- V. burejaeticum Regel et A, - JQ JQ JX JX JX JX JX V. carlesii Hemsl. Ex Forb. & Hemsl.

Transcription

1 Appendix A: Voucher information and Genbank accession numbers for Viburnum species sampled. Species are grouped by clade name (Clement and Donoghue 2011) and missing data are indicated by a "." For each species the Collector, Collector number and Herbarium where voucher is located is provided. Herbaria abbreviations are as follows: A: Arnold Arboretum, K: Kew Royal Botanic Garden, MO: Missouri Botanical Garden, NY: New York Botanical Garden; WTU: University of Washington Herbarium, YU Yale University Herbarium. Clade Species Collector Collection No. H psbatrnh rpl32 ITS trnk matk rbcl ndhf trncy6 trnsg petbd Coriacea V. coriaceum Bl. L. Averyanov et al. VH3300 MO HQ HQ HQ HQ HQ HQ HQ HQ HQ V. cylindricum Buch. Ham. ex D. Don Boufford et al A AY HQ AY AY HQ EF V. hebanthum Wight & Arn. J. Klackenberg 32 NY HQ HQ HQ HQ HQ HQ HQ HQ HQ Lantana V. bitchiuense Makino Arnold Arboretum A AA living collection JX JX JX JX JX JX JX JX JX JX V. buddleifolium C.H. 7533B, Wright Arnold Arboretum A JQ JQ JQ JX JX JX JX JX V. burejaeticum Regel et 37595A, Herder Arnold Arboretum A JQ JQ JQ JQ JX JX JX JX JX V. carlesii Hemsl. Ex Forb. & Hemsl. R.C. Winkworth 24 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ V. lantana 1 L. R.C. Winkworth 26 YU AY HQ AY AY HQ HQ HQ HQ EF HQ AA living B V. lantana 2 L. Arnold Arboretum collection JQ JQ JX JQ JX JQ JX JX JX V. macrocephalum Fortune M.J. Donoghue 101 YU HQ HQ EF EF HQ HQ HQ HQ HQ HQ V. mongolicum Rehder M.J. Donoghue s.n. YU HQ HQ EF EF HQ HQ HQ HQ HQ HQ V. rhytidophyllum Hemsl. Ex Forb. & Hemsl. R.C. Winkworth 8 YU HQ HQ AY AY HQ HQ HQ HQ HQ HQ V. schensianum Maxim. Boufford et al A HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ cultivated V. utile Hemsl. Egolf 2336E plant AY HQ AY AY HQ HQ HQ HQ EF HQ V. veitchii C.H. Wright Bouffourd et al A HQ HQ HQ HQ HQ HQ HQ HQ HQ Lentago V. cassinoides L. Arnold Arboretum 87485A, A HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ V. elatum Benth. M.J. Donoghue 472 YU AY HQ AY AY HQ HQ EF HQ V. lentago L. R.C. Winkworth 21 YU AY HQ AY AY HQ HQ HQ HQ EF HQ V. prunifolium L. R.C. Winkworth 13 YU AY HQ AY AY HQ HQ HQ HQ EF HQ V. rufidulum Raf. R.C. Winkworth 14 YU AY HQ AY AY HQ HQ HQ HQ EF HQ Lobata V. acerifolium L. R.C. Winkworth 27 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ V. kansuense Batalin Boufford et al A AY HQ AY AY HQ HQ HQ HQ EF HQ V. orientale Pall. Merello et al MO HQ HQ EF EF HQ HQ HQ HQ EF HQ Lutescentia V. colebrookeanum Wall. Parker 3220 A HQ HQ HQ HQ HQ HQ HQ HQ V. lutescens Blume Wu et al. WP531 A HQ HQ HQ HQ HQ HQ HQ HQ HQ Mollodontotinus A, V. bracteatum Rehder Arnold Arboretum A HQ HQ HQ HQ HQ HQ HQ HQ V. ellipticum Hook. M.J. Donoghue NVI AY AY AY HQ HQ HQ HQ HQ HQ V. molle Michx. R.C. Winkworth 5 YU AY HQ AY AY HQ HQ HQ HQ EF B, V. rafinesquianum Schult. Arnold Arboretum A JX JX JX JX JX JX JX JX JX JX Opulus V. edule (Michaux) Raf. NVI AY AY AY HQ HQ EF V. koreanum Nakai H. Yamaji 5170 MO HQ EF EF EF V. opulus L. W.L. Clement 250 YU HQ HQ HQ HQ HQ HQ HQ HQ V. sargentii Koehne R.C. Winkworth 17 YU AY HQ AY AY HQ HQ HQ HQ EF HQ A, V. trilobum Marshall Arnold Arboretum AA HQ HQ HQ HQ HQ HQ HQ HQ EF HQ Oreinodontotinus V. caudatum Greenm. M.J. Donoghue 64 YU HQ HQ HQ HQ HQ HQ , A V. dentatum 1 L. Arnold Arboretum JQ JQ JX JQ JX JX JX JX V. dentatum 2 L. R.C. Winkworth 33 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ AA living V. dentatum 3 L. Arnold Arboretum C collection JX JX JX JX JX JX JX JX JX A, A V. dentatum 4 L. Arnold Arboretum JQ JQ JQ JQ JX JQ JX JX JX B, A V. dentatum 5 L. Arnold Arboretum JQ JQ JQ JQ JQ JQ JX JX JX V. hartwegii Benth. M.J. Donoghue 486 YU AY HQ AY AY HQ HQ HQ HQ HQ V. loeseneri Graebn. M.J. Donoghue 2547 YU HQ HQ HQ HQ HQ HQ HQ HQ V. stenocalyx Hemsl. M.J. Donoghue 60 YU HQ HQ HQ HQ HQ HQ HQ HQ V. triphyllum Benth. P.W. Sweeney, W.L. Clement, J. Yepez. PWS1783 YU HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ Pseudotinus V. furcatum Bl. ex Hook.f. & Thomson Tsugaru & Takashi MO AY HQ AY AY HQ HQ HQ HQ EF HQ V. lantanoides Michx. R.C. Winkworth 2 YU AY HQ AY AY HQ HQ HQ HQ EF HQ V. nervosum D. Don Boufford et al A AY HQ AY AY HQ HQ HQ HQ EF V. sympodiale Graebn. Lai & Shan 4529 MO HQ HQ EF EF HQ HQ HQ EF HQ Punctata V. lepidotulum Merr. & Chun Lau A HQ HQ HQ HQ HQ591740

2 V. punctatum Buch.Ham. SinoBritish Ex D. Don Expedition 133 A HQ HQ HQ HQ HQ HQ HQ HQ HQ Sambucina V. inopinatum Craib Rock 1603 A HQ HQ JQ HQ V. ternatum Rehder Bartholomew et al A HQ HQ HQ HQ HQ HQ HQ HQ HQ Solenotinus V. amplificatum J. Kern Pereira et al. JTP 677 A HQ HQ HQ JQ V. awabuki Hort.Berol. Ex K. Koch Liu 141 A HQ HQ HQ HQ HQ HQ HQ HQ V. brachybotryum Hemsl. Ex Forb. & Hemsl. NVI HQ HQ HQ V. chingii P.S. Hsu Bartholomew et al. 973 A HQ HQ HQ HQ HQ V. erubescens Wall. Boufford et al A AY HQ AY AY HQ HQ HQ HQ HQ HQ V. farreri Stearn R.C. Winkworth 18 YU AY HQ AY AY HQ HQ HQ HQ EF HQ V. odoratissimum Ker Gawl. R. Olmstead 118 WTU AY HQ AY AY HQ HQ HQ HQ HQ V. oliganthum Batalin Bouffourd et al A HQ HQ HQ HQ HQ HQ HQ HQ V. sieboldii Miq. R.C. Winkworth 3 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ V. subalpinum Hand. Mazz. Heng A HQ HQ HQ HQ HQ HQ HQ HQ V. suspensum Lindl. R.C. Winkworth 36 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ Succodontotinus V. betulifolium Batalin Boufford et al A HQ HQ HQ HQ HQ HQ HQ V. cf. corylifolium Hook f. AA living & Thomson Arnold Arboretum 10399A collection JX JX JX JX JX JX JX JX JX V. dilatatum Thunberg R.C. Winkworth 19 YU AY HQ AY AY HQ HQ HQ HQ HQ V. erosum Thunberg R.C. Winkworth 16 YU AY HQ AY AY HQ HQ HQ HQ EF HQ V. flavescens W.W. Smith Boufford et al A HQ HQ HQ HQ HQ HQ HQ JX HQ V. foetidum Wall. C.H. Lin 563 MO HQ HQ HQ EF HQ HQ HQ JX HQ V. hupehense Rehder Bartholomew et al A HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ V. ichangense Rehder Bartholomew et al A HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ V. japonicum Spreng NVI AY HQ AY AY HQ HQ HQ HQ HQ HQ V. lobophyllum Graebn. R.C. Winkworth 25 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ V. luzonicum Rolfe Shen 673 A HQ HQ HQ HQ HQ HQ JX HQ JX HQ V. melanocarpum P.S. Hsu R.C. Winkworth 12 YU AY HQ AY AY HQ HQ HQ HQ HQ V. sempervirens K. Koch Hu & But A HQ HQ HQ HQ HQ HQ HQ V. setigerum Hance M.J. Donoghue 102 YU HQ HQ HQ EF HQ HQ HQ HQ HQ HQ V. wrightii Miquel Yonekura 1362 A HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ Tinus V. atrocyaneum C.B. Clarke Boufford et al A HQ HQ HQ HQ HQ HQ HQ HQ HQ HQ V. calvum Rehder Li & Soukup 934 A HQ HQ HQ HQ HQ HQ HQ HQ JX HQ V. cinnamomifolium Rehder Olmstead 120 WTU AY HQ AY AY HQ HQ HQ HQ HQ HQ V. propinquum Hemsl. M.J. Donoghue 100 YU HQ HQ EF EF HQ HQ HQ HQ V. rigidum Ventenat Stearn 1116 A HQ HQ HQ HQ HQ HQ HQ HQ V. tinus L. R.C. Winkworth 35 YU AY HQ AY AY HQ HQ HQ HQ HQ HQ Tomentosa V. plicatum Thunberg R.C. Winkworth 10 YU AY HQ AY AY HQ HQ HQ HQ EF HQ Unassigned V. clemensiae Kern J. Beaman K AY HQ AY AY HQ HQ HQ HQ EF HQ V. taiwanianum Hayata W.H. Hu et al MO HQ HQ EF EF HQ HQ HQ HQ HQ V. urceolatum S. & Z. M.J. Donoghue NVI AY HQ AY AY HQ HQ HQ HQ HQ592053

3 Appendix B. Quantifying leaf shape: We assembled leaf images for 88 focal taxa both from living Viburnum plants at the Arnold Arboretum of Harvard University (Jamaica Plain, MA) and from herbarium specimens in the Harvard University Herbaria (HUH) and the Yale Herbarium in the Peabody Museum of Natural History (YU). For the living collections (39 taxa) we photographed the adaxial surface of individual leaves on a white background with a scale bar. For herbarium specimens (49 taxa), we photographed entire herbarium sheets and then digitally traced and excised individual leaves, which were scaled and cropped into their own image files. We digitally cropped the petiole from each leaf, and scaled, reoriented, and binarized these images so that each contained a black silhouette of the adaxial lamina surface on a white background. The resulting dataset included individual leaf images for each taxon. We quantified the leaf outline directly using Elliptical Fourier analysis (EFA). This allows the outline of any closed, twodimensional shape to be quantified numerically, as a series of coefficients describing a set of harmonic ellipses [1]. McLellan and Endler [2] found that EFA described complex leaf shapes better than either singleparameter metrics of outline shape or traditional multivariate analysis of linear landmark measurements. We used the software suite SHAPE ver. 1.3 [3] to quantify the leaf outline using EFA for the lamina silhouette in each leaf image. We extracted the lamina outline as a chain code, rotationally aligning all outlines along a common homologous vector (leaf midrib), and calculated the elliptic Fourier descriptors for 300 harmonic ellipses for each laminar outline; this was the maximum number of harmonics possible with this software, producing the best possible fit to the original outline. In EFA, size can be represented as a function of the 0th Fourier harmonic. However, our preliminary analyses found size to be by far the greatest component of overall shape, obscuring other components of leaf morphology. Therefore we set SHAPE to scale our elliptic Fourier descriptors so that the 0th harmonic was constant across all leaves, which effectively controlled for variability in size. We then considered size separately from overall shape, measuring leaf area directly from the leaf images. We used principal component analysis (PCA) to reduce the large set of potentially correlated Fourier descriptors to a smaller set of uncorrelated principal components. These principal components defined independent axes of shape variation in lamina outline, allowing the principal component scores for a leaf to be treated as

4 continuous morphological characters. Because traditional PCA relies on the assumption that all observations are independent, we also performed a phylogenetic PCA [4] using the phylogeny described above. Because these analyses are far more computationally intensive, we analyzed a 10% subsample of the 300 harmonic ellipses (the first 120 EFD coefficients, describing 30 harmonic ellipses). Our phylogenetic PCA analysis produced nearly identical results to the nonphylogenetic analyses: in both cases, the first two principal component axes together accounted for > 95% of the variation in leaf shape observed in Viburnum. These axes showed a slight yet significant phylogenetic signal. Multivariate lambda for the principal component axes was near zero (λ 0.045). Nevertheless, the multivariate lambda value calculated was a better fit (AIC= ) than a multivariate lambda of zero (AIC= ). We conducted all subsequent regression analyses using the phylogenetic PCA axes only. References: 1 Kuhl, F. & Giardina, C Elliptic Fourier features of a closed contour. Computer Graphics and Image Processing 18, McLellan, T. & Endler, J The relative success of some methods for measuring and describing the shape of complex objects. Systematic Biology 47, Iwata, H. & Ukai, Y SHAPE: a computer program package for quantitative evaluation of biological shapes based on elliptic Fourier descriptors. J. Hered. 93, Revell, L. J Sizecorrection and principal components for interspecific comparative studies. Evolution 63, (doi: /j x)

5 Supplemental Table 1. Testing for correlated trait evolution using maximum likelihood. Dependent indicates correlated evolution; bestchosen model highlighted in bold. Independent Dependent Dependent Dependent Equalrates Symmetricrates Allratesdifferent lnlik AIC lnlik AIC lnlik AIC lnlik AIC HabitatA x MarginTypeA HabitatA x MarginTypeB HabitatA x LeafHabitA HabitatA x LeafHabitB HabitatB x MarginTypeA HabitatB x MarginTypeB HabitatB x LeafHabitA HabitatB x LeafHabitB HabitatC x MarginTypeA HabitatC x MarginTypeB HabitatC x LeafHabitA HabitatC x LeafHabitB HabitatA: tropical vs. cloud forest, warm temperate, and cold temperate (TR vs. CL+WT+CT); HabitatB: cold temperate vs. tropical, warm temperate, and cloud forest (CT vs. TR+WT+CL); HabitatC: tropical and cloud forest vs. warm temperate and cold temperate (TR+CL vs. WT+CT); MarginTypeA: entire margins vs. prominent, regular teeth and reduced/irregular teeth (EN vs. PT+RT); MarginTypeB: prominent, regular teeth vs. reduced/irregular teeth and entire margins (PT vs. RT+EN); LeafHabitA: evergreen vs. leafexchanger and seasonally deciduous (EV vs. LE+SD); LeafHabitB: seasonally deciduous vs. leafexchanger and evergreen (SD vs. LE+EV).

6 Supplemental Table 2. Phylogenetic regressions between continuous response variables and discrete character states. Bestchosen models are highlighted in bold. Habitat Leaf Habit Margin Type CT vs. TR+WT+CL TR+CL vs. WT+CT TR vs. CL+WT+CT EV vs. LE+SD SD vs. LE+EV PT vs. RT+EN EN vs. PT+RT PC1 Brownian loglik AIC pvalue OU alpha loglik AIC pvalue Lambda lambda loglik AIC pvalue Leaf Brownian loglik Size AIC pvalue OU alpha loglik AIC pvalue Lambda lambda loglik AIC pvalue CT= cold temperate; WT= warm temperate; CL= cloud forest; TR=tropical; EV=evergreen; LE=leafexchanger; SD=seasonally deciduous; EN=entire margins; RT=reduced/irregular teeth; PT=prominent, regular teeth.

7 Supplemental Table 3. Phylogenetic regressions between sets of continuous variables. Bestchosen model highlighted in bold. Brownian OrnsteinUhlenbeck Lambda loglik AIC pvalue alpha loglik AIC pvalue lambda loglik AIC pvalue PC1~MAT PC1~T season PC1~cold quarter PC1~warm quarter PC1~MAP PC1~D teeth PC1~size Size~MAT Size~ T season Size~MAP D teeth ~MAT D teeth ~T season D teeth = # teeth*cm 1 ; MAT = Mean Annual Temperature; T season = Temperature seasonality (standard deviation of monthly temperatures*100); cold quarter= Mean Temperature of the coldest quarter; warm quarter=mean Temperature of the warmest quarter; MAP=Mean Annual Precipitation

8 percentage of Viburnum species with toothed leaves Ecuador 0/8 Sabah 0/4 Guatemala 3/8 Luzon 2/6 Puebla 4/5 Taiwan 7/12 Arkansas 5/6 Kyushu 9/11 Asia The Americas Hokaido 5/5 Maine 6/ o Latitude ( N) Supplementary Figure 1. Latitudinal gradient in leaf margins in Viburnum. In both the Old World and the New World there are more species with entire margins (or highly reduced teeth) at lower latitudes, and more with prominently toothed margins at high latitudes. Each dot marks the midpoint of the latitudinal range of the named geographic region; ratio indicates the number of species with teeth/total number of species in the region.

9 probability of being tropical probability of being evergreen Supplementary Figure 2. Correlations between probabilites of character states at internal nodes in Viburnum probability of having entire margins