Fig. 1. Geographic distribution and variation in external morphology of Distichodus species diversity. Map of Africa divided into ichthyofaunal provinces (originally defined by Roberts [47], modified by Lévêque [48], and redrawn according to new hydrological basin mapping published by FAO [74]): Congo Basin (CB), East Africa (EA), Nilo-Sudan (NS), Lower Guinea (LG), South Africa (SA), and West Africa (WA). Shaded area represents Distichodus extent of occurrence. Inset bar charts indicate number of Distichodus species present in each ichthyofaunal province: endemic (red) and total (blue) (when more than endemics). Inset frame fish photographs illustrate the extent of variation in body shape, size, and coloration in Distichodus species (from top to bottom: D. hypostomatus, D. sexfasciatus, D. lussoso, D. antonii, D. affinis, D. shenga, D. decemmaculatus).
Fig. 2. Alternative log-normally distributed priors used to account for temporal uncertainty of calibration nodes. Each prior probability density function (PDF) is characterized by a hard minimum bound of 18 Ma (set by the age of the calibration fossil), a standard deviation (σ) of 0.5, and a variable mean (μ) (in real space) that probabilistically models the extent to which the node age spreads into the past: μ=19 (black), μ=24 (blue), and μ=29 (red). The lower limit of the x-axis interval defining the area shaded under each curve corresponds to its 95th percentile soft maximum bound (P95 SMB): 20 Ma (black), 30 Ma (blue), and 40 Ma (red).
Fig. 3. Total-evidence Distichodus phylogeny as inferred by likelihood in RAxML. Colored circles on nodes indicate degree of clade support as determined by bootstrap values (BS). Nodes labeled A and B represent the two main infrageneric clades. Outgroup taxon (Paradistichodus dimiatus) not shown.
Fig. 4. Summary tree of the RAxML Distichodus phylogeny highlighting interspecific relationships.
Fig. 5. Distichodus species tree generated using the coalescence-based method SVDquartets.
Fig. 6. Distichodus species tree generated using the coalescence-based method ASTRAL-III.
Fig. 7. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 8, intermediate in terms of calibration node (D, crown) and P95 SMB (30 Ma). Divergence-time estimates are represented by the mean ages of clades. Light red bars correspond to 95% highest posterior density (HPD) intervals of mean node ages. Calibration (fossil-based) node indicated by a dagger (†). Colored circles on nodes indicate degree of clade support as determined by posterior probabilities: black > 0.95, 0.95 ³ blue ³ 0.75, red < 0.75. Outgroup taxon (Nannocharax ansorgii) not shown.
Fig. 8. A spatiotemporal reconstruction of Distichodus range evolution. Based on the optimal DEC* model (M1; CB-as-source) and input chronogram resultant from BEAST2 analysis 8 (Fig. 7). Ichthyofaunal provinces color-coded and abbreviated as in Fig. 1.
Fig. 1S. enc1 Distichodus phylogeny as inferred by likelihood in RAxML. Colored circles on nodes indicate degree of clade support as determined by bootstrap values (BS). The identity of leaves (terminals) not printed on the tree is specified by the species name (in bold) at the base of the most recent labeled ancestral node from which the sample descends. Names in bold black correspond to those species resolved as monophyletic (when multiple individuals were available), whereas those in bold green indicate that, while most of the sampled specimens fall into the clade subtended by that node, some samples fall outside the clade, and therefore the species is not resolved as monophyletic. Outgroup taxon (Paradistichodus dimiatus) not shown.
Fig. 2S. glyt Distichodus phylogeny as inferred by likelihood in RAxML. Same contextual information as in Fig. 1S.
Fig. 3S. myh6 Distichodus phylogeny as inferred by likelihood in RAxML. Same contextual information as in Fig. 1S.
Fig. 4S. sh3px3 Distichodus phylogeny as inferred by likelihood in RAxML. Same contextual information as in Fig. 1S.
Fig. 5S. mtDNA (co1, cr, cytb, nd2) Distichodus phylogeny as inferred by likelihood in RAxML. Same contextual information as in Fig. 1S.
Fig. 6S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 1. Same contextual information as in Fig. 7.
Fig. 7S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 2. Same contextual information as in Fig. 7.
Fig. 8S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 3. Same contextual information as in Fig. 7.
Fig. 9S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 4. Same contextual information as in Fig. 7.
Fig. 10S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 5. Same contextual information as in Fig. 7.
Fig. 11S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 6. Same contextual information as in Fig. 7.
Fig. 12S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 7. Same contextual information as in Fig. 7.
Fig. 13S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 9. Same contextual information as in Fig. 7.
Fig. 14S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 10. Same contextual information as in Fig. 7.
Fig. 15S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 11. Same contextual information as in Fig. 7.
Fig. 16S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 12. Same contextual information as in Fig. 7.
Fig. 17S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 13. Same contextual information as in Fig. 7.
Fig. 18S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 14. Same contextual information as in Fig. 7.
Fig. 19S. A time-scaled phylogeny of Distichodus. Chronogram resulting from BEAST2 analysis 15. Same contextual information as in Fig. 7.
Fig. 20S. A spatiotemporal reconstruction of Distichodus range evolution. Based on the optimal DEC* model (M1; CB-as-source) and input chronogram resultant from BEAST2 analysis 5. Ichthyofaunal provinces color-coded and abbreviated as in Fig. 1.
Fig. 21S. A spatiotemporal reconstruction of Distichodus range evolution. Based on the optimal DEC* model (M1; CB-as-source) and input chronogram resultant from BEAST2 analysis 14. Ichthyofaunal provinces color-coded and abbreviated as in Fig. 1.
- Table Captions
Table 1. Taxa, voucher specimens (catalog and tissue numbers), and GenBank accession numbers for the gene sequences included in the analyses. Institutional abbreviations: AMCC (Ambrose Monell CryoCollection, AMNH), AMNH (American Museum of Natural History), CU (Cornell University Museum of Vertebrates), SAIAB (South African Institute for Aquatic Biodiversity), MRAC (Royal Museum for Central Africa).
Table 2. Alternative BEAST2 analyses (1-15) for co-estimating phylogeny and divergence times in Distichodus resulting from variable calibration strategies (calibration node, stem vs. crown group, and 95th percentile [P95] soft maximum bound [SMB] of calibration prior).
Table 3. Results from alternative BEAST2 analyses. Estimated mean ages (in Ma) and associated 95% HPD intervals of select nodes: D+P = MRCA of Distichodus & Paradistichodus; D = MRCA of Distichodus species; Dne + Dro = MRCA of D. nefasch & D. rostratus; DA = Distichodus subclade A; DB = Distichodus subclade B. P95 SMB = 95th percentile soft maximum bound (in Ma), as a proxy for the maximum node age constraint.
Table 4. Results from DEC* analysis of geographic range evolution on the Distichodus phylogeny. Results are presented for each of the three analyses based on different BEAST2 input chronograms (derived from analyses 5, 8, and 14). Comparison of alternative models (biogeographic hypotheses) and their support as assessed via Akaike weights. M0 (unconstrained, dispersal to and from the Congo Basin); M1 (allowing only dispersal out of the Congo Basin); M2 (allowing only dispersal into the Congo Basin); dispersal (d); extinction (e); number of parameters (k); Akaike information criterion (AIC); Akaike Weights (AW).