Evidence for a Close Relationship Between Iostephane and Viguiera (: ) Author(s): Edward E. Schilling and Jose L. Panero Source: American Journal of Botany, Vol. 78, No. 8 (Aug., 1991), pp. 1054-1062 Published by: Botanical Society of America Stable URL: http://www.jstor.org/stable/2444894 Accessed: 09/11/2009 13:53

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=botsam.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Botanical Society of America is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Botany.

http://www.jstor.org AmericanJournal of Botany 78(8): 1054-1062. 1991.

EVIDENCE FOR A CLOSE RELATIONSHIP BETWEEN IOSTEPHANE AND VIGUIERA (ASTERACEAE: HELIANTHEAE)1

EDWARD E. SCHILLING2 AND JOSE L. PANERO Departmentof Botany, University of Tennessee, Knoxville, Tennessee 37996-1100; and Grinnell College, Grinnell,Iowa 50112-0806

The phylogeneticrelationship of Iostephaneisassessed using data from morphology, flavonoid chemistry, and chloroplastDNA and nuclear ribosomal DNA restrictionfragment analysis. Morphologicalevidence supportsplacement of Iostephanein subtribeHelianthinae, but fails to clarifythe placementof the genuswithin this assemblage.Further evidence for the placement of Iostephanein subtribeHelianthinae is provided by the presencein all species of the genus of floralflavonoids of the chalcone/auronetype, whichprovides a distinctivetrait for the subtribe within the tribe Heliantheae.Analysis of chloroplastDNA from two species of Iostephane,L heterophyllaand I. madrensis,in comparisonto Viguieraand relatedgenera indicates that the restrictionsite patternswith 16 enzymes for the Iostephanespecies are virtually identical to one another as well as to those of Viguierasect. Maculatae. Data from restrictionfragment patternsof nuclearrDNA are concordantwith the resultsfrom chloroplastDNA in suggesting a direct relationship between the two groups. The close phylogenetic relationship between Iostephaneand Viguierasect. Maculatae suggestedby the DNA restrictionfragment data was not suggestedby any other set of data.

IostephaneBenth. is a small genus of about Viguieraand its related genera, including He- four species endemic to Mexico. The genus lianthusL., HeliomerisNutt., PappobolusS. F. forms a readily recognizableassemblage based Blake, Pers., and Desf. ex on its subscapiform, heavy-rooted perennial Gmelin. In addition to morphological char- herbhabit and preferencefor pine, oak, or pine- acters, Viguieraand related genera have been oak forests between 1,500 and 3,000 m (Stro- shown to be characterizedby the shared pres- ther, 1983). Because there is variability be- ence of floral chalcone/auronepigments (Har- tween species for the presence or absence of a borne and Smith, 1978; Schilling, 1983; Rie- pappus, a characteroften consideredto be im- seberg and Schilling, 1985; Schilling and portant at the generic level in Asteraceae, the Panero, 1988; Schilling, Panero, and Bohm, genus has only recently been circumscribedin 1988; Schilling and Spooner, 1988), a type of its current constitution (Strother, 1983). The flavonoid that is otherwiseuncommon in tribe relationshipsof the genus have been suggested Heliantheae except in subtribe Coreopsidinae to lie with subtribe Helianthinae (Robinson, (Crawfordand Stuessy, 1981). Data from chlo- 1981; Strother, 1983) based on reproductive roplast DNA (cpDNA) restriction fragment characters,including the presenceof pales, her- variation provides furtherevidence for the re- maphroditic disk flowers, striate and carbon- latedness of this group, and suggests that Vi- ized achenes, sterile rays, a single stigmatic guiera as currentlyconstituted forms a para- surface,and style ducts outside the veins in the phyleticassemblage relative to the othergenera shaft (Robinson, 1981). (Schilling and Jansen, 1989). Robinson (1981) suggests that the subtribe The currentstudy is an attemptto determine Helianthinaeis one of the most concisely char- the phylogenetic relationships of Iostephane. acterized subtribes of the tribe Heliantheae. Initially, a single sample of the genus was an- The bulk of subtribe Helianthinae is formed alyzed as a possible outgroupto polarize chlo- by a group of species currently classified as roplast DNA restriction site changes within Viguieraand related genera. Additional sam- ' Received for publication 24 October 1990; revision ples were then analyzedto verify the somewhat accepted22 March 1991. surprisingresult that its chloroplast DNA re- The authorsthank John Strother,Robert K. Jansen,and striction site pattern is very similar to that of Mary Lou Schmid for assistance with various phases of Viguierasect. Maculatae (S. F. Blake) Panero the project;the StrybingArboretum for providinga sample & E. Schilling,a groupthat is pheneticallyvery of Iostephaneheterophylla; and Loren H. Rieseberg for rDNA restriction site map data from . This different. Additional data from morphology, work was supportedby NSF grantBSR 88-06513 to EES. chemistry,and restrictionfragment analysis of 2Author for correspondence. a second molecule, nuclear ribosomal DNA 1054 August 1991] SCHILLING AND PANERO-IOSTEPHANE/VIGUIERA RELATIONSHIP 1055

(rDNA), were then sought to provide further TABLE 1. Charactersscored for cladisticanalysis of Vi- perspective on these results. guiera, lostephane, Helianthus,Tithonia and Simsia

1. Habit (Oa = woody; 1 = herbaceous) MATERIALSAND METHODS 2. Habit (Oa = not scapose; 1 = scapose) 3. Chromosome number (Oa = 18; 1 = 17; 9 = 8) The following samples of Iostephane were 4. lobed (Oa = absent; 1 = present) analyzed for DNA restriction site variation, 5. vein course (0 = straight/sinuous;1 = zigzag) using living grown at the University of 6. Leaf vein areolation (0 = well developed; 1 = im- Tennessee greenhouses (all originating from perfect) 7. Synflorescencestructure (0 = dichasial; 1 = mono- Mexico; vouchers at TENN): L heterophylla: chasial) Durango, Panero and Schilling 1549 (I1); 8. Head shape(0 = hemispherical;1 = campanulate/cy- Breedlove44199 (12;grown from seed provid- lindrical) ed by Strybing Arboretum);Guerrero, Schil- 9. Phyllaryshape (Oa = narrowedgradually; 1 = narrowed ling and Panero 86-A (13); Jalisco, Schilling abruptly) and Panero 88-53 (14);and L madrensis:Gua- 10. Pale apex shape (0 = entire; 1 = trifid) najuato, Schilling and Panero 88-22 (IM). 11. Ray corollaUV nectarguides (0 = absent;1 = present) 12. Disk corolla shape (Oa = straight;1 = bulbous at base Analysis of morphology and floral flavonoids of throat) utilized these samples, as well as the following, 13. Disk corolla lobe color (Oa = yellow; 1 = black) using material collected from natural popula- 14. Disk corolla lobe sclerified cells (0 = absent; 1 = tions: L papposa, Oaxaca, Panero et al. 617; present) L trilobata,Oaxaca, Schilling and Panero 88- 15. Disk corolla elongate moniliform glands (0 = rare; I = present) 34. 16. Disk corolla tube glands (0 = present; 1 = absent) For morphological studies, the 33 samples 17. Stamenconnective pubescence (0 = none; 1 = present) of Viguiera,Helianthus, Heliomeris, Simsia, 18. Crystalsin staminalfilaments (0 = few; 1 = abundant) and Tithonia (species names in Fig. 1; sources 19. Anther color (0 = black; 1 = orange) in Table 1 of Schilling and Jansen, 1989) an- 20. Anthercollar cells (0 = thickened; 1 = heavily thick- alyzed previously for cpDNA variation were ened) 21. Antherendothecial cells (0 = polarized;1 = partially consideredto be exemplarsof their respective polarized) taxa, and were analyzedtogether with samples 22. Style appendage(0 = minute; 1 = well developed) of Iostephanespecies. A total of 27 characters 23. Style branchshape (0 = deltoid; 1 = tapering) (Table 1) was scored for these samples, in- 24. Ray ovary (0 = short; 1 = long, narrow) cluding ones used traditionallyfor taxonomic 25. Disk achene pappus (0 = persistent; 1 = deciduous; delimitation as well as novel charactersfrom 9 = absent) leaf venation and floralmicromorphology (the 26. Disk achene pappus (0 = present; 1 = absent) 27. Disk (Oa = I = latter discussed fully in Panero, in press). De- achene type biconvex; very flattened) letion of sampleswith identical scoresfor these a Plesiomorphic character state, polarized relative to charactersresulted in a data set of 25 samples Flourensia. used for cladistic analysis (deleted samples in- cluded Helianthus giganteus, two samples of nol: acetic acid: water (3:1:1) and 15%acetic Viguieradentata, V. deltoidea, V. incana, V. acid. Flavonoid spots, visualized with uv light, puruana, V. tomentosa, and V. triangularis). were cut from chromatogramsand extracted The genus Flourensia was utilized as an out- in MeOH, followed by Sephadex LH-20 col- group to polarize characterchanges. Wagner umn chromatography.Compound identifica- parsimonywas performedutilizing the PAUP tion was based on color under uv light, uv program (version 2.4, D. Swofford, Illinois spectraldata, and cochromatographywith au- Natural History Survey) implemented,on an thentic compounds identified from previous IBM-PC.Most parsimonioustrees were sought studies of Helianthus and Viguiera. Localiza- by implementing the multiple parsimony tion of compounds of the chalcone and aurone (MULPARS) with global branch swapping types was done by visual inspection of dried, (SWAP = GLOBAL)options. A strict consen- intact floweringheads that were flooded with sus tree was generated using the CONTREE ammonia vapor, which produces a red color program. reaction with these compounds. Methodsfor flavonoid analysisgenerally fol- Methods for DNA restrictionfragment anal- lowed Mabry, Markham,and Thomas (1970) ysis generally followed Jansen and Palmer and Markham (1982). Dried flowering heads (1988). Preparationsof total DNA were made (ca. 1 g) were ground and extracted overnight from 1.5 g of fresh leaves by the procedureof in methanol. The extract was filtered,concen- Doyle and Doyle (1987). Restriction endonu- trated, and chromatographedon Whatmann clease digestions (utilizing the following en- 3MM paper two-dimensionally using t-buta- zymes:AvaI, BamHI, BanI, BanlI, BclI, BglII, 1056 AMERICAN JOURNAL OF BOTANY [Vol. 78

Flourensia (Outgroup) 10 12 25 11 18 -o-x-o Helianthus angustifoliu 20 21 22 23 26 -o-x - 16 26 - o| -0-o - x-o 6 16 Tithonia rotundifolia

4 26 lostephane heterophylla 12 18 2 13 15 13 - x - o x lostephane trilobata

lostephane madrensis 9 11 6 -o-- o Viguiera dentata x - 7 8 4 5 19 20 27 -x Simsia calva

8 14 17 Simsia foetida 13 1 23 24 - Viguiera sessilifolia -o0 --oo- o 13 26 -o-0 o Viguiera ovata 8 10 26 0-0-0o Viguiera potosina 16 7 18 -o Viguiera szyszylowiczii -o--o - 20 -o Viguiera flava 16 13 Pappobolus argenteus

25 Pappobolus nigrescens 19 24 3 6 - o-0o Pappobolus imbaburensis

Viguiera grammatoglossa 1 26 o Heliomeris multiflora -o 4 9 o-0 Viguiera pinnatilobata 7 8 0 Viguiera adenophylla 15 L-0 -Viguiera eriophora 19 Viguiera insignis

5 9 11 g Viguiera parishii -0o-0o-0o 16 ? Viguiera carterae

Fig. 1. Strictconsensus tree generatedfrom parsimonyanalysis using morphologicaldata, showingrelationships of Iostephaneand relatedgenera. Characters numbered as in Table 1. 1, apomorphy;0, parallelism;x, reversal.

BstNI, BstXI, ClaI, EcoRI, EcoRV, HaeII, for filter hybridizationexperiments. Mapping HincII, HindIII, NcoI, NsiI), agarose gel elec- offragmentswas done relativeto dataofJansen trophoresis,and bidirectionaltransfer of DNA and Palmer (1987, 1988) and Schilling and fragmentsfrom agarosegels to Zetabind(AMF Jansen (1989). Because the cpDNA restriction CUNO) nylon filters were performed as de- fragment site pattern for Iostephane samples scribedin Palmer(1 9 8 6) and Jansenand Palm- was almost identical to that of samples of Vi- er (1987). Preparation of digoxigen-labeled guiera sect. Maculatae, it was possible to place probes and filter hybridizations followed the these unequivocally on the cladogram from manufacturer's instruction ("Genius Kit," restriction-site mutation data generated by Boehringer Mannheim Biochemicals, India- Schilling and Jansen (1989). Ribosomal DNA napolis). The 22 cloned restriction fragments variation was detected by hybridization to of lettuce cpDNA were combined into batches plasmids containing a single 18S-28S rDNA August 199 1] SCHILLING AND PANERO-IOSTEPHANE/VIGUIERA RELATIONSHIP 1057

TABLE 2. Chemistryand distributionof lostephaneflo- TABLE 3. ChloroplastDNA restrictionsite mutationsdif- ralflavonoidsa ferentiating samples of lostephane and Viguiera sect. Maculataefrom other membersof Viguiera and re- Quer- Quer- latedgenera, numberedfollowingSchillingand Jansen cetin cetin Core- 3-gluco- 3-glucu- opsin/ (1989) Species side ronide Sulfureinb Ligules Disk Probe' Enzyme Mutation I. heterophylla + + + - + I. madrensis + + + - + 13. 18.8 Kb HincII 4.6 = 4.3 [+0.3] L papposa + + + + + 19. 14.7 Kb BcJI 3.8 + 2.7 = 6.5 L trilobata + + + + + 26. 14.7 Kb EcoRV 1.4 + 2.1 = 3.5

a 47. 10.6 Kb - Deletion +, compound present;-, compound not detectable. 61. 10.6 Kb NsiI 0.5 + 0.5 = 1.0 bCoreopsin spontaneouslyisomerizes to form sulfurein, 72. 7.7 Kb HincII 7.2 = 4.6 + 2.6 so the two compounds occur togetherin extracts. 75. 7.0 Kb BamHI 3.7 + 3.7 = 7.4 105. 6.7 Kb NsiI 16.7 = 12.2 + 4.5 repeatunit from Helianthus argophyllus(orig- 109. 6.3 Kb BanIl 1.2 + 2.5 = 3.7 inally isolated by Mike Arnold), and mapped 113. 6.3 Kb EcoRI 0.9 [+0.3] = 1.2 relative to data from 128. 4.6 Kb EcoRI 1.8 [+0.1] = 1.9 Helianthus of Rieseberg 144. 9.0 Kbb BcJI 4.0 + 2.6 = 6.6 (unpublisheddata). 148.c 3.8 Kb AvaI 3.2 = 2.2 + 1.0 a From lettuce SacI cpDNA clone bank (Jansen and RESULTS Palmer, 1987), listed by fragment size. b Fragment from petunia 15.4 Kb region not cloned from Morphologicalanalysis -Parsimony analy- lettuce. sis of morphological data using PAUP pro- c Mutation newly reported in this paper. duced a total of 54 equally parsimoniousmin- imum length trees. A strict consensus tree produced from these is shown in Fig. 1. This that most of its internalnodes are unresolved; tree has a length of 67 steps and a consistency only Simsia and Iostephane are retained as index of 0.40. An estimate of the homoplasy defined groups (not shown). excess ratio (HER) for this tree was calculated using equations(4) and (5) from Archie (1990), Floral flavonoid chemistry -Samples of all and this produceda value of HER = 0.43. This species of Iostephane produced identical pro- value suggests that there is a high proportion files for floral nonanthocyanic flavonoids, of phylogenetically random variation in the which included two quercetin glycosides and data set. In agreementwith cpDNA data, the the isomeric chalcone/auronepair coreopsin/ woody habit appearsto be plesiomorphicwith- sulfuerin (Table 2). This is notable given the in this group, with the basal clades including differencesbetween species in ray flower col- primarilywoody taxa. One of these, V. subg. oration(L heterophyllahas purplerays whereas Bahiopsis, is placed as the sister group to the the otherspecies have yellow ones). Therewere, remainingtaxa. Although this is also in agree- however, significantdifferences in localization ment with cpDNA data, it should be noted of the chalcone/auronepigments. These com- that this position is influenced by the choice pounds were produced throughout the ligule of an outgroupwith a chromosome base num- of I. trilobata and I. papposa, whereas they ber of x = 18. Another basal clade is formed were restricted to the disk corolla lobes in I. by the samples of V. sect. Maculatae. Iosteph- madrensisand I. heterophylla(Table 2). Thus ane is placed with other herbaceoustaxa in a it is expected that the coloration of the ligule large terminal clade as the sister group to He- detected by pollinators sensitive to uv light is lianthusand Tithonia.These taxa arenot, how- differentfor I. madrensiscompared to the other ever, well resolved, and most of the defining two species with yellow ligules. charactersare parallelisms or reversals. This situation can be illustratedby considerationof Chloroplast DNA analysis -All of the Io- the effectof addingan additionalcharacter (the stephane samples shared the same ten restric- presenceor absence of foliar subsessile glands) tion site mutations (Table 3) that distinguish that was removed because of its high level of Viguierasect. Maculatae relative to V. subg. homoplasy, and because chemical studies of Bahiopsis (the "Baja CaliforniaGroup"), and individual taxa have suggestedthat the loss of lacked any additional site mutations charac- such glandshas occurredrepeatedly in different teristic of other taxa of Viguieraand related groups (Schilling, 1983; Schilling and Panero, genera. No apomorphic mutations were de- 1988; Schilling,Panero, and Bohm, 1988). Af- tected that would distinguish Iostephanefrom ter addition of this character,the herbaceous V. sect. Maculatae, although some samples of clade in the consensus tree becomes altered so both groups have individual autapomorphies. 1058 AMERICAN JOURNAL OF BOTANY [Vol. 78

Viguiera Viguiera Helianthus, Pappobolus, subgenus section Simsia, Tithonia and

Bahiopsis Maculatae Iostephane various Viguiera

19+ 6 samples HM VD VE VI VP VA I3 I1 I2 I4 IM samples

t105 -109 R [4 apomorphies [21 apomorphies] 144 at next node]

[14 apomorphies] 75

-26 [5 apomorphies] 72

Fig. 2. Single most parsimoniousWagner tree of Viguieraand related species based on cpDNA restrictionsite mutation data, from Schillingand Jansen (1989), showingplacement of samples of Iostephane(numbered as in text). Mutations(numbered as in Table 3) or numbersof mutations shown along branches.HM, Heliomeris mutiflora;VA, V. adenophylla;VD, V. dentata;VE, V. eriophora;VI, V. insignis;VP, V. puruana.

Hence, the lostephane samples can be placed bands in the Iostephane samples that would (Fig. 2) to form an unresolved polychotomy, provide possible evidence of polymorphism together with V. adenophylla and the other for rDNA molecules. samples of V. sect. Maculatae, in the most parsimoniousWagner tree producedby Schil- DISCUSSION ling and Jansen (1989). No additional homo- plasy is introduced into this tree by inclusion Comparisonof morphologicaland DNA re- of the samples of Iostephane. striction site data suggests that Iostephane is phyletically close to a taxon, Viguiera sect. Nuclear rDNA analysis-Variation in re- Maculatae, from which it is pheneticallyquite strictionsites readilyinterpreted relative to data distinct. Like many genera of Asteraceae, Io- from Helianthus(Rieseberg, unpublished data) stephane is recognizedby the presence of dis- was available for the following enzymes: BclI, creteautapomorphies. There are, however, rel- BglII, EcoRI, EcoRV, HindlIl, NsiI. Other atively few characters that are potential enzymes either producedno variabilityamong synapomorphiesto representits relationships samples or produced band patterns that were to other taxa. This problem is exacerbatedin not easily interpretedwithout further experi- Asteraceae because of the high degree of ho- mentation. In addition to restrictionsite vari- moplasy exhibited by other characters. ation, there were also differencesin the length Cladistic analysis of morphology, including of the basic repeatunit. The postulated rDNA both traditional taxonomic charactersas well restrictionsite mutations are shown relative to as novel ones from leaf venation and floral the most parsimonious cpDNA tree (Fig. 3). micromorphology,fails to clarifythe relation- Except for some size variation, the restriction ships of Iostephane (Fig. 1). In the consensus fragmentpatterns of the samples of Iostephane tree, Iostephaneis placed in a clade formed by were the same as those of samples of V. sect. the herbaceous members of Viguieraand re- Maculatae (Fig. 3). There were no additional lated genera. Although Iostephane is further August 1991] SCHILLING AND PANERO-IOSTEPHANE/VIGUIERA RELATIONSHIP 1059

Viguiera Viguiera

subgenus section

Bahiopsis Maculatae Iostephane Helianthus

6 samples HM VD VE VI VP VA I3 I1 I2 I4 IM 20+ samples

-[+ NsiI] + IIII -[- Bcl I] EcoRII -[+ [+ EcoRI]

[+ NsiI[- EcoRI]

[+ EcoRV] [- HindIII] [+ NsiI] [+ NsiI] II

Fig. 3. Postulated restrictionsite mutations in nuclear ribosomal DNA, showing site gains (+) and losses (-), plotted on most parsimoniousWagner tree based on chloroplastDNA data (Fig. 2). Dashed lines show parts of tree that are unsupportedby nuclearribosomal DNA information.Abbreviations as in Fig. 2. placed as a sister group to Helianthus and Ti- which is clearly distinguished by autapomor- thonia, all of the charactersthat support this phies in growth habit but lacks any of the dis- placement exhibit a high degree of homoplasy tinctive apomorphies that characterizeother in the analysis. The overall high level of ho- named taxa of this assemblage. moplasy and lack of resolution provided by Variabilityand homoplasy in morphological the cladistic analysis, however, suggest that featurescomplicate recognitionof many of the morphologicaldata provide little information named taxa of Viguieraand relatedgenera. For regardingthe relationshipsof taxa of Viguiera instance, the taxa of Pappobolusdo not form and relatedgenera. The estimates of the excess a monophyletic group in the cladistic analysis homoplasy ratio (Archie, 1990) indicate that of morphology (Fig. 1), even though consid- the data set includes significant amounts of erations of biogeography and overall mor- phylogeneticallyrandom variation. This is no- phology (Panero, in press) suggest that Pap- table in that exemplars of taxa were used; in- pobolus is monophyletic. In several other clusion of all members of each taxon could be groups, including Helianthus, Simsia, Titho- expectedto raiselevels of homoplasy and lower nia, Viguiera sect. Maculatae, and V. subg. overall resolution even further. The lack of Bahiopsis,there is similar variationin defining resolution of relationships between taxa pro- features, although it is not reflected in the vides confirmationthat the currenttaxonomy cladogram because only a subset of taxa was of this group is essentially a phenetic one. In- analyzed. dividual taxa are recognized primarily on the The presence in Iostephane heads of com- basis of autapomorphies, and there are few pounds of the chalcone/auronetype (Table 2) morphological synapomorphies to indicate provides supportingevidence that Iostephane phylogenetic relationshipsbetween such taxa. shouldbe placedtaxonomically in subtribeHe- This is particularly notable for Iostephane, lianthinae, near Viguieraand related genera. 1060 AMERICAN JOURNAL OF BOTANY [Vol. 78

Compoundsof the chalcone/auronetype have cea, Ximenesia, Helianthella, Pionocarpus, a limited distribution in Asteraceae, but are Rudbeckia, Gymnolomia), but never in Vi- characteristicof Viguieraand related genera guiera (Strother, 1983). Among the morpho- (Crawfordand Stuessy, 1981). Floralflavonoid logical charactersutilized in the currentstudy chemistrydoes not, however, provide evidence (Table 1), only one, presenceof distinctive mo- to indicate more specificallythe relationships niliform trichomes on the corolla, links Io- of Iostephane within this assemblage. Al- stephanewith V. sect. Maculatae, and the cla- though Schilling and Spooner (1988) reported distic analysis of morphological data would the presence of quercetin 3-glucuronideto be suggestthis to be a parallelism (Fig. 1). characteristicof Simsia, this glycosidehas now Two alternative hypotheses that might be been found in various species of Viguieraas consideredto explain the similarityin cpDNA well (unpublisheddata). There were no differ- restrictionfragment patterns between Iosteph- ences in compound profilesbetween species of ane and V. sect. Maculatae are wide hybrid- Iostephane (Table 2), but the distribution of ization and lineage sorting, but neither appear compounds within the head provides another to be very likely. It might be hypothesizedthat characterto distinguishamong the yellow-rayed the habit of Iostephane originatedfrom a pa- species, with I. madrensislacking ligule com- ternal parentin an initial hybridizationevent, poundswhereas I. trilobataand I. papposapro- whereas the maternally inherited cpDNA ge- duce them. nome is retained from a member of V. sect. Data from cpDNA restriction fragment Maculatae. An initial problem with this ex- analysis also support the relationship of Io- planation is that the subscapose growth habit stephane to Viguieraand related genera, and of Iostephane is observed in only a few, rela- provide evidence of a direct relationship to tively specializedmembers of relatedtaxa, no- Viguierasect. Maculatae (Fig. 2). Iostephane tably members of Helianthus ser. Angustifolii and V. sect. Maculatae share four unique re- such as H. carnosus and H. radula (Heiser et striction mutations (threelosses and one gain) al., 1969), and that the few crosses that have relativeto all othertaxa of Viguieraand related been attemptedbetween Helianthusand either genera that have been examined to date (Fig. V. sect. Maculatae or Iostephane have failed 2; Table 3), generating a confidence level of (unpublisheddata). Further,data from rDNA 100%when assessed by the bootstrap method restriction site variation are completely con- (Felsenstein,1985; Schillingand Jansen, 1989). sistent with cpDNA data in indicatinga direct The two groups also share six further restric- relationship between Iostephane and V. sect. tion sites relative to V. subg. Bahiopsis, the Maculatae (Fig. 3); there is no evidence from apparentsister group to other members of Vi- this biparentally inherited molecule, such as guiera and related genera, and lack any of the polymorphismfor or recombinationof restric- additionalmutations that characterizevarious tion site patterns,to indicatethat hybridization other members of this assemblage (Schilling between widely differentparents has been in- and Jansen, 1989). volved in the origin of Iostephane. It might That there should be a direct phylogenetic also be proposed that lineage sorting from a relationship between Iostephane and V. sect. polymorphic ancestral pool of cpDNA ge- Maculatae is surprising,given the great differ- nomes has resulted in retention of the same ences in morphology that occur between the genomes in phyletically differentgroups. Ad- two groups.Members of Iostephaneare all sub- ditional survey (Schilling, unpublished data) scapose perennial herbs, whereas those of V. of cpDNA restriction site patterns in over 70 sect. Maculatae are all large shrubs or small other samples of Helianthus,Pappobolus, Sim- trees (Figs. 4, 5). Although both groups are sia, Tithonia,and Viguieraspecies, all of which endemic to Mexico, there are also differences wouldpresumably be descendedfrom this same in habitat preferencebetween them, with Io- ancestralgene pool, has failed to indicate any stephaneoccurring in pine and pine/oak forests evidence for the retention of such a cpDNA (Strother, 1983), while members of V. sect. genome elsewhere. Maculatae occur primarily in tropical decid- Thereare some notableparallels between the uous forests and extend only rarely into oak situation of Iostephane and Viguieraand that or oak/pineforests (Panero and Schilling,1988). of another group of Asteraceae,the Hawaiian Previous taxonomic treatments of Iostephane silversword alliance. Data from cpDNA re- have also failed to suggesta direct relationship strictionfragment analysis suggest that the most with any member of Viguiera(e.g., Strother, speciose genus of the silverswordalliance, Du- 1983). For example, taxa of Iostephane have bautia,may be paraphyleticrelative to the oth- been at times variously placed in a total of er two genera, Argyroxiphiumand Wilkesia eight other genera(Coreopsis, Simsia, Echina- (Baldwin, Kyhos, and Dvorak, 1990). In par- August 1991] SCHILLING AND PANERO-IOSTEPHANE/VIGUIERA RELATIONSHIP 1061

4 5

Figs. 4, 5. Habit of Iostephaneand Viguierasect. Maculatae.4. Iostephaneheterophylla in Jalisco, Mexico. Plant is ca. 6 dm tall. 5. Viguierapuruana (arrows) in Michoacan, Mexico. Human in picture is ca. 1.8 m tall. ticular, cpDNA data (Baldwin, Kyhos, and For example, it appears that Solanum is para- Dvorak, 1990) suggest that Wilkesia, which phyletic relative to Lycopersicon (tomatoes), occurs in dry scruband woodland habitatsand with the potatoes (S. sect. Petota) being the produces strikingcauline, yuccalikerosettes of immediate sister group to tomatoes (Spooner, monocotlike leaves, has as its immediate sister Anderson, and Jansen, 1990). Similarly, the group a lineage of Dubautia species that are monotypic Heterogaura is the sister species to rain forest trees and shrubs with typical dicot a member of Clarkia (Sytsma and Gottlieb, habit and leaf venation (Carr, 1985). Diver- 1986a, b). Soltis, Soltis, and Bothel (1990) have gence in the Hawaiian silversword alliance as suggested that in herbaceous Saxifragaceae most a whole has produceda variety ofgrowth forms monotypic, phenetically distinct taxa are likely that inhabit a range of habitats and exhibit a to be phyletic ingroups to larger, diverse gen- broad range of physiological diversity (Carr, era. There has not yet appeared any consensus 1985; Robichaux et al., 1990). Because the in how to resolve such conflicts between phe- montane topographyof Mexico may result in netic and phylogenetic groups in plants. Schil- the occurrence of high altitude habitat "is- ling and Jansen (1989) note that other genera lands," particularlyduring periods of extreme related to Viguiera, such as Helianthus, Pap- climate, it might be hypothesizedthat the same pobolus (= Helianthopsis), Tithonia,and Sim- types of forces that produced the radiation in sia, are recognized primarily by distinctive the Hawaiian silverswordalliance may also be morphological features (autapomorphies), responsible for the divergence of Iostephane leaving a paraphyletic assemblage not char- and V. sect. Maculatae. acterized by any apomorphies (but mostly The case of Iostephane provides another of sharing one or more sympleisomorphies) as a growinglist of examples wherethe phylogeny Viguiera. In the current situation, Iostephane inferrable from cpDNA data is not in agree- differs from Viguiera by autapomorphies in ment with the current classification scheme. features of the achene (shape and type of pap- 1062 AMERICAN JOURNAL OF BOTANY [Vol. 78

pus) and the involucre that have traditionally The systematic identification of flavonoids. Springer- been considered to be important generic level Verlag, New York. charactersin tribe Heliantheae. Similar types MARKHAM, K. R. 1982. Techniques of flavonoid iden- tification. Academic Press, London. of variation in pappus type have been noted PALMER, J. D. 1986. Isolation and structural analysis of in Pappobolus(Panero, in press), and there is chloroplast DNA. Methods in Enzymology 118: 167- considerablevariability in the involucrewithin 186. Viguiera.A classificationin agreementwith the PANERO, J. L. In press. Systematics of the Andean genus cpDNA phylogenywould be achieved by sub- Pappobolus (Asteraceae: Heliantheae). Systematic merging Iostephane within Viguiera, but it ap- Botany Monographs. , AND E. E. SCHILLING.1988. Revision of Viguiera pears advisable to await resolution of the sect. Maculatae (Asteraceae: Heliantheae). Systematic broader problems of generic delimitation in- Botany 13: 371-406. volving Viguiera and related genera before RIESEBERG,L. R., AND E. E. SCHILLING.1985. Floral making a formal taxonomic proposition. flavonoids and ultraviolet patterns in Viguiera (Com- positae). American Journal of Botany 72: 999-1004. ROBICHAUX,R. H., G. D. CARR,M. LIEBMAN,AND R. W. LITERATURE CITED PEARCY. 1990. Adaptive radiation of the Hawaiian silversword alliance (Compositae-Madiinae): ecolog- ical, morphological, and physiological diversity. An- ARCHIE, J. W. 1990. Homoplasy excess ratios: new in- nals of the Missouri Botanical Garden 77: 64-72. dices for measuring levels of homoplasy in phyloge- ROBINSON,H. 1981. A revision ofthe tribal and subtribal and a critique of the consistency netic systematics limits of the Heliantheae (Asteraceae). Smithsonian 38: 253-269. index. Systematic Zoology Contributions to Botany 51: 1-102. BALDWIN, KYHOS,AND B. G., D. W. J. DVORAK.1990. SCHILLING,E. E. 1983. Flavonoids of Helianthus series Chloroplast DNA evolution and adaptive radiation Angustifolii. Biochemical Systematics and Ecology 11: in the Hawaiian silversword alliance (Asteraceae-Ma- 341-344. diinae). Annals of the Missouri Botanical Garden 77: , AND R. K. JANSEN. 1989. Restriction fragment 96-109. analysis of chloroplast DNA and the systematics of CARR, G. D. 1985. A monograph of the Hawaiian Ma- Viguiera and related genera (Asteraceae: Heliantheae). diinae (Asteraceae): Argyroxyphium, Dubautia and American Journal of Botany 76: 1769-1778. Wilkesia. Allertonia 4: 1-123. , ANDJ. L. PANERO. 1988. Flavonoids of Viguiera CRAWFORD, D. J., AND T. F. STUESSY. 1981. The taxo- series Brevifoliae. Biochemical Systematics and Ecol- nomic significance of anthochlors in the subtribe Co- ogy 16: 417-418. reopsidinae (Compositae, Heliantheae). American ,J. L. PANERO,ANDB. A. BOHM. 1988. Flavonoids Journal of Botany 68: 107-117. of Viguiera section Maculatae. Biochemical System- DOYLE, J. J., AND J. L. DOYLE. 1987. A rapid DNA atics and Ecology 16: 413-416. isolation procedure for small quantities of fresh leaf ,AND D. M. SPOONER.1988. Floral flavonoids and tissue. Phytochemical Bulletin 19: 11-15. the systematics of Simsia. (Asteraceae: Heliantheae). FELSENSTEIN, J. 1985. Confidencelimits on phylogenies: Systematic Botany 13: 572-575. an approach using the bootstrap. Evolution 39: 783- SOLTIs,D. E., P. S. SOLTIS,AND K. D. BOTHEL. 1990. 791. Chloroplast DNA evidence for the origins of the HARBORNE, J. B., AND D. M. SMITH.. 1978. Anthochlors monotypic Bensoniella and Conimitella (Saxifraga- and other flavonoids as honey guides in the Com- ceae). Systematic Botany 15: 349-362. positae. Biochemical Systematics and Ecology 6: 287- SPOONER,D. M., G. J. ANDERSON,AND R. K. JANSEN. 291. 1990. Chloroplast DNA phylogeny of tomatoes, po- HEISER, C. B., JR., D. M. SMITH, S. B. CLEVENGER, AND tatoes and pepinos (Solanum subgenus Potatoe). W. C. MARTIN, JR. 1969. The North Americansun- American Journal of Botany 77: 156. flowers (Helianthus). Memoirs of the Torrey Botanical STROTHER,J. L. 1983. Pionocarpus becomes Iostephane Club 22: 1-218. (Compositae: Heliantheae): a synopsis. Madrono 30: JANSEN, R. K., AND J. D. PALMER. 1987. Chloroplast 34-38. DNA from lettuce and Barnadesia (Asteraceae): struc- SYTSMA,K. J., ANDL. D. GOTTLIEB.1986a. Chloroplast ture, gene localization, and characterization of a large DNA evidence for the derivation of the genus Het- inversion. Current Genetics 11: 553-564. erogaura from Clarkia (Onagraceae). Proceedings of , AINID . 1988. Phylogenetic implications of the National Academy of Science of the United States chloroplast DNA restriction site variation in the Mu- of America 83: 5554-5557. tisieae (Asteraceae). American Journal of Botany 75: , AND . 1986b. Chloroplast DNA evolution 753-766. and phylogenetic relationships in Clarkia sect. Peri- MABRY, T. J., K. R. MARKHAM, AND M. B. THOMAS. 1970. petasma (Onagraceae). Evolution 40: 1248-1262.