Boletín de la Sociedad Botánica de México 56: 77-88, 1995 DOI: 10.17129/botsci.1466 Bol. Soc. Bot. México 56: 77-88 (1995)
Systematics and character evolution of the genus Yucca L. (Agavaceae): Evidence from morphology and molecular analyses
KAREN H. CLARY AND BERYL B. SIMPSON. Department of Botany, The University ofTexas, Austin, Texas, 78713-7640
Abstract. Yucca L. (the Desert Soaptree, Agavaceae) consists of 45 perennial species distributed primarily in the deserts of Mexico and the United States. Although severa! taxonomic treatments of yuccas exist; the-phylogeny of the group is poorly known. It is unclear which taxa retain primitive characters and how characters have evolved. In addition, relationships oftaxa within the genus are uncertain. We compare our phylogeny based upon reproductive characters with a chloroplast phylogeny (Hanson and Rieseberg, 1991 ; Hanson, 1993) and with data from the ITS region of the nuclear genome (Bogler, this volume). The molecular analyses of the chloroplast genome by Hanson and Rieseberg (1991) led to a phylogeny on ly partially congruent with the traditional morphology-based phylogeny (McKelvey, 1938, 1947) because the cpDNA analysis indicated that chloroplast capture among distantly related, sympatric/parapatric species has been a factor in the evolution of the group. Our results indicate that Yucca in the traditional sen se is paraphyletic, that sect. Chaenocarpa is more basal than sect. Yucca ( =Sarcocarpa), and that Yucca (=Hesperoyucca) whipplei falls outside of Yucca. Key words: Agavaceae, evolution, molecular phylogeny, systematics, Yucca. Resumen. Yucca L. (Agavaceae) se compone de 45 especies perennes distribuidas principalmente en los desiertos de México y los Estados Unidos. Aunque existen varios tratamientos taxonómicos de Yucca, la filogenia del género es aún poco conocida. No está claro qué taxa retienen caracteres primitivos y cómo han evolucionado los caracteres. Además, las relaciones entre los taxa del genero son inciertas. En éste artículo, comparamos la filogenia basada en caracteres reproductivos con una filogenia del ADN del cloroplasto (cpADN) (Hanson y Rieseberg, 1991; Hanson, 1993) y con los datos de la region ITS del genoma nuclear (Bogler, este volumen). Los análisis moleculares del genoma del cloroplasto (Hanson y Rieseberg, 1991) produjeron una filogenia solo parcialmente congruente con la filogenia tradicional basada en morfología (McKelvey, 1938, 1947) porque la filogenia basada en cpADN indica que la captura del cloroplasto entre especies genéticamente distantes (simpátricas o parapátricas) ha influido en Ja evolución del grupo. Nuestros resultados indican que Yucca tal como se conoce tradicionalmente es parafilética, y que la seccción Chaenocarpa es más basal que la sección Yucca ( =Sarcocarpa), y que Yucca ( = Hesperoyucca) whipplei se encuentra fuera de los límites del genero Yucca. Palabras clave: Agavaceae, evolución, filogenia molecular, sistemática, Yucca.
INTRODUCTION Indeed, substantial convergence of characters among of distantly related taxa appears to have obscured species The genus Yucca is composed of approximately 45 New relationships. Since the evolution of Yucca likely parallels World species, most of the which grow in the arid regions the climatic diversity of the New World deserts of the ofMexico and the western United States. The genus, one of Northern Hemisphere, this group is of special interest. Our the most widespread of ali the North American desertmonocots, research thus focuses on the evolution of characters in exhibits a high leve! of endemism where it occurs. Striking response to increasing aridity. We thereforeexamine characters in appearance with sword-shaped lea ves and massive, creamy, such as plant habit (tree, shrub, rosette (=acaulescent)), fruit whiteflowers,plants ofthis genus arean importantcomponent succulence (dry, fleshy), leaf rigidity (rigid, flaccid), and of desert habitats. They provide animals with food and shelter and are an integral part of the Iives of traditional, leaf margination (denticulate, smooth) which vary not only desert-dwelling American peoples. The budding stalks, and across species of Yucca, but also within other New World later, blossoms and fruits, are harvested and eaten. The monocot genera such as Agave, Nolina, Hesperaloe and leaves are a traditional source of cordage for rope and twine Dasylirion. Yucca is of particular importance in assessing and the hardened, woody flower stalks are used forconstruction changes of these features across aridity gradients because it purposes. is the only genus in which the en tire suite of characters listed Although these plants have obviously been used by abo ve is fo un d. Tracing the pattern of morphological features people since before the Conquest, little is understood about correlated with increased aridity might provide evidence the relationships between species or the patterns of speciation. abouthistorical specialization in all these genera. In unraveling
______Clary KH, Simpson BB. 1995. Systematics and character evolution of the genus Yucca (Agavaceae): Evidence from morphology and molecular analyses. Boletín de la Sociedad Botánica de México 56: 77-88. 78 KAREN H. CLARY AND BERYL B. SIMPSON the evolution we need to answer whether the ancestors of In dealing with the evolutionary history of Yucca, we Yucca were fleshy-fruited, tropical trees, or small herb-like first ask whether the genus as traditionally circumscribed rosettes with dry fruits. (Trelease, 1902; McKelvey, 1938, 1947; Webber, 1953; Our analyses assess whether the traditional Matuda and Piña, 1980) is monophyletic. Specifically, we circumscription of the genus is natural. Taxonomists have question if Y. whipplei (sect. Hesperoyucca) is a «true» tended to define the genus based upon reliance on its major yucca. Y. whipplei has been placed both within (McKelvey, pollinator, the Yu cca moth (Tegeticula), fruit type, and 194 7; Webber, 1953) and outside of Yucca (Trelease, 1902) stigma morphology (Table 1). (Two monotypic sections, by various taxonomists. Trelease considered it a genus Hesperoyucca and Clistocarpa are pollinated, respectively, (Hesperoyucca) closely allied to both Yucca and Hesperaloe. by two distinct species of Tegeticula moths. Ali of the «Y. whipplei» shares reproductive characters with both Yucca species of the two remaining sections (Chaenocarpa and andHesperaloe (Table 2). (In both these genera, Hesperaloe, Yucca [ =Sarcocarpa}) are poll inated by the species complex, stamen characters are similar: the filaments reflex towards Tegeticula yuccasella). the ovary, they are the same length as the pistil, and the
TABLE l. Morphological characters and pollinators which define traditional Yucca sections (sensu McKelvey, 1938, 1947).
Yucca Section
Hesperoyucca Clistocarpa Sarcocarpa Chaenocarpa
Cha rae ter ha bit shrub/rosette tree/shrub tree/shrub tree/shru b/rosette stigma capitate cleft deeply cleft deeply cleft fruit dry spongy fle shy dry lf. margin denticulate denticulate denticulate/ denticulate/ smooth smooth pollinator T. maculata T. synthetica T. yurcasella T. yuccasel/a
Table 2. Shared reproductive characters of Yucca, Hesperoyucca and anthers dehisce laterally, not apically. Besides pollination Hespera/oe. by a Yucca moth, it shares with Yucca a micro-reticulate pollen exine (Palacios-Chávez, 1978; Álvarez and Kohler, 1987), a large paniculate display and large, campanulate Character Yucca Hesperoyucca Hesp era/oe creamy-whitish flowers. Yucca whipplei possess time autapomorphies, a capitate stigma, tufted anthers and pallen Filaments reflex which is shed in a gelatinous mass rather than as dry toward ovary X X individual grains). Filaments sa me Ata higher taxonomic leve! we consider whether or len gth as ovary X X not the sections, based on fruit types, are monophyletic. An thers dehiscc Does fruit type reflect common ancestry or is it the result of apically X X homoplasy? Does fruit type correlate with habit? Most Yucca moth fleshy-fruited taxa are arboreal, most dry-fruited taxa are pollinator X X Poli en small shrubs or rosettes. Based upon the distribution of morphology X X these two characters among species, researchers (Trelease Large panicle 1902; Tidwell and Parker, 1990) hypothetized that the genus display X X arase as sizeable tropical trees (sect. Clistocarpa) and was Campan ulat e, widespread during the warmer, climatically equitableperiods pendan! flowers X X of the Eocene and Miocene (Tidwell and Parker, 1990; Dott Capitale stigma X and Prothero, 1994). Subsequent genetic diversification Pollen shed as occurred in response to the complex development of North gelatinous mass X American deserts. The Sonaran Desert of the Mexican SYSTEMATICS AND CHARACTER EVOLUTION OF THE GENUS YUCCA (AGAVACEAE) 79
Platea u is thought to have been a refugium during the last Ice the southern Mohave, while sect. Clistocarpa (trees with Age (Van Devender, 1990) and subsequently a center of spongy fruits) occurs entirely within the Mojave. substantial radiation afterwards (Trelease, 1902; Piña, 1980). Determining the ancestral morphology of Yucca is Present-day diversity and distributional patterns are thought speculative. The oldest Yucca-like fossils date from the to be the result of Holocene warming and drying trends Miocene of Nevada, USA (Tidwell and Parker, 1990). The which stimulated range expansion northward, where subsequent ancient fossils of the extinct Protoyucca shadishii are most diversification and reduction of the tree form to shrub and similar to the modern, massive arborescent species, Yucca rosette forms occurred. brevifolia (theJoshua tree), which today inhabits the Mohave The hypotheses of how Yucca diversified are based desert. We question which part of the lineage this fossil upon the fact that the various growth forms now present marks, whether it is basal, intermediate or more derived. within the genus show apparent distributional trends on The architecture of the flower strongly reflects the longitudinal/latitudinal gradients (Table 3). Species atlower mutualistic interaction between the moth and the plan t. For latitudes are trees while species at the higher latitudes and example, the perianth encloses the mating moths during the elevations form grÓund-hugging rosettes. For example, day. During pollination, the pollen is presented by anthers members of sect. Yucca are fleshy-fruited trees of Mexico, which sit atop sturdy filaments and support the weight of the with the center of diversity in the Sonoran desert (Matuda moth's movements as she collects the pollen in to a ball with and Piña, l 980;Piña, 1980). In contrast, mostofthe members her tentacles. Later the moth steadies herself between two of sect. Chaenocarpa are small, freeze-tolerant rosettes filaments in order to position her body for oviposition. The (with dry fruits) which are distributed in the drylands to the sturdy, tapering style is split in to three spreading stigma north and east: the Chihuahuan Desert, the Trans Pecos lobes, which open to receive the pollen mass that the moth region, the Rocky Mountains, the Colorado Plateau, and the pushes down the throat of the style to effect pollination. Great Plains. Two centers of diversity occur, one in Texas, MATERIALS AND METHODS the other on the Colorado Plateau. From south (Chihuahua) to north (Canada), species change from small trees to shrubs We include in our morphological analysis, characteristics of to simple rosettes. Yucca whipplei, a species complex with the ovary, style, stigma, filaments and anthers (34 characters) rosettes and dry fruits, are restricted to Baja California and that had not previously been examined in a systematic way.
TABLE 3. Geographical distribution of Yucca in arid lands of northen Mexico and the United States by habit and fruit type (=Sections).
Region Yucca Section Habit
Baja California, Sarcoca171a Tree Mexico Hesperoyucca Rosette
Sonaran Desert, Sarcocarpa Tree Mexico Sonaran Desert, Sarcocarpa Tree, Shrub USA Chihuahuan Desert, Sarcocarpa Tree Mex ico Chaenocarpa Shrub
Chihuahuan Desert, Sarcocarpa Tree, Rosette USA Chaenocarpa Shrub, Rosette
Trans Pecas, Chaenocarpa Shrub, Rosette USA Rocky Mts., Chaenocarpa Shrub, Rosette USA Mohave Desert, Sarcocarpa Tree, Shrub USA Clistocarpa tree Hesperoyucca Rosette
Colorado Pl ateau, Sarcocarpa Rosette USA Chaenocarpa Shrub, Rosette
Great Plains, Chaenocarpa Rosette USA 80 KAREN H. CLARY AND BERYL B. SIMPSON
We added synapomorphous characters (14) from analyses of Cladistic anal yses were performed using the P A UP algorithm the inflorescence, habit, and leaf shape, rigidity and (Swofford, 1993), Options -unordered characters, ACCTRAN, margination. A morphological character data matrix was from which a strict consensus tree (Fig. 1) of the 54 most made of these characters from 38 taxa (Appendix I, II). parsimonious 268 step trees (CI 0.358) was computed.
FIGURE 1. Strict consensus tree of 54 equally parsimonious 268 step trees from 38 taxa and 47 characters, CI =0.358. Percentages represent values computed from a boots trap 50% majority-rule consensus tree using 100 replicates, MAXTREES setting = 75 , MULPARS option and unrooted trees. Dashed lines represent bootstrap values below 50%. Taxa are denoted by collection numbers.
J] Y. rígida 177 e CY. rostrata "Ce;· ------LY. thompson 121 o D> Y. reverchon 390 (1) Y. arkansana 347 _ _ _ Y. glauca 303 ¡- --r--Y. campes 152 --~Y. interme 148 -- Y. louisian JC { G) _____ -CY. kanabensis 215 ii) e Y. angustis 132 (') ~Y . elata 115 ~ Chaeno. ______i==y. constricta 387 m
Y. standleyi 139 aD> (1) ------cy. coah 190 --r Y. harriman 225 Y. verdiensis 213 ------CY.filam~n Y. utahensis 220 J] 59° Y. queret type e 59% "'2. ______-CY. rupicola 385 (') o Y. pallida 393 ii) .------·v. linear 375 (1) CY. endlich 379 r- -c=v. trec 3392.3 --l Y. schid E42592 Sarco. '------v. filifera 380 Y. baccata 126 -- - Y. thorn 841292 EY. carner 4692 ------Y. brev 041892 ~ Clisto. Y. schotti k72594 '------·v. torreyi C4292 Sarco. '------Y. aloifolia 394 ------C[3~:;¡~~:~:,!'::3 • Hespero. 79% Hesp. parvi 106
~------A. striata 344 SYSTEMATICS AND CHARACTER EVOLUTION OF THE GENUS YUCCA (AGAVACEAE) 81
Characters were run unordered and unweighted with Agave (1938, 1947) and computed another strict consensus tree striata selected as an outgroup. To compare the resolution (Fig. 2) of 48 trees using 339 steps (CI 0.398). These results between our data set and one using more generalized characters, were compared with the traditional classification (Trelease, we added 21 additional characters described by McKelvey 1902; McKelvey, 1938, 1947; Matuda and Piña, 1980).
FIGURE 2. Strict consensus tree of 48 equally parsimonius 339 step trees from 38 taxa and 68 characters, CI =0.398. Percentages represent values computed from a bootstrap 50% majority-rule consensus tree using 100 replicates, MAXTREES setting = 100, MULPARS option and unrooted trees. Dashed lines indicated bootstrap values below 50%. Taxa are denoted by collection numbers.
Y. rigida 177 Y. rostrata 117 l:J ------e== Y. thompson 121 e "2. Y. reverchon 390 g Chaeno. Y. rupicola 385 Q;" (1) Y. pallida 393 Y. queret type Y. linear 375 Y. schid E42592 Y. filifera 380 Y. baccata 126 Sarco. --- Y. thorn 841292 ---L~J Y. carner 4692 Y. schotti k72594 Y. trec 3392.3 Y. torreyi C4292 .....-- --r-r-f" Y. brev 041892 ~ Clisto . 74% Y. aloifolia 394 Y. arkansana 347 Y. glauca 303 Y. campes 152 Y. interme 148 ---r! Y. louisian JC ---L _____ c Y. kanabensis 215 --- Y. angustis 132 - Y. elata 115 Chaeno. 51% ------~ Y. constricta 387 Y. standleyi 139 Y. coah 190 ------eY. harriman 225 ------j r Y. verdiensis 213 ------e=::= Y. filamen Y. utahensis 220 Y. endlich 379 1111!1 Sarco. ------t:C Hesp. funifera 253 91% Hesp. parvi 106 Y. whippl 842192 • Hespero. A. striata 344 82 KAREN H. CLARY AND BERYL B. SIMPSON
The relationships derived from these analyses were intergeneric analyses the Agavaceae (Fig. 4) using the compared with molecular chloroplast DNA analyses of Intercistronic Transcribed Spacer Region (ITS) of the nu Yucca (Hanson, 1993, Fig. 3) taxa and outgroups and with clear genorne (Bogler 1994, this volume).
FIGURE 3. Strict consensus of the 100 most parsimonious 74-step character weighted trees for Yuccoideae based on chloroplast DNA restriction site and length mutations, CI =0.982 , showing discordance between previous taxonomic treatment of Yucca , reproduced from Hanson, 1993. Numbers refer to synapomorphic mutations. Homoplasious mutations are underlined. Boxed taxa in section Sarcocarpa belong to series Baccatae; boxed taxa in section Chaenocarpa belong to series Rupicolae.
Y.camerosana - ll ll -.-- Y.faxoniana Í' Y.schottii 87 t t'iU.irii6#i2'd 1 lt4fuillchiaim: 1 Y.filifera Y. elephantipes 57 60 65 1 Wgrat1iiifWra:I Sarc 1 1 1 1 Y.valida 83 1 22 35 42 53 Y.torreyi 1 1 1 1 1 84 1 1 Y.treculeana 1 Y.madrensis Y.periculosa Y.decipiens
~ Y.aloifolia 1 Y.flaccida ~ ~ 1 Pf;batc-ai4:I ..-Sarc 1 1 Y.brevifolia ..-Clist 15 86 1 1 Y.angustissima 1 2 12 44 58 .82. Y.baileyi Chaen 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Y. cons tricta 22. ']JJ. 104 .a! Y.schidigera ..-Sarc 1 rii#i@W#d Y.elata 7 11 33 1 1 1 '1 Y.glauca Y.filamentosa 18 32 80 89 18. .82. 1 1 1 1 1 Y.navajoa Chaen 1 1 101102 103
lo. 22. ]!). 1 !III Hesp 1 1 1 1 Y. whipplei ..- 17 lo. ~ 76 77 90 96 100 1 1 1 1 1 1 1 H.funifera 1 1 1 1 1 1 1 H. nocturna H.parviflora A.palmen A.lecheguilla ' SYSTEMATICS AND CHARACTER EVOLUTION OF THE GENUS YUCCA (AGAVACEAE) 83
RESULTS Generic relationships: Ali of the analyses place sect. The molecular and morphological trees are congruent in the Hesperoyucca outside of Yucca (Figs. 1- 4), and three ofthe following ways: four analyses, (Figs. 1, 3 and 4) place it on a branch with
FIGURE 4. Combined ITS 1 and ITS 2 data sets, 50% majority rule consensus tree from 100 bootstrap replications (Bogler, 1994 ). The number of base substitutions supporting each branch are given above the line, the bootstrap values are below the Iines.
Beschomeria yuccoides
84% Furcraea pubescens Agave dasylirioides Agave striata Agave attenuata Agave lechuguilla 97% 61% Agave americana o Polianthes geminiflora Polianthes pringlei 99% Prochnyanthes mexicana Manfreda scabra Manfreda virginica
9 Yucca elata 92% Yucca treculeana _,_... 9 Hesperaloe funif era 80% 96% 7 .. Yucca whipplei Camassia scilloides 84 KAREN H. CLARY AND BERYL B. SJMPSON
Hesperaloe indicating that it shares a more recent ancestor are informative. We are expanding the morphological analysis with Hesperaloe than with Yucca sensu stricto. to include fruít and seed characters, and are performing a Sectional relationships: The cladistic analyses are molecular analyses of the 580 base pair ITS (Intercistronic partially congruent with traditional taxonomy. However, Transcribed Spacer Regions) sequence. These analyses will be when the traditionally-named sections are superimposed upon combíned to evaluate further the evolutionary path of Yucca L. the trees, it is clear that none of the sections are monophyletic ACKNOWLEDGEMENTS although the majority of taxa fa]] within the traditional groupings (Figs.1,2). Inoneofthese, Yuccabrevifolia (sect. Clistocarpa) We thank Javier Fuertes Aguilar for his help with the is buried deep within sect. Yucca, which appears to be a Spanish translation of the abstract and Youngong Kim for derived grouping within the Yucca lineage. Taxa of the dry help with the graphics in Fig. 1 and Fig. 2. Michael Hanson fruited, small shrub and rosette, denticulate-leaved series provided Fig. 3 and David Bogler provided Figure 4. This Rupicolae (sect. Chaenocarpa) are basal. research is supported by a Doctoral Dissertation Improvement Series: None of the traditionally-named series are Grant (DEB-9410882) from the National Science Foundation. monophyletic when they are superimposed upon the trees. LITERA TURE CITED As expected, the morphology trees are more congruent with the traditional circumscription than is the cpDNA tree. In Alvarez A, Kiihler E. 1987. Morfología del polen de las Agavaceae y the morphology tree (Fig. 1), ali Chaenocarpa taxa with algunos géneros afines. Grana 26:25-46. smooth-leaved margins fall within a single derived clade. Bogler DJ.1994. Taxonomy and phylogeny of Dasylirion (Nolinaceae). PhD thesis. Austin: University of Texas. Taxa of ser. (ali with denticulate-leaved margins) Rupicolae Bogler DJ, Neff JL, Simpson BB. In press. Multiple origins of the are at the base the two clades within Yucca . Yucca-Yucca moth association. Proceedings of the National Differences between trees: The best resol ved tree is the Academy of the United States of America 92. tree based upon reproductive characters alone (Fig. 1). lt Dott RH Jr., Prothero DR.1994. Evolution ofthe Earth. 5th Ed. New indicates that the dry-fruited, denticulate-leaved taxa share York: McGraw-Hill. pleisiomorphic characters. The most dissimilar tree is the Hanson MA.1993. Dispersed unidirectional introgression from Yucca cpDNA tree (Fig 3). The relationships between taxa are less schidigera into Y. Baccata (Agavaceae). PhD. California: The Claremont Graduate School. resolved: The majority of terminal taxa are at the ends of Hanson MA, Rieseberg L. 1991. Evolutionary history of Yucca: polytomies. In one clade, taxa of three fruit types cluster discordance between morphological and DN A evidence. Paper together. In another, a dry-fruited cladecontains both denticulate presented at the 42nd AIBS Annual Meeting, San Antonio, lea ved and smooth-marginate taxa. Hanson and Rieseberg Texas. ( 1991) has noted that sympatric taxa cluster together in these Hillis DM, Moritz C. 1990. Molecular systematics. Sunderland, clades possibly reflecting past introgressive events involving Massachusetts: Sinauer Associates, !ne., Publishers. a «chloroplast capture,» or transfer, between sympatric taxa. Matuda E, Piña LI. 1980. Las plantas mexicanas del género Yucca. Estado de México, México: Serie Fernando de Alvalxtlilxochitl, CONCLUSIONS Colección Miscelánea. McKelvey SD. 1938. Yuccas ofthe Southwestern United States, Part At the generic leve!, taxa of the Y. whipplei complex share l. Jamaica Plain, Massachusetts: The Arnold Arboretum of more characters with Hesperaloe forming a clade that is Harvard University. sis ter to Yucca. That it possesses the same pollination syndrome McKelvey SD. 1947. Yuccas of the Southwestern United States, Par/ as Yucca raises interesting questions about host shifts within !l. Jamaica Plain, Massachusetts: The Arnold Arboretum of a coevolutionary context (Bogler et al., in press). Harvard University. Of the traditionally named sections, both Yucca and Palacios-Chávez R. 1978. Morfología de los granos de polen de las especies Mexicanas más comunes del género Yucca. Cactáceas Chaenocarpa are shown to be paraphyletic because neither y Suculentas Mexicanas 23:3-8. section contains both an ancestor ADN ali of its descendants Piña LI.1980. Geographic distribution of the genus Yucca. Cactus and (Hillis and Moritz, 1990). However, smaller monophyletic Succulent Journal (U.S.) 52: 277-281. clades of !'leshy, or dry, fruited taxa in partreflect the traditional Swofford D. 1993. PAUP: phylogenetic analysis using parsimony. circumscription. In addition, theJoshua tree (sect. Clistocarpa) Versión 3.1.1. Champaign: lllinois Natural History Survey. falls within sect. Yucca indicating that its spongy fruit is Trelease W. 1902-1911. The Yuccae. Annual Report of the Missouri derived, and that the species arose fairly late in the evolution Bownical Garden 3:27-133. Tidwell WD, Parker LR.1990. Protoyucca shadishii, gen. et. sp. nov., of the lineage. Cladistic analysis suggests the ancestor of an arborescent monocotyledon with secondary growth from Yucca was small and that fleshy-fruited tree forms arose later. the Middle Miocene of Northwestern Nevada, USA. Review of We caution that these results are preliminary in nature. Paleobotany and Palynology 62:79-95. Both morphological and molecular data sets are based upan Van Devender TR. 1990. Late Quaternary vegetation and climate of relatively few characters. Ofthe 64 mutations in thecpDNA tree the Sonoran Desert, United States and Mexico. Chapter 8. En: (Fig. 3), only 38 are synapomorphic and phylogenetically Packrat Middens, the last 40,000 years of biotic change. informative (Hanson, 1993:31). Of the 47 characters in the Tucson: University of Arizona Press. Webber JM. 1953. Yuccas of the Southwest. USDA, Agriculture morphological anal y ses (Fig. 1),41 are informa ti ve. Of the 68 Monographs 17:1-97. characters for the morphology tree represcnted by Fig. 2, 60 SYSTEMATICS AND CHARACTER EVOLUTION OF THE GENUS YUCCA (AGAVACEAE) 85
APPENDIX 1. Master data matrix of 38 Yucca taxa and 79 characters. Characters excluded in 68 character matrix (Fig. 1): # 54, 57, 58, 59, 60, 61, 67, 68, 70, 71, 73, 78, 79. Characters excluded in 47 character matrix (Fig. 2): #48-79
Ag stri 344 20115025401132011100005221001101100410100211111000 He funi 253 010110144200?3?10104015000010101?10400000020001000 He parv 106 00011014300110010104012000010101110400000320001000 Yu whip B42192 00001024541031300114012100000010210320000211001000 Yu rigi 171 11013022000111111012012000000000?10210010211010111 Yu rost 117 12014002410112111102010000000000010200?10211010101 Yu thom 121 12014002110112111001010000000000010220010211010112 Yu reve 390 12015022410112111002010000000000010220010211012000 Yu rupi 385 12013022410112111111010000000000010121010011011000 Yu pall 393 13015022410112121111010000000000010221010201011000 Yu line 375 110150224101???11101012000000000111121010211010100 Yu arka 347 11111203320022?210020110000100000011??011010012000 Yu loui JC 12112202200012111001011000010000110120011000012000 Yu fila 341 11211222400112101002022000010000011220011201012000 Yu coah 190 12012222400122111002000000010000000220011010011000 Yu elat 115 11012202400112101001001000010000000101010210010110 Yu cons 387 12212222400112111001001000010000000201010210012000 Yu verd 213 11213222400112111002010000010000010220010210012020 Yu utah 220 11011222400012111001020000010000010221010211012000 Yu kana 215 11112202400112111011020000010000001220010110012000 Yu stan 139 11011002300112111103010000010000002200010210012000 Yu angu 132 11112402400112111001020000010000000220010000012000 Yu harr 225 11011422400122111002012000010000001220010010011000 Yu glau 303 121111034000222110010120000100000111????0010011000 Yu camp 152 11013305300012111001012000010000021111010010012000 Yu inte 148 11111302300022111001011000010000021120010010012000 Yu endl 379 00112022300113121003122000?10000012220000300011000 Yu schi E42592 100150024000101110111120002100001111000003000101?0 Yu trec 3392.3 100120224001???11101110000210000?112????????010110 Yu bacc 126 21212021410011111110012000210000112200010000011000 Yu torr C4292 10011022410112111001022000210000?11100010Ól0010110 Yu aloi 394 11212202400112111102022000200000011121?10011010011 Yu brev D41892 10010023420012111010020000100002012100010011010222 Yu quer type ??0130?2500?????10?00100000000???1?1000?01110101?? Yu scho K72594 100120024400???1101202100021000???2200010010010000 Yu thor B41292 210120224101???010100110002100???12100010?100100?? Yu cam4692 310120034101???011260120002100???10100010?00010201 Yu fili 380 100150225000???1101102100021000??11100010000010211 86 KAREN H. CLARY AND BERYL B. SIMPSON
APPENDIX 1. Cont.
Table 2, continued
Ag stri 344 ???1??0???04??40?????211????0 He funi 253 ???1??0?0?05??10??3??2101???1 He paiv 106 ?111?00?02050?504?31?21011001 Yu rigi 177 ???1??0???0????0?????210????1 Yu whip B42192 1011110?031510201220221112111 Yu rost 117 ???1??0???0????0?????210????1 Yu thom 121 ???1??0???0????0?????210????1 Yu reve 390 ???1?20???0????1?????210?1??1 Yu rupi 385 20110?00000610?0220222100?111 Yu pall 393 ???1??0???0????0?????210????1 Yu line375 ???0??0???0????0?????000????1 Yu arka 347 ???1??0???0????0?????210????1 Yu loui JC ???1??0???0????0?????210????1 Yu fila ???1??0???0????0??????10????1 Yu coah 190 ???1??0???0????0?????210????1 Yu elat 115 3011010?000310403102221001111 Yu cons 387 ???1??0???0????0?????210????1 Yu verd 213 ???1??0???0????0?????210????1 Yu utah 220 ???1??0???0????0?????210????1 Yu kana 215 ???1??0???0????0?????210????1 Yu stan 139 ???1??0???0????0?????210????1 Yu angu 132 ???1??0???0????0?????210????1 Yu harr 225 ???1??0???0????0?????210????1 Yu glau 303 ???1??0???0????0?????210????1 Yu camp 152 ???1??0???0????0?????210????1 Yu inte 148 ???1??0???0????0?????210????1 Yu endí 379 ???0??0???0????0?????210????1 Yu schi E42592 0(01)000001010212201111?00000111 Yu trec 3392.3 01000101000410201011100000111 Yu bacc 126 11000010110121201101000000111 Yu torr C4292 11001201010320301111100000111 Yu aloi 394 ???O??O???O????O???????O????l Yu brevD41892 10001101020513401221211000111 Yu quer type ???1 ??0???0????0?????210????1 Yu scho K72594 ?1000001000222?01011100000111 Yu thorB41292 ??000010100220201001000000111 Yu cam 4692 01110010110021101001100000111 Yu fili 380 ???0??0????0?????000????1 SYSTEMATICS AND CHARACTER EVOLUTION OF THE GENUS YUCCA (AGAVACEAE) 87
Appendix 2. Characters and character states from mophological analysis. Characters are unordered. Ali characters are synapomorphic.
l. Pistil length: 0=0.5-<2 cm; 1=2-<4.5 cm; 2=4.5-<6.0 cm 2. Papillose portion ofstyle, length: O=O.l-<0.5 cm; 1=0.5- 10. Ovary shape: O=oblong, cylindric; l=slender, tapering; 2=ovate; 3=ellipsoid; 4=obovate 11. No. of deeply divided ovary lobes: 0=3; 1=6 12. Ovary bears impressions of stamens: O=yes; 1=no 13. Style color: O=yellowish; 1=white; 2=green; 3=red/maroon 14. Ovary color: O=yellowish; l=white; 2=green; 3=red/maroon 15. Stigma color: O=yellowish; 1=white; 2=green; 3=red/maroon 16. Length of pistil relative to stamens: 0=0.8-<1.3 times; 1=1.3- Appendix 2.Cont. 3 3. Poli en color: O=pale yellow; 1=egg yolk yellow; 2=lemon yellow 34. Infloresence type: O=strictly racemose; 1=strictly paniculate; 2=characteristics ofboth types 3 5. lnflorescence exertion: O=high abo ve lea ves; 1=at ornear leve! ofleaves; 2=within lea ves 36. Widest part of stamen: O=uniform thickness; 1=top; 2=above middle; 3=rniddle; 4=at base 3 7. Anther basal lo bes morphology @ anthesis: O=sagitate; 1=elongate; 2=rounded 38. Anthertip extension: O=extended outward beyond filament; l=not extended 39. Anther papillae at tip : O=present; 1=absent 40. Anthers wider than filament@ anthesis: O=yes; 1=no 41 . Leaf rigidity, upper half: O=rigid; 1=not rigid 42. Leafflexion, upper half, ventral surface: O=concave; l=convex; 2=flat; 3=conduplicate 43 . Leaf spine morphology: O=acute (short); l=acuminate (long, tapering); 2=neither 44. Leaf fibers composition: O=of single fibers; 1=of multiple fibers 45. Flowers pedicillate: O=yes; l=no 46. Base of corolla a funnel (as in Y. whipplei): O=yes; l=no 47. Caulescence: O= caulescent; 1 = acaulescent; 2 = subacaulescent. 48 . Plant height: O= < 2.5m; 1 = 2.5 - 6m; 2 = 6 - 9m; 49 . Number of stems: O= mostly 1 - 2; 1=mostly2 - 6; 2 = mostly 6 - 24. 50. Stem branching pattem: O= no branching; 1 =once or twice; 2 = few branched. 51. Leaf shape: O= elliptic; 1=lanceolate;2 = ovate-lanceolate; 3 =linear; 4 = linear-lanceolate. 52. Blade length range: O= 0.19m - 0.75m mostly; 1=0.50m - l.5m mostly. 53. Inflorescence length: O= < l. 5m; 1 = > l .Sm. 55 . No. of inflorescence branchlets: O= < 30; 1= > 30. 56. No. ofbracts on scape: O= < 12;1= > 12. 62. Fruit shape: O = tapered above and below; 1 = plump, rounded throughout; 2 = plump below, tapered at top; 3 = slender, long, tapered for most of its length; 4 = oblong-cylindric, rarely tapered; 5 = ovoid; 6 = asymmetrical; 7 = obovoid. 63. Fruit length: O= < 3cm; 1=3 - 8cm; 2 = > 8 cm. 64. Fruit tip shape: O= beaked; 1 = lacking beak; 2 = wide beak; 3 = rounded beak. 65. Inflorescence tinge: O= pure white; 1 =rose or lilac; 2 = reddish purple; 3 = cream; 4 = greenish; 5 = none. 66. Leaftwist: O= not twisted; 1 = twisted. 69. Perianth shape: O= campanulate; 1globose; 2 = spreading; 3 rotate. 72 . Fmit pulp texture: O= fleshy; 1 = spongy; 2 = dry. 74. Monocarpy: O= no; 1 = yes. 75 . Inflorescence branchlet length: O= short; 1 =long. 76. Fmit dehiscence: O= indehiscent; 1 = septicidal; 2 = loculicidal. 77. Flower color: O = red ; 1 = white, cream or green.