INFORMATION TO USERS The most advanced technology has been used to photo graph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are re produced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
University Microfilms International A Bell & Howell Information Com pany 3 0 0 North Z eeb R oad, Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600
Order Number 9011282
Evolution of Erigeron (Compositae) andPeperomta (Piperaceae) in the Juan Fernandez Islands, Chile
Valdebenito, Hugo Alberto, Ph.D.
The Ohio State University, 1989
UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106
EVOLUTION OF ERIGERON (COMPOSITAE) AND PEPEROMIA (PIPERACEAE) IN , THE JUAN FERNANDEZ ISLANDS, CHILE
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of the Ohio State University
By
Hugo Alberto Valdebenito, B.S.
The Ohio S tate University
1989
D issertation Committee: Approved by Ralph E .J. Boerner Daniel J . Crawford Tod F. Stuessy Thomas N. Taylor Adviser Department of Botany To my Family ACKNOWLEDGMENTS
Numerous people should be acknowledged fo r their contributions during the tenure of this study. Dr. Tod F. Stuessy, my major professor, is gratefully acknowledged for his encouragement and guidance throughout my graduate studies. Members of my committee, Drs. Daniel J. Crawford, Ralph E.J. Boerner, and Thomas N. Taylor are given special thanks for their comments and criticisms on this dissertation. This study was supported by grants from NSF doctoral dissertation improvement grant INT-8317088, Sigma X i, and Graduate Student Alumni Research Award from the Ohio State University. Travel to herbaria at New York Botanical Garden, Smithsonian In stitu tio n , Field Museum of Natural History and DePaw University was made possible by an Herbarium
Travel Award from the American Society of Plant Taxonomists. Field work in Bolivia and Peru was possible by a Tinker Foundation Travel Award which is gratefully acknowledged. I am especially grateful to my fellow graduate students in systematics for their aid and companionship in this study. A special thanks is extended to the people in Juan Fernandez, Chile, as well as to the field assistants in
Bolivia and Peru. I also wish to thank the curators of many herbaria from which specimens were borrowed (BISH, BM, CONC, F, GB, LP, MO, and NY).
iii Finally, and most especially, I express my sincerest appreciation to my wife, Herma. Without her support, and care of Cristobal this work never would have been possible.
iv VITA
January 10, 1953 ...... Born - Concepcion, Chile 1976 ...... B.S., Universidad de Concepcion, Concepcion, Chile 1977-1981 ...... Graduate Teaching A ssistant, Botany Department, Universidad de Concepcion, Chile 1981 ...... Llcenciado in Marine Biology, Universidad de Concepcion, Concepcion, Chile 1982-Present ...... Graduate Associate, The Ohio State U niversity, Columbus 1988 ...... Summer Internship, Harvard University Herbaria
PUBLICATIONS Rivera, P. and H. Valdebenito. 1979* Diatoms collected a t the mouths of the Chivilingo, Laraquete and Carampangue riv ers, Chile. Gayana (Botanica) 35. 98pp. Valdebenito, H., M. B ittner, P.G. Sammes, M. S ilv a, and W.H. Watson. 1982. A compound with antim icrobial ac tiv ity isolated from the red seaweed Laurenoia c h ile n sis. Phytochemistry 21: 1456-1457 ______, M. B ittner, R. Zemelman, and M. Silva. 1982. Some components from the Chilean marine algae Laurenoia chilensis. Rev. Lat. Amer. Quimica 13: 37-49. , T.K. Lowrey, and Tod F. Stuessy. 1986. A new species of Erlgeron (Compositae, Astereae) from Chile. Brittonia 38: 1-3* Bittner, M., F. Gonzalez, H. Valdebenito, M. Silva, V.J. Paul, W. Fenical, M. Chen and J. Clardy. 1986. A novel tetracyclic polyketal from the marine red alga Laurenoia ch ilen sis. Tetrahedron Lett. 28: 4031-4032. v Valdebenito, H., T.F. Stuessy, and Daniel J. Crawford. In press. Synonymy in Peperomia (Piperaceae) results in biological disjunction between Pacific and Atlantic Oceans. Brittonia. ______, D. Crawford, M. Silva and T.F. Stuessy. 1984. Phylogenetic relationships of Erigeron (Astereae) in the Juan Fernandez Islands, Chile. Ohio J. Sci. (Abstr.) 84: 5. ______, T.F. Stuessy, and D. Crawford. 1985. Evolution of the genus Erigeron in the Juan Fernandez Islands, Chile. Amer. J. Bot. (A bstr.) 72: 974. ______. 1986. Flavonoid chemistry of Peperomia in the Juan Fernandez Islands. Ohio J. Sci. (Abstr.) 86: 2. ______, and Tod F. Stuessy. 1987• The genus Peperomia (Piperaceae) in the Juan Fernandez Islands and in the South American continent: A phenetic study. Ohio J. Sci. (Abstr.) 87: 7» , T.F. Stuessy and D.J. Crawford. 1987- The genus Peperomia in the Juan Fernandez Islands: A phenetic and cladistic study. Amer. J . Bot. (A bstr.) 74: 760-761. , and Tod F. Stuessy. 1988. Phytogeography of Peperomia b erteroana. An interesting case of transoceanic specific d ifferen tiatio n . Ohio J . Sci. (A bstr.) 88: 13.
• 1988. Seed morphology and relationships of four endemic species of Peperomia (Piperaceae) of the Juan Fernandez Islands, Chile. Ohio J. S ci. 88: 13. , T.F. Stuessy, and D.J. Crawford. 1989. Evolution of Erigeron in the Juan Fernandez Islands: New considerations. Ohio J. Sci. 89: 7.
FIELD OF STUDY
Major F ield: Botany Studies in Plant Systematica and Island Biology. Professor Tod F. Stuessy and Daniel J. Crawford.
vi TABLE OF CONTENTS
ACKNOWLEDGEMENTS...... i l l
VITA...... V LIST OF TABLES...... ix LIST OF FIGURES...... xi INTRODUCTION...... 1 CHAPTER PAGE
I . EVOLUTION OF ERIGERON (COMPOSITAE) IN THE JUAN FERNANDEZ ISLANDS...... 3 Introduction ...... 3 M aterials and Methods ...... 5 Results ...... 14 Discussion ...... 32 Literature Cited ...... 50 I I . A NEW SPECIES OF ERIGERON (COMPOSITAE: ASTEREAE) FROM THE JUAN FERNANDEZ ISLANDS, CHILE...... 61 I I I . EVOLUTION OF PEPEROMIA (PIPERACEAE) IN THE JUAN FERNANDEZ ISLANDS, CHILE...... 67 Introduction ...... 67 M aterials and Methods ...... 69 Results ...... 90 Disoussion ...... 103 References ...... 125 IV. SYNONYMY IN PEPEROMIA (PIPERACEAE) RESULTS IN BIOLOGICAL DISJUNCTION BETWEEN PACIFIC AND ATLANTIC OCEANS...... 133 V. A NEW BIOGEOGRAPHIC CONNECTION BETWEEN ISLANDS IN THE ATLANTIC AND PACIFIC OCEANS...... 140
vii APPENDIX A. A new species of Erigeron (Compositae: Astereae) from Chile...... 155 LIST OF REFERENCES...... 158
viii \
LIST OF TABLES
TABLE PAGE
1. Character and states of species of Erigeron used in the phenetic analysis ...... 7 2. Characters and their states as defined for dadistic analysis of Erigeron ...... 10 3* Data matrix of characters and states in species of Erigeron for cladistic analysis ...... 11 J). Summary of distribution of flavonoids in insular endemic and continental species of Erigeron ...... 25 5* D istribution of flavonoids present in endemic species of Erigeron in the Juan Fernandez Islands ...... 26 6. Distribution of flavonoids in selected species of Erigeron from South America ...... 29 7. Populations of Erigeron (Compositae) on the Juan Fernandez Islands for which chromosome numbers were determined ...... 30 8. Summary of distinguishing characteristics of the subgenera of Peperomia ...... 73 9* Character and character states used in the phenetic analysis of Peperomia ...... 81 10 Characters and their states used in the cladistic analysis of Peperomia ...... 82 11. Data matrix of characters and states in species of Peperomia for cladistic analysis ...... 83 12. Code numbers and names of flavonoids found in Penernmi a ...... 98 ix 13. Distribution of flavonoids in endemic species of Peperomia in the Juan Fernandez Islands ...... 99 14. Distribution of flavonoids in selected species of Peperomia from South A m e ric a ...... 102 15. Populations of Peperomia (Piperaceae) in the Juan Fernandez Islands and continental South America for which chromosome numbers were determined ...... 104 16. Occurrence of flavones in Peperomia tristanensis and P. berteroana on Masatierra (MT) and Masafuera (MF) ...... 144
x LIST OF FIGURES
FIGURES PAGE
1. Phenogram of species of Erigeron using UPGMA...... 15 2. Collections of Erlaeron fernandezianus from Masatierra examined for flavonoid chemistry ...... 17 3’ Collections of Erigeron from Masafuera examined fo r flavonoid chemistry ...... 19 4. Hypothetical evolutionary relationships among endemic species of Erigeron in the Juan Fernandez Islands ...... 22 5. Changes in the flavonoid system in Erigeron. superimposed upon a tree of relationships based on morphological features ...... 39 6. Evolution of the genus Erigeron in the Juan Fernandez Islands ...... 42 7. A ltitu d in al range o f six species of Erigeron in Masafuera ...... 45 8. Erigeron stuessyi ...... 63 9. Fruits of subgenera of Peperomia ...... 71 10—13 • Fruits of endemic species of Peperomia in the Juan Fernandez islands ...... 77 14. Collections of Peperomia in M asatierra studied for flavonoid chemistry ...... 85 15. Collections of Peperomia in Masafuera studied fo r flavonoid chemistry...... 87 16. Phenogram of species of Peperomia using UPGMA...... 91
xi 17* Cladogram of hypothetical evolutionary relationships among endemic especies of Peperomia in the Juan Fernandez Islands ...... 94 18. Evolution of flavonoid chemistry in endemic taxa of Peperomia in the Juan Fernandez islands ...... 114 19- Phylogenetic hypothesis of the genus Peperomia in the Juan Fernandez Islands ...... 119 20. Location of the Juan Fernandez and T ristan da Cunha archipelagos...... 142 21. Principal components analysis o f Peperomia present in Juan Fernandez and Tristan da Cunha archipelagos ...... 146 A1. Erigeron campanensis ...... 156
xii INTRODUCTION
Evolutionary phenomena in flowering plants are sometimes more easily investigated on oceanic islands than in continental areas. These small isolated areas have been the sites of dramatic evolutionary differentiation and adaptations. Oceanic islands contain unique plants, often quite different from close mainland relatives, which have diverged into different ecological niches in a restricted geographical area. Numerous studies have established the potential of understanding phylogeny and evolutionary processes in island biotas. One example of an oceanic archipelago remarkable for the high degree of morphological adaptive radiation and endemism of the native flora is the Juan Fernandez Islands. These islands are located about
600 kms west of mainland Chile, at approximately 33 S latitude, and they consist of two main islands: M asatierra (= Is la Robinson Crusoe) and Masafuera (= Isla Alejandro Selkirk). The Juan Fernandez archipelago can provide especially useful information on patterns of phylogeny and modes of speciation in the native flora for several reasons: It is close to the principal continental source area; it has only two main islands, all volcanic and of known ages (3>7-4.4 million years for Masatierra and 1-2.4 my for Masafuera); Masafuera is 150 kms west from Masatierra which is nearest the principal source area of propagules; and the flora is sufficiently small so that a comprehensive understanding of the evolution of the endemic flora can be realistically achieved. Among the taxa of the Juan Fernandez Islands of especial interest are the endemic species of Erigeron (Compositae) and Peperomia
(Piperaceae). Six species of Erigeron are endemic to the Juan Fernandez archipelago. Five are restricted to the younger island,
Masafuera: ,E. ingae. E. luteoviridis, JE. rupicola. J2. stu e ssy i. and J2. turricola. The sixth, E. fernandezianus. is found on Masafuera and on the older island, Masatierra. Likewise, there are four species of Peperomia in the archipelago: P.. berteroana. P. fernandeziana. F> margaritifera. and P.. skottsbergii. All are endemic except P.. fernandeziana which is also known from southern Chile. A morphological, chemical and cytological study of both genera was undertaken, with the following objectives: (1) to determine the closest mainland relatives of the insular taxa; (2) to determine the number of introduction(s) from which the insular species.originated, and (3) to reveal the phylogenetic relationships among the endemic island species. In addition, a new endemic species of Erigeron is described fo r the archipelago. Finally, an strik in g example of broad transoceanic p attern of d istrib u tio n in Peperomia is documented. C h a p te r I
EVOLUTION OF ERIGERON (COMPOSITAE) IN THE JUAN FERNANDEZ ISLANDS, CHILE
INTRODUCTION
The Juan Fernandez islands, located 600 kms off the coast of Chile at latitude 33°S, consist of two major islands, Masatierra (MT, or Isla Robinson Crusoe) and Masafuera (MF, or Isla Alejandro Selkirk), separated in an east-west line by 150 kms of ocean. Its unusual flora with 147 native species of angiosperms, 69$ of which are endemic (Skottsberg 1956), has been the source for a number of systematic and evolutionary studies with emphasis on a variety of aspects: chromosomal studies (Sanders et al. 1983; Spooner et al. 1987), studies of flavonoid evolution (Pacheco et al. 1985), enzyme electrophoresis to quantify genetic variation (Crawford et al. 1987* 1988), evolution and biogeography (Sanders et a l. 1987), studies on the monotypic endemic family Laotoridaoeae (Crawford et al. 1986; Lammers et al. 1986), survey of the conservation status of the endemio flora (Sanders et al. 1982; Stuessy et al. 1989a), and patterns of phylogeny (Stuessy et al. 1989b). The Juan Fernandez archipelago is a good natural system in which to examine evolutionary trends, e ith e r in morphology and/or chemistry, because: the approximate geological ages of the islands are known (3.7-4.4 million years for Masatierra and 1-2.4 my for Masafuera, Stuessy et al. 1984); the source for the most of the flora is mainland South America; thejre are only two major islands of small size (ca. 75 and 58 km, respectively); and the flora is of manageable size. Among all the native taxa of the islands, the largest family are the Compositae, which comprise 20$ of the flo ra . Erigeron L. (Compositae) is represented by six endemic species in the archipelago: 12. fernandezianus (Colla) Harling, E. ingae Skottsb., JE. luteoviridis Skottsb., E. rupicola Phil., E. stuessvi Valdebenito, and E. turricola Skottsb. This is one of the most important generic systems in the archipelago from the point of view of speciation. One species, Erigeron fernandezianus. is common to both islands, and the remaining five are restricted to Masafuera, the younger island. These taxa differ in leaf shape (from lanceolate to spathulate), flower head size and number, and growth form (semiglobose subshrubs to erect shrubs over 1 m tall). Ecological differentiation also occurs. Different species are found in environments ranging from near sea level to over 1 ,000 m, and from dry and exposed areas near the coast to more wet environments, usually oovered by fog. These observations support the concept of adaptive radiation of Erigeron in the Juan Fernandez Islands (Skottsberg 1953; Solbrig 1962; p ers. observ.), a mode of evolution common in island archipelagos. A morphological, chemical and cytological study of Erigeron was undertaken with the following objectives: (1) to determine the closest 5 mainland relatives of the island endemics; (2) ascertain the number of introduction(s) from which the insular species originated; (3) to reveal the phylogenetic relationships among the endemic island species; and (4) to interpret likely modes of speciation. To help resolve these relationships, numerical taxonomic and cladistic studies based on morphology and in terp retatio n s based on flavonoid chemistry were completed.
MATERIALS AND METHODS
The morphological data used to determine phenetic and cladistic relationships in Erigeron were obtained from consultation of herbarium m aterial (BISH, BM, CONC, GB, GH, LP, MO, NY, and US). Samples of endemic species of Erigeron were collected during four expeditions to the Juan Fernandez Islands in Jan-Feb and Nov-Dee, 1980; January-February, 1984 and 1986. Mean values of 10 measurements of three plants from each taxon were scored for quantitative characters.
Phenetic analysis— Twenty-four species of Erigeron constitute the 24 OTUs (sensu Sneath and Sokal, 1973) in the phenetic analysis. These are: E. andicola DC. [AND]; E. campanensis Valdebenito. Lowrey and Stuessy [CAM]; E. c ilia r is P h il. [OIL]; E. oinereus Hook, and Arn. [CIN], Ei. fasoioulatus Colla [FAS]; E^. fernandezianus (Colla) Harling [FER] (MT, MF);. E. inoaicus Solbrig [INC], E. inoertus (d'Urv.) Skottsb. [INS]; E. ingae Skottsb. [ING] (MF); E. karwinskianus DC. [EAR]; JU. leptonetalus Phil. [LEP]; JSj. lentorhlzon DC. [LEP]; E^ luteoviridis Skottsb. [LUT] (MF), E. luxurians (Skottsb.) Solbrig [LUX]; E. myosotis Pers. [MYO]; E. othonnaefolius Hook, e t Arn. [OTH]; E. pazensis Sch. Bip. ex Rusby [PAZ]; E. pratensis Phil. [PHA]; E. rosulatus Weddell [ROS]; E. rupicola Phil. [RUP] (MF); E. stuessyi Valdebenito [STU], and E. tu rrio o la . F ifty -fiv e morphological characters were scored for each taxon (table 1). The Basic Data Matrix can be requested from the senior author. This matrix was analyzed U3ing numerical techniques included in the NT-SYS package (Rohlf et a l ., 1972). Data were standardized and both correlation and distance m atrices were computed. Cluster analysis was performed on the matrix using the unweighted p a ir group method with arithm etic means (UPGMA). The cophenetic correlation coefficient, which measures the amount of d isto rtio n of the phenogram from the m atrices, was computed a fter the cluster analysis.
Cladistic analyses— In order to determine phylogenetic relationships among the taxa of Erigeron. a set of sixteen characters (7 vegetative and 9 reproductive) were utilized. These, and their numerical assignments are in Table 2. The basic data matrix is shown in Table 3* The different criteria for hypothesizing evolutionary directionality of character states have been reviewed by several workers (e .g ., Crisoi and Stuessy 1980; De Jong 1980; Stevens 1980; Arnold 1981; Watrous and Wheeler 1981; Bishop 1982; Stuessy and C risci 1984; Maddison .et a l. 1984; Donohue and Cantino 1984). The outgroup comparison criterion (Crisoi and Stuessy 1980) was used to assess the polarities for all 16 characters used in the cladistic analysis of 7
Table 1. Characters and states of species of Erigeron used In the phenetic analysis. All quantitative measurements are In mm unless stated otherwise.
HABIT and STEM. 1, HABIT, herb (0), subshrub (1), shrub (2). 2, Height (cm). 3» STEM Dlam; 4, Cross-section, terete (0), sulcate (1); 5, Vestlture, glabrous (0), slightly pubescent (1), strigose (2), pilose (3); 6, Number of shoots, one (0), 2-3 (1), more than 3 (2); 7, Plant shape, elongate (0), semiglobose (1); 8, Stem orientation, acaulescent (0), ascending (1), caespitose (2), caulescent (3); 9, Woody trunk, absent (0), present (1); 10, Branching pattern, unbranched (0), branched at the base (1), branched toward the top (2). LEAVES. 11, Petiole length, sessile (0), 1-3 (1), 4-8 (2), more than 9 (3); 12, Basal rosette, absent (0), present (1); 13, Differentiation In basal and cauline leaves, undifferentiated (0),' differentiated (1); 14, Basal leaf disposition, solitary (0), laxed (1), in dose clusters (2); 15, Leaf scars, absent (0), present (1); 16, Basal leaf length; 17» width; 18, vestiture, glabrous (0), slightly pubescent (1), strigose (2), pilose (3); 19, Texture, smooth (0), slightly rugose (1); 20, Color, light-green (0), green (1), grayish-green (2), greenish-brown (3), light-brown (4); 21, Shape, linear-lanceolate (0), lanceolate (1), obovate-lanoeolate (2), obovate (3), spathulate (4), elongate-lanceolate (5); 8
Table 1. Continued.
22, Margin, entire (0), slightly toothed (1), toothed (2), serrate (3), widely toothed (4); 23, Inflorescence length/leaf length ratio: 1:1 (0), 3:2 (1), 2:1 (2), 3:1 (3), 5:1 (4); 24, Trichome shape, prickle-shaped (0), awl-shaped (1), absent or pustulate-like (2); 25, length (u); 26, Cauline leaf shape, mostly linear (0), usually lanceolate (1). HEADS. 27, Capitulum disposition, end of branch (0), scape (1), branch and scape (2); 28, Nature of scape, woody (0), herbaceous (1); 29, Peduncle, length; 30, Involucre, width, 31i height (cm); 32, Inflorescence arrangement, solitary (0), cymose (1), paniculate (2); 33i Number capitula/inflorescence, one (0), 2-3 (1), more than three
(2). ‘
PHYLLARIES. 34, Shape, linear (0), linear-lanceolate (1), lanceolate (2), triangular-lanceolate (3); 35, length; 36, width; 37> Number of series, one (0), two (1), three or more (2); 38, Midrib black (0), purple (1), yellow (2); 39, Apex and margin color, homogeneous (0), apex reddish (1), margin and apex reddish (2); 40, Vestiture, glabrous
(0), slightly pubescent (1), pubescent (2), highly pubescent (3). LIGULATE FLORETS. 41, Number of se rie s one (0 ),.1 -2 (1 ), 2 (2 ), 2-3 (3), 3 (4); 42, Tube length; 43, Ligule length; 44, width; 45, Number of lobes, 1 (0), 1-3 (1), 3-4 (2); 46, Stamen/filament length ratio. 9
Table 1. Continued.
TUBULAR FLORETS. 47, Tube length; 48, tube shape, constricted (0), unconstricted (1); 49, Stamen/filament length ratio. ACHENES (ligulate florets). 50, Length; 51, Vestlture, glabrous (0), slightly pubescent (1), strigose (2), highly pilose (3); 52, Shape cross section, terete (0), compressed (1), 53, Pappus length; 54, Carpopodium height (u); 55, V estlture, glabrous (0 ), pubescent
(1). 1 0
Table 2. Characters and their states (with numerical assignments) as defined for cladistic analysis of Erigeron.
No. Descriptor Prim itive (0) Advanced (1)
1 Habit two or more stems one main stem 2 Rosette absent present 3 Nature of roots diffuse fascicled 4 Leaf vestiture pubescent glabrous 5 Leaf margins dentate/serrate entire 6 Leaf venation prominent midvein prominent midvein only and secondaries 7 Shape of cauline leaves mostly linear lanceolate 8 Stigmatic bhanch shape triangular gradually tapers toward the tip 9 Number heads/inflorescence 1-5 more than 5 10 Involucre width (mm) .10 and over under 10
11 Involucral bract shape ovate lin e a r
12 Series involuoral bracts 2-3 2 13 Diso corolla tube shape constricted unconstricted
14 Stigma/tubQ ratio 0.5 0.5 (llgulate flowers) 15 Achene cross-section te re te compressed (tubular flowers)
16 Pappus bristles, length twice as long about equal or relative to the achene or longer shorter 11
Table 3. Data matrix of characters and states in species of Erigeron for cladistic analysis. Names of taxa described by the first three letters of the specific epiphets (cf. materials and methods). Ancestor (ANCE) = Erigeron campanensis. E. fasciculatus. E. karwinskianus. E. leptorhizon. E. luxurians. and E,. othonnaefolius.
Taxa Characters
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
ANC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 FER 0 0 1 1 1 1 1 0 11 1 1 1 0 1 1 1NG 0 0 1 0 0 1 1 0 0 0 1 1 1 0 1 1 LUT 0 0 1 1 0 1 1 0 1 0 1 1 1 1 1 1 RUP 1 0 0 1 1 1 0 000 1 0 1 0 1 0 STU 1 0 0 0 1 1 0 1 0 0 1 0 1 1 1 0 RUP 0 1 1 0 0 1 1 0 0 0 1 1 1 1 1 1 12 Erigeron. One Peruvian continental species (E^ leptorhizon) and five Chilean continental species (EL campanensis. E. fasciculatus. E. karwinskianus. E. luxurians. and E^ othonnaefolius) were the species most similar to the island taxa (according to the phenetic results in this study, fig. 1) and therefore were used as the hypothetical ancestors in the present study. For character states that were non-uniform in the outgroup taxa, the most predominant condition was regarded as most primitive. The final step in the phylogenetic analysis of Erigeron was cladogram construction. A number of both manual (Hennig 1966; F arris 1970) and computer methods (Farris 1970; Vagner 1980; Swofford 1985) are available for tree construction. For Erigeron all data were analyzed by'PAUP version 2.4 (Swofford 1985), which infers phylogeny by maximum parsimony. The program was run using CLOSEST, BANDB, ROOT=OUTGROUP which calculates the shortest tree, indicates the most parsimonious trees, and roots the insular taxa ingroup by the outgroup. One aspect of the selection and use of characters in the phenetic and cladistic analyses requires discussion. Most of the continental taxa, and some of the insular species (e.g., E. fernandezianus, E. luteoviridis. and E. rupicola), contain a broad range of morphological variation. For example, 62 characters were initially selected for inclusion in the phenetic analysis, but several species (e.g., the E. andicola complex, E. eouadoriensis. E. fernandezianus. E. karwinskianus. ,E. lanoeolatus. E. leptorhizon. and EL. luxurians) have many states of some characters (e.g., number of ligulate flowers, shape of stigmatio branches in ligulate flowers, and lobe number in the ligulate flowers). The same is true in the phylogenetic analysis. For example, leaf shape was in itia lly selected, but several taxa ( e .g ., E. ciliaris, E. ecuadoriensis. E. fernandezianus. E. lanceolatus. E. leptorhizon. and E. rupicola) have many states of this character (e.g., lanceolate to lanceolate-pinnatifid or lanceolate-spathulate). The same was found in shoot and achene vestiture. Erigeron fernandezianus. E. karwinskianus. E. lanceolatus. and luxurians have from slightly to very pubescent leaves and achenes. This kind of v a ria b ility , which seems common in Erigeron (Cronquist, 1947; Nesom, 1980), causes difficulty in selecting characters and states. The problem was resolved by eliminating characters judged as too variable, or in some instances (leaf vestiture, involucral bract shape, cauline leaf shape) by coding the most prevalent condition.
Flavonoid analysis.— Material for flavonoid chemical analysis was obtained from vegetative parts of continental populations from Chile and Peru collected in Jan-Feb 1986. In addition, material was obtained by permission from selected species in the following herbaria: BM, CONC, GH, LP, and NY. The endemic species in the Juan Fernandez archipelago were obtained during the four expeditions already mentioned, in which 57 populations were collected (fig. 2 and 3) • Thirty-eight populations of E^. fernandezianus were sampled from both islands but principally from Masatierra (26 populations). In addition, two population^ of E. ingae, E. luteoviridis and E. stuessyi as well as 12 collections of E. rupicola were sampled from Masafuera. Specimens were air-dried in the field. Vouchers are on deposit at OS with u duplicates at CONC. The isolation and characterization of flavonoids were conducted using standard techniques described in Pacheco et al. (1985). Identification of the flavonoids was by UV spectral analysis (Mabry et al., 1970).
Chromosome counts-— Flower buds were collected from plants in the field during the collecting trips already mentioned. The buds were fixed in modified Carnoy's solution (4: chloroform: 3 ethanol: 1 glacial acetic acid, v/v) in the field, transferred to 70$ ethanol once in the laboratory and refrigerated. Anthers were macerated in acetocarmine, squashed and the stained tissue mounted in Hoyer's medium. Chromosome numbers were determined from m eiotic microsporocytes v ia phase contrast microscopy and documented with camera lucida drawings. Voucher specimens are deposited at OS.
RESULTS
Phenetic analysis.— The results of the cluster analysis of island and continental species of Erigeron reveal three main groupings (fig. 1). The first contains the six insular species with EL leptorhizon from the continent (Group A). The second cluster contains two subgroups which include five species from Central Chile (B) joined by taxa from Ecuador, Peru and Bolivia (C). The third major cluster (0) contains the rest of the South Amerioan species distributed along the Andes from 30 S down to Tierra del Fuego, which forms a part of the E. andioola oomplex described by Solbrlg (1962). Erigeron lanceolatus. 15
Fig. 1. Phenogram of species of Erigeron using UPGMA. Dots indicate species present in the Juan Fernandez Islands. Scale indicates level of correlation. Capital letters correspond to subgroupings (see text). 16
I i 1 I I I • 0.2 0 0.2 0.4 0.6 0.8
Fig. 1 17
Fig. 2. Collections of Erigeron fernandezianus from Masatierra examined for flavonoid chemistry. Numbers refer to collections of taxa listed in table 5; otherwise correspond to herbarium collections.
» 18 19
Fig. 3• Collections of Erigeron from Masafuera examined for flavonoid chemistry. Numbers refer to collections of taxa listed in table 5; otherwise correspond to herbarium collections. ^ E. fernandezianus; ^ E. ingae; O i- luteoviridis; Oi?* rupicola; H E. stuessyl;'^^' E. turricola. 20
Fig. 3 21 from Bolivia and Northern Argentina, is also included.
Cladistic analysis.— From the phenetic analyses, six continental species were selected as hypothetical ancestors to root the tree(s): Erigeron leptorhizon and the group formed by five species found in Central Chile (E^ campanensis. E. fasciculatus. E. karwinskianus. E. luxurians. and Ej_ othonnaefolius) . Several analyses were done either with the Chilean species and E. leptorhizon as the ancestral group or with the Chilean species only as the ancestral group. In either situation, the cladistic relationships among the species were the same (fig. 4). The resulting tree has a total of eighteen character state changes with two of these representing parallelisms (4 and 14, representing leaf vestiture and ratio of length of stigma/tube of ligulate flowers, respectively). The high consistency index of 0.89 re fle c ts a low lev el of homoplasy in the evolution of the group. The presence of four synapomorphies distinguishing the insular species from the ancestral group supports the hypothesis that the Juan Fernandez taxa are monophyletic. Two main evolutionary lines (clades) can be distinguished: (1) an association of J3. rupicola and E. stuessyj (first principal branch); and (2) an assemblage of E. ingae. E. turrioola. E. luteoviridis and E. fernandezianus (second branch). In the first prinoipal branch, E. rupicola shows the fewest derived character state changes and may be regarded as the most primitive taxon for the whole group. The four insular species in the second clade (Erigeron fernandezianus. E. ingae. E. luteoviridis. and E. turrioola) are united based by four oharacter states: nature of roots (3), oauline 22
Fig. 4. Hypothetical evolutionary relationships among endemic species of Erigeron 'in the Juan Fernandez Islands. For description of characters see table 2. Single bars indicate synapomorphies; double bars parallelisms. 23
STU RUP ING TUR LUT FER
Fig. 4 2 4 leaf shape (7)» series of involucral bracts (12), and length of pappus bristles relative to the achene (16). In this lineage, E!. ingae and El. turrioola are differentiated by the ratio of length of stigma /tube ratio of ligulate flowers (14), a trait that apparently has had an independent origin (parallelism) in E. stuessvi. and by the rosette arrangement of the leaves (2) in the latter. Furthermore, _E. fernandezianus and E. luteoviridis are shown to be closely related, also corroborated by the cluster analysis (fig . 1). These two taxa show the greatest character state changes, and E. fernandezianus may be regarded as the most advanced taxon in the archipelago.
Flavonoid analysis-- Identities of the flavonoids from all collections analyzed are: quercetin 7-0-glucoside, 6-hydroxy quercetin 7-0-glueoside, quercetin 8-C-glycoside, quercetin 7-0-glycoside, luteolin, 6-0-methoxy luteolin, 6-methoxy luteolin 7-0-galactoside, luteolin 7-0-glucoside, luteolin 7-O-diglucoside, aplgenin 7-0-glucoside, and acaoetin 7-0-diglucoside 6-C-glucoside (table 4).
The f ir s t four are flavonols and th e remmaining seven are flavones. Glycosides and C-glycosyl derivatives are present in both classes of compounds. The flavonols were based on quercetin, and the flavones on either luteolin, apigenin, or acacetin derivatives. Sugar substitution of the basic compounds encountered was as 7-0-monoglycosides of glucose and galactose. There is qome interpopulational variation in all taxa (table 5) as well as a general pattern of species-specific profiles (table 4). The exceptions are Erigeron ingae. E. luteoviridis and E. turrioola. which Table 4. Summary of distribution of flavonoids in insular endemic and continental species o f Erigeron. A = apigenin; Ac = acacetin; L =,luteolin; Q = quercetin. ■ ....I...... r " 1 a * FUNKS S v to sid H C -olvcoflm nols Aolvcone Glycosides C-alvcosvlflavones No. Htydrcxy 0 D 7*0- 0 C - 0 7-0- L 6-0- 6-eethyl L L 7-0- L 7-0- A 7-0- Ac 7-0-diglu to ll.* 7- E. tmmSezimus 26 ** ♦ f ♦ ♦ ♦ * ♦ NHFUBA f . fertundnims12 7 ♦ 7 7 * ♦ * ♦ * E. in g e 2 ♦ ♦ ♦ * ♦ * ♦ ♦ * ♦ L luttoriridis 2 ♦ * * * ♦ ♦ ♦ ♦ ♦ E. rupicalt 12 «* ♦ * ♦ ♦ ♦ E. stueayi 1 ♦ ♦ ♦ ♦ E. turriolci 2 ♦ * ♦ * ♦ ♦ ♦ ♦ * ♦ IMDUND SOUTH AHERIGA E. cmpumis 1 ♦ ♦ E. fixiculttus 2 ♦ ♦ ♦ * ♦ ♦ ♦ ♦ * E. kminskizna 1 ♦ ♦ » ♦ ♦ * * E. leptorhizon 3 ♦ ♦ * * ♦ ♦ ♦ + E. luxuries 3 * * ♦ ♦ * ♦ ♦ * ♦ * Total number of collections analyzed for each taxon. Table 5. Distribution of flavonoids present in endemic species of Erigeron in the Juan Fernandez ______Islands. A » apigenin; ftc = acacetin; L = luteolin; 0 - quercetin.______FUWHS______„ ______FlfllDES______GLYCOSIDES Enlycollivonols ftalvcont _ Glycosides ______C-olucosvlflranes Sjpecits toll, (tap Hlydrmy 0 0 7-0- 0 0 7-0-C- L 6-0- 6-aethyl L L 7-0- ft 7-0- L 7-0- Ac 7-0-diglucoside not. nos. 7-0-olucosidt olucoside C-olucoside qlucoside wWww L 7-OnalKtoside olucoside olucoside diolucoside 6 C-olucoside HASATIERRA E. (tnunhzims 3000 1 ♦ 5000 2 5143 3 ♦ 5153 4 ♦ 5219 5 5244 6 6220 7 ♦ 6221 8 6222 9 6228 10 6236 11 6237 12 ♦ 6238 13 6239 14 ♦ 6325 15 6424 16 6478 17 ♦ 6478A 18 ' 6491 19 ♦ 6499 20 ♦ 6503 21 ♦ 6508 22 6530 23 6542 24 ♦ 6543 25 ro 6S59 26 ♦ o Table 5 (continued). jyrons flM O B HI. Ch1ko(1«¥Qho1» Aalvcont 61vcotidw C-ducotYlfl4wn« SfeaciH Coll. ItaP 6 fcydroty Q 0 7-0- 0 0 7-0-C- L 6-0* 6-oethyl L L 7-0- A 7-0- L 7-0- Ac 7-0-di9lucoside not. m , 7-0-alucoii* olucotid* C-ol«o*iO» alu co ii* m toan L 7-0-o«l«to»ide oluco»id» alucoeide diglucoiidt 6 C-glucosi* mavuBM £ f m u n i n i m n5054 27 6397 28 6417 29 815B 30 « 6212 31 ♦ 8214 32 4 4 6243 33 4 4 9067 34 4 4 9139 35 4 4 4 9162 36 9315 37 4 4 9455 38 4 £ inpt 9110 39 4 4 4 9640 40 4 4 4 4 £ IidtonrUis9100 41 4 4 4 9161 42 4 £ n f i c e l i 5064 43 4 8107 44 4 4 8113 45 8121 46 4 4 8134 47 4 4 8136 48 4 813B 49 4 8215 50 4 8216 51 4 8450 52 4 4 8460 S3 8508 54 4 £ stmayi 9247 55 £ t m r i c o l i 9337A 56 4 ♦ 9357 57 4 ♦ ♦ 28 have identical arrays of compounds (table 4). Erigeron rupicola and E,. stuessyl show similar flavonoid profiles with the exception of two flavonol glycosides in the latter (table 4). Interestingly. Erigeron fernandezianus exhibits flavonoid variation between islands (table 5). Collections from Masatierra lack C-glycoflavonols but they have luteolin 7-0-diglucoside. The broadest interpopulational variation is in E. leptorhizon and E. ingae. perhaps due in part to having herbarium and fresh material for extraction. The leaves of the putative continental ancestors analyzed yielded similar flavonoids to those found in the insular taxa, with the exception of E. campanensis which has only two flavones (table 6). Chromosome counts— Previous chromosome reports of Erigeron on the islands, as well as new counts from 20 populations representing six taxa in the archipelago, are listed in table 7* All are n, = 27. These are first reports for Erigeron ingae and E. luteoviridis. Erigeron turrioola was counted once previously as E. cf. rupicola (Spooner et a l. 1987). The counts confirm previously reported numbers (Sanders et a l. 1983; Spooner et a l. 1987) and give cytologlcal information for additional populations in the islands. Additional counts from each species were highly desirable before accurate judgments could be made regarding evolutionary relationships. Table 6. Distribution of flavonoids in selected species of Erigeron from South America. A = apigenin; Ac = acacetin; L “ luteolin; Q ** quercetin. FUWBBLS FUMKES GLYCOSIDES C-olKoflivonoli Aolvcone 61vcosidw C-qlucosvlflevcnes $ ic c in Coll. lo c. 6-frydroxy 0 0 7-0- 0 0 7-0-C- I 6-0- G-wthyl L L 7-0- A 7-0- L 7-0- Ac 7-0-diglucoside • > no. 7-0-oluto»i4e elereoide C-oliicoiidt alucoiide oettav L 7-0-«loctc«ide olucoside olucoside diolucoside 6 C-olucoside IHDUM) £ cmpemsis 6 51068 Chile £. tuciculitia J 3452 Chile S 9708 Chile £ krvinsiimus IB 38606 Chile £ liptorhiion A 13747 Peru VB 352 Peru U 5924 Peru £ /a n r ia s POP 1026 Chile POP 1049 Chile __J 1799 Chile Collectors: A ■ Asplund; Q ■ Goraventa; J “ Jiles; MO ■ Marticorenait Quezada; POP - Pacheco, Campos, N eira it Pacheco; S = Stuessy; VB * Valdebenitoi t Benavente; W =» Meberbauer. ro vO 30 Table 7. Populations of Erigeron (Compositae) on the Juan Fernandez Islands for which chromosome numbers were determined. All are n = 27. Vouchers deposited at OS. Taxon Voucher E. fernandezianus ••MASAFUERA: So 3645. (Colla) Harling MASAFUERA: Between Quebrada Casas and Q. Vacas. VL 8216: Q. Pasto', SG 9025: Q. Casas. SVLG 9047. 9067; Q. Pasto, LR 9436. •MASATIERRA: Carbonera de Torres. SMSV 5320: Q. V illagra. SMSV 5524. ••MASATIERRA: So 3510. •••MASATIERRA: Trail from Mlrador Selkirk to Valle Villagra, SCPVRL 6238. v 6542 . E. ingae Skottsb. MASAFUERA: Cordon Atravesado, LR 9247: Q. Pasto. SL 9419: Cordon Inocentes. SS 9544: Q. Tongo. LR 9640. E. luteoviridis Skottsb. MASAFUERA: Q. Guaton, SV 9113, 9315: Los Inocentes, L 9357; Q. Pasto, SL 9408. E. ruoioola Phil. •MASAFUERA: Chorro Dona Maria. SARS 5064 MASAFUERA: Between Q. Chica Varadero and Q. Varadero, VL 8106; Tierras Blancas. VL 8121. SR 8453: Q. Vacas, A8368; Q. Tongo. SR 8460: Q. Angosta. VLS 8508: Q. Guaton. LG 9110. E. stuessyi Valdebenito •••MASAFUERA: Q. Las Casas. PR 6401. E. turrioola Skottsb. MASAFUERA: Between Q. Vacas and Q. Guaton. SRL 9337* Table 7 Continued. * = Sanders et a l. 1983. *« = Solbrig, Anderson, Kyohs and Raven, 1969. = Spooner et a l. 1987* A= Arriagada; Cs Crawford; L= Landero; M= Marticorena; P= Pacheco; R= Ruiz; Sa= Sanders; Ss Stuessy; So= Solbrig; Vs Valdebenito. 32 DISCUSSION Relationships of mainland species and island endemics The genus Erigeron is represented in South America by the two sections Leptostelma and Erigeron (Solbrig, 1962), which are separated mainly on characters of habit and distribution. Species of section Leptostelma are large herbs up to 4 m high with no basal rosette, profuse foliage, and heads arranged in loose cymose secondary inflorescences. Members of th is section are found in wet and marshy localities at altitudes below 1 ,000 m in northeastern and southeastern Brazil, southern Argentina, Uruguay, and Paraguay. Members of section Erigeron have basal rosettes, and mono- or polycephalous floral scapes up to 1.5 m tall. Species of this section are found at altitudes ranging from sea level to 5,000 m in mostly dry, rocky habitats. This is the larg est and most variable of the North and South American sections of the genus (Cronquist, 1947; Solbrig, 1962). It comprises 24 species (two of them recently described, Valdebenito et al. 1986; Valdebenito, in prep.) including one species complex (Erigeron andicola. E. ciliaris. E. oinereus. S. leptopetalus. and E. myosotis) which has been considered to contain several intermediate forms (Solbrig, 1962). The d istrib u tio n of th is section in South America is along the Andes from Ecuador to Tierra del Fuego including the Falkland or Mai vines islands (E. inoertus) and the Juan Fernandez archipelago. To investigate taxonomic and geographic relationships among the endemic 33 insular species and those elsewhere, therefore, one must consider members of both sections. In a phenetic analysis based on the total number of species (29) of Erigeron present in South America (Solbrig, 1962; Valdebenito e t a l. 1986) those taxa belonging to section Leptostelma (E. camposportol Cabrera, E. maximus (D. Don.) DC., E. meyeri Cabrera, E. tucumanensis Cabrera, and E. tweediel Hook, et Arn.) grouped in one cluster. Due to these results, the subsequent analyses were done on the remaining 24 species included in section Erigeron only. Based on morphological and phenetic analyses, six continental species were proposed as most closely related to the insular species. This group is composed of five species from Central Chile and one from Southern Peru, with the latter seemingly most closely related. This taxon, Ej. leptorhizon. was consistantly grouped phonetically with species present in the islands. The coastal occurrence of E. leptorhizon offers good possibilities as an ancestral stock, but geographically it is at considerable distance north of the islands. It is also possible that Erigeron developed from ancestors from Chile, which also cluster close to the island taxa. The presence of thick leaves (sometimes thinner and lanceolate to spathulate in E. leptorhizon and E. karwinskianus), dentate to entire margin, smooth leaf surface, heads arranged in loose paniculate synflorescence or solitary, and overall morphological resemblanoe (perennials from a slender rootstock with several ascending stems from the base ranging 0.35-1 m ta ll), suggest a olose relationship among these continental taxa and the insular speoles. 34 Little is known about the cytology of Erigeron in South America which might provide insights on genomic relationships. Of the approximately 24 species currently recognized in section Erigeron. only nine, six from the archipelago which are a ll n = 27, have been studied eytologically. Only two species of the putatively ancestral group determined in the phenetic analysis have been examined: Erigeron leptorhizon DC. with n = 18 (Diers 1961) which is confirmed in the present study [Peru, Lomas de Lachay, Valdebenito and Benavente 352], and E. karwinskianus DC. Several counts have been reported for the la s t species, though not from South America: n. = 27 I (Turner, E llison, and King, 1961), n. = ca. 27 (Turner, Powell, and King, 1962), n. = 5 II & 17 I (Keil and Stuessy, 1975), and 2n = 36 (Anderson et a l. 1974). Considering the v ariatio n in chromosome number that has been documented in this species, it has been suggested to be apomictic through part of i t s range from Mexico to northern South America (Solbrig 1962). Two of the three remaining continental taxa, in separate clusters in the phonogram (fig. 1), have been reported as n = 27: Erigeron ecuadoriensis Hier. (Jansen and Stuessy 1980), and Erigeron lanceolatus Wedd. var. subacaule Wedd., (Turner, Powell and Cuatrecasas 1967)* The remaining speoies, E. myosotis Pers., has been reported with n = 18 (Solbrig et al. 1964). It is difficult to know, therefore, whether polyploidy aooompanied the evolution of the species of Erigeron in the islan d s. I t appears these endemics are a ll hexaploids on an ancestral base of x = 9, but whether they came ohromosomally unchanged from an .n = 27 ancestor is still uncertain. Previously flavonoid chemistry had not been surveyed for the South 35 American species of Erigeron. and only two species from elsewhere had been studied: Erigeron acris L. and E,. annus (L.) Pers. (Kaneta et al. 1978). The compounds reported were luteolin 7-0-glucoside in the former, and apigenin 7-0-rhamnoglucoside and quercetin in the latter. Therefore, our results represent the first reports of flavonoids for South American species. In general, the flavonoid data co rrelate well with the taxa grouped in the phenogram based on morphology. The morphological closeness of the six endemic taxa is substantiated by their similar array of flavonoids. Vi thin this group, E. ingae and E.. turrieo'la are very close (E. luteoviridi3 is also similar) as supported by the presence of identical flavonoids and flavones. Likewise, the only species present in both islands, E. fernandezianus, has basically the same flavonoid profile with the exception of two C-glycoflavonols absent on populations from Masatierra and present in only two populations from Masafuera. Similarly, the conformity of flavonoid profiles of Erigeron stuessyi and E. rupicola is consistent with the results of the phenetic and cladistic analyses, which indicate a close relationship of these two taxa. ' They are separable only in the lack of two flavonol glycosides in E. stuessyi. Except for E. campanensis. the flavonoid affinities of the putatively ancestral group and the endemic taxa are strong. Seven out of ten compounds are shared by the species in this group. Erigeron fasoioulatus and E. luxurians, two of the continental species, are particularly noteworthy in similarity of flavonoids. The only compound missing in this group, but oocurring in some endemic taxa, is quercetin 7-0-C-glycoside. 36 Number of introductions to the archipelago There has been some problem in determining the number of introductions of Erigeron to the Juan Fernandez Islands. Solbrig (1962) pointed out that .E. ruplcola may have speciated through a separate introduction from the mainland, because he judged it to be not closely related to the other endemic species. Skottsberg (1922, p. 182), however, reported that all the island species are closely related to each other, in spite of E. rupicola being somewhat unlike the rest. In addition, he pointed out that none of the continental Chilean species seemed to be related closely to the island endemics but rather to some species from Oceania, e .g ., J3. rapensis F.B.H. Brown, and IS. lepidotus Less., (which is now referred to Tetramolopium: Bentham 1873» Lowrey 1986) and the re st of South America: E. uliginosus Benth. and J2. heteromorphus Rob. from Colombia and Mexico, respectively. According to the dadistic analysis (fig. 4), floral features (e.g., shape of involucral bracts [11], disc corolla tube shape [133 and achene shape [15] as well as vegetative features such as leaf venation [6]), lend support to the monophyly of this group and its separation from the putative ancestor from mainland Chile. It is worth mentioning that the cluster analysis makes a compact insular assemblage separated from continental species at the 0.06 phenon level, which in turn also formq a defined group of taxa from mainland Chile separated from the rest of the South American speoies. The uniformity in ohromosome number, n = 27, in a ll the endemic taxa (tab le 7) provides 37 additional evidence of this close relationship. Relationships among endemic species To investigate phylogenetic relationships in Erigeron within the archipelago requires making two assumptions. First, it is assumed that all six species in the islands have 'developed from only one introduction from the continent, and therefore, that the group is monophyletic. Second, it is assumed that a cladogram generated based on morphological characters (fig. 4) can be used to represent the most parsimonious evolution of Erigeron in the islands and does reflect the actual phylogenetic relationships among the taxa. Four synapomorphies [nos. 6(1), 11(1), 13(1), and 15(1)] link two principal clades. First, Erigeron rupioola and E. stuessyi are distinguished from the other Juan Fernandez species by the globose habit arising from one short stem, and spathulate leaves with entire margins, in this first clade, which apparently diverged at some e a rlie r time during the evolution of the genus in the archipelago, E. rupioola is separated from J2. stuessyi by its highly pubescent habit (4), a trait that may have evolved independently in the rest of the species. The second clade is supported by synapomorphies of fascicled roots (3), lanceolate oauline leaves (7), two series of individual bracts (2), and the presence pappus bristles do not equal in length to the achene (16).. Erigeron ingae and E. turricola are basal members of this evolutionary line leading to the more advanced species, E. luteoviridis and 12. fernandezianus. the latter species being the most 38 advanced taxon in the archipelago. Evolution of the flavonoid system Even though no method has been universally accepted to evaluate flavonoid data in a phylogenetic manner (i.e., Bate-Smith 1962; Crawford 1978; Giannasi and Crawford 1986; Gornall and Bohm 1978; Gornall et al 1979; Harborne, 1977; Stuessy and Crawford, 1983), some views regarding use of flavonoids for inferring phylogenetic relationships in plants have been discussed (Crawford, 1978; Giannasi, 1978; Gornall and Bohm 1978; Harborne 1977, and Young 1981). Because of the uncertainty about general trends of evolution in flavonoids, the most conservative approach is to correlate flavonoid distribution with evolutionary patterns determined by morphology (Pacheco et al. 1985) as well as by relative positions of flavonoid classes in biosynthetic pathways (Giannasi 1978). If we accept the cladogram for the insular taxa of Erigeron as being valid (fig. 4), the most parsimonious evolution of the flavonoid system in the taxa reveals a trend in reduction in flavonoid complexity (fig. 5) from the center of dispersion, South America, to the site of colonization of the genus in the Juan Fernandez archipelago, Masafuera. The ancestors of the endemic species had C-glycosylated flavones and flavonols which were lost in the .E. stuessyi and J3. rupioola clade. Furthermore, flavonols were lost as the evolution of E. stuessyi occurred. Proceeding one step further, the evolutionary line leading to the four remaining Insular species shows a gain in two flavonols. 3 9 Fig. 5. Changes in the flavonoid system in Erigeron, superimposed upon a tree of relationships based on morphological features (fig. 4). 4 0 STU RUP ING TUR LUT FER f l a v o n o l s f l a v o n o l s Fig. 5 41 Here the presence of an extra A-ring hydroxylation at the 6-position (e.g., quercetin 6-hydroxy 7- 0-glucoside) correlates with morphological specialization by being regarded as an advanced condition in general within the angiosperms (Harborne 1967, 1972). During the evolution of J3. fernandezlanus in the past 1-2 million years, there have been no major flavonoid changes except for the absence of C-glycosylation in flavonols. However, some populations on Masatierra also do still show an absence of C-glycosylation. Evolution of Erigeron in the archipelago The results of the phenetic and cladistic analyses, in conjunction with the geological age of the islands of the archipelago (Stuessy et al. 1984) and its presumptive geological history (Sanders et al. 1987), allow an evolutionary hypothesis for Erigeron in the Juan Fernandez Islands to be formulated (fig. 6). The endemic species in the islands, which are all closely related, could have been derived from propagules from the ancestral continental stock in one introduction to Masafuera, the younger island, in the past two million years. All potential mainland relatives are coastal, with Erigeron leptorhizon common in the dry hills (lomas) along the Southern Peruvian coast, and the Chilean taxa found on hills near the coast or on beachs from Concepolon to Coquimbo. How propagules successfully traversed the 690 km of ocean from the continent to Masafuera is Impossible to know with assurance. However, the setose pappus on the achenes suggests seed dispersal by wind or by F ig. 6 . Evolution of the genus Erigeron in the Juan Fernandez Islands. Tree shows phylogenetic relationships among the endemic species. ^ South America E. laptorhlxan STU RUPING TUR LUT FER 1 f i * •• E. campananala / faaelculalua kararlnaklanua E. luxuriant (4 Minton year*} Fig. 6 VjJ u attachment to bird feathers, modes common in many species of Astereae (Carlquist 1974; Venable and Levin 1983)* Dispersal to the island of Masafuera by cyclonic easterlies from the continent seems especially lik ely . Once the ancestor became established in the islands, the insular taxa appear to have been derived by adaptive radiation. Masafuera provides a broad variety of habitats concentrated into a relatively small geographic area. Therefore, isolation of small populations in these diverse habitats can lead to rapid evolution (Crawford, Whitkus and Stuessy 1987, also see Carlquist 1974; Carr and Kyhos 1981 , 1986; G ilJett and Lim 1970; Gardner 1976; Helenurm and Ganders 1985; for examples in Hawaii). According to the dadogram in fig. 4 , three pairs of species can be differentiated. Each taxon within the species-pair is differentiated by an ecological shift from dry or relatively dry and exposed environments to more humid and protected sites, a type of shift rather common among plant species on oceanic islands (Carlquist 1974). The first species group is formed by Erigeron rupioola and E. stuessyi. Erigeron rupioola. the most primitive taxon probably derived from a continental ancestor growing on open fields near the beach, is a common speoies growing a t lower a ltitu d e s from sea level up to 300 m (fig. 7) on open coastal cliffs. Its related species, E. stuessyi. diverged in to more protected and humid environments represented by deep ravines ( e .g ., Quebrada Casas). I t i s not unlikely th at E. stuessyi evolved recently from E. rupioola. Only two populations of the former at the 200 m band have been reported, and only two at 950 m on Cordon Fig. 7« Altitudinal range of six species of Erigeron in Masafuera. Dots correspond to individual collections. ELEVATION IN METERS O ion 200 800 400 600 600 700 600 900 1000 1100 1200 1300 1400 E. fernandezianus E. i n g a e E. luteoviridis E. rupicota E. stuessyi E. turricola Fig. 7 ■F- O 4 7 Atravesado at a point much higher up in the same ravine. Shrouded in clouds most of the time, Los Inocentes and Cordon Atravesado represent the highest part of the island, over 1,000 m, where three species occur: E. ingae. E. tu rric o la , and E. lu te o v irid is. The first two species form the second pair of closely related taxa differentiated by their habitats. Erigeron ingae commonly occurrs in Lophosoria covered rocky outcrops (pers. observ.) whereas E.. turricola grows scattered on open and d rier steep c lif f s , sometimes in small crevices of the rocks. They show some expected differences: IS. turricola. resisting strong winds, is smaller with the leaves in rossette arrangement with a strong peduncle supporting the single capltulum. Erigeron ingae is a subshrub, 30 cm ta ll, with several shoots covered with highly pubescent leaves, and with the heads in a paniculate arrangement. Erigeron luteoviridis. and the very closely related EL fernandezianus. represent the third pair of species differentiated ecogeographically. Erigeron luteoviridis occurs from 1050 m and up (fig. 7) in regions usually covered by ferns (e.g., Lophosoria quadripinnata. Thyrsopteris elegans). On the other hand, Erigeron fernandezianus is well differentiated by having the broadest distribution in the island (fig. 3 and 7)• Usually it is found growing in open sunny places, especially rooky w alls and slopes (Kunkel 1957; Johow 1896; Skottsberg 1922; pers. observ.). This species, which is the most c la d ia tio a lly advanced taxon (fig . 4), occupies the open intermediate elevation habitats. After this divergence, the dispersal of Erigeron fernandezianus to AS Masatierra probably occurred during historical time. This archipelago, discovered in 167^» served during the 18th and 19th century as water, wood, and food supplies for whalers and buccaneers, as well as for a permanent population on Masatierra (Vicuna Mackenna 1883; Woodward 1969)* Erigeron fernandezianus may have been brought to M asatierra after continous visits by fishermen to Masafuera, where there is an abundance of lobster, a commercial product now shipped to mainland Chile. The major occurrence of J5. fernandezianus on Masatierra occurs along trails that go from the Village (estimated population 500 inhabitants) to Quebrada de V illagra (fig . 2). These endemic species of Erigeron are the only group of taxa th at may have come o rig in ally from mainland South America and speciated in situ on the younger island (Masafuera) followed by subsequent dispersal of one of its species (E. fernandezianus) to the older island (Masatierra). It is important to keep in mind, therefore, that inferences about progenitor and derivative taxa on islands must be made carefully because phytogeographic trends may not always be most parsimonious. Furthermore, Erigeron also contributes to a further understanding of the evolution for the entire flora of the archipelago by suggesting the Important role that adaptive radiation has played during speciation. It would be of interest to conduct compatibility studies and artificial hybridizations among the insular species. There is no information regarding modes of reproduction in Erigeron. with the exception of some Isolated information reporting apomixis in some 4 9 continental taxa (Holmgren 1919; Cronquist 1947; Solbrig 1962). Crossing studies would indicate whether reproductive barriers, in addition to ecogeographical differences, have been involved in the differentiation of this genetically cohesive insular group (n. = 27) with apparently no hybrids. Preliminary studies (unpubl.) have shown complete s ta in a b ility and uniformity in pollen grain size and normal chromosome pairing and segregation in meiosis, both of which support the motion of high pollen viability within the six endemic species. It has been shown in several insular groups of Compositae (e.g., Bldens. Ganders and Nagata 1984; Tetramolopium. Lowrey 1986; Lowrey and Crawford 1985; Lipoohaeta. Gardner 1976; the silversw ords, Carr 1985; Dendroseris. Crawford et al. 1987) that natural interspecific hybridization appears to be rare or non-existent in spite of the lack of gametic or post-zygotic iso latin g mechanisms. Geography and ecology appear to be the main factors responsable for preventing hybridization and maintaining species integrity, which seem to be also the principal isolating factors in Erigeron of the Juan Fernandez Islands. 5 0 Literature Cited ANDERSON, L. C., D. W. KYHOS, T. MOSQUIN, A. M. POWELL and P. H. RAVEN. 1974. Chromosome numbers in Compositae. IX. Haplopappus and other Astereae. Amer. J. Bot. 61:665-671. ARNOLD, E. N. 1981. Estimating phylogenies at low taxonomic levels. Z. Zool. Syst. Evolut.-forsch. 19:1-35. BATE-SMITH, E. C. 1962. The phenolic constituents of plants and their taxonomic significance. I. Dicotyledons. Bot. J. Linn. Soc. 58:95-73. BENTHAM, G. 1873* Notes on the classification, history and geographic d istrib u tio n of Compositae. Bot. J. Linn. Soc. 13:335-577. BISHOP, H. J. 1982. Criteria for the determination of the direction of character state changes. Zool. J. Linn. Soc. 74:197-206. CARLQUIST, S. '1974. Island biology. New York: Columbia University Press. 51 CARR, G. D. 1985. Habitat variations in the Hawaiian Madiinae (Heliantheae) and its relevance to generic concepts in the Compositae. Taxon 34:22-25. —— and D. W. KYHOS. 1981. Adaptive radiation in the Hawaiian silversword alliance (Compositae-Madiinae). I. Cytogenetics of spontaneous hybrids. Evolution 35:543-556. a n d . 1986. Adaptive radiation in the Hawaiian silverword alliance (Compositae - Madiinae). II. Cytogenetics of artificial and natural hybrids. Evolution 40:959-976. CRAWFORD, D. J. 1978. Flavonoid chemistry and angiosperm evolution. Bot. Rev. 44:431-456. — —, T. F. STUESSY and M. SILVA. 1986. Leaf flavonoid chemistry and the relationships of the Laotoridaceae. PI. Syst. Evol. 153:133-139. —— , and M. SILVA. 1987. Allozyme divergence and the evolution of Dendroseris (Compositae: Lactuceae) on the Juan Fernandez Islands. Syst. Bot. 12:435-443. , — — and — —. 1988. Allozyme variation in Chenopodium sanctae-olarae. an endemic speoies of the Juan Fernandez Islands, Chile. Bioohem. Syst. Eool. 16:279-284. 52 , r. WHITKUS and T. F. STUESSY. 1987. Plant evolution and speciation on Oceanic Islands. Pp. 183-199 in: D ifferentiation patterns in higher plants. ed. K. M. Urbanska. London: Academic Press. CRISCI, J . V. and T. F. STUESSY. 1980. Determining prim itive character states for phylogenetic reconstruction. Syst. Bot. 5:112-135. CRONQUIST, A. 1947* Revision of the North American species of Erigeron. Brittonia 6:121-302. De JONG, R. 1980. Some trends for evolutionary and phylogenetic studies. Z. Zool. Syst. Evolut.-forsch. 18:1-23. DIERS, L. 1961. Der Anteil an Polyploiden in den Vegetationsgurteln der Westkordillere Perus. Z. Bot. 49:437-488. D0N0GHUE, N. J. and P. D. CANTINO. 1984. The logic and lim itatio n s of the outgroup substitution approach to dadistic analysis. Syst. Bot. 9:192-202. FARRIS, J . S. '1970. Methods fo r computing Wagner tre e s. Syst. Bot. 19:83-92. 53 GANDERS, F. R. and K. M. NAGATA. 1984. The role of hybridization in the evolution of Bidens on the Hawaiian Islands. Pp. 179-184 in Plant biosystematios. ed. W. F. Grant. Toronto: Academic Press. GARDNER, R. C. 1976. Evolution and adaptive radiation in Lipochaeta (Compositae) of the Hawaiian Islands. Syst. Bot. 1:383-391. GIANNASI, 0. E. 1978. Systematic aspects of flavonoid biosynthesis and evolution. Bot. Rev. 44:339-429. — and D. J. CRAWFORD. 1986. Biochemical systematica. II. A reprise. Evol. Biol. 20:25-248. GILLET, G. W. and E. K. S. LIM. 1970. An experimental study of Bidens (Asteraceae) in the Hawaiian Islands. Univ. California Publ. Bot. 56:1-63. GORNALL, R. J. and B. A. BOHM. 1978. Angiosperm flavonoid evolution: a reappraisal. Syst. Bot. 3:353-368. , and R. DAHLGREN. 1979. The distribution of flavonoids in the angiosperms. Bot. Not. 132:1-30. HARBORNE, J. B.' 1967* Comparative biochemistry of the flavonoids. Aoademic Press: New York. 5 4 —— 1972. Evolution and function of flavonoids in plants. Recent Adv. Phytochem. 4:107-141. —— . 1977. Flavonoids and the evolution of the angiosperms. Biochem. Syst. Ecol. 5:7-22. HELENURM, K. and F. R. GANDERS. 1985. Adaptive radiation and genetic d iffe re n tia tio n in Hawaiian Bidens. Evolution 39:753-765. HENNIG, W. 1966. Phylogenetic systematica. Translated by D. D. Davis and R. Zangeri. Urbana: University of Illin o is Press. HOLMGREN, I. 1919. Zytologische Studien uber die Fortpflanzung bei den Gattungen Erigeron und Eupatorium. Kongl. Svenska Vetenskapsakad. Handl. 59(7):1-118. JANSEN, R. K. and T. F. STUESSY. 1980. Chromosome counts of Compositae from Latin America. Amer. J . Bot. 67:595-594. J0H0W, F. 1896. Estudios sobre la flora de las Islas de Juan Fernandez. Santiago: Imprenta Cervantes. KANETA, M., H. HIKICHI, S. ENDO and N. SUGIYAMA. 1978. Id en tificatio n of flavone's in sixteen Compositae species. Agric. Biol. Chem. 42:457-477* 55 KEIL, D. and T. F. STUESSY. 1975* Chromosome counts of Compositae from the United S tates, Mexico, and Guatemala. Rhodora 77:171-195. KUNKEL, G. 1957. Uber einige endemische Kompositen der Flora fernandeziana. Ber. Schweiz. Bot. Ges. 67:428-457. LAMMERS, T. G., T. F. STUESSY and M. SILVA. 1986. Systematic relationships of the Lactoridaceae, an endemic family of the Juan Fernandez Islands, Chile. PI. Syst. Evol. 152:243-266. LOWREY, T. K. 1986. A biosystematic revision of Hawaiian Tetramolopium (Compositae: Astereae). Allertonia 4:203-265. — — and D. J. CRAWFORD. 1985. Allozyme divergence and evolution in Tetramolopium (Compositae: Astereae) on the Hawaiian Islands. Syst. Bot. 10:64-72. MABRY, T. J ., K. MARKHAM and M. THOMAS. 1970. The systematic Identifioation of flavonoids. New York: Springer-Verlag. MADDISON, V. P ., M. J. D0N0GHUE and D. R. MADDISON. 1984. Outgroup analysis and parsimony. Syst. Zool. 33:83-103* NESOM, G. L. 1980. A revision of the eppapose species of Erigeron (Asteraoeae - Astereae). Ph.D. disseration, University of North 5 6 Carolina, Chapel Hill. PACHECO, P ., D. J . CRAWFORD, T. F. STUESSY and M. SILVA. 1985. Flavonoid evolution in Robinaonia (Compositae) of the Juan Fernandez Islands. Amer. J. Bot. 72:989-998. ROHLF, J . , J. KISHPAUGH, and D. KIRK. 1972. NT-SYS numerical taxonomy system of multivariate statistical programs. Stony Brook: State University of New York. SANDERS, R. W., T. F. STUESSY and C. MARTICORENA. 1982. Recent changes in the flora of the Juan Fernandez Islands, Chile. Taxon 31:284-289. -——, ——- , and M. SILVA. 1987. Phytogeography and evolution of Dendroseris and Robinsonia. tree-Compositae of the Juan Fernandez Islands. Opera Bot. 92:195-215. ______— and R. RODRIGUEZ. 1983. Chromosome numbers from the flora of the Juan Fernandez Islands. Amer. J. Bot. 70:799-810. SKOTTSBERG, C. 1922. The phanerogams of the Juan Fernandez Islands. Pp. 95-240 in Nat, h i s t . Juan Fernandez. 2: Botany, p art 2. Uppsala: A'lmqvist & Wiksells. . 1953. The vegetation of the Juan Fernandez Islands. Pp. 57 792-960 in Nat, h is t. Juan Fernandez. 2: Botany, p art 6. Uppsala: Almqvist & Wiksells. - —— . 1956. Derivation of the flora and fauna of Juan Fernandez and Easter Island. Pp. 193—437 in Nat. h ist. Juan Fernandez. 1: Geography, geology, origin of island life , part 5. Uppsala: Almqvist & Wiksells. SNEATH, P. H. A. and R. R. SOKAL. 1973. Numerical taxonomy. San Francisco: W. H. Freeman and Co.. SOLBRIG, 0. T. 1962. The South American species of Erigeron. Contr. Gray Herb. 191:3-79. , l . C. ANDERSON, D. W. KYHOS and P. H. RAVEN. 1969* Chromosome numbers in Compositae VII: Astereae III. Amer. J. Bot. 56:348-353• , ——- , — —, —— and L. RUDENBERG. 1964. Chromosome numbers in Compositae. V. Astereae. II. Amer. J. Bot. 51:513-519* SPOONER, D. M., T. F. STUESSY, D. J . CRAWFORD and M. SILVA. 1987. Chromosome numbers from the flora of the Juan Fernandez Islands. II. Rhodo'ra 89:351-356. STEVENS, P. F. 1980. Evolutionary polarity of oharacter states. Ann. 58 Rev. Ecol. Syst. 11:333-358. STUESSY, T. F. and D. J. CRAWFORD. 1983. Flavonoids and phylogenetic reconstruction. PI. Syst. Evol. 143:83-107. ----- , and C. MARTICORENA. 1989a. Patterns of phylogeny in the endemic vascular flo ra of the Juan Fernandez Islands, Chile. Syst. Bot. (in press). and J. V. CRISCI. 1984. Problems in the determination of evolutionary directionality of character-state change for phylogenetic reconstruction. Pp. 71-87 in Cladistics: Perspectives on the reconstruction of evolutionary history. eds. T. Duncan and T. F. Stuessy. New York: Columbia University Press. , K. FOLAND, J. SUTTER, R. W. SANDERS and M. SILVA. 1984. Botanical and geological significance of potasslum-argon from the Juan Fernandez Islands. Science 225:49-51. , c . MARTICORENA, D. J . CRAWFORD and M. SILVA. 1989b. Plant conservation in the Juan Fernandez Islands. In Plant conservation in ceanio islan d s, ed. D. Bramwell. London: Academic P ress, (in p ress). SWOFFORD, D. L. 1985. PAUP: Phylogenetic analysis using parsimony version 2.4. Champaign: Illinois Natural History Survey. 5 9 TURNER, B. L ., W. L. ELLISON and R.M. KING. 1961. Chromosome numbers in the Compositae. IV. North American species with phyletic interpretations. Amer. J. Bot. 48:216-223. —— , A. M. POWELL and J . CUATRECASAS. 1967• Chromosome numbers in Compositae. XI. Peruvian species. Ann. Missouri Bot. Gard. 54:172-177- ——- , ----— and R. M. KING. 1962. Chromosome numbers in the Compositae. VI. Additional Mexican and Guatemalan species. Rhodora 64:251-271. VALDEBENITO, H., T. K. LOWREY and T. F. STUESSY. 1986. A new species of Erigeron (Compositae: Astereae) from Chile. Brittonia 38:1—3- VENABLE, D. L. and D. A. LEVIN. 1983* Morphological dispersal structures in relation to growth habit in the Compositae. PI. Syst. Evol.'143:1-16. VICUNA MACKENNA, B. 1883. Juan Fernandez. Historia verdadera de la is la de Robinson Crusoe. 2 vols. Santiago. WAGNER, W.1 1980. Origin and philosophy of the groundplan divergence method of cladistios. Syst. Bot. 30:1-11. 6 0 WATROUSi L. and Q. WHEELER. 1981. The out-group comparison method of character analysis. Syst. Zool. 30:1-11. WOODWARD, R ., Jr. 1969• Robinson Crusoe^ Islan d . Chapel H ill: University of North Carolina Press. YOUNG, D. A. 1981. The usefulness of flavonoids in angiosperm phylogeny: some selected examples. Pp. 205-232 in Phytochemistry and angiosperm phylogeny. eds. D. A. Young and D. S. Siegler. New York: Praeger S c ie n tific . Chapter II A NEW SPECIES OF ERIGERON (COMPOSITAE: ASTEREAE) FROM THE JUAN FERNANDEZ ISLANDS, CHILE In the flora of the Juan Fernandez Archipelago, the genus Erigeron is represented by five endemic species (Skottsberg, 1956). Four are restricted to the younger island, Masafuera: Erigeron ingae Skottsb., E. luteoviridis Skottsb., E. rupioola Phil., and E. turricola Skottsb. The fifth, E. fernandezianus (Colla) Harling, is found on Masafuera and on the older island, Masatierra. During expeditions in January and February of 1984 and 1986, it became clear that several collections of Erigeron from Masafuera represented a new taxon worthy of formal recognition. Erigeron stuessyi Valdebenito, sp. nov. (Fig. 8) Herbae suffruticosae perennes caespitosae, 9-20 cm altae. Folia rosulata, spathulata, apice acuta, pubescentia, margine Integra vel serrato-dentata, 5-7 cm longa, 0.7-1*2 cm lata. Capitula laxe paniculata 3-6 florlbus, pedunoulis tenuibus villosis 4-6 cm longis. Involucrum turbinatum, 7-8 mm altum et 9-10 mm diametro. Bracteae involuorum 4 mm longae, 1 mm latae, lanoeolatae, aoutae, dorso dense 61 62 hirsutae; bracteae interiorae subglabrae, margine dense ciliatae. Flosculi radiorum 35-42; ligulae albae tridentatae 4 mm longae, tubo 2 mm longo. Flosculi discorum 40-48, hermaphroditi; corolla tubulosa, 4-5 mm longa, 5-lobult, lutea. S ty li 1-1.5 mm longi, ad apicem attenuati pubescentes. Pappus albus, 2.1 mm longus. Achaenia compressa 1 mm longa, marginibus pubescentibus. Chromosomatum numerus n = 27. Caespitose, perennial herb, 9-20 cm ta ll. Stems few from the base, slender, woody, brown, 2-3 mm diam, forming dense semiglobose mats, the bases covered with old leaves. Leaves spathulate, highly pubescent, mainly rosulate at the end of the stems, blades 5-7 cm long, 0.7-1.2 cm wide, at apex broadly acute, with margins entire, occasionally dentate toward the tip; petiole gradually expanding into the lamina, 1-1.6 cm long, 1-2 mm wide, pubescent. Capitulescence paniculate, with 3-6 heads; peduncles up to 8 cm long, pubescent, leafless or with a few short bracts 1-1.5 cm long; involucre turbinate, 7-8 mm ta ll, 9-10 mm diam; phyllaries 38-44, imbricate, 2-seriate, lanceolate, 4 mm long, 1 mm wide pubescent; outer bracts hyaline with dark brown midrib, and with margin pubescent. Ray florets 35-42, uniseriate; corollas white, with limb 4 mm long, 1 mm diam, and tube 2 mm long. Diso florets 40-48, hermaphroditic; corollas narrowly funnelform, 5-lobed, yellow, 4-5 mm long, 0.8 mm diam; style 1-1.5 mm long, with stlgmatio branches tapering gradually towards the tip, covered entirely by collecting hairs. Pappus white, 2.1 mm long. Aohenes compressed, 1 mm long pubesoent along the margins. 6 3 Fig. 8. Erigeron stueaayi. A. Habit. B. Detail of a leaf. C. Head. D. Ray floret. E. Disc floret. F. Style branch of disc floret. A, Ruiz & Lammers 8019. B-F, Landero & Ruiz 9247. V'M F1g. 8 65 Typus: JUAN FERNANDEZ. Masafuera: Cordon Atravesado down into Quebrada Vacas, east side of upper part of steep valley, 950 m, 25 Jan 1986; A. Landero & E. Ruiz 9247 (Holotype: OS). Additional specimens examined: Juan Fernandez. Masafuera: Quebrada Casas, 250 m, 15 Jan 1986, E. Ruiz & T. Lammers 8019 (OS), 218 m, P. Pacheco & E. Ruiz 6401 (OS). Erigeron stuessyi is most similar to E. rupicola from which i t . differs in vestiture, leaf color, leaf and capitulum size, and stigmatic branches of the disc florets. Erigeron stuessyi has pubescent leaves, bracts and stems, which contrasts strongly with pubescence lacking in E. rupicola. Further, in herbarium collections E. stuessyi has dark green leaves whereas in E. rupicola they appear light brown. In Erigeron stuessyi the leaves are 5-7 cm long, with margins entire, occasionally dentate, and with apex acute. In contrast, E. rupicola has smaller leaves (3*5-5 cm long), with margins entire and the apex obtuse. Capitula are arranged in a panicle in E. stuessyi: they are solitary in E. rupicola. The involucre is 7-9 mm diam and 4.5-5.5 mm tall in E. rupicola whereas in E. stuessyi it is 9-12 mm diam and 6-7 mm ta ll. Stigmatic branches are triangular and tipped with oolleoting hairs covering only the tip in E. rupicola whereas in E. stuessyi they gradually taper towards the tip and are completely covered by the h a irs. In addition to morphological differences, ecological separation also occurrus between Erigeron stuessyi and E. rupicola. The former 6 6 occurs a t altitudes of 850 to 1200 m in Cordon Atravesado, a region covered by fog; and also at 200 m in Q. Casas. The latter is a common species on the coastal cliffs from sea level to 300 m. It is a pleasure to name this species of Erigeron for Dr. Tod F. Stuessy in recognition of his contributions to the knowledge of the Juan Fernandez flora. Literature Cited Skottsberg, C. 1956. Derivation of the flora and fauna of Juan Fernandez and Easter Island. Nat. Hist. Juan Fernandez and Easter Is. 1: 193-437. Chapter III EVOLUTION OF PEPEROMIA (PIPERACEAE) IN THE JUAN FERNANDEZ ISLANDS, CHILE INTRODUCTION Evolutionary phenomena in flowering plants are more easily investigated' on oceanic islands than in continental areas. These small areas contain unique plants which are often quite different from mainland relatives. A number of studies have established the potential of understanding phylogeny and evolutionary processes on island biotas, principally in the Galapagos Islands, the Hawaii archipelago, and the Canary Islands (e.g. C arlquist, 1974; Kunkel, 1976; Gorman, 1979; .Carr and Kyhos, 1986). One of the best oceanic archipelagos in which to study patterns of phylogeny and modes of speciation is the Juan Fernandez Islands (33°S 80°W). There are several reasons for this suitability (Stuessy et al. 1984): 1) i t is separated only 600 km from the most likely source of new propagules in Southern South America; 2) i t has only two main volcanic islands, Masatlerra (MT) which is known to be 4 million years (my) old, and Masafuera (MF) which is 2 my; and 3) these islands have a high degree of endemism, more than 69$ of the native angiosperm flora 6 8 (Skottsberg, 1956). In spite of this suitability for interpreting evolutionary patterns and processes, only a few in-depth studies have been done in the islands, and these have dealt principally with the Compositae (Pacheco e t a l., 1985; Crawford et a l ., 1987; Sanders e t a l., 1987). Other families in the archipelago, however, also offer good oportunities for studying evolutionary phenomena. The family Piperaceae is one such groups, having endemics in both islands as well as one species occurring in the islands as well as in the continent. Piperaceae in the Juan Fernandez Islands are represented by four species of Peperomia (Ruiz and Pavon, 1794): P.. berteroana Miq. (MT, MF), P. fernandflsiana Miq. (continental Chile, MT, MF), .P. maraaritifera Hook. (MT), and P.. skottsberaii C.DC. (MF). Previous work on these taxa has focused on the taxonomy (Skottsberg, 1947) and phytogeography (Skottsberg, 1946, Valdebenito, in prep.). However, no studies have examined the processes of evolution in the form of explicit analysis of phylogeny or modes of speciation. The purposes of this paper, therefore, are to: (1) determine the taxa in South America or elsewhere most closely related to the island endemics; (2) estimate the number of introduction(s) from which the insular species originated; (3) reconstruct the phylogeny of the speoies on the arohipelago, and (4) interpret in a preliminary way the likely modes of speciation. Morphology, oytology and flavonoid ohemistry were used to achieve these objectives. 6 9 MATERIALS AND METHODS The morphological data used to determine phenetic and cladistic relationships in Peperomia were obtained from consultation of herbarium m aterial (BM, CONC, F, GB, GH, MO, NY, and US). In addition, fie ld work in the Juan Fernandez Islands was conducted in Jan-Feb and Nov-Dee 1980, and Jan-Feb 1984 and 1986. Plant samples from the continent were also collected during an expedition to Bolivia and Peru in Jan-Feb 1987. Sixty-five populations from a ll four species of Peperomia were sampled from both islands, and 14 collections were made in continental South America. PHENETIC AND CLADISTIC ANALYSES Phylogenetic reconstruction of the Juan Fernandez endemic taxa of Peperomia and those elsewhere required first a determination of the taxa most closely allied to those in the islands. To ascertain which group(s) were most closely related, a survey of the genus was carried out in different'herbaria (F, GH, MO, NY, US). Due to the large size of the genus and polymorphic nature of many species, the recognition of subgenera and seotlons has been problem atical. Mlquel (1843-44), Candolle (1869)> Dahlstedt (1900), Trelease (1930), Trelease and Yuncker (1950) and Yunoker (1974) have been the prinoipal workers in the genus. They have recognized a variable number of subgeneric groups based on vegetative features (e.g., leaf shape arrangement) as well as reproductive features (e.g., infloresoence and fruit morphology). In the present paper, the classification system of Dahlstedt (1900) was followed. In spite of being based only on species of Peneromia from Central and South America, in my opinion i t represents the most stable and adequate classification of the genus. Nine subgenera are recognized, and in addition, the subgenus Hawaiia described by Yuncker (1933) can be added. Considering fru it structure (fig. 9) as well as vegetative features (table 8), eight of these ten subgenera, Acrooarpidium. Erasmia, Mioropiper, Ogmooarpidium. Panicularia. Pleurooarpidium. Ryncophorum, and Hawaiiana cannot be considered as related to the insular species, grouped here under subgenus Tildeniidium (Skottsberg, 1947). Members of this subgenus have subglobose fruits with a short oblique apex and subapical stigma (figs. 10-13), a condition not found in the other eight subgenera. In Panicularia the drupe is somewhat of the same type (fig. 9-f) as in the insular species; except that it is stalked. Furthermore, there is differentiation in the general habit (e.g., peltate leaves) of taxa of subgenus Panioularia. and in a terminal conspicuous umbellate inflorescence, very unlike the lanceolate leaves and paniculate inflorescences of the insular species. The remmaining two subgenera sensu Dahlstedt, i.e ., Sphaerooarpidium and Tildenla. however, do seem to be more closely related to the endemic species of Peperomia of the Juan Fernandez Islands. The phenetic and oladistic analyses, therefore, fooused on these two subgenera. One aspect of the selection and use of characters in the phenetic and oladistic analyses requires discussion. Most of the continental 71 Fig. 9. a-k, Fruits of subgenera of Peperomia. a Acrooarpidium [P. hispidula (Hutchison & Bennett 4747)].- b Erasmia [P. lanoifolia (Cuatrecasas 22135)].- c Hawaiiana [JP. hesperomanni (Heller 2826)].- d Mioropjper [P> quadrifolia (Nee 23314)].- e Qgmocarpldium [P.. pellucida (Alverson, White & Shepherd 3 68)].- f Panicularia [P.. fra s e ri (Plowman & Alcorn 14327)].- g Plerooarpidium [P.. tenella (Cuatrecasa 22207)].- h Ryncophorum [P.. distachya (Schunke 4896).- i Sphaerocarpidium [P. galioides (Nee 22933)].- j Tildenia [P. umbilinai-.a(Pringle 13181)].- k Tildeniidiun [P. berteroana (Skottsb. 96)]. (a ll X30) 7 2 F1g. 9 Table 8. Summary of distinguishing charactersistics of the subgenera of ter Dajil-stgdt_(_19pO) and Yuncker (1933).______' r - 5 UB5 ENEPA Character Ac roc air pi di um Erasmi a Hawaii ina Durat ion Annual Perennial Annual/perennial Habit Erect/prostrate Erect/prostrate Ascending Leaf A lternate. A lternate Vert ic illa te arrangement Inflorescence Terminal/axillary; Terminal/axillary; T erm in al/axil1ary; sim ple simple/paniculate sim p le. with basal bract Fruit Ovoid, covered with Elongate, cylindrical, Ovoid to obovoid; hairs, stipitate, with longitudinal ribs; apex pointed, with apical stigma on a apex oblique, broad and two subapical slightly oblique flattened, with stigma stigmas (Fig 9-c). appendage (Fig. at base of apex (Fig. 9 - a ) . 9 -b ). Tabic 8 continued SUBGENERA Character Micropiper Ogmocarpidiurn Panicularia Durat ion Perennial Annual Perennial Habit Ascending/decumbent Ascending Erect Leaf A ltern ate/ A lternate A ltern ate/ arrangement verticillate verticillate Inflorescence Terminal/axillary Terminal/axillary Terminal; pan ic u la te / umbellate Fruit Drupe gradually Rounded, with Ovoid-el1ipt ic, tapering to a longitudinal ribs; convex or conical; straight apex; stigma on an apical apical disc; apical stigma; appendage (Fig. 9-e) simple apical presence of pseudo- stigma (Fig. 9-f). cupule (Fig. 9-d). Table 8 continued SUB6ENERA" Character PIeurocarpidiurn Rhyneophorum Sphaerocarpidium Durat ion Annual Perennial Annual/perennial Habit P rostrate Erect/decumbent Erect/prostrate Leaf A lternate A ltern ate/ A ltern ate/ arrangement vert ic illa te verticillate Inflorescence Terminal; Terminal/axillary; Terminal/axillary; s im p le simple simple/paniculate with basal bract Fruit Obovate, laterally Qbovoid, with apical Ovat e-rounded, ribbed, with basal process; stigma at the with oblique apex; pedicel equal to the base of the process stigma lateral length of the drupe; (Fig. 9-h). (Fig. 9—i). apical stigma on a conico-curvate appendage (Fig. 9 - g ) . Table 8 continued SUBGENERA Character Til deni a Til deniidium Durat ion Annual/perennial Perennial Habit Erect/prostrate Erect ascending Leaf A l t e r n a t e A l t e r n a t e / arrangement vert icill ate Inflorescence Terminal/axillary; Terminal/axillary; simple/paniculate paniculate with with basal bract basal bract Fruit Ovate; apical Ovoid-subglobose, stigma on a conical oblique apex; or discoid apex stigma subapical Figs. 10 -13. Fruits of endemic species of Peperomia in the Juan Fernandez Islands; SEM.~ Fig. 10. JP. fernandeziana Miq. (Stuessy, Matthei, Sanders and Valdebenito 5363).- Fig. 11. P. berteroana Miq. (Skottsb. 96).- Fig. 12. £. margaritifera Bert. (Skottsb. 220).- Fig. 13. P.. skottsbergii C.DC. (Stuessy, Sanders and Rodriguez 5037). (all SEM X120) taxa present either a broad range of morphological variation or invariant characters. For example, 57 characters were initially selected for inclusion in the phenetic analysis, but several of them were either polymorphic characters (e.g., number of shoots, leaf shape, leaf and shoot pubescence, leaf arrangement, length of inflorescence, number of flowers per inflorescence) or invariant (e.g., floral features such as position, shape and texture of the floral bract; structure of the stigma and stamen; and relative position of the ovary). The same is true in the phylogenetic analysis. For example, inflorescence arrangement, anther shape, drupe shape, number of leaf nerves, and rooting system have many states within at least some of the taxa. This kind of v a ria b ility , which is common in Peperomia (Burger, 1971; Yuncker, 1933), causes d iffic u lty in selecting characters and states for quantitative analysis of any kind. The problem was resolved mainly by eliminating such characters, or in some instances (e.g., inflorescence arrangement) by coding the most prevalent state. Phenetic analysis To construot a basic data matrix for phenetic analysis, twenty-seven characters (nineteen vegetative and eight reproductive) were measured on 16 OTUs representing subgenera Sphaerocarpidium, Tildenia, and the insular species. Mean values of 10 measurements of three specimens' from each taxon were used to represent the range of morphological variation. The 16 OTUs are: Peperomia alata R. & P. [ALAT]; P. berteroana Miq. [BERT]; P. blanda (Jacq.) H.B.K. [BLAN]; P. 8 0 cillobraoteata C.DC. [CILI]; P. coouimbensis Skottsb. [COQU]; P.. d o e llli P hil. [DOEL]; P.. fernandeziana Miq. [FERN]; P.. galioides H.B.K. [GALI]; P.. h irsu ta C.DC. [HIRS]; P.. lignescens C.DC. [LIGN]; P. m argaritifera Bert, ex Hook [MARG]; P.. martiana Miq. [MART]; P.. mexioana Miq. [MEXI]; P.. nummularioides Griseb. [NUMM]; P. petrophila C.DC. [PETR]; P. sk o ttsb erg ii C.DC. [SKOT]. The characters and their states are enumerated in Table 9* The basic data matrix is available upon request from the senior author. The data matrix was analyzed using numerical techniques included in the NT-SYS package of Rohlf, Kishpaugh, and Kirk (1972). Data were standardized and both correlation and distance matrices were computed. Cluster analysis was performed on the matrix using the unweighted pair group method with arithmetic means (UPGMA). The cophenetic correlation coefficient, which measures the amount of distortion of the phonogram from the matrices, was computed after the cluster analysis. Computations were carried out at the Instructional Research Computer Center of The Ohio State University. Cladistio analysis The interpretation of phylogenetic relationships in Peperomia has been based primarily on 15 morphological characters divided into two dlsorete states (table 10). The basic data matrix is shown in Table 11. Different orlteria for hypothesizing evolutionary directionality of character states have recently been reviewed by several workers 81 Table 9* Characters and character states used In the phenetic analysis of Peperomia. Numbers or letters in parentheses indicate the qualitative states or scales for quantitative measurements. Stem. 1. Orientation: prostrate (0), semierect (1), erect (2). 2. Succulence: succulent (0), very succulent (1). 3* Texture: smooth (0), verrucose (1). 4. Branching pattern: monopodial (0), sympodial (1). 5. Vestiture: glabrous (0), pubescent (1). 6. Diameter (mm). 7- Height (cm). Leaves. 8. Petiole length (cm). 9• Arrangement: alternate (0), verticillate (1). 10. Shape: spathulate (0), ovate to ovate-lanceolate (1), elliptic-lanceolate (2). 11. Venation: one main vein (0), two to three (1), four or more (2). Upper leaves. 12. Length (cm). 13* Width (cm). Lower leaves. 14. Length (cm). 15. Width (cm). Leaf vestiture. 16. Upper surface: glabrous (0), pubescent (1). 17* Lower surface: glabrous (0), pubescent (1). Inflorescence. 18. Flower abundance: scarce (0), abundant (1). 19* Spike configuration: peduncle I only (0), peduncle I and I I (1). 20. Peduncle Length I (cm). 21. Length peduncle I I (cm). 22. Spike ram ification: so lita ry (0 ), two to four (1). 23. Length of complete inflorescence (cm). 24. Spike arrangement: terminal (0), lateral (1), terminal/axillary (2). Fruit. 25. Stigma position: apical (0), subapical (1), lateral (2). 26. Position of apex: erect (0), oblique (1). 27. Habitat: terrestrial (0), epiphytic (1)> terrestrial/epiphytic (2). 8 2 Table 10 . Characters and their states (with numerical assignments) used In the cladlstlc analysis of Peperomia. Character states Primitive (0) Derived (1) No Characters 1 Stem surface Smooth Verrucose 2 Stem configuration Sympodial Monopodial 3 Stem diam Under 8 mm 8 mm or over 4 Leaf persistence Long persistent Deciduous 5 Leaf arrangement A lternate Vertieillate 6 Leaf attachment Petiolate Subsessile 7 Inflorescence arrangement Simple Umbellate/Cymose 8 Inflorescence position Axillary Terminal 9 Inflorescence basal bract Absent Present 10 Stigma length Short Large 11 Trace of second stigma Absent Sometimes present 12 Stigma position Lateral Subapical 13 Position of ovary apex Oblique Slightly oblique 14 Persistence Annual Perennial 15 Habitat Epiphytic Terrestrial 83 Table 11. Data matrix of characters and states in species of Peperomia for cladistic analysis. Names of taxa described by the first four letters of the specific epiphets (cf. Materials and Methods). Taxa Characters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ANCE 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MARG 1 1 0 0 0 0 1 0 1 0 0 1 1 1 BERT 0 1 1 • 1 1 1 1 0 1 1 1 1 1 1 SKOT 0 1 0 0 0 0 1 1 1 0 0 1 1 1 8 4 (e.g. C risci and Stuessy 1980; Stevens, 1980; Watrous and Wheeler, 1981; Bishop, 1982; Maddison e t a l. 1984; Donahue and Cantino, 1984). The out-group comparison method has been relied upon to assess the apomorphlc state of each character. I t has been assumed th a t the species of Peperomia found In the Islands have been derived from continental relatives. According to the phenetic analysis as well as additional morphological and comparative studies, representatives of subgenus Sphaerocarpidium Dahlstedt (Section Alternifoliae) were judged to be the most lik e ly outgroup taxon. The final step in the phylogenetic analysis of Peperomia was cladogram construction. All data were analyzed by PAUP version 2.4 (Swofford, 1985), which infers phylogeny by maximum parsimony. The program was run using CLOSEST, ORDERED ALL, BANDB, R00T=OUTGROUP which calculates the shortest tree, orders the character states, finds all the most parsimonious trees, and roots the insular taxa by using Section Alternifoliae (subgenus Sphaerocarpidium) as the outgroup. Flavonold analysis All plant samples were colleoted during four expeditions to the Juan Fernandez Islands in Jan-Feb and Nov-Dee 1980, and Jan-Feb 1984 and 1986. Vegetative parts of 51 populations from all four species of Peperomia were sampled from both islands (Figs. 14 and 15). In addition, two populations of JP. fernandeziana and one population of P. coquimbensis were colleoted in continental Chile. Samples o f Peperomia alata. P.. galioides and P. olens also were collected from Bolivia. 8 5 Fig. 14. Collections of Peperomia in Masatierra studied for flavonoid chemistry. Q P. berteroana; ^ P . fernandeziana: ^^Z* margaritifera. Numbers refer to collections of taxa listed in table 13. 1 km 8 6 <.... 1 i F ig . 14 8 7 Fig. 15. Collections of Peperomia in Masafuera studied for flavonoid chemistry. Q P. berteroana: ^P,. fernandeziana: | P. skottsberaii. Numbers refer to collections of taxa listed in table 13 • 1 km F ig . 15 8 9 Specimens were a ir-d rie d in the fie ld . Vouchers are on deposit at OS. Flavonoid iso la tio n and purification were conducted using techniques described in Pacheco et al. (1985) and are based on 2-D chromatography and thin layer chromatographic procedures. Identification of flavonoids was accomplished using UV spectral analysis (Mabry et al., 1970). The presence of sulfate in flavonoids was detected initially by their appearance as arrow- or comet-shaped rather than circular spots on 2-D paper chromatography. Their presence was comfirmed by high voltage paper electrophoresis applying 1.1 Kv for one hour using a BIO-RAD Model 1420B high voltage power supply (Al-Khubaizi, 1977). For purifying the sulfated flavonoids, two dimensional paper chromatography combined with polyamide chromatography (TLC) was employed. Standard spectroscopic methods were then applied (Al-Khubaizi, 1977). Chromosome counts Materials for cytological study were obtained during collecting trips to the Juan Fernandez Islands in 1980, 1984 and 1986, and to Bolivia in 1987. Flower buds were fixed in modified Carnoy's solution (4 chloroform: 3 ethanol: 1 glaoial acetic aoid, v/v). Anthers were stained with acetocarmine and squashed in a drop of Hoyer's solution. Chromosome numbers were determined from meiotic microsporooytes via phase-contrast microscopy and documented with camera luoida drawings. Voucher specimens are deposited a t OS. 9 0 F ru it morphology Fruits of the insular species were obtained from personal collections as well as herbarium specimens. In addition, fruits of 50 species of Peperomia. representing the Subgenera Acrocarpidium, Erasmia. Micropiper. Qgmocarpidium, Panicularia. Pleurocarpidium. Rynoophorum. Sphaerocarpidium and Tildenla. sensu Dahlstedt (1900); Tildeniidium sensu Skottsberg (1947) and Hawaiiana sensu Yuncker (1933) were also included fo r comparative purposes. F ruits were mounted on aluminum stubs with silv e r paint. The samples were coated with 250 A of a gold/palladium alloy in a Technics Hummer III Coater, observed at 20kV with an Hitachi S-500 SEM, and photographed with Polaroid Type 55/N film. RESULTS Phenetio analysis Phenetic evaluation was needed to determine the closest mainland relatives of the island endemics of Peperomia so that the phylogeny of the island taxa could be estimated. The results of the cluster analysis, which had a cophenetic correlation coefficient of 0.88, are presented in fig. 16. Analysis of the resulting phenogram reveals two major clu sters-at the -0.380 phenon lev el. The uppermost inoludes the group of endemic species of the Juan Fernandez Islands (Group A) joined first by members of subgenus Sphaerocarpidium Seotlon Alterniifoliae 91 Fig. 16. Phonogram of species of Peperomia using UPGMA. Dots indicate species present in the Juan Fernandez Islands. Scale indicates level of correlation. Capital letters correspond to subgroupings (see text). 9 2 PER • 8K0 • MAR • ALA B MAT CIL LIO PET £ MEX FER • COQ DOE QAL NUM JL BLA F HIR T - 0.4 o 0.4 0.8 F ig . 16 93 (Group B) and ultimately by taxa of subgenus Tildenia. section Eutildenia (Group C). The second major cluster includes taxa of subgenus Sphaerocarpidium (Section Verticillatae) with the exception of P. nuTnmiiiflfioides (subgenus Mloropiper) (Group E). Peperomia fernandeziana. present in both islands as well as in continental Chile, falls within Group D with other Chilean species (e.g., P. coouimbensis and JP. d o e llii) as well as P. galioides which is widely d istrib u te d in South America. Finally, Group F consists of taxa belonging to subsection Platyphyllae within section Verticillatae. With regard to the closest relatives of the insular endemics, therefore, these results indicate members of section Alternifoliae, subsections Macrophyllae and Microphvllae (e.g., P. alata and P.. martiana. respectively) as the most likely allies. Cladistic analysis From the phenetic studies, it is clear that the four species of Peperomia in the Juan Fernandez archipelago are separated into two groups: Peperomia berteroana. £. margaritifera and P. sk o ttsb e rg ii: and P. fernandeziana. I t was assumed, th erefo re, that they represent independent phyletlc lines, and therefore the three-species group was treated separately in the cladistic analysis. Employment of the PAUP program resulted in generation o f one most parsimonious tree (fig. 17) involving 15 oharacter state changes. The Consistency Index (Cl), which is the ratio of the oharacter state 94 Fig. 17* Cladogram o f hypothetical evolutionary relationships among endemic species of Peperomia in the Juan Fernandez Islands. For description of characters see table 10. Bars indicate synapomorphies. 95 MAR BER SKO 10 12 13 14 Fig. 17 changes without homoplasies to the total number of character state changes, was 1. This high value of Cl reflects the presence of autapomorphies (uniquely derived character states) and synapomorphies only. These results suggest a strong differentiation in one main clade of the endemic in su lar taxa from the ancestral stock by six synapomorphies, supporting the option of recognizing subgenus Tildeniidium (Skottsberg, 1947) to include the endemics. Another option would be the establishment of a new section in subgenus Sphaerocarpidium. but a comprehensive monograph of the genus would be required before such an alternative could be justified. One synapomorphy' link P.. skottsbergii and P.. berteroana. and these are differentiated by 1 and 6 autapomorphies, respectively. Peperomia margaritifera showed the fewest derived conditions and may be regarded as the most primitive species of this group in the archipelago. Flavonoid analysis The flavonoids isolated from insular and continental species of Peperomia examined are listed in table 12. The interpopulational occurrence of flavonoids within species in the Juan Fernandez Islands and among species that occur in the continent are shown in table 13 and table 14, respectively. Thirty-eight flavonoids Including 2 flavone aglycones, 6 flavone sulfates, 8 flavone glycosides, 19 C-glycoflavones and 3 flavonol glyoosides were encountered among the insular and continental species of Peperomia. The flavones were based on either 97 luteolin, diosmetin, acacetin or apigenin, and the flavonols on quercetin or kaempferol. Substitutions of the basic compounds encountered were 7-0-m ethylation, 7-0-hydroxylation and 7-0-mono and disulfate groups. This is the first report of flavonoids in the genus as well as the first report of flavonoids with sulfate groups in Piperaceae. Flavones presented either 7-0-mono or 7-0-diglycosides, and the flavonols 3-0-monoglycosides only. The sugars involved are glucose, arabinose, rhammnose and xylose. The only monosaccharide present in flavonols was glucose. In general, the distribution of flavonoids in the species occurring in the islands reveals species-specific profiles as well as greater diversity (table 13) than in the continental species (table 14). Thirty-eight compounds were present in the former and only 18 (with a more sim ilar array of compounds) in the la tte r . Among the insular species, some interpopulational variation is present in all taxa, but it is most pronounced in P> skottsberaii. In addition, there is interpopulational differentiation within the same species between islands (e.g., P. berteroana) as well as between Islands and the continent (e.g., P.. fernandeziana). The greatest variety of substitution patterns was observed in C-glyoosylflavones: mono and di-C-glycosyl derivatives were observed in luteolin, diosmetin, acacetin and apigenin (table 12). These compounds, with the exception of orientin (compound 28) and vitexin (compound 29), were tentatively Identified by resistance to hydrolysis and by comparison of chromatographic and spectral behavior with published data. The small amounts available for study prevented further 98 Table 12. Code numbers and names of flavonoids found in Peperomia. FLAVONE AGLYCONES. 1. Acacetin. 2. Luteolin 7-0-methoxyde. FLAVONE GLYCOSIDES. 3* Acacetin 7-0-dlgalactoside. 4. Diosmetin 7-0-monoglucoside. 5. Diosmetin 7-0-diglueoside. 6. Diosmetin 7-0-diglycoside. 7. Diosmetin 7-0-arabinoglucoside. 8. Diosmetin 7-0-rhamnoglucoside. 9* Luteolin 7-0-monoglucoside. 10. Luteolin 7-0-rhamnoglucoside. C-GLYCOSYLFLAVONES. 11. Acacetin C-monoglucoside. 12. Acacetin C-diglycoside. 13* Acacetin C-diarabinoside. 14. Acacetin C-arabinoglucoside. 15. Acacetin C-arabinoxyloside. 16. Apigenin C-glucoside. 17* Apigenin C-dlglycoside. 18. Apigenin C-diarabinoside. 19* Apigenin C-arabinoxyloside. 20. Diosmetin C-glucoside. 21. Diosmetin C-monoglycoside. 22. Diosmetin C-arabinoglucoside. 23. Luteolin C-monoglycoside. 24. Luteolin C-diglycoside. 25. Luteolin C-diarabinoside. 26. Luteolin C-arabinoglucoside. 27. Luteolin C-arabinogalactoside. 28. Orientin. 29. Vitexin rhamnoside. SULFATE FLAVONES. 30. Aoacetin 7-sulfate. 31* Acacetin 7-glucosidedisulfate. 32. Apigenin 7-glucosidesulfate. 33* Diosmetin 7-sulfate. 34. Diosmetin 7-glucosidesulfate. 35. Luteolin 7-sulfate. FLAVONOLS. 36. Kaempferol 3-methoxy-7-0-glucoside. 37* Kaempferol 3 ,7-0-diglucoside. 38* Quercetin 7-methoxy-3-0-glucoside. I Table 13. Distribution of flavonoids in endemic species of P e p e ro m ia in the Juan Fernandez Islands. Collection numbers those of Stuessy et al. For flavonoid identity see Table 12. ftalvcones filvcosidw C-olvtosvlHavones Sul fated flavones Flavonols C oll. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 IB 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 P. btrttrom Hiq. i featicrra 5067 1 ♦ + ♦ ♦ ♦ ♦ ♦ 5204 2 > ♦ » ♦ ♦ ♦ ♦ 5212 3 ♦ ♦ ♦ t ♦ ♦ ♦ 6332 4 ♦ ♦ ♦ * ♦ 6335 5 ♦ ♦ ♦ ♦ ♦ ♦ f ♦ 6S17 6 ♦ t ♦ ♦ ♦ 6527 7 ♦ f ♦ ♦ + f 6550 8 ♦ ♦ ♦ ♦ ♦ ♦ ♦ + 6391 9 ♦ ♦ ♦ ♦ 4 i ♦ t 6594 10 ♦ ♦ t ♦ ♦ ♦ ♦ ♦ ♦ Ibufwra 5034 II ♦ ♦ ♦ ♦ ♦ ♦ f ♦ ♦ ♦ ♦ 8004 12 ♦ ♦ ♦ ♦ t ♦ ♦ e ♦ ♦ 8200 13 ♦ ♦ ♦ ♦ ♦ ♦ ♦ t * t 8260 14 f ♦ f t ♦ ♦ ♦ * * f 8402 15 ♦ ♦ t ♦ ♦ ♦ + * ♦ ♦ 8491 16 ♦ ♦ t ♦ ♦ * ♦ ♦ ♦ * 8513 17 » ♦ ♦ ♦ ♦ ♦ + ♦ ♦ ♦ sO sO Table 13 contiued FIWDNES flnlvconw S lv to s ite C-olvtosvlflawnes Sul fated flavones Flavcnols P o ll. fcP 1 2 3 4 S 6 7 8 9 10 11 12 13 14 IS 16 17 18 19 20 21 ?• finvndaim Hiq. Ih s itic rra 5097 IB ♦ 4 4 4 5363 19 f ♦ 4 4 4 4 6498 20 4 ♦ 4 4 4 4 6579 21 f t f ♦ » 4 4 4 4 t e a fiora 3045 22 ♦ ♦ ♦ ♦ 4 4 4 4 8093 23 ♦ ♦ ♦ 4 4 4 8101 24 ♦ ♦ f 4 4 4 0184 25 ♦ ♦ t 4 4 4 8211 26 ♦ ♦ ♦ 4 4 4 8213 27 ♦ ♦ 4 4 4 4 8229 28 ♦ ♦ a 4 4 4 8242 29 * * t 4 4 4 8295 30 ♦ ♦ 4 4 4 8309 31 t ♦ f 4 4 4 9015 32 f 4 4 4 9145 33 ♦ 4 4 9184 34 4 4 4 4 4 4 9192 35 4 4 4 4 4 4 9295 36 4 4 4 4 4 4 9342 37 4 4 4 4 4 4 9437 38 4 4 4 4 4 4 9512 39 4 4 4 4 4 4 9560 40 4 4 4 4 4 100 9607 41 4 4 4 4 4 4 9654 42 4 4 4 4 4 Table 1 3 continued FUNKS Aalwoms Slvcosides C-alvcosvl Havanas Sil fated flavones Flavonols m - % 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 IB 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 3B P. trgritifm fc r t. H asatierra 5206 43 4 4 4 4 6541 44 4 4 4 4 P. ikottitergii C.DC. fbsafw ra 63BB 45 4 4 4 4 4 4 4 * 6402 46 ♦ 4 4 4 8159 47 44 4 4 4 4 4 8197 48 4 4 4 8207 49 ♦ 4 4 8272 50 4 4 4 4 4 4 4 4 4 8404 51 4 4 4 4 4 4 4 101 Table 14. Distribution of flavonoids in selected species of Peperomia from South America. For flavonoid identity see Table 12. Collection Flavone flolvcone C-olvtosvlflavones ______Flavonols Species______InuBber locality 1 ______4 5 9 11 14 16 IB IS 20 24 25 26 27 28 36 37 38 P. a/ata W 323 Bolivia ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ * P. coquirbem is POP 1031 Chile ♦ ♦ * ♦ ♦ + P. fernandaim POP 1043 Chile ♦ ♦ ♦ ♦ + ♦ ♦ ♦ ♦ ♦ ♦ PCKM043A Chile ♦ ♦♦♦♦*♦ ♦ ♦ ♦ ♦ P. gal hides VH 350 Bolivia ♦ ♦ ♦ ♦ ♦ ♦ P. olens W! 358 Bolivia ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ F’CNP = Pacheco, Campos, Neira, Pacheco Wl = Valdebenito and Martinez 103 charact eri zat ion. C-glycosyl derivatives and flavone glycosides were present in most of the species analyzed. However, flavonols were mostly restricted to continental taxa whereas sulfated compunds were present only in P.. berteroana and P. skottsbergii (table 13). Chromosome numbers Seventeen new chromosome counts have been made for Peperomia of the Juan Fernandez Islands, plus three from continental Chile and two from Bolivia (table 15). First reports are from Peperomia coouimbensis (n = ca. 23), P. galioides (n = 22 + 2) and P_. olens (n = 22). Other counts confirm previously reported numbers and give cytological information for other conspecific populations in the islands. Chromosome number determinations have been very useful as indicators of relationship and evolutionary directionality (Jackson, 1971; Stuessy, 1971; Lammers, 1988); therefore, these additional counts are very important to document cytological constancy within taxa. The circa reports are accurate to plus or minus one chromosome. DISCUSSION Relatives o f the island endemics It is diffioult to determine the true evolutionary affinities in a genus as large and taxonomioally poorly understood as Peperomia. It is Table 15. Populations of Peperomia (Piperaceae) in the Juan Fernandez Islands and continental South America for which chromosome numbers were determined. Vouchers deposited at OS. Taxon Gametic chromosome Voucher number P. berteroana Miq. **ca. 22 MASATIERRA: Valle V illagra, LRV 6527: Valle Ingles, SCVL 6550 •22-24 MASAFUERA: Q. Casas, SARS 5034: ca. 22 MASAFUERA: Q. del Tongo, LG 8491: W aterfall West of Vicente Porras, VLS 8513. ca. 23 MASAFUERA: Q. Sandalo, VL 8154: Q. Pasto, SL 8402. .P. coQuimbensls Skottsb. ca. 23 CHILE: Parque Nacional Fray Jorge, Los C rista le s, PCNP 1031. fernandeziana Miq. •22+24. MASATIERRA: between Plazoleta del Tunque and El Camote, MAU 9080. ••23+2 MASATIERRA: path from Valle Villagra to Mirador Selkirk, SC 6498 ca. 22 MASATIERRA: Q. Casas, RS 8101: Q. Vaoas, VL 8188: Quebrada Tongo, LR 9651. 22+2 MASAFUERA: Q. Inocentes, VL 8229: Q. Pasto, SL 9184: Q.Sandalo, SL 9274: Q. Larga, SR 9437. SR 9446. SR 9464: Q. Varadero, SGS 9560. 105 Table 15. Continued. Taxon Gametic chromosome Voucher number (a) P. fernandeziana Mia. ca. 22 CHILE: Parque Nacional Fray Jorge, Los C ristales. PCNP 1043. 1043A. P. Kalioides H.B.K. 22+2 BOLIVIA: Dpto. La Paz, Prov. Sud Yungas, VM 350. P. marearitifera Bert •ca.24 MASATIERRA: Q. Villagra S 5206. ex Hook P. olens C.DC. 22 BOLIVIA: Dpto. La Paz, Prov. Sud Yungas, VM 358. P. skottsbereii C.DC. ca.22-24 MASAFUERA: Q. Sandalo, VL 8159. ca. 23 MASAFUERA: Q. Varadero, LR 8272: Q. Pasto. SL 8404. **ca. 23 MASAFUERA: Q. Casas PR 6399 *ca.24 MASAFUERA: Q. Casas. SARS 5037 (a) numbers represent bivalents. * = Sanders et al., 1983. •• = Spooner e t a l ., 1987. LG = Landero & Gaete; L = Landero; LR = Landero & Ruiz; LRV = Landero, Ruiz & Valdebenito; MAU = Marticorena, Arrlagada & Ugarte; PCNP = Pacheco, Campos, Nelra & Pacheco; PR = Parra, Rodriguez; RS = Ruiz & Sepulveda; SARS = Stuessy, Arrlagada, Rodriguez & Sanders; SC = Stuessy, Crawford; SCVL = Stuessy, Crawford, Valdebenito & Landero; SG s Stuessy & Gaete; SGS = Stuessy, Gaete & Sepulveda; SL = Stuessy & Lammers; SR = Stuessy & Ruiz; SS = Stuessy & Sepulveda; SV = Stuessy & Valdebenito; SVLG = Stuessy, Valdebenito, Landero & Gaete; VL = Valdebenito & Lammers; VLS = Valdebenito, Lammers & Sepulveda; VM = Valdebenito & Martinez. 106 one of the largest genera of flowering plants for which a natural c la ssific a tio n is not available. Several factors have caused extreme difficulties in the study of this group: 1) abundance and diversity of species; 2) wide range of morphological variation; 3) absence of clear contrasting characters in some taxa; and 4) the description of new species based on occurrence in distinct political areas without detailed studies on previously described similar taxa. However, in light of the data here acumulated on members of Peperomia. some points can be made regarding relationships of the endemic taxa with those elsewhere in the continent. Two subgenera sensu Dahlstedt have been mentioned as closely related to the in su lar endemics (cf. phenetic and c la d istic analyses): Tlldenia and Sphaerooarpidium. Therefore, it is of interest to examine the relationships of the endemic species of the Juan Fernandez Islands with these two subgenera as well as the evolutionary history and biogeography of the genus. Relationships to Tildenia.— This subgenus, as limited by Dahlstedt, does not appears to be a natural group. Dahlstedt distinguished two sections. Section Hemirhynchophorum can easily be dismissed as anoestral to the endemic Juan Fernandez taxa by being over 35 cm tall, very succulent herbs (vs. under 40 cm tall herbs in the Juan Fernandez Islands), sometimes branched (vs. unbranched), large alternate leaves (vs. medium), leaves lanoeolate to oblong-lanceolate, acuminate at the apex (vs. oblanceolate), and pinnately 7-nerved (vs. palmately 3-5-nerved). The most striking difference is the obovold ellipsoidal 107 f r u it which terminates in a d isc-lik e beak with an apical stigma. Section Eutildenia consists of stemless herbs, leaves round to ovate, usually peltate, forming a tufted cluster. However, a group of species displayed in the cluster analysis (fig. 16, Group C), shows succulent and erect stems, (deciduous leaves downward the stem, alternate, ovate-lanceolate or linear-lanceolate (vs. oblaneeolate, elliptic-lanceolate in Juan Fernandez), long petioles (vs. subsessile to sessile) shows certain affinity with the insular taxa. However, the ovate and ovate-cylindrical fruits, with an erect scutellum and apical stigma in simple inflorescences, argue against a close relationship with the endemic taxa of the Juan Fernandez Islands. Only two species of this subgenus have been counted chromosomally: Peperomia argyreia E. Morr. (2n = 20; Bai and Subramanian, 1964; and 2n = 22; Smith 1966; Dasgupta and Datta 1976); and P. arifolia Miq. (2n = 24; Blot, 1960). Relationships to Sphaerocarpidium.— Of the four species of Peperomia in the Juan Fernandez, P. fernandeziana Miq., which is also found in continental Chile, it is a typical member of subgenus Sphaerocarpidium. The sessile globose-ovoid fruit, with oblique apex, and lateral stigma (fig. 9-i and 10) are important features for including this species in the subgenus. In addition, the erect stem, verticlllate leaves, ovats=elliptic to broadly ovate or almost suborbicular leaves make this species most closely allied to taxa in section Verticillatae. Subsection Plabyphyllae. The remmaining three insular species form a distinct group. They are distinguished principally by their paniculate lnfloresoence and 108 fru it morphology which includes small, pulvinate (p en icillate) stigma at the subapex of the style (figs. 11-13)- In habit, these species also are more related to subgenus Sphaerocarpidium (table 8). Which section of Sphaerocarpidium is closer to the endemics is difficult to assign, but the phenetic results suggests that members of Section Alternifollae are a likely candidate. That the insular species could be related to the f ru it morphology of Sphaerocarpidium at first seems unlikely. However, closer examination of the fruits belonging to taxa within Sphaerocarpidium. shows variation from a typical fruit for the subgenus [e.g., P. galioides Kunth, and P. inaequalifolia Ruiz and Pavon; see Dahlstedt (1900) [lam. XI, fig s. 7 and 6, respectively], to a reduced scutellum and subapical stigma [e.g., P.. blanda (Jacq.) H.B.K., and P.. increscens Miq., see Dahlstedt (1900) lam XI, fig s 4 and 12 respectively] and even a disk-like scutellum with more apical stigma (e.g. J?. rotundifolia (L.) D ahlst., P. acuminata (L.) D ahlst., see Dahlstedt (1900) [lam. XI, fig. 9]* In addition, the presence of thin (e.g., P. alata. P. glabella) to succulent alternate leaves (e.g., JP. acuminata). leaves elliptic-lanceolate to broadly ovate-lanceolate, suberect to erect sucoulent stems, sometimes zig-zag upward (e.g. P. alata), presence of spikes in groups of two or more arranged in more or less paniculate branching clusters (e.g. P. alata. P. acuminata). suggest a close relationship. Regarding 'chromosome numbers, Peperomia rotundifolia. the only species in the ancestral stock which has been cytologlcally studied, exhibits n = 11 (Smith, 1966), offering the possibility that polyploidy 109 accompanied the origin of the new species in the archipelago (n. = 22-24; table 15). However, because no extensive chromosome counts are available for the suspected continental relatives of the endemic taxa, no definite conclusions can be drawn about chromosome evolution into the origin of the island taxa. Another potential problem is that chromosome numbers may vary within taxa. For example, JP. v e rtlc illa ta (L .) A. D ietr. n = 12 (Blot 1960) and P.. galioldes H.B.K. as n = 22, (Huynh 1965) and n = 24 (Diers 1961). Both taxa belong to Section . Verticillatae. Likewise, no clear pattern of relationships can be drawn from the flavonoid data alone. Both sections of subgenus Sphaerocarpidium surveyed for flavonoids show similar flavonoid arrays among them as well as within the insular taxa, with the characteristic of a reduced number of compounds in the continental ones (table 14). In addition, Peperomia a la ta . Seotion A lte rn ifo lia e . has nine flavonoids, of which six are C-glycoflavone derivatives including luteolin, diosmetin, and acacetin. Four of these compounds are also present in P.. galioldes and five in P. olens. members of section Verticillatae. Furthermore, Peperomia alata, and P.. galioldes and P.. olens (section Verticillatae) share a kaempferol glucoside (compound 36) recorded also in insular and continental populations of P. fernandeziana (tables 13 and 14). Phytogeography and evolutionary history of Peperomia.— In addition to the morphological, oytologlcal and flavonoid analyses, present-day geographic distribution patterns and fossil evidence can help in determining evolutionary relationships between the Insular taxa with 110 those in the continent. The only record of pollen for Piperaceae come from Piper in the Eocene of Argentina (Menendez, 1972), and P ip e rite s. from the upper Cretaceous and Miocene of Costa Rica and Trinidad (Trelease, 1930). According to Raven and Axelrod (1974), judging from the distribution of the possibly most closely related family, the Saururaceae, the Piperaceae may have had a Laurasian origin, perhaps reaching South America via Africa. On the other hand, Trelease (1930) points out th a t the genus may be o f South American o rig in . This is supported by the preponderance in South America of the larger subgenera such as Sphaerocarpidium and Micropiper (which also have Old World members) as w ell as by the exclusive presence in America of groups such as P anioularia. Rhynchophorum and T ildenia. Skottsberg (19*16, 1947) suggested that the strongest relationships of the endemics in the Juan Fernandez Islands were with Peperomia leptostachya and P.. urvelleana A. Rich. (P. enllicheri Miq.) from the Old World (e.g., Australia and New Zealand), and the Pacific (Java, Lord Howe Island and Norfolk Island) and not to tro p ica l American taxa. However, according to the present study and monographic works of the genus in the Old World (e.g., Dull, 1973 [Africa], Quisumbing, 1930 [Indo-Malaysia, Phillipines] and Allan, 1961 [New Zealand]) most taxa are sim ilar to subgenus Tildenia sensu Dahlstedt. Drupes with no posterior process, and a simple, apical stigama (fig. 9-j) are suggestive of this relationship as well as the tufted rosette of leaves with terminal or developing axillary inflorescences Regarding species of Peperomia in the Pacific (e.g., Micronesia, Hawaii, Polynesia, and F iji) which have been monographied in p a rt by 111 Yuncker (1933» 1936, 1938, 1943, 1959), the most striking difference is the ovoid to obovoid-turbinate, mostly rounded or pointed apex with always a divided apical stigma (fig . 9 -c ). In addition, chromosome numbers of Hawaiian species are quite different to those present in the Juan Fernandez Islands (Skottsberg, 1955): Peperomia eekana C.DC., n = ca.36; P.. ervthroclada C.DC., n = 28; P. hawaiiensis C.DC., n = 42 = 46; P.. hesperomannii Haura, n = 48, 66; and P.. l l l i i f o l i a C.DC., n = ca.42. Therefore, based on morphology and chromosome counts, there does not seem to be any good reason to link the Juan Fernandez endemics with the Old World or Pacific ones. Evolution of the Juan Fernandez endemic taxa Relationships among endemic species—. To investigate phylogenetic relationships in Peperomia within the archipelago requires making two assumptions. First, it is assumed that P.. fernandeziana and the three endemic species of Peperomia in the Juan Fernandez Islands (Peperomia berteroana. JP. margaritifera and P.. skottsbergii) have developed from two Independent Introductions from the continent, and that each group is monophyletio. Evidence for this point of view is that there are conspeoific populations of P> fernandeziana in the continent and that the endemic taxa of Peperomia in the islands are more sim ilar to each other morphologically than to any other continental Peperomia species (supported by the phenetio results, fig. 16, and their inclusion in an independent subgenus by Skottsberg (1946). In addition, the three endemic species probably originated in the Juan Fernandez Islands, 112 meaning that speciation occurred there. Second, it is assumed that the cladogram generated based on morphological relationships (fig. 17) represents the most parsimonious evolution of Peperomia in the islands and does reflect the actual evolutionary affinities of the taxa. According to the tree generated in the phylogenetic analysis (fig. 17), Peperomia margaritifera, which is endemic to the older island, shows the fewest derived character states and therefore, represents the most primitive taxon in the islands. Furthermore, several species in the ancestral stock (e.g., 2* alata. J?. cooperl. and P.* mvrtifolia) are facultative epiphytes and P. margaritifera is the only species in the archipelago with such a condition. One synapomorphy, involving habitat, holds P.. berteroana and P. skottsbergil together, the latter being the most recently derived and found only on the younger island of Masafuera. Peperomia berteroana is present on both islands. Thus, Peperomia skottsbergil is more closely related cladistically to JP. berteroana than to P. margaritifera. However, the autapomorphies of succulent stems with no lower leaves on the stem (a "palm-like" condition), subsessile leaves in a verticillate arrangement, presence of a larger stigma and sporadic presence of a second stigma, help define the patristically divergent Peperomia berteroana. This striking number of autoapomorphies (6) differentiating this taxon is concordant with the cluster analysis in which P.. berteroana is divergent from the P. margaritifera and P.. skottsbergil group at the 0.2 level (fig. 16), the la tte r two 'species having more than 70$ of characters in common. This represents a good example of a discrepancy between phenetlo and cladistio analyses, which ordinarily tend to be congruent (Gilmartin, 113 1983), in this case divergent due to conspicuous patristic evolutionary change. It is hypothesized that rapid evolution followed the speciation of P.. berteroana in the Juan Fernandez archipelago. Flavonoid evolution.— Flavonoid evolution in the endemic in su lar taxa (P. berteroana. £. margaritifera and P.. skottsbergil) can be determined by superimposing flavonoid compounds on the hypothesized phylogeny based on morphology (fig. 18). Assuming that Peperomia margaritifera. P. berteroana. and JP. skottsbergil. all have descended from a common ancestor, the most parsimonious evolution of the flavonoid system in the taxa under study can be shown as toward an increase in the number of compounds as well as increased structural complexity (fig. 18). Peperomia margaritifera as well as P.. alata from the ancestral stock, have C-glycoflavones and flavonols, with flavone aglycone and glycosides present only in the latter species. This suggests that these classes of compounds might be ancestral in the insular group. Unfortunately, no comparison with other species within the ancestral section can be made. Only Piper has been surveyed for flavonoids in the Piperaceae, and 5-0-methylflavone, flavanone, C-glycoflavone, dihydroflavonols and chalcones have been documented (Gornall e t a l., 1979). However, some considerations about flavonoid evolution can be made with available data. With regard to Peperomia berteroana and P.. skottsbergil. they form an homogeneous group whose flavonoid profile is based on flavones, with the flavonols being replaced by "advanced" su lfate flavones (Harborne, 1975; Gornall and Bohm, 1978; Del Pero 114 Fig. 18. Evolution of flavonoid chemistry in endemic taxa of Peperomia in the Juan Fernandez Islands. MAR BER SKO FLAVONES F- Sulfat«d FLAVONES FLAVONOLS 116 Martinez, 1985), sometimes conjugated to a glycosyl group. Once thought to be present only in herbaceous or otherwise morphologically advanced families (Harborne, 1975), sulfated flavonoids have been shown to occur rather commonly in flowering plants (Harborne 1977a). These compounds occur in about 200 species belonging to 21 different families (Harborne, 1977a; Young, 1981). Nine of these families belong to the Monocotyledons and 12 to the Dicotyledons, this being the first report in the Piperaceae. Although the function and distribution of sulfated flavonoids is not well known, it has been reported that they are especially frequent in plants inhabiting either fresh or salt water or saline habitats, such as salt marshes or alkaline deserts (Harborne, 1975, 1977a; Harborne and Villiams, 1976). Perhaps due to the b asaltic nature of the cliffs in Juan Fernandez (Quensel, 1920), the rock may weather to produce an alkaline substrate (Ehlers and Blatt, 1982) which may have been the selective factor yielding production of sulfated compounds in Peperomia. It is interesting to note that during evolution of Peperomia berteroana. some flavonoid differences have developed between populations on Masatierra and Masafuera (Table 13). Flavonoid and morphological studies have shown a close relationship between this species and P. tristanensis of Tristan da Cunha island group in the Atlantio Ocean (Valdebenito et al., in prep.). The unique sulfated flavonoid (diosmetin derivative) as well as three unique C-glycosylflavones (acacetin derivatives which replace apigenin, luteolin and diosmetin derivatives) found in P> skottsbergil from Masafuera may have occurred sifter the split in the line leading to 117 i t and P.. berteroana. Thus, the flavonoid data are in accord with morphological interpretations. It seems that a gradual reduction in the number of classes of compounds, and hence, in biosynthetic steps, leads to the presence of only flavone derivatives as well as a C-glycoacacetin derivative in P,. skottsbergil as it evolves in isolation on Masafuera, the younger island of the archipelago. Regarding Peperomia fernandeziana. which has been dispersed independently from mainland Chile, forms a reasonably homogeneous group of populations whose flavonoid profile is based on common flavones, C-glycoflavones, as well as flavonols. The only trend noticeable in the island populations is toward an increase in the number of sugars in apigenin derivatives (table 13> compounds 16 to 19) as one moves from the continent to the older island, and a reduction in the number of populations with acacetin derivatives (compounds 11 to 15). Evolution of Peperomia in the Juan Fernandez Arohipelago.—- Exactly how and when the founding populations of Peperomia arrived in the Juan Fernandez archipelago is not known. However, taking into account the age of the arohipelago as well as its geological history (Stuessy et al., 1984), and the probable origin of the species from South Amerioa, an evolutionary hypothesis for the genus can be offered (fig. 19)* Once the anoestor of the endemio taxa of Peperomia arrived in the island, the insular taxa could have been derived by adaptive radiation. The great variety of habitats available on the islands when the 1 1 8 original colonizations occurred (Sanders et al., 1987) may have correlated with the morphological, chemical and physiological adaptations to environmental conditions exhibited now by taxa of Peperomia in the continent (Yuncker and Gray, 1934; Kaul, 1977; Virzo de Santo et a l., 1983; Ting et a t., 1985; Nishio and Ting, 1987; Schmidt and Werner, 1987). Furthermore, Sanders et al. (1987) have suggested that Masatierra was originally a larger island. Therefore, the original habitats might have been more varied than seen now, and adaptive radiation might hav yed a more significant ole in the origin of the flora than one low perceive. Although the variation in growth form is not overwhel species of Peperomia vary in habitat preference from facult ;.ve epiphytism on Myrceugenia in damp shady areas (e.g., P.. margaritifera) to occurrence on open cliff faces (P.. berteroana and P.. skottsbergil). Endemic to the o ld e r island, Peperomia m argaritifera was the f i r s t species which underwent differentiation in the archipelago. Its closer ties to the putative ancestral stock based on morphology, flavonoid chemistry, and habitat seems warrant this disposition. Peperomia berteroana. which' is abundant on both islands, could have d ifferen tiated firs t on M asatierra and l a tte r dispersed to Masafuera. , * At the same time, P. skottsbergil. a scarce herb, could have derived from a 2* margaritifera-like ancestor in Masatierra with dispersalto the younger island, Masafuera, and subsequent speoiation. Peperomia berteroana and 'P.. skottsbergii grow in rocky canyon walls, particularly on crevices, up to 550 m for the former and 200 m for the latter. Although the two occur In the same kind of environment, JP. berteroana 119 Fig. 19. Phylogenetic hypothesis of the genus Peperomia in the Juan Fernandez Islands. Mainland MASAFUERA MASATIERRA South America Fig. 19 121 and P. skottsbergil have substantially diverged morphologically. The former has succulent erect stems and thick verticillate leaves, which are general adaptations to drier conditions. Skottsberg (1946) showed the presence of a multiple epidermis as well as hydathodes, water-retentive features also present in continental Peperomia growing in exposed s ite s (Kaul, 1977)* On the other hand, Peperomia skottsbergil has thinner stems ascending up to 30 cm, commonly two or three and thin alternate leaves, which correlate with its occurrence on less exposed ridges. How members of the ancestral stock became dispersed to Masatierra is not known, but some agents can be suggested. I t has been argued (Stuessy et al, 1984), that the Juan Fernandez Islands have probably never been connected by land either to the continent or to one another. Wind, sea and animals have been generally accepted (Carlquist, 1974) as possible methods of dispersal to oceanic islands. In the case of Peperomia in the Juan Fernandez Islands, the most likely dispersal mode is by birds. One characteristic of the genus that makes it suitable for long-distanoe dispersal is the production of sticky fruits (Ridley, 1930, Skottsberg, 1946; pers. observ.). The fruits of Peperomia have a pericarp oovered with viscid papillae making them adhesive, and probably these oan be dispersed among the feathers of migratory birds that visit both islands. A number of continental and seabirds have been described as v isitin g the Juan Fernandez arohipelago (Mayer de Schauensee, 1966; Johnson, 1965, 1967; S ohlatter, 1987)* Continental birds that visit both islands inolude: Buteo polyosoma. Faloo peregrinus. and Faloo sparverius (Falconiidae) ; Turdus falklandii 1 2 2 (Turdidae), and Vanellus chilensis (Caradriidae). Additionally, sea birds that connect both islands are: Pterodroma externa (Procellaridae), Phaeton lepturus and P.. rubricanda (Phaethontidae). Dispersal from Masatierra to Masafuera across 150 km of ocean may also have been effected by birds. Speciation.— The study of modes of plant speciation involves, among i others, obtaining knowledge of reproductive modes via detailed breeding and isozyme studies of the taxa under consideration (Sanders et al., 1987)* Unfortunately, little is known about reproductive modes in Peperomia to allow inferences about modes of speciation, but a few comments of a general nature can be made at this point. Of the eight modes of speciation outlined by Grant (1981), quantum and geographic speciation seem the most likely modes operative in Peperomia in the Juan Fernandez. On the basis of inferred phylogeny and distributional patterns, it is hypothesized that through quantum speciation, a novel species (e.g., £. margaritifera) evolved rapidly on Masatierra from a random subset of the ancestor’s gene pool (taxa of Section Alternifoliae). Thus, once the founder population became established and was se ttle d on M asatierra i t dispersed to Masafuera which provided an ideal setting for rapid geographic isolation, especially i f aocompanied by s e lf-fe rtiliz a tio n . Allopolyploid speciation seems unlikely because of the uniformity of chromosome counts within the insular taxa (n = 22; table 15). Related taxa in Seotlon Alternifoliae (e.g., ?. ro tu n d ifo lia (L.) H.B.K., Smith, 1966) have n = 11; therefore, the endemic species are probably tetraploids 123 but it is unknown whether tetraploidy accompanied speciation in the islan d s. Regarding mode of reproduction of Peperomia, it is not known if self-pollination or outcrossing predominates in the Piperaceae. A study of four Piper species and one of Pothomorphe in Costa Rica showed that insects as well as rain are involved in pollination (Semple, 1974). Wind-pollination is unlikely because of the glutinous nature of the pollen. If pollinators are not available, there is evidence that self-fertilization may occur yielding good fruits (Martin and Gregory, 1962). The protogyny of Piperaceae flowers (the stigmas being exserted several days before anther dehisoence) encourages outcrossing, although the timing is not sufficiently different to prevent pollination of flowers on the same spike by each other. If apomixis occurs, it is rare. Kanta (1961) in an embryological study in Piper nigrum, recorded that only 10$ of all ovules developed healthy embryos due to a lack of pollination. It has been suggested (Semple, 1974) that members of Hymenoptera (e .g ., ants) may aid in d istrib u tin g pollen within a simple spike and occasionally in distributing pollen between plants. It is interesting to mention that six species of Formicidae have been described for the Juan Fernandez Islands (Wheeler, 1924). This combination of self-fertilization and occasional outcrossing is highly successful and characteristic of a large number of colonizing species (Allard, 1965). Given these facts, in addition to the geographic features of both islands of deep canyons, especially as now seen on Masafuera, the essential characteristics are present for rapid geographic isolation accompanied by self-fertlizatlon as may have occurred in the endemic species of Piperaceae of the Juan Fernandez Islands. 125 REFERENCES AL-KHUBAIZI, M., 1977: Sulfated and non-sulfated flavonoids from Flaverla. Sartwellia and Haploesthes. ~ M.A. thesis, University of Texas, Austin. ALLAN, H. H ., 1961: Piperaceae. — In ALLAN, H. H. (Ed.): Flora of New Zealand I, pp. 173-174. — Wellington: R. E. Owen. ALLARD, R., 1965: Genetic systems associated with colonizing ability in predominantly self-pollinated species. — In BAKER, H. G., STEBBINS, G. L ., (Eds.): The genetics of colonizing species, pp. 105-120. — New York: Academic Press. BAI, G. V. S ., SUBRAMANIAN, D., 1964: Cytotaxonomical studies of South Indian Piperaceae.-- Cytologia 50: 583-592. BISHOP, M. J ., 1982: Criteria for the determination of the direction of character sta te changes. — Zool. J. Linn. Soc. 74: 197-206. BLOT, J., 1960: Contribution a 1 'etude cytologlque du genre Peperomia. — Rev. Gen. Bot. 67: 522-535. BURGER, W . C ., 1971: Piperaceae. — In BURGER, W. C ., (Ed.): F lora Costarioensis. — Fieldiana, Bot. 35: 6-79. CARLQUIST, S ., 1974: Island biology. — New York: Columbia University Press. CARR, G. D ., KYH0S, D. W. , 1986: Adaptive radiation in the Hawaiian silverword allianoe (Compositae: Hadiinae). II. Cytogenetics of a r tif ic ia l and natural hybrids. — Evolution 40: 959-976. CRAWFORD, D. J ., STUESSY, T. F . , SILVA, M., 1987: Allozyme divergence and the evolution of Dendroserls (Compositae: Lactuceae) on the Juan Fernandez Islands. — Syst. Bot. 12: 435-443. CRISCI, J . V., STUESSY, T. F ., 1980: Determining prim itive character states for phylogenetic reconstruction. — Syst. Bot. 5: 112-135. DAHLSTEDT, H., 1900: Studien uber Sud- und Central-Amerikanische Peperomien. — Xongl. Svenska Vetenskapsakad. Handl. 33: 1-218. DASGUPTA, A., DATTA, P. C., 1976: Cytotaxonomy of Piperaceae. — Cytologia 41: 6 97-706. CANDOLLE, A. C. P. d e, 1869: Piperaceae. — In CANDOLLE, A. L. P. P. de, (Ed.): Prodromus XVI, part 1, pp. 235-471. — Paris: Victoris Masson et F i l i i . Del PER0 MARTINEZ, M. A., 1985: Flavonoid relationships between the Commelinaceae and related fam ilies. — Biochem. Syst. Ecol. 13: 253-256. DIERS, L., 1981: Der Anteil on Polyploiden in den vegetationsgurteln der Westkordillere Perus. — Z. Bot. 49: 437-488. DONOGHUE, M. J . , CANTIN0, P. D ., 1984: The logic and lim itatio n s of the outgroup substitution approach to cladistic analysis. — Syst. Bot. 9: 192-202. DULL, R., 1973: Die Peperomla-Arten Afrlkas. -- Bot. Jahrb. Syst. 93: 56-129. DUNCAN, T., BAUM, B., 1981: Numerical phenetlcs: its uses in botanical systematios. — Ann. Rev. Ecol. Syst. 12: 387-404. EHLERS, E. G., BLATT, H., 1982: Petrology, igneous, sedimentary and 1 2 7 metamorphic. — San Francisco: V. H. Freeman. GORMAN, M., 1979: Island ecology. — London: Chapman and Hall. GORNALL, R. J ., BOHM, B. A., 1978: Angiosperm flavonoid evolution: a reappraisal. — Syst. Bot. 3: 353-368. , ------1 DAHLGREN, R ., 1979: The d istrib u tio n of flavonoids in the angiosperms. — Bot. Not. 132: 1-30. GRANT, V., 1981: Plant speciation. 1st edn. — New York: Columbia University Press. HARBORNE, J. B., 1975: Flavonoid sulphates: a new class of sulphur compounds in higher plants. — Phytochemistry 14: 1147-1155. —, 1977: Flavonoid sulphates: a new class of natural product of ecological significance in p lan ts. - - Progr. Phytochem. 4: 189 - 208 . ------, WILLIAMS, C., 1976: Flavonoid patterns in leaves of the Gramineae. — Biochem. Syst. Ecol. 4: 267-280. HUYNH, K. L., 1965: Contribution a 1 'etude caryologique et embryologique des phanerogames du Perou. — Denkschr. Schweiz. Naturf. Ges. 85: 1-178. JACKSON, R. C., 1971: The karyotype in systematics. — Ann. Rev. Ecol. Syst. 2: 327-368. JOHNSON, A. W. , 1965: The birds of Chile I. — Buenos Aires: Platt Estableoimientos Graficos, S.A. — —, 1967: The birds of Chile II. — Buenos Aires: Platt Estableoimientos Graficos, S.A. KANTA, K., 1961: Embryolessness in Piper nigrum L. — Phytomorphology 11: 304-306. 128 KAUL, R. B., 1977: The role of the multiple epidermis in foliar succulence of Peperomia (Piperaceae). — Bot. Gaz. 138: 213-218. KUNKEL, G., (E d .)f 1976: Biogeography and ecology in the Canary Islands. — The Hague: W. Junk. LAMMERS, T. G., 1988: Chromosome numbers and th e ir systematic implications in Hawaiian Lobelioideae (Campanulaceae). — Amer. J. Bot. 75: 1130-113^. LOWREY, T. K., CRAWFORD, D. J . » 1985: Allozyme divergence and evolution in Tetramolopium (Compositae: Astereae) on the Hawaiian Islands. — Syst. Bot. 10: 64-72. MABRY, T. J ., MARKHAM, K. R., THOMAS, M. B., 1970: The systematic identification of flavonoids. — New York: Springer-Verlag. MADDISON, W. P ., DONOHUE, M. J . , MADDISON, D. R ., 1984: Outgroup analysis and parsimony. — Syst. Zool. 33: 83-103. MARTIN, F. W., GREGORY, E ., 1962: Mode of po llin atio n and factors affecting fruit set in Piper nigrum L. in Puerto Rico. — Crop Sci. 2: 295-299. MENENDEZ, C. A., 1972: Estudios paleobotanicos en la Argentina, avanoes, problemas y perspectivas. — Mem. Symp. I Congr. Latino-Amer. Bot., pp. 61-97* MEYER de SCHAUENSEE, R., 1966: The species of b ird s of South America and their distribution. — Philadelphia: Livingston Publishing Co. MIQUEL, F. A. W., 1843-1844: System piperacearum. — Rotterdam: H. A. Kramers. NISHIO, J. N., TING, I . P . , 1987: Carbon flow and metabolic speciation in the tissue layers of the orassulaoean acid metabolism plant, Peperomia oamptotrloha. — Plant Physiol. 84 : 600-6 04. PACHECO, P ., CRAWFORD, D., STUESSY, T. F ., SILVA, M., 1985: Flavonoid evolution in Robinsonia (Compositae) of the Juan Fernandez Islands. — Amer. J. Bot. 72: 989-998. QUENSEL, P ., 1920: Additional comments on th e geology o f the Juan Fernandez Islands. — In SKOTTSBERG, C., (Ed.): The natural history of Juan Fernandez and Easter Island, part 1, pp. 37-87. — Uppsala: Almqvist and Wiksells. QUISUMBING, E ., 1930: Philippine Piperaceae.— Philipp. J. Sci. 43: 1-246. RAVEN, P. H ., AXELROD, D. I ., 1974: Angiosperm biogeography and past continental movements. — Ann. Missouri Bot. Gard. 61: 539-673. RIDLEY, H., 1930: The dispersal of plants throughout the world. — Ashford: Reeve and Co. R0HLF, J ., KISHPAUGH, J . , KIRK, D ., 1972: NT-SYS numerical taxonomy system of multivariate statistical programs. — Stony Brook: State University of New York. RUIZ, H., PAVON, J ., 1797: Flora peruvianae et chilensis prodromus. 2nd edn. — Rome. SANDERS, R. W., STUESSY, T. F., MARTICORENA, C., SILVA, M., 1987: Phytogeography and evolution of Dendroseris and Robinsonia. tree-Compositae o f the Juan Fernandez Islands. — Opera Bot. 92: 195-215. SCHLATTER, R. P., 1987: Conoolmiento y situacion de la ornitofauna en las islas ooeanicas Chilenas. — In CASTILLA, J. C., (Ed.): Islas ooeanicas Chilenas: oonooimiento cientiflco y necesidades de investigaciones, pp. 271-285. — Santiago: Ediciones Universidad Catolica de Chile. SCHMIDT, J. E ., 1987: Response of the succulent leaves of Peperomia magnoliaefolia to dehydration. — Plant Physiol. 83: 190-194. SEMPLE, K. S., 1974: Pollination in Piperaceae. — Ann. Missouri Bot. Gard. 61: 868-871. SKOTTSBERG, C., 1946: Peperomia berteroana Miq. and P.. tristan en sis Christoph., an interesting case of disjunction. — Acta Horti Gotob. 16: 251-288. , 1947: The genus Peperomia in Chile. — Acta Horti Gotob. 17: 1-47. —— , 1955 ; Chromosome numbers in Hawaiian flowering plan ts. — Ark. Bot. 3: 63-70. —- — , 1956: Derivation o f the ora and fauna of th e Juan Fernandez and Easter islands, part 1. the Juan Fernandez Islands. — In SKOTTSBERG, C., (Ed.): The natural history of Juan Fernandez and Easter Island, pp. 193-405. — Uppsala: Almqvist and Viksells. SMITH, J . B., 1966: Chromosome numbers in Peperomia Ruiz and Pavon (Piperaceae) and note on chromosome number of Piper magnifloum Trelease. Kew Bull. 20: 521-526. STEVENS, P. F., 1980: Evolutionary polarity of character states. — Ann. Rev. Ecol. Syst. 11: 333-358. STUESSY, T. F ., 1971: Chromosome numbers and phylogeny in Melampodium (Compositae). — Amer. J. Bot. 58: 732-736* , FOLAND, K ., SUTTER, J . , SANDERS, R ., SILVA, M. , 1984: Botanical and geological significance of potassium-argon dates from the Juan Fernandez Islands. — Science 225: 49-51. SWOFFORD, D. L ., 1985: PAUP: phylogenetic analysis using parsimony, version 2.4. — Champaign: Illinois Natural History Survey. TING, I ., BATES, L. A., O'REILLY, L ., DENIRO, M. J . , 1985: Physiological and isotopic aspects of photosynthesis in Peperomia. — Plant Physiol. 78 : 246-249. TRELEASE, W., 1930: The geography of American peppers. — Proc. Amer. Philos. Soc. 69: 309-327. ----, YUNCKER, T. 6., 1950: The Piperaceae of northern South America. 2 vols. — Urbana: University of Illin o is Press. VIRZO De SANTO, A ., ALFANI, A., RUSSO, G., FI0RETT0, A., 1983: Relationship between CAM and succulence in some species of Vitaceae and Piperaceae. — Bot. Gaz. 144: 342-346. WATROUS, L., WHEELER, Q ., 1981: The out-group comparison method of character analysis. — Syst. Zool. 30: 1-11. WHEELER, W. M., 1924: Formicidae from Easter Island and Juan Fernandez. — In SKOTTSBERG, C., (Ed.): The natural history of Juan Fernandez and Easter Island, part 2, p. 419* — Uppsala: Almqvist and Wiksells. YOUNG, D. A., 1981: The usefulness of flavonoids in angiosperm phylogeny: some seleoted examples. — In YOUNG, D. A., SIEGLER, D. S., (Eds.): Phytochemistry and angiosperm phylogeny, pp. 205-232. — New York: Praeger Scientific. YUNCKER, T. G . 1933: Revision of the Hawaiian species of Peperomia. — Bernice P. Bishop Mus. Bull. 112: 1-133* , 1936: Peperomia. — In SMITH, A., (Ed.): Fijian plant studies.— Bernice P. Bishop Mus. Bull. 141: 25-47. , 1938: Revision of the Micronesian species of Peperomia. — Occas. Pap. Bernice Pauahi Bishop Mus. 14: 7-25. , 1943: Piperaceae. — In YUNCKER, T. G., (Ed.): The flo ra of Niue Island. — Bernice P. Bishop Mus. Bull. 178: 44-45. -i 1959: Piperaceae of Micronesia. — Occas. Pap. Bernice Pauahi Bishop Mus. 22: 83-108. , 1974: The Piperaceae of Brazil III. Peperomia: taxa of uncertain status. — Hoehnea 4: 71-413. , GRAY, W. D., 1934: Anatomy of Hawaiian Peperomias. — Occas. Pap. Bernice Pauahi Bishop Mus. 10: 1-19. C hapter IV SYNONYMY IN PEPEROMIA (PIPERACEAE) RESULTS IN BIOLOGICAL DISJUNCTION BETWEEN PACIFIC AND ATLANTIC OCEANS During studies of patterns of phylogeny and modes of speciation in the endemic taxa of the Juan Fernandez Islan d s, Chile (Stuessy, Sanders and Silva, 1984; Lammers et al., 1986; Sanders et a l., 1987) a conspicuous transoceanic distributional pattern has been noted in the genus Peperomia (Piperaceae). The Juan Fernandez Islands are located 660 km west of continental Chile (C) at approximately 33°S, 80°W. There arc two principal volcanic islands separated by 150 km: Masatierra (MT) which is known to be 4 m.y. old; and Masafuera (MF) which is 1-2 m.y. (Stuessy et al., 1984). In this archipelago, there are four distinctive Peperomia species: P. berteroana Miquel (MT, MF), P. fernandeziana Miquel (C, MT, MF), J?. margaritifera Hook. (MT), and P. skottsbergii C. DC. (MF). Peperomia berteroana, the species of most interest h ere, was f i r s t described by Miquel (1843) from material colleoted from Masatierra. The Tristan da Cunha archipelago also is relatively young (Oilier, 1984; Preece, 1986) and oonsists of three islands: Tristan (0.5 m.y.), Inaccessible (2.9 Q.y., where P.. tristanensis ocours), and Nightingale (18 m.y.). During the Norwegian Scientific Expedition 1937-1938, 133 134 Christophersen (1944) collected a specimen of Peperomia on Inaccessible Island, which was treated as a new species, P. tris ta n e n sis . This species is the only Piperaceae inhabiting that archipelago, which is situated in the middle of the South Atlantic Ocean (37°S, 12°W). On the basis of evidence presented by Skottsberg (1946), Christophersen (1968) agreed that P.. berteroana and P.. tristanensis are closely related, but specifically distinct, taxa. It is the purpose of this paper to show that the morphological differences between these taxa do not justify recognizing two separate species, as has been done in the past. Two subspecies are herein recognized: Peperomia berteroana subsp. berteroana confined to the Juan Fernandez Islands, and P.. berteroana subsp. tristanensis found only in the Tristan da Cunha Archipelago. The results of the present comparative study of both taxa are derived from examination of herbarium material covering as wide a morphological and geographical range as possible, supplemented by field observations on the Juan Fernandez Archipelago (expeditions in 1980, 1984, and 1986). Twenty-seven vegetative and reproductive characters were studied in detail. Of these characters, leaf shape, number of principal leaf veins, blade vestiture and spike arrangement show discontinuities. In all other respects, the two taxa are indistinguishable. Even in these diagnostic features, however, some overlap occurs. For example, the Tristan da Cunha material has oblanceolate leaves as do most populations from the Juan Fernandez Islands. The number of principal leaf veins is five in the Tristan da Cunha specimens and three in most populations from Juan Fernandez, but 1 3 5 several populations on Masafuera (e.g . Pacheco and Ruiz 6407. 6412) have both patterns. Regarding leaf vestiture, populations from Juan Fernandez have blades that are pubescent above only on th e principal veins, and are completely pubescent below. Some populations on Masafuera, however, show the same pattern as the Tristan da Cunha specimens in which the blade is pubescent below only on th e principal veins. The fourth distinctive character, spike arrangement, is uniformly subumbellate in the Juan Fernandez Islands specimens, but varies from umbellate to paniculate in the T ristan da Cunha populations. In light of the significant morphological overlap between the two described species, J?. berteroana and P.. tristanensis. therefore, it seems prudent to unite them within the same species as disjunct subspecies (of P. berteroana). Within this new perspective, the species show one of the broadest natural disjunctions among flowering plants, with the two known population systems separated by more them 5,000 kms. Peperomia berteronana Miq- subsp. tristanensis (E. Christophersen) Valdebenito, comb, et stat. nov.—- Peperomia tristanensis E. Christophersen, Results Norw. Sclent. Exped. Tristan da Cunha 1937-1938, 11:5. 1944. — TYPE: Tristan da Cunha Archipelago, Inaccessible Island, above waterfall at Salt Beach, 150 m, 2 Mar 1938, E. Christophersen 2592 (holotype: 0! [sheet No 1]; isotypes: KEW! 0!). Additional specimens examined: Tristan da Cunha. Inaccessible. 136 Brought by inhabitant and planted in garden, 30 Dec 1937, Christophersen 462 (0), 502 (0); above Salt Beach, by waterfall, 21 Jan 1938, Glass and Christophersen 1278 (A, BOL, 0, PRE, WELT); gorge on the N side of waterfall stream, 4 Nov 1982, Hall 73 (KEW); above second f a ll in w aterfall gulch, 200 m, 28 Jan 1983, Hall 308 (KEW). Peperomia berteroana Miquel subsp. berteroana — Peperomia berteroana Miauel. Syst. Pip. 114. 1843. TYPE: Masatierra, Apr 1830, (J. G« Bertero 1492 (holotype: G; iso types: GH! KEW!). Additional specimens examined: Juan Fernandez. MASATIERRA: Valle V illagra, 530 m, Landero e t a l. 6527 (OS); La Damajuana, 480 m, Lopez 36 (C0NC); S elkirks' Lookout, 540 m, Marticorena e t al. 9024 (OS); Quebrada de V illagra, mas abajo del Salto de La Pulga, 375 m, Marticorena et a l. 9042 (OS); ca. 200 yards from pass of Camote, 1800 ft, Meyer 9616 (UC); Skottsberg s.n. (GB); Portezuelo de Villagra, 550 m, Skottsberg 38 (GB, 0); Quebrada Salsipuedes, 600 m, Skottsberg 96 (0; GB); Quebrada de Villagra, 200 m, Skottsberg 261 (0); Puerto Ingles, Cordon Central, 380 m, Skottsberg 321 (GB, 0); ridges on El Yunque, 550 m, Skottsberg 628 (GB, 0); El Camote, 550 m, Sparre 25 (C0NC); Near base of Damajuana, 370 m, Stuessy and Sanders 5087 (OS); down W slope from Mirador Selkirk into Quebrada V illagra, 420 m, Stuessy and Sanders 5204 (OS); steep wooded slope ca. 100 m SW of Mirador toward Piramide, 530 m, Stuessy and Sanders 5212 (OS); 2/3 up trail to La Carbonera de Torres, 500 m, Stuessy et al. 5314 (OS); Valle Villagra, 430 m, Stuessy et al. 6517 (OS); East side of Cerro Damajuana, 440-525 m, Valdebenito e t a l. 6321 (OS), 6332 (OS); V alle Ingles, 330 m, Valdebenito and Landero 6550 (OS). 6594 (OS): Puerto Frances, 400-410 m, Valdebenito and Landero 6663 (OS), 6667 (OS); Lookout, 30 m, Weber 23606 (C0NC). MASAFUERA: Quebrada Casas, 160—265 m, Pacheoo and Ruiz 6388 (OS), 6407 (OS); Quebrada Vacas, North fork, 60 m, Pacheco and Ruiz 6412 (OS); Quebrada Casas, 160 m, Ruiz and Lammers 8004 (OS); North branch of Quebrada Varadero, 100-150 m, Ruiz and Landero 8260. 8277 (OS); Quebrada Vacas, Skottsberg 394 (GB); Quebrada Casas, 110 m, Stuessy et al. 5034 (OS); Quebrada Pasto, 20 m, Stuessy and T.ammftra 8402 (OS); Quebrada Sandalo, 320 m, Valdebenito and Lammers 8154 (OS); Quebrada Vacas, 250-600 m, Valdebenito and Lammers 8172 (OS), 8200 (OS); a t base of high waterfall just west of Vicente Porras, 75 m, Valdebenito 8513 (OS). 138 LITERATURE CITED CHRISTOPHERSEN, E. 1944. New Phanerogams from Tristan da Cunha. Results Norw. Sclent. Exped. Tristan da Cunha. 11:1-5. ______1968. Flowering plants from Tristan da Cunha. Results Norw. Sclent. Exped. Tristan da Cunha. 55:1-29. LAMMERS, T.G., T. F. STUESSY and M. SILVA. 1986. Systematic relationships of the Lactoridaceae, an endemic family of the Juan Fernandez Islands, Chile. PI. Syst. Evol. 152: 243-266. M1QUEL, F.A. 1843. Systema Piperacearum. Rotterdam: H.A. Kramers. OLLIER, C.D. .1984. Geomorphology o f South A tlantic volcanic islands. Part I : The T ristan da Cunha Group. Z. Geomorphol. 28 : 367-382. PREECE, R.C., K.D. BENNETT and J.R . CARTER. 1986. The Quaternary paleobotany of Inaccessible Island (Tristan da Cunha group). J. Biogeography 13:1-33* SANDERS, R.H., T.F. STUESSY, C. MARTICORENA and M. SILVA. 1987. Phytogeography and evolution of Dendroseris and Robinsonia. tree-Compositae of the Juan Fernandez Islands. Opera Bot. 92: 1 3 9 195-215. SKOTTSBERG, C. 1946. Peperomia berteroana Miq. and P,. trista n e n sis Christoph., an Interesting case of disjunction. Acta Horti Gotob. 16: 251-288. STUESSY, T.F., R.W. SANDERS and M. SILVA. 1984. Phytogeography and evolution of the Juan Fernandez Islands: a progress report. Pp. 55-69. In: F .J. Radovsky, P.H. Raven and S.H. Sohmer (e d s.), Biogeography of the Tropical Pacific. A.S.C. and B.P. Bishop Museum, Lawrence, Kansas. ______, K.A. FOLAND, J.F. SUTTER, R.W. SANDERS and M.O. SILVA. 1984. Botanical and geological significance of Potassium-Argon dates from the Juan Fernandez Islands. Science 255: 49-51. Chapter V A NEW BIOGEOGRAPHIC CONNECTION BETWEEN ISLANDS IN THE ATLANTIC AND PACIFIC OCEANS The affinities between floras of widely separated regions of the world, as well as modes of dispersal of elements in these floras, have long been of interest in evolutionary and phytogeographical studies (1). While investigating phylogenetic patterns and modes of speciation in the endemic plants of Juan Fernandez Islands, Chile (2) we have documented a strikingly broad transoceanic pattern of distribution in the plant genus Peperomia (Piperaceae). The Juan Fernandez islands are located 600 km west of continental Chile at approximately 33°S, 80°W (Fig. 20). The two principal volcanic islands are Masatierra (MT), which is approximately 4 m.y. old, and Masafuera (MF), which is 150 km farther west, and 1-2 m.y. old (3). Four distinctive species of Peperomia occur in this archipelago: P. margarltifera (MT), P. akottsbergii (MF), J?. fernandeziana (Chilean mainland, MT, MF) and P. berteroana (MT, MF) ooour in this archipelago. Comparative morphological studies reveal that the latter species is nearly identical to P. tristananaia. which was treated as a synonym of P. berteroana by Christophersen (1968) (4) and Groves (5). Peperomia U 0 U 1 tristanensis occurs on Inaccessible Island of the Archipelago of Tristan da Cunha, which is in the middle of the South Atlantic Ocean (37°S, 12°W) (Fig. 20). This archipelago is relatively young (6) and consists of three islands Tristan (0.5 m.y.), Inaccessible (2.9 m.y.), and Nightingale (18 m.y.). The study of disjunctions involves: (1) the taxonomic problem of determining whether disjunct taxa are conspecific, or at least close relatives, (2) determining the closest relatives or the disjunct taxa, and (3) ascertaining how the disjunctions may have arisen. Because the present study focuses on the extent of morphological and chemical differentiation in relation to geographical distribution, the results are based on-herbarium material of Peperomia berteroana and relatives. Observations and collections were made during three field expeditions to the Juan Fernandez Islands (1980, 1984, 1986). Twenty-seven morphological characters, including both vegetative and reproductive features, were analyzed using numerical techniques (7). Leaf-shape, number of principal leaf veins, blade vestiture, and spike arrangement showed the most notable discontinuities. Species were compared fo r flavonoid compounds sequestered in th e ir leaves. A total of fourteen flavones, five with sulfate groups, were Isolated and identified through high voltage paper electrophoresis and 2-D paper ohromatography (Table 16). There is a pattern of island-specific, rather than taxon or archipelago-specific flavone p ro file s. The .most notable flavonoid d iffe re n tia tio n is seen between plants on Masatierra and those of the Masafuera-Inaccessible island group. Even though the latter two islands belong to two arohipelagos Fig. 20. Location of the Juan Fernandez and Tristan da Cunha archipelagos. Juan Farnandaz Tristan da Cunha 1000 km T r ista n 10 km 5 km 3 U Fig. 20 I Table 16. Occurrence of flavones in Peperomia tristanensis and P. berteroana on Masatierra CMT) and Masafuera (MF). FLAVONES S D e c ie s 1 2 3 4 5 6 7 8 9 10 11 12 1 3 1 4 P. berteroana (MT) + + + + + + + + + + P. berteroana (MF) + + + + P. tristanensis + + + + + + + + + + + A = apigenin; Ac = acacetin; D = diosmetin; L = luteolin; 1 = Acacetin; 2 = Ac 7-sulfate; 3 = D 7-0-monoglucoside; 4 = D 7-0-rhamnoglucoside; 5 = L 7-0-monoglucoside; 6 = L O-diarabinoside; 7 = L 7-sulfate; 8 = A 7-glucosidesul fate; 9 — L 7—0—rhamnogl ucoside;' 10 = 0 7—0—glucoside; 11 = Ac 7-glucosidedisulfate; 12 = D C-arabinog1ucoside; 13 = D 7-0-arabinoglucoside; 14 = D 7-sulfate. 14-5 separated by more than 5,000 km., p lan ts of P_. berteroana occurring in them have four compounds in common, none of which is present in the M asatierra populations. There is some interpopulational variation in flavonoids of plants from Masatierra, but not so in populations from Masafuera and Inaccessible. In addition, principal components analysis (fig . 21) disclosed some populations of P. berteroana from Masafuera closer to specimens from Tristan da Cuhna, suggesting morphological ties between populations from both archipelagos. Therefore, flavone distribution and morphological sim ilarity suggest that Tristan da Cunha and Juan Fernandez Islands populations of P.. berteroana are best recognized at the subspecific level (the new combinations have been proposed elsewhere). This particular disjunct distribution could be explained by one of four hypotheses: (1) parallel evolution from a common ancestor in South America; (2) convergent evolution from different ancestors; (3) continental drift; and (4) long-distance dispersal by birds. If the first hypothesis were true, P. berteroana and P. tristanensis would represent subspecies having evolved in parallel from a common ancestral population which probably spread from South America to both arohipelagos by means of bird dispersal within the past 3 m.y. This time frame is suggested by the known ages of the islands where Peperomia now grows, and by the very slight morphological and chemical differentiation between the Juan Fernandez and Tristan da Cunha populations. No linking populations, or a common anoestral population with the U 6 Fig. 21. Principal components analysis of Peperomia present In Juan Fernandez and Tristan da Cunha archipelagos. o £. berteroana (Masatierra); Q £. berteroana (Masafuera); ^ £. tristanensis. U 7 0.774 o • • o ° o 0.440 • • CM • O o o oo 8 H o O o OP < O O o o 0 o 0. 8*0 - oo 0.007 - 0.700 - 0.00 0 - 0.000 - 0.104 0.04 7 0.140 0 .4 4 0 0.04 0 0.0 00 FACTOR 1 F1g, 21 148 characteristics shown by the taxa on both islands have been discovered in South America despite extensive searches in different herbaria rich in collections of Peperomia. Therefore, the ancestor either does not ex ist in South America, has became extinct (fo ssils of Peperomia are not known from these antecedent localities), or it has not yet been discovered. The hypothesis of convergent evolution to account for the disjunct distribution of these two subspecies of Peperomia seems unlikely because both taxa are morphologically very similar (8) and sequester nearly the same array of compounds. This would mean that convergence would have occured in both morphological and chemical features, which seems improbable. A third hypothesis, involving continental drift, plate tectonics, and ocean-floor spreading, seems unlikely as an explanation of the distributional pattern of the subspecies of Peperomia. At the close of the Cretaceous (65 m.y.), when the Piperaceae, as well as probably a majority of flowering plant families, were in existence (9), about 800 km probably separated Africa and South America at their closest points, and neither the Juan Fernandez nor Tristan da Cunha Arohipelago is older than 18 m.y. Therefore, the Tristan da Cunha Archipelago appeared when the South A tlantic Ocean already was opened to i t s fu ll extent. Furthermore, the Juan Fernandez Islands have never been connected to continental Chile (2). The final .hypothesis, long distance dispersal of the subspecies of Peperomia. seems most conoordant with all available information. One characteristic of the genus Peperomia that makes it suitable for this U 9 mode of dispersal is the production of sticky fruits (8; 10, pers. observ.). The fruits of Peperomia have a pericarp covered with viscid papillae making them adhesive, and they probably can be attached to and dispersed among the feathers of migrating birds that visit both archipelagos. Such birds include Fregetta grallaria subsp. melanolerea. (the Tristan Storm Petrel) which has been recorded as breeding in both archipelagos (11), and Pterodroma externa, the (Juan Fernandez Petrel) endemic to Masafuera, which probably only touches the Tristan Archipelago accidentaly in its circumpolar wanderings. Th'e position of the Tristan da Cunha archipelago on the mid-Atlantic Ridge suggests th a t there could have been numerous islands that may have served as stepping stones between South America and Africa during the gradual opening of the Atlantic. However, the poverty and general composition of the Tristan flora, as compared to that of South America and Africa, do not support this possibility. Prominent members of the fern flo ras of both South America and A frica are absent (12). Migration across archipelagos would be expected to result in a richer' flora than the islands have and one that contains a better representation of the prominent elements of the adjacent continents. Therefore, because the f ru its of Peperomia berteroana are stick y , and because birds visit and breed on both archipelagos, i t seems most likely that long-distance dispersal is the explanation for the distribution of these two insular subspecies of Peperomia in separate oceans. In addition, there is evidenoe (13) that self-fertilization 150 may occur in at least some species of the family Piperaceae, and this feature would be advantageous for island colonization (14). Therefore, Peperomia has several characteristics conducive to success following long-distance dispersal. Furthermore, there is a marked similarity in the ecology on Tristan da Cunha and the Juan Fernandez archipelago; Masafuera as well as the islands in the Tristan da Cunha group have an alpine and sub alpine flora (15). It seems most probable that the present distribution of the subspecies of Peperomia berteroana can be explained by long-distance dispersal, perhaps initially from the Juan Fernandez Islands, and specifically from Masafuera. The genus Peperomia is very well distributed on Masatierra and Masafuera, but this is not so in Inaccessible where it is an exceedingly rare plant, being known only from the valley leading up from Salt Beach. Phylogenetic relationships based on morphological and flavonoid data (16) suggest that the populations on Tristan da Cunha are more derived than those of Juan Fernandez. Of the four characters separating Peperomia berteroana subsp. bertoana and subsp. tristanensis. only the latter subspecies has paniculate spikes, a derived character state. Regarding complexity of flavonoid chemistry and phyletic advancement, Peperomia populations on Masafuera and Inaccessible have a more biosynthetically complex chemical pattern as well as more complex compounds than those from Masatierra (Table 16). In addition, the Tristan da Cunha plants represent a subset of the morphological variation found in plants on Masafuera. These observations support the view that the former two island populations were derived from sources on Masatierra. It has 151 been suggested (17) that in general the presence of flavones only, complex O-glycosylation, and presence of flavonoid bisulfates, represent derived character states in a phylogenetic context. Therefore, if character states can be used to suggest derived taxa and derived distributional areas, then P. berteroana subsp. tristanensis would be viewed as derived from P.. berteroana subsp. berteroana from the Juan Fernandez Islands. It is clear that Peperomia is of considerable interest for the biogeographer. Even though there exist a number of amphitropical Pacific examples, (e.g., the genus Rhamphogyne, Compositae, with one species in the Mascarene Islands and another in New Guinea (18)) as well as disjunctions between Hawaii and the center-South P acific ( e .g ., Charpentiera. Amaranthaceae, (19) and Tetramolopium. Asteraceae, (20), none of these disjunctions is between d ifferen t oceans. The present documentation represents a disjunction of about 5,000 km with the South American continent lying between the two populations. The only other known flowering plant found in both archipelagos (as two d ifferen t v a rie tie s), is apparently Empetrum rubrum Vabl. (Empetraeeae), which is also found in Southern South America (4). This example also may be due to long-distance dispersal by birds, but further in-depth comparative studies in the genus are needed before this hypothesis oan be confirmed. At present, Peperomia berteroana appears to represent one of the longest disjunot distributions among flowering plants. 152 REFERENCES 1. P.F. Stevens, Notes Roy. Bot. Gard. Edinburgh 30. 341 (1970); D.M. Moore and A.O. Chater, Bot. Not. . 124. 316 (1971); D.M. Moore, in Taxonomy. Phytogeography and Evolution. D.H. Valentine, Ed. (Academic Press, London, 1972), pp. 115-138. 2. T.G. Lammers, T.F. Stuessy, M.O. Silva, PI. Syst. Evol. 152. 243 (1986); R.W. Sanders, T.F. Stuessy, C. Marticorena, M.O. Silva, Opera Bot. 92. (1987); T.F. Stuessy, K.A. Foland, J.F. Sutter, R.W. Sanders, M.O. Silva, Science. 255. 49 (1984). 3> T.F. Stuessy, K.A. Foland, J.F. Sutter, R.W. Sanders, M.O. Silva, Science. 255. 49 (1988). 4. E. Christophersen, in Results Norw. Soient. Exped. Tristan da Cunha. L. Christensen, Ed. (Oslo, 1968) ££., pp. 1-29. 5. E. Groves, Bull. Brit. Mus. (nat. Hist), (Bot).. 8., 333 (1981). 6. C.D. Oilier, Z. Geomorphol. 28. 367 (1984); R.C. Preeoe, K.D. Bennett, J.R. C arter, .J. Biogeogr. 13. 1 (1986). 7. J ., Rohlf, J. Kishpaugh, D. Kirk, NT-SYS. Numerical taxonomy 153 system of multivariate statistical programs, (State University • N.Y., 1972). 8. C. Skottsberg, Acta Horti Gotob. 16. 251 (1946). 9. P.H. Raven, and D.I. Axelrod, Ann. Missouri Bot. Gard.. 61. 539 (1974). 10. H.N. Ridley. The dispersal of plants throughout the world. (Reeve and Co., Ltd., 1930). 11. Y. Hagen, in Results Norw. S clent. Exped. Tristan da Cunha. L. Christensen, Ed. (Oslo, 1952) 20, pp. 1-237* 12. A.F. Tryon, Brit. Fern Gaz., £, 269 (1966). 13* F.W. M artin, L.E. Gregory, Crop. S c i., 2, 295 (1962); K.S. Semple, Ann. Missouri Bot. Gard., 61., 868 (1974). 14. H.G. Baker, Am. Naturalist. 93. 255 (1959). 15. C. Skottsberg, Nat. Hist. Juan Fernandez Easter Is. 1_, 193 (1956). 16. H. Valdebenito, unpublished cladistio investigation involving continental and island species. 154 17. R.J. Gornall, B.Bohm, Syafc. Bot.. 1, 353 (1978); J.B. Harborne, Bioohem. Syat. Eool.. 7 (1977). 18. J. Th. Roster, Blumea. 22. 207 (1975). 19. S.H. Sohmer, B ritto n ia . 24 . 283 (1972). 20. S. C arlquist, Island Biology. (Columbia University Press, N.Y., 1974), p. 575. APPENDIX A Bnttonia, 38(1), 1986, pp. 1-3. © 1986, by the New York Botanical Garden, Bronx, NY 10458 A NEW SPECIES OF ERIGERON (COMPOSITAE: ASTEREAE) FROM CHILE H u g o V a l d e b e n it o , T im o t h y K . Lo w r e y , a n d T o d F . S t u e ssy Valdebenito, Hugo (Department of Botany, Ohio State University, Columbus, OH 43210), Timothy K. Lowrey (University of the Witwatersrand, Johannesburg. South Africa), and Tod F. Stuessy (Department of Botany, Ohio State University, Columbus, OH 43210). A new species of Erigeron (Compositae: Astereae) from Chile. Bnttonia 38:1-3.1936.—Erigeron campanensis, sect. Erigeron, is described as new from Central Chile (Cerro Campana, near Valparaiso). It is most similar to E. fasciculatus, from which it differs in leaf shape, leaf arrangement, pubescence of leaves, stems, and phyllaries, and stigmatic branches of disc florets. In the course of systematic investigations of species of Erigeron from the Juan Femandez Islands and adjacent southern South America, it became clear that several collections of what had been called Erigeron fasciculatus Colla represent a discordant taxon worthy of formal recognition. These collections are all from Cerro Campana in central Chile (30°22'S; 71°12'W). The principal differences between Erigeron fasciculatus and the Cerro Campana populations, here named E. campanensis, are in leaf shape, leaf arrangement, pubescence pattern, apex of the stigmatic branches of the disc florets, and length of tube of the ray corollas. Erigeron campanensis has spathulate leaves not in fascicles as in E. fasciculatus, and it has little or no pubescence on stems, leaves, and bracts, which contrasts strongly with the abundant pubescence in the latter. Further, in E. campanensis the apex o f stigmatic branches o f the disc florets is rounded and the length of the tube of the ray corollas is 1.5 to 2 mm long, whereas in E. fasciculatus the stigmatic branches are broadly acute and the corolla tube is 2 to 3 mm long. Although E. fasciculatus does contain morphological variation among its known populations, none of the collections examined in this study extends into the suite of features of this new taxon. In examination of the South American species of Erigeron, Solbrig (1962, p. 25) remarked that: The collections from Cerro Campana near Valparaiso are quite interesting. Although the species has in general slightly to heavily pubescent leaves and stems, the specimens from Cerro Campana are almost completely glabrous. In addition the leaves are spathulate, at least the basal ones, and not in fascicles. Nevertheless, in view of the scarcity of material, it is thought best not to ascribe taxonomic status to this material for the present. More extensive herbarium material now demonstrates geographical, morpho logical, and ecological evidence for the recognition of £. campanensis as a distinct species. Erigeron campanensis Valdebenito, Lowrey, and Stuessy, sp. nov. F ig . A l. Suflrutex asccndens, 31-40 cm altus. Caules basi lignosi, bruneoli, glabri, rhizomalibus gracilibus. Folia altema, spathulala, bruneolo-viridia, 18-22 mm longa, 3-3 mm lata, glabra, apice obtusa, basi attenuata, marginibus integris, brevitercilialis; pctiolus gradalim in laminam dilatatus, 3-3 mm longus, 1-2 mm latus, glabrer. Capitulescentia paniculata (interdum simplex), e 4-10 capitulis constans; pedunculi usque ad 8 cm longi, glabri vel minute glandulosi, bracteati, bracteis lineari-lanceolatis, quasi subulatis, sparsim glanduloso-pubescenlibus; involucrum turbinatum, 5-7 mm altum, 8-10 mm latum; phyllaria 28-36, imbricata, glabra, 2-3-seriata, lineari-lanceolata, 3-8 mm longa, 1 m m lata, bracteis extemis sparsim glandulosis, ad marginem hyalinis luteolis, in centra atroviridibus, laterale 156 BR1TTO N IA 5 mm 5 mm '■ " o Fig« Al • Erigeroncampanensis. campanensis. A. Habit. B. Leaf. C. Head. D. Ray floret. E. Disc floret. D and E same scale. D raw n from Morrison A H'agenknecht 17153 (GH). luteolis. Flosculi radiali 18-27, uniseriati; corollae albae, limbo 4-5 mm longo, 1-2 mm lato, tubo 1.5-2 mm longo. Flosculi disci 32—47 , hermaphroditi; corollae anguste infundibuliformes, S-lobatae, albae, 3.8-4 mm longae, 0.6-0.8 m m diametro; stylus 0.6-0.7 mm longus, apice appendiculis rotun- datis 0.2 mm longo et pilis instructus. Pappus capillaris, albus, 2.5 mm longus. Achenia compressa, pubescentia, 1-1.5 mm long*. Ascending suffruticose sh ru b , 31-40 cm tall. Stems near base woody, light- brown, glabrous, with slender rhizomes. Leaves alternate, spathulate, brownish- 157 VALDEBENITO ET AL.: ERIGERON green, 18-25 mm long, 3-5 mm wide, at apex obtuse, at base attenuate, with margins entire, glabrous, shortly ciliate; petiole gradually expanding into the lam ina, 3-5 m m long, 1-2 mm wide, glabrous. Capitulescence paniculate (occasionally simple), with (1) 4-10 heads; peduncles up to 8 cm long, glabrous or minutely glandular; bracts linear-lanceolate, becoming nearly subulate, sparsely glandular- pubescent; involucre turbinate, 5-7 mm high, 8-10 mm wide; phyllaries 28-36, imbricate, glabrous, 2-3-seriate, linear-lanceolate, 5-8 mm long, 1 mm wide; outer bracts sparsely glandular, with margin hyaline, with central portion of phyllary dark green, and outer portion pale yellow. Ray florets 18-27, uniseriate; corollas white, with limb 4-5 mm long, 1-2 mm wide, and tube 1.5-2 mm long. Disc florets 32-47, hermaphrodite; corollas narrowly funnelform, 5-lobed, white, 3.8- 4 mm long, 0.6-0.8 mm diam; style 0.6-0.7 mm long, with rounded apical ap pendages 0.2 mm long and covered with collecting hairs. Pappus capillary, white, 2.5 mm long. Achenes compressed, pubescent, 1-1.5 mm long. T y p u s : CHILE. Prov. Valparaiso: Cerro Campana, 15 km NE of Granizos, very rocky habitat, 1550 m, 17 Jan 1939, J. Morrison & R. Wagenknecht 17153 (kolotype: UC; is o ty p e s : CONC, GH). Additional specimens examined: CHILE. Prov. Valparaiso: Cerro Campana, 3! Jan 1931, Gar- avenla 1875 (GH), 16 Jan 1932, Garaventa 2797 (CONC), 19 Jan 1936, Garavema 3114 (CONC), 3 Feb 1958, Garaventa 6496 (CONC), 19 Feb 1928, Looser 586 (GH). Erigeron campanensis has been found only on Cerro Campana in very rocky habitats over 1500 m. Although E. campanensis and E. fasciculatus are morpho logically similar, they apparently do not occur sympatrically, the former only growing on Cerro Campana and the latter found along the Chilean coast from Valparaiso to Coquimbo on hills or beach. Acknowledgments Thanks are due to the curators of CONC, GH, and UC for the loan of specimens; David Dennis for drawing Figure 1; Clodomiro Marticorena for a review of the manuscript; Ohio State University for a Postdoctoral Fellowship to T.J.L. during which time the initial phase of this work was done; and to NSF for grant support to T.F.S. (BSR-8306436) and H.V. (INT-8317088). Literature Cited Solbrig, O. T. 1962. The South American species of Erigeron. Contr. Gray Herb. 191: 3-79. LIST OP REFERENCES AL-KHUBAIZI, M. 1977- Sulfated and non-sulfated flavonoids from Flaveria. Sartwellia and Haploesthes. M.A. thesis, University of Texas, Austin. ALLAN, H.H. 1961. Piperaceae. Pp. 173-174 in Flora o f New Zealand .1, ed. H.H. Allan. Wellington: R.E. Owen. ALLARD, R. 1965. Genetic systems associated with colonizing ability in predominantly self-p o llin ated species. Pp. 132-153 in The genetics of colonizing species, eds. H.G. Baker and G.l. Stebbins. New York: Academic Press. ANDERSON, L.C., D.W.KYHOS, T. MOSQUIN, A.M. POWELL and P.H. RAVEN. 1974. Chromosome numbers in Compositae. IX. Haplopappus and other Astereae. Amer. J . Bot. 61:665-671 * ARNOLD, E.N. 1981. Estimating phylogenies at low taxonomic levels. Z. Zoll. Syst. Evolut.-forsch. 19:1-35. BAI, G.V.S. and SUBRAMANIAN, D. 1964. Cytotaxonomical studies of South Indian Piperaceae. Cytologia 50:583-592. BAKER, H.G. 1959* The contribution of autoecological and genecological studies to our knowledge of the past migrations of plants. Am. Naturalist 93:255-272. BATE-SMITH, E.C. 1962. The phenolic constituents of plants and their taxonomic significance. I. Dicotyledons. Bot. J. Linn. Soo. 58:95- 173* BENTHAM, G. 1873* Notes on the classification, history and geographic d istrib u tio n of Compositae. Bot. J. Linn. Soc. 13:335-577* 158 159 BISHOP, M.J. 1982. Criteria for the determination of the direction of character state changes. Zool. J. Linn. Soc. 74:197-206. BLOT, J. 1960. Contribution a 1'etude cytologique du genre Peperomia. Rev. Gen. Bot. 67:522-535. BURGER, W.C. 1971. Piperaceae. Pp. 6-79 in Flora Costaricensis. ed. W.C. Burger; Fieldiana, Bot. 35. CANDOLLE, A.C.P. de. 1869. Piperaceae. Pp.235-471 in Prodromus XVI. ed. A.L.P.P. Candolle. Paris: Victoris Masson et Filii. CARLQUIST, S. 1974. Island Biology. New York: Columbia University Press. CARR, G.D. 1985. Habitat variations in the Hawaiian Madiinae (Heliantheae) and its relevance to generic consepts in the Compositae. Taxon 34:22-25. ______and D*W. KYHOS. 1981. Adaptive rad iatio n in the Hawaiian silversword alliance (Compositae-Madiinae). I. Cytogenetics of spontaneous hybrids. Evolution 35:543-556. and ______. 1986. Adaptive radiation in the Hawaiian silversword alliance (Compo'sltae-Madiinae). II. Cytogenetics of artificial and natural hybrids. Evolution 40:959-979* CHRISTOPHERSEN-, E. 1944. New Phanerogams from Tristan da Cunha. Results Norw. Solent. Exped. Tristan da Cunha 11:1-5. ______. 1968. Flowering plants from T ristan da Cunha. Results Norw. Solent. Exped. T ristan da Cunha 55:1-29 CRAWFORD, D.J. 1978. Flavonoid ohemistry and angiosperm evolution. Bot. Rev. 4if:431-456. 160 , T.F. STUESSY and M. SILVA. 1986. Leaf flavonoid chemistry and the relationships of the Lactoridaceae. PI. Syst. Evol. 153: 133-139. , ______and M. SILVA. 1987. Allozyme divergence and the evolution of Dendroseris (Compositae: Lactuceae) on the Juan Fernandez Islands. Syst. Bot. 12:435-443. , ______and . 1988. Allozyme variation in Chenopodium sanctae-olarae. an endemic species of the Juan Fernandez Islands, Chile. Biochem. Syst. Ecol. 16:279-284. , R. WHITKUS and T.F. STUESSY. 1987. Plant evolution and speciation on Oceanic Islan d s. Pp. 183-199 in : D ifferentiation patterns in higher plants. ed. K.M. Urbanska. London: Academic Press. CRISCI, J.V. and T.F. STUESSY. 1980. Determining primitive character states for phylogenetic reconstruction. Syst. Bot. 5:112-135. CRONQUIST, A. 1947. Revision of the North American species of Erigeron. Brittonia 6:121-302. DAHLSTEDT, H. 1900. Studien uber Sud- und Central-Amerikanische Peperomien. Kongl. Svenska Vetenskapsakad. Handl. 33:1-218. DASGUPTA, A. and P.C. DATTA. 1976. Cytotaxonomy of Piperaceae. Cytologia 41 :697-706. De JONG, R. 1980. Some trends for evolutionary and phylogenetic studies. Z. Zool. Syst. Evolut.-forsch. 18:1-23. Del PERO MARTINEZ, M.A. 1985. Flavonoid relationships between the Commellnaceae and related fam ilies. Biochem. Syst. Ecol. 13:253-256. DIERS, L. 196t. Der Anteil an Polyploiden in den Vegetationsgurtein der Westkordillere Perus. Z. Bot. 49:437-488. 161 DONOGHUE, M.J. and P.D. CANTINO. 1984. The logic and lim itations of the outgroup substitution approach to cladistic analysis. Syst. Bot. 9:192-202. DULL, R. 1973* Die Peperomia-Arten Afrikas. Bot. Jahrb. Syst. 93: 56-129. DUNCAN, T. and B. BAUM. 1981. Numerical phenetics: its uses in botanical systematics. Ann. Rev. Ecol. Syst. 12:387-404. EHLERS, E.G. and H. BLATT. 1982. Petrology, Igneous, sedimentary and metamorphic. San Francisco: H.H. Freeman. FARRIS, J.S . 1970. Methods for computing Wagner tre e s. Syst. Bot. 19:83-92. GANDERS, F.R. and K.M. NAGATA. 1984. The ro le of hybridization in the evolution of Bidens on the Hawaiian Islands. Pp. 179-184 in Plant biosystematics. ed. W.F. Grant. Toronto: Academic Press. GARDNER, R.C. 1976. Evolution and adaptive radiation in Lipochaeta (Compositae) of the Hawaiian islan d s. Syst. Bot. 1:383-391* GIANNASI, D.E. 1978. Systematic aspects of flavonoid biosynthesis and evolution. Bot. Rev. 44:339-429. and D.J. CRAWFORD. 1986. Biochemical systematics. II. A reprise. Evol. Biol. 20:25-248. GILLET, G.W. and E.K.S. LIM. 1970. An experimental study of Bidens (Asteraoeae) in the Hawaiian islands. Univ. California Publ. Bot. 56:1-63. GORMAN, M. 1979. Island ecology. London: Chapman and Hall. G0RNALL, R.J. and B.A. B0HM. 1978. Angiosperm flavonoid evolution: a reappraisal. Syst. Bot. 3:353-368. 162 , ______and R. DAHLGREN. 1979« The distribution o f flavonoids in the angiosperms. Bot. Not. 132:1-30. GRANT, V. 1981. Plant speciation. 1st. edn. New York: Columbia University Press. GROVES, E. 1981. Vascular plant collections from the T rista n da Cunha group of islands. Bull. Brit. Mus. (Nat. Hist.), Bot. 8:333-420. HAGEN, Y. 1952. Birds of Tristan da Cunha. Results Norw. Scient. Exped. T ristan da Cunha 20:1-237. HARBORNE, J.B. 1967- Comparative biochemistry of the flavonoids. New York: Academic Press. . 1972. Evolution and function of flavonoids in plants. Recent Adv. Phytoehem. 4:107-141. . 1975. Flavonoid sulphates: a new class of sulphur compounds in higher plants. Phytochemistry 14:1147-1155. . 1977. Flavonoid sulphates: a new class of natural products of ecological significance in plants. Progr. Phytoehem. 4:189-208. . 1977. Flavonoids and the evolution of the angiosperms. Biochem. Syst. Ecol. 5:7-22. and C. WILLIAMS. 1976. Flavonoid patterns in leaves of the Gramineae. Biochem. Syst. Ecol. 4:267-280. HELENURM, K. and F.R. GANDERS. 1985. Adaptive radiation and genetic differentiation in Hawaiian Bidens. Evolution 39:753-765. HENNIG, W. 1966. Phylogenetic systematics. Translated by D.D. Davis and R. Zangeri. Urbana: University of Illinois Press. HOLMGREN, I. 1919. Zytologisohe Studien uber die Fortpflanzung bei den Gattungen Erigeron und Eupatorium. Kongl. Svenska Vetenskapsakad. Handl. 59:1-118. 163 HUYNH, K.L. 1965. Contribution a l 1etude caryologique et embryologique des phanerogames du Ferou. Denkschr. Schweiz. Naturf. Ges. 85:1-178. JACKSON, B.C. 1971. The karyotype in system atics. Ann. Bev. Ecol. Syst. 2:327-368. JANSEN, B.K. and T.F. STUESSY. 1980. Chromosome counts of Compositae from Latin America. Amer. J. Bot. 67:595-594. JOHNSON, A.W. 1965. The birds of Chile I . Buenos Aires: P latt Establecimientos Graficos, S.A. ______. 1967• The birds of Chile II. Buenos Aires: Platt Establecimientos Graficos, S.A. JOHOW, F. 1896. Estudios sobre la flo ra de la s isla s de Juan Fernandez. Santiago: Imprenta Cervantes. KANETA, M., H. HIKICHI, S. ENDO and N. SUGIYAMA. 1978. Id en tificatio n of flavones in sixteen Compositae species. Agric. Biol. Chem. 42: 457-477. KANTA, K. 1961. Embryolessness in Piper nigrum L. Phytomorphology 11:304-306. KAUL, B.B. 1977* The role of the multiple epidermis in foliar succulence of Peperomia (Piperaceae). Bot. Gaz. 138:213-218. KEIL, D. and T.F. STUESSY. 1975. Chromosome counts of Compositae from the United States, Mexico, and Guatemala. Bhodora 77:171-195. KOSTEB, J.T. 1975. The compositae of New Guinea IV. Additions and corrections to I, II, and III. Blumea 22:207-217. KUNKEL, G. 1957. Uber einige endemisohe Kompositen der Flora fernandezlana. Ber. Schweiz. Bot. Ges. 67:428-457* 164 ______. 1976. Biogeographv and ecology In the Canry Islands. The Hague: W. Junk. LAMMERS, T.G. 1988. Chromosome numbers and their systematic im plications in Hawaiian Lobelioideae (Campanulaceae). Amer. J. Bot. 75:1130-1134. ______f T.F. STUESSY, and M. SILVA. 1986. Systematic relationships or the Lactoridaceae, an endemic family of the Juan Fernandez Islands, Chile. PI. Syst. Evol. 152:243-266. LOWREY, T.K. 1986. A biosystematic revision of Hawaiian Tetramolopium (Composita: A stereae). Allertonia 4:203-265. ______and D.J. CRAWFORD. 1985. Allozyme divergence and evolution in Tetramolopium (Compositae: Astereae) on the Hawaiian Islands. Syst. Bot. 10:64-72. MABRY, T .J ., K. MARKHAM and M. THOMAS. 1970. The systematic identification of flavonoids. New York: Springer-Verlag. MADDISON, W.P., M.J. D0N0GHUE and D.R. MADDISON. 1984. Outgroup analysis and parsimony. Syst. Zool. 33 : 83-103. MARTIN, F.W. and L.E. GREGORY. 1962. Mode of pollination and facto rs affecting fruit set in Piper nigrum L. in Puerto Rico. Crop Sci. (Madison) 2:295-299. MENENDEZ, C.A.- 1972. Estudios paleobotanicos en la Argentina, avances, problemas y perspectivas. Mem. Symp. I Congr. Latino-Amer. Bot. pp. 61-97. MEYER de SCHAUENSEE, R. 1966. The species of birds o f South America and their distribution. Philadelphia: Livingston Publishing Co. MIQUEL, F.A.W. • 1843-1844. Syatema Piperacearum. Rotterdam: H.A. Kramers. MOORE, D.M. and A.O. CHATER. 1971. Studies of bipolar disjunct species I. Car ex . Bot. Not. 124:316-334. 165 . 1972. Connections between cool temperature flo ra s, with p articular reference to Southern South America. Pp. 115-138 in Taxonomy. Phytogeography and Evolution, ed. D.H. Valentine. London: Academic Press. NESOM, G.L. 1980. A revision of the eppapose species of Erigeron (Asteraceae-Astereae). Ph.D. dissertation, University of North Carolina, Chapel H ill. NISHIO, J.N. and I.P . TING. 1987. Carbon flow and metabolic speciation in the tissue layers of the crassulacean acid metabolism in plants, Peperomia camptotricha. Plant Physiol. 84:600-604. OLLIER, C.D. 1984. Geomorphology of South A tlantic volcanic islands. Part I: The Tristan da Cunha Group. Z. Geomorphol. 28:367-382. PACHECO, P., D.J. CRAWFORD, T.F. STUESSY and M. SILVA. 1985. Flavonoid evolution in Robinsonia (Compositae) of the Juan Fernandez Islands. Amer. J. Bot. 72:989-998. PREECE, R.C., K.D. BENNETT, and J.R. CARTER. 1986. The Quaternary paleobotany of Inaccessible Island (Tristan da Cunha group). J. Biogeogr. 13:1—33 - QUENSEL, P. 1920. Additional comments on the geology of the Juan Fernandez Island?. Pp. 37-87 in The natural history of Juan Fernandez and Easter Islan. ed. C. Skottsberg. Uppsala: Almqvist and Wiksells. QUISUMBING, E. 1930. Philippine Piperaceae. Philipp. J. Sci. 43: 1-246. RAVEN, P.H. and D .I. AXELROD. 1974. Angiosperm biogeography and past continental movements. Ann. Missouri Bot. Gard. 61:539-673* RIDLEY, H. 1930. The dispersal of plants throughout the world. Ashford: Reeve and Co. 166 ROHLF, J . , J. KISHPAUGH, and D. KIRK. 1972. NT-SYS numerical taxonomy system of m ultivariate s ta tis tic a l programs. Stony Brook: State University of New York. RUIZ, H. and J. PAVON. 1797- Flora peruvianae et chilensis prodromus. 2nd edn. Rome. SANDERS, R.W., T.F. STUESSY and C. MARTICORENA. 1982. Recent changes in the flora of the Juan Fernandez Islands, Chile. Taxon 31:284-289. , ______, ______and M. SILVA. 1987* Phytogeography and evolution of Dendroseris and Robinsonia, tree-Compositae of the Juan Fernandez Islands. Opera Bot. 92:195-215. , ______and R. RODRIGUEZ. 1983. Chromosome numbers from the flora of the Juan Fernandez Islands. Amer. J. Bot. 70:799-810. SCHLATTER, R.P. 1987. Conocimiento y situacion de la ornitofauna en las islas oceanicas Chilenas. Pp. 271-285 in Islas oceanicas Chilenas: conocimiento cientifico y necesidades de investigaciones. ed. J.C. Castilla. Santiago: Ediciones Universidad Catolica de Chile. SCHMIDT, J.E . 1987. Response of succulent leaves of Peperomia magnoliaefolia to dehydration. Plant Physiol. 83:190-194. SEMPLE, S.K. 1974. Pollination in Piperaceae. Ann. Missouri Bot. Gard. 61:868-871. SKOTTSBERG, C. 1922. The phanerogams of the Juan Fernandez Islands. Pp. 95-240 in Nat, h i s t . Juan Fernandez. 2: Botany, part 2. Uppsala: Almqvist & Wlksells. ______. 1946. Peperomia berteroana Miq. and P. trista n e n sis C hristoph., an in terestin g case of disjunction. Acta Hortl Gotob. 16: 251-288.- . 1947. The genus Peperomia in C hile. Aota Horti Gotob. 17:1-47. 167 ______• 1953. The vegetation of the Juan Fernandez islands. Pp. 792-960 in Nat, h is t. Juan Fernandez. 2: Botany, p art 6. Uppsala: Almqvist & Wiksells. . 1955. Chromosome numbers in Hawaiian flowering plants. Ark. Bot. 3:63-70. . 1956. Derivation of the flora and fauna of Juan Fernandez and Easter island. Pp. 193-437 in Nat, h i s t . Juan Fernandez. I: Geography, geology, origin of islan d l i f e , part 5. Uppsala: Almqvist & W iksells. SMITH, J.B. 1966. Chromosome numbers in Peperomia Ruiz y Pavon (Piperaceae) and notes on chromosome number of Piper magnifioum Trelease. Kew Bull. 20:521-526. SNEATH, P.H. A. and R.R. SOKAL. 1973. Numerical taxonomy. San Francisco: W.H. Freeman and Co. SOHMER, S.H. 1972. Revision of the genus Charpentiera (Amaranthaceae) Brittonia 24:283-312. SOLBRIG, O.T. 1962. The South American species of Erigeron. Contr. Gray Herb. 191:3-7 9. , L. C. ANDERSON, D.W. KYHOS and P.H. RAVEN. 1969. Chromosome numbers in Compositae VII: Astereae III. Amer. J. Bot. 56:348-353* , .______, ______and L. RODENBERG. 1964. Chromosome numbers in Compositae. V. A stereae. I I . Amer. J. Bot. 51:513-519. SPOONER, D.M., T.F. STUESSY, D.J. CRANFORD and M. SILVA. 1987. Chromosome numbers from the f lo r a of the Juan Fernandez Islands. I I . Rhodora 89:351-356. STEVENS, P.P. 1970. Agaurla and Agarista: An example of tropioal transatlantic affinity. Notes Roy. Bot. Gard. Edinburgh 30:341-359* 168 . 1980. Evolutionary polarity of character states. Ann. Rev. Ecol. Syst. 11:333-358. STUESSY, T.F. 1971* Chromosome numbers and phylogeny in Melampodium (Compositae). Amer. J. Bot. 58:732-736. and D.J. CRAWFORD. 1983. Flavonoids and phylogenetic reconstruction. PI. Syst. Evol. 143:83-107* , ______and C. MARTICORENA. In press. Patterns o f phylogeny in the endemic vascular flora of the Juan Fernandez islands, Chile. , and J.V. CRISCI. 1984. Problems in the determination of evolutionary directionality of character-state change for phylogenetic reconstruction. Pp. 71-87 in Cladistics: Perspectives on the reconstruction of evolutionary history, eds. T. Duncan and T.F. Stuessy. New York: Columbia University Press. , K.A. FOLAND, J.F. SUTTER, R. SANDERS, and M. SILVA. 1984. Botanical and geological significance of Potassium-Argon dates from the Juan Fernandez Islands. Science 225:49-51 . , C. MARTICORENA, D.J. CRAWFORD and M. SILVA. In p ress. Plant conservation in the Juan Fernandez Islands. In Plant conservation in oceanic islands, ed. D. Bramwell. London: Academic Press. , R.W. SANDERS and M. SILVA. 1984. Phytogeography and evolution of the Juan Fernandez Islands: a progress report. Pp. 55- 69 in Biogeoaraphy of the Tropical Pacific, e d s. F.J. Radovsky, P.H. Raven and S.H. Scbmer. Kansas: A.S.C. and B.P. Bishop Museum. SWOFFORD, D.L. 1985. PAUP. Phylogenetic analysis using parsimony version 2.4. Champaign: Illinois Natural History Survey. TING, I . , L.A. BATES, L. O'REILLY and M.J. DENIRO. 1985. Physiological and isotopio aspects of photosynthesis in Peperomia. Plant Physiol. 78:246-249. TRELEASE, W. 1930. The geography of American peppers, Proc. Amer. Philos. Soc. 69:309-327. 169 and T.G. YUNCKER. 1950. The Plperaceae of northern South America. Urbana: University of Illin o is Press. TRYON, A.F. 1966. Origin of the fern flora of Tristan da Cunha. Brit. Fern Gaz. 9:269-276. TURNER, B.L., V.L. ELLISON and R.M. KING. 1961. Chromosome numbers in the Compositae. IV. North American species with phyletic interpretations. Amer. J. Bot. 48:216-223. , A.M. POWELL and J. CUATRECASAS. 1967. Chromosome numbers, in Compositae. XI. Peruvian species. Ann. Missouri Bot. Gard. 54:172- 177. , ______and R.M. KING. 1962. Chromosome numbers in the Compositae. VI. Additional Mexican and Guatemalan species. Rhodora 64:251-271. VALDEBENITO, H., T.K. LOWREY and T.F. STUESSY. 1986. A new species of Erlgeron (Compositae: Astereae) from Chile. Brittonia 38:1-3. VENABLE, D.L. and D.A. LEVIN. 1983* Morphological dispersal structures in relation to growth habit in the Compositae. PI. Syst. Evol. 143:1-16. VICUNA MACKENNA, B. 1883* Juan Fernandez. H istoria verdadera de la isla de Robinson Crusoe. 2 vols. Santiago. VIRZO Oe SANTO, A., A. ALFANI, G. RUSSO and A. FIORETTO. 1983. Relationship between CAM and succulence in some species of Vitaceae and Piperaceae. Bot. Gaz. 144:342-346. WAGNER, W. 1980. Origin and philosophy of the groundplan divergence method of dadistios. Syst. Bot. 30:1-11. WATROUS, L. and Q. WHEELER. 1981. The out-group comparison of character analysis. Syst. Zool. 30:1-11. 170 WHEELER, W.M. 1924. Formicidae from Easter Island and Juan Fernandez. Pp. 419 In The natural history of Juan Fernandez and Easter Island, ed. C. Skottsberg. Uppsala: Almqvist and Wiksells. WOODWARD, R., J r. 1969• Robinson Crusoe's Isla n d . Chapel H ill: University of North Carolina Press. YOUNG, D.A. 1981. The usefulness of flavonoids in angiosperm phylogeny: some selected examples. Pp. 205-232 in Phytochemistry and angiosperm phylogeny. eds. D.A. Young and D.S. Sieger. New York: Praeger S c ie n tific . YUNCKER, T.G. 1933. Revision of the Hawaiian species of Peperomia. Bernice P. Bishop Mus. Bull. 112:1-133. . 1936. Peperomia. Pp.25-47 in Fi.Han plant studies, ed. A. Smith. Bernice P. Bishop Mus. Bull. 141. ______. 1938. Revision of the Micronesian species of Peperomia. Occas. Pap. Bernice Pauahi Bishop Mus. 14:7-25. . 1943. Piperaceae. Pp. 44-45 in The flora of Niue Island. ed. T.G. Yuncker. Bernice P. Bishop Mus. B ull. 178. . 1959- Piperaceae of Micronesia. Occas. Pap. Bernice Pauahi Bishop Mus. 22:83-108. ______. 1974. The Piperaceae of Brazil I I I . Peperomia: taxa of uncertain status. Hoehnea 4:71-413 • and GRAY, W.D. 1934. Anatomy of Hawaiian Peperomias. Occas. Pap. Bernice Pauahi Bishop Mus. 10:1-19*