = 1. Scripts in the programming language Python http://www.python.org/ were generated using the online helper http://www.reelab.net/lagrange. Because of the large surface of each area (continent-level areas), the maximum number of ancestral areas was limited to two.
The genus Cocos presented a special problem because it is widely distributed across all areas. Such highly polymorphic states generally inject a high level of ambiguity into the analyses, as was the case in preliminary analyses here. Following the recommendations of Ronquist , we allocated a putative ancestral area coding to Cocos (area A) based on the findings of Meerow et al. .
The probable ancestral biome at the crown node of palms was reconstructed under a maximum likelihood method using the Markov k-state 1 parameter model(Mk1) model of character evolution implemented in Mesquite version 2.74 [10, 111]. We assigned genera into three different biome state categories following Olson et al. . State 0: predominantly 'tropical and subtropical moist broadleaf forests' biome (that is, TRF); state 1: 'mangrove'; state 2: a general category that contains genera not belonging to any of the two first categories (that is, not TRF-restricted). State 2 is broad in its definition encompassing all other biomes for palms. When a genus occurred in both TRF and non-TRF biomes it was coded as ambiguous (see Additional file 1 for the original data used to perform this analysis). The mangrove category was included to take into account the ecology of Nypa. The phylogenetic signal of the biome character was tested by randomising the tips of the phylogeny 1,000 times in Mesquite  in order to create a null distribution of the number of steps under the maximum parsimony criterion and the 99% confidence intervals. This null distribution was compared to the observed number of steps necessary to explain the occurrence of each character on the phylogeny. In this case, the observed value fell outside the 99% confidence interval confirming that the biome category is phylogenetically conserved.
We are grateful to Jim Doyle, Roy Erkens, Alan Graham, Madeline Harley, David Mabberley, Jim Miller, Bob Morley, Vincent Savolainen, Selena Smith and Gary Upchurch for their advice during the preparation of this manuscript. Alexander Antonelli and one anonymous reviewer are gratefully acknowledged for their comments and suggestions. Gregory Plunkett and Ken Carol kindly provided logistical support for our analyses. Tylor Durand is thanked for his help in preparing the photo of Figure 2. This work was supported by a grant from the UK Natural Environment Research Council to WJB.
The New York Botanical Garden
Royal Botanic Gardens
Institut de Recherche pour le Développement (IRD), UMR DIA-DE, DYNADIV researche group
Gentry A: Tropical forest biodiversity: distributional patterns and their conservational significance.Oikos 1992, 63:19–28.View Article
Whitmore TC: An introduction to tropical rain forests. Oxford, UK: Clarendon Press; 1998.
Hill JL, Hill RA: Why are tropical rain forests so species rich? Classifying, reviewing and evaluating theories.Prog Phys Geog 2001, 25:326–354.
Asner GP, Rudel TK, Aide TM, Defries R, Emerson R: A contemporary assessment of change in humid tropical forests.Conserv Biol 2009, 23:1386–1395.PubMedView Article
Pennington RT, Richardson JE, Lavin M: Insights into the historical construction of species-rich biomes from dated plant phylogenies, neutral ecological theory and phylogenetic community structure.New Phytol 2006, 172:605–616.PubMedView Article
Moritz C, Patton JL, Schneider CJ, Smith TB: Diversification of rainforest faunas: an integrated molecular approach.Annu Rev Ecol Syst 2000, 31:533–563.View Article
Jaramillo C, Rueda MJ, Mora G: Cenozoic plant diversity in the neotropics.Science 2006, 311:1893–1896.PubMedView Article
Jaramillo C, Hoorn MC, Silva S, Leite F, Herrera F, Quiroz L, Dino R, Antonioli L: The origin of the modern Amazon rainforest: implications from the palynological and paleobotanical record. In Amazonia, Landscape and Species Evolution. Edited by: Hoorn MC, Wesselingh FP. Oxford, UK: Blackwells; 2010:317–334.
Upchurch GR, Wolf JA: Mid-Cretaceous to early tertiary vegetation and climate: evidence from fossil leaves and woods. In The Origin of Angiosperms and Their Biological Consequences. Edited by: Friis EM, Chaloner WG, Crane PR. Cambridge, UK: Cambridge University Press; 1987:75–105.
Davis CC, Webb CO, Wurdack KJ, Jaramillo CA, Donoghue MJ: Explosive radiation of malpighiales supports a mid-Cretaceous origin of modern tropical rain forests.Am Nat 2005, 165:E36-E65.PubMedView Article
Johnson KR, Ellis B: A tropical rainforest in Colorado 1.4 million years after the Cretaceous-Tertiary boundary.Science 2002, 296:2379–2383.PubMedView Article
Wing SL, Herrera F, Jaramillo CA, Gómez-Navarro C, Wilf P, Labandeira CC: Late Paleocene fossils from the Cerrejón formation, Colombia, are the earliest record of neotropical rainforest.Proc Natl Acad Sci USA 2009, 106:18627–18632.PubMedView ArticlePubMed Central
Jacobs BF: Palaeobotanical studies from tropical Africa: relevance to the evolution of forest, woodland and savannah biomes.Philos Trans R Soc Lond, B 2004, 359:1573–1583.View Article
Morley RJ: Origin and Evolution of Tropical Rain Forests. New York, USA: John Wiley & Sons; 2000.
Raven HP, Axelrod DI: Angiosperm biodiversity and past continental movements.Ann Missouri Bot Gard 1974, 61:539–673.View Article
Wallace AR: Tropical Nature, and Other Essays. London, UK: Macmillian; 1878.View Article
Richards PW: The Tropical Rain Forest: an Ecological Study. 2nd edition. Cambridge, UK: Cambridge University Press; 1996.
Wolfe JA: Some aspects of plant geography of northern hemisphere during Late Cretaceous and Tertiary.Ann Missouri Bot Gard 1975, 62:264–279.View Article
Zachos J, Pagani M, Sloan L, Thomas E, Billups K: Trends, rhythms, and aberrations in global climate 65 Ma to present.Science 2001, 292:686–693.PubMedView Article
Morley R: Cretaceous and Tertiary climate change and the past distribution of megathermal rainforests. In Tropical Rainforest Responses to Climatic Changes. Edited by: Bush MB, Flenley J. Chichester, UK: Praxis Publishing; 2007:1–31.View Article
Wilf P, Cuneo NR, Johnson KR, Hicks JF, Wing SL, Obradovich JD: High plant diversity in Eocene South America: evidence from Patagonia.Science 2003, 300:122–125.PubMedView Article
Stebbins GL: Flowering Plants: Evolution Above the Species Level. Cambridge, MA: Harvard University Press; 1974.
Haffer J: Speciation in Amazonian forest birds.Science 1969, 165:131–137.PubMedView Article
Gentry A: Neotropical floristic diversity: phytogeographical connections between Central and South America, Pleistocene climatic fluctuations, or an accident of the Andean orogeny?Ann Missouri Bot Gard 1982, 69:557–593.View Article
Richardson JE, Pennington RT, Pennington TD, Hollingsworth PM: Rapid diversification of a species-rich genus of neotropical rain forest trees.Science 2001, 293:2242–2245.PubMedView Article
Janssens SB, Knox EB, Huysmans S, Smets EF, Merckx V: Rapid radiation ofImpatiens(Balsaminaceae) during Pliocene and Pleistocene: result of a global climate change.Mol Phylogenet Evol 2009, 52:806–824.PubMedView Article
Erkens RHJ, Chatrou LW, Maas JW, van der Niet T, Savolainen V: A rapid diversification of rainforest trees (Guatteria; Annonaceae) following dispersal from Central into South America.Mol Phylogenet Evol 2007, 44:399–411.PubMedView Article
Dransfield J, Uhl NW, Asmussen CB, Baker JW, Harley MM, Lewis CE: Genera Palmarum: The Evolution and Classification of Palms. Kew, UK: Kew Publishing; 2008.
Phillips O, Miller JS: Global Patterns of Plant Diversity: Alwyn H. Gentry's Forest Transect Data Set. Volume 84. Louis, MO: Missouri Botanical Press; 2002.
Kahn F, de Granville J-J: Palms in Forest Ecosystems of Amazonia. Berlin, Germany: Springer-Verlag; 1992.View Article
Bjorholm S, Svenning JC, Skov F, Balslev H: Environmental and spatial controls of palm (Arecaceae) species richness across the Americas.Glob Ecol Biogeogr 2005, 14:423–429.View Article
Kreft H, Sommer JH, Barthlott W: The significance of geographic range size for spatial diversity patterns in neotropical palms.Ecography 2006, 29:21–30.View Article
Tomlinson PB: The uniqueness of palms.Bot J Linn Soc 2006, 151:5–14.View Article
Tomlinson PB: Systematics and ecology of the Palmae.Annu Rev Ecol Syst 1979, 10:85–107.View Article
Kvacek J, Herman AB: Monocotyledons from the early Campanian (Cretaceous) of Grünbach, lower Austria.Rev Paleobot Palyno 2004, 128:323–353.View Article
Daghlian CP: A review of the fossil record of monocotyledons.Bot Rev 1981, 47:517–555.View Article
Crié L: Studies on the silicified palms of the Cretaceous formation of Anjou [translated from French].Bull Soc Etudes Sci Angers 1892, 21:97–103.
Anderson CL, Janssen T: Monocots. In The TimeTree of Life. Edited by: Hedges BS, Kumar S. New York, NY: Oxford University Press; 2009:203–212.
Janssen T, Bremer K: The age of major monocot groups inferred from 800 + rbcL sequences.Bot J Linn Soc 2004, 146:385–398.View Article
Wikström N, Savolainen V, Chase MW: Evolution of the angiosperms: calibrating the family tree.Proc Natl Acad Sci USA 2001, 268:2211–2220.
Savolainen V, Anstett MC, Lexer C, Hutton I, Clarkson JJ, Norup MV, Powell MP, Springate D, Salamin N, Baker WJ: Sympatric speciation in palms on an oceanic island.Nature 2006, 441:210–213.PubMedView Article
Trénel P, Gustafsson MHG, Baker WJ, Asmussen-Lange CB, Dransfield J, Borchsenius F: Mid-Tertiary dispersal, not Gondwanan vicariance explains distribution patterns in the wax palm subfamily (Ceroxyloideae: Arecaceae).Mol Phylogenet Evol 2007, 45:272–288.PubMedView Article
Gunn FB: The phylogeny of the Cocoeae (Arecaceae) with emphasis onCocos nucifera.Ann Missouri Bot Gard 2004, 91:505–522.
Cuenca A, Asmussen-Lange CB, Borchsenius F: A dated phylogeny of the palm tribe Chamaedoreeae supports Eocene dispersal between Africa, North and South America.Mol Phylogenet Evol 2008, 46:760–775.PubMedView Article
Roncal J, Borchsenius F, Asmussen-Lange CB, Balslev H: Divergence times in the tribe Geonomateae (Arecaceae) coincide with Tertiary geological events. In Diversity, Phylogeny and Evolution of Monocotyledons. Edited by: Seberg O, Peterson AT, Barfod A, Davis AM. Aarhus, Denmark: Aarhus University Press; 2010:245–265.
Baker WJ, Savolainen V, Asmussen-Lange CB, Chase MW, Dransfield J, Forest F, Harley MM, Uhl NW, Wilkinson M: Complete generic-level phylogenetic analyses of palms (Arecaceae) with comparisons of supertree and supermatrix approaches.Syst Biol 2009, 58:240–256.PubMedView Article
Drummond AJ, Ho SYW, Phillips MJ, Rambaut A: Relaxed phylogenetics and dating with confidence.PloS Biol 2006, 4:699–710.View Article
Ree RH, Moore BR, Webb CO, Donoghue MJ: A likelihood framework for inferring the evolution of geographic range on phylogenetic trees.Evolution 2005, 59:2299–2311.PubMed
Ree RH, Smith SA: Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis.Syst Biol 2008, 57:4–14.PubMedView Article
Crane PR, Friis EM, Pedersen KR: The origin and early diversification of angiosperms.Nature 1995, 374:27–33.View Article
Friis EM, Pedersen KR, Crane PR: When Earth started blooming: insights from the fossil record.Curr Opin Plant Biol 2005, 8:5–12.PubMedView Article
Upchurch GR, Wolfe JA: Cretaceous vegetation of the western interior and adjacent regions of North America. In Evolution of the Western Interior Basin. Volume 39. Edited by: Kaufman EG, Caldwell WGE. St. John's, Canada: Geological Association of Canada; 1993:243–281.
Uhl NW, Dransfield J: Genera Palmarum. Lawrence, KS, USA: Allen Press; 1987.
Tiffney BH: Seed size, dispersal syndromes, and the rise of the Angiosperms: evidence and hypothesis.Ann Missouri Bot Gard 1984, 71:551–576.View Article
Wing SL, Boucher LD: Ecological aspects of the Cretaceous flowering plant radiation.Ann Rev Earth Planet Sci 1998, 26:379–421.View Article
Burnham RJ, Johnson KR: South American palaeobotany and the origins of neotropical rainforests.Philos Trans R Soc Lond, B 2004, 359:1595–1610.View Article
Wolfe JA, Upchurch GR: North American nonmarine climates and vegetation during the late Cretaceous.Palaeogeogr Palaeoclimatol Palaeoecol 1987, 61:33–77.View Article
DiMichele WA, Gastaldo RA: Plant paleoecology in Deep Time1.Ann Missouri Bot Gard 2008, 95:144–198.View Article
Beerling DJ, Woodward Fl: Vegetation and the Terrestrial Carbon Cycle: Modelling the First 400 Million Years. Cambridge, UK: Cambridge University Press; 2001.View Article
Linder HP: The radiation of the Cape flora, southern Africa.Biol Rev 2003, 78:597–638.PubMedView Article
Linder HP: Evolution of diversity: the Cape flora.Trends Plant Sci 2005, 10:536–541.PubMedView Article
Lavin M, Herendeen PS, Wojciechowski MF: Evolutionary rates analysis of Leguminosae implicates a rapid diversification of lineages during the tertiary.Syst Biol 2005, 54:575–594.PubMedView Article
Couvreur TLP, Pirie MD, Richardson JE, Chatrou LW, Saunders RMK, Su YCF, Erkens RHJ: Early evolutionary history of the flowering plant family Annonaceae: steady diversification and boreotropical geodispersal.J Biogeogr 2011, 38:664–680.View Article
Bremer B, Eriksson T: Time tree of Rubiaceae: phylogeny and dating the family, subfamilies, and tribes.Int J Plant Sci 2009, 170:766–793.View Article
Nee S, May RM, Harvey PH: The reconstructed evolutionary process.Phil Trans R Soc Lond B 1994, 344:305–311.View Article
Nee S, Mooers AO, Harvey PH: Tempo and mode of evolution revealed from molecular phylogenies.Proc Natl Acad Sci USA 1992, 89:8322–8326.PubMedView ArticlePubMed Central
Harvey P, Pagel M: The Comparative Method in Evolutionary Biology. Oxford, UK: Oxford University Press; 1991.
Pybus OG, Harvey PH: Testing macro-evolutionary models using incomplete molecular phylogenies.Proc R Soc B Biol Sci 2000, 267:2267–2272.View Article
Cusimano N, Renner SS: Slowdowns in diversification rates from real phylogenies may not be real.Syst Biol 2010, 59:458–464.PubMedView Article
Ricklefs RE: Historical and ecological dimensions of global patterns in plant diversity.Biol Skr 2005, 55:583–603.
Hoorn C, Wesselingh FP, ter Steege H, Bermudez MA, Mora A, Sevink J, Sanmartín I, Sanchez-Meseguer A, Anderson CL, Figueiredo JP, Jaramillo C, Riff D, Negri FR, Hooghiemstra H, Lundberg J, Stadler T, Särkinen T, Antonelli A: Amazonia through time: Andean uplift, climate change, landscape evolution, and biodiversity.Science 2010, 330:927–931.PubMedView Article
Couvreur TLP, Chatrou LW, Sosef MSM, Richardson JE: Molecular phylogenetics reveal multiple tertiary vicariance origins of the African rain forest trees.BMC Biol 2008, 6:54.PubMedView ArticlePubMed Central
Schulte P, Alegret L, Arenillas I, Arz JA, Barton PJ, Bown PR, Bralower TJ, Christeson GL, Claeys P, Cockell CS, Collins GS, Deutsch A, Goldin TJ, Goto K, Grajales-Nishimura JM, Grieve RA, Gulick SP, Johnson KR, Kiessling W, Koeberl C, Kring DA, MacLeod KG, Matsui T, Melosh J, Montanari A, Morgan JV, Neal CR, Nichols DJ, Norris RD, Pierazzo E, et al.: The Chicxulub asteroid impact and mass extinction at the Cretaceous-Paleogene boundary.Science 2010, 327:1214–1218.PubMedView Article
Wilson R, Heinrichs J, Hentschel J, Gradstein SR, Schneider H: Steady diversification of derived liverworts under tertiary climatic fluctuations.Biol Lett 2007, 3:566–569.PubMedView ArticlePubMed Central
Venditti C, Meade A, Pagel M: Phylogenies reveal new interpretation of speciation and the Red Queen.Nature 2010, 463:349–352.PubMedView Article
McKenna DD, Farrell BD: Tropical forests are both evolutionary cradles and museums of leaf beetle diversity.Proc Natl Acad Sci USA 2006, 103:10947–10951.PubMedView ArticlePubMed Central
Dransfield J, Uhl NW, Asmussen CB, Baker JW, Harley MM, Lewis CE: A new phylogenetic classification of the palm family, Arecaceae.Kew Bull 2005, 60:559–569.
Dransfield J, Rakotoarinivo M, Baker WJ, Bayton RP, Fisher JB, Horn JW, Leroy B, Metz X: A new Coryphoid palm genus from Madagascar.Bot J Linn Soc 2008, 156:79–91.View Article
Harley MM: A summary of fossil records for Arecaceae.Bot J Linn Soc 2006, 151:39–67.View Article
Harley MM, Baker WJ: Pollen aperture morphology in Arecaceae: application within phylogenetic analyses, and a summary of the fossil record of palm-like pollen.Grana 2001, 40:45–77.View Article
Magallón SA: Dating lineages: molecular and paleontological approaches to the temporal framework of clades.Int J Plant Sci 2004, 165:S7-S21.View Article
Palmer AR, Geissman J: Geologic Time Scale. Boulder, CO: The Geological Society of America; 1999.
Berry EW: The Upper Cretaceous and Eocene floras of South California and Georgia. Volume 84. Reston, VA: United States Geological Survey; 1914.
Pan AD, Jacobs BF, Dransfield J, Baker WJ: The fossil history of palms (Arecaceae) in Africa and new records from the late Oligocene (28–27 Mya) of North-Western Ethiopia.Bot J Linn Soc 2005, 151:69–81.View Article
Schrank E: Palynology of the Yesomma Formation in Northern Somalia: a study of pollen spores and associated phytoplankton from the late Cretaceous Palmae Province.Palaeontogr Abt B Palaeophytol 1994, 231:63–112.
Rull V: Biogeographical and evolutionary considerations of Mauritia (Arecaceae), based on palynological evidence.Rev Paleobot Palyno 1998, 100:109–122.View Article
Campbell JD, Fordyce RE, Grebneff A, Maxwell PA: Fossil coconuts from mid-Cenozoic shallow marine sediments in southern New Zealand.Geol Soc New Zeal Misc Publ 2000, 108A:21.
Médus J: Palynology of the Tertiary sediments from meridional Senegal [translated from French].Pollen et Spores 1975, 17:545–601.
Gomez-Navarro C, Jaramillo C, Herrera F, Wing SL, Callejas R: Palms (Arecaceae) from a Paleocene rainforest of northern Colombia.Amer J Bot 2009, 96:1300–1312.View Article
Gee CT: On the fossil occurrence of the mangrove palmNypa. In Paleofloristics and Paleoclimatic Changes in the Cretaceous and Tertiary. Edited by: Knobloch E, Kvacek Z. Prague, Czech Republic: Geological Survey Publisher; 1990:315–319.
Ediger VS, Bati Z, Alisan C: Paleopalynology and paleoecology ofCalamus-like disulcate pollen grains.Rev Paleobot Palyno 1990, 62:97–105.View Article
Baker WJ, Dransfield J: Towards a biogeographic explanation of the calamoid palms. In Monocots: Systematics and Evolution. Edited by: Wilson KL, Morrison DA. Melbourne, Australia: CSIRO; 2000:545–553.
Muller J: Palynology of the Pedawan and Plateau sandstone formations (Cretaceous - Eocene) in Sarawak, Malaysia.Micropaleontology 1968, 14:1–37.View Article
Friis EM, Pedersen KR, Crane PR: Araceae from the Early Cretaceous of Portugal: evidence on the emergence of monocotyledons.Proc Natl Acad Sci USA 2004, 101:16565–16570.PubMedView ArticlePubMed Central
Suchard MA, Weiss RE, Sinsheimer JS: Bayesian selection of continuous-time Markov chain evolutionary models.Mol Biol Evol 2001, 18:1001–1013.PubMedView Article
Ricklefs RE: Phylogenetic perspectives on patterns of regional and local species richness. In Tropical Rainforests: Past, Present, & Future. Edited by: Bermingham E, Dick CW, Moritz C. Chicago, IL: University of Chicago Press; 2005:16–40.
Rabosky DL: Likelihood methods for detecting temporal shifts in diversification rates.Evolution 2006, 60:1152–1164.PubMed
Rabosky DL: LASER: a maximum likelihood toolkit for detecting temporal shits in diversification rates from molecular phylogenies.Evol Bioinform Online 2006, 2:257–260.
Yule GU: A mathematical theory of evolution based on the conclusions of Dr. J.C. Willis. Volume 213. Philos Trans R Soc Lond, B; 1924:21–87.
Govaerts R, Dransfield J: World checklist of palms. Kew, UK: Royal Botanic Gardens; 2005.
Tiffney BH, Manchester SR: The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the Northern Hemisphere tertiary.Int J Plant Sci 2001, 162:S3-S17.View Article
Ali JR, Aitchison JC: Gondwana to Asia: plate tectonics, paleogeography and the biological connectivity of the Indian sub-continent from the Middle Jurassic through latest Eocene (166–35 Ma).Earth Sci Rev 2008, 88:145–166.View Article
Meerow AW, Noblick L, Borrone JW, Couvreur TLP, Mauro-Herrera M, Hahn WJ, Kuhn DN, Nakamura K, Oleas NH, Schnell RJ: Phylogenetic analysis of seven WRKY genes across the Palm subtribe Attaleinae (Areceaceae) identifiesSyagrusas sister group of the Coconut.PLoS ONE 2009, 4:e7353.PubMedView ArticlePubMed Central
Olson DM, Dinerstein E, Wikramanayake ED, Burgess ND, Powell GVN, Underwood EC, D'Amico JA, Itoua I, Strand HE, Morrison JC, Loucks C, Allnutt T, Ricketts TH, Kura Y, Wettengel W, Kassem K: Terrestrial ecoregions of the world: a new map of life on Earth.Bioscience 2001, 51:933–938.View Article
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.