Results
Divergence dating
The dated phylogeny for the Stasimopus genus in the Karoo can be seen in Figure 2. Samples indicated in red text (and the grey bar on the right) could not be identified to species level as they are juveniles. Divergence estimates indicate that the genus Stasimopus in the Karoo dates to between 48.30 - 84.15 MYA, towards the beginning of the Paleocene under the birth-death speciation model (Figure 2, node i). This range is shared by the birth-death incomplete and coalescent models (Figures S1 and S2). The yule speciation model gives older node ages across the entire phylogeny (Figure S3, Table 1). Due to the consistent dating of the birth-death, birth-death incomplete and coalescent models, the discussion will revolve around these dates.
The phylogeny is divided into 11 clades for the ease of discussion (labelled A - K). Clade A is the oldest clade across all models. This clade is well supported and consists of a species complex of S. erythrognathus and S. karooensis as well as a single individual representing S. theaei. This clade is dated to 24.13 – 50.25 MYA (Figure 2, node ii). The age of the species complex is younger 9.36 - 22.35 MYA (Figure 2, node iii), the inclusion of S. theaei and the unidentified samples make the clade significantly older. Clade B is a well-supported monophyletic clade, consisting of S. astutus andS. finni. The species S. astutus is young from the Miocene (6.89- 16.46 MYA) (Figure 2, node iv), but Clade B is dated between 22.69 - 50.78 MYA due to the inclusion of S. finni (Figure 2, node v). Clade C is dated within the Oligocene, the individual species cannot be accurately dated due to small sample sizes. Stasimopus dylani (Clade D) (Figure 2, node vi), S. venterstadensis (Clade I) (Figure 2, node iiv) and S. unispinosus (Clade H) (Figure 2, node iiiv) are all dated to the Miocene. Clade F (S. maraisi ) is one of the older species dated between 16.2 – 39.32 MYA (Figure 2, node ix), followed by clade G (S. hamartia ) between 22.53 – 43.56 MYA (Figure 2, node x). Stasimopus leipoldti (Clade K) is very recently diverged in the Pleistocene (Figure 2, node xi). This may however, be inaccurate due to the small sample size and that the samples are from the same locality. Clades E and J are comprised of only unidentifiable juveniles. It is however apparent that these specimens likely comprise separate species.
Demographic history
The haplotypes and demographic results for CO1 are presented in Figure 3 and Table 2 and the data for 16S and EF-1ɣ in Figures S4, S5 and Tables S8, S9, respectively. The haplotype networks are largely in agreement with the phylogeny, except for the placement of the S. venterstadensis (Clade I, light green) individuals in the CO1 network (Figure 3) which are split and group with S. maraisi (Clade F, dark blue) and S. astutus (Clade B, coral) separately. The species complex of clade A (Figure 2) appears to be present across all three haplotype networks. The networks reveal that there is a higher degree of genetic differentiation in the CO1 gene within species than the other two gene regions, this is apparent from the higher number of mutational steps between haplotypes. EF-1ɣ consistently has the least number of mutational steps between haplotypes within species. This is further substantiated by the differences in k, S and πvalues across the three gene regions for each species (Table 2, Tables S8 and S9). This should however be interpreted with caution as all sample sizes (except for S. maraisi ) are quite low.
The haplotype diversity is high across all species considered for the CO1 gene region (Table 2). Stasimopus unispinosus has a lower haplotype diversity for the 16S and EF-1ɣ regions, both of 0.6 (Tables S4 and S5). Nucleotide diversity is relatively low across species and gene regions (>0.1) (Table 2, S8 and S9).
In the EF-1ɣ network (Figure S5) the S. maraisi network shows a clear expansion pattern, but this is not shown in the other two gene regions. This expansion of S. maraisi for the EF-1ɣ gene region is further proven by the significant negative result of the Tajima’s D (-1.825*) and Fu & Li’s D* (-3.140*) statistics. The only other significant result is for S. hamartia for 16S, where Fu & Li’s D* is significant (1.418).
There is clear phylogeographic structuring across all three gene regions. Most species appear to be isolated to one or a few locations, except for S. maraisi which is widespread across the region.Stasimopus leipoldti was only found in one location in the Tankwa Karoo, but the type locality for the species is Clanwilliam (Western Cape), over 150 km away. A similar trend is seen for S. erythrognathus which was in found in the eastern Karoo, but the type locality is Worcester (Western cape), 700 km away.
There are many species occurring sympatrically at multiple locations. These include S. maraisi and S. hamartia ; S. maraisi and S. dylani ; S. theaei and S. finni ;S. erythrognathus and S. karooensis ; S. astutus ,S. mandelai and S. dylani ; S. maraisi and Clade J. Clade J likely comprises a new species based on the degree of genetic variation, but no adults were found to confirm their identity.
Mantel test
The Mantel tests yielded significant results for all tests conducted (p<0.05). The results by locality indicate the same weak positive correlation for both software packages (Ade: r =0.1413, p =0.0022; Vegan: r =0.1413, p =0.0032). The results for the mantel test across the various species are not significant (Ade: r =0.0157, p =0.4333; Vegan: r =0.0157, p =0.437).