Evolution of Philodendron (Araceae) species along Neotropical biomes

Abstract

Philodendron is the second most diverse genus of the Araceae, a tropical monocot family with significant morphological diversity along its wide geographic distribution in the Neotropics. Although evolutionary studies of Philodendron were conducted in recent years, the phylogenetic relationship among its species remains unclear. Additionally, analyses conducted to date suggested the inclusion of all American representatives of a closely related genus, Homalomena, within the Philodendron clade. A thorough evaluation of the phylogeny and timescale of these lineages is thus necessary to elucidate the tempo and mode of evolution of this large Neotropical genus and to unveil the biogeographic history of Philodendron evolution along the Amazonian and Atlantic Rain Forests, as well as open dry forests of South America. To this end, we have estimated the molecular phylogeny for 68 Philodendron species, which consists of the largest sampling assembled to date aiming the study of the evolutionary affinities. We have also performed ancestral reconstruction of species distribution along biomes. Finally, we contrasted these results with the inferred timescale of Philodendron and Homalomena lineage diversification. Our estimates indicate that American Homalomena is the sister clade to Philodendron. The early diversification of Philodendron took place in the Amazon Forest from Early to Middle Miocene, followed by colonization of the Atlantic Forest and the savanna-like landscapes, respectively. Based on the age of the last common ancestor of Philodendron, the species of this genus diversified by rapid radiations, leading to its wide extant distribution in the Neotropical region.

Full text

A peer-reviewed version of this preprint was published in PeerJ on 24
March 2016.
View the peer-reviewed version (peerj.com/articles/1744), which is the
preferred citable publication unless you specifically need to cite this preprint.
Loss-Oliveira L, Sakuragui C, Soares MdL, Schrago CG. (2016) Evolution of
Philodendron (Araceae) species in Neotropical biomes. PeerJ 4:e1744
https://doi.org/10.7717/peerj.1744

Evolution of Philodendron (Araceae) species along Neotropical
biomes
Leticia Loss-Oliveira, Cassia CMS Sakuragui, Maria de Lourdes Soares, Carlos G Schrago
Philodendron is the second most diverse genus of the Araceae, a tropical monocot family
with significant morphological diversity along its wide geographic distribution in the
Neotropics. Although evolutionary studies of Philodendron were conducted in recent years,
the phylogenetic relationship among its species remains unclear. Additionally, analyses
conducted to date suggested the inclusion of all American representatives of a closely
related genus, Homalomena, within the Philodendron clade. A thorough evaluation of the
phylogeny and timescale of these lineages is thus necessary to elucidate the tempo and
mode of evolution of this large Neotropical genus and to unveil the biogeographic history
of Philodendron evolution along the Amazonian and Atlantic Rain Forests, as well as open
dry forests of South America. To this end, we have estimated the molecular phylogeny for
68 Philodendron species, which consists of the largest sampling assembled to date aiming
the study of the evolutionary affinities. We have also performed ancestral reconstruction of
species distribution along biomes. Finally, we contrasted these results with the inferred
timescale of Philodendron and Homalomena lineage diversification. Our estimates indicate
that American Homalomena is the sister clade to Philodendron. The early diversification of
Philodendron took place in the Amazon Forest from Early to Middle Miocene, followed by
colonization of the Atlantic Forest and the savanna-like landscapes, respectively. Based on
the age of the last common ancestor of Philodendron, the species of this genus diversified
by rapid radiations, leading to its wide extant distribution in the Neotropical region.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

1Evolution of Philodendron (Araceae) species along Neotropical biomes
2
3Leticia Loss-Oliveira1, Cassia M. Sakuragui2, Maria L. Soares3 and Carlos G. Schrago1*
4
51Department of Genetics and 2Department of Botany, Federal University of Rio de Janeiro, Rio
6de Janeiro, RJ, Brazil
73Instituto Nacional de Pesquisas da Amazônia, Manaus, AM, Brazil
8
9
10
11
12
13 *Address for correspondence:
14 Carlos G. Schrago
15 Universidade Federal do Rio de Janeiro
16 Instituto de Biologia, Departamento de Genética, CCS, A2-092
17 Rua Prof. Rodolpho Paulo Rocco, SN
18 Cidade Universitária
19 Rio de Janeiro, RJ
20 CEP: 21941-617
21 BRAZIL
22 Phone: +55 21 2562-6397
23 Phone: +55 21 4063-8278
24 Email: carlos.schrago@gmail.com
25
26
27
28
29 Running title: Diversification of Philodendron
30
31 Keywords: biogeography; South America; Amazon; Andes; dispersal; supertree
32
33
34
35
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

36 Abstract
37
38 Philodendron is the second most diverse genus of the Araceae, a tropical monocot family with
39 significant morphological diversity along its wide geographic distribution in the Neotropics.
40 Although evolutionary studies of Philodendron were conducted in recent years, the phylogenetic
41 relationship among its species remains unclear. Additionally, analyses conducted to date
42 suggested the inclusion of all American representatives of a closely related genus, Homalomena,
43 within the Philodendron clade. A thorough evaluation of the phylogeny and timescale of these
44 lineages is thus necessary to elucidate the tempo and mode of evolution of this large Neotropical
45 genus and to unveil the biogeographic history of Philodendron evolution along the Amazonian
46 and Atlantic Rain Forests, as well as open dry forests of South America. To this end, we have
47 estimated the molecular phylogeny for 68 Philodendron species, which consists of the largest
48 sampling assembled to date aiming the study of the evolutionary affinities. We have also
49 performed ancestral reconstruction of species distribution along biomes. Finally, we contrasted
50 these results with the inferred timescale of Philodendron and Homalomena lineage
51 diversification. Our estimates indicate that American Homalomena is the sister clade to
52 Philodendron. The early diversification of Philodendron took place in the Amazon Forest from
53 Early to Middle Miocene, followed by colonization of the Atlantic Forest and the savanna-like
54 landscapes, respectively. Based on the age of the last common ancestor of Philodendron, the
55 species of this genus diversified by rapid radiations, leading to its wide extant distribution in the
56 Neotropical region.
57
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

58 Introduction
59 Philodendron is an exclusively Neotropical genus, comprising 482 formally recognized
60 species (Boyce & Croat, 2013). Their geographic distribution range from Northern Mexico to
61 Southern Uruguay (Mayo et al., 1997), consisting mainly of the biomes of the Amazonian and
62 Atlantic Rain Forests and also the open dry forests of South America. According to Olson et al.’s
63 (2001) classification of terrestrial biomes, South American open dry forests are composed of the
64 Cerrado (savanna-like landscapes) and Caatinga biomes (Croat, 1997, Mayo, 1988, Mayo, 1989,
65 Sakuragui et al., 2012a) (Figure 1). Philodendron species richness is especially significant in
66 Brazil, where 168 species were described thus far (Sakuragui et al., 2012b).
67 Although Philodendron presents a significant morphological plasticity, wide leaf
68 variation and several types of habits (Sakuragui et al., 2012b, Coelho, 2000), the inflorescence
69 morphology of its representatives is largely conserved. The unisexual flowers in the spadix are
70 clustered in male, female and sterile zones; located at the basal, median and superior portions,
71 respectively (Figure 1b). The spadix, in nearly all of its extension, is surrounded by the spate
72 (Sakuragui, 2001).
73 Currently, Philodendron species are grouped into three subgenera according to its floral
74 and vegetative morphology and anatomy (Mayo, 1991, Mayo, 1988, Croat, 1997), namely,
75 subgenus Meconostigma (Schott) Engl., which consists of 21 species (Gonçalves & Salviani,
76 2002, Croat et al., 2002, Mayo, 1991); subgenus Pteromischum (Schott) Mayo, with 75 species
77 (Coelho, 2000) and subgenus Philodendron (Mayo, 1986), comprising approximately 400
78 species (Coelho, 2000, Croat, 1997).
79 Because of the wide geographic range, patterns of distribution along niches, as well as the
80 characteristic morphology, interest in investigating Philodendron systematics and evolution has
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

81 increased in the last decades (Sakuragui et al., 2005, Mayo, 1986, Grayum, 1996, Croat, 1997).
82 Morphological and anatomical characters of flowers has been of special interest for phylogenetic
83 analysis due to their high level of variability (Sakuragui, 1998). However, the plasticity and
84 convergence of these characters in Philodendron may increase the probability of homoplasies
85 (Mayo, 1986, Mayo, 1989).
86 Recently, Gauthier et al. (2008) investigated the phylogenetic relationships of
87 Philodendron species based on three molecular markers, sampling a total of 49 species. This
88 work comprised the largest taxon sampling of the genus to date. In accordance to previous
89 analysis (Barabé et al., 2002, Mayo et al., 1997), authors questioned the monophyly of
90 Philodendron, suggesting the inclusion of all American species of the morphologically similar
91 genus, Homalomena Schott, within the Philodendron clade. Homalomena species occur in
92 America and Asia and its geographic distribution partly overlaps with Philodendron in the
93 Neotropics. The inference of the evolutionary relationships between Philodendron and
94 Homalomena has a significant biogeographic appeal. If American Homalomena species are
95 indeed more closely related to Philodendron than to Asian Homalomena, a single colonization
96 event should be considered. Unveiling the evolutionary relationships between those lineages is
97 thus necessary to elucidate their origin and subsequent diversification.
98 Besides phylogeny, several issues regarding Philodendron evolution remain unclear. For
99 example, the historical events that led to the wide geographic occurrence along biomes need a
100 thorough analysis. In this sense, investigating the evolutionary affinities of a large sample of
101 Philodendron species will shed light on how this lineage diversified along the Amazonian and
102 Atlantic Rain Forests, as well as South American open dry forests biomes, namely, the Cerrado
103 and Caatinga. To this end, we have performed an ancestral area reconstruction of Philodendron
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

104 and Homalomena species and estimated the divergence times from a phylogeny inferred from the
105 largest Philodendron dataset composed to date. We were able to address the timing and pattern
106 of Philodendron diversification in selected Neotropical biomes with a focus on the evolutionary
107 relationships between the three Philodendron subgenera.
108
109
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

110 Materials and Methods
111 Taxon and gene sampling
112 We have sequenced new data for 110 extant species of Philodendron and 16 species of
113 Homalomena of the following molecular markers: the nuclear 18S and external transcribed
114 spacer (ETS); and the chloroplast trnL intron, trnL-trnF intergenic spacer, the trnK intron and
115 maturase K (matK) genes. Additionally, 13 outgroup species were analyzed, comprising the
116 genera Cercestis, Culcasia, Colocasia, Dieffenbachia, Heteropsis, Montrichardia, Nephthytis,
117 Furtadoa and Urospatha. Outgroup choice was based on the close evolutionary affinity of these
118 genera to Philodendron, as suggested by previous studies. The complete list of species included
119 in this study, the voucher and GenBank accession numbers were listed in Tables 1 and 2 of the
120 Supplementary Material.
121 Ancestral biome reconstruction is dependent on the estimated phylogeny and the
122 current geographic distribution of sampled species terminals. Thus, taxon sampling may impact
123 the inference of ancestral species distribution along biomes. As indicated in Table 1
124 (Supplementary Material), we have sampled all P. subg. Meconostigma species; 82 P. subg.
125 Philodendron species and 7 P. subg. Pteromischum species. Our sampling strategy is
126 representative of the current Philodendron diversity. Although ~75% of the sampled species are
127 P. subg. Philodendron in our analysis, ~82% of Philodendron species consist of P. subg.
128 Philodendron (Boyce & Croat, 2013, Sakuragui et al., 2012a).
129
130 DNA isolation, amplification and sequencing
131 Genomic DNA was isolated with QIAGEN DNeasy Blood & Tissue kit from silica-dried
132 or fresh leaves. Primers used for amplification and sequencing were listed in Table 3 of
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

133 Supplementary Material. Sequencing reactions were performed in the Applied Biosystems
134 3730xl automatic sequencer and edited with the Geneious 5.5.3 software.
135 Alignment and phylogenetic analysis
136 Molecular markers were individually aligned in MAFFT 7 (Katoh & Standley, 2013) and
137 then manually adjusted in SeaView 4 (Gouy et al., 2010). We estimated individual gene trees
138 (Fig. 1, SM) for each molecular marker in MrBayes 3.2.2 (Huelsenbeck & Ronquist, 2001,
139 Ronquist & Huelsenbeck, 2003) using the GTR + G substitution model. The Markov chain
140 Monte Carlo (MCMC) algorithm was ran twice for 10,000,000 generations, using four chains.
141 Chains were sampled every 100th cycle and a burn-in of 20% was applied. A supertree was
142 estimated from the tree topologies of each molecular marker using the PhySIC_IST algorithm,
143 available at the ATGC-Montpellier online server (http://www.atgc-montpellier.fr/physic_ist/).
144 Only clades with posterior probability  85% were considered to estimate the supertree. We have
145 used this approach to avoid the impact of missing data in phylogeny estimation (Scornavacca et
146 al., 2008). As PhySIC_IST calculates non-plenary supertrees, it removes taxa with significant
147 topological conflict and/or with small taxon sampling (Scornavacca et al., 2008). The final
148 supertree was thus composed of 89 terminals, as 50 terminals were discarded due to conflicting
149 resolutions.
150 In order to assess the stability of the (Philodendron + American Homalomena) clade, we
151 have calculated the log-likelihoods of alternative topological arrangements in PhyML 3.0
152 (Guindon et al., 2009) using the species sampling of the supertree. We have tested the following
153 topologies: (I) (American Homalomena (P. subg. Philodendron + P. subg. Meconostigma); (II)
154 (P. subg. Meconostigma (P. subg. Philodendron + American Homalomena) and (III) (P. subg.
155 Philodendron (P. subg. Meconostigma + American Homalomena). The significance of the
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

156 difference in log-likelihoods between topologies was tested with the approximately unbiased
157 (AU) and the Shimodaira-Hasegawa (SH) tests implemented in CONSEL 1.2.0 (Shimodaira &
158 Hasegawa, 2001).
159 Divergence time inference
160 Dating Philodendron evolutionary history is difficult mainly because of the scarcity of
161 the fossil record (Loss-Oliveira et al., 2014). For instance, Dilcher and Daghlian (1977), based
162 on fossilized leaves, described a putative P. subg. Meconostigma fossil from the Eocene of
163 Tennessee (56.0 – 33.9 Ma). However, Mayo (1991) identified the referred fossil as a Peltranda.
164 Thus, we have decided not to use this fossil as calibration information. Alternatively, in order to
165 estimate divergence times, we have assigned a prior on the rate of nucleotide substitution. We
166 were then prompted to infer the evolutionary rates of plastid coding regions of monocots using a
167 large sample of publicly available chloroplast genomes. Nuclear genes were excluded from
168 dating analysis because of the absence of prior information on evolutionary rates.
169 To estimate monocots substitution rate, we used chloroplast genomes from 154 Liliopsida
170 species retrieved from the GenBank (Table 4). All orthologous coding regions were concatenated
171 into a single supermatrix. Maximum likelihood phylogentic reconstruction was implemented in
172 RaxML 7.0.3 (Stamatakis, 2006) under GTR model. Molecular dating of monocots (Liliopsida)
173 was conducted under a Bayesian framework, using fossil information obtained from Iles et al.
174 (Iles et al., 2015) (Table 5). Because the number of terminals used was large, rate estimation was
175 conducted with the MCMCTree program of PAML 4.8 package (Yang, 2007) using the
176 approximate likelihood calculation (dos Reis & Yang, 2011) and the uncorrelated model of
177 evolution of rates. In MCMCTree, posterior distributions were obtained via MCMC; chains were
178 sampled every 500th cycle until 50,000 trees were collected. We performed two independent
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

179 replicates to check for convergence of the estimates. Calibration information for Liliopsida was
180 entered as minimum and maximum bounds of uniform priors. The estimated mean substitution
181 rate was inferred at 3.26 x 10-9 substitutions/site/year (s/s/y). This value is significantly higher
182 than the previous estimate of Palmer (1991), which reported an average substitution rate of 0.7 x
183 10-9 s/s/y for angiosperm platids. As the credibility interval of our estimate was large, we
184 adopted a Gaussian prior for evolutionary rates with a 95% highest probability density (HPD)
185 interval that included maximum and minimum values of our estimate and that of Palmer’s.
186 Dating analysis of Philodendron and Homalomena species was performed in BEAST
187 using a relaxed molecular clock with evolutionary rates modeled by an uncorrelated lognormal
188 distribution; the GTR + Gmodel of sequence was applied. MCMC algorithm was ran for
189 50,000,000 generations and sampled every 1,000th cycle, with a burn-in of 20%.
190
191 Biome shifts
192 To unveil how Philodendron species colonized the Amazon forest, Atlantic Forest,
193 Cerrado and Caatinga, we conducted a Bayesian Binary MCMC (BBM) (Yu et al., 2012,
194 Ronquist & Huelsenbeck, 2003) implemented in Reconstruct Ancestral State in Phylogenies 2.1b
195 (RASP) software (Yu et al., 2012). The input tree topology was the supertree estimated in
196 PhySIC_IST. BBM chains were ran for 10,000,000 generations and were sampled every 1000th
197 cycle. State frequencies were estimated under the F81 model with gamma rate variation.
198 Information on the occurrence of each Philodendron species along Neotropical biomes was
199 obtained from Sakuragui et al. (2012b) and from the (Team) CATE Araceae (http://araceae.e-
200 monocot.org).
201
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

203 Results
204 The Homalomena genus was not recovered as monophyletic; the Asian Homalomena
205 clustered within a single group and the American representatives clustered independently, as
206 sister to Philodendron species (Figure 2). Although our analysis failed to support the monophyly
207 of Philodendron with significant statistical support, the topological arrangement in which
208 Philodendron is a monophyletic genus was significantly supported by the AU and SH tests (p <
209 0.05, Figure 3, Table 6SM). Within Philodendron, subg. Meconostigma was recovered as
210 monophyletic (Fig. 2, node D), whereas subg. Philodendron was recovered as polyphyletic (Fig.
211 2, node E). Finally, the monophyly of P. subg. Pteromischum was not inferred, because
212 Pteromischum species clustered with P. subg. Philodendron species.
213 We estimated that the last common ancestor (LCA) of Philodendron diversified in the
214 Amazon Forest (Fig. 4, node B) at ca. 8.6 Ma (6.8 – 12.1Ma) 95% HPD. Thus, we inferred that
215 the LCA of Philodendon diversified from Middle to Late Miocene. This also suggests that the
216 divergence between Philodendron and the American Homalomena occurred in a short period of
217 time after this American lineage diverged from the Asian Homalomena (Figure 4, nodes B and
218 A, respectively).
219 The earliest events of Philodendron diversification occurred exclusively in the Amazon
220 Forest (e.g., Fig. 4, nodes C, D, E, F). The ancestors of Atlantic Forest lineages were inferred to
221 have been distributed in the Amazon (Fig. 4, nodes I, J and nodes G, H). This pattern of
222 Amazonian ancestry of Atlantic Forest lineages was also observed in some terminal branches.
223 For instance, from node K to P. loefgrenii and from node L to P. imbe.
224 On the other hand, the majority of Cerrado species evolved from Atlantic Forest
225 ancestors (Fig. 4, nodes J and M; node N to P. rhizomatosum and P. pachyphyllum). In subgenus
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

226 Meconostigma, the age of early species diversification into Atlantic Forest was dated at 3.7 Ma
227 (5.6 – 2.7 Ma) (Fig. 4, node J), whereas in the P. subg. Philodendron early lineage
228 diversification occurred at 4.1 Ma (5.5 – 3.0 Ma) (Fig. 4, node J). Therefore, during a period of
229 5.0 – 6.0 Ma, Philodendron species occupied exclusively the Amazon Forest. The diversification
230 into Cerrado biome occurred later, at approximately 1.7 Ma (3.3 – 1.1 Ma) (Fig. 4, node M).
231
232
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

233 Discussion
234 Phylogenetic relationship between Philodendron and Homalomena
235 In this study, Asian Homalomena was recovered as sister to the (Philodendron +
236 American Homalomena) clade, and Furtadoa mixta clustered with the Asian Homalomena clade.
237 The evolutionary affinities of American Homalomena, P. subg. Meconostigma and P. subg.
238 Philodendron were not strongly supported. However, the topological arrangement in which
239 Philodendron is a monophyletic genus was statistically significant by the AU and SH tests,
240 suggesting the monophyly of Philodendron.
241 Previous studies have reported conflicting results concerning the monophyly of
242 Philodendron and the phylogenetic status of American Homalomena (Figure 5). For instance,
243 Barabé et al. (2002), based on the trnL intron and the trnL-trnF intergenic spacer, proposed P.
244 subg. Philodendron as a paraphyletic group and was unable to solve the (P. subg. Meconostigma
245 + Asian + American Homalomena) polytomy (Figure 5A). Gauthier et al. (2008) recovered the
246 American Homalomena as sister to Philodendron and the Asian Homalomena as sister to the
247 (American Homalomena + Philodendron) clade, although their Bayesian analysis inferred a
248 paraphyletic Philodendron, with P. subg. Pteromischum sister to the American Homalomena
249 (Figure 5B and 5C, respectively). Alternatively, Cusimano et al. (2011) recovered a
250 monophyletic Philodendron, with Homalomena as sister lineage of Furtadoa (Figure 5D).
251 Recently, Yeng et al. (2013) estimated the Homalomena phylogeny based on the nuclear ITS
252 marker and also sampled Philodendron species. In the ML and Bayesian trees reported in their
253 study, P. subg. Pteromischum was closely related to the American Homalomena, whereas P.
254 subg. Meconostigma and P. subg. Philodendron were recovered as sister taxa (Figure 5E).
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

255 Discrepancies between previous works and our analysis may be due to different choice of
256 phylogenetic methods, markers and taxon sampling. Gauthier et al. (2008) was the only study
257 intended to investigate specifically the systematics of Philodendron genus. When compared to
258 their analysis, our study included a larger sampling of taxa and molecular markers with the aim
259 of estimating the phylogeny of Philodendron and Homalomena species; it is also the first
260 analysis that used a supertree approach to this end.
261 Our phylogeny characteristically presents short branch lengths within the Philodendron
262 clade. The high frequency of polytomies indicates the genetic similarity among terminals, which
263 is further corroborated by the ease in obtaining artificial hybrids between different species. This
264 corroborates a scenario of low genetic differentiation and low reproductive isolation (Carlsen,
265 2011).
266 Philodendron diversification may also consist of several recent rapid radiation events.
267 Phylogenetic reconstruction under this scenario is challenging, because of a significant amount
268 of substitutions is needed to accumulate within short periods of time (Maddison and Knowles,
269 2006). However, morphological variation of Philodendron is remarkable, which seems
270 contradictory considering the previously discussed features. However, it has been extensively
271 discussed that morphological variation is not a suitable proxy for genetic variation (e.g.,
272 Prud’Homme et al., 2011; Houle et al., 2010). Many environmental and epigenetic factors may
273 can increase phenotypic variation even in the absence of DNA sequence variation (Prud’Homme
274 et al., 2011). Evidently, we cannot rule out the possibility that DNA regions that present
275 significant genetic differences between species were not sampled in this work.
276 Diversification of Philodendron and Homalomena
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

277 Although the chronology of Philodendron divergence was not extensively focused by
278 previous studies, Nauheimer et al. (2012) analyzed the global history of the entire Araceae
279 family based on a supermatrix composed of 5 chloroplast markers and several well-established
280 calibration points. Their analysis included a single Philodendron species and estimated age of the
281 Philodendron/Asian Homalomena divergence at approximately 20.0 Ma (ranging from 31.0 –
282 9.0 Ma). This study, however, also included a single species of Asian Homalomena.
283 The wide range of the posterior distribution credibility intervals of Nauheimer et al.
284 (2012) hampers the proposition of putative biogeographic scenarios for the evolution of
285 Philodendron, American and Asian Homalomena. Differences between their timescale and the
286 divergence times proposed in this study might therefore be due to methodological differences
287 caused by their reduced taxonomic sampling. Nevertheless, both our estimate of the age of the
288 Philodendron divergence from Asian Homalomena and that of Nauheimer et al. (2012) suggests
289 that this event took place when South America was essentially an isolated continent.
290 The isolation of the South American continent persisted from approximately 130.0 Ma
291 (Smith & Klicka, 2010) to 3.5 Ma (Vilela et al., 2014), with the rise of the Panamanian land
292 bridge. Therefore, from the Early to Middle Miocene there was no land connection with North
293 America, Asia or Africa (Oliveira et al., 2010). If dispersal, rather then vicariance, is the most
294 plausible hypothesis to explain Philodendron and American Homalomena colonization of the
295 Neotropics, hypotheses on the possible routes of colonization should be investigated. Based on
296 the continental arrangement during the Miocene, we propose that the dispersal of Philodendron
297 and American Homalomena ancestor could have followed four possible routes (Figure 6): (1)
298 from Asia to North America through the Bering Strait; (2) from Africa to the Neotropics by
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

299 crossing the Atlantic ocean; (3) from Asia to Neotropics by crossing Pacific ocean; and (4) from
300 Asia to Neotropics , also by crossing the Atlantic ocean.
301 The Araceae fossil record is currently assigned to Florida, Russia, Germany, United
302 Kingdom, Venezuela, Yemen, Colombia and Canada (Shufeldt, 1917, Berry, 1936, Bogner et al.,
303 2005, Chandler, 1964, Dorofeev, 1963, As-Saruri et al., 1999, Wilde & Frankenhauser, 1998,
304 2005, Wing et al., 2009, Stockey et al., 2007). However, as none of the fossil specimens was
305 described as closely related to Philodendron or Homalomena, the Araceae fossil record fails to
306 corroborate any dispersal hypothesis in particular.
307 Considering route 1, although the Bering Strait have connected Asia to the North
308 America during most of the Cenozoic period (Butzin et al., 2011), there is no evidence of extant
309 Philodendron and Homalomena in North America or North Asia. Route 2 involves long distance
310 oceanic dispersal through ca. 2,000 km – the minimum distance between Africa and the
311 Neotropics (Oliveira et al., 2010) – through Atlantic paleocurrents, which were probably stronger
312 than Pacific currents. This hypothesis is congruent with the clustering of Philodendron and
313 American Homalomena into a single clade, assuming Africa as the center of diversification of
314 Asian and American Homalomena, as well as Philodendron. However, we should conisder that
315 the last recent common ancestor of Philodendron and Homalomena was distributed in Africa. On
316 the other hand, this hypothesis is corroborated by the distribution of the extant Philodendron and
317 Homalomena species. Givnish and colleagues (2004) also suggested two long distance dispersal
318 events through the Atlantic, but in the opposite direction. Their analysis indicated that
319 Bromeliaceae and Rapateaceae arose in the Guayana Shield of northern South America and reached
320 tropical west Africa via long-distance dispersal at ca. 6–8 Ma.
321 When considering long distance dispersal events, it is crucial to evaluate their viability as
322 related with the plant’s ability to produce dispersal structures that would tolerate aquatic and
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

323 saline conditions for long periods of time (Lo et al., 2014). Although such features have not been
324 evualuated for Philodendron and Homalomena, some Homalomena species inhabits swamp
325 forests and open swamps. Thus, features that would favor their survival in waterlogged
326 environments could also influence their maintenance in seawater.
327 Although route 3 is geographically unlikely due to the 8,000 km distance between Asia
328 and the Neotropics through the Pacific Ocean (Oliveira et al., 2010), it cannot be completely
329 discarded, because it is corroborated by the extant distribution of Homalomena and
330 Philodendron. Finally, route 4 suggests the dispersal through the Atlantic ocean from Asia to the
331 Neotropics. This is also an improbable hypothesis, because the African continent would act as a
332 barrier between Asia and the Neotropics, requiring the dispersal through both the Indian and the
333 Atlantic oceans.
334 The extant distribution of Philodendron and Homalomena species and the scarcity of
335 fossil information challenge the proposition of a scenario for the origin of Philodendron and
336 American Homalomena in the Neotropics. However, the biological and geographical information
337 provided to date indicates a long distance oceanic dispersal through the Atlantic, as suggested by
338 route 2, as the most plausible hypothesis to explain Philodendron and American Homalomena
339 colonization of the Neotropics.
340 Early diversification of Philodendron species
341 According to our analysis, the last common ancestor of Philodendron and the American
342 Homalomena was distributed in the Amazon Forest about 8.6 Ma (11.1 – 6.8 Ma) during the
343 Middle/Late Miocene. Interestingly, this time estimate is very close to the age of the divergence
344 between the (Philodendron/American Homalomena) clade from the Asian Homalomena (Fig. 4,
345 node A). The Middle and Late Miocene were characterized by wetland expansion into western
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

346 Central Amazonia, which fragmented the rainforest and formed extensive wetlands (Jaramillo et
347 al., 2010). According to our analysis, Philodendron earliest divergence events took place in this
348 scenario. The Amazon forest, from the Late Miocene to the beginning of Pliocene, was
349 composed of a diverse and well-structured forest. The Amazon river landscape was well
350 established, this probably allowed the extensive development of the Amazonian terra firme
351 forest (Jaramillo et al., 2010). This scenario is compatible with the biology of extant species of
352 Philodendron, because a well-structured forest would allow the development of epiphyte and
353 hemiepiphyte species, such as Philodendron.
354 Philodendron diversification along Neotropical biomes
355 Our results suggest that Philodendron species occurred exclusively at the Amazon forest
356 for ca. 5.0 – 6.0 Ma. During the Pliocene, as result of the glacial cycles, climate cooling and
357 drying permitted the expansion of the open savanna areas, mostly represented by the ‘dry
358 diagonal’, which is constituted by the Caatinga, Cerrado and Chaco biomes. This consisted of a
359 crucial event, because it resulted in the isolation of the Atlantic forest in the east coast of South
360 America (DaSilva & Pinto-da-Rocha, 2013), which is synchronous to the inferred age of the
361 early diversification of Philodendron in this biome. This also corroborates the hypothesis that the
362 Atlantic forest taxa present a closer biogeographic relationship with the Amazon forest, as
363 proposed by Amorim and Pires (1996) and Eberhard and Bermingham (2005). After the
364 separation between Atlantic and Amazon Forests during the Pliocene, species dispersal might
365 have been common through the forest patches (DaSilva and Pinto-da-Rocha, 2013).
366 Roig-Juñent and Coscarón (2001) and Porzecanski and Cracraft (2005) suggested that the
367 Atlantic rainforest also presents similarities in organismal composition with the Cerrado biome.
368 This association would have been a result of dispersal events through gallery forests. The history
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

369 of the formation of Cerrado biome is still uncertain (Zanella, 2013, Werneck, 2011), but our
370 analysis indicated that the ancestors of Philodendron clades from the Cerrado were distributed in
371 the Atlantic forest. Therefore, we also corroborate the hypothesis of lineage dispersal from the
372 Atlantic Forest to the Cerrado biome. These events would have occurred after the colonization
373 the Atlantic Forest by Philodendron species.
374
375 Final considerations on Philodendron evolution
376 Given the significant morphological diversity of Philodendron, its widespread
377 distribution in the Neotropics and the age of the Araceae family (~140.0 Ma, Nauheimer et al.,
378 2012), it would be expected that the origin of this genus was older. In sharp contrast, we have
379 estimated phylogenies with very short branch lengths and very recent divergence times. A
380 similar scenario was reported by Carlsen and Croat (2013) for Anthurium, which is the most
381 diverse Araceae genus, and also by Nagalingum and colleagues (2011) for cycads. Therefore, the
382 inferred tempo and mode of evolution of Philodendron species were reported in several plant
383 families.
384
385 Conclusion
386 The present work was the first attempt to establish a chronological background for the
387 diversification of this highly diverse genus and to suggest possible routes of colonization of the
388 ancestors of Neotropical Philodendron and Homalomena. Philodendron was statistically
389 supported as a monophyletic genus, sister to American Homalomena by AU and SH tests. The
390 last common ancestor of Philodendon diversified from the Middle to the Late Miocene in the
391 Amazon Forest, where the earliest events of Philodendron diversification occurred. Amazon was
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

392 also the exclusive biome occupied by Philodendron species during a 5.0 – 6.0 million years
393 period. Atlantic Forest lineages of P. subg. Meconostigma and P. subg. Philodendron diverged
394 from Amazonian ancestors. On the other hand, the majority of Cerrado species evolved from
395 Atlantic Forest ancestors, from the Late Miocene to the Pliocene.
396
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

397 Acknowledgements
398 We thank Petrobrás and INPA for allowing field expeditions in their biological reserves. We also
399 thank Alexandre Antonelli for valuable contributions in the manuscript text.
400
401 References
402
403 Amorim, D. S. & Pires, M. R. S. (1996) Neotropical biogeography and a method for maximum
404 biodiversity estimation. (Bicudo, C. E. M. & Menezes, N. A., eds.). pp. 183-219. CNPq,
405 Campos do Jordão.
406 As-Saruri, M. L., Whybrow, P. J. & Collinson, M. E. 1999. Geology, fruits, seeds, and
407 vertebrates (Sirenia) from the Kaninah Formation (Middle Eocene), Republic of Yemen.
408 Fossil Vertebrates of Arabia: 443-453
409 Barabé, D., Bruneau, a., Forest, F. & Lacroix, C. 2002. The correlation between development of
410 atypical bisexual flowers and phylogeny in the Aroideae (Araceae). Plant Systematics
411 and Evolution 232: 1-19.
412 Berry, E. W. 1936. Tertiary plants from Venezuela. Proceedings of the United States National
413 Museum 83: 335-360.
414 Bogner, J., Hoffman, G. L. & Aulenback, K. R. 2005. A fossilized aroid infructescence,
415 Albertarum pueri geno. nov. et sp. nov., of Late Cretaceous (late Campanian) age from
416 the Horseshoe Canyon Formation of southern Alberta, Canada. Canadian Journal of
417 Earth Sciences 85: 591-598.
418 Boyce, P. & Croat, T. (2013) The Überlist of Araceae: Totals for published and estimated
419 number of species in aroid genera. pp. 7-9.
420 Butzin, M., Lohmann, I. G. & Bickertl, T. 2011. Miocene ocean circulation inferred from marine
421 carbon cycle modeling combined with benthic isotope records. Paleoceanography 26.
422 Carlsen, M. (2011) Understanding the origin and rapid diversification of the genus
423 Anthurium Schott (Araceae), integrating molecular phylogenetics, morphology and
424 fossils. Vol. Ph.D. pp. 159. University of Missouri, USA.
425 Carlsen, M. M. & Croat, T. B. 2013. A Molecular Phylogeny of the Species-Rich Neotropical
426 Genus <I>Anthurium</I> (Araceae) based on Combined Chloroplast and Nuclear DNA.
427 Systematic Botany 38: 576-588.
428 Chandler, M. E. J. (1964) The Lower Tertiary Floras of Southern England. In: A summary and
429 survey of findings in the light of recent botanical observations, Vol. 4. pp. British
430 Museum (Natural History), London.
431 Coelho, M. A. N. 2000. Philodendron Schott (Araceae): morfologia e taxonomia das espécies da
432 Reserva Ecológica de Macaé de Cima - Nova Friburgo, Rio de Janeiro, Brasil.
433 Rodriguesia 51: 21-68.
434 Cohen, K.M., Finney, P.L., Gibbard, P.L. & Fan, J.-X. 2015. The ICS International
435 Chronostratigraphic Chart. Episodes 36 (3).
436 Croat, T. B. 1997. A revision of Philodendron subgenus Philodendron (Araceae) for Mexico and
437 Central America. Annals of the Missouri Botanical Garden 84: 311-704.
438 Croat, T. B., Mayo, S. J. & Boss, J. 2002. A new species of Brazilian Philodendron subgenus
439 Meconostigma (Araceae). Aroideana 25: 63-66.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

440 Cusimano, N., Bogner, J., Mayo, S. J., Boyce, P. C., Wong, S. Y., Hesse, M., Hetterscheid, W.
441 L. a., Keating, R. C. & French, J. C. 2011. Relationships within the Araceae: Comparison
442 of morphological patterns with molecular phylogenies. American Journal of Botany 98:
443 654-668.
444 DaSilva, M. & Pinto-da-Rocha, R. (2013) História biogeográfica da Mata Atlântica: Opiliões
445 (Arachnida) como modelo para sua inferência. In: Biogeografia da América do Sul –
446 Padrões e Processos, (Carvalho, C. J. B. & Almeida, E. A. B., eds.). pp. 306. Roca.
447 Dilcher, D. L. & Daghlian, C. P. 1977. Investigations of Angiosperms from the Eocene of
448 Southeastern North America: Philodendron Leaf Remains. American Journal of Botany
449 64: 526-526.
450 Dorofeev, P. I. 1963. Tretichnye Flory Zapadnoi Sibiri (Tertiary Floras of Western Siberia).
451 Moskva, Izd-vo Akademii nauk SSSR.
452 dos Reis, M. & Yang, Z. H. 2011. Approximate Likelihood Calculation on a Phylogeny for
453 Bayesian Estimation of Divergence Times. Molecular Biology and Evolution 28: 2161-
454 2172.
455 Eberhard, J. R. & Bermingham, E. 2005. Phylogeny and comparative biogeography of
456 Pionopsitta parrots and Pteroglossus toucans. Molecular Phylogenetics and Evolution 36:
457 288-304.
458 Gauthier, M.-p. L., Barabe, D. & Bruneau, A. 2008. Molecular phylogeny of the genus
459 <i>Philodendron</i> (Araceae): delimitation and infrageneric classification. Botanical
460 Journal of the Linnean Society 156: 13-27.
461 Givnish, T.J., Millam, K.C., Evans, T.M. J.C., Hall, J.C., Pires , Berry, P.E. & Sytsma K.J. 2004.
462 Ancient vicariance or recent long-distance dispersal? Inferences about phylogeny and
463 South American-African disjunction in Rapateaceae and Bromeliaceae based on ndhf
464 sequence data. International Journal of Plant Sciences 165: 35–54.
465 Gonçalves, E. G. & Salviani, E. R. 2002. New species and changing concepts of Philodendron
466 subgenus Meconostigma (Araceae). Aroideana 25: 2-15.
467 Gouy, M., Guindon, S. & Gascuel, O. 2010. SeaView version 4: A multiplatform graphical user
468 interface for sequence alignment and phylogenetic tree building. Mol Biol Evol 27: 221-4.
469 Grayum, M. H. 1996. Revision of Philodenron Subgenus Pteromischum (Araceae) for Pacific
470 and Caribean Tropical America. Systematic Botany Monographs 47: 234.
471 Guindon, S., Delsuc, F., Dufayard, J. F. & Gascuel, O. 2009. Estimating maximum likelihood
472 phylogenies with PhyML. Methods Mol Biol 537: 113-37.
473 Houle, D., Govindaraju, D.R. & Omholt, S. 2010. Phenomics: the next challenge. Nature
474 Reviews Genetics 11: 855-866.
475 Huelsenbeck, J. P. & Ronquist, F. 2001. MRBAYES: Bayesian inference of phylogenetic trees.
476 Bioinformatics (Oxford, England) 17: 754-755.
477 Iles, W. J. D., Smith, S. Y., Gandolfo, M. A. & Graham, S. W. 2015. Monocot fossils suitable for
478 molecular dating analyse. Botanical Journal of the Linnean Society.
479 Jaramillo, C., Hoorn, C., Silva, S. A. F., Leite, F., Herrera, F., Quiroz, L., Dino, R. & Antonioli,
480 L. (2010) The origin of the modern Amazon rainforest: implications of the palynological
481 and palaeobotanical record. (Hoorn, C. & Wesselingh, F., eds.). pp. 457-457. Wiley-
482 Blackwell.
483 Katoh, K. & Standley, D. M. 2013. MAFFT multiple sequence alignment software version 7:
484 Improvements in performance and usability. Molecular Biology and Evolution 30: 772-
485 780.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

486 Lo, E.Y.Y., Norman, C.D. & Sun, M. 2014. Phylogeographic pattern of Rhizophora
487 (Rhizophoraceae) reveals the importance of both vicariance and long-distance oceanic
488 dispersal to modern mangrove distribution. BMC Evolutionary Biology 14:83.
489 Loss-Oliveira, L., Calazans, L. S., de Morais, E., Mayo, S. J., Schrago, C. G. & Sakuragui, C. M.
490 2014. Floral evolution of Philodendron subgenus Meconostigma (Araceae). PLoS One 9:
491 e89701.
492 Maddison, W.P. & Knowles, L.L. 2006. Inferring phylogeny despite incomplete
493 lineage sorting. Systematic Biology 55(1): 21-30.
494 Mayo, S. J. (1986) Systematics of Philodendron Schott (Araceae) with special reference to
495 inflorescence characters. Vol. Ph.D. pp. 673. University of Keating, UK.
496 Mayo, S. J. 1988. Aspectos da evolução e da geografia do gênero Philodendron Schott
497 (Araceae). Acta Botanica Brasilica 1: 27-40.
498 Mayo, S. J. 1989. Observations of gynoecial structure in Philodendron (Araceae). Botanical
499 Journal of the Linnean Society 100: 139-172.
500 Mayo, S. J. 1991. A revision of Philodendron subgenus Meconostigma (Araceae). Kew Bulletin
501 46: 601-681.
502 Mayo, S. J., Bogner, J. & Boyce, P. 1997. The genera of Araceae, 1 ed. Royal Botanical Garden,
503 Kew.
504 Nagalingum, N.S., Marshall, C.R., Quental, T.B., Rai, H.S., Little, D.P. & Mathews, S. 2011.
505 Recent synchronous radiation of a living fossil. Science 334: 796-799.
506 Nauheimer, L., Metzler, D. & Renner, S. S. 2012. Global history of the ancient monocot family
507 Araceae inferred with models accounting for past continental positions and previous
508 ranges based on fossils. New Phytologist 195: 938-950.
509 Oliveira, F. B., Molina, E. C. & Marroig, G. (2010) South american primates, developments in
510 Primatology: progress and prospects. pp. 547-547. Springer Science, Chicago.
511 Olson, D. M., Dinerstein, E., Wikramanayake, E. D., Burgess , N. D., Powell, G. N.,
512 Underwood, E. C., D’amico, J. A., Itoua, I., Strand, H. E., Morrison, J. C., Loucks, C. J.,
513 Allnutt, T. J., Ricketts, T. H., Kura, Y., Lamoreux, J. F., Wettengel, W. W., Hedao, P. &
514 Kassem, K. R. 2001. Terrestrial Ecoregions of the World A New Map of Life on Earth.
515 Bioscience 51.
516 Opitz, W. 2005. Classification, natural history, and evolution of the genus Aphelocerus Kirsch
517 (Coleoptera : Cleridae : Clerinae) - Abstracts. Bulletin of the American Museum of
518 Natural History: 6-128.
519 Porzecanski, A. L. & Cracraft, J. 2005. Cladistic analysis of distributions and endemism
520 (CADE): Using raw distributions of birds to unravel the biogeography of the South
521 American aridlands. Journal of Biogeography 32: 261-275.
522 Prud’Homme, B., Minervino, C., Hocine, M. Cande, J.D., Aouane, A., Dufour, H.D.,
523 Kassner, V.A. & Gompel, N. 2011. Body plan innovation in treehoppers through the
524 evolution of an extra wing-like appendage. Nature 473: 83-86.
525 Roig-Juñent, S. & Coscarón, S. 2001. Biogeographical history of the Neotropical and
526 Neoantarctic. Revista del Museo Argentino de Ciencias Naturales 3: 119-134.
527 Ronquist, F. & Huelsenbeck, J. P. 2003. MrBayes 3: Bayesian phylogenetic inference under
528 mixed models. Bioinformatics 19: 1572-4.
529 Sakuragui, C. M. (1998) Taxonomia e filogenia das espécies de Philodendron seção Calostigma
530 (Schott) Pfeiffer no Brasil. pp. 238-238.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

531 Sakuragui, C. M. 2001. Biogeografia de Philodendron seção Calostigma ( Schott ) Pfeiffer (
532 Araceae ) no Brasil. Acta Scientiarum: 561-569.
533 Sakuragui, C. M., Calazans, L. S. B. & Soares, M. L. (2012a) Philodendron Schott In: List of
534 species of flora of Brazil. pp.
535 Sakuragui, C. M., Calazans, L. S. B. & Soares, M. L. (2012b) Philodendron Schott In: List of
536 species of flora of Brazil. Vol. 2014. pp. Botanical Garden of Rio de Janeiro.
537 Sakuragui, C. M., Mayo, S. J. & Zappi, D. 2005. Taxonomic revision of Brazilian species of
538 Philodendron Section Macrobelium. Kew Bulletin 60: 465-513.
539 Scornavacca, C., Berry, V., Lefort, V., Douzery, E. J. & Ranwez, V. 2008. PhySIC_IST:
540 cleaning source trees to infer more informative supertrees. Bmc Bioinformatics 9: 413.
541 Shimodaira, H. & Hasegawa, M. 2001. CONSEL: for assessing the confidence of phylogenetic
542 tree selection. Bioinformatics 17: 1246-7.
543 Shufeldt, W. D. 1917. Fossil birds found at Vero, Florida. Florida State Geological Survey
544 Annual Report 9: 35-42.
545 Smith, B. T. & Klicka, J. 2010. The profound influence of the Late Pliocene Panamanian uplift
546 on the exchange, diversification, and distribution of New World birds. Ecography 33:
547 333-342.
548 Stamatakis, A. 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with
549 thousands of taxa and mixed models. Bioinformatics 22: 2688-90.
550 Stockey, R. A., Rothwell, G. W. & Johnson, K. R. 2007. Cobbania corrugata gen. et comb. nov.
551 (Araceae): a floating aquatic monocot from the Upper Cretaceous of western North
552 America. American Journal of Botany 94: 609-624.
553 Team, E. CATE Araceae. pp.
554 Vilela, J. F., Mello, B., Voloch, C. M. & Schrago, C. G. 2014. Sigmodontine rodents diversified
555 in South America prior to the complete rise of the Panamanian Isthmus. Journal of
556 Zoological Systematics and Evolutionary Research 52: 249-256.
557 Werneck, F. P. 2011. The diversification of eastern South American open vegetation biomes:
558 Historical biogeography and perspectives. Quaternary Science Reviews 30: 1630-1648.
559 Wilde, V. & Frankenhauser, H. 1998. The Middle Eocene plant taphocoenosis from Eckfeld
560 (Eifel, Germany). Review of Palaeobotany and Palynology 101: 7-28.
561 Wing, S. L., Herrera, F., Jaramillo, C. A., Gómez-Navarro, C., Wilf, P. & Labandeira, C. C.
562 2009. Late Paleocene fossils from the Cerrejón Formation, Columbia, are the earliest
563 record of Neotropical rainforest. Proceedings of the National Academy of Sciences 106:
564 18627-18632.
565 Yang, Z. H. 2007. PAML 4: Phylogenetic analysis by maximum likelihood. Molecular Biology
566 and Evolution 24: 1586-1591.
567 Yeng, W. S., Jean, T. P., Kiaw, N. K., Othman, A. S., Boon, L. H., Ahmad, F. B. & Boyce, P. C.
568 2013. Phylogeny of Asian Homalomena ( Araceae ) Based on the ITS Region Combined
569 with Morphological and Chemical Data Phylogeny of Asian Homalomena ( Araceae )
570 based on the ITS Region Combined with Morphological and Chemical Data. 38: 589-
571 599.
572 Yu, Y., Harris, A. J. & He, X. J. (2012) RASP (Reconstruct Ancestral State in Phylogenies) 2.1b.
573 pp.
574 Zanella, F. C. (2013) Evolução da biota da Diagonal de Formações Abertas Secas da América do
575 Sul. (Carvalho, C. J. B. & Almeida, E. A. B., eds.). pp. 306-306. Roca, São Paulo.
576
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

577 Figure legends
578 Figure 1A. Geographic distribution of Philodendron species along the Neotropical biomes of
579 Amazon, Atlantic Forest, Cerrado and Caatinga. B. Philodendron inflorescence and the flower
580 zones.
581 Figure 2. Supertree of Philodendron and Homalomena species.
582 Figure 3. Phylogeny of Philodendron and Homalomena corroborated by the approximately
583 unbiased (AU) test.
584 Figure 4. Ancestral biome reconstructions and divergence time estimates of Philodendron and
585 Homalomena lineages. The epoch intervals followed the international chronostatigraphic chart
586 (Cohen et al., 2015) and are indicated by dashed lines.
587 Figure 5. Phylogenetic relationships between Philodendron and Homalomena recovered by
588 previous studies. A. Barabé et al. (2002); B. Gauthier et al. (2008) using the maximum
589 parsimony method; C. Gauthier et al. (2008) using Bayesian analysis; D. Cusimano et al. (2011);
590 E. Yeng et al. (2013).
591 Figure 6. Putative dispersal routes of the ancestor of Philodendron and American Homalomena
592 to the Neotropical region during the Miocene.
593
594
595
596
597
598
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Figure 1(on next page)
Figure 1
A. Geographic distribution of Philodendron species along the Neotropical biomes of Amazon,
Atlantic Forest, Cerrado and Caatinga. B. Philodendron inflorescence and the flower zones.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Amazon Forest
Atlantic Forest
Caatinga
Cerrado
A B
Male zone
Male sterile
zone
Female zone
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Figure 2(on next page)
Figure 2
Supertree of Philodendron and Homalomena species.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Culcasia rotundifolia
Cercestis camerunense
Cercestis afzelii
Cercestis sp.
Heteropsis flexuosa
Dieffenbachia elegans
Nephthytis poissoni
Nephthytis afzelii
Cercestis kawennianus
P. pinnatifidum
P. cannifolium
P. acutatum
P. malesevichiae
P. hastatum
P. rhizomatozum
P. cordatum
P. aemulum
P. pachyphyllum
P. eximium
P. renauxi
P. martianum
P. lindenii
P. asplundii
P. auriculatum
P. erubescens
P. oblongum
P. wurdackii
P. wendlandii
P. angustisectum
P. heleniae
P. schottii talamancae
P. davidsonii
P. smithii
P. stenophyllum
P. ruizii
P. imbe
P. distantilobum
P. placidum
P. tripartitum
P. hylaeae
P. squamiferum
P. sagittifolium
P. krugii
P. pedatum
P. billietiae
P. loefgrenii
P. ornatum
P. micranthum
P. gloriosum
P. verrucosum
P. lazorii
P. grandipes
P. panamense
P. elaphoglossoides
P. wittianum
P. fragrantissimum
P. linnaei
P. callosum
P. crassinervium
P. brevispathum
P. insigne
P. camposportoanum
P. longilaminatum
P. edmundoi
P. goeldii
P. leal-costae
P. tweedianum
P. saxicola
P. adamantinum
P. williamsii
P. venezuelense
P. speciosum
P. paludicola
P. dardanianum
P. xanadu
P. brasiliense
H. panamensis
H. crinipes
H. wendlandii
H. picturata
H. erythropus allenii
F. mixta
H. griffthii
H. cochinchinensis
H. humilis
H. aromatica
H. pendula
H. rubescens
P. findens
0.002
A
B
C D
E
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Figure 3(on next page)
Figure 3
Phylogeny of Philodendron and Homalomena corroborated by the approximately unbiased
(AU) test.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

P. subg. Meconostigma
P. subg. Philodendron + Pteromischum
American Homalomena
Asian Homalomena
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Figure 4(on next page)
Figure 4
Ancestral biome reconstructions and divergence time estimates of Philodendron and
Homalomena lineages. The epoch intervals followed the international chronostatigraphic
chart (Cohen et al., 2015) and are indicated by dashed lines.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

(F) Crotundifolia
(F) Ccamerunense
(F) Cafzelii
(F) Ckawennianus
(F) Cercestis
(BC) Hflexuosa
(C) Delegans
(F) Npoissoni
(F) Nafzelii
(CD) Ppinnatifidum
(C) Scannifolium
(BCDE) Pacutatum
(C) Pmalesevichiae
(C) Pfindens
(B) Phastatum
(D) Prhizomatosum
(B) Pcordatum
(B) Paemulum
(D) Ppachyphyllum
(B) Peximium
(B) Prenauxi
(B) Pmartianum
(C) Plindenii
(C) Pasplundii
(CD) Pauriculatum
(C) Perubescens
(BCD) Poblongum
(C) Pwurdachii
(C) Pwendlandii
(C) Pangustisectum
(C) Psagittifolium
(C) Pheleniae
(C) Pschottiitalamancae
(C) Pdavidsonii
(C) Psmithii
(C) Pstenophyllum
(C) Pruizii
(B) Pimbe
(C) Pdistantilobum
(C) Pplacidum
(C) Ptripartitum
(C) Phylaeae
(C) Psquamiferum
(C) Pkrugii
(BCDE) Ppedatum
(C) Pbillietiae
(B) Ploefgrenii
(BCE) Pornatum
(C) Pmicranthum
(C) Pgloriosum
(C) Pverrucosum
(C) Plazorii
(CD) Pgrandipes
(C) Ppanamense
(C) Pelaphoglossoide
(C) Pwittianum
(BC) Pfragrantissimum
() Plinnaei
(C) Pcallosum
(B) Pcrassinervium
(CD) Pbrevispathum
(BC) Pinsigne
(CD) Pcamposportanum
(B) Plongilaminatum
(B) Pedmundoi
(C) Pgoeldii
(BE) Plealcostae
(D) Ptweedianum
(D) Psaxicola
(D) Padamantinum
(B) Pwilliamsii
(C) Pvenezuelense
(B) Pspeciosum
(B) Ppaludicola
(D) Pdardanianum
(G) Pxanadu
(D) Pbrasiliense
(G) Hpanamensis
(C) Hcrinipes
(C) Hwendlandii
(C) Hpicturata
(C) Herythropusallenii
(A) Hrubescens
(A) Fmixta
(A) Hgriffthii
(A) Hcochinchinensis
(A) Hhumilis
(A) Haromatica
(A) Hpendula
Culcasia rotundifolia
Cercestis camerunense
Cercestis afzelii
Cercestis sp.
Heteropsis flexuosa
Dieffenbachia elegans
Nephthytis poissoni
Nephthytis afzelii
Cercestis kawennianus
P. pinnatifidum
P. cannifolium
P. acutatum
P. malesevichiae
P. findens
P. hastatum
P. rhizomatozum
P. cordatum
P. aemulum
P. pachyphyllum
P. eximium
P. renauxi
P. martianum
P. lindenii
(F)
(F)
(F)
(F)
(F)
(BC)
(C)
(F)
(F)
(CD)
(C)
(BCDE)
(C)
(C)
(B)
(D)
(B)
(B)
(D)
(B)
(B)
(B)
(C)
(B)
(C)
(C)
(C)
(C)
(C)
(C)
(C)
(B)
(C)
(D)
(D)
(D)
(B)
(C)
(B)
(B)
(G)
(G)
(C)
(A)
(C)
(CD)
(C)
(BCD)
(C)
(C)
(C)
P. asplundii
P. auriculatum
P. erubescens
P. oblungum
P. wurdackii
P. wendlandii
P. angustisectum
(BCDE)
(C)
(C)
(C)
(C)
(C)
(C)
(C)
(B)
(C)
(C)
(C)
(C)
(C)
(C)
(C)
P. heleniae
P. schottii talamancae
P. davidsonii
P. smithii
P. stenophyllum
P. ruizii
P. imbe
P. distantilobum
P. placidum
P. tripartitum
P. hylaeae
P. squamiferum
P. sagittifolium
P. krugii
P. pedatum
P. billietiae
(BCE)
(CD)
(BC)
(CD)
(C)
(B)
(CD)
(BC)
(CD)
(B)
P. loefgrenii
P. ornatum
P. micranthum
P. gloriosum
P. verrucosum
P. lazorii
P. grandipes
P. panamense
P. elaphoglossoides
P. wittianum
P. fragrantissimum
P. linnaei
P. callosum
P. crassinervium
P. brevispathum
P. insigne
P. camposportoanum
P. longilaminatum
(BE)
(D)
(D)
P. edmundoi
P. goeldii
P. leal-costae
P. tweedianum
P. saxicola
P. adamantinum
P. williamsii
P. venezuelense
P. speciosum
P. paludicola
P. dardanianum
P. xanadu
P. brasiliense
H. panamensis
H. crinipes
(C)
(C)
(C)
(A)
(A)
(A)
(A)
(A)
(A)
H. wendlandii
H. picturata
H. erythropus allenii
H. rubescens
F. mixta
H. griffthii
H. cochinchinensis
H. humilis
H. aromatica
H. pendula
12.8Ma
9.2Ma
12.2Ma
8.3Ma
10.6Ma
7.6Ma
5.5Ma
3.7Ma
5.0Ma
2.9Ma
Miocene
Middle Late
6.8Ma
4.2Ma
Pliocene
11.6 5.3 2.6 Ma
AC
BC
BCD
BD
BF
CD
CF
LEGEND
*
A
Legend
= Asian Tropical Moist Forest
B= Atlantic Forest
C= Amazon Forest
D
E
F
G
= Cerrado
= Caatinga
= African Tropical Moist Forest
= Unknown
A
B
C
D
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Figure 5(on next page)
Figure 5
Phylogenetic relationships between Philodendron and Homalomena recovered by previous
studies. A. Barabé et al. (2002); B. Gauthier et al. (2008) using the maximum parsimony
method; C. Gauthier et al. (2008) using Bayesian analysis; D. Cusimano et al. (2011); E. Yeng
et al. (2013).
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

P. subg. Meconostigma
Asian Homalomena
P. subg. Pteromischum
P. subg. Philodendron
P. subg. Philodendron
AB
CDE
P. subg. Philodendron
P. subg. Meconostigma
P. subg. Pteromischum
American Homalomena
Asian Homalomena
P. subg. Philodendron
P. subg. Meconostigma
P. subg. Pteromischum
American Homalomena
Asian Homalomena
Philodendron
Homalomena
Furtadoa
P. subg. Philodendron
P. subg. Meconostigma
P. subg. Pteromischum
American Homalomena
Asian Homalomena
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

Figure 6(on next page)
Figure 6
Putative dispersal routes of the ancestor of Philodendron and American Homalomena to the
Neotropical region during the Miocene.
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016

11
2
3
3
4
...
...
PeerJ PrePrints | https://doi.org/10.7287/peerj.preprints.1708v1 | CC-BY 4.0 Open Access | rec: 3 Feb 2016, publ: 3 Feb 2016