1
Appendix S2 - Supplementary Results Biogeographic inference We report below the likelihood of four different biogeographical delimitations at the continental scale obtained with the software Lagrange (Table S1); and the likelihood values obtained for regional biogeographical models differing in the probability of dispersal between non-neigbour areas (Fig. S3). Table S1. Log-likelihood scores obtained with Lagrange reconstructions for different continental delimitations (the four models have the same number of parameters). Area delimitation codes correspond to: A, Asia minor grouped with Asia and Beringian Asian region with North America; B, Asia minor grouped with Europe and Beringian Asian region with Asia; C, Asia minor grouped with Europe and Beringian Asian region with North America; D, Asia minor and Beringian Asian region grouped with Asia (see Fig. S2 for a graphical representation of the area delimitations).
Area delimitation
Log-likelihood
A)
A
B
C
D
50.2 ±1.5
56.4 ±2.2
51.9 ±1.6
55.3 ±2.2
B)
Fig. S4 Log-likelihood values obtained with Lagrange for dispersal different probabilities (A: 0.1-0.9; B: 0.01-0.09) for non-neighbour areas (mean and standard deviation over 100 trees).
2
Climatic vicariance analyses We performed a set of statistical analyses implemented in SEEVA to test whether climatic vicariance may have played an important role in Androsace diversification. Either a low or non-significant climatic divergence was detected for most of the phylogenetic nodes of Androsace (see below Table S2). Only 5 nodes showed a high divergence (D>0.70) for several climatic variables: (1) the separation of Pomatosace (endemic to W China) from all the other species in the American-European clade (node 18 in Fig. S4); (2) the split between two annual species (A. elongata and A. occidentalis) and the rest of perennial-cushions species in the American-European clade (node 20); (3) the divergence between Douglasia and its sister clade composed of the two SW European species A. cantabrica and A. adfinis (node 37); (4) the separation of A. cantabrica (a perennial from the Cantabric range) and A. adfinis (a perennial endemic to the SW Alps) (node 38); and (5) the divergence of the perennials A. spinulifera (found in the Hengduan range), A. tapete (Himalayas and Tibet) and A. mariae (Hengduan range and Tibet) (node 11). Only 4 of the nodes with a high D for at least one climatic variable corresponded to sister species. These results suggest that climatic vicariance has played a minor role (or none at all) in Androsace diversification. Table S2. Index of climatic divergence (D, 0-1) estimated for 5 climatic variables and for each node in the phylogeny; * indicates nodes showing significant differences between sister groups following a Bonferroni correction (P ≤ 0.0012); those with high D values (>0.7) are written in bold. Node numbers correspond to Fig. S4. Index of divergence Node
Mean annual
Temperature
Temperature
seasonality
Isothermality
Annual
Precipitation
precipitation
seasonality
1
0,19
0,40
0,09*
0,21
0,09
2
0,58
0,24
0,46
0,47
0,19
3
0,01
0,10*
0,28*
0,40*
0,62*
4
0,06
0,04
0,25
0,66*
0,13
5
0,08
0,20
0,41
0,03
0,00
6
0,05
0,78
1,00*
0,32
0,00
7
0,15*
0,48*
0,17
0,30*
0,40*
8
0,56
0,58
0,63
0,63
0,62
9
0,44*
0,50*
0,78*
0,57*
0,06
10
0,20*
0,71*
0,35*
0,21*
0,09*
11
0,61*
0,00
0,72*
0,76*
0,00
12
0,15
0,00
0,22
0,02
0,00
3
13
0,10
0,76*
0,10
0,09
0,10
14
0,02
0,87*
0,15*
0,03
0,13*
15
0,11
0,12
0,09
0,11
0,13
16
0,64*
0,00
0,59
0,47
0,00
17
0,11
0,00
0,10
0,01
0,00
18
0,50*
0,70*
0,44*
0,63*
0,64*
19
0,06*
0,47
0,56
0,57
0,46
20
0,79*
0,36*
0,78*
0,76*
0,76*
21
0,01
0,17*
0,68*
0,06*
0,41*
22
0,16*
0,20*
0,10*
0,34*
0,34*
23
0,28*
0,03*
0,08*
0,15*
0,01*
24
0,10
0,26*
0,50*
0,03
0,96*
25
0,57
0,81
0,44
0,27
0,02
26
0,25
0,09
0,18
0,16*
0,17
27
0,03
0,36
0,32
0,57*
0,62*
28
0,15*
0,19*
0,09*
0,09*
0,07*
29
0,63*
0,87*
0,06
0,39*
0,36*
30
0,49*
0,59*
0,58*
0,45*
0,64*
31
0,04
0,02
0,06*
0,06*
0,17*
32
0,51
0,51
0,32
0,53
0,43
33
0,21
0,54
0,22
0,04
0,08
34
0,06
0,06
0,02
0,08
0,11
35
0,17
0,04
0,01
0,42
0,41
36
0,31
0,02
0,19
0,33
0,22
37
0,61*
0,86*
0,82*
0,80*
1,00*
38
0,87*
1,00*
0,07
0,70*
1,00*
39
0,62
0,71
0,57
0,61
0,04*
40
0,41
0,67*
0,00
0,01
0,17*
41
0,03
0,00
0,00
0,02
0,04
42
0,00
0,00
0,00
0,00
0,00
43
0,00
0,00
0,00
0,00
0,00
Fig. S5. Cladogram corresponding to a Bayesian 50%-majority rule consensus phylogenetic tree showing phylogenetic relationships between the 51 species for which climatic data was available (species without climatic data were pruned). Nodes are numbered to serve as a reference for Table S2.
4
5
Diversification analyses We report below the likelihood results of four trait-dependent diversification models run independently for the two main clades of Androsace (Table S3), and the lineage-throughtime plot for the American-European clade of Androsace species which shows a pattern of density-dependence in Europe and the decoupling of diversity dynamics in North American species (Fig. S5). Table S3. AIC scores obtained for different trait-dependent diversification models for each of the two main clades of Androsace. PB means pure birth model; BD is “birth-death model”; PB-Form and BD-Form correspond respectively to a PB and a BD model with different speciation rates for the three life forms of Androsace. AIC scores Clade
PB
BD
PB-Form
BD-Form
Central-Asian
282.7 ±1.2
282.2 ±2.0
266.3 ±2.2
272.3 ±2.2
279.6 ±6.5
281.3 ±6.9
270.8 ±8.0
276.0 ±8.3
AmericanEuropean
6
Table S4. Dispersal traits measured for some Androsace species. The number of seeds measured, mean seed mass and life form are given for each species. Seed providers are abbreviated as follows: MP for Mojmir Pavelka, VH for Vojtěch Holubec and AG for the Androsace Group of the Alpine Garden Society. Seeds from MP and VH were collected in natural populations and the locality of collection is reported. Seeds from AG were collected in botanical or private gardens which location was not provided.
Taxon Androsace adenocephala A. bisulca var. aurata A. bisulca var. brahmaputrae A. carnea A. chaixii A. cylindrica A. fedschenkoi A. halleri A. hausmanii A. hedraeantha A. hirtella A. incana A. integra A. jacquemontii A. limprichtii A. mairei A. mariae A. mathildae A. maxima A. multiscapa A. muscoidea A. neuwirthii A. obtusifolia A. pubescens A. pyrenaica A. rigida A. rioxana A. robusta var. purpurea A. salicifolia A. sempervivoides A. spinulifera A. tanggulashanensis A. tapete A. umbellata A. vandelii A. villosa A. yargongensis Douglasia montana
Seed number 28 52 33 107 118 68 47 91 24 116 16 23 91 4 27 32 63 32 26 12 13 63 32 31 23 18 14 16 58 4 17 50 11 6 87 57 37 40
Seed mass (mg) 0,57 0,27 0,71 1,80 0,27 2,09 0,49 1,76 0,64 1,58 2,28 0,62 0,78 0,5 0,98 0,83 0,26 1,28 1,08 1,88 0,9 0,13 0,76 0,47 0,67 1,64 2,31 0,69 1,17 0,5 1,31 0,63 1,2 0,15 0,34 1,11 1,11 2,89
Life form
Locality of collection
Provider
Perennial Cushion Cushion Perennial Short Perennial Short Perennial Cushion Perennial Perennial Perennial Short Perennial Perennial Perennial Cushion Cushion Short Perennial Cushion Cushion Cushion Cushion Cushion Perennial Perennial Cushion Short Cushion Perennial Cushion Cushion Short Cushion Perennial Cushion Cushion
Tibet Sichuan, China Tibet
VH VH VH AG AG AG AG AG MP AG AG MP AG AG VH VH VH AG VH MP AG MP VH VH AG VH AG AG AG MP VH MP MP AG AG MP VH AG
Alps, Italy
Siberia, Russia
Sichuan, China Sichuan, China Qinghai, China Xinjiang, China Taurus, Turkey Qinghai, China Alps, France Alps, France Yunnan, China
Himachal Pradesh, India Sichuan, China Qinghai, China Sichuan, China
Taurus, Turkey Yunnan, China
