reduction in chemical potential by forming thecrystalline phase. In the second scenario, weassume that oxygen has first adsorbed to theLS interface, driven by the reduction in inter-face energy (24, 28), having diffused along theinterface from the triple junction. Once oxygenadsorbs, a critical-size nucleus (in the form ofa step) forms at the triple junction and sweepsacross the LS (0001) interface, forming a new(0006) layer of a-Al2O3 in its wake. Oxygen thenagain adsorbs to the new LS interface, suppliedfrom the dissolving outer rim of the nanowire.The nucleation rate of steps is relatively low un-der the present experimental conditions, leadingto single steps that sweep across the interface (nodefects resulting from the intersection of stepsor multiple nucleation events were detected).Both of these reactions would be very fast; theformation of a new (0006) layer after oxygendiffusion has started requires only ~0.16 s. Ourexperimental observations cannot distinguishbetween these two scenarios.

Our in situ HRTEM observations have re-vealed an oscillatory growth process that is notobvious from ex situ observations of grown nano-wires. These observations provide an explanationfor the (0001) growth mechanism and why it isnot continuous. What distinguishes this reactionfrom most VLS processes is that the catalystdroplet is not directly involved in supplying allthe necessary components for growth. Instead,oxygen for the new (0006) layer is supplied bytransient sacrificial dissolution of the top rim ofthe nanowires via facet formation. The compo-nents of interface tensions at the triple junctionalso promote formation of the facets and probablydrive the nucleation of steps of solid (0006),which sweep across the LS (0001) interface, re-sulting in growth of a new (0006) layer. Forma-tion of the (0006) layer is relatively fast becausethe Al atoms on the liquid side of the LS surface

are already in positions similar to those in the un-derlying bulk sapphire. The oxygen collectedon the LV surface is used to eradicate the facetsat the outer, top rim of the nanowires once acomplete (0006) layer is formed. It is this processthat is the rate-limiting step.

References and Notes1. E. I. Givargizov, J. Cryst. Growth 31, 20 (1975).2. F. M. Ross, J. Tersoff, M. C. Reuter, Phys. Rev. Lett. 95,

146104 (2005).3. Y. Hao, G. Meng, Z. L. Wang, C. Ye, L. Zhang, Nano Lett.

6, 1650 (2006).4. L. E. Jensen et al., Nano Lett. 4, 1961 (2004).5. S. Kodambaka, J. Tersoff, M. C. Reuter, F. M. Ross,

Science 316, 729 (2007).6. V. Valcárcel, A. Souto, F. Guitián, Adv. Mater.

10, 138 (1998).7. W. F. Li et al., Phys. Stat. Sol. A 203, 294 (2006).8. T. Kuykendall et al., Nat. Mater. 3, 524 (2004).9. E. A. Stach et al., Nano Lett. 3, 867 (2003).10. Y. Wu, P. Yang, J. Am. Chem. Soc. 123, 3165

(2001).11. S. Hofmann et al., Nat. Mater. 7, 372 (2008).12. G. Lee et al., Angew. Chem. 48, 7366 (2009).13. B. Tian, P. Xie, T. J. Kempa, D. C. Bell, C. M. Lieber,

Nat. Nanotechnol. 4, 824 (2009).14. S. H. Oh, Y. Kauffmann, C. Scheu, W. D. Kaplan, M. Rühle,

Science 310, 661 (2005); published online6 October 2005 (10.1126/science.1118611).

15. See supporting material on Science Online.16. V. Laurent, D. Chatain, C. Chatillon, N. Eustathopoulos,

Acta Metall. 36, 1797 (1988).17. D. Chatain, W. C. Carter, Nat. Mater. 3, 843

(2004).18. E. J. Siem, W. C. Carter, D. Chatain, Philos. Mag. 84,

991 (2004).19. H. Wang, G. S. Fischman, J. Appl. Phys. 76, 1557

(1994).20. H. Saka, K. Sasaki, S. T. Tsukimoto, S. Arai, J. Mater. Res.

20, 1629 (2005).21. D. J. Siegel, L. G. Hector Jr., J. B. Adams, Phys. Rev. B 65,

085415 (2002).22. W. D. Kaplan, Y. Kauffmann, Annu. Rev. Mater. Res. 36,

1 (2006).23. A. Hashibon, J. Adler, M. W. Finnis, W. D. Kaplan,

Interface Sci. 9, 175 (2001).24. G. Levi, W. D. Kaplan, Acta Mater. 50, 75 (2002).

25. The reaction volume was calculated by assuming a ringof triangular cross section. Calculation of the volumeyields ~1.7 × 10−20 cm3. Given that the number ofoxygen atoms per unit volume of a-Al2O3 is 8.63 × 1022

atoms cm−3, the number of oxygen atoms containedwithin the reaction volume is ~1.46 × 103 atoms. Thenumber of oxygen atoms needed to deposit a monolayerof close-packed oxygen at the LS (0001) interface wascalculated by assuming a square and a circular shapeof the interfaces. The area density of a hexagonal close-packed oxygen layer is 1.6 × 1015 atoms cm−2. Theamount of oxygen needed to deposit a monolayer ofclose-packed oxygen is ~1.6 × 103 and ~1.2 × 103

atoms for the square and the circular interfaces,respectively.

26. S. M. Roper et al., J. Appl. Phys. 102, 034304(2007).

27. K. W. Schwarz, J. Tersoff, Phys. Rev. Lett. 102,206101 (2009).

28. G. Levi, W. D. Kaplan, J. Am. Ceram. Soc. 85, 1601(2002).

29. The in situ TEM experiments were carried out using theStuttgart high-voltage microscope (JEM-ARM 1250) at theMax-Planck-Institut für Metallforschung. We thankF. Phillipp, R. Höschen, and U. Salzberger for technicalsupport, and D. Chatain and the participants of the 3Phase, Interface, Component Systems (PICS3) meeting forstimulating discussions. Supported by the GermanScience Foundation via the Graduiertenkolleg InnereGrenzflächen (GRK 285/3), the German-Israel Fund(grant I-779-42.10/2003), the Nano and Steel FusionResearch Program funded by the POSCO (S.H.O.) BasicScience Research Institute grant 2.0006028 (S.H.O.), theBasic Science Research Program through the NationalResearch Foundation funded by the Ministry ofEducation, Science and Technology (grant 2010-0005834) (S.H.O.), the excellence cluster NanosystemsInitiative Munich (C.S. and S.H.O.), and the MaterialsSciences and Engineering Division of the U.S.Department of Energy (M.F.C. and W.L.).

Supporting Online Materialwww.sciencemag.org/cgi/content/full/330/6003/489/DC1SOM TextFigs. S1 to S5ReferencesMovies S1 to S3

7 April 2010; accepted 16 September 201010.1126/science.1190596

Species Selection MaintainsSelf-IncompatibilityEmma E. Goldberg,1 Joshua R. Kohn,2 Russell Lande,3 Kelly A. Robertson,1Stephen A. Smith,4 Boris Igic 1*Identifying traits that affect rates of speciation and extinction and, hence, explain differences inspecies diversity among clades is a major goal of evolutionary biology. Detecting such traits isespecially difficult when they undergo frequent transitions between states. Self-incompatibility,the ability of hermaphrodites to enforce outcrossing, is frequently lost in flowering plants, enablingself-fertilization. We show, however, that in the nightshade plant family (Solanaceae), specieswith functional self-incompatibility diversify at a significantly higher rate than those without it.The apparent short-term advantages of potentially self-fertilizing individuals are therefore offsetby strong species selection, which favors obligate outcrossing.

Theastonishing diversity of flowering plantshas led to numerous hypotheses linkingthe propensity for diversification to mor-

phological, life history, and physiological char-

acteristics (1–4). Although it is clear that nosingle character can entirely explain the variationin plant diversity (2, 4), many traits potentiallyassociated with an increase in species richness

may have been overlooked because of method-ological shortcomings (5, 6).

The overwhelming majority of floweringplants function as both males and females, butdespite the potential for self-fertilization in her-maphrodites, most of these plants predominantlyor exclusively mate with other individuals (out-cross) (7, 8). Outcrossing is often enforced byself-incompatibility (SI): the ability of individualplants to recognize and reject their own pollen.Each SI system characterized to date relies on

1Department of Biological Sciences, University of Illinois atChicago, 840 West Taylor Street, M/C 067, Chicago, IL 60607,USA. 2Division of Biological Sciences, University of California atSan Diego, 9500 Gilman Drive, La Jolla, CA 92093, USA.3Division of Biology, Imperial College London, Silwood ParkCampus, Ascot, Berkshire, SL5 7PY, UK. 4National EvolutionarySynthesis Center 2024 West Main Street, Durham, NC 27705,USA.

*To whom correspondence should be addressed. E-mail:boris@uic.edu

www.sciencemag.org SCIENCE VOL 330 22 OCTOBER 2010 493

REPORTS

on

Oct

ober

21,

201

0 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

complex genetic mechanisms (9). Their value isprimarily attributed to the avoidance of inbreed-ing depression (10), but the short-term disad-vantages of SI, which limits an individual plant'sability to reproduce if outcross pollen is not avail-able, is expected to result in frequent transitions toself-compatibility (SC) (8). Mate limitation andthe automatic advantage of selfing (10, 11) eachfavor the spread of mutants that break down SI.Indeed, angiosperm families in which SI is foundalso contain significant proportions of SC species(12). The two states are commonly interspersedwithin clades, implying frequent evolutionary tran-sitions. Independent data, both genetic and phylo-genetic (12, 13), suggest that the transition fromSI to SC is one of the most common evolutionaryshifts in flowering plants (14). Studies of the ge-netic basis of SI indicate that novel SI systemsoriginate very rarely (9, 15). The combined evi-dence implies that SI is frequently lost by speciesand that an identical form of SI is only rarely, ifever, regained (12).

In Solanaceae species with SI, strong negativefrequency-dependent selection maintains dozensof alleles at the locus controlling self-recognition(S-locus), where it has preserved ancestral poly-morphism over 40 million years (13, 16, 17). Afterpollination, if the haploid pollen allele matcheseither allele of the diploid female reproductiveorgan, fertilization is prevented. Individuals withcommon alleles therefore have fewer potentialmates, and those with rare alleles more, so thatalleles ordinarily lost by drift are instead main-tained over long time scales (16). The outcome isa pattern of ancestral polymorphism shared notonly across species boundaries but among distant-ly related genera. Randomly sampled alleles fromdifferent SI species of Solanaceae are often moreclosely related than those within species (16, 17).This provides strong evidence that a polymorphicand therefore functional S-locus was present inthe common ancestor and passed on to descend-ent SI species in the family (17).

Although the SI mechanism has been fre-quently lost and not regained within Solanaceae(13), it occurs at an intermediate frequency. Theproportion of SI species is declining with eachtransition to SC, so SI may be destined for ex-tinction. But the SI mechanism of Solanaceae is~90 million years old, nearly as old as eudicots(15), and SI may instead persist if the diversifi-cation rate of SI lineages exceeds that of SClineages by at least the SI-to-SC transition rate(17). In this scenario, the loss of SI may be count-ered by a greater difference between speciationand extinction rates associated with the presenceof SI relative to SC. A strong association betweenSI and increased diversification would provideunusually compelling evidence for a causal re-lationship, because frequent character changes toSC uncouple the evolution of the focal characterfrom others and reduce the chance that an inferredconnection is spurious (18).

We examined the effect of SI on diversifi-cation in Solanaceae. The nightshades contain

many agriculturally important species (such astomato, potato, chile, and tobacco), so phyloge-netic and breeding system data are extensive.Among the estimated 2700 species in the fam-ily, approximately 57% are SC, 41% are SI, andless than 2% have separate male and femaleplants (19).

We constructed a phylogeny through an ini-tial alignment matrix of 998 species and eightgenes, and we assigned character states (SI or SC)for 616 species (19). Subsequent analyses con-centrated on a monophyletic clade termed “x =12” for its uniform base chromosome number,containing approximately 2300 species, includingthe subfamilies Solanoideae and Nicotianoideae(20) [we also conducted analyses on the wholefamily (19)]. Both character states and phylo-genetic data are better sampled in this clade thanin the rest of the family, yielding 356 speciespresent in the tree andwith knownmating systems

(Fig. 1). This sample did not statistically differ inSI frequency from the total available data set of616 species (P > 0.3), and our analyses accountedfor sampling incompleteness (21).

Irreversible loss of SI in Solanaceae willcause the frequency of SI to decline to zero if rI ≤rC + qIC, where rI and rC are net diversificationrates (the difference between speciation andextinction rates) for SI and SC lineages, respec-tively, and qIC is the per-lineage rate of transitionfrom SI to SC. Alternatively, the loss of SI maybe offset by more rapid diversification of SIlineages when rI > rC + qIC. In this latter case,the equilibrium proportion of SI species, pSI, ispositive and equal to 1 − qIC/(rI − rC) (17).

We analyzed character evolution and diversi-fication on the maximum likelihood tree (Fig. 1)and also on 100 bootstrap trees to assess robust-ness to phylogenetic uncertainty (19). Rates ofspeciation (lI and lC; the subscripts denote SI

Fig. 1. Maximum likelihood tree of phylogenetic relationships among 356 species of Solanaceae.Higher ranks are indicated around the perimeter of the tree. Purple and turquoise tip colors denote SIand SC extant species, respectively. The root age is 36 million years. Inset panels display posteriorprobability distributions and 95% credibility intervals of reconstructed rates of character evolution (thetime unit is millions of years). (A) BiSSE estimates of transition, speciation, and extinction parameters(qIC << mI < lI << lC < mC). (B) Net diversification rate—the difference between speciation andextinction rates—associated with each state. (C) Schematic summary of estimated rate parameters. Formethods, species names, character states, and further results, see (19).

22 OCTOBER 2010 VOL 330 SCIENCE www.sciencemag.org494

REPORTS

on

Oct

ober

21,

201

0 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

and SC states, respectively), extinction (mI and mC),and loss of SI (qIC) were estimated with Markovchain Monte Carlo analyses under the binary-state speciation and extinction (BiSSE) model(19, 22), which explicitly incorporates differingspeciation and extinction rates associated withthe two states of a binary character. Because SIhas apparently not been regained in this clade(13), we fixed the rate of transitions from SC toSI, qCI, to be zero in our main analyses [althoughwe also considered bidirectional transitions (19)].

We find a higher rate of speciation for SCthan SI lineages, but this apparent advantage iserased by an SC extinction rate that exceeds thatof SI lineages and even the SC speciation rate[mI < lI < lC < mC, with posterior probability 0.99(Fig. 1A)]. SI lineages consequently show asubstantially higher rate of net diversificationthan do SC lineages [rI > rC, with posterior prob-ability 1.0 (Fig. 1B)]. Furthermore, this diversi-fication difference is large enough to counter theirreversible loss of SI [rI > rC + qIC, with posteriorprobability 1.0 (Fig. 2A)]. The equilibrium fre-quency of SI is therefore positive [pSI > 0.25,with posterior probability 0.99 (Fig. 2B)], demon-strating that the macroevolutionary advantage ofSI is enough for it to persist indefinitely, despiteits loss through transitions to SC. These resultsare robust to phylogenetic uncertainty and againstthe assumption that SI is not regained (19).

The higher rate of diversification associatedwith SI means that selection is acting amongspecies, which can produce trends that are notpredictable from studies of selection among in-dividuals within species. Natural selection can actonmany levels of evolutionary hierarchy (includ-ing genes, individuals, populations, and species),with the potential for cascading effects of se-lection from any level to those above and below(18, 23–26). Species, as units, may vary in her-itable traits associated with different rates of mul-tiplication. A higher form of selection may thusbe manifested as a trait-dependent net diversifica-tion rate. There is wide acceptance of the existenceof higher-order selection (18, 25, 26) but rela-tively sparse data on its importance (25, 26).

Our results point to a strong role for speciesselection in determining the evolutionary dynam-ics and distribution of mating systems amongspecies. Both theoretical and empirical studies onthe shape of outcrossing rate distributions haveoverlooked the possibility that species selectionmay produce an arbitrary proportion of selfing,mixed-mating, and outcrossing species (8, 27).

It remains unknown why strict avoidance ofself-fertilization yields such strong long-term evo-lutionary advantages. Outcrossing plants, whichgenerallymaintain high effective population sizesand recombination rates, may have increased poly-morphism, lower linkage disequilibrium, moreefficient response to purifying selection, moreextensive geographic distribution of genetic di-versity, and lower rates of extinction (28–31).Various mechanisms other than SI, such as thephysical and temporal separation of sexes withinhermaphrodite flowers, or single-sex individuals,can effect outcrossing in the absence of SI (3).Moreover, although SI may be a relatively ef-ficient mechanism for obligate outcrossing (32),delayed selfing ability and plastic sex expressionafforded by other breeding systems may addi-tionally provide reproductive assurance (7).

At short time scales, the expression of inbreed-ing depression can oppose the spread of muta-tions that break down SI, but many ecologicalfactors favor their rapid and frequent fixation (7).This is problematic because natural selectioncannot maintain features slated for future use(18). We find that SC is associated with a macro-evolutionary disadvantage from high extinctionrates, despite its high speciation rates relative toSI. Both theoretical and empirical evidence linkinbreeding with an increased probability of ex-tinction (27,29,31). Self-fertilization and increasedinbreeding are also closely tied to higher among-population differentiation (30), which may in turnlead to escalating speciation rates.

We conclude that species selection opposesthe short-term advantages provided by the abilityof flowering plants to self-fertilize, and that vol-atile dynamics underlie the scattered phyloge-netic distribution of SC plant species. A strong

relationship between traits, includingSI, and specia-tion and extinction may result in the unevenpatterns of diversification rates observed amongclades of flowering plants (2, 4), especially whencoupled with time heterogeneity and historicalcontingency. These results provide evidence thatthe evolutionary origin and maintenance of SImay play an important role in shaping the immenseextant diversity of flowering plants (18, 33).

References and Notes1. C. Darwin, F. Darwin, A. C. Seward, More Letters of

Charles Darwin: A Record of His Work in a Series ofHitherto Unpublished Letters ( J. Murray, London, 1903).

2. S. Magallón, M. J. Sanderson, Evolution 55, 1762 (2001).3. S. C. H. Barrett, Philos. Trans. R. Soc. London Ser. B 358,

991 (2003).4. T. J. Davies et al., Proc. Natl. Acad. Sci. U.S.A. 101,

1904 (2004).5. W. P. Maddison, Evolution 60, 1743 (2006).6. E. E. Goldberg, B. Igić, Evolution 62, 2727 (2008).7. C. Goodwillie, S. Kalisz, C. G. Eckert, Annu. Rev. Ecol.

Evol. Syst. 36, 47 (2005).8. B. Igić, J. R. Kohn, Evolution 60, 1098 (2006).9. V. Franklin-Tong, Self-Incompatibility in Flowering

Plants: Evolution, Diversity, and Mechanisms(Springer, Berlin, 2008).

10. E. Porcher, R. Lande, Evolution 59, 46 (2005).11. J. W. Busch, D. J. Schoen, Trends Plant Sci. 13, 128 (2008).12. B. Igić, R. Lande, J. R. Kohn, Int. J. Plant Sci. 169, 93

(2008).13. B. Igić, L. Bohs, J. R. Kohn, Proc. Natl. Acad. Sci. U.S.A.

103, 1359 (2006).14. G. L. Stebbins, Flowering Plants (Belknap Press,

Cambridge, MA, 1974).15. J. E. Steinbachs, K. E. Holsinger, Mol. Biol. Evol. 19,

825 (2002).16. T. R. Ioerger, A. G. Clark, T. H. Kao, Proc. Natl. Acad. Sci.

U.S.A. 87, 9732 (1990).17. B. Igić, L. Bohs, J. R. Kohn, New Phytol. 161, 97 (2004).18. G. C. Williams, Natural Selection: Domains, Levels,

and Challenges (Oxford Univ. Press, New York, 1992).19. Materials, methods, and further results are available as

supporting material on Science Online.20. R. G. Olmstead, L. Bohs, Acta Hortic. 745, 255 (2007).21. R. G. FitzJohn, W. P. Maddison, S. P. Otto, Syst. Biol. 58,

595 (2009).22. W. P. Maddison, P. E. Midford, S. P. Otto, Syst. Biol. 56,

701 (2007).23. R. C. Lewontin, Annu. Rev. Ecol. Syst. 1, 1 (1970).24. S. J. Gould, Science 216, 380 (1982).25. D. Jablonski, Annu. Rev. Ecol. Evol. Syst. 39, 501 (2008).26. D. L. Rabosky, A. R. McCune, Trends Ecol. Evol. 25, 68 (2010).27. D. J. Schoen, J. W. Busch, Int. J. Plant Sci. 169,

119 (2008).28. G. L. Stebbins, Am. Nat. 91, 337 (1957).29. D. J. Schoen, A. H. Brown, Proc. Natl. Acad. Sci. U.S.A.

88, 4494 (1991).30. J. L. Hamrick, M. J. W. Godt, Philos. Trans. R. Soc. London

Ser. B 351, 1291 (1996).31. S. Glémin, E. Bazin, D. Charlesworth, Proc. R. Soc.

London Ser. B 273, 3011 (2006).32. K. Karoly, Am. Nat. 144, 677 (1994).33. M. S. Zavada, T. N. Taylor, Am. Nat. 128, 538 (1986).34. We thank R. FitzJohn, A. Mooers, R. Olmstead, T. Paape,

L. Popović, D. L. Shade, and R. Ree for help anddiscussions. Data are deposited in the Dryad databasewith accession no. 1888. Funded by NSF grant DEB-0919089 to B.I.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/330/6003/493/DC1Materials and MethodsSOM TextFigs. S1 and S2Tables S1 to S4References

1 July 2010; accepted 25 August 201010.1126/science.1194513

0.0 0.2 0.4 0.6 0.8 1.0

01

23

45

6

rate

prob

abili

ty d

ensi

ty

rI − rC − qIC

0.0 0.1 0.2 0.3 0.4 0.5

02

46

8

frequency

pSIA B

Fig. 2. Further analyses of the posterior rate distributions in Fig. 1. (A) The diversification rate of SI issufficiently higher than that of SC for it to persist deterministically (shown by positive values here) despitefrequent and irreversible loss. (B) The equilibrium frequency of SI is consequently always positive.

www.sciencemag.org SCIENCE VOL 330 22 OCTOBER 2010 495

REPORTS

on

Oct

ober

21,

201

0 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

www.sciencemag.org/cgi/content/full/330/6003/493/DC1

Supporting Online Material for

Species Selection Maintains Self-Incompatibility E. E. Goldberg, J. R. Kohn, R. Lande, K. A. Robertson, S. A. Smith, B. Igic∗

*To whom correspondence should be addressed. E-mail: boris@uic.edu

Published 22 October 2010, Science 330, 493 (2010)

DOI: 10.1126/science.1194513

This PDF file includes:

Materials and Methods SOM Text Figs. S1 and S2 Tables S1 to S4 References

Materials and methods

Phylogeny construction

Alignment and tree construction pipeline

The procedure for reconstructing phylogenetic relationships among a subset of Solanaceae

closely followed the one developed and reported in detail by Smith and colleagues (S1, S2).

We implemented this pipeline in Python (v2.5) with the BioPython (v1.48) module (S3) and

using the BioSQL (v1.0.1) database schema (S4). Here, we present the most important aspects

of this procedure.

We found eight loci within Solanaceae with more than 200 taxa represented in GenBank.

Six of these genes are encoded in the chloroplast genome: atpB (ca. 300 entries), matK (520),

ndhF (500), rbcL (220), trnK (210), trnL-trnF (700), and two in the nuclear genome: GBSSI

(790), ITS (910). We removed taxa with identical (duplicated) names and produced a first-

pass multiple alignment using MAFFT (S5), which resulted in a 1084-taxon, 16,405bp (eight-

gene) matrix. Gene-partitioned phylogenetic reconstructions were performed using RAxML

(v7.0.4) (S6). Subsequent analyses were performed on the maximum likelihood (ML) tree found

in a heuristic search and on 100 bootstrap replicates. We manually curated the tree obtained after

the initial pass. The resulting matrix contained 995 ingroup taxa, plus five outgroup species of

Convolvulaceae (Fig. S1).

The resulting trees were then pruned to include only those species contained within the

“x=12” clade, the best-sampled monophyletic group in the family. Cultivars and known hybrid

varieties were removed from the analyses, as were those species for which the self-incompatibility

status is unknown. The final dataset contained 356 species (Fig. S2).

In addition, we present below abbreviated results for the dataset consisting of all 397 species

in the family on the tree and with known character states.

1

Relaxed molecular clock implementation

In order to ultrametricize our phylogenetic trees and obtain node and tip dates proportional to

true time (rather than substitutions per site), we used the semiparametric method developed

by Sanderson (S7). This method accommodates rate variation among lineages by combining a

parametric model with different nucleotide substitution rates on every branch with a nonpara-

metric roughness penalty. The λ parameter for penalized likelihood calculations determines the

contribution of this penalty. It was estimated using a cross-validation procedure implemented

in r8s (v1.71) (S7). We applied 13 values of λ, spaced evenly on a logarithmic scale from 1 to

1000, after randomly pruning the ML tree to 71 taxa (for computational feasibility) 10 times. In

each pruned set, the best estimate was log10 λ = 2.25, which we used for penalized likelihood

rate-smoothing on the ML and bootstrap trees.

Accurate absolute divergence dates are not crucial for our analyses. Nevertheless, we ob-

tained approximate values in order to make the units of the results of our character evolution

analyses meaningful for comparative study. We constrained the age of the focal crown group

(“x=12” clade) to 36 Ma based on the node date obtained by Paape and colleagues (S8). This

dating procedure relies on the use of two key constraints derived from dating and identification

of Solanum- and Physalis-like seeds (S9). Those two dates were used to anchor a normally

distributed age prior (µ = 10.0 Ma, σ = 4.0 Ma; 95% CI = [3.4, 16.6]), along with the date

for stem group age of Solanaceae (divergence from Convolvulaceae; µ = 52 Ma, σ = 5.2 Ma;

95% CI = [43, 60]). The divergence dates we obtained are comparable to others recovered with

different strategies [e.g., (S10, S11)].

Character evolution analyses

Testing for a correlation between character state and net diversification rate has commonly

relied on associations of clade age and number of species, either through the use of sister clade

2

or regression analyses (S12–S14). These methods consider only clades in which most tips

are in the same state and assume that ancestors within the clade had that state. Because the

transition rate from SI to SC is high relative to the family’s age (S15), few non-polymorphic

clades are available, and such tests will have low statistical power (S13, S16). We therefore

employ the recently-developed BiSSE method for estimating from a phylogeny speciation and

extinction rates that depend on character states (S17). We include the correction for incomplete

sampling (S18): the probability of being sampled in the tree is 0.162 for SI species and 0.150

for SC species. By explicitly modeling speciation and extinction associated with the two states

of a binary character, BiSSE allows us to quantify the effects of state-dependent diversification

and state transitions.

Character state encoding

For macroevolutionary models of binary character evolution to be applied to breeding system

evolution, each species must be classified as either SI or SC, and this can be difficult (S19). We

use the same general principles as described elsewhere (S20). The best available data, however

sparse, indicate that the breakdown of SI is so common that no SI species studied in any detail

fails to reveal segregating SC mutants, SC populations, or sister species. Our intention is to

measure the average rate at which an entire lineage, ancestrally SI (see main text and (S20)),

transitions to SC. Therefore, we encode as SI all “polymorphic” taxa that have not entirely

transitioned to SI. This encoding should yield a conservative estimate of the advantage of SI

lineages. The breeding system data were available for 356 species represented on the phylogeny

of “x=12” Solanaceae. Of those, 196 were described in the literature and encoded by us as SC.

Of the remaining species encoded as SI, 116 were unambiguously described as SI, 17 as both

SI and SC, 25 as dioecious, and one each as SI/SC/dioecious and SI/dioecious.

3

Bayesian analysis

We performed a Bayesian analysis of BiSSE on the Solanaceae ML phylogeny. Compared

with maximum likelihood parameter estimation, this approach better handles the potentially-

rough likelihood surfaces and provides a natural means of testing how meaningful parameter

differences are. Although a Bayesian analysis across a posterior set of trees would be more

philosophically consistent, obtaining such a set was not computational feasible due to the large

number of species and sites in the phylogenetic analysis.

We obtained 1,000,000 post-burn-in samples of slice-sampling Markov chain Monte Carlo

(MCMC) on the maximum likelihood tree and 10,000 samples on each of the 100 bootstrap

trees. Priors on each parameter were exponentially distributed and broad, with rate 0.3. We

used the conditional likelihood root state assumption (S18): because the rate of gain of SI was

fixed to zero, this effectively fixes as SI the root and hence the other internal nodes leading to

SI tips. Analyses were conducted with the R package diversitree (S21, S22). Once posterior

distributions of the five model parameters were obtained (Fig. 1A), further analyses (Fig. 1B,

Fig. 2) simply involved computing additional quantities from each MCMC sample. Reported

posterior probabilities are the proportion of MCMC samples for which a statement was true.

Supporting text

Additional character evolution results

Results from maximum likelihood tree

Mean values, standard deviations, and the standard errors of the mean (determined from time

series analysis as implemented in the R package coda (S23)) are reported in Table S1. The

results for λC − λI and µC − µI provide further quantification of our finding that SC lineages

have a higher rate of speciation and an even higher rate of extinction than do self-incompatible

lineages. The results for rI , rC and their difference show not only a higher rate of net diversifi-

4

cation for SI lineages, but in fact negative net diversification for SC lineages. The equilibrium

frequency of SI is only slightly less than the observed frequency and is certainly positive, show-

ing that SI is not bound for deterministic extinction due to its loss through transitions to SC.

Correlations between the five directly estimated parameters are reported in Table S2. The

speciation and extinction rates for each state are very strongly correlated, consistent with the

much tighter estimates of net diversification rates than of speciation and extinction rates sep-

arately (Fig. 1AB, Table S1). The moderate correlations between the transition rate and the

rest of the rates reflects the balance between state change and diversification in explaining the

observed state frequencies.

Results from bootstrap trees

The results reported above include uncertainty from the parameter estimation process but not

uncertainty in the tree topology or branch lengths. Variation in individual parameter estimates

across the bootstrap trees is comparable to uncertainties in estimation on the ML tree (Table S3).

For net diversification rates, however, variation across bootstrap trees is much less (Table S3),

likely because diversification rate is determined largely by the frequency of tip states and the

root age, both of which are constant across the bootstrap set.

Our main results are robust to phylogenetic uncertainty. On 92 out of the 100 bootstrap

trees, µI < λI < λC < µC is true with a posterior probability of at least 0.90. On all 100

bootstrap trees, the following statements are true with posterior probability 0.90: rI > 0.35,

rC < −0.01, rI − rC − qIC > 0.12, pSI > 0.27.

Results with regain of SI allowed

Although we feel we are justified in fixing the rate of SI gain, qCI , to zero in our analyses, we

understand that results without that restriction may be of interest. We therefore also ran the

BiSSE MCMC analysis on the ML tree without the general qCI = 0 constraint, but incorporat-

5

ing information on specific lineages in which reversals were shown not to occur. For the seven-

teen species on our tree in which trans-specific polymorphism has been documented (Solanum

chilense, S. peruvianum, S. carolinense, S. chacoense, Lycium californicum, L. andersonii,

L. parishii, L. ferocissimum, L. cestroides, Physalis cinerascens, P. crassifolia, P. longifolia,

Witheringia solanacea, Nicotiana alata, N. glauca, Iochroma australe, Eriolarynx lorentzii),

every node that is an ancestor of that species was fixed to the SI state by setting its conditional

likelihood of SC to zero in all tree likelihood calculations.

The estimate of qCI is positive but small, and the estimates for other parameters (Table S4)

are similar to those with qCI = 0 (Table S1). Our main conclusions all hold under this anal-

ysis. The ordering of the speciation and extinction rates is the same (the posterior probability

of µI < λI < λC < µC is 0.99). Diversification is greater for SI lineages (rI > rC with pos-

terior probability 1.0). The equilibrium frequency of SI is positive (pSI > 0.30 with posterior

probability 0.90; the pSI computation is somewhat more complicated than the expression given

in the main text when reversals are allowed (S17)). Compared with the results in which qCI is

fixed to zero, the diversification advantage of SI is smaller when regain of SI is allowed but the

equilibrium frequency of SI is slightly larger because the rate of loss of SI is smaller.

Results from whole-family analysis

On the full Solanaceae tree, pruned only to contain species with known SI status, overall sam-

pling intensity is somewhat less than that of the “x=12” clade used in the main analyses: the

sampling proportions are 0.123 for SI and 0.168 for SC. Inclusion in the tree is much less

uniform outside of the “x=12” clade (e.g., Petunia is very well represented but the other deep

genera are not), so we did not use the whole-family in the main analyses.

Nevertheless, it is worthwhile to ask if our results for the clade spanning Solanum and

Nicotiana also hold for what is known of the entire family. We find that the ordering of the

6

speciation and extinction rates is not as clear (on the ML tree, the posterior probability of

µI < λI < λC < µC is only 0.61), but the transition rate qIC is always much less than these

rates, and all of our other conclusions hold. Specifically, with posterior probabilities of at least

0.99, rC < −0.09, rI > 0.26, rI > rC + qIC , and pSI > 0.32.

7

Dinetus truncatusConvolvulus arvensis

Ipomoea purpureaIpomoea batatas

Evolvulus glomeratus

Schizanthus grahamiiSchizanthus hookeri

Schizanthus candidusSchizanthus lacteus

Schizanthus integrifolius

Schizanthus alpestris

Schizanthus laetusSchizanthus porrigens

Schizanthus litoralis

Schizanthus tricolor

Schizanthus parvulus

Schizanthus pinnatus

Tsoala tubiflora

Metternichia principisCoeloneurum ferrugineumEspadaea amoenaGoetzea elegans

Goetzea ekmanii

Henoonia myrtifolia

Duckeodendron cestroides

Vestia lycioides

Vestia foetida

Sessea corymbifloraCestrum megalophyllum

Cestrum strigilatumCestrum acutifolium

Cestrum virgaurea

Cestrum dasyanthum

Cestrum guatemalense

Cestrum pacayense

Cestrum regelii

Cestrum aurantiacum var macrocalyx

Cestrum aurantiacum

Cestrum roseum

Cestrum oblongifolium

Cestrum thyrsoideum

Cestrum miradorense

Cestrum fasciculatum

Cestrum elegans

Cestrum endlicheri

Cestrum laxum

Cestrum fulvescens

Cestrum poasanum

Cestrum sp Montero 213

Cestrum fragileCestrum chiriquianum

Cestrum irazuense

Cestrum luteovirescens

Cestrum nocturnum

Cestrum milciomejiaeCestrum sphaerocarpum

Cestrum violaceumCestrum parqui

Cestrum glanduliferum

Cestrum tuerckheimiiCestrum tomentosum

Cestrum mortonianum

Cestrum rigidum

Cestrum pittieri

Cestrum inclusum

Cestrum macrophyllum

Sessea vestita

Sessea stipulataStreptosolen jamesonii

Browallia eludens Browallia speciosa

Protoschwenkia mandonii

Salpiglossis sinuata Combera paradoxa

Pantacantha ameghinoi

Melananthus guatemalensis

Schwenckia lateriflora

Schwenckia glabrata

Nierembergia frutescens

Schwenckia americanaLeptoglossis linifolia

Plowmania nyctaginoides

Hunzikeria texana

Bouchetia erecta

Bouchetia anomala

Nierembergia calycina

Nierem

bergia angustifolia

Nierem

bergia aristata

Nierem

bergia veitchii

Nierem

bergia rigida

Nierem

bergia spathulata

Nierem

bergia repens

Nierem

bergia rivularis

Nierem

bergia pinifolia

Nierem

bergia graveolens

Nierem

bergia micrantha

Nierem

bergia hatschbachii

Nierembergia linariifolia var linariifolia

Nierembergia hippom

anica

Nierembergia scoparia

Nierembergia ericoides

Nierembergia tucum

anensis

Nierembergia pulchella var pulchella

Nierem

bergia browallioides

Nierem

bergia boliviana

Nierem

bergia pulchella var macrocalyx

Leptoglossis darcyana

Brunfelsia americana

Brunfelsia uniflora

Fabiana imbricata

Calibrachoa sellowiana

Calibrachoa pygmaea

Calibrachoa parviflora

Petunia axillaris

Petunia axillaris subsp axillaris

Petunia interior

Petunia integrifolia subsp inflata

Petunia axillaris subsp parodii

Petunia bonjardinensis

Petunia integrifolia var depauperata

Petunia littoralis

Petunia integrifolia

Petunia reitzii

Petunia saxicola

Petunia scheideana

Petunia integrifolia var integrifolia

Petunia guarapuavensis

Petunia exserta

Petunia occidentalis

Petunia axillaris subsp subandina

Petunia bajeensis

Petunia riograndensis

Petunia altiplana

Petunia mantiqueirensis

Petunia secreta

Brow

allia americana

Sym

onanthus bancroftii

Sym

onanthus aromaticus

Anthocercis gracilis

Anthocercis littorea

Anthocercis ilicifolia

Anthocercis sylvicola

Anthocercis intricata

Anthocercis angustifolia

Anthocercis viscosa

Anthotroche w

alcottii

Anthotroche pannosa

Anthotroche blackii

Anthotroche m

yoporoides

Gram

mosolen dixonii

Gram

mosolen truncatus

Cyphanthera odgersii

Cyphanthera m

icrophylla

Crenidium

spinescens

Cyphanthera anthocercidea

Duboisia hopw

oodii

Duboisia m

yoporoides

Cyphanthera albicans

Duboisia leichhardtii

Anthocercis racem

osa

Anthocercis m

yosotidea

Nicotiana trigonophylla

Nico

tiana o

btu

sifolia

Nicotiana palm

eri

Nicotiana bigelovii

Nicotiana clevelandii

Nico

tiana ta

bacu

m

Nico

tiana d

iglu

ta

Nico

tiana to

mento

sa

Nico

tiana ka

waka

mii

Nico

tiana to

mento

siform

is

Nico

tiana se

tchellii

Nico

tiana o

tophora

Nico

tiana sp

egazzin

ii

Nico

tiana lin

earis

Nico

tiana m

iersii

Nico

tiana p

auciflo

ra

Nico

tiana a

ttenuata

Nico

tiana a

cum

inata

Nico

tiana co

rymbosa

Nico

tiana la

ngsd

orffii

Nico

tiana b

onarie

nsis

Nico

tiana fo

rgetia

na

Nico

tiana x sa

ndera

eN

icotia

na lo

ngiflo

ra

Nico

tiana p

icillaN

icotia

na p

lum

bagin

ifolia

Nico

tiana a

lata

Nico

tiana a

fricana

Nico

tiana fra

gra

ns

Nico

tiana d

ebneyi

Nico

tiana sim

ula

ns

Nico

tiana m

egalo

siphon

Nico

tiana in

gulb

aN

icotia

na u

mbra

tica

Nico

tiana d

idepta

Nico

tiana e

xigua

Nico

tiana ro

sula

ta

Nico

tiana g

oodsp

eedii

Nico

tiana g

osse

iN

icotia

na ve

lutin

a

Nico

tiana a

mple

xicaulis

Nico

tiana m

aritim

a

Nico

tiana e

xcelsio

rN

icotia

na b

enth

am

iana

Nico

tiana su

ave

ole

ns

Nico

tiana ca

vicola

Nico

tiana e

astii

Nico

tiana ro

tundifo

lia

Nico

tiana o

ccidenta

lisN

icotia

na h

esp

eris

Nico

tiana sylve

stris

Nico

tiana n

udica

ulis

Nico

tiana re

panda

Nico

tiana sto

cktonii

Nico

tiana n

eso

phila

Nico

tiana a

caulis

Nico

tiana g

lauca

Nico

tiana n

octiflo

ra

Nico

tiana p

etu

nio

ides

Nicotiana cordifolia

Nicotiana solanifolia

Nicotiana paniculata

Nicotiana knightiana

Nicotiana rustica

Nicotiana benavidesii

Nicotiana raim

ondii

Nicotiana glutinosa

Nicotiana arentsii

Nicotiana undulata

Nicotiana w

igandioides

Nicotiana thyrsiflora

Exo

deco

nus m

iersii

Lyci

um m

asca

rene

nse

Lyci

um in

tric

atum

Lyci

um te

tran

drum

Lyci

um b

arba

rum

Lyci

um rut

heni

cum

Lyci

um c

hine

nse

Lyci

um d

epre

ssum

Lyci

um e

urop

aeum

Lyci

um p

ilifo

lium

Lyci

um o

xyca

rpum

Lyci

um a

frum

Lyci

um a

reni

cola

Lyci

um h

orrid

um

Lyci

um s

chiz

ocal

yx

Lyci

um fe

roci

ssim

um

Lyci

um c

iner

eum

Lyci

um s

p N

309

Lyci

um g

rand

ical

yx

Lyci

um a

moe

num

Lyci

um s

trand

veld

ense

Lyci

um te

nue

Lyci

um a

ustra

le

Lyci

um g

arie

pens

e

Lyci

um p

umilu

m

Lyci

um d

ecum

bens

Lyci

um a

cutif

oliu

m

Lyci

um e

enii

Lyci

um s

haw

ii

Lyci

um s

chw

einf

urth

iiLy

cium

bos

ciifo

lium

Lyci

um h

irsut

um

Lyci

um v

illos

umLy

cium

pru

nus

spin

osa

Gra

bowsk

ia g

lauc

a

Gra

bowsk

ia b

oerh

aviifol

ia

Gra

bowsk

ia d

uplic

ata

Gra

bow

skia

obt

usa

Lycium

pub

erul

um

Lycium

mac

rodo

n

Lycium

coo

peri

Lycium

sho

ckle

yi

Lycium

pal

lidumLy

cium

gilli

esia

num

Lycium

am

eghino

i

Lycium

chilens

e

Lycium

cilia

tum

Lycium

cha

nar

Lyciu

m d

eser

ti

Lyciu

m fu

scum

Lyciu

m m

inut

ifoliu

m

Lyciu

m s

teno

phyll

um

Lycium

cal

iforn

icum

Lycium

nod

osum

Lycium

vim

ineu

m

Phrod

us b

ridge

sii

Lyciu

m c

arolin

ianu

m va

r caro

linia

num

Lycium

minim

um

Lycium

ath

ium

Lycium exs

ertum

Lycium

frem

ontiiLycium sp

202

Lycium anderso

nii

Lycium paris

hii

Lycium texanum

Lycium torre

yi

Lycium berla

ndieri subsp

parviflo

rum

Lycium berla

ndieri

Lycium berla

ndieri subsp

berlandieri

Lyciu

m e

longatu

m

Lyciu

m in

faustu

m

Lyciu

m a

meric

anum

Lyciu

m b

revip

es

Lyciu

m ra

chidocla

dum

Lyciu

m ca

rolin

ianum va

r quadrif

idum

Lyciu

m sa

ndwicense

Lyciu

m te

nuispinosu

m

Lyciu

m le

iosp

ermum

Lyciu

m c

estro

ides

Lyciu

m c

unea

tum

Lyciu

m m

oron

gii

Phrod

us m

icro

phyllu

s

arolfili

ss

es

an

alo

N

Nola

na tara

paca

na

Nola

na g

aya

na

Nola

na tom

ente

lla

Nola

na s

p D

illon 8

790

Nola

na p

ilosa

Nola

na c

onfin

is

Nola

na jo

hnst

onii

Nola

na iv

ania

na

Nola

na ly

cioid

es

Nola

na p

alli

da

Nola

na thin

ophila

Nola

na v

olc

anic

a

Nola

na c

err

ate

ana

Nola

na in

tonsa

Nola

na d

iffusa

Nola

na s

edifo

lia

Nola

na c

livic

ola

Nola

na m

olli

sN

ola

na a

plo

cary

oid

es

Nola

na s

als

olo

ides

Nola

na r

am

osi

ssim

a

Nola

na s

p D

illon 8

591

Nola

na a

lbesc

ens

Nola

na toco

pill

ensi

s

Nola

na p

eru

viana

Nola

na g

lauca

Nola

na le

pto

phy

lla

Nol

ana

flacc

ida

Bar

gem

ontia

sph

aero

phy

lla

Nol

ana

villo

sa

Nol

ana

sp D

illon

572

9

Nola

na d

ivarica

ta

Nola

na in

cana

Nola

na s

p 4

13

Nola

na w

erd

erm

annii

Nola

na c

rass

ulif

olia

No

lan

a s

ten

op

hyl

la

No

lan

a b

als

am

iflu

aN

ola

na

lin

eari

folia

No

lan

a p

tero

carp

a

Nola

na b

acc

ata

Nola

na p

arv

iflora

Nola

na e

legans

Nola

na s

p D

illon 8

690

Nola

na p

ara

doxa

Nola

na a

cum

inata

Nola

na r

upic

ola

Nola

na c

arn

osa

Nola

na fili

folia

Nola

na r

ost

rata

Nola

na c

oele

stis

Nola

na s

p 4

14

Nola

na s

path

ula

ta

Nola

na p

licata

Nola

na w

eis

siana

Nola

na in

flata

Nola

na a

ticoana

Nola

na s

p L

eiv

a e

t al 2

212

Nola

na h

um

ifusa

Nola

na u

rubam

bae

Nola

na s

p D

illon 8

789

Nola

na s

p Q

uip

usc

oa 2

785

Nola

na s

caposa

Nola

na la

xa

Nola

na a

renic

ola

Nola

na a

danso

nii

Nola

na g

ala

pagensi

s

iis

eillig

xal

yh

por

elc

S7

58

05

sh

oB

dn

a e

eN

ps

xal

yh

por

elc

S

ailofit

an

da

xal

yh

por

elc

S

sn

ec

se

nip

s x

aly

hp

orel

cS

Atro

pa

be

llad

on

na

su

cit

ug

nat

s

ud

osi

nA

su

dir

ul

su

do

sin

A

acil

oinr

ac

ailo

po

cS

siral

ubi

dn

ufni

ani

alh

co

sy

hP

se

diol

as

yh

p a

nial

hc

os

yh

Psil

atn

eiro

ani

alh

co

sy

hP

Sco

po

lia ja

po

nic

a

aci

tu

gn

at

aik

sla

we

zrP

sis

ne

nis

eht

na

por

tA

su

eru

a s

um

ay

cs

oy

Hr

egi

n s

um

ay

cs

oy

H

Hyo

scyam

us m

uticu

s

mur

otr

es

ed

su

ma

yc

so

yH

su

na

ev

ob

su

ma

yc

so

yH

sul

lis

up

su

ma

yc

so

yH

Hyo

scyam

us a

lbu

s

Latu

a p

ubiflo

raJa

boro

sa in

tegrifo

lia

Jab

oro

sa sa

tivaJa

boro

sa sq

uarro

sa

Schultesianthus m

egalandrus

Schultesianthus le

ucanthusSolandra brachycalyx

Solandra grandiflora

Dyssochroma viridiflo

ra

Merinthopodium neuranthum

Markea ulei

Markea panamensis

Hawkesiophyton ulei

Juanullo

a mexicana

Juanullo

a aurantiaca

Mandragora officinarum

Mandragora caulescens

Mandragora chinghaiensisNicandra physalodes

Brugmansia aurea

Brugmansia sanguinea

Datura leichhardtii

Datura metel

Datura inoxia

Datura stramonium

Iochroma cardenasianum

Witheringia cuneata

Witheringia coccoloboides

Larnax hunzikeriana

Cuatresia riparia

Cuatresia exiguiflora

Cuatresia colombianaIochroma sp Smith 317

Iochroma nitidum

Iochroma umbellatum

Iochroma grandiflorum

Iochroma sp Smith 370

Iochroma sp Smith 337

Iochroma squamosum

Iochroma lehmanniiIochroma stenanthumIochroma loxense

Iochroma cyaneum

Iochroma cornifolium

Iochroma confertiflorum

Acnistus arborescensIochroma salpoanum

Iochroma sp Smith 353

Iochroma edule

Iochroma ellipticum

Iochroma gesnerioides

Iochroma sp Smith 476

Iochroma calycinum

Iochroma fuchsioides

Dunalia spinosa

Iochroma parvifoliumEriolarynx fasciculataEriolarynx lorentzii

Iochroma australe

Dunalia brachyacantha

Vassobia dichotoma

Larnax hawkesiiLarnax peruviana

Larnax suffruticosaDeprea orinocensis

Larnax psilophyta

Larnax cuyacensis

Larnax grandiflora

Deprea glabraDeprea bitteriana

Deprea paneroi

Vassobia fasciculataDunalia fasciculata

Vassobia breviflora

Dunalia obovata

Dunalia spathulata

Dunalia solanacea

Saracha quitensis

Saracha punctata

Aureliana fasciculata var fasciculata

Deprea sylvarum

Larnax subtrifloraLarnax sachapapa

Larnax sylvarum

Mellissia begoniifolia

Withania frutescensWithania somnifera

Withania sp W009Withania coagulans

Athenaea sp Bohs 91Aureliana fasciculata

Nothocestrum latifoliumNothocestrum longifolium

Discopodium penninerviumTubocapsicum anomalum

Tzeltalia amphitrichaTzeltalia calidaria

Brachistus stramonifolius

Witheringia solanacea

Witheringia meianthaWitheringia macrantha

Leucophysalis viscosa

Witheringia mexicana

Oryctes nevadensis

Quincula lobataPhysalis microphysa

Physalis sp P099

Physalis minima

Physalis ignotaPhysalis cordata

Physalis longifolia var subglabrataPhysalis pruinosa

Physalis angulataPhysalis pubescensPhysalis lagascae

Physalis mexicana

Physalis ixocarpa

Physalis philadelphica var immaculata

Physalis sp P149

Physalis aff philadelphica MW 2004

Physalis philadelphica

Physalis microcarpa

Physalis nicandroidesPhysalis patula

Physalis chenopodiifolia

Physalis lassa

Physalis angustiphysaPhysalis peruviana

Physalis aff heterophylla MW 2004

Physalis viscosaPhysalis mollis

Physalis pumila

Physalis heterophylla

Physalis virginiana

Physalis arenicolaPhysalis lanceolata

Physalis coztomatl

Physalis sordida

Physalis glutinosa

Physalis hederifolia

Physalis longifolia

Physalis caudella

Physalis hintonii

Physalis campanulata

Physalis greenmanii

Physalis hederifolia var puberula

Physalis angustifolia

Physalis walteri

Physalis cinerascens

Physalis curassavica

Physalis mendocina

Physalis fuscomaculata

Physalis aequata

Physalis minimaculata

Margaranthus solanaceus

Physalis lanceifoliaPhysalis acutifoliaPhysalis crassifoliaChamaesaracha coronopus

Chamaesaracha sordidaPhysalis arborescensPhysalis melanocystisPhysalis alkekengi

Physalis carpenteriLeucophysalis grandifloraLeucophysalis nana

Jaltomata grandiflora

Jaltomata procumbens

Jaltomata auriculata

Jaltomata sinuosa

Salpichroa tristis

Salpichroa origanifolia

Nectouxia formosa

Lycianthes shanesii

Lycianthes biflora

Lycianthes denticulata

Lycianthes lysimachioides

Lycianthes synanthera

Lycianthes sanctaeclarae

Lycianthes pringlei

Lycianthes beckneriana

Lycianthes tricolor

Lycianthes multiflora

Lycianthes furcatistellata

Lycianthes pseudolycioides

Lycianthes glandulosaLycianthes nitida

Lycianthes jalicensis

Lycianthes surotatensisLycianthes lenta

Lycianthes geminiflora

Lycianthes heteroclita

Lycianthes stephanocalyx

Capsicum lanceolatum

Capsicum ciliatum

Capsicum rhom

boideum

Capsicum gem

inifolium

Capsicum lycianthoides

Capsicum parvifolium

Capsicum cam

pylopodium

Capsicum flexuosum

Capsicum

schottianum

Capsicum

hunzikerianum

Capsicum

recurvatum

Capsicum

villosum

Capsicum

pereirae

Capsicum

annuum

Capsicum

tovarii

Capsicum

pubescens

Capsicum

galapagoense

Capsicum

baccatum

Capsicum

coccineum

Capsicum

chacoense

Capsicum

minutiflorum

Capsicum

ceratocalyx

Capsicum

chinense

Capsicum

frutescens

Capsicum

annuum var annuum

Capsicum

cardenasii

Capsicum

eximium

Lycianthes pilifera

Lycianthes lycioides

Lycianthes rantonnei

Lycianthes saltensis

Lycianthes fasciculata

Lycianthes jelskii

Lycianthes acapulcensis

Lycianthes dejecta

Lycianthes peduncularis

Lycianthes rzedowskii

Lycianthes moziniana var margaretiana

Lycianthes ciliolata

Lycianthes acutifolia

Lycianthes inaequilatera

Lycianthes radiata

Lycianthes amatitlanensis

Lycianthes asarifolia

Jaltomata hunzikeri

Jaltomata dentata

Solanum

thelopodium

Sola

num

ule

anu

m

Sola

num

trizygum

Sola

num

phase

olo

ides

Sola

num

evolvu

lifoliu

m

Sola

num

fraxin

ifoliu

m

Sola

num

taenio

trichum

Sola

num

murica

tum S

ola

num

appendicu

latu

m

Sola

num

palu

stre

Sola

num

etu

bero

sum

Sola

num

more

lliform

e

Sola

num

claru

m

Sola

num

circaeifo

lium

Sola

num

polya

deniu

m

Sola

num

pin

natise

ctum

Sola

num

bulb

oca

stanum

Sola

num

ehre

nberg

ii

Sola

num

card

iophyllu

m

Sola

num

jam

esii

Sola

num

stenophyllid

ium

Sola

num

arn

ezi

i

So

lanu

m y

un

ga

sense

So

lanu

m c

ha

coe

nse

So

lanu

m le

pto

phye

s

So

lanu

m o

plo

cense

es

nejir

at m

un

alo

S

Sola

num

infu

ndib

ulif

orm

eS

ola

num

bert

haulti

iS

ola

num

verr

uco

sum

Sola

num

spars

ipilu

mS

ola

num

avi

lesi

i

Sola

num

oka

dae

Sola

num

ugentii

mut

no

dor

cim

mu

nal

oS

ie

nre

v m

un

alo

S

So

lanu

m k

urt

zia

nu

m es

ne

sa

nira

m m

un

alo

S

mu

nis

ob

ma

mu

nal

oS

mu

na

ello

dn

ac

mu

nal

oS

iiv

os

ak

ub

mu

nal

oS

mu

sor

eb

ut m

un

alo

S

elu

aci

ver

b m

un

alo

SS

ola

nu

m lig

nica

ule

So

lanu

m a

cha

cach

en

se

So

lanu

m sa

nto

lalla

em

umi

ssi

xal

m

un

alo

S

So

lanu

m p

am

pa

sen

sem

uil

ofi

na

hp

ar

mu

nal

oS

mu

pra

cy

xo

mu

nal

oS

mui

lof

in

omi

rg

a m

un

alo

S

es

ne

mol

ad

nut

m

un

alo

S

mut

aru

dn

ap

bu

s m

un

alo

Sm

un

aib

mol

oc

mu

nal

oS

mu

cin

oci

gn

ol

mu

nal

oS

mut

aro

mra

mie

cal

oiv

mu

nal

oS

Sola

num

alb

icans

Sola

num

dem

issum

Sola

num

aca

ule

Sola

num

iopeta

lum

Sola

num

doddsii

Sola

num

ala

ndia

e

Sola

num

andre

anum

Sola

num

mosco

panum

Sola

num

hje

rtingii

Sola

num

stolo

nife

rum

Sola

num

inca

mayo

ense

Sola

num

schenckii

Sola

num

guerre

roense

Sola

num

hougasii

Sola

num

spegazzin

ii

Sola

num

com

merso

nii

Sola

num

multiin

terru

ptu

m

Sola

num

acro

scopicu

m

Sola

num

imm

ite

Sola

num

chom

ato

philu

mS

ola

num

pasco

ense

Sola

num

piu

rae

Sola

num

scabrifo

lium

Sola

num

lycopersico

ides

Sola

num

sitiens

Sola

num

habro

chaite

s

Sola

num

pennellii

Sola

num

chm

iele

wskii

Sola

num

neorickii

Sola

num

lycopersicu

m

Sola

num

lycopersicu

m va

r cera

siform

e

Sola

num

cheesm

ania

e

Sola

num

gala

pagense

Sola

num

pim

pin

ellifo

lium

Sola

num

chile

nse

Sola

num

peru

vianum

Sola

num

och

ranth

um

Sola

num

jugla

ndifo

lium

Solanum

pubigerum

Solanum

aligerum

Solanum

nitidum

Solanum

amygdalifolium

Solanum

crispum

Solanum

riojense

Solanum

americanum

Sola

num

ptych

anth

um

Solanum

opacum

Solanum

fiebrigii

Solanum

physalifolium

Solanum

palitans

Solanum

tripartitum

Solanum

ipomoeoides

Solanum

laxum

Solanum

calileguae

Solanum

dulcamara

Solanum

wallacei

Solanum

seaforthianumS

olanum caesium

Solanum

triflorum

Solanum

villosum

Solanum

retroflexum

Solanum

nigrumSolanum

quadrangulare

Solanum

terminale

Solanum

aggregatum

Solanum

symonii

Solanum

linearifolium

Solanum

aviculare

Solanum

laciniatum

Solanum

trisectum

Solanum

herculeum

Solanum

multifidum

Solanum

montanum

Sola

num

allo

phyl

lum

Sola

num

more

llifo

lium

Sola

num

mapir

iense

Sola

num

anom

alu

mS

ola

num

nem

ore

nse

Sola

num

repta

ns

Sola

num

hoehnei

Sola

num

fusi

form

e

Sola

num

melis

saru

m

Sola

num

tobagense

Sola

num

div

ers

ifoliu

m

Sola

num

oxy

phyl

lum

Sola

num

tegore

Sola

num

circi

natu

m

Sola

num

circi

natu

m s

ubsp

circi

natu

m

Sola

num

tenuis

eto

sum

Sola

num

pro

teanth

um

Sola

num

circi

natu

m s

ubsp

ram

osu

m

Sola

num

occ

ultu

m

Sola

num

sib

undoy

ense

Sola

num

endopogon s

ubsp

guia

nensi

s

Sola

num

caja

num

ense

Sola

num

falla

x

Sola

num

calid

um

Sola

num

obliq

uum

Sola

num

rose

um

Sola

num

mate

rnum

Sola

num

beta

ceum

Sola

num

unilo

bum

Sola

num

cyl

indricu

m

Sola

num

gla

uco

phyl

lum

Sola

num

stu

ckert

ii

Sola

num

confu

sum

Sola

num

hib

ern

um

Sola

num

lute

oalb

um

Sola

num

am

ota

pense

Sola

num

pendulu

m

Sola

num

cory

mbifl

oru

m

Sola

num

latif

loru

m

Sola

num

exi

guum

Sola

num

caco

smum

Sola

num

dip

loco

nos

Sola

num

pin

eto

rum

Sola

num

sci

adost

ylis

Sola

num

hav

anense

Sola

num

och

rophy

llum

Sola

num

aphy

odendro

n

Sola

num

capsi

cast

rum

Sol

anum

rov

irosa

num

Sola

num

arg

entin

um

Sola

num

pse

udoca

psi

cum

Sola

num

arb

ore

um

Sol

anum

turn

eroi

des

Sol

anum

ads

cend

ens

Sol

anum

sch

lech

tend

alia

num

Sol

anum

rug

osum

Sol

anum

mau

ritia

num

Sol

anum

cor

dove

nse

Sol

anum

lepi

dotu

m

Sol

anum

abu

tiloi

des

Sol

anum

inel

egan

s

Sol

anum

ref

ract

umS

olan

um b

icor

ne

Sol

anum

wen

dlan

dii

Sol

anum

cla

ndes

tinum

Sol

anum

del

itesc

ens

Sol

anum

am

eric

anum

sub

sp n

odifl

orum

Sol

anum

pol

ygam

um

Solan

um a

ccre

scen

s

Solan

um ro

bust

um

Solan

um s

tagn

ale

Solan

um m

icro

phyllu

m

Solan

um a

grar

ium

Solan

um s

tena

ndru

m

Solan

um c

arol

inen

se

Solanu

m c

onditum

Solan

um c

ompt

um

Solanu

m ja

maice

nse

Solanum

atu

rense

Solanu

m a

dhaerens

Solanum m

ultispinum

Solanum cr

otonoides

Solanum g

lutinosu

m

Solanum cr

initipes

Solanum p

aniculatu

m

Solanum to

rvum

Solanum la

nceolatu

m

Solan

um d

rym

ophi

lum

Solan

um b

aham

ense

Solan

um c

ampe

chie

nse

Solanum sp

Goldblatt

12426

Solanum h

ispidum

Solanum erianthum

Solanum ra

cem

osum

Solanum elaeagnifoliu

m

Solanum tridyn

amum

Solanum hindsianum

Solanum sejunctum

Solanum asymmetriphyllumSolanum cf centrale AMH 1

Solanum centraleSolanum quadriloculatum

Solanum esuriale

Solanum pugiunculiferum

Solanum nummularium

Solanum nemophilum

Solanum cunninghamii

Solanum carduiformeSolanum leopoldense

Solanum cataphractum

Solanum petraeum

Solanum sp Martine 805

Solanum vansittartenseSolanum tudununggaeSolanum aff vansittartense Fryxell 3421

Solanum dioicum

Solanum eremophilum

Solanum ellipticum

Solanum cleistogamum

Solanum aff echinatum Bohs 2727

Solanum echinatumSolanum tetrathecumSolanum orbiculatum

Solanum heteropodiumSolanum oedipus

Solanum melanospermumSolanum clarkiae

Solanum beaugleholei

Solanum phlomoides

Solanum lasiophyllum

Solanum chippendaleiSolanum diversiflorum

Solanum hystrixSolanum hoplopetalum

Solanum petrophilum

Solanum campanulatumSolanum prinophyllum

Solanum stupefactum

Solanum nobileSolanum brownii

Solanum curvicuspeSolanum neoanglicum

Solanum pancheri

Solanum sandwicenseSolanum incompletumSolanum vaccinioides

Solanum stelligerum

Solanum chenopodinum

Solanum cookii

Solanum corifoliumSolanum ferocissimumSolanum parvifolium

Solanum furfuraceumSolanum cinereum

Solanum coacti

liferu

m

Solanum renschii

Solanum giganteum

Solanum thruppii

Solanum aculeastrum

Solanum coagulans

Solanum schimperia

num

Solanum kwebense

Solanum marginatum

Solanum richardii

Solanum incanum

Solanum arundo

Solanum cumingii

Solanum ovigerum

Solanum melongenaSolanum lichtensteinii

Solanum linnaeanum

Solanum sessilistellatumSolanum campylacanthum

Solanum panduriforme

Solanum violaceum

Solanum supinum

Solanum manaense

Solanum macrocarpon

Solanum dasyphyllum

Solanum vespertilio

Solanum lidii

Solanum rigescens

Solanum capense

Solanum tomentosum

Solanum hastifolium

Solanum cyaneopurpureum

Solanum anguivi

Solanum aethiopicum

Solanum virginianum

Solanum myo

xotrichum

Solanum pyraca

nthos

Solanum mahorie

nse

Solanum heinianum

Solanum tolia

raea

Solanum hieronym

i

Solan

um sisym

briifo

lium

Solan

um c

rinitu

m

Solan

um m

itlens

e

Solan

um ly

coca

rpum

Solan

um ro

stra

tum

Solan

um citr

ullifo

lium

Solan

um w

right

ii

Sol

anum

mam

mos

um

Sol

anum

pla

tens

eSol

anum

cap

sico

ides

Sol

anum

ace

rifol

ium

Sol

anum

tenu

ispi

num

Sol

anum

atro

purp

ureu

m

Sol

anum

via

rum

Sol

anum

acu

leat

issi

mum

Sol

anum

myr

iaca

nthu

m

Sol

anum

inca

rcer

atum

Sol

anum

pal

inac

anth

umSol

anum

stra

mon

iifol

ium

var

iner

me

Sol

anum

ses

silif

loru

m

Sol

anum

qui

toen

se

Sol

anum

felin

um

Sol

anum

lasi

ocar

pum

Sol

anum

upo

ro

Sol

anum

repa

ndum

Sol

anum

can

didu

m

Sol

anum

stra

mon

iifol

ium

Sol

anum

ves

tissi

mum

Sol

anum

hyp

orho

dium

Sol

anum

pse

udol

ulo

Solan

um h

irtum

Sol

anum

pec

tinat

um

Sol

anum

hirs

utissim

um

Sol

anum

tequ

ilens

e

0.02 substitutions / site

995 Solanaceae + five outgroup Convolvulaceae taxafrom a sparse eight-gene, 16405-site alignment

Figure S1: Phylogenetic relationships among 995 species of Solanaceae, rooted with five

species of Convolvulaceae.

8

Symonanthus bancroftiiSymonanthus aromaticus

Nicotiana obtusifoliaNicotiana bigelovii

Nicotiana clevelandii

Nicotiana tomentosa

Nicotiana tomentosiformis

Nicotiana setchellii

Nicotiana otophora

Nicotiana linearisNicotiana miersiiNicotiana attenuata

Nicotiana corymbosa

Nicotiana langsdorffii

Nicotiana bonariensis

Nicotiana forgetiana

Nicotiana longiflora

Nicotiana plumbaginifolia

Nicotiana alataNicotiana africana

Nicotiana fragrans

Nicotiana debneyi

Nicotiana simulans

Nicotiana megalosiphon

Nicotiana ingulba

Nicotiana umbratica

Nicotiana exigua

Nicotiana rosulata

Nicotiana goodspeedii

Nicotiana gossei

Nicotiana velutina

Nicotiana amplexicaulis

Nicotiana m

aritima

Nicotiana excelsior

Nicotiana bentham

iana

Nicotiana suaveolens

Nicotiana cavicolaNicotiana rotundifolia

Nicotiana occidentalis

Nicotiana hesperis

Nicotiana sylvestris

Nicotiana nudicaulis

Nicotiana repanda

Nicotiana stocktonii

Nicotiana acaulis

Nicotiana glaucaNicotiana noctiflora

Nicotiana petunioides

Nicotiana cordifolia

Nicotiana solanifoliaNicotiana paniculataNicotiana knightiana

Nicotiana rustica

Nicotiana benavidesii

Nicotiana glutinosaNicotiana undulataNicotiana wigandioides

Lycium tetrand

rumLyci

um a

reni

cola

Lyci

um h

orrid

umLy

cium

fero

ciss

imum

Lyci

um c

iner

eum esnedlevdnarts

muicyLLy

cium

gar

iep

ense

muli

mup

muic

yL

Lycium hirsutum

Lycium villosum

Grabow

skia duplicataLycium

pallidum

Lycium californicum

Lycium m

inimum

Lycium exsertum

Lycium frem

ontii

Lycium and

ersoniiLycium

parishii

Lycium b

erlandieri

Lycium cestroides

Nolana ivaniana

Nolana elegans

Nolana paradoxa

Nolana acum

inataN

olana rupicola

Nolana rostrata

Nolana plicata

Nolana aticoana

Nolana hum

ifusaN

olana laxaN

olana adansonii

Atropa belladonna

Anisodus tanguticus

Hyoscyam

us niger

Hyoscyam

us muticus

Hyoscyam

us albus

Jaborosa integrifolia

Sol

and

ra g

rand

iflor

aD

ysso

chro

ma

virid

iflor

a

Man

dra

gora

off

icin

arum

Bru

gman

sia

aure

aB

rugm

ansi

a sa

ngui

nea

Dat

ura

leic

hhar

dtii

Dat

ura

met

el

Dat

ura

inox

iaD

atur

a st

ram

oniu

m

With

erin

gia

cune

ata

With

erin

gia

cocc

olob

oide

s

Cua

tres

ia ri

paria

Cua

tres

ia e

xigu

iflor

a

Ioch

rom

a cy

aneu

m

Acn

istu

s ar

bore

scen

s

Ioch

rom

a ge

sner

ioid

es

Ioch

rom

a fu

chsi

oide

s

Erio

lary

nx lo

rent

zii

Ioch

rom

a au

stra

le

Dun

alia

bra

chya

cant

ha

Vass

obia

bre

viflo

ra

Dun

alia

sol

anac

ea

With

ania

som

nife

ra

With

ania

coa

gula

ns

Brac

hist

us s

tram

onifo

lius

With

erin

gia

sola

nace

a

With

erin

gia

mei

anth

a

With

erin

gia

mac

rant

ha

With

erin

gia

mex

ican

a

Physa

lis p

ruino

sa

Physa

lis an

gulat

a

Physa

lis p

ubes

cens

Physa

lis p

hilad

elphic

a

Physa

lis vi

scosa

Physa

lis he

terophy

lla

Physa

lis lo

ngifo

lia

Physa

lis ci

nera

scen

s

Physa

lis la

nceif

olia

Phys

alis

cras

sifol

ia

Jalto

mat

a gr

andi

flora

Jalto

mat

a pr

ocum

bens

Jalto

mat

a au

ricul

ata

Jalto

mat

a si

nuos

a

Salpich

roa o

rigan

ifolia

Capsicum la

nceolatum

Capsicum rh

omboideum

Capsicum parvi

folium

Capsicum fle

xuosum

Capsicum sc

hottianum

Capsicum vil

losumCapsicum annuum

Capsicum tovarii

Capsicum pubescens

Capsicum galapagoenseCapsicum baccatum

Capsicum chacoense

Capsicum chinense

Capsicum frutescens

Capsicum cardenasii

Capsicum exim

ium

Lycia

nthes ra

ntonnei

Lycian

thes cilio

lata

Lycian

thes asa

rifolia

Jaltomata hunzikeri

Jaltomata dentata

Solanum trizygumSolanum fraxinifolium

Solanum taeniotrichumSolanum muricatum

Solanum appendiculatum

Solanum palustre

Solanum etuberosum

Solanum morelliforme

Solanum clarum

Solanum polyadenium

Solanum pinnatisectum

Solanum bulbocastanum

Solanum ehrenbergii

Solanum cardiophyllumSolanum jamesii

Solanum stenophyllidium

Solanum chacoense

Solanum leptophyes

Solanum oplocense

Solanum tarijense

Solanum infundibuliform

e

Solanum berthaultii

Solanum verrucosum

Solanum sparsipilum

Solanum m

icrodontum

Solanum vernei

Solanum kurtzianum

Solanum bukasovii

Solanum tuberosum

Solanum santolallae

Solanum laxissim

um

Solanum oxycarpum

Solanum agrim

onifolium

Solanum colom

bianum

Solanum longiconicum

Solanum violaceim

armoratum

Solanum dem

issum

Solanum acauleSolanum iopetalum

Solanum andreanum

Solanum moscopanum

Solanum hjertingii

Solanum stoloniferum

Solanum guerreroense

Solanum hougasii

Solanum spegazzinii

Solanum commersonii

Solanum chomatophilum

Solanum piurae

Solanum lycopersicoidesSolanum sitiens

Solanum habrochaites

Solanum pennellii

Solanum chmielewskiiSolanum neorickii

Solanum lycopersicum var cerasiforme

Solanum cheesmaniae

Solanum galapagenseSolanum pimpinellifolium

Solanum chilenseSolanum peruvianum

Solanum ochranthumSolanum juglandifolium

Solanum crispum

Solanum americanum

Solanum ptychanthumSolanum opacum

Solanum physalifoliumSolanum palitans

Solanum tripartitum

Solanum dulcamaraSolanum wallacei

Solanum seaforthianum

Solanum triflorumSolanum villosumSolanum retroflexum

Solanum nigrum

Solanum terminaleSolanum symoniiSolanum aviculare

Solanum laciniatum

Solanum trisectum

Solanum

allophyllum

Solanum

melissarum

Solanum

tobagense

Solanum

diversifolium

Solanum

oxyphyllum

mutanicric munalo

S Sol

anum

occ

ultu

mS

olan

um s

ibun

doy

ense

Sol

anum

caj

anum

ense

xall

af mu

nalo

S

Solanum

obliquum

Solanum

roseumS

olanum m

aternumS

olanum b

etaceumS

olanum unilobum

Solanum

glaucophyllumS

olanum stuckertii

Solanum

confusumS

olanum hibernum

Solanum

luteoalbum

Solanum

amotapense

Solanum

pendulum

Solanum

corymbiflorum

Solanum

latiflorum

Solanum

exiguum

Solanum

cacosmum

Solanum

diploconosS

olanum pinetorum

Solanum

sciadostylis

Sol

anum

ap

hyod

end

ron

Sol

anum

arg

entin

um

Sol

anum

pse

udoc

apsi

cum

Sol

anum

ad

scen

den

sS

olan

um m

aurit

ianu

m

Sol

anum

cor

dov

ense

Sol

anum

abu

tiloi

des

Sol

anum

wen

dlan

dii

Sol

anum

pol

ygam

um

Sola

num

robu

stum

Sola

num

sta

gnal

eSola

num

car

oline

nse

Solanu

m ja

maic

ense

Solanu

m cr

oton

oides

Solanum

pan

iculat

um

Solanu

m to

rvum

Solan

um d

rym

ophil

um

Solan

um c

ampe

chien

se

Solanum hisp

idum

Solanum

erian

thum

Solanum tri

dynam

um

Solanum hindsia

num

Solanum asymmetriphyllum

Solanum nemophilum

Solanum cunninghamii

Solanum carduiformeSolanum leopoldense

Solanum cataphractumSolanum petraeumSolanum vansittartenseSolanum tudununggae

Solanum dioicumSolanum cleistogamumSolanum echinatum

Solanum melanospermum

Solanum clarkiae

Solanum beaugleholei

Solanum chippendaleiSolanum diversiflorum

Solanum campanulatum

Solanum prinophyllum

Solanum sandwicense

Solanum incompletum

Solanum cinereum

Solanum giganteum

Solanum aculeastrum

Solanum marginatum

Solanum incanum

Solanum melongena

Solanum linnaeanum

Solanum macrocarpon

Solanum dasyphyllum

Solanum vespertilio

Solanum lidiiSolanum tomentosum

Solanum cyaneopurpureum

Solanum anguivi

Solanum aethiopicum

Solanum virginianum

Solanum pyra

canthos

Sola

num

sisy

mbr

iifoliu

m

Solan

um c

rinitu

m

Solan

um ly

coca

rpum

Solan

um ro

stra

tum

Sola

num

citr

ullifo

lium

Sol

anum

mam

mos

umS

olan

um p

late

nse

Sol

anum

cap

sico

ides

Sol

anum

tenu

ispi

num

Sol

anum

atr

opur

pure

um

Sol

anum

via

rum

Sol

anum

acu

leat

issi

mum

Sol

anum

myr

iaca

nthu

m

Sol

anum

pal

inac

anth

um

Sol

anum

ses

silif

loru

m

Sol

anum

qui

toen

se

Sol

anum

felin

umSo

lanu

m la

sioc

arpu

m

Sola

num

repa

ndum

Sol

anum

can

didu

mSo

lanu

m s

tram

oniif

oliu

m

Sola

num

ves

tissi

mum

Sola

num

hyp

orho

dium

Sola

num

pse

udol

ulo

Sola

num

hirt

um

Sola

num

pec

tinat

um

10 million years

356 “x=12” Solanaceae taxa with known breeding system status

Figure S2: Phylogenetic relationships among 356 species of Solanaceae used in the analyses

of character evolution. Branch lengths were obtained with penalized likelihood smoothing,

implemented in r8s (S7), with the crown group age fixed at 36 Ma.

9

mean std dev std errλI 2.933 0.467 0.049λC 5.199 0.738 0.063µI 2.319 0.556 0.059µC 5.433 0.757 0.065qIC 0.555 0.138 0.010λC − λI 2.266 0.899 0.083µC − µI 3.114 1.011 0.095rI 0.613 0.141 0.010rC −0.234 0.075 0.002rI − rC 0.847 0.200 0.012rI − rC − qIC 0.292 0.075 0.002pSI 0.345 0.043 0.002

Table S1: Summary of MCMC results from the ML tree, showing the mean, standard deviation,

and standard error from time-series analysis. The first five quantities are the parameters directly

estimated by the BiSSE analysis.

λI λC µI µC qICλI —λC −0.067 —µI 0.977 −0.142 —µC −0.076 0.995 −0.165 —qIC −0.510 0.352 −0.677 0.414 —

Table S2: Parameter correlations on the ML tree. Correlations between qIC and net diversifica-

tion are 0.984 and −0.719 for rI and rC , respectively.

10

λI λC µI µC qIC rI rCML 0.467 0.738 0.556 0.757 0.138 0.141 0.075BS 0.548 0.812 0.547 0.869 0.124 0.059 0.033

Table S3: Standard deviations of estimated parameters, computed from MCMC on the ML tree

(upper row) and from the median values on the set of bootstrap trees (lower row).

mean std dev std errλI 3.031 0.486 0.013λC 5.108 0.744 0.018µI 2.516 0.563 0.015µC 5.284 0.769 0.018qIC 0.484 0.128 0.002qCI 0.021 0.011 0.000rI − rC 0.690 0.187 0.003pSI 0.355 0.044 0.000

Table S4: Summary of MCMC results from the ML tree when regain of SI is not prohibited.

11

Supporting references

S1. S. A. Smith, C. W. Dunn, Bioinformatics 24, 715 (2008).

S2. S. A. Smith, J. M. Beaulieu, M. J. Donoghue, BMC Evol Biol 9, 37 (2009).

S3. P. J. A. Cock, et al., Bioinformatics 25, 1422 (2009).

S4. http://www.biosql.org.

S5. K. Katoh, K.-i. Kuma, H. Toh, T. Miyata, Nucleic Acids Res 33, 511 (2005).

S6. A. Stamatakis, Bioinformatics 22, 2688 (2006).

S7. M. J. Sanderson, Mol Biol Evol 19, 101 (2002).

S8. T. Paape, et al., Mol Biol Evol 25, 655 (2008).

S9. M. J. Benton, The fossil record 2 (Chapman and Hall (London), 1993).

S10. S. Magallon, M. J. Sanderson, Evolution 55, 1762 (2001).

S11. Y. Wang, et al., Genetics 180, 391 (2008).

S12. C. Mitter, B. Farrell, W. B. J, Am Nat 132, 107 (1988).

S13. M. J. Sanderson, M. J. Donoghue, Trends Ecol Evol 11 (1996).

S14. R. E. Ricklefs, Ecology 87, 2468 (2006).

S15. B. Igic, R. Lande, J. R. Kohn, Int J Plant Sci 169, 93 (2008).

S16. J. C. Heilbuth, Am Nat 156, 221 (2000).

S17. W. P. Maddison, P. E. Midford, S. P. Otto, Syst Biol 56, 701 (2007).

12

S18. R. G. FitzJohn, W. P. Maddison, S. P. Otto, Syst Biol 58, 595 (2009).

S19. C. Goodwillie, S. Kalisz, C. G. Eckert, Ann Rev Ecol Evol Syst 36, 47 (2005).

S20. B. Igic, L. Bohs, J. R. Kohn, Proc Natl Acad Sci USA 103, 1359 (2006).

S21. R Development Core Team, R: A Language and Environment for Statistical Computing,

R Foundation for Statistical Computing, Vienna, Austria (2009). ISBN 3-900051-07-0.

S22. R. G. FitzJohn, diversitree: diversitree: comparative phylogenetic tests of diversification

(2010). R package version 0.4-5.

S23. M. Plummer, N. Best, K. Cowles, K. Vines, coda: Output analysis and diagnostics for

MCMC (2010). R package version 0.13-5.

13