Fig 1.
Representatives of Dendrosenecio illustrating the conspicuous variation in growth form and morphology in this genus (Photos: Abel Gizaw).
(a-d): Distinct growth form differentiation between two types of high-altitude habitats on Mt Kenya. Dendrosenecio keniensis (KEN-Kn, a & b) is a low-grown (< 1.5 m), procumbent plant that branches close to the ground and occurs on constantly water-saturated soils, whereas D. keniodendron (KND-Kn, c & d) is an erect giant with tall stems (up to 7 m) that branch high above the ground, occurring on well-drained soils. (e-h): Plants at lower altitudes tend to have thin stems with only few old leaves kept for insulation, exemplified by plants of D. battiscombei on Mt Kenya (BAT-Kn, e & f) and D. erici-rosenii on Mt Ruwenzori (ERI-Ru, g & h). (i-o): Plants at higher altitudes tend to have thick stems with a prominent layer of insulating old leaves, exemplified by plants of D. adnivalis on Mt Ruwenzori (ADN-Ru, i & j), D. kilimanjari on Mt Kilimanjaro (KIL-Ki, k & l) and D. elgonensis on Mt Elgon (ELG-El, m, n & o).
Table 1.
Species and subspecies of Dendrosenecio accepted by Knox [24] and their altitudinal and geographical range.
Fig 2.
Sampling sites and main genetic structuring in Dendrosenecio based on the total AFLP dataset of 460 successfully analysed plants (109 populations representing all species and all but one of the subspecies accepted by Knox [24]).
All species are represented with samples from all mountains from where they have been recorded, except that 1) KND has also been reported from Mt Aberdare, and 2) ERI has also been reported from some other Western Rift mountains located close to the two mountains investigated here. Colours represent the four main genetic groups inferred from STRUCTURE analyses, and the pie-charts show the proportion of admixture averaged over all individuals (corresponding to K = 4 in Fig 3). The pie-charts closest to the map are based on the total number of plants successfully analysed from each mountain (i.e. no taxonomic assignment). The pie-charts on the right and the left sides are based on taxonomic assignment according to morphology (ADN, D. adnivalis; BAT, D. battiscombei; BRA, D. brassiciformis; CHE, D. cheranganiensis; ELG, D. elgonensis; ERI, D. erici-rosenii; JOH, D. johnstonii; KEN, D. keniensis; KND, D. keniodendron; KIL, D. kilimanjari; MER, D. meruensis). The abbreviated species name is followed by abbreviated mountain name and number of populations (individual plants) successfully analysed. In three cases, plants originating from different mountains were assigned to the same species, indicated with brackets. The stippled lines represent the Great Rift Valley system.
Fig 3.
(a) Assignment of the 460 individual plants in the total AFLP dataset to genetic groups using STRUCTURE analyses for K = 4 to K = 12. Species and mountain names are abbreviated as in Fig 2. WRZ–Western Rift Zone mountains, ERZ–Eastern Rift Zone mountains. (b) Principal Coordinates Analysis (PCoA) based on Dice’s coefficient of similarity between AFLP phenotypes in the total dataset; colours show genetic groups for K = 4.
Fig 4.
Separate principal coordinates analyses (PCoA) based on Dice’s coefficient of similarity between AFLP phenotypes for subsets of the total dataset to assess details of the admixture between three of the four main genetic groups.
(a) The WRZ-Ruwenzori and WRZ/ERZ genetic groups; (b) The ERZ-Kenya/Aberdare and WRZ/ERZ genetic groups. Plots to the left show mountains, plots to the right show taxonomic annotations. Abbreviations as in Fig 2.
Fig 5.
Separate principal coordinates analyses (PCoA) based on Dice’s coefficient of similarity between AFLP phenotypes for subsets of the total dataset representing each of the four main mountain groups.
Plots to the left show mountains, plots to the right show taxonomic annotations. Abbreviations as in Fig 2, except in 4b where KIL co-Ki (D. kilimanjari ssp. cottonii), KIL ki-Ki (D. kilimanjari ssp. kilimanjari) and in 4d where CHE ch-Ch (D. cheranganiensis ssp. cheranganiensis), CHE da-Ch (D. cheranganiensis ssp. dalei), ELG el-EL (D. elgonensis ssp. elgonensis) and ELG ba-EL (D. elgonensis ssp. barbatipes) as indicated in S1 Appendix.
Fig 6.
Neighbor-Net diagram computed from pairwise Fst values as measures of distance among 109 populations of Dendrosenecio, calculated from the total AFLP dataset comprising 460 individual plants successfully analysed for AFLP.
Colours show the main genetic groups inferred in the STRUCTURE analyses. Branch supports were estimated using 1000 bootstrap replicates; no major branches obtained >50% support. Abbreviations as in Fig 2.
Table 2.
Genetic diversity and rarity in Dendrosenecio based on 455 AFLP markers scored in 460 individual plants from 109 populations, representing all 11 species accepted by Knox [24].
Table 3.
Non-hierarchical and hierarchical analyses of molecular variance (AMOVAs) in Dendrosenecio based on 455 AFLP markers scored in 460 individuals from 109 populations (representing all 11 species accepted by Knox [24]).