Origins of dioecy in the Hawaiian flora. (2024)

INTRODUCTION

The evolution of separate male and female plants in populations(dioecy) has occurred independently in many floras and in diverse taxa,and its repeated evolution has been of particular interest. The majorityof flowering plant species are hermaphroditic, and worldwide only[approximately equal to] 4% of flowering plant species are dioecious(Yampolsky and Yampolsky 1922). The incidence of dioecy variesconsiderably in different regional floras (summarized in Steiner 1988),including values as low as 2.8% in California (Fox 1985) to 12-13% ofspecies in New Zealand (Godley 1979; 18% of genera, Lloyd 1985; also seeWebb and Kelley 1993). The Hawaiian flora is of particular interest forstudies of dioecy. Gilmartin (1968, using Hillebrand's florawritten in 1888) reported only 5% dioecy in the Hawaiian flora, but morerecently Carlquist (1974), using data from a variety of sourcesincluding his own investigations, reported that 27.7% of the nativeHawaiian angiosperm species and varieties were dioecious, a figure twiceas high as that for the next highest flora of New Zealand.

Hypotheses on selective-forces promoting the evolution of dioecyinclude those that suggest that dioecy has evolved as a mechanism toavoid inbreeding depression as well as those that suggest that resourceallocation, sexual selection, and ecological factors are important(reviewed in Bawa 1980, Thomson and Brunet 1990). Because of its highfrequency of dioecy, the Hawaiian flora has been cited as criticalevidence in support of some theories on factors promoting the evolutionof dioecy (Baker 1967, Carlquist 1974, Bawa 1980, Thomson and Barrett1981, Baker and Cox 1984). Carlquist (1966, 1974) suggested that theadvantages of outcrossing with dioecy were sufficiently high in insularhabitats that they outweighed the disadvantages of needing individualsof both sexes to establish populations after long distance dispersal. Asa consequence, Carlquist suggested that the high-incidence of dioecy inHawaii was in part a result of dioecious colonists. In contrast, Baker(1967) contended that self-compatible hermaphrodites were much morelikely to colonize after long-distance dispersal because a singlepropagule was sufficient to start a population (Baker's law). Hesuggested that the high incidence of dioecy in the Hawaiian Islands wasthe result of autochthonous (in situ) evolution of dioecy (Baker 1967),although in later work (Baker and Cox 1984) he suggested severalmechanisms that allow establishment by dioecious colonists. Thomson andBarrett (1981) suggested that the high levels of autochthonous evolutionof dioecy in the Hawaiian Islands support the importance of outcrossingas a factor. Bawa (1980) used the Hawaiian flora to support hishypothesis of a correlation of dioecy with pollination by smallgeneralist insects and with fruit dispersal by birds.

Analysis of the Hawaiian flora can offer special insights into theevolution of dioecy because the great isolation of the archipelago([approximately equal to]4000 km from the nearest large mass of NorthAmerica) has limited the number of angiosperm colonists. Previousestimates suggest that only [approximately equal to]272-282long-distance colonists gave rise to the current native flowering plants(Fosberg 1948, Wagner et al. 1990, Wagner 1991). Phylogeneticconsiderations that have presented problems in analyses of other floras(Donoghue 1989) can be addressed by analysis of presumed colonists aswell as extant species. With hypotheses about the colonists'breeding systems and lineages descended from these colonists, in lieu ofmore detailed phylogenies for most taxa, it is possible to distinguishcurrent dioecious species that arose from dioecious colonists from thosespecies where dioecy evolved autochthonously (in situ) within theHawaiian Islands.

We analyzed current taxonomic information (Wagner et al. 1990; seealso Sakai et al. 1995) on the breeding systems of known (extant andrecently extinct) native species of the Hawaiian Islands as well as thebreeding systems and lineages of colonists of the Hawaiian Islands withtwo objectives. The first objective was to report breeding systemdistributions of the Hawaiian flora, in light of recent systematic work,including better knowledge of breeding systems. The second objective wasto distinguish whether the high incidence of dioecy in the Hawaiianflora results from (1) high rates of successful colonization bydioecious colonists (of endemic and indigenous species), (2) greaternumbers of species in dioecious lineages than hermaphroditic lineages,or (3) evolution of dioecy in situ from hermaphroditic colonists. In thelatter case, comparison of the ecological conditions associated withdioecious and hermaphroditic species may be especially relevant forascertaining causal factors in the evolution of dioecy (see also Sakaiet al. 1995).

TABLE 1. Comparisons of breeding system distributions fromCarlquist(1974) and Wagner et al. (1990). Letters represent the sex of theflowers (M = male, F = female, H = hermaphroditic) and parenthesesindicate the types of flowers found on the same plant. Ellipsesindicate categories that were not included in Carlquist's (1974)analysis; N/A = not applicable.

 Species Genera

Breeding system 1974 1990 19741990

Dioecy (M) (F) 27.7 14.7 15.311.4Gynodioecy (F) (H) 2.6 3.8 2.71.3Subdioecy (M) (F) (rare H) ... 0.6 ...0Polygamodioecy (M, rare H) (F, rare H) 0.4 1.6 0.91.8Hermaphroditism (H) 56.9 62.4 64.462.0Monoecy (M, F) 5.0 7.6 7.210.9Andromonoecy (M, H) 2.5 4.5 4.15.2Gynomonoecy (F, H) 4.4 3.9 3.62.6Polygamomonoecy (M, F, rare H) 0.5 0.1 1.80

Dimorphism 30.7 20.7 18.914.4Monomorphism 69.3(*) 78.7 81.180.8Unknown ... 0.5 ...0.4Mixed genera (dimorphic and monomorphic spp.) N/A N/A ...4.4N 1490(*) 971 222229

* Values were calculated using the sum of individual sexualconditions (Carlquist, 1974: Table 13.1) rather than his dimorphictotal (based on N = 1449) or the data given (N = 1430). Note thatCarlquist's (1974) totals include species plus varieties; our 1990data include species but do not count infraspecific taxa.

MATERIALS AND METHODS

Information was taken from the Manual of the Flowering Plants ofHawai'i (Wagner et al. 1990; referred to hereafter as the Manual)with some updating (see Appendix; see also Sakai et al. 1995).Terminology of breeding systems follows that given in the Manual.Following Lloyd (1980), we also use the term sexually dimorphic tocollectively refer to taxa with dioecious, subdioecious,polygamodioecious, or gynodioecious systems; monomorphic refers to taxawith hermaphroditic, monoecious, andromonoecious, gynomonoecious, orpolygamomonoecious systems (see Table 1 for definitions). Breedingsystems were generally taken as those listed in the Manual. If data onbreeding systems in the Manual were ambiguous, specimens (BISH, US) werere-examined or authorities for those taxa were consulted when possible(Schiedea, A. Sakai and S. Weller; Wikstroemia, S. Mayer; Bidens, F.Ganders). Five species were omitted from analyses because their breedingsystems were unknown; Poa also was excluded from the generic-levelanalysis because the breeding systems of the native species are notknown. Most genera were easily classified with respect to breedingsystem because species within them all had the same breeding system orat least were all monomorphic or all dimorphic. Genera comprised ofspecies with both dimorphic and monomorphic breeding systems wereclassified as mixed.

Native species included endemic species (found naturally only in theHawaiian Islands), indigenous species (found naturally in the HawaiianIslands as well as elsewhere), and also those species that were noted inthe Manual as questionably naturalized (i.e., possibly native orquestionably introduced by colonizing Polynesians). Each indigenousspecies was counted as one colonization, even for species such asScaevola sericea that have colonized the islands on multiple occasions.Presumed original colonists were derived from consideration of twoprevious estimates of colonization events in the Hawaiian archipelago(Fosberg 1948, Carlquist 1974), and from phylogenetic relationshipsreported in the Manual by over 50 contributors, including more recentinformation communicated to us by them. Colonists for taxa without aspecialist were inferred from morphological studies conducted (by W. L.Wagner and D. R. Herbst) in preparation of the Manual and fromconsultation of taxonomic works with more general discussions ofrelationships. Explicit phylogenetic discussion of Hawaiian angiospermlineages has begun to emerge only recently (Baldwin et al. 1990; Wagnerand Funk 1995; Weller et al. 1995; F. Ganders, unpublished manuscript).

In the absence of more specific information, we assigned the breedingsystem of the colonist based on genera or species related to the endemicHawaiian taxa; in most cases, however, the closest sister taxon of theHawaiian species is unknown. For nonendemic genera, we used the generalconditions present in extra-Hawaiian species, unless more specificrelationships within the genus could be determined. In more difficultcases, the breeding system of the colonist was assigned only a moregeneral designation (e.g., monomorphic or dimorphic). Ten colonists hadunknown breeding systems and were omitted from analyses of breedingsystems, thus making our estimate of dimorphism in the colonists aconservative one.

Because the data set used in this paper is not available in anysource in its entirety, a comprehensive list of the presumed colonistsand their resulting Hawaiian lineages is listed in the Appendix. TheAppendix also includes additional attributes for each presumed colonistthat relate to ecological correlates of breeding systems in the Hawaiianflora (Sakai et al. 1995). Presumed colonists with breeding systems thatwere difficult to determine or where our determination differed fromearlier works (e.g., Carlquist 1974, Bawa 1982; others in Appendix) arediscussed in more detail in the notes to the Appendix.

RESULTS

Of the 971 native species, 14.7% are dioecious and 20.7% aredimorphic, proportions that are the highest of any flora studied, butfar lower than Carlquist's earlier estimates that includedinfraspecific taxa (1974: Chapter 13; Table 1). Our results differbecause we did not use infraspecific taxa (there was no variation inbreeding system at that level), and because recent taxonomic changes(Wagner et al. 1990) reduced the total number of both hermaphroditic anddioecious taxa, but disproportionately affected dimorphic taxa,especially infraspecific taxa in Loganiaceae, Pittosporaceae, Rubiaceae,and Rutaceae. Changes also result in small part to more detailed studiesof breeding systems. At the generic level, 11.4% are dioecious; 14.4%are dimorphic. Strictly dioecious genera have a mean of 3.04species/genus (N = 24, SD = 4.03); strictly hermaphroditic genera have amean of 3.65 species/genus (N = 142, SD = 6.27).

The 971 native Hawaiian species are the result of speciation from 291presumed colonists, a number slightly higher than that of previousestimates (Fosberg 1948, Carlquist 1974, Wagner et al. 1990, Wagner1991). Six colonists gave rise to more than one genus, and in 45 genera,the species are the result of more than one colonization. Two-thirds ofthe colonists (194/ 291) are represented by only a single species. Ofthe 119 indigenous colonists, all but 10 are represented by only onespecies. From those 10 colonists [Lepidium (Brassicaceae), Gahnia andMariscus (Cyperaceae), Eugenia (Myrtaceae), Boerhavia and Pisonia(Nyctaginaceae), Peperomia (Piperaceae), Portulaca (Portulacaceae), and2 colonists of Korthalsella (Viscaceae)], indigenous species arepresumed to have given rise directly to endemic species and bothindigenous and endemic species are extant.

At the other extreme, one colonist in the Campanulaceae has givenrise to four genera (Clermontia, Cyanea, Delissea, and Rollandia) with atotal of 91 species ([greater than]9% of the total flora) and the onecolonist of Stenogyne, Phyllostegia and Haplostachys (Lamiaceae) hasresulted in 53 species. Other species-rich lineages include those forMelicope (47 species, Rutaceae), the silversword complex (28 species ofDubautia, Argyroxyphium, and Wilkesia, Asteraceae), the HawaiianAlsinoideae (Schiedea and Alsinidendron, 26 species, Caryophyllaceae),Hedyotis (20 species, Rubiaceae), Myrsine (20 species, Myrsinaceae), andone colonist of Peperomia (20 species, Piperaceae).

Dimorphic colonists resulting in only dimorphic Hawaiian speciesaccount for 10% (29/291) of the colonists, suggesting that dimorphism ishigh in part because colonists were dimorphic (Table 2). Over half(111/201) of current dimorphic species arose from dimorphic colonists.Monomorphic colonists resulting in only monomorphic species constitute82% (238/291) of the colonists. Of the colonists leading to endemic[TABULAR DATA FOR TABLE 2 OMITTED] species, 11% (19/173 colonists) weredimorphic and gave rise to only dimorphic species. Of the indigenouscolonists, 8% (10/119 colonists) were dimorphic. Colonists that gaverise to endemic species were no more likely than indigenous colonists tobe dimorphic (N = 285, [mean] = 0.64, df = 1, P = 0.43).

Although 10% of the colonists were dimorphic, 20.7% of currentspecies are dimorphic, indicating that either dimorphic lineages haveled to more species/ colonist, or that dimorphism has arisenautochthonously [TABULAR DATA FOR TABLE 3 OMITTED] (in situ). Ouranalysis suggests that the latter is true; approximately one-third(31.8%) of current dimorphic species arose from monomorphic colonists.At least 12 monomorphic colonists evolved dimorphism autochthonously[the Hawaiian Alsinoideae (Schiedea), Bidens, Broussaisia, Cyrtandra,Hedyotis, Neraudia, Perrottetia, 2 Psychotria (Rubiaceae) colonists,Psydrax, Santalum, and Wikstroemia (Table 3)]. In Neraudia, dioecyappears to have evolved from monoecy. In the two Psychotria colonists,separate sexes probably were derived from heterostyly. In the other ninecolonists, separation of the sexes appears to have evolved fromhermaphroditism or in lineages with hermaphroditism and gynodioecy. All12 of these lineages (with the exception of Psydrax) contain onlyendemic species. In five other cases, dimorphism may have evolved frommonomorphism, but we have conservatively listed the breeding system ofthe colonist as unknown (Table 3).

Other colonists had changes in breeding system within monomorphic orwithin dimorphic systems (Table 3). In two cases [Rhus (Anacardiaceae)and Melicope (Rutaceae)], evolution in the opposite direction apparentlyhas occurred, and hermaphroditism has arisen from a presumablyfunctionally dimorphic ancestor.

Lineage size was similar for dimorphic colonists that gave rise toonly dimorphic species ([mean] = 2.3 dimorphic species/colonist) andmonomorphic colonists that gave rise to only monomorphic species ([mean]= 2.9 monomorphic species/colonist; Fig. 1), but the 12 monomorphiccolonists that gave rise to dimorphic species were significantlydifferent in lineage size ([mean] = 9.25 species/ colonist, Fig. 1; N =267, df = 2, [mean] = 11.0, P = 0.004, Kruskal-Wallis test). A number ofcolonists with unknown breeding systems could affect this latterdistribution if they were to be included. Because of this, we did nottry to distinguish why the lineages that evolved dimorphism wereapparently larger. Lineages that evolved dimorphism may be largerbecause factors associated with speciation may also favor the evolutionof dimorphism in species-rich lineages. Alternatively, each species mayhave a similar probability of evolving dimorphism regardless of lineagesize, and thus larger lineages will tend to have more autochthonousevolution of dimorphism, simply because they have more species.

DISCUSSION

Our work shows that the incidence of dioecy in the Hawaiian Islandsis not as high as originally estimated, but remains the highest of anyflora where similar data are available. The percentage of dioeciousspecies is slightly higher than that of New Zealand (12-13%), anotherinsular flora with some tropical elements (Godley 1979, Webb and Kelly1993). The high incidence of dioecy in the Hawaiian Islands is theresult of a number of dimorphic colonists as well as autochthonousevolution of dimorphism within the archipelago. Of the 291 presumedcolonists, 10% were dimorphic, considerably higher than the worldwideaverage for dioecy of [approximately equal to]4% (Yampolsky andYampolsky 1922). Over half of the native Hawaiian flora has Malesian,Austral, or Pacific affinity (Fosberg 1948; W. L. Wagner, unpublisheddata), and this higher percentage of dioecy may reflect a higherincidence of dimorphism in the source floras, although these areas(especially Malesia) are not well studied. Other sources includepantropical elements as well as temperate areas (North America,Australia, New Zealand) and a few boreal elements (Wagner et al. 1990).Further studies of breeding systems (particularly in tropical sourcefloras) are needed to determine if dioecious colonists areover-represented relative to the source flora, or conversely, ifhermaphroditic taxa are disproportionately represented as colonists, aspredicted by Baker's (1967) law. The number of dimorphic coloniststo the Hawaiian Islands, however, suggests that dioecy has not been aseverely limiting factor in dispersal and colonization of the HawaiianIslands. Lloyd (1985) also found that most of the dimorphism in the NewZealand flora resulted from dimorphic colonists; only 5 of the 72dimorphic genera were the result of autochthonous evolution ofdimorphism.

The high incidence of dimorphism in the Hawaiian Islands has notresulted from different rates of speciation of dimorphic and monomorphiccolonists as suggested by Bawa (1982), at least as measured by thecurrent number of species per colonist. Two-thirds of the colonizationsresulted in only one species, but in a few notable cases (Asteraceae,Campanulaceae, Caryophyllaceae, Lamiaceae, Myrsinaceae, Rubiaceae, andRutaceae), colonizations have led to a remarkable number and diversityof species, for both dimorphic and monomorphic colonists. Carlquist(1974: 526) reported that the average number of species per genus wastwice as great in dioecious genera as the flora at large; in contrast,we found similar numbers of species per genus for monomorphic anddimorphic genera, with apparently more species per colonist in lineageswhere dimorphism has arisen in situ. This difference related in part totaxonomic differences in the data sets used. Further studies are neededto elucidate whether evolution of dimorphism has been associated withchanges in addition to breeding system that resulted in greaterspeciation in these lineages than in lineages of colonists that did notevolve dimorphism.

Although over half of the dimorphism in current species is the resultof dimorphic colonists, selective pressures have presumably beensufficient to promote a diversity of pathways to dimorphism in theHawaiian flora, and about one-third of the dimorphic species occur inlineages from a monomorphic colonist. Dioecious and dimorphic breedingsystems apparently have been derived from heterostyly, from monoecy, ormost commonly directly from hermaphroditism or from hermaphroditism viagynodioecy. In two cases, hermaphroditism apparently arose from afunctionally dimorphic ancestor. Further study of the presumed colonist,phylogeny, and breeding system of these taxa (Rhus, Melicope) is needed.In general, the assumption has been that evolution of breeding systemsaway from dioecy is extremely difficult, and very few cases of this havebeen documented [e.g., androdioecy from dioecy in Datisca (Datiscaceae),Rieseberg et al. 1992]. Hermaphroditic plants in some Hawaiianpopulations of Wikstroemia (Thymeliaceae) may be secondarily derived ashybrids from individuals with different modes of control of dioecy orfrom the breakdown of dioecy (Mayer and Charlesworth 1992). Within theHawaiian Alsinoideae, some species may have secondarily evolvedhermaphroditism from gynodioecious ancestors (Wagner et al. 1995, Welleret al. 1995). Strong self-incompatibility in the Hawaiian flora is knownonly in one lineage of the Hawaiian Madiinae (Compositae; Carr et al.1986). In general, the evolution of dioecy and the apparent lack ofself-incompatibility in the endemic flora support the notion that dioecymay be easier to evolve than self-incompatibility (Thomson and Barrett1981, Charlesworth 1985), although few Hawaiian taxa have beeninvestigated for the occurrence of self-incompatibility.

The limited number of lineages comprising the Hawaiian angiospermflora also creates a unique opportunity for detailed studies withinlineages of the number of independent origins of dioecy and associatedtraits (e.g., Sakai et al. 1995, Wagner et al. 1995, Weller et al.1995). Better knowledge of phylogenetic patterns and further ecologicalstudies, particularly within those groups evolving dioecyautochthonously, are needed to determine causality.

ACKNOWLEDGMENTS

This work was supported in part by NSF grants BSR88-17616,BSR89-18366, DEB92-07724 (S. G. Weller and A. K. Sakai, Co-P.I.s), witha Research Experience for Undergraduates supplement for D. M. Ferguson.It was also supported by a Smithsonian Institution Scholarly Studiesgrant to W. L. Wagner. We wish to thank R. Shannon for help withword-processing, F. Ganders for providing information on Bidens, and S.Mayer for information on Wikstroemia, S. Weller for helpful commentsthroughout the project, and D. Charlesworth, J. Brunet, and V. Eckhartfor manuscript review.

LITERATURE CITED

Backer, C. A., and R. C. Bakhuizen van den Brink, Jr. 1968. Flora ofJava (Spermatophytes only) III. Wolters-Noordhoff N. V., Groningen, TheNetherlands.

Baker, H. G. 1967. Support for Baker's law - as a rule.Evolution 21:853-856.

Baker, H. G., and P. A. Cox. 1984. Further thoughts on dioecism andislands. Annals of the Missouri Botanical Garden 71:244-253.

Baldwin, B. D., W. Kyhos, and J. Dvorak. 1990. Chloroplast DNAevolution and adaptive radiation in the Hawaiian silversword alliance(Asteraceae - Madiinae). Annals of the Missouri Botanical Garden77:96-109.

Bawa, K. S. 1980. Evolution of dioecy in flowering plants. AnnualReview of Ecology and Systematics 11:15-40.

-----. 1982. Outcrossing and the incidence of dioecism in islandfloras. American Naturalist 119:866-871.

Brown, F. B. H. 1935. Flora of southeastern Polynesia. III.Dicotyledons. Bernice P. Bishop Museum Bulletin 130:1-386.

Carlquist, S. 1966. The biota of long-distance dispersal. IV. Geneticsystems in the floras of oceanic islands. Evolution 20:433-455.

-----. 1974. Island biology. Columbia University Press, New York, NewYork, USA.

Carr, G. D., E. A. Powell, and D. W. Kyhos. 1986.Self-incompatibility in the Hawaiian Madiinae (Compositae): an exceptionto Baker's rule. Evolution 40:430-434.

Charlesworth, D. 1985. Distribution of dioecy andself-in-compatibility in angiosperms. Pages 237-268 in J. J. Greenwoodand M. Slatkin, editors. Evolution - essays in honour of John MaynardSmith. Cambridge University Press, Cambridge, Great Britain.

Conn, B. J. 1980. A taxonomic revision of Geniostoma subg. Geniostoma(Loganiaceae). Blumea 26:245-364.

Ding Hou. 1978. Anacardiaceae. Flora Malesiana I. 8(3): 395-548.

Donoghue, M. J. 1989. Phylogenies and the analysis of evolutionarysequences, with examples from seed plants. Evolution 43:1137-1156.

Eliasson, U. 1988. Floral morphology and taxonomic relations amongthe genera of Amaranthaceae in the New World and the Hawaiian Islands.Botanical Journal of the Linnean Society 96:235-283.

Fosberg, F. R. 1939. Taxonomy of the Hawaiian genus Broussaisia(Saxifragaceae). Occasional Papers of the Bernice P. Bishop Museum15(4):49-60.

-----. 1948. Derivation of the flora of the Hawaiian Islands. Pages107-119 in E. C. Zimmerman. Insects of Hawaii. Volume 1. University ofHawaii Press, Honolulu, Hawaii, USA.

Fox, J. F. 1985. Incidence of dioecy in relation to growth form,pollination and dispersal. Oecologia 67:244-249.

Gardner, R. C. 1979. Revision of Lipochaeta (Compositae: Heliantheae)of the Hawaiian Islands. Rhodora 81:291-343.

Gilmartin, A. J. 1968. Baker's law and dioecism in the Hawaiianflora: an apparent contradiction. Pacific Science 22: 285-292.

Godley, E. J. 1979. Flower biology in New Zealand. New ZealandJournal of Botany 17:441-466.

Hanco*ck, J. F., Jr., and R. S. Bringhurst. 1979. Hermaphroditism inpredominately dioecious populations of Fragaria chiloensis (L.) Duchn.Bulletin of the Torrey Botanical Club 106:229-231.

Hanco*ck, J. F., Jr., and R. S. Bringhurst. 1980. Sexual dimorphism inthe strawberry Fragaria chiloensis. Evolution 34:762-768.

Hillebrand, W. 1888. Flora of the Hawaiian Islands: a description oftheir phanerogams and vascular cryptogams. Carl Winter, Heidelberg,Germany. Facsimile edition, 1981, Lubrecht & Cramer, Monticello, NewYork, USA.

Hutchinson, J. 1967. The genera of flowering plants (Angiospermae).Dicotyledones. Volume II. Clarendon, Oxford, England.

Lam, H. J. 1938. Monograph of the genus Nesoluma (Sapotaceae). Aprimitive Polynesian endemic of supposed Antarctic origin. OccasionalPapers of the Bernice P. Bishop Museum 14(9): 127-165.

Lloyd, D. G. 1980. Sexual strategies in plants: a quantitative methodfor describing the gender of plants. New Zealand Journal of Botany18:103-108.

-----. 1985. Progress in understanding the natural history of NewZealand plants. New Zealand Journal of Botany 23:707-722.

Mayer, S., and D. Charlesworth. 1992. Genetic evidence for multipleorigins of dioecy in the Hawaiian shrub Wikstroemia (Thymelaeaceae).Evolution 46:207-215.

Molina, Ana Maria. 1978. El genero Gunnera en la Argentina y elUruguay (Gunneraceae). Darwiniana 21:473-489.

Oliver, W. R. B. 1935. The genus Coprosma. Bernice P. Bishop MuseumBulletin 132:1-207.

Rieseberg, L. H., M. A. Hanson, and C. T Philbrick. 1992. Androdioecyis derived from dioecy in Datiscaceae: evidence from restriction sitemapping of PCR-amplified chloroplast DNA fragments. Systematic Botany17:324-336.

Sakai, A. K., W. L. Wagner, D. M. Ferguson, and D. R. Herbst. 1995.Biogeographical and ecological correlates of dioecy in the Hawaiianflora. Ecology 76:000-000.

Skottsberg, C. 1945. The flower of Canthium. Arkiv for botanikutgivet av k. svenska vetenskapsakademien 32A(5): 1-12.

Smith, A. C. 1988. Flora vitiensis Nova. A new flora of Fiji(Spermatophytes only). Volume 4. Pacific Tropical Botanical Garden,Lawai, Hawaii, USA.

Sohmer, S. H. 1972. Revision of the genus Charpentiera(Amaranthaceae). Brittonia 24:283-312.

Staudt, G. 1962. Taxonomic studies in the genus Fragaria.Typification of Fragaria species known at the time of Linnaeus. CanadianJournal of Botany 40:869-886.

-----. 1967. Die Genetik und evolution der heterozie in der gattungIII. Untersuchungen an hexa-und oktoploiden arten. Zeitschrift furPlanzenzuchtung 59:83-102.

Steiner, K. E. 1988. Dioecism and its correlates in the Cape flora ofSouth Africa. American Journal of Botany 75: 1742-1754.

Thomson, J. D., and S. C. H. Barrett. 1981. Selection foroutcrossing, sexual selection, and the evolution of dioecy in plants.American Naturalist 118:443-449.

Thomson, J. D., and J. Brunet. 1990. Hypotheses for the evolution ofdioecy in seed plants. Trends in Ecology and Evolution 5:11-16.

Wagner, W. L. 1991. Evolution of waif floras: a comparison of theHawaiian and Marquesan archipelagos. Pages 267-284 in E. C. Dudley,editor. The unity of evolutionary biology. Volume 1. Proceedings of theFourth International Congress of Systematic and Evolutionary Biology.Dioscorides Press, Portland, Oregon, USA.

Wagner, W. L., and V. Funk, editors. 1995. Hawaiian bio-geography:evolution on a hot spot archipelago. Smithsonian Institution Press,Washington, D.C., USA.

Wagner, W. L., D. R. Herbst, and S. H. Sohmer. 1990. Manual of theflowering plants of Hawai'i. University of Hawaii Press and BishopMuseum Press (Special Publication 83), Honolulu, Hawaii, USA.

Wagner, W. L., S. G. Weller, and A. K. Sakai. 1995. Phylogeny andbiogeography in Schiedea and Alsinidendron (Caryophyllaceae). Pages221-258 in W. L. Wagner and V. A. Funk, editors. Hawaiian biogeography:evolution on a hot spot archipelago. Smithsonian Institution Press,Washington, D.C., USA.

Webb, C. J., and D. Kelly. 1993. The reproductive biology of the NewZealand flora. Trends in Ecology and Evolution 8:442-447.

Weller, S. G., W. L. Wagner, and A. K. Sakai. 1995. A phylogeneticanalysis of Schiedea and Alsinidendron (Caryophyllaceae: Alsinoideae):implications for the evolution of breeding systems. Systematic Botany20:315-337.

Yampolsky, C., and H. Yampolsky. 1922. Distribution of sex forms inthe phanerogamic flora. Bibliotheca Genetica 3: 1-62.

[TABULAR DATA FOR APPENDIX OMITTED]