CLASSIFICATION OF THE GradyL. Webster2 EUPHORBIACEAE'

ABSTRACT

The familyEuphorbiaceae appears to be monophyletic,despite proposals for segregate families. The Euphorbiaceae displaya greatvariety of growthforms, including at least 17 "models" of Hallk. Anatomicalcharacters particularly usefulfor classificationinclude wood structure,laticifer type, trichomes, and stomata.Inflorescences are basically dichasial,and pseudanthiahave evolved several times. Pollen nuclear numberand exine structureprovide useful criteriafor characterizing genera, tribes, and subfamilies.Structure of the seed coat is characteristicfor the family and does not provideevidence for a polyphyleticorigin of the family.Pollination is prevailinglyentomophilous, and seed dispersalby ants is importantin manytaxa. Geographicdistribution patterns of generashow a concentrationof primitivetaxa in Africaand Madagascar,although in subfamilyCrotonoideae there is evidenceof a neotropicalcenter. Disjunctionsbetween Africa and South Americaare common.Bentham's hypothesis of an Old World originof the familyappears well supported.The basic distributionpatterns appear to reflectearly (Cretaceousand Paleogene) dispersalacross land or narrowwater barriers and spectacularbut rathertrivial instances of long-distancedispersal in the late Tertiaryand Pleistocene;Tertiary high-latitude dispersals via the Beringland bridgeappear to have been relativelyinsignificant.

The Euphorbiaceae, although one of the largest uninformative about relationships, and the later dicot familiesand conspicuous throughoutthe trop- ones of Bentham (1880) and Pax (1890), who ics, have been relatively neglected by systematists accepted the general framework of Mueller with in the 20th century. While other families such as relativelyminor exceptions. Later revisions by Pax Compositae, Leguminosae, and Solanaceae have & Hoffmann (1931) and the treatment by Huru- been the subjects of various symposia, the very sawa (1954) continued to reflectthe original ideas firstinternational conference with a major focus of Mueller, despite considerable changes. Although on Euphorbiaceae was held at Kew in 1986 (Jury my recent classification (Webster, 1975) appears et al., 1987). The Kew symposium, on the Eu- very differentfrom that of Mueller at firstglance, phorbiales, was heavily biochemical in orientation I was struck by how many of the great Swiss and focused to a considerable extent on relation- systematist's ideas still survive. Even though he ships between the Euphorbiaceae and other fami- went astray in his major subdivisions (based on lies. The present symposium in St. Louis is the first cotyledon shape) of the "Stenolobeae" and "Pla- in which the classification of the family and its tylobeae," Mueller arranged the genera of Eu- constituentinfrafamilial taxa is the major focus of phorbiaceae into subtribes that, in an impressively attention. large number of instances, reflectphylogenetic af- The historyof classificationsystems for the Eu- finityas presently understood. In reviewing our phorbiaceae at the subfamilial and tribal level has effortsat this symposium to improve the classifi- been reviewed in the Kew symposium on Euphor- cation of Euphorbiaceae, it seems evident to me biales (Webster, 1987). In my opinion, the two that it is Mueller's foundation we are standing on. major milestones in this history were the classifi- In my classificationof 1975, the 300 genera of cations of Adrien Jussieu (1824), who identified Euphorbiaceae were grouped into 52 tribes in five the major series of genera that (after much later subfamilies, with several of the tribes divided into revision) correspond roughlyto currentsubfamilies, subtribes. The classification presented at this con- and Jean Mueller (1866), who provided the first ference shows only a limited number of changes, detailed classificationof the familyinto subfamilies, even though it is quite probable that futureresearch tribes, and subtribes. The originalityof Mueller is will show that substantial modifications will have particularly strikingin comparison with the earlier to be made. As has oftenbeen remarked, the linear systemof Baillon (1858), which was disappointingly arrangement of taxa in a classification is an im-

I I wishto thankRobert Rhode forhis assistancein preparationof the manuscriptand illustrations.Outline maps wereprovided through the courtesyof Daniel Axelrod. 2Section of Plant Biology,University of California,Davis, California95616, U.S.A. ANN. MISSOURI BOT. GARD. 81: 3-32. 1994.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 4 Annals of the MissouriBotanical Garden

perfect way of indicating phylogenetic relation- Ixonanthaceae are accepted as possible "out- ships. Cladistic analysis of the familyis badly need- groups," as speculated earlier (Webster, 1967), ed in order to provide a clear idea of the evolution then it is feasible to compile a list of characters in of characters and to test the implicit phylogenetic the Euphorbiaceae for which plesiomorphic (prim- hypotheses that are lodged in the classification itive) and apomorphic (advanced) states are defined scheme. Unfortunately,the informationavailable (Table 1). for the Euphorbiaceae is imperfectand this, along An important contributionto identifyingsignif- with the size of the group, makes it difficultto icant synapomorphiesin subfamilyPhyllanthoideae proceed. In my discussion of the familyat the Berlin has been made by Levin (1986c) in his cladistic congress in 1987 (unpublished), I provided some analysis of the subfamilybased on foliarcharacters. informallyconstructed "family trees" that reflect- Pollen studies on the subfamily, mostly still un- ed evolution of characters and used these as the published (Punt, 1987) or in this symposium (Levin basis for a schematic model of the evolutionary & Simpson, 1994a, b; Nowicke, 1994), show great and biogeographic history of the subfamilies and promise of adding important informationfor con- tribes. The present essay is intended to set forth structingphyletic models. The really major lacuna this model in greater detail, in the hope that it will in our knowledge, for this and other subfamilies, suggest topics fordiscussion and criticism.In doing is in biochemical data. Although Hegnauer (1962- so, I emphasize that, while our classificationsystem 1973) and Gibbs (1974) have compiled informative is a descendant of the system of Jean Mueller in resumes of reports fromthe biochemical literature, 1866, my scenario for the evolution and bioge- this informationis frustratingto the systematist, ography of the Euphorbiaceae is indebted to the because it demonstrates a great biochemical di- perspicacious essay of Bentham (1878), who from versity in the Euphorbiaceae but at the same time extremelyimperfect data deduced a model that has reveals a meager and erratic level of sampling. As proved to be almost clairvoyant in its anticipation a consequence, biochemical data are stillof limited of later work on the biogeography of major angio- usefulness for meaningfulcomparisons at the tribal sperm taxa. and subfamilial levels. In my 1987 review of the classification of the family, the Euphorbiaceae were implicitlytreated RELATIONSHIPS BETWEEN THE SUBFAMILIES as a monophyletic group by the inclusion of such Since the perceptive discussion of Bentham proposed segregate families as Bischofiaceae and (1878), it has been accepted by later workers (e.g., Hymenocardiaceae. However, for those interested Pax, 1924), that the subfamilyPhyllanthoideae is in relationships within and outside the Euphorbi- the primitive group from which other subfamilies aceae, it should be noted that this delimitation of are derived. Within the Phyllanthoideae, genera the familyis not accepted by all of my colleagues. withinthe tribe Wielandieae, such as Heywoodia Hymenocardiaceae are accepted as a family by and Savia, appear to represent relicts of the orig- Leonard & Mosango (1985) and by Radcliffe-Smith inal euphorbiaceous complex that arose probably (1987). Furthermore, Radcliffe-Smith accepts in the late Cretaceous. These primitiveEuphorbi- Pandaceae and retains three genera (Antidesma, aceae are dioecious shrubs or small trees of seasonal Bischofia, and Uapaca) only provisionally within forest habitats in tropical latitudes. They have, as the Euphorbiaceae. Jensen (1994) suggests that expected, relativelyunspecialized flowerswith well- on the basis of serological data there appear to be developed petals, floraldisk, and a pistillode in the two main groups of Euphorbiaceae: Phyllanthoi- staminate flower. The pistillate flower usually has deae + Oldfieldioideae; and Acalyphoideae + Cro- a 3-locular ovary, with two anatropous ovules be- tonoideae + Euphorbioideae. This binary arrange- neath an obturator. The fruitis a 3-celled capsule, ment of "biovulate" and "uniovulate" taxa parallels and the seeds have a dry testa, copious endosperm, the arrangement proposed by Mahlberg (1987) on and a large embryo with a radicle much shorter the basis of laticifer morphology. On a phenetic than the cotyledons. As far as the evidence goes, basis, this basic dichotomy appears intuitivelyrea- the chromosome base number is n = 13 in these sonable, but it remains to be seen whether it will "archaic" taxa of subfamily Phyllanthoideae. An- be substantiated by critical phylogenetic studies. atomically, the relict Phyllanthoideae have such Meeuse (1990) went much furtherand proposed unspecialized features as vessel elements with sca- recognizing nine families, with Euphorbiaceae be- lariformperforation plates, paracytic stomata, and ing restrictedmainly to the "uniovulate" subfam- unicellular nonglandular trichomes. Laticifers ap- ilies Acalyphoideae, Crotonoideae, and Euphor- pear to be absent. If Geranialean families such as bioideae. There is not sufficientspace here to analyze

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 5 1994 Classificationof the Euphorbiaceae

TABLE 1. List of primitive(plesiomorphic) and derived(apomorphic) characters in the Euphorbiaceae.

Character Primitivestate Derivedstate 1. Habit Trees/shrubs Herbs/vines 2. Branching Monopodial Sympodial 3. Phyllotaxy Alternate Opposite 4. Leaf shape Simple Lobed; compound 5. Leaf venation Pinnate Palmate 6. Stipules Present Absent 7. Vessel perforation Scalariform Simple 8. Vascular rays Multiseriate Uniseriate 9. Internalphloem Absent Present 10. Laticifers Absent Present 11. Trichomes Simple Stellate;lepidote 12. Foliar glands Absent Present 13. Inflorescence Axillary Terminal 14. Calyx aestivation Imbricate Valvate; reduplicate 15. Petals Present;free Absent;connate 16. Disk Present Absent 17. Stamennumber 5-10 1-4; over 10 18. Filaments Free Connate 19. Antherdehiscence Longitudinal(vertical) Horizontal 20. Pollen nuclei 2-nucleate 3-nucleate 21. Pollen exine Semitectate Tectate; intectate 22. Pollen apertures Colpi Pores; inaperturate 23. Aperturenumber 3 4 or more 24. Pistillode Present Absent 25. Carpel number 3-5 1-2; 6 or more 26. Stylebranches Bifid Unlobed;multifid 27. Styleunion Free Connate 28. Ovule number 2/locule 1/locule 29. Ovule configuration Anatropous Hemitropous 30. Embryosac Monosporic Disporic;tetrasporic 31. Fruit Dehiscent Indehiscent 32. Seeds Ecarunculate Carunculate 33. Seed testa Dry Fleshy 34. Endosperm Present Scantyor absent 35. Cotyledon/radiclelength 2+ Less than 2 36. Cotyledon/radiclewidth 2+ Less than 2

his provocative arguments, but I believe that he ual tribes or subfamilies, particular morphological has overemphasized the seed-coat data of Corner characters have proven to be especially useful. So (1976) and that the case for fragmentationof the far only pollen morphology has provided micro- family remains unpersuasive. Clearly, though, the morphological characters that are useful in deter- monophyly of the Euphorbiaceae as a whole, and mining systematic affinitiesin all five subfamilies, of the individual subfamilies, needs to be more as shown by the works of Erdtman (1952) and convincingly demonstrated in the course of any Punt (1962, 1987). However, the studies of Levin detailed analysis of the infrafamilialtaxa. (1986a, b, c) demonstrate that foliarvenation pro- vides systematicallyvaluable characters in subfam- ily Phyllanthoideae, and this may prove to be true REVIEW OF SYSTEMATIC CRITERIA for other characters as well. The survey of chro- At present, the overwhelmingpreponderance of mosome numbers by Hans (1973) indicates that the data available for interpretingthe patterns of these may well yield clues to defininglines of sys- phylogeny within the Euphorbiaceae comes from tematic affinity,but at present the available data gross vegetative and floral morphological charac- are frustratinglyincomplete. Biochemical data from ters that have been used as diagnostic taxonomic comparative studies of secondary compounds such features. In evaluating relationshipswithin individ- as alkaloids and terpenes can hardly fail to produce

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 6 Annals of the Missouri Botanical Garden

results of great interest,especially at the tribal and succulent Euphorbiaceae show additional variations generic levels, but at present we have only tan- that cannot be easily assigned to any of the Halle talizing scraps of data from scarcely more than a models. All of the subfamilies except for the Old- handful of genera. fieldioideaeshow more or less comparable diversity, In all of the classificationsof the Euphorbiaceae and there appear to be differencesbetween some proposed before 1975, the major criteria were of the subfamilies that may be phylogenetically drawn fromdetails of gross morphologyobservable significant. with the naked eye or a dissecting lens. Even after In subfamily Phyllanthoideae branching pat- demonstratingthe systematic importance of pollen terns tend to be monopodial and inflorescences characters by Erdtman (1952), the systems of Hu- axillary; the prevalent unspecialized patterns are rusawa (1954), Hutchinson (1969), and Airy Shaw those shown by the models of Attims, Aubreville, (1972, 1975) stilldepended almost entirelyon the and Rauh. A strikingtendency is that toward mono- same data sets available to Baillon, Mueller, Ben- podial plagiotropic branching (Roux's model) as in tham, and Pax. In order to expedite future work Flueggea, Phyllanthus(input), and Drypetes. A on classificationof the family,it seems appropriate furtherspecialization of this pattern yields the de- in this essay to review the characters that show ciduous, floriferous phyllomorphic branchlets promise in interpreting evolutionary trends and (Cook's model) found in several genera of tribe lines of affinity.Ecological characters are not dis- Phyllantheae subtribeFlueggeinae, includingBrey- cussed, except for those relating to pollination and nia, Glochidion,Phyllanthus, and Sauropus. This dispersal with a possibly significanteffect on the highly specialized growth form, firstdocumented evolution of reproductive structures. by Dingler (1885) and designated as "phyllanthoid branching" by Webster (1956), has been thor- oughlyanalyzed by Roux ( A. GROWTH FORM 1968), Bancilhon ( 1971), and Rossignol & Rossignol (1985). The Euphorbiaceae display an extraordinary The subfamily Oldfieldioideae resembles the range of growth forms, perhaps equaling or sur- Phyllanthoideae in having relatively unspecialized passing any other angiosperm family(Halle, 1971). monopodial branching patterns that agree with the Trees and shrubs predominate, as one would expect models of Attims and Rauh. However, Stachys- of a primarilytropical family, but herbs have evolved temon polyandrus (F. Muell.) Benth., and perhaps independently in all five subfamilies. The greatest a few other taxa, appear to fitthe model of Koriba, diversity of growth form is seen in the genera in which extension of the main axis is interrupted Phyllanthus (Webster, 1956) and Euphorbia by a terminal inflorescence, leading to sympodial (Croizat, 1972; Cremers, 1977), which show a wide growth. In the three uniovulate subfamilies,deter- range of adaptations to relatively mesic and xeric minate/syrnpodial growth patterns predominate. habitats, respectively. Among the Acalyphoideae studied, about half (20 In recent years the architecture of tropical trees genera) show the unspecialized patterns of Attims has been studied on a comparative basis, so that and Rauh, but a number of genera followthe mod- for the firsttime the bewilderingdiversity of form els of Koriba and Leeuwenberg, in which the main in tropical taxa can be comprehended through the axis is terminated by an inflorescence and branch- scheme of about two dozen basic models proposed ing is more or less dichasial in nature. In the by Halle and coworkers (Halle & Oldeman, 1970; Crotonoideae, the model of Koriba seems to be the Halle et al., 1978). Determination of the mem- most common, and that of Rauh is about as prev- bership of a particular taxon in the Halle/Oldeman alent as that of Attims. Finally, in the Euphor- model system is difficultand sometimes impossible bioideae, which are possibly the latest subfamilyto from examination of herbarium specimens; obser- evolve, the branching patterns are more special- vations of the germinationand ramificationof seed- ized. Only six out of 32 genera show the patterns lings or saplings may be required. Despite the con- of Attims and Rauh, as compared to 15 for the siderable number of Euphorbiaceae that have been models of Koriba and Leeuwenberg. Furthermore, studied, the sampling of taxa in the subfamiliesand the models of Prevost and Nozeran, almost or en- tribes is still incomplete. However, furtherstudies tirelyabsent in the other subfamilies,are well rep- may provide importantinsights into the phylogeny resented with nine genera. Even though these fig- of suprageneric taxa. ures are influenced by the large number of models At the present time, no less than 17 models have found in Euphorbia, the divergence of the sub- been reported for the woody Euphorbiaceae; and family Euphorbioideae in growth pattern is clear. a significantnumber of herbaceous, scandent, and The vegetative and reproductive morphology of

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 7 1994 Classificationof the Euphorbiaceae

Euphorbia has received a great deal of attention, model. Cremers (1977) demonstrated 10 different and a strikingsequence of modificationsof growth models in Madagascan Euphorbia, and there ap- form can be traced within the tribe Euphorbieae. pear to be some species that do not fitany model. It is particularlynotable that the branching pattern Even worse, species such as Euphorbia stenoclada of Anthostema is referredto the model of Nozeran Baillon and E. abyssinica Gmel., which conform by Halle & Oldeman (1970); the ramificationin to the models of Attims and Rauh, have surely that genus recapitulates the origin of the cyathium been derived from ancestral taxa with the more froma concrescent verticel of floriferousbranches, "specialized" models of Prevost and Nozeran. with the abortive axis surmounted by a pistillate It is understandable that correlations between flower. Nozeran (1953) had appropriately applied architectural models and ecological conditions are the term "pre-cyathium" to the bisexual spikes of imperfect and often obscure, in view of the fact Hippomaneae such as Sapium or Stillingia. A that the vegetative growth patterns interact with whorl of four or five of these "spikes" (which are reproductive adaptations in a complicated fashion. really thyrses) produced on a plant growing ac- Certain models, such as Cook's and Nozeran's, cording to the model of Nozeran could indeed have clearly predominate in rainforesthabitats; but oth- given rise to the primitivecyathium, although there ers, such as Attims's, are found in both rainforests are still problems in the absence of any living and deserts. Evidently the Euphorbiaceae have en- ''missing link" in specifyinghow the lateral and tered such a wide variety of habitats during their basal flower in the Hippomanean thyrse becomes radiation that no simple correlations can be ex- the terminal flower in the cyathium (see Gilbert, pected. Nevertheless, the coevolution of branching 1994). patterns and reproductive structures in some taxa One can argue that the ultimate point in veg- of Euphorbiaceae is sufficientlystriking that more etative evolution in the Euphorbiaceae is attained attention needs to be given to field observation of in Chamaesyce, where the seedling axis is usually the architectural models. abortive above the cotyledons and the axial portion of the plant is virtuallyall sympodial. In the abor- B. ANATOMY tion of the main axis, the patternsuggests the model of Nozeran, but the subsequent dichasial pattern Anatomical characters have been regarded as is a two-dimensional plagiotropic variant of the importantin the classificationof the Euphorbiaceae model of Leeuwenberg. Hurusawa (1954) has since the essay of Pax (1884), which firstestab- claimed that the conformationof axes in Chamae- lished the importance of laticifers as a character syce represents the terminus of a reduction series for definingnatural groups. Pax used laticifertype beginning in Euphorbia subgenus Esula and pro- (articulated vs. nonarticulated), phloem characters, gressingthrough subgenus Agaloma by retardation and trichome type as the major anatomical char- of the main axis relative to the lateral anthocladia. acters to support his redivision of the major su- However, in a recent study,Hayden (1988) showed prageneric taxa. These studies were considerably that the lateral branches in Chamaesyce arise at extended in the followingdecade by Radlkofer and the cotyledonary nodes, and that the apparent ho- his students. Radlkofer (1870) emphasized ana- mology with the terminal pleiochasial whorl in Eu- tomical characters in a paper establishing the new phorbia is doubtful. euphorbiaceous genus Pausandra; and a series of The adaptive significance of the growth forms his students surveyed most of the family anatom- in the Euphorbiaceae is not as clear as their utility ically (Rittershausen, 1892; Froembling, 1896; as systematic characters (a common situation in Rothdauscher, 1896; Herbert, 1897). The first many angiosperm taxa). Part of the difficultyis really critical anatomical survey of the familywas that the changes in some ramificationpatterns are provided by Solereder (1899), who reviewed the due to simple shiftsin relative rates of meristematic contributionsof Radlkofer's group and added many activity, so that reversals of trends and conver- observations of his own. Gaucher (1902) published gences appear common indeed. A strikinginstance an independent anatomical survey of the family of this, for example, is shown by the distribution that curiously omits all mention of the works of of Troll's model in the Euphorbiaceae: in the Phyl- the German school. He was criticized for this by lanthoideae, it is found in woody taxa of Briedelia Solereder (1908), who also pointed out mistakes and herbaceous Phyllanthus, where it appears de- in his observations. However, Gaucher's study did rived from Attims's model; but in the Euphorbioi- have the merit of providing a comparative review deae, it seems to have evolved in Pedilanthus via of anatomical characters arranged by tribe. Attims's model and in Actinostemon via Prevost's Anatomical studies of wood by Pax and others,

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 8 Annals of the Missouri Botanical Garden

up to the turn of the century, seem mainly to have ticifersin the single genus Jatropha: single-celled been made on twig material. The firstimportant idioblasts, coenocytic laticiferswithout cross-walls, contributionbased on studies of mature wood from and classic "articulated" laticifers.Their results in tree trunksappears to be that of Janssonius (1929, some respects contradict the findings of Rao & 1930), based largely on Indonesian taxa. Jansson- Malaviya (1964). Mahlberg (1975, Mahlberg et ius pointed out the divergent wood structure of al., 1987), in a series of studies showed the sys- Daphniphyllum as an indication of its lack of tematic importance of laticifersin Euphorbia, par- relationship to the Euphorbiaceae. The most in- ticularly with respect to the type of starch grains teresting of his findingswas his designation of a produced. Only recently, however, has Rudall group of genera of subfamily Phyllanthoideae (1987) provided the firstmodern survey of laticifer (Aporosa, Baccaurea, Drypetes,and Putranjiva) types within the family and amply demonstrated as characterized by thick-wallednonseptate fibers, the importance of laticifers in evaluating relation- abundant metatracheal parenchyma, and scalari- ships. Rudall's survey indicates that articulated formvessel perforations.Bamber (1974) amplified laticifers are confined to subfamily Crotonoideae the work on fiber morphology by showing that and nonarticulated ones to the Crotonoideae and thick-walledfibers ("type II") also occur in genera Euphorbioideae, except for a few reports (needing of subfamily Oldfieldioideae (Dissiliaria and Pe- confirmation) of nonarticulated laticifers in the talostigma). Mennega (1987) expanded the survey Acalyphoideae. Mahlberg et al. (1987) interpreted and showed the presence of type II fibers in a this laticifer data radically by proposing a subdi- number of additional genera of Phyllanthoideae and vision of the Euphorbiaceae into three familiesbased Oldfieldioideae. Hayden (1980) reviewed the wood on the groups Phyllanthoideae, Acalyphineae, and anatomy of subfamily Oldfieldioideae as part of a Hippomanoineae of Pax (1884). Clearly there is comprehensive anatomical survey. still an urgent need for more intensive sampling of Metcalfe & Chalk (1950), in the most recent tribes in the uniovulate Euphorbiaceae for mor- comprehensive anatomical review of the entire phological and anatomical data. family,created some anatomical groups that partly Another important source of systematic data is correspond to accepted subfamilial and tribal con- furnished by trichomes. Mueller (1866) was the cepts but in a number of respects are contradictory. firstto emphasize trichomes as a major character Their "Group A" of the Phyllanthoideae includes in his revision of the family. Froembling (1896) the genera with the syndrome of characters first reviewed the trichome types in the Crotonoideae, identifiedby Janssonius:thick-walled (Bamber "type which show the greatest diversity among the sub- II") fibers,abundant parenchyma, and scalariform families. The subsequent neglect of trichome char- vessel perforations.It appears that "Group A" taxa acters in the Euphorbiaceae is indicated by the fact such as tribes Wielandieae and Antidesmeae may that the last detailed systematic review is that of indeed have the most anatomically primitivechar- Solereder (1899). Inamdar & Gangadhara (1977) acters in the Euphorbiaceae. However, there are presented a survey of euphorbiaceous trichomes, also some outstanding discrepancies between the but their sample of only 53 species was too limited arrangement of Metcalfe & Chalk and the classi- to be systematically meaningful. A more compre- fication accepted currently. The most notable is hensive sample of 250 species by Rao & Raju the position of Acalypha, which is grouped by them (1985) provided a succinct review of the systematic (and Janssonius) with the Phyllanthoideae; but this distributionof trichometypes. Most of the trichome disposition is contradicted by so much other evi- yariation of systematic interestis found in subfam- dence that the resemblances of Acalypha to the ilies Acalyphoideae and Crotonoideae, where var- Phyllanthoideae must be due entirely to conver- ious kinds of malpighiaceous, stellate, and lepidote gence. The studies of Mennega (1987) on wood types occur; in the other three subfamilies simple anatomy in the Phyllanthoideae in general sup- unicellular or multicellular hairs overwhelmingly ported the relationshipsin the classificationof Web- predominate. ster (1975), although some discrepancies remain. The most unusual trichomes found in the Eu- Although Pax (1884) placed great emphasis on phorbiaceae are stinginghairs, which have evolved laticifer type in defining the major infrafamilial independentlyin the Acalyphoideae and Crotonoi- subdivisions, the use of laticifers to address prob- deae. The hairs in Cnidoscolus (Crotonoideae-Ma- lems at higher ranks was subsequently neglected nihoteae) are of the common Urtica type, with a despite the plea by Solereder (1908) for new de- deciduous bulbous tip (Thurston & Lersten, 1969). velopmental and comparative studies of laticifers. Breckon (1975) has described considerable vari- Dehgan & Craig (1978) reported a variety of la- ation in morphology of the stinging hairs in Mex-

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 9 1994 Classificationof the Euphorbiaceae

ican species of Cnidoscolus. Within the Acaly- centage of anomocyticstomata in Chamaesyce phoideae, stinginghairs of a fundamentallydifferent (Raju & Rao, 1987). Not unexpectedly,Kakkar type are known fromthe tribe Plukenetieae. These & Paliwal (1974) founda considerablediversity were carefully described by Knoll (1905) as com- of stomataltypes in Euphorbia. Dehgan (1980) posed of a central crystaliferouscell of subepider- foundthat the two subgeneraof Jatrophacan be mal origin surrounded by a jacket of three epi- distinguishedby stomataltype and also surfaceof dermal cells; Knoll's interpretation of these hairs(smooth versus verruculose). trichomes has been confirmed by the ultrastruc- Foliarvenation, a neglectedcharacter, has only tural studies of Thurston (1976). The common recentlybeen used by Levin (1986a, b, c) to pro- possession of this unique type of trichome (along videnew insights into relationships within subfamily with other characters such as columnar compound Phyllanthoideae.Similar studies on the othersub- styles) provides a powerful argument for assigning familieswould no doubt be illuminating.So far, Dalechampia, which has been previously kept in almostall ofthe work in othersubfamilies has been a separate tribe, to the Plukenetieae. done on tribe Euphorbieae (Sehgal & Paliwal, The Euphorbiaceae are also impressively rich 1974). The characteristicjacketed venation in in foliar glands, which appear to have arisen de Chamaesyce has attractedinterest because of its novo within the family, as they are absent in the correlationwith C4 photosynthesis(Webster et al., Wielandieae and other apparently archaic taxa in 1975). the Phyllanthoideae. In both the Phyllanthoideae Althoughthe vast majorityof Phyllanthoideae and Oldfieldioideae, foliar glands are restricted in and Euphorbioideaehave simpleentire leaves, the occurrence, but they are very common in the uni- Acalyphoideaeand Crotonoideaehave predomi- ovulate subfamilies.Diffuse pellucid-punctate glands natelytoothed or lobed leaves. Hickey & Wolfe have been reported in the unrelated genera Clutia (1975) have designatedthe Euphorbiaceaeas hav- (Acalyphoideae) and Suregada (Crotonoideae). ing a "Dilleniid"leaf typeon the basis of actinod- Discrete embedded laminar glands are widespread, romousvenation and "Violoid" marginalteeth, in occurring in many Acalyphoideae (e.g., in Acaly- whichthe medial vein expands into a glandular pheae, Alchorneae, Bernardieae, Chrozophoreae, terminationwithout a deciduousapical seta. How- Omphaleae, and Plukenetieae). In the Crotonoi- ever, in the most primitivePhyllanthoideae the deae, glands most often occur at the junction of leaves are entireand camptodromous,so the "Vio- lamina and petiole (e.g., in Crotoneae, Manihoteae, loid" teethwould appear to have arisen de novo Micrandreae). In Jatropha (Dehgan & Webster, withinthe family.Furthermore, the foliarteeth in 1979), and many other Crotonoideae and Acaly- Euphorbiaceaeare variable;for example, decidu- phoideae, glands are very diverse; glands on leaf ous setae occur in the Hippomaneae,which would margins often appear to intergrade with petiolar make them"Theoid" insteadof "Violoid." A sur- glands. vey of foliarvenation and toothtypes in the Aca- Bernhard (1966), in a survey of the morphology lyphoideaeand Crotonoideaewill be necessaryto and anatomy of euphorbiaceous foliar glands, has verifythe claim by Hickey& Wolfethat the leaves suggested that the marginal glandular teeth in many of Euphorbiaceaebetray a "Dilleniid" insteadof Euphorbiaceae may be of a separate origin from "Rosid" affinity. the laminar or petiolar glands. Belin-Depoux & Vascular anatomyof petioleshas been studied Clair-Maczulajtys (1974, 1975) elucidated the an- fromcross sections by Dehay (1935), who showed atomical structure of petiolar glands of Aleurites considerable variation in stelar configurations. withlight microscopy and TEM, and Belin-Depoux However,Dehay did not attemptto untanglevari- (1977) showed that the foliar glands of Alchornea ation correlatedwith leaf size fromthat due to cordata Benth. (Acalyphoideae) are fundamentally systematicaffinity, and it seems unlikelythat this similar to those of Aleurites (Crotonoideae). characterwill prove to be of generalutility. How- Stomatal development withinthe Euphorbiaceae ever,on a morerestricted level, Miller & Webster shows considerable variation, as indicated by Raju (1962) used differencesin petiolarsteles to sepa- & Rao (1977). They found an overwhelmingpre- rate Cnidoscolus from Jatropha, and Dehgan dominance of paracytic stomata in the woody taxa (1982) foundthat different degrees of petiolarste- of most of the 17 tribes sampled, with anisocytic lar dissectionwere significantat the sectionaland stomata becoming prominentmainly in herbaceous subsectionallevel in Jatropha. Phyllanthoideae (as indicated earlier for Phyllan- A significantfoliar structure in many Euphor- thus by Webster, 1956). The most aberrant sit- biaceae are the stipulesinserted at the base of the uation reported by Raju & Rao is the high per- petiole(Uhlarz, 1978). In many of the Phyllan-

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 10 Annals of the MissouriBotanical Garden

thoideae and occasional genera (e.g., Manihot) the flowersin axillaryclusters are producedcy- elsewhere,the stipulesare early deciduous (ca- mosely.The thyrse,often mistakenly described as ducous),even if they are ratherlarge. Phyllanthoid a spike,is verycommon in all fivesubfamilies. In stipulesare entire,but in the Acalyphoideaeand subfamilyEuphorbioideae there is a strikingre- Crotonoideaethey can becomelobed, dissected, or ductionseries from the diffuse"paniculate" inflo- glandular.In a considerablenumber of taxa, stip- rescences of Senefelderathrough the thyrsesof ules have become reduced or even entirelysup- Hippomaneaesuch as Sapium or Stillingia to the pressed.Sometimes one or a fewspecies in a genus pseudanthia(cyathia) of the Euphorbieae.Exam- (e.g., Hyeronimaalchorneoides F. Allemao)have ples of pseudanthiaor "subpseudanthia"can be stipules,whereas they are absentin relatedspecies. foundin isolatedgenera in mostof the subfamilies: Manyspecies of Crotonlack stipules,whereas they Phyllanthoideae(Uapaca), Oldfieldioideae(Pseud- are large and conspicuousin others.Most notably, anthus),Acalyphoideae (Dalechampia, Pera), and in Euphorbia presenceor absence of stipulesis a Euphorbioideae(tribe Euphorbieae). In the tribe diagnosticcharacter for some sections and sub- Euphorbieaethere has been the greatest"radia- genera. Althoughstipules presumably play some tion" of pseudanthialstructure, with several cy- adaptive role in protectingstem apices and leaf athialgenera apparentlyevolved from a common primordia,no convincingexplanation has been giv- ancestor. en fortheir presence or absence in closelyrelated Floral formin the Euphorbiaceae shows tre- taxa. mendousdiversity, partly as a consequenceof in- dependentevolution of staminate and pistillateflow- ers and partlyas the result of adaptationto a C. INFLORESCENCES AND FLOWERS varietyof vectors (wind, bees, flies,butterflies, Since Euphorbiaceaehave unisexualflowers that birds,and mammals).A masterlyillustrated survey are oftenreduced, grouping of flowersinto inflo- of floralgross morphology, with emphasis on on- rescences is oftenimportant for pollination,and togeny,was made by Baillon (1858), the most inflorescencestructure is an importantsystematic perceptivemorphologist of the 19th-centuryspe- characterin manyinstances. Monoecious and di- cialistson the family.Since Baillon's workthere oecious flowerproduction are widespread,with has been only one other comprehensivesurvey, monoeciousperhaps somewhatmore common.It that of Michaelis(1924). However,Venkata Rao seems likelythat dioecy is primitivein the family, morerecently (1971, 1972) studiedvascular anat- judgingfrom sexual dispositionsin "basal" taxa omy of a considerablenumber of genera and at- (e.g., Wielandieae) in the subfamilyPhyllanthoi- temptedto relatehis findingsto the phylogenyof deae. However,this seems to have been readily the family.Unfortunately, he sometimesgeneral- reversible,since monoecioustaxa appear to have ized excessivelyfrom single instances, as whenhe givenrise to dioeciousones in Croton,Euphorbia, suggestedthat the flowers of Jatropha are themost Phyllanthus,and othergenera. Inflorescencesas primitivein the familybecause theyare oftenher- a wholeare proterogynous,although within some maphroditic;this is contradictedby the specialized genera(e.g., Euphorbia) transitionscan be found pollengrains and secretoryapparatus in thatgenus. fromproterogyny to proterandry.Pollination of VenkataRao pointedout a numberof trendsin Euphorbiaceaeis mainlyentomophilous, but wind floralmorphology within the Euphorbiaceae,the pollinationhas evolvedindependently in all of the most obviousbeing that of reductionin size and subfamilies(except in the Euphorbioideae,where numberof parts. Reduction is sometimescorrelated anemophilymay be absent):Hymenocardia (Phyl- withshift to wind pollination,as in tribeAcaly- lanthoideae);Picrodendron (Oldfieldioideae); Aca- pheae; but in otherinstances (Phyllantheae, Eu- lypha and a considerablenumber of othergenera phorbieae)it clearlyis not. As noted by Venkata (Acalyphoideae);and Borneodendronand Ere- Rao, thereare also trendsto increasethe number mocarpus (Crotonoideae).Wind-pollinated taxa of floralparts; here again, thereis no correlation usuallyhave spicateinflorescences, but so do many withanemophily, since thehighest numbers of sta- entomophiloustaxa. mens, for example,occur in Croton(entomophi- It seems apparentthat inflorescencestructure lous) and Ricinus (anemophilous).A particularly in the Euphorbiaceaeis fundamentallydichasial. evidentreduction tendency is loss of the corolla, This is most strikinglyevident in genera of Cro- whichis so prevalentthat Euphorbiaceae are often tonoideaesuch as Cnidoscolusand Jatropha,which placed in "apetalous" groups. This affectsboth have elaborate compounddichasia. However,in sexes but is moststriking for the pistillateflowers; the Phyllanthoideae,it can usually be seen that petaliferousstaminate flowers and apetalouspistil-

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 11 i1994 Classificationof the Euphorbiaceae

late flowersoccur in most species of Crotonand (1968) have shown that the vascularization of the in other Crotonoideae,as well as many Acaly- disk is highly variable; and Webster (1956) found phoideae (e.g., Agrostistachydeae,Chrozopho- that in Phyllanthus the degree of vascularization reae). Two subfamilies,Oldfieldioideae (except seems to be correlated with mass; small disk-seg- Croizatia) and Euphorbioideae,are apetalous in ments are usually nonvascularized. The prepon- bothsexes. In the presumablywind-pollinated old- derance of evidence seems to be against the view fieldioidgenera Androstachysand Stachyandra, of Michaelis and in favor of the classical concept the staminateflowers have a chaotic organization of Eichler (1875) that the floral disk is of recep- of tepals and stamensthat led Leroy (1976) to tacular nature and not an independent whorl of proposea curioustheory suggesting that the sta- the flower. minateflowers in these genera are modifiedinflo. The euphorbiaceous androecium displays a pro- rescences. tean diversity that has provided many characters The calyx in Euphorbiaceaevaries greatlyin used by Mueller (1866) and later workers to di- configurationand numberof parts.Where the co- agnose subtribes, genera, and infrageneric taxa. rolla is absent,petaloid calyces may occur, as in Michaelis (1924) hypothesized that the primitive Cnidoscolus and Manihot. Croizat (1943) sug- condition in the family was an androecium of an gestedthat the perianth of Cnidoscolusis corolline indefinitenumber of stamens in many whorls. Ven- and that the calyx had been lost; however,the kato Rao & Ramalakshmi (1968), in contrast, sug- evidencefrom vascular traces in Manihot contra- gested that the primitive condition is 10 stamens dictsthis idea (VenkataRao & Ramalakshmi,1968; in two whorls, as in Jatropha. However, on the Venkata Rao, 1971). In past systemsof classifi- basis of the prevalent condition in the Wielandieae cationof the Euphorbiaceae,great stress has been and associated tribesof the Phyllanthoideae, I think placed on aestivationof the calyx (whetherimbri- that an androecium of a single whorl of five or six cate or valvate), a practice initiatedby Mueller stamens is most likely to be ancestral within the (1866). AlthoughBaillon (1873) rightlyobjected family.Mueller (1866), in definingsubtribes, placed to the excessiveweight accorded the characterby great emphasis on the insertion of the stamens; in Mueller,it was stillextensively used by Pax (1890) practice, this usually depended on whether they and Pax & Hoffmann(1931). Comparisonof the were inserted around a pistillode or disk. However, aestivationcharacter with evidence from other in some groups such as the Drypeteae it is difficult characters(wood anatomy,trichomes, and pollen) to distinguisha pistillode from a confluent central stronglysuggests that it mustbe used withcaution. disk; and in Crotonoideae such as Cnidoscolus or The euphorbiaceouscorolla, where present,is Jatropha, it is not easy to tell whether the slender variablein size and color;large showy red or white projections at the top of the staminal column rep- flowersoccur mainlyin the Crotonoideae(e.g., in resent staminodes or a pistillode. Aleurites,Jatropha). The petals are usually dis- Mueller was sharply criticizedby Baillon (1873) tinct,but may be coherentor connate in some for placing undue emphasis on anther form and Crotonoideaesuch as Jatropha(Dehgan & Web- dehiscence as generic characters. Because of tran- ster,1979) and someAleuritideae (Leonard, 1962). sitionalsituations, distinctions between extrorse and VenkataRao & Ramalakshmi(1968) reportedthat introrse dehiscence are sometimes difficultto ver- vascularizationof petals (with three bundles) is ify. There are all degrees of shift of the plane of usually similarto that of the sepals, althoughit anther dehiscence, not only between extrorse and maybe obsoletein thereduced petals of Codiaeum introrse but between basifixed/versatile and lon- and Croton. gitudinal/horizontal.The most bizarre specializa- The floraldisk (nectary) is a conspicuousfeature tions are found in the Acalypheae (subtribe Lasio- of many euphorbiaceousflowers. Baillon (1858) coccinae), where the filamentsof the stamens are regardedit as receptacularin natureand was fol- united into branching fascicles and the number of lowedin thisinterpretation by mostlater workers anthers proliferated,reaching the extreme degree (e.g., Bentham,1880; Pax & Hoffmann,1931). of elaboration in Ricinus. Although there has been On theother hand, Michaelis (1924) proposedthat much profitlessspilling of ink in speculations about the disk is of staminodialorigin, partly because of the androecium of Ricinus, it tells us nothingabout theapparent obdiplostemony in flowersof taxa such the role of telomes in the angiosperm flower, as as Jatropha, and partlybecause of the stami- pointed out by van der Pijl (1952). Venkata Rao nodelikeappearance of disk-segmentsin various & Ramalakshmi (1968) made the fantastic sug- genera such as Chiropetalum,Clutia, and Por- gestion that the branched stamen complex of Rici- anthera. However,Venkata Rao & Ramalakshmi nus is primitive and that conventional dithecous

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 12 Annals of the MissouriBotanical Garden

anthers are derived by reduction; abnormal super- published). The importance of palynological char- numerarily lobed anthers in Jatropha curcas L. acters is furtherhighlighted by the studies in this are supposed to represent an intermediate condi- issue on Oldfieldioideae, Crotonoideae, and tribe tion. In fact, the derivation of the androecium of Plukenetieae. Ricinus is clearly suggested by comparison with The great pollen diversity in Euphorbiaceae the related genus Adriana, which has a large num- makes it difficultto identifyconsistent trends. As ber of completelyseparate stamens. Althoughmore one might expect, taxa that are presumed to be careful morphological study is needed, it appears wind-pollinatedshare a tendency toward reduction that Airy Shaw (1980) may have been correct in of colpi and sexinous ornamentation, e.g., in Hy- suggestingthat the fascicled stamens in Homonoia menocardia (Phyllanthoideae), Longetia (Oldfield- and Lasiococcus were derived independentlyfrom ioideae), Acalypha (Acalyphoideae), and Borneo- those in Ricinus. dendron (Crotonoideae). However, Ricinus is a Pollen grains of the Euphorbiaceae are produced partial exception; although the sexinous ornamen- in tetrads by simultaneous cytokinesis, as in most tation is reduced, the grains have very prominent dicots (Davis, 1966), and are always monads at colpi. Within all of the subfamilies except the Eu- maturity. At anthesis the pollen grains may be phorbioideae, there is a tendency for multiplication either binucleate or trinucleate. Webster & Rupert and shorteningof apertures that leads to polyporate (1973) showed that the nuclear condition of pollen grains; in the Crotonoideae, the apertures become is a significant taxonomic character in defining obsolete. Crotonoid sexinous ornamentation is a major phyletic lines in the family: the Phyllan- unique synapomorphy within subfamily Crotono- thoideae, Oldfieldioideae, and Acalyphoideae ap- ideae. Transmission electron micrography will no pear to be entirely binucleate, while in the Cro- doubt provide even more valuable insightinto tribal tonoideae and Euphorbioideae both binucleate and relationships within the subfamilies. trinucleate pollen occur. In a study of the tribe The gynoecium in Euphorbiaceae is almost as Euphorbieae, Webster, Rupert & Koutnik (1982) diverselymodified as the androecium. Althoughthe showed that trinucleate pollen originated several carpel number is three in the majority of taxa, it times and that the distributionof pollen nuclear is reduced to two in a number of genera and to number is an important criterion in defining sub- one in several, including Antidesma (in part), Cro- genera and sections. These observations support tonopsis,Drypetes (in part), and Jatropha. On the "Schiirhoff-BrewbakerLaw," which postulates the other hand, it is increased to six or more in that the shiftfrom binucleate to trinucleate pollen species of a number of genera, including Glochi- is irreversible. dion, Hippomane, Hura, and Phyllanthus.It is The morphology of pollen grains in Euphorbi- not even certain that 3-carpellate gynoecia rep- aceae is so diverse that only the Acanthaceae can resent the plesiomorphic condition in Euphorbi- rival it. Although Mueller (1874) illustrateda con- aceae, since withinthe Wielandieae there are gen- siderable number of pollen grains of Brazilian Eu- era characterized by four carpels (Heywoodia) and phorbiaceae, he made no use of pollen morphology five(Wielandia). In Margaritaria carpelnumber in definingtaxa at any level. Pax (1884) dismissed fluctuates between two and six, and varies from the systematic utility of pollen despite the paper three to six in the single species M. nobilis L.f. by Radlkofer (1870) emphasizing the pollen char- (Webster, 1979). acters of Pausandra. We owe to Erdtman (1952) Stylar variation is marked and has been used the firstconvincing demonstrationthat pollen mor- often for specific, generic, and even tribal diag- phology offers decisive clues to systematic rela- noses. In the majority of genera the styles are tionshipswithin the Euphorbiaceae. Erdtman's sug- elongated and bifid (often so deeply as to be bi- gestion that genera with crotonoid sexine patterns partite); in scattered genera and in the subfamily should be associated irrespective of their positions Euphorbioideae, they are undivided in most genera in the system of Pax & Hoffmann(1931) has been (but usually bifidin Euphorbia). On the other hand, supported by Punt (1962) and later workers. K6h- the two style branches may be furthersubdivided ler (1965), on the basis of a survey of pollen of so that each style may have four or eight stigmas, the biovulate taxa, was able to make a persuasive or even more; this is particularly common in Cro- argument for recognizing a new subfamily, the ton. Usually the styles are connate, if at all, only Oldfieldioideae. The reclassification of subfamilies at the base, but in the tribe Plukenetieae they are and tribes by Webster (1975) was strongly influ- often connate to the tip into a long stylar column enced by these light microscopic studies and by which may be modified in bizarre shapes. Finally, unpublished SEM observations (Webster et al., un- in a number of genera (e.g., Endospermum, Glo-

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 13 1994 Classificationof the Euphorbiaceae

chidion,Tetrorchidium) the stylesare reducedto phorbia, and tetrasporictypes in Acalypha, Eu- sessile, oftenpetaloid, stigmas. Rather unexpect- phorbia, and Mallotus. A characteristicof most edly, stylarmorphology is not clearly correlated euphorbiaceousfemale gametogenesis is the early withpollination vectors: anemophiloustaxa may disappearanceof theantipodals, which in the older have eitherhighly dissected styles (Acalypha) or literatureled to a numberof misinterpretationsof undividedentire styles (Alchornea). embryosac developmentin genera such as Phyl- Ovular charactershave been regardedas sig- lanthusand Codiaeum (Maheshwari,1942). Apo- nificantwithin the Euphorbiaceaesince A. Jussieu mixisis rare in Euphorbiaceae,even thoughthe (1824) originallyestablished the distinctionbe- earliestreport of apogamyis fromthe genus Coe- tweenbiovulate and uniovulatetaxa as a diagnostic lebogyne(Smith, 1841); ithas beenverified mainly criterion.There is no evidenceof reversionfrom in temperatespecies of Euphorbia such as E. dul- the presumedderived condition, uniovulate, to the cis L. (Carano, 1926; Cesca, 1961; Kapil, 1961). biovulatecondition. However, it remainsunclear whetherthe transition from gynoecia with biovulate D. FRUIT AND SEEDS locules to gynoecia with uniovulatelocules oc- curredonly once. In otherwords, we cannot yet The fruitin most Euphorbiaceaeis an explo- be certainthat the uniovulate taxa, treatedby Pax sivelydehiscent schizocarp, but indehiscentfruits (1890) as subfamilyCrotonoideae, represent a have evolvedrepeatedly in all of the subfamilies. monophyleticgroup. The schizocarpousfruit appears to be theprimitive As pointed out by Baillon (1858), ovules of state, both in the familyas a whole and within Euphorbiaceaeare characteristicin theirepitro- individualgenera. In Phyllanthus(Webster, 1956- pous orientationand the presenceof an obturator 1958) thereis a particularlycomplete morpholog- thatfits over the micropyleof one or both ovules ical series:typical capsules (mostsections) to bac- and fillsthe gap betweenthe stylarconducting cate fruits(sect. Anisonema),drupaceous but de- tissue and the nucellus. Withinthe Phyllanthoi- hiscent(sect. Emblica), and finallydrupaceous and deae, some taxa (e.g., Aporusa, Drypetes)have a indehiscent(sect. Cicca). It mustbe admitted,how- single massive obturator,while others(e.g., An- ever, that this linear sequence does not clearly drachne,Phyllanthus) have tenuousseparate ob- representmorphological transition within a single turators.Schweiger (1905) surveyedthe variation phyleticline. in ovularform in the Euphorbiaceaeand foundit Berg(1 975a, b) has providedthe most thorough to be ofconsiderable importance systematically. In descriptionof the development, anatomy, and ecol- some Phyllanthoideae,especially tribe Phyllan- ogy of the typicaleuphorbiaceous fruit as repre- theae, the ovules are distinctlyhemitropous, in- sentedin the Australiangenus Micrantheum. Berg stead of anatropousas in the great majorityof relatedmany of the distinctivefeatures of the eu- generain the family.Singh (1962) has contrasted phorbiaceouspistillate flower, such as the vascular the anatropousand hemitropoustypes and illus- patternin the columella,to adaptive aspects of tratedan unusual variationin Trevia, wherethe fruitdehiscence and seed dispersal.He regarded ovule is ventrallyadnate to the placentaand lacks the placentationin Micrantheumas not typically a distinctfunicle. The ovulesof Crotonand Trevia axile because the three columellarbundles lie on illustratedby Singh are characteristicof the sub- differentradii fromthe dorsal bundles; he thus familiesCrotonoideae and Acalyphoideaerespec- describedthe columellarportion of the axis of the tively;in the Crotonoideaethe ovule tends to be ovaryas "an internalcarpophore congenitally fused elongatedand has a nucellarbeak projectingbe- withdown-hanging lobes fromcarpels attachedto yondthe micropyle,whereas in the Acalyphoideae itstip." However,as shownin thedescriptions and the ovule is less elongatedwith a thickenedinner illustrationsof Venkata Rao & Ramalakshmi integumentand withouta projectingnucellus. (1968), the threebasal centralstrands in the axis Ovules in the Euphorbioideaeconform to those of each give rise to pairedstrands that represent the the Acalyphoideae. ventralsof the carpel margins; this strongly implies Considerableembryological work has been done thatthe placentationin the euphorbiaceousflower on the Euphorbiaceae(Davis, 1966; Rao, 1970), is typicallyaxile, and Berg's ingenioushypothesis showingthat embryosac developmentis of the is unnecessary. normal Polygonum type in most of the family. Anotherprovocative suggestion of Berg is that Aberrantdevelopment has been reportedonly in the primaryfunction of the obturatoris not to the Acalyphoideaeand Euphorbioideae:disporic facilitatepassage of the pollentube but ratherto (Allium) types occur in Chrozophora and Eu- displace the seed lowerinto the coccus formore

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 14 Annals of the MissouriBotanical Garden

effectiveseed dispersal. Against this view may be subfamilies differ from the Phyllanthoideae in cited the great variation in obturator dimensions sometimes having tegmic vascular bundles, this withinthe capsular-fruitedtaxa (although it is pos- character is inconsistent: Corner's groups 1 and 3 sible that size variation could be correlated with contain representatives of all three uniovulate sub- differentialgrowth of the ovary overall). families,and group 2 contains both Acalyphoideae Euphorbiaceous seeds show variation in size, and Crotonoideae. Nevertheless, it is quite possible shape, seed-coat ornamentation and anatomy, en- that this character is phylogenetically significant. dosperm development, and configuration of the Most of the genera with tegmic vasculature belong embryo; seed characters are taxonomically impor- to the Crotonoideae, and Gelonium, witha "tegmic tant, especially at specific and generic levels. In pachychalaza," may be closest to the ancestral the biovulate taxa, there are typically two seeds condition. per locule, but the number is often reduced to one Although the caruncle is by no means a ubiq- (e.g., in Amanoa, Meineckia, and Savia). In bi- uitous feature of the euphorbiaceous seed Baillon ovulate taxa with drupaceous fruits, only one or (1873) censured Mueller for using it as a generic two seeds per fruitmay develop (e.g., in Antides- character it stillis of considerable taxonomic and ma, Aporusa, Baccaurea, and Drypetes). Seed evolutionary interest. Bresinsky (1963) and Berg size varies greatly, from less than 1 mm in her- (1975a, b) described the structural and functional baceous species of Chamaesyce or Phyllanthus to features of the caruncles in Euphorbia and Mi- over 4 cm in Hevea and Omphalea. crantheum. As Schweiger (1905) demonstrated, Netolitzky(1926), Wunderlich (1968), and Cor- the caruncle always arises from the micropylar ner (1976) have shown that the anatomical struc- region (exostome) of the outer integument. In Eu- ture of the seed coat is of considerable importance phorbia, Micrantheum, and doubtless the other in relating the Euphorbiaceae to other families, as genera with carunculate seeds, the caruncle func- well as providing clues about infrafamilialrelation- tions as an elaiosome to attract ants for dispersal. ships. The testa (from the outer integument) per- Berg noted that in Micrantheum it may also play sists in the mature seed and sometimes is thickened a role in the build-up of tension within the devel- and fleshy (e.g., in Baccaurea, Mallotus, Mar- oping capsule, leading up to the explosive dehis- garitaria, Melanolepis, Sapium, and Tetrorchi- cence. dium). The mechanical layer of the seed is formed Carunculate seeds have been recorded from all from the outer epidermis of the tegmen (inner of the subfamilies except the Phyllanthoideae (and integument),and Corner reported that it is different even in that subfamilythere are sometimes micro- in the uniovulate and biovulate taxa. In the Phyl- pylar outgrowthsthat could be regarded as incip- lanthoideae, the mechanical cells consist of cuboid ient caruncles). However, their distributionwithin (or only slightlyradially elongate) sclerotic cells or each subfamilyis inconsistent; closely related gen- of tangentiallyelongated fibers. In contrast, in the era may differin presence or absence of a caruncle, uniovulate taxa the mechanical layer is usually while in various genera (e.g., Euphorbia, Stil- formedof radially elongated palisade sclerenchyma lingia) some species have a caruncle and others ("Malpighian") cells. On the basis of this difference lack it. Caruncles are invariably lacking on seeds in seed structure, Corner referred the uniovulate from indehiscent fruits, which accounts for part taxa (Crotonoideae sensu Pax) to a Malvalean af- (but by no means all) of the spottiness in reports finityand the Phyllanthoideae to a Celastralean of distribution. affinity,but frankly admitted that the picture is A seed character that appears particularly im- not crystal-clear. portant in the Phyllanthoideae is the presence or In my opinion, the primitive condition in the absence of endosperm, which develops as the Nu- family would seem to be the fibrousexotegmen as clear type (Rao, 1970). It is copious in most genera, found in Phyllanthoideae such as Andrachne, An- but scanty or absent in relatively primitivegenera tidesma,Aporusa, Baccaurea, and Briedelia. The such as Actephila, Amanoa, Spondianthus,and palisade cells in Glochidion seem to be derived Wielandia. In exalbuminous seeds, the cotyledons fromthe shortened fibersfound in Breynia, Flueg- are often distinctly folded. Outside the Phyllan- gea, and Phyllanthus; but contrary to Corner's thoideae, seeds normally have well-developed en- interpretation,the condition in Glochidion does dosperm, but there are a few exceptions (e.g., not seem to be transitional to that found in the Elateriospermum in the Crotonoideae). uniovulate taxa. Instead, the tegmen with cuboidal Cotyledon characters have been given excep- mechanical cells of Drypetes may be closer to the tional weight in the classificationof Euphorbiaceae main line of specialization. Although the uniovulate since they were strongly emphasized by Mueller

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 15 1994 Classificationof the Euphorbiaceae

(1866), who erected his two primary infrafamilial number of x = 10, with other numbers originating divisions,Platylobeae and Stenolobeae, on the basis in a descending aneuploid series. Jones & Smith of cotyledon shape. Baillon (1873) rightlyobjected (1969) reporteda surprisingdiversity of basic num- to this overemphasis on the cotyledon character, bers in Monadenium: x = 12, 16, 17, 18, and and his rejection of Mueller's scheme has been 19. Since the prevalent number in Monadenium substantiated by the palynological studies of Erdt- is x = 18, and the lowest number of x = 12 occurs man (1952), Punt (1962), and K6hler (1965), in a species (M. echinulatum Stapf) ranked as which clearly demonstrated that the Stenolobeae highlyspecialized in the monograph of Bally (1961), are an artificial group. One can sympathize with the most probable hypothesisis that these numbers Bentham (1880) in his decision to maintain the are of polyploid origin from an ancestor with x = Stenolobeae as a tribe because of the strikingcor- 9, followedby a mostlydecreasing aneuploid series. relation between cotyledon shape, ericoid habit, Unfortunately,all of the cytologically known po- and Australasian distribution.As a matter of fact, tential succulent ancestral species in Euphorbia it is still not clear why the character of narrow subg. Euphorbia are reported to have x = 10, so cotyledon shape should be almost confined to Aus- the case remains not proven. tralia, except for the anomalous North American genus Reverchonia (Webster & Miller, 1963). SUMMARY OF TAXONOMIC CHARACTERS IN THE EUPHORBIACEAE E. CYTOLOGICAL CHARACTERS At this writing, only one comprehensive phy- Perry (1943) was the firstto provide a survey logenetic analysis of Euphorbiaceae using both phe- of chromosome numbers for the Euphorbiaceae, netic and cladistic methods has been made. Levin but his sample was very small, and a considerable (1986b, c) classified the genera of subfamilyPhyl- number of his counts have proved to be erroneous. lanthoideae primarilyon the basis of leaf anatom- Three decades later, Hans (1973) presenteda much ical data. A cladistic study of Oldfieldioideae based more extensive (and reliable) survey of chromo- on pollen characters has recently been completed some numbers; and there have been a few addi- (Levin & Simpson, 1994a, b). However, critical tional surveys since then, such as that of Urbatsch phylogenetic analysis of the family is really just et al. (1975). Although the percentage of species beginning. Considerably more morphological, an- sampled is stillvery low (below 5%), there are now atomical, and biochemical (not to mention molec- data on a sufficientnumber of genera to discuss ular) work needs to be done before the data set is the possible implications for phylogeny and clas- adequate. sification. Hans concluded that there are two fun- In Table 1 are listed a basic set of primitiveand damental basic numbers in the family: x = 7 and derived character states for the Euphorbiaceae, x = 13, of which the latter is presumed to be based on both the literature and personal obser- derived from the firstby doubling and aneuploid vations. It is focused on the Phyllanthoideae as the reduction. Although this scenario cannot be dis- sister group of the other subfamilies, and also the proved, it is notable that x = 13 is the prevailing sister group of the unknown collateral (or ancestral) number in the "lower" Phyllanthoideae, and it closest related family. I have used a pastiche of seems unwarranted to invoke a hypothetical an- several potential "ancestral" families as a com- cestral complement of x = 7. Hans's reason for posite sister group; this includes families of Linales doing so is probably a consequence of his following and Sapindales (sensu Cronquist, 1981) such as Bentham's (1880) tribal arrangement, and unfor- Ixonanthaceae, Irvingiaceae, and Linaceae. The tunately the worst mistake in Bentham's treatment ovular structure in these Linales is suggestively is his inclusion of the Buxaceae (with x = 7) as similar to that in the Euphorbiaceae (Webster, one of the tribes. 1967). Other possible families,such as the Thyme- Hans (1973) is probably correct in concluding laeaceae, should be tested as sister groups, and this that x = 11 is the original base number in the might well alter the list of characters. Acalyphoideae and Crotonoideae. The same con- Furtherrefinement of this crude and preliminary clusion can also be made for the Euphorbioideae, list of character states in Euphorbiaceae can be as x = 11 is prevalent in the Hippomaneae. Hans expected from more intensive study of fossils as regarded x = 7 and 10 as the primary base num- well as living plants. Since I reviewed the earlier bers for Euphorbia, but this implies that the genus reports of euphorbiaceous fossils (Webster, 1967), is polyphyletic,which is scarcely warranted. In my data provided by fossil wood and fruits has been opinion, the evidence clearly suggests an original supplemented by discoveries of euphorbiaceous

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 16 Annals of the MissouriBotanical Garden

flowers in the Eocene. Crepet & Daglian (1982) guez and his collaborators have not yet demon- have described a genus Hippomaneoidea fromthe strated convincingly that the benefitfrom visits of Eocene Claiborne formation that appears similar Polistes wasps is not merely incidental (or super- to the modern genus Senefeldera. This raises the fluous) with respect to a pollination system that encouraging possibilitythat furtherdiscoveries may depends on visitations by a variety of generalist provide a picture of the past evolution of euphor- visitors. It should be noted, though, that Eremo- biaceous reproductive structuresand a firmerbasis carpus, a specialized derivative group surely de- for polarizing character states. Although the sys- rived from within Croton (and better treated as a tematic affinitiesof fossil fruitshave proved more section of that genus), does appear to be wind- difficultto establish, the study of Mazer & Tiffney pollinated. The situation in Croton suberosus could (1982) on fruitsof Hippomaneae from the North represent a preliminary step toward anemophily American Eocene indicates that it may be possible and merits furtherexperimental study. to correlate flower and fruit characters in fossil The lack of specificityin most euphorbiaceous taxa of Euphorbiaceae. flower/pollinatorrelationships clearly has not im- peded diversification in floral and inflorescence THE ECOLOGICAL AND BIOGEOGRAPHIC BACKGROUND structure;however, the number of genera that have received even superficial study is very small. A 1. POLLINATION number of instances of more unusual adaptations Despite the reproductive diversity in the Eu- for pollination have been demonstrated, and no phorbiaceae, the limited studies on pollination in- doubt many remain to be detected. Pollination by dicate that the flowers of the majority of genera Diptera is no doubt usually unspecialized, as re- are not highly specialized for a single kind of pol- corded by Hauman-Merck (1912) for a species of lination vector. Although wind pollination is pre- Sapium in Argentina,where species of genera such sumably the norm in a limited number of genera as Eristalis and Sarcophaga were recorded re- (e.g., Acalypha, Macaranga, Mallotus) withan- moving nectar from extra-floralnectaries on the emophilous characteristics such as reduced peri- floral bracts; as Hauman suggested, a similar syn- anth and nectar, it has been demonstrated exper- drome may occur in other taxa of tribe Hippo- imentallyonly in a few genera such as Mercurialis maneae, but there are few recent observations. In (Daumann, 1972) and Ricinus (Alex, 1957). Per- contrast to this, however, Warmke (1952) dem- haps the typical situation might be that demon- onstrated that Hevea in Brazil is naturally polli- strated forthree species of Chamaesyce in Arizona nated by midges (Heleidae). Although furtherver- by Ehrenfeld (1976), who collected 175 species of ificationof this is desirable, it appears that Hevea insects, including small bees, flies, and wasps. A may depend on midges of a single family for pol- similar spectrum of visitors was reported for a lination, and this might well prove to be the case North American and a South American species of for other euphorbiaceous genera. Jatropha by Simpson (1977), and for Jatropha In a few genera of Euphorbiaceae, floral struc- gossypiifolia L. in India by Reddi & Reddi (1983). ture has become modified for pollination by Lepi- Steiner (1983) reported visits by a variety of bees, doptera. Perkins et al. (1975) reported nocturnal flies,and butterfliesto the flowersof Croton draco pollination of Cnidoscolus texanus (Muell. Arg.) Chain. & Schltdl. in Panama. Recently it has been Small in Oklahoma by two species of hawkmoths. claimed (Dominguez & Bullock, 1989; Dominguez However, Breckon, in observations on four species et al., 1989) that the Mexican species Croton of Cnidoscolus in Mexico (pers. comm.), found pseudoniveus Lundell and C. suberosus HBK are pollination primarilyby sphingids in C. herbaceus wind-pollinated,even though nectaries are present (L.) I. M. Johnston,but mostly by butterfliesin C. in both staminate and pistillate flowers. The ar- aconitifolius (Miller) I. M. Johnston and C. mul- gument is made that the role of nectar in Croton tilobus (Pax) I. M. Johnston; the most important suberosus is to attract wasps that deter lepidop- butterflyfamilies were Papilionidae and Nymphal- teran predators from attacking the foliage. How- idae, but significantnumbers of visits were made ever, other reports such as that of Jose & Inamdar by taxa of Heliconidae, Pieridae, and Hesperiidae. (1989) on the role of floral nectaries in Croton In Brazil the situation in Cnidoscolus is similar, bonplandianus Baill. support adaptation for insect except that near Jequie, Bahia, visits to Cnidosco- pollination. In my opinion, the arguments for wind lus urnigerus Pax were seen to be almost entirely pollinationin Croton, in view of the floraland pollen by hummingbirds(Webster, unpublished). The pe- morphology, are not entirely persuasive; Domin- culiar urn-shaped corolla and the androecium of

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 17 1994 Classificationof the Euphorbiaceae

distinct(not monadelphous) stamens appear to have bracts are usually glandless. In contrast, the cy- evolved in this one species concomitantlywith a athium of the Euphorbieae appears derived from shiftfrom lepidopteran to ornithophilouspollina- the spicate thyrse of the Hippomaneae, in which tion. the nectaries of the floral bracts have become in- Pollinationby small Hymenopterais doubtless corporated as the nectaries of the pseudanthium; verywidespread in the Euphorbiaceae.In Califor- theories about the nature of the cyathium have nia, wasps were foundto be the most important been reviewed by Dressler (1957) and Webster pollinatorsfor Croton and Euphorbiaby Moldenke (1967). Other pseudanthia, of quite a different (1976). In Suriname,pollination of Tragia lesser- nature, are known in Uapaca (Phyllanthoideae) tiana (Baill.) Muell. Arg. appearedto be primarily and Pera (Acalyphoideae); but nothing is known by Melipona (Webster & Dobson, unpublished), about their reproductive biology. althoughspecies of Trigonawere common visitors The pseudanthium of Euphorbia and Chamae- to this and otherEuphorbiaceae. Further obser- syce is generally pollinated by small generalist in- vationsof pollinationby smallwasps and bees may sects (flies, wasps, bees) as reported by Ehrenfeld help to elucidate some of the variationsin eu- (1976) and earlier workers. However, the type phorbiaceousfloral structure that at presentappear species of Euphorbia subgenus Poinsettia, Eu- mysterious. phorbia pulcherrima, has very conspicuous red Larger bees presumablyare major pollinators bracts subtending the cyathia (forming a second- forEuphorbiaceae with relatively large and showy order pseudanthium); this, and the copious nectar, flowers,such as Aleurites,Manihot, and species suggest hummingbirdpollination. Indeed, Dressler of Croton;however, critical field observations are (1957) clearly illustrated how the more or less stillalmost entirely lacking. The major exception radially symmetrical cyathium of Euphorbia has involvesDalechampia, which has small flowers been highly modified into the bilaterally symmet- aggregatedinto a conspicuouspseudanthium that rical cyathium of Pedilanthus that is pollinated in mostspecies producesresin for a floralreward mainly by hummingbirds. (Webster& Webster,1972). Armbruster& Web- At least fragmentary accounts of most major ster(1979) demonstratedthat in Mexicopollination pollination syndromes of Euphorbiaceae were pub- of one species, Dalechampia magnistipulata lished by the end of the 19th century. However, Webster & Armbruster,was affectedby female Steiner (1981, 1983) demonstrated mammal pol- euglossinebees visitingthe inflorescenceto gather lination of Mabea occidentalis Benth. by bats the resin; in the other species, D. spathulata (Glossophaga, Carollia) and the red woolly pos- (Scheidw.)Baill., whichhas modifiedresin glands, sum (Caluromys), which visit the inflorescences to visitswere by male euglossinebees gatheringvol- obtain nectar from the extrafloralnectaries on the atileterpenoids. Subsequently, in a seriesof papers bracts. An even more remarkable instance of bat Armbruster(1984; 1988; Armbruster& Herzig, pollination has been reported by Steiner (1982) in 1984) describedpollination of a large numberof another genus of Hippomaneae, the sandbox tree Dalechampia species, whichdepend on visitsby (Hura). In contrast to Mabea, in which the sta- anthidiineand euglossinebees. The considerable minate and pistillate flowers are associated in the speciationin Dalechampia (over 100 species)may same inflorescence, in Hura the fleshy staminate plausiblybe relatedto exploitationof thisunique inflorescencesand the solitarypistillate flowers with "niche" in the pollinationspectrum of tropical bizarrely enlarged stigmas behave as separate pol- plants. Armbruster(1992) showed a fascinating lination units. Steiner suggested that the staminate correlationbetween phylogeny of taxa in Dale- "cones are taken by bats mistaking them for champia and evolutionof pollinationmechanisms; fleshy fruits certainly the most striking case of he convincinglydemonstrated that pollinationby "mistake pollination" in the Euphorbiaceae. fragrance-collectingeuglossine bees has evolved These examples indicate that the great diversity independentlyin threedifferent species lineages, of euphorbiaceous floral structure is at least partly The othermajor occurrence of pseudanthialin- related to a long history of shiftingadaptation to florescencesin the Euphorbiaceaeis in the tribe differentpollinators. However, because the flowers Euphorbieae,where modification of floral structure of most Euphorbiaceae are less than 1 cm in di- has occurredalong an entirelydifferent pathway. ameter, the effectiveunit of pollination has often As indicatedby Venkata Rao (1971), the inflores- been the inflorescence or subinflorescence. Func- cence structureof Dalechampia may be derived tional studies of evolution of reproductive struc- fromthe thyrseof Acalypheae,in whichthe floral tures must therefore take into account the inflo-

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 18 Annals of the MissouriBotanical Garden

TABLE 2. Taxonomicdistribution of disseminuletypes tralianoldfieldioid genus, Micrantheum,is men- in the Euphorbiaceae.A numberof generahave notbeen tioned,in whichdispersal of the carunculateseeds scoreddue to lack of definiteinformation. dependedon attractionof ants to the elaiosome. Berg'sstudy remains unique in itssynthesis of field Num- Fruit observationson ant dispersalof Micrantheum seeds ber Seeds Seeds inde- of carun-ecarun- Testa his- witha carefulanatomical study of fruitand seed Subfamily genera culate culate fleshy cent development;he also providedcomparative data on myrmecochoryin severalother Australian gen- Phyllanthoideae 58 0 38 5 12 Oldfieldioideae 28 11 14 0 2 era of Euphorbiaceae(1975a, 1981). Since a very Acalyphoideae 118 18 77 13 5 rudimentaryelaiosome occurs in some Phyllan- Crotonoideae 61 19 29 5 8 thoideae(e.g., Phyllanthus),it is notunreasonable Euphorbioideae 42 16 18 1 4 to suggestthat selectionfor ant dispersalwas an Totals 307 64 176 24 31 importantfactor in the evolutionof fruitand seed morphologyin the Euphorbiaceae.Berg plausibly suggestedthat the diagnosticcharacter syndrome of the euphorbiaceousgynoecium and fruit in- rescence as a whole, as well as its constituentparts. cludingthe reducedovule number,obturator, the Furthermore, as discussed below, adaptations for axisbecoming a persistentcolumella, and theclosed dispersal may also have a "retrospective" effect ventralwall of the coccus representa specialized in projecting developmental restraints back onto derivationfrom an ancestralstate of nonexplosive the flowers. capsular fruits,and that morphologicalchanges involvedin dispersalhave as a consequencerelayed 2. DISPERSAL fromthe fruitback intothe flower.Berg also sug- The prevalent fruit type in the primitive Eu- gested,somewhat less persuasively,that this char- phorbiaceae (Phyllanthoideae), and indeed in most acteristic euphorbiaceous gynoecium/fruit/seed of the taxa of the family,is the 3-carpellate schizo- charactersyndrome arose in a tropicalmonsoon carp that dehisces explosively into three segments climate;this could be so, but thereare manytaxa (cocci) each with one or two seeds. Typically, the withtypical explosive fruits in the tropicalrain- central axis of the flower(carpel margins and pla- forest.Intensive investigations on othergenera are centae) remains in the fruit as a columella after badlyneeded in orderto test Berg's hypotheses. the cocci have dispersed. In the subfamily Phyl- Some of the relationshipsinvolved in the adap- lanthoideae, the individual seeds may have a dry tive aspects of dispersalare givenin Tables 2-4. or fleshy testa, but there is never the large con- It is notable fromTable 2 that "unspecialized" spicuous elaiosome (caruncle) found in other sub- seeds (i.e., ecarunculate,with dry seed coat) pre- families; the caruncle appears to be a novelty (in- dominatein all fivesubfamilies. However, in the dependently evolved?) in the subfamilies Oldfieldioideaeand Euphorbioideaethere are near- Oldfieldioideae and Acalyphoideae. ly as many carunculateas ecarunculategenera. Under the description of fruit and seed mor- The Acalyphoideaehave the lowestpercentage of phology, Berg's (197 5b) detailed study on the Aus- the foursubfamilies with carunculate seeds, but in compensationhave the largest numberof seeds withfleshy testa. TABLE 3. Relationof dispersaltype to geographic Geographically(Table 3), thereare some inter- distributionof genera: carunculateseeds are treatedas estingdifferences in average dispersaltype, since myrmecochorous;fleshy seeds and indehiscentfruits as theneotropical and Australasiangenera have many ornithochorous.Many genera,not enumeratedin the ta- more myrmecochorousgenera than ornithocho- ble, are probablyautochorous. rous,whereas in the restof the paleotropicalarea the situationis reversed. Myrmeco- Ornitho- chorous chorous In view of the interesttropical ecologists have taken in the role of dioecy as a reproductivead- Neotropics 18 10 aptation,it is ofsome interest to comparethe types Africa/Madagascar 8 21 of fruitsand seeds with the monoeciousversus TropicalAsia 6 28 dioeciousconditions within Euphorbiaceae (Table Australasia 21 6 Holarctic 4 0 4). The data supportthe predictions of Bawa (1980), Givnish(1982), and Steiner(1988) thatdioecious Totals 57 65 taxa shouldhave a higherpercentage of ornitho-

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 19 1994 Classificationof the Euphorbiaceae

chorous disseminules than do monoecious taxa. In TABLE 4. Relationof sexual expressionto dispersal contrast, there appears to be no significantdiffer- type in genera of Euphorbiaceae.M indicatesgenus is ence in the expectation of monoecy or dioecy in monoecious;M/D, monoeciousor dioecious(in same or myrmecochorous taxa, or in the "unspecialized" differentspecies of a genus); D, consistentlydioecious. taxa, which are presumably mostly autochorous. Seeds Seeds/ Although Berg is surely correct in suggesting carun- fruits Totals that the autochorous tricoccous fruitsrepresent a Seeds culate fleshy for primitivestate, and myrmecochorydeveloped sub- dry (myrme-(ornith- each sequently, furtherevolution to fleshy,more or less (auto- cocho- ocho- sub- indehiscent fruits has occurred in all of the sub- Subfamily chorous) rous) rous) family families. Within the Phyllanthoideae there are also Phyllanthoideae a number of fruitsthat appear adapted fordispersal M 11 0 3 14 by birds: examples include berries in Breynia, and M/D 5 0 1 6 drupesin Drypetes and Uapaca. In Phyllanthus D 22 0 9 31 (Webster, 1956), there is a particularly marked Oldfieldioideae progression from schizocarps (subgenus Isocladus) M 1 3 0 4 to berries (subgenus Kirganelia) and finally to M/D 1 1 0 2 drupes (subgenus Cicca). Within dry fruits,there D 12 7 3 22 has also been an elaboration of seeds with sarco- Acalyphoideae testa; this is very often reddish, as in Macaranga M 44 4 6 54 or Mallotus, but in Margaritaria it is blue or green. M/D 3 0 1 4 The baccate fruits in Euphorbiaceae are watery D 30 14 11 55 and dispersed by nonspecialist frugivores (Snow, Crotonoideae 1981). Presumably, drupaceous fruits are dis- M 15 persed by specialist frugivores,but we have very 13 4 32 M/D 0 2 0 2 little data on Euphorbiaceae. A striking develop- D 14 4 9 27 ment in Acalyphoideae and Crotonoideae is the evolution of large oily seeds in genera such as Euphorbioideae Caryodendron,Hevea, Joannesia,Aleurites, and M 14 13 0 27 Omphalea. The seeds of Hevea are explosively M/D 2 2 1 5 D 2 1 4 dispersed to great distances (up to 45 m according 7 to van der Pijl, 1982) and then dispersed by rivers Totals in the Amazon basin. Hevea and these other large- M 85 33 13 131 seeded taxa are mostly rainforestdominants (Om- M/D 11 5 3 19 phalea is a liana), but another group of genera D 80 26 36 142 such as Homonoia, rheophytes that are adapted 176 64 52 292 to extensive periods of submergence, also presum- ably depend mainly on water dispersal. According to Gottsberger(1978), none of the Euphorbiaceae biaceae appear to have been ecarunculate, and observed in the Amazon basin (including Hevea) well-developed caruncles are unknown in the Phyl- are dispersed by fish, but rather the seeds are lanthoideae. It is interestingthat carunculate seeds preyed upon. Wind dispersal is rare in Euphorbi- occur in "basal" taxa of the Oldfieldioideae (Tetra- aceae and occurs mainly in taxa with samaroid coccus), Acalyphoideae (Clutia), and Crotonoideae fruits adapted for open tropical woodland (e.g., (Micrandra, Manihot). Seeds with sarcotesta are Hymenocardia); there are no good examples of found in relatively "advanced" genera (e.g., Mar- winged seeds in the Euphorbiaceae. garitaria, Breynia) in Phyllanthoideae, but in rel- Bresinsky (1963), on the basis of studies of atively "basal" genera in Acalyphoideae (Cheilosa) elaiosomes, hypothesized that seed morphology in and Crotonoideae(Tetrorchidium, Klaineanthus). temperate taxa with myrmecochory has been de- It appears that diaspore evolution has followed a rived from ornithochoryin tropical taxa. This as- tortuous path, with many reversals between dry/ sumes that seeds with sarcotesta (tropical forms) fleshy and carunculate/ecarunculate testa. There have evolved into carunculate seeds (temperate is a general tendency for seeds with sarcotesta to forms), with Mercurialis showing an intermediate occur in closed forestedhabitats and for those with condition.However, the real situationappears much caruncules to occur in open dry habitats. What is more complex. Primitively,seeds in the Euphor- evident, however, is there is no unilinear direction

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 20 Annals of the MissouriBotanical Garden

TABLE 5. Summaryof diversityin subfamiliesand tribesof Euphorbiaceaeas measuredby numbersof native and endemicgenera (endemic genera indicated in parentheses).Tropical Arabia is countedas part of Africa,Pacific islandswith Australasia.

Subfamily America Africa Madagascar Asia Australasia Phyllanthoideae Wielandieae 4 (3) 4 (3) 3 (2) 1 (0) 1 (0) Amanoeae 1 (0) 2 (1) Bridelieae 2 (0) 2 (0) 2 (0) 2 (0) Phyllantheae 8 (3) 14 (4) 6 (1) 10 (1) 7 (0) Drypeteae 1 (0) 3 (2) 1 (0) 2 (1) 1 (0) Antidesmeae 6 (5) 7 (4) 4 (1) 4 (1) 5 (2) Hymenocardieae 1 (1) 1 (0) 1 (0) Bischofieae 1(0) 1(0) Totals 21(12) 33 (14) 16(4) 21(3) 17(2) Oldfieldioideae Croizateae 1 (1) Podocalyceae 3 (3) Caletieae 1(1) 2(1) 13 (12) Picrodendreae 4 (4) 4 (3) 3 (2) 1 (1) Totals 8 (8) 5 (4) 3 (2) 3 (2) 13 (12) Acalyphoideae Clutieae 1 (1) Pogonophoreae 1 (0) 1(0) Chaetocarpeae 1 (0) 1 (0) 1 (0) 2 (1) Pereae 1 (1) Cheiloseae 2 (1) 1 (0) Erismantheae 3 (2) 1 (0) Dicoelieae 1 (1) Galearieae 2 (1) 2 (0) 1 (0) Ampereae 2 (2) Agrostistachydeae 2 (2) 2 (1) 1 (0) Chrozophoreae 5 (4) 2 (0) 6 (4) 1 (0) Caryodendreae 2 (2) 1 (1) Bernardieae 2 (2) 2 (2) 1 (1) Pycnocomeae 2 (1) 2 (1) 4 (2) 2 (0) Epiprineae 2 (1) 2 (0) 8 (6) Adelieae 4 (4) 1 (1) Alchorneae 5 (4) 1 (0) 3 (2) 1 (0) 2 (2) Acalypheae 3 (1) 11 (6) 10 (5) 14 (4) 11 (3) Plukenetieae 11 (8) 3 (1) 3 (1) 5 (3) Omphaleae 1(0) 1 (0) 1 (0) 1 (0) 1(0) Totals 36 (26) 33 (17) 22 (9) 52 (26) 23 (7) Crotonoideae Micrandreae 4 (4) Manihoteae 2 (2) Adenoclineae 3 (2) 3 (2) 1 (0) Gelonieae 1 (0) 1 (0) 1 (0) 1(0) Elateriospermeae 1 (1) Jatropheae 3 (2) 2 (1) 3 (2) Codiaeae 4 (4) 1 (0) 8 (5) 5 (3) Trigonostemoneae 1(0) 1(0) Ricinocarpeae 1 (1) 6 (6) Crotoneae 4 (3) 2 (1) 1 (0) 2 (1) 1 (0) Ricinodendreae 2 (1) 1 (0) 1 (0) Aleuritideae 2 (2) 7 (6) 3 (2) 4 (3) 1 (0) Totals 22 (19) 18 (11) 6 (2) 23 (13) 15 (9)

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 21 1994 Classificationof the Euphorbiaceae

TABLE 5. Continued.

Subfamily America Africa Madagascar Asia Australasia Euphorbioideae Stomatocalyceae 1 (1) 2 (2) 1 (0) 1 (0) Hippomaneae 14 (10) 4 (2) 1 (0) 5 (1) 4 (0) Pachystromateae 1 (1) Hureae 4 (4) Euphorbieae 4 (2) 7 (4) 3 (0) 2 (0) 4 (2) Totals 24 (18) 13 (8) 4 (0) 8 (1) 9 (2) Grandtotals 111 (83) 102 (54) 51 (17) 107 (45) 77 (32)

of seed-coat specialization in any of the subfamilies some of the most strikingdisjunct or relict distri- except for the Phyllanthoideae, which lack myr- butions. There do not appear to be any critically mecochory, and the Oldfieldioideae, which are documented maps of suprageneric taxa in the lit- overwhelminglymyrmecochorous (Table 2). The erature, and since the discussion of Bentham there geographical distributionof myrmecochorous and have been few analyses besides the study of the ornithochorous diaspores (Table 3) is interesting, distributionof Jatropheae by Pax (1910), and (par- with the neotropical and Australian regions mainly tially) of the Euphorbieae by Croizat (1972) and myrmecochorous, and the paleotropical region or- Leach (1976). nithochorous. Particularlynotable is the prevalence A review of the geographic distributionof the of seeds withfleshy testa in the paleotropical region, tribes and genera of subfamily Phyllanthoideae and the overwhelming predominance of myrme- brings out the strikingprominence of Africa and cochorous dispersal in Australia, confirmingthe Madagascar (Table 5). The highly disjunct distri- statements of Berg (1975a, b, 1981). bution of the primitivegenus Savia is striking(Fig. Little has been studied so far on the effects of 1); including the related genus Heywoodia, the seed predation on dispersal systems in Euphorbi- diversity in Africa/Madagascar is clearly greater aceae. The seed-coat toxins of such genera as Ric- than the New World. The occurrence of the relict inus could be explained as a deterrentto predators, genus Wielandia in Madagascar and the Sey- but this apparently has not been studied in the field chelles is noteworthy; together with the endemic in Africa, where natural predators might be ex- genera Blotia and Petalodiscus, it indicates a ma- pected. However, there is a suggestive study by jor center in Malagasia (Fig. 2). Discocarpus and Cook et al. (1971) on the effectsof seed predation Lachnostylis apparently form a vicariant pair in by doves on the seed morphology of Croton setig- South America and Africa, and the neotropical and erus Hook. in California. Here there is a kind of paleotropical species of Savia seem equally vicar- "balanced polymorphism" between gray unpat- iant. This South American/African disjunction is terned seeds with toxins produced by plants in the repeated at the level of groups of species withina deserts and mottled palatable seeds that are pro- number of other genera of Phyllanthoideae, such duced by coastal populations. Cook et al. (1971) as Amanoa, Drypetes,Margaritaria, Meineckia, argued plausibly that the mottled seed patterns of and Phyllanthus. It also occurs in other subfam- coastal plants are maintained by the adaptive ad- ilies, as noted in Table 5. It seems particularly vantage of camouflage, whereas the unpatterned informativeto look at the distributionof the more seeds of desert areas indicate lower levels of pre- primitivegenera withineach subfamily.It is notable dation. The situation is reminiscent of that in Ric- that in the Oldfieldioideaethe "basal" tribesCroiza- inus, which also has notoriouslyvariable seed-coat teae and Podocalyceae are entirelyneotropical (Fig. patterns. 3), whereas the Caletieae are entirely Old World, and only the Picrodendreae (Fig. 4) are represented in both hemispheres. Within the Picrodendreae, 3. GEOGRAPHIC DISTRIBUTION the genera Celaenodendron,Piranhea, and Old- A detailed analysis of the geographic distribution fieldia appear to be vicariants that link America patternsof the 50 tribesof Euphorbiaceae is clearly and Africa in the manner of Savia. The Laurasian beyond the scope of this essay. However, it seems genus Tetracoccus is taxonomically isolated, but appropriate here to outline some of the major geo- clearly seems to belong in the Podocalyceae; it graphic patterns of the fivesubfamilies and indicate may represent the taxon in which the important

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 22 Annals of the MissouriBotanical Garden

PHYLLANTHOIDEAE Wielandieae

FIGURE 1., Distributionof Phyllanthoideae,tribe Wielandieae: 1, Heywoodia (dottedline, Africa);2, Savia (continuousline); 6, Actephila (dashed line); 9, Chonocentrum(dotted line, America). apomorphy of carunculate seeds firstappeared in tropicalgenus, Nealchornea, two African,Ham- the family. ilcoa and Plagiostyles, and one oriental,Pime- In the Acalyphoideae, there is one relict neo- lodendron.It is interestingthat the "advanced" tropical/Africangenus, Pogonophora, one of east tribe Euphorbieae (Fig. 10) has a distributional and southern Africa, Clutia, one neotropical and patternsomewhat paralleling the Stomatocalyceae, Asiatic, Chaetocarpus, and three Asiatic, Chei- withtribe Anthosteminae in Africa/Madagascar, losa, Neoscortechinia,and Trigonopleura(Figs. the Neoguillauminiinaein Australia/NewCaledo- 5, 6). Africa does not appear to be the epicenter nia, and the Euphorbiinae(via shrubbyspecies of for oldfieldioidsand especially acalyphoids that it subg. Esula) with an apparent focus in Africa/ is for the Phyllanthoideae. It is notable that in the Madagascar. To me, this suggeststhat as stated Acalyphoideae, in contrast to the biovulate sub- by Raven & Axelrod (1974), deploymentof a families, basal taxa such as Cheiloseae and Chae- significantnumber of tribesof Euphorbiaceaemay tocarpeae are well developed in the Oriental region. have takenplace beforeCretaceous plate tectonic The Crotonoideae (Figs. 7, 8) differfrom all the movementsseparated the taxa of the familyinto other subfamilies in a clear predominance of neo- their presentdisjunct situations. Fossil evidence tropical relict genera: Ditta, Glycydendron, He- (summarizedfor North America by Taylor,1990) vea, Micrandra, Micrandropsis,and Tetrorchi- is stilllimited, but the fact that fossilsof the rel- dium. However, Tetrorchidium also occurs in atively"advanced" tribeHippomaneae have been Africa, and Adenocline and Klaineanthus are con- recordedfrom the Eocene (Crepet& Daglian,1982; finedthere. The Crotonoideae thereforeshow con- Dilcher & Manchester,1988) suggeststhat the siderable parallelism to the Phyllanthoideae, except originand initialspread of all fivesubfamilies may that the center of gravity is definitelyto the west go back into the Cretaceous,or the earliestTer- of the Atlantic rather than to the east. Finally, the tiary.For example,the pollen recordsfrom Aus- primitiveEuphorbioideae of tribe Stomatocalyceae tralasiadiscussed by Martin(1974, 1978) indicate (Fig. 9) give an ambiguous picture, with one neo- that relatively"advanced" genera of Oldfieldioi- deae were in place in Australiain the Paleocene, whichsurely implies differentiation ofthe subfamily 3 Distributionsare based on referencescited in Webster well back intothe Cretaceous. pri- (1994) and examinationof herbariumspecimens, It is not easy to summarizethe overall distri- marilyat DAV. Genusnumbers are the same as in Web- ster(1994). Distributionsare necessarilysomewhat gen- butionalhistory of the Euphorbiaceae,but I am eralizedand do not showminor gaps. struckby the remarkable insightfulness ofthe essay

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 23 1994 Classificationof the Euphorbiaceae

PHYLLANTHOIDEAE Wielandieae

10

FIGURE 2.3 Distributionof Phyllanthoideae,tribe Wielandieae (continued): 3, Gonatogyne,(dotted line, Brazil); 4, 5, Blotia, Petalodiscus (dottedline, Madagascar); 7, Discocarpus (dashedline, America); 8, Lachnostylis(dashed line,Africa); 10, Wielandia (solidline). by Bentham (1878), who presented the firstmodel Order was in the Old World, but that several of for the geographic patterns involved in the origin the principal forms were differentiatedand widely and radiation of the Euphorbiaceae. Bentham's spread before that remote period when the present words deserve to be quoted: ". . . we may be led impassible barriers opposed by the Atlantic and to conjecture that the most ancient home of the Pacific did not exist, or were crossed over in some

OLDFIELDIOIDEAE Croizateae Podocalyceae

63 ~ ~ 6

FIGURE 3.3 Distributionof Oldfieldioideae.Tribe Croizateae: 61, Croizatia. TribePodocalyceae: 62, Podocalvx: 63, Tetracoccus;64, Paradrypetes.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 24 Annals of the MissouriBotanical Garden

OLDFIELDIOIDEAE

Picrodendreae -

FIGURE 4.3 Distributionsof Oldfieldioideae,tribe Picrodendreae: 79, Celaenodendron(dot, NorthAmerica); 80, Piranhea (dashed line); 81, Picrodendron(solid line, NorthAmerica); 82, Parodiodendron(dot, South America); 83, Oldfieldia(solid line, Africa). mannerof which no plausibleexplanation has been of the new Americanforms." Later in this essay suggested.It would also appear that interchange Benthamclearly stated "that the most ancient home of formsbetween the principalOld-World centres ofthe order was in theOld World,whence it spread of differentiation. . . continuedlong afterthe in- in veryremote times to America." terpositionof the obstaclespreventing the spread This scenarioproposed by Benthamover a cen-

ACALYPHOIDEAE Clutieae Pogonophoreae Cheiloseae

FIGURE 5.3 Distributionof Acalyphoideae.Tribe Clutieae: 89, Clutia (solid line, Africa).Tribe Pogonophoreae: 90, Pogonophora(dashed line). Tribe Cheiloseae (solid line, Asia; rangesof the twogenera combined); 94, Cheilosa; 95, Neoscortechinia.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 25 1994 Classificationof the Euphorbiaceae

ACALYPHOIDEAE Chaetocarpeae

1\91

92

FIGURE6.3 Distributionof Acalyphoideae, tribe Chaetocarpeae: 91, Trigonopleura(dashed line); 92 Chaetocarpus (solidline). tury ago now appears as a prescient anticipation like most plant geographers of his time did not of models of intercontinental relationships based reject "continental drift" as an explanation. The on plate tectonic models. Pax (1924), although not very limited paleontological record still does not mentioning Bentham's essay, accepted much the provide critical evidence for Euphorbiaceae, but same overall explanation of intercontinental"land the distributionsand relationships of the taxa are bridge" distributionsin Euphorbiaceae, and un- now better understood, so that Bentham's model

CROTONOIDEAE Micrandreae Manihoteae

209

f20 217

FIGURE 7*3 Distributionof Crotonoideae.Tribe Micrandreae(dotted line, genericranges combined);205, Mi- crandra [and 206, Micrandropsis; 207, Cunuria; 208, Hevea]. Tribe Manihoteae(northern and southernlimits): 209, Manihot; 210, Cnidoscolus.Tribe Gelonieae:217, Suregada.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 26 Annals of the MissouriBotanical Garden

CROTONOIDEAE Adenoclineae

2131

FIGURE 8.3 Distributionof Crotonoideae,tribe Adenoclineae:211, Glycydendron(solid line, America); 212, Klaineanthus(solid line, Africa); 213, Tetrorchidium(dashed line); 214, Adenocline(dotted line, Africa); 215, Ditta (dottedline, North America); 216, Endospermum(solid line, Asia & Australasia).

can be refined. It now appears that Africa/Mad- the Acalyphoideae and Euphorbioideae, the trail agascar retains the largest number of primitivetaxa of clues leads back to a joint South American/ of Phyllanthoideae, and is the most likely center African center, but a narrower designation of orig- of origin for the family,but South America would inal locality must await furtherstudies. seem to be the cockpit of initial differentiationfor With regard to the role of long-distance dispersal both the Oldfieldioideaeand Crotonoideae. For both in possibly accounting for these geographic pat-

EUPHORBIOIDEAE Stomatocalyceae

FIGURE 9.3 Distributionof Euphorbioideae,tribe Stomatocalyceae: 272, Plagiostyles (solid line, Africa);273, Pimelodendron(solid line, Asia); 274, Hamilcoa (dottedline); 275, Nealchornea (solidline, America).

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 27 1994 Classificationof the Euphorbiaceae

EUPHORBIOIDEAE Euphorbieae -

301

300 302

FIGURE 10.3 Distributionof Euphorbioideae, tribe Euphorbieae. Subtribe Anthosteminae: 300, Anthostema(dashed line,Africa & Madagascar);301, Dichostemma(solid line, Africa). Subtribe Neoguillauminiinae: 302, Neoguillauminia (solidline, New Caledonia); 303, Calycopeplus (dashed line,Australia). terns, which Thorne (1972) and more recently appears correct in claiming that American/African Carlquist (1983) impliedmay be operative formany disjuncts such as Mayaca, Pitcairnia, and Sac- tropical intercontinentaldisjunctions, it is notable coglottis represent recent long-distance dispersal, that the "basal" taxa in the Phyllanthoideae, Old- but he goes too far in concluding that this expla- fieldioideae,and Acalyphoideae have capsular fruits nation may be extended to all of the taxa that link with dry seed-coats, which certainly do not suggest America with Africa. The transoceanic vicariant a high degree of "preadaptation" for transoceanic patterns of genera and tribes in the Euphorbiaceae dispersal. Only in the Adenoclineae of the Crotonoi- agree with those of many other tropical families deae (Fig. 8) and the Stomatocalyceae of the Eu- mapped and discussed by Axelrod (1970, 1972, phorbioideae (Fig. 9) do fleshyfruits or seeds occur 1975); these strongly suggest that the differenti- that seem more likely to have been transportedby ation of genera and tribes was decisively influenced birds. The weedy nature of many. Euphorbiaceae by ocean-floor spreading. would indeed appear to predispose them for success There is a thirdalternative explanation fortrans- in establishmentafter transoceanic dispersal. How- Atlantic or trans-Pacific distributionsin Euphor- ever, the patterns reviewed above suggest that in biaceae, and that is overland migration at high the Euphorbiaceae, as in many other angiosperm latitudes during the early Tertiary; Wolfe (1972) families,the spectacular ocean crossings have been proposed that some tropical groups crossed into achieved by taxa (e.g., Euphorbia) that appear to America fromAsia duringthe Eocene by the Bering have evolved later in family evolution, at a time land bridge. It is notable, however, that the genera when the continents were known to be widely sep- he cited Meliosma,Sagaretia, and Saurauia- arated. As suggested by Schuster (1976), the pro- are cloudforest rather than lowland tropical taxa; gressive refinementof dispersal capabilities during the suggested presence of Lecythidaceae and My- the course of the Tertiary leads to the paradox risticaceae is more significant,but requires confir- that the most widely dispersed genera are the latest mation. The distributionof Euphorbiaceae does not evolved, and so the dramatic cases of long-distance offermuch support for this Beringian alternative transoceanic dispersal often noted in the literature migrationalpathway. In the Phyllanthoideae, a pos- tell us a great deal about the geologically recent sible example is Flueggea (Webster, 1984); how- stocking of oceanic islands but very littleabout the ever, this genus has also colonized Hawaii (Hayden, migrationalhistory of less highly evolved taxa that 1987). There are also possible trans-Pacificor trans- were dispersing in the late Cretaceous or earliest Atlantic links in the subtribe Antidesminae. The Tertiary. Because of this paradox, Thorne (1973) Oldfieldioideae show no indication of possible use

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 28 Annals of the MissouriBotanical Garden

of the Alaskan corridor. In the Acalyphoideae and LITERATURE CITED Crotonoideae, an intrusionfrom Asia into America AIRY SHAW, H. K. 1972. The Euphorbiaceaeof Siam. is suggested by the Tertiary records of Aleurites Kew Bull. 26: 191-363. and Macaranga (Wolfe & Leopold, 1967). There . 1975. The Euphorbiaceaeof Borneo. Kew is littlesign of this in living taxa, however; perhaps Bull. Add. Ser. 4: 1-245. . 1980. The Euphorbiaceaeof New Guinea. the best example is in Crotonoideae tribe Codiaeae, Kew Bull. Add. Ser. 8: 1-253. where the North American genus Acidocroton ap- ALEX, A. H. 1957. Pollinationof some oil seed crops pears to be a vicariant of the Asian Blachia. In by honeybees. Texas A. & M. Coll. Agr. Exp. Sta. the Euphorbiaceae, therefore,an Eocene migration Progr.Rep. 1960. ANTON, R. 1974. Etudechimiotaxonomique sur le into America from Eurasia seems to have been an genre Euphorbia(Euphorbiac6es). These, Universit6Louis unimportantpathway. For explaining transoceanic Pasteur,Strasbourg. links at the tribal level, a convincing model will ARMBRUSTER, W. S. 1982. Seed productionand dis- have to involve plate tectonics as a major deter- persal in Dalechampia (Euphorbiaceae):Divergent minant. patternsand ecologicalconsequences. Amer. J. Bot. 69: 1429-1440. Although the overall ecology of the Euphorbi- 1984. The role of resinin angiospermpolli- aceae has not been reviewed in detail, it is worth- nation:Ecological and chemicalconsiderations. Amer. while noting that genera of subfamily Phyllan- J. Bot. 71: 1141-1160. thoideae withmany plesiomorphic characters (e.g., 1988. Multilevelcomparative analysis of the morphology,function, and evolutionof Dalechampia Heywoodia, Savia) occur as shrubs or small trees (Euphorbiaceae)blossoms. Ecology 69: 1746-1761. in seasonal woodlands or scrub formations in the 1992. Phylogenyand the evolutionof plant- Southern Hemisphere. Except for the Acalyphoi- animalinteractions. BioScience 42: 12-20. deae, the "basal" genera in the other subfamilies & A. L. HERZIG. 1984. Partitioningand shar- occur in rainforestor at least more mesic habitats; ing of pollinatorsby foursympatric species of Dal- echampia (Euphorbiaceae)in Panama. Ann. Mis- thus the spectacular efflorescence of xerophytes souriBot. Gard. 71: 1-16. such as the African succulent Euphorbieae, Jatro- & G. L. WEBSTER. 1979. Pollinationof two pha, and Manihot, appears to be a later devel- species of Dalechampia (Euphorbiaceae)in Mexico opment. In all five subfamilies,insect pollination is by euglossinebees. Biotropica11: 278-283. AXELROD, D. I. 1970. Mesozoic paleogeographyand the plesiomorphic condition, and wind pollination earlyangiosperm history. Bot. Rev. 36: 277-319. has evolved independently within each subfamily. . 1972. Ocean-floorspreading in relationto Relatively few Euphorbiaceae have become canopy ecosystematicproblems. In: R. T. Allen & F. C. trees in rainforest;rather, the familyhas specialized James (editors),A Symposiumon Ecosystematics. in an opportunisticlife-style involving colonization Univ. Ark.Mus. Occ. Pap. 4: 15-68. 1975. Plate tectonicsand problemsof angio- of fragmented, shifting,or extreme (e.g., desert) spermhistory. M6m. Mus. Nat. Hist. Nat. II. 88A: habitats. 72-86. In closing this essay, I feel obliged to emphasize BAILLON, H. 1858. Etude G6n6raledu Groupedes Eu- the large amount of conjecture that has been nec- phorbiac6es.Victor Masson, Paris. essary in discussing the major patterns of tribal . 1873. Nouvelles observationssur les Eu- phorbiac6es.Adansonia I. 11: 72-138. differentiationand continental deployment in the BALLY, P. R. 0. 1961. The genusMonadenium. Ben- Euphorbiaceae. Perhaps we can keep our feet on teli,Berne. the ground if we imagine our reconstructed phy- BAMBER, R. K. 1974. Fibretypes in woodsof Euphor- logenies as a fireworksdisplay: colorful, intricate, biaceae. Austral.J. Bot. 22: 629-634. cleverly contrived, and evanescent. There may be BANCILHON, L. 1971. Contributiona l'6tude taxono- mique du genrePhyllanthus (Euphorbiac6es). Bois- a correspondence here between the mental activity siera 18: 1-81. of the systematistand the master-plan of evolution: BAWA, K. S. 1980. Evolutionof dioecy in flowering ifwe allow the rockets to representfamily ancestors plants.Ann. Rev. Ecol. Syst. 11: 15-39. (or plesiomorphic states), and their trajectory the BELIN-DEPOUX, M. 1977. Introductiona l'6tude des onward rush in evolutionary time, then the first glandesfoliaires de l'Alchorneacordata (Juss.) Muell. Arg.(Euphorbiaceae). Rev. G6n.Bot. 84: 127-136. burst of the rocket's red glare will represent the & D. CLAIR-MACZULAJTYS. 1974. Introduction formationsof tribes, subsequent pops genera, and a l'6tudedes glandesfoliaires de l'Aleuritesmoluc- the final transient fizzles, the species. The sedate cana Willd. (Euphorbiac6e)I. La glandeet son on- appearance of cladograms in our scholarly papers togenese.Rev. G6n. Bot. 81: 335-351. represents an austere, abstract schematization of & . 1975. Introductiona l'6tudedes glandesfoliaires de l'Aleuritesmoluccana Willd. (Eu- the turbulent, convoluted, and ultimately partly phorbiac6e)II. Aspectshistologiques et cytologiques unknowable historyof the Euphorbiaceae and other de la glande p6tiolairefonctionelle. Rev. G6n. Bot. plant taxa. 82: 119-155.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 29 1994 Classificationof the Euphorbiaceae

BENTHAM, G. 1878. Notes on Euphorbiaceae.J. Linn. stiubungs6kologievon Mercurialis L. Preslia 44: Soc. London,Bot. 17: 185-267. 308-315. 1880. Euphorbiaceae.In: G. Bentham& J. D. DAVIS,G. 1966. SystematicEmbryology of Angio- Hooker(editors), Gen. P1. 3: 239-340. sperms.Wiley & Sons, New York. BERG, R. 1975a. Myrmecochorousplants in Australia DEHAY, C. 1935. L'appareil libero-ligneuxfoliare des and theirdispersal by ants.Austral. J. Bot. 23: 475- Euphorbiac6es.Ann. Sci. Nat. Bot. X. 17: 147-190. 508. DEHGAN, B. 1980. Applicationof epidermal morphology * 1975b. Fruit,seed, and myrmecochorousdis- to taxonomicdelimitations in the genusJatropha L. persalin Micrantheum(Euphorbiaceae). Norwegian (Euphorbiaceae).Bot. J. Linn. Soc. 80: 257-278. J. Bot. 22: 173-194. . 1982. Comparativeanatomy of the petiole * 1981. The role of ants in seed dispersalin and infragenericrelationships in Jatropha(Euphor- Australianlowland heathland. Pp. 51-59 in R. L. biaceae). Amer.J. Bot. 69: 1283-1295. Specht(editor), Heathlands and RelatedShrublands: & M. E. CRAIG. 1978. Types of laticifersand AnalyticalStudies [Ecosystemsof the World 9B]. crystalsin Jatropha and theirtaxonomic implica- Elsevier,Amsterdam. tions.Amer. J. Bot. 65: 345-352. BERNHARD, F. 1966. ContributionAl'6tude des glandes & G. L. WEBSTER. 1979. Morphologyand foliaireschez les Crotonoidees(Euphorbiacees). Mem. infragenericrelationships of the genus Jatropha (Eu- Inst. Fond. Afr.Noire 75: 71-156. phorbiaceae).Univ. Calif.Publ. Bot. 74: 1-73. BOR, J. & R. N. KAPIL. 1976. Anatropyand ontogeny DILCHER, D. L. & S. R. MANCHESTER. 1988. Investi- of the bitegmicovule in ChrozophoraA. H. L. Jus- gationsof angiospermsfrom the Eocene of North sieu(Euphorbiaceae). Acta Bot. Neerl. 25: 385-400. America: A fruitbelonging to the Euphorbiaceae. BRECKON, G. 1975. Cnidoscolus,Section Calyptosolen TertiaryRes. 9: 45-58. (Euphorbiaceae)in Mexico and Central America. DINGLER, H. 1885. Die Flachsprosseder Phaneroga- Ph.D. DissertationTitle. Universityof California, men. ErstesHeft: Phyllanthus sect. Xylophylla.T. Davis. Ackermann,Munich. BRESINSKY, A. 1963. Bau, Entwicklungsgeschichteund DOMINGUEZ, C. A. & S. H. BULLOCK. 1989. La re- Inhaltsstoffeder Elaiosomen.Bibl. Bot. 126: 1-54. producci6nde Crotonsuberosus (Euphorbiaceae) en BULLOCK, S. H. 1982. Componentesfenol6gicos del luz y sombra.Rev. Biol. Trop. 37: 1-10. sistemade cruzamientomon6ico de Cnidoscolusspi- , R. DIRzo & S. H. BULLOCK. 1989. On the nosus(Euphorbiaceae) en Jalisco.Bol. Soc. Bot. Mex. functionof floralnectar in Crotonsuberosus (Eu- 42: 1-9. phorbiaceae).Oikos 56: 109-114. CARANO, E. 1926. Ulterioriosservazioni su Euphorbia DRESSLER, R. L. 1957. The genus Pedilanthus (Eu- dulcis L. in rapportocol suo comportamentoapomit- phorbiaceae).Contr. Gray Herb. 182: 1-188. tico. Ann. Bot. Roma 17: 50-79. EHRENFELD, J. 1976. Reproductivebiology of three CARLQUIST, S. 1970. Wood anatomyof Hawaiian, Mac- species of Euphorbia subgenus Chamaesyce (Eu- ronesianand other species of Euphorbia. Bot. J. phorbiaceae).Amer. J. Bot. 63: 406-413. Linn. Soc. 63 (suppl.): 181-193. EICHLER, A. W. 1875. Bliitendiagramme.Wilhelm Eng- . 1983. Intercontinentaldispersal. Sonderbd. elmann,Leipzig. naturwiss.Ver. Hamburg7: 37-47. ERDTMAN, G. 1952. PollenMorphology and PlantTax- CESCA, G. 1961. Ricerchecariologiche ed embriologiche onomy.Almqvist & Wiksells,Uppsala. sulleEuphorbiaceae. I-su alcunibiotipi di Euphor- FROEMBLING, W. 1896. Anatomische-systematische bia dulcis L. della Toscana. Caryologia14: 79-96. Untersuchungvon Blattund Axe der Crotoneenund 1969. Cytologicaland embryologicalstudies Euphyllantheen.Inaug.-Dissert., Gebr. Gotthelft, in the genus Euphorbia: Euphorbia epithymoides Cassel. L. Proc. Natl. Inst. Sci. India (Nat.) 35: 139-152. GAUCHER, L. 1902. Recherchesanatomiques sur les COOK,A. D., P. R. ATSATT & C. A. SIMON. 1971. Doves Euphorbiacees.Ann. Sci. Nat. Bot. 15: 161-309. and dove weed: Multipledefenses against avian pre- GIBBS, 1974. Chemotaxonomyof FloweringPlants. 4 dation.Bioscience 21: 277-281. vols. McGill-Queen'sUniv. Press, Montreal& Lon- CORNER, E. J. H. 1976. The Seeds of Dicotyledons.2 don. vols. CambridgeUniv. Press, Cambridge. GILBERT, M. G. 1994. The relationshipsof the Eu- CREMERS, G. 1977. Architecturev6getative de quelques phorbieae.Ann. MissouriBot. Gard. 81: 283-288. especesmalgaches du genreEuphorbia L. Bull.Jard. GIVNISH, T. J. 1982. Outcrossingversus ecological con- Bot. Nat. Belg. 47: 55-81. straintsin the evolutionof dioecy.Amer. Nat. 119: CREPET, W. L. & C. P. DAGLIAN. 1982. Euphorbioid 849-865. inflorescencesfrom the Middle Eocene Claibornefor- GOTTSBERGER, G. 1978. Seed dispersalby fishin the mation.Amer. J. Bot. 69: 258-266. inundatedregions of Humaita,Amazonia. Biotropica CROIZAT, L. 1943. New or criticalEuphorbiaceae of 10: 170-183. Brazil.Trop. Woods 76: 11-14. HALLE, F. 1971. Architectureand growthof tropical 1972. An introductionto the subgenericclas- trees exemplifiedby the Euphorbiaceae.Biotropica sificationof Euphorbia L. withstress on the South 3: 56-62. Africanand Malagasy species. III. Webbia 27: 1- & R. A. OLDEMAN. 1970. Essai sur l'archi- 221. tectureet la dynamiquede croissancedes arbres CRONQUIST, A. 1981. An IntegratedSystem of Classi- tropicaux.Masson, Paris. ficationof FloweringPlants. ColumbiaUniv. Press, , R. A. OLDEMAN & P. B. TOMLINSON. 1978. New York. TropicalTrees and Forests:An ArchitecturalAnal- DAUMANN, E. 1972. Zur Bliitenmorphologieund Be- ysis. SpringerVerlag, Heidelberg.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 30 Annals of the MissouriBotanical Garden

HANS, A. S. 1973. Chromosomalconspectus of the relationto angiospermtaxonomy. Pp. 272-295 in Euphorbiaceae.Taxon 22: 591-636. P. K. K. Nair. (editor),Glimpses in Plant Research, HAUMAN-MERCK, L. 1912. Observationsd'6thologie flo- vol. 5. Vikas Publ. House, New Delhi. rale sur quelques especes argentineset chiliennes. KNOLL, F. 1905. Die Brennhaareder Euphorbiaceen- Rec. Inst. Bot. Leo Errera 9: 3-20. GattungenDalechampia und Tragia. Sitzungsber. HAYDEN, W. J. 1977. Comparativeanatomy and sys- Kaiserl. Akad. Wiss. Math.-Nat.Klasse 114: 29- tematicsof Picrodendron,genus Incertae sedis. J. 48. ArnoldArb. 58: 257-279. K6HLER, E. 1965. Die Pollenmorphologieder Biovu- 1980. SystematicAnatomy of Oldfieldioideae latenEuphorbiaceae und ihre Bedeutung fur die Tax- (Euphorbiaceae). Doctoral DissertationTitle, Uni- onomie.Grana Palynol.6: 26-120. versityof Maryland. LEACH, L. C. 1976. Distributionaland morphological 1987. The identityof the genus Neowawraea studiesof the tribe Euphorbieae (Euphorbiaceae) and (Euphorbiaceae).Brittonia 39: 268-277. theirrelevance to its classificationand possibleevo- 1988. Ontogenyof the cotyledonaryregion lution.Excelsa 6: 3-19. of Chamaesyce maculata (Euphorbiaceae).Amer. LEONARD,J. 1962. Euphorbiaceae.Fl. Congo 8(1): 1- J. Bot. 75: 1701-1713. 214. , W. T. GILLIS, D. E. STONE, C. R. BROOME & & M. MOSANGO. 1985. Hymenocardiaceae. G. L. WEBSTER. 1984. Systematicsand palynology Flore d'AfriqueCentrale: Spermatophytes. of Picrodendron:Further evidence for relationship LEOPOLD, E. B. & H. D. McGINITIE. 1972. Develop- withthe Oldfieldioideae(Euphorbiaceae). J. Arnold mentand affinitiesof Tertiaryfloras in the Rocky Arbor.65: 105-127. Mountains.Pp. 147-200 in A. Graham (editor), HEGNAUER, R. 1962-73. Chemotaxonomieder Pflan- Floristicsand PaleofloristicsofAsia and EasternNorth zen: Eine Ubersichtfiber die Verbreitungund der America.Elsevier, Amsterdam. SystematischeBedeutung der Pflanzenstoffe.6 vols. LEROY, J.-F. 1976. Recherchessur la natureet l'origine Birkhauser,Basel. de la fleurangiospermienne: Interpretation des struc- HERBERT, H. 1897. AnatomischeUntersuchung von tures dans un groupe singulierd'Euphorbiaceae. Blattund Axe der Hippomaneen.Inaug.-Diss., Mu- Compt. Rend. Hebd. S6ances Acad. Sci., Ser. D. nich. 283: 247-250. HICKEY, L. J. & J. A. WOLFE. 1975. The bases of LEVIN, G. A. 1986a. Systematicfoliar morphology of angiospermphylogeny: Vegetative morphology. Ann. Phyllanthoideae(Euphorbiaceae), I. Conspectus.Ann. MissouriBot. Gard. 62: 538-589. MissouriBot. Gard. 73: 29-85. HURUSAWA, I. 1954. Eine nochmaligeDurchsicht des 1986b. Systematicfoliar morphology of Phyl- herk6mmlichenSystems der Euphorbiaceenim wei- lanthoideae(Euphorbiaceae). II. Pheneticanalysis. terenSinne. J. Fac. Sci. Univ. Tokyo,Bot. 6: 209- Ann. MissouriBot. Gard. 73: 86-98. 342. 1986c. Systematicfoliar morphology of Phyl- HUTCHINSON, J. 1969. Tribalismin the familyEuphor- lanthoideaeIII. Cladisticanalysis. Syst. Bot. 11: 515- biaceae. Amer.J. Bot. 56: 738-758. 530. INAMDAR, J. A. & M. GANGADHARA. 1977. Studieson & M. G. SIMPSON. 1994a. Phylogeneticim- the trichomesof some Euphorbiaceae.Feddes Re- plicationsof pollenultrastructure in the Oldfieldioi- pert.88: 103-1 1 1. deae (Euphorbiaceae).Ann. MissouriBot. Gard. 81: JANSSONIUS, H. H. 1929. A contributionto the natural 203-238. classificationof the Euphorbiaceae. Trop. Woods 19: & M. G. SIMPSON. 1994b. Phylogeneticre- 8-11. lationshipsof Didymocistus and Hymenocardia(Eu- 1930. Mikrographiedes Holzes der auf Java phorbiaceae).Ann. MissouriBot. Gard. 81: 239- vorkommendenBaumarten [Euphorbiaceae, pp. 442- 244. 835]. Brill,Leiden. MAHESHWARI,P. 1942. The embryo-sacof Euphorbia JENSEN, U., I VOGEL-BAUER & M. NITSCHE. 1994. Legu- heterophyllaL.-A reinvestigation.Proc. Indian minlikeproteins and the systematicsof the Euphor- Acad. Sci. 15B: 158-166. biaceae. Ann. MissouriBot. Gard. 81: 160-179. MAHLBERG,P. G. 1975. Evolutionof the laticiferin JONES, K. & J. B. SMITH. 1969. The chromosome Euphorbia as interpretedfrom starch grain mor- identityof MonadeniumPax and SynadeniumPax phology.Amer. J. Bot. 62: 577-583. (Euphorbiaceae).Kew Bull. 23: 491-493. , D. G. DAVIS, D. S. GALITZ & G. D. MANNERS. JOSE, T. & J. A. INAMDAR. 1989. Structure,ontogeny 1987. Laticifersand theclassification of Euphorbia: and biologyof nectariesin Croton bonplandianus The chemotaxonomyof Euphorbia esula L. Bot. J. Bail. Beitr.Biol. Pflanzen64: 157-165. Linn. Soc. 94: 165-180. JURY, S. L., T. REYNOLDS, D. F. CUTLER & F. J. EVANS MARTIN,H. A. 1974. The identificationof some Ter- (editors). 1987. The Euphorbiales:Chemistry, tax- tiarypollen belonging to the familyEuphorbiaceae. onomic& economicbotany. Bot. J. Linn. Soc. 94: Austral.J. Bot. 22: 271-291. 1-326. 1978. Evolutionof the Australianflora and JUSSIEU, A. 1824. De Euphorbiacearumgeneribus med- vegetationthrough the Tertiary:Evidence from pol- icisque earumdemviribus tentamen. Paris. len. Alcheringa2: 181-202. KAKKAR, L. & G. S. PALIWAL. 1974. Studieson the MAZER, S. J. & B. H. TIFFNEY. 1982. Fruitsof Weth- leaf anatomyof Euphorbia L. Epidermis.Proc. In- erelliaand Palaeowetherellia(?Euphorbiaceae) from dian Nat. Sci. Acad. 40B: 55-67. Eocene sedimentsin Virginiaand Maryland.Brittonia KAPIL, R. N. 1961. Some embryologicalaspects of 34: 300-333. Euphorbia dulcis L. Phytomorphology11: 24-36. MEEUSE, A. D. J. 1990. The Euphorbiaceaeauct. plur.: & A. K. BHATNAGAR. 1980. Embryologyin An unnaturaltaxon. Eburon,Delft.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions Volume 81, Number1 Webster 31 1994 Classificationof the Euphorbiaceae

MENNEGA,A. M. W. 1987. Wood anatomyof the the Symposiumon ComparativeEmbryology of the Euphorbiaceae,in particularof the subfamilyPhyl- Angiosperms.Bull. IndianNat. Sci. Acad. 41: 136- lanthoideae.Bot. J. Linn. Soc. 94: 111-126. 141. METCALFE,C. R. & L. CHALK. 1950. Anatomyof the RAO,A. R. & M. MALAVIYA. 1964. On the latex-cells Dicotyledons.Clarendon Press, Oxford. and latexof Jatropha.Proc. IndianAcad. Sci. 60B: MICHAELIS, P. 1924. BliitenmorphologischeUntersu- 95-106. chungenan den Euphorbiaceen.Bot. Abh. Jena 3: RAO,P. R. & V. S. RAJU. 1985. Foliar trichomesin 1-150. thefamily Euphorbiaceae. Pp. 128-136 in B. Singh MILLER, K. & G. L. WEBSTER. 1962. Systematicpo- & M. P. Singh(editors), Trends in Plant Research. sitionof Cnidoscolus and Jatropha. Brittonia14: Govil & Kumar,Dehra Dun. 174-180. RAVEN, P. H. & D. J. AXELROD. 1974. Angiosperm MOLDENKE, A. R. 1976. Californiapollination ecology biogeographyand past continentalmovements. Ann. and vegetationtypes. Phytologia 34: 305-361. MissouriBot. Gard. 61: 539-673. MUELLER, J. 1866. Euphorbiaceae.DC. Prodr. 15(2): REDDI, E. U. B. & C. S. REDDI. 1983. Pollination 189-1261. ecologyof Jatrophagossypiifolia (Euphorbiaceae). . 1874. Euphorbiaceae[part 2]. In: Martius, Proc. Indian Acad. Sci. (Plant Sci.) 92: 215-231. Fl. Bras. 11(2): 293-750. RITTERSHAUSEN,P. 1892. Anatomisch-systematische NETOLITZKY, F. 1926. Anatomieder Angiospermen- Untersuchungvon Blatt und Axe der Acalypheen. Samen. GebrilderBorntraeger, Berlin. Inaug.-Dissert,.Munich. NOWICKE, J. W. 1994. A palynologicalstudy of Cro- ROssIGNOL,L. & M. ROssIGNOL. 1985. Architectureet tonoideae(Euphorbiaceae). Ann. Missouri Bot. Gard. tendances6volutives dans le genrePhyllanthus (Eu- 81: 245-269. phorbiaceae).Bull. Mus. Nat. Hist. Nat. IV. 7: 67- NOZERAN, R. 1953. Sur quelques fleursmales d'Eu- 80. phorbiacees.Rec. Trav. Lab. Bot. Univ. Montpellier ROTHDAUSCHER,H. 1896. Ueberdie anatomischenVer- Bot. 6: 99-114. haltnissevon Blatt und Axe der Phyllantheen(mit PAX,F. 1884. Die Anatomieder Euphorbiaceen in ihrer Auschlussder Euphyllantheen).Inaug.-Dissert., Cas- Beziehungzum System derselben. Bot. Jahrb. 5: 384- sel (also publishedin Bot. Centrabl.68). 421. Roux, J. 1968. Sur le comportementdes axes aeriens . 1890. Euphorbiaceae.In: A. Engler & K. chez quelques plantes a rameaux v6g6tatifspoly- Prantl(editors), Die Natiirl.Pflanzenfam. ed. 1, 3(5): morphes;le concept de rameaux plagitropes.Ann. 1-119. Sci. Nat. Bot. XII. 9: 109-256. * 1910. Euphorbiaceae-Jatropheae.In A. Eng- RUDALL, P. J. 1987. Laticifersin Euphorbiaceae-A ler (editor),Das PflanzenreichIV. 147. I (Heft42): conspectus.Bot. J. Linn. Soc. 94: 143-163. 1-148. SCHUSTER,R. M. 1976. Plate tectonicsand its bearing * 1924. Die Phylogenieder Euphorbiaceae. on the geographicalorigin and dispersalof angio- Bot. Jahrb.59: 129-184. sperms.Pp. 48-138 in C. B. Beck (editor),Origin & K. HOFFMANN. 1931. Euphorbiaceae.In A. and EarlyEvolution of Angiosperms. Columbia Univ. Engler & K. Prantl(editors), Die Natuirl.Pflanzen- Press, New York. fam. ed. 2. 19c: 11-233. SCHWEIGER,J. 1905. Beitragezur Kenntnis der Samen- PERKINS, G., J.R. ESTES & R. W. THORP. 1975. Pol- entwicklungder Euphorbiaceen.Flora. 94: 339- linationof Cnidoscolustexanus (Euphorbiaceae)in 379. south-centralOklahoma. Southw. Nat. 20: 391-396. SEHGAL,L. & G. S. PALIWAL. 1974. Studieson theleaf PERRY, B. 1943. Chromosomenumber and phyloge- anatomyof Euphorbia-VIII. Generalconclusions netic relationshipsin the Euphorbiaceae.Amer. J. and systematicconsiderations. Phytomorphology 24: Bot. 30: 527-543. 141-151. PIJL, L. VANDER. 1952. The stamensof Ricinus. Phy- SIMPSON, B. B. 1977. Breedingsystems of dominant tomorphology2: 130-132. perennialplants of twodisjunct warm desert ecosys- 1982. Principlesof Dispersal in HigherPlants, tems.Oecologia 27: 203-226. 3rd ed. Springer-Verlag,Berlin. SINGH, R. P. 1962. Formsof ovulesin Euphorbiaceae. PUNT, W. 1962. Pollen morphologyof the Euphorbi- Pp. 124-128 in Plant Embryology,A Symposium. aceae withspecial referenceto taxonomy.Wentia New Delhi. 7: 1-116. SMITH, J. 1841. Noticeof a plantwhich produces per- 1987. A surveyof pollenmorphology in Eu- fectseeds withoutany apparentaction of pollenon phorbiaceaewith special referenceto Phyllanthus. the stigma.Trans. Linn. Soc. London 18: 512. Bot. J. Linn. Soc. 94: 127-142. SNOW, D. W. 1981. Tropical frugivorousbirds and RADCLIFFE-SMITH,A. 1987. Segregatefamilies from the theirfood plants: A worldsurvey. Biotropica 13: 1- Euphorbiaceae.Bot. J. Linn. Soc. 94: 47-66. 14. RADLKOFER, L. 1870. Ueber Pausandra, ein neues SOLEREDER,H. 1899. SystematischeAnatomie der Di- Euphorbiaceen-Genus.Flora 53: 81-95. cotyledonen.Enke, Stuttgart. RAJU,V. S. & P. N. RAO. 1977. Variationin the . 1908. SystematischeAnatomie der Dicoty- structureand developmentof foliarstomata in the ledonen.Erganzungsband. Enke, Stuttgart. Euphorbiaceae.Bot. J. Linn. Soc. 75: 69-97. STEINER, K. E. 1981. Nectarivoryand potentialpolli- & . 1987. The taxonomicuse of the nationby a Neotropicalmarsupial. Ann. Missouri basal stomataltype in the generic delimitationof Bot. Gard. 68: 505-513. Chamaesyce (Euphorbiaceae).Feddes Repert. 98: 1982. MistakePollination of Hura crepitans 137-141. (Euphorbiaceae)by FrugivorousBats. DoctoralDis- RAO,A. R. 1970. Euphorbiaceae.In: Proceedingsof sertationTitle. Universityof California,Davis.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions 32 Annals of the MissouriBotanical Garden

* 1983. Pollinationof Mabea occidentalis(Eu- 1957. A monographicstudy of the West phorbiaceae)in Panama. Syst. Bot. 8: 105-117. Indianspecies of Phyllanthus.J. ArnoldArbor. 38: * 1988. Dioecismand itscorrelates in theCape 51-80, 170-198, 295-373. floraof South Africa. Amer. J. Bot. 75: 1742-1754. . 1958. A monographicstudy of the West TAYLOR, D. W. 1990. Paleobiogeographicrelationships Indianspecies of Phyllanthus.J. ArnoldArbor. 39: of angiospermsfrom the Cretaceousand earlyTer- 49-100, 111-212. tiaryof the North Americanarea. Bot. Rev. 56: . 1967. The genera of Euphorbiaceaein the 279-417. southeasternUnited States. J. ArnoldArbor. 48: THORNE, R. F. 1972. Major disjunctionsin the geo- 303-430. graphicranges of seed plants.Quart. Rev. Biol. 47: 1975. Conspectusof a new classificationof 365-411. the Euphorbiaceae.Taxon 24: 593-601. * 1973. Floristicrelationships between tropical 1979. A revisionof Margaritaria (Euphor- Africaand tropicalAmerica. Pp. 27-47 in B. J. biaceae). J. ArnoldArbor. 60: 403-444. Meggerset al. (editors),Tropical Forest Ecosystems 1984. A revisionof Flueggea (Euphorbi- in Africaand SouthAmerica: A ComparativeReview. aceae). Allertonia3: 259-312. SmithsonianInstitution, Washington. * 1987. The saga of the spurges:A reviewof THURSTON, E. L. 1976. Morphology,fine structure and classificationand relationshipsof the Euphorbiales. ontogenyof the stingingemergences of Tragia ra- Bot. J. Linn. Soc. 94: 3-46. mosa and T saxicola (Euphorbiaceae).Amer. J. Bot. 1994. Synopsisof the genera and suprage- 63: 710-718. nerictaxa ofEuphorbiaceae. Ann. Missouri Bot. Gard. & N. R. LERSTEN. 1969. The morphologyand 81: 33-144. toxicologyof plant stinginghairs. Bot. Review 35: & K. MILLER. 1963. The genusReverchonia 393-412. (Euphorbiaceae).Rhodora 65: 193-207. UHLARZ, H. 1978. Ueber die Stipularorganeder Eu- & E. RUPERT. 1973. Phylogeneticsignificance phorbiaceae,unter besonderer Berucksichtigung ihr- ofpollen nuclear number in the Euphorbiaceae.Evo- er Rudimentation.Trop. Subtrop.Pflanzenwelt. 23: lution27: 524-531. 1-65. & B. D. WEBSTER. 1972. The morphology URBATSCH, L. E., J. D. BACON, R. L. HARTMAN, M. C. and relationshipsof Dalechampia scandens (Eu- JOHNSTON, T. J. WATSON JR. & G. L. WEBSTER. phorbiaceae).Amer. J. Bot. 59: 573-586. 1975. Chromosomenumbers for North American , W. V. BROWN & B. N. SMITH. 1975. Sys- Euphorbiaceae.Amer. J. Bot. 62: 494-500. tematicsof photosyntheticcarbon fixation pathways VENKATA RAO,C. 1971. Anatomyof the inflorescence in Euphorbia. Taxon 24: 27-33. of some Euphorbiaceae.Bot. Not. 124: 39-64. , E. RUPERT & D. KOUTNIK. 1982. Systematic 1972. Floral anatomyof Ricinocarpus pin- significanceof pollennuclear numberin Euphorbi- ifoliuswith some observationson the phylogenyand aceae, tribeEuphorbieae. Amer. J. Bot. 69: 407- centreof origin of Euphorbiaceae. Adv. P1. Morphol. 415. 1972: 85-91. WOLFE, J. A. 1972. An interpretationof AlaskanTer- & T. RAMALAKSHMI. 1968. Floralanatomy of tiaryfloras. Pp. 201-233 in A. Graham (editor), the Euphorbiaceae-I. Some non-cyathiumtaxa. J. Floristicsand PaleofloristicsofAsia and EasternNorth Indian Bot. Soc. 47: 278-300. America.Elsevier, Amsterdam. VINDT, J. 1960. Monographiedes Euphorbiaceesdu . 1975. Some aspects of plant geographyof Maroc. Deuxiemepartie: Anatomy. Trav. Inst. Sci. theNorthern Hemisphere during the late Cretaceous Ch6rifien,Bot. 19: 219-533. and Tertiary.Ann. MissouriBot. Gard. 62: 264- VOGEL, C. 1986. PhytoserologischeUntersuchungen 279. zur Systematikder Euphorbiaceae;Beitrage zur in- WOLFE, J. A. & E. B. LEOPOLD. 1967. Neogene and frafamiliarenGliederung und zu Beziehungenim ex- earlyQuaternary vegetation of northwesternNorth trafamiliarenBereich. Dissert. Bot. 98: 1-124. Americaand northeasternAsia. Pp. 193-206 in D. WARMKE, H. E. 1952. Studieson naturalpollination M. Hopkins(editor), The BeringLand Bridge.Stan- of Hevea brasiliensisin Brazil.Science 116: 474- fordUniv. Press, Stanford. 475. WUNDERLICH, R. 1968. Some remarkson the taxo- WEBSTER, G. L. 1956. A monographicstudy of the nomicsignificance of the seed coat. Phytomorphology West Indianspecies of Phyllanthus. J. Arnold Arbor. 17: 301-311. 37: 91-122, 217-268, 340-359.

This content downloaded from 169.237.8.36 on Mon, 29 Dec 2014 19:08:16 PM All use subject to JSTOR Terms and Conditions