A Revision of Margaritaria (Euphorbiaceae)
Total Page:16
File Type:pdf, Size:1020Kb
JOURNAL OF THE ARNOLD ARBORETUM HARVARD UNIVERSITY VOLUME 60 US ISSN 0004-2625 Journal of the Arnold Arboretum Subscription price $25.00 per year. Subscriptions and remittances should be sent to Ms. E. B. Schmidt, Arnold Arboretum, 22 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A. Claims will not be accepted after six months from the date of issue. Volumes 1-51, reprinted, and some back numbers of volumes 52-56 are available from the Kraus Reprint Corporation, Route 100, Millwood, New York 10546, U.S.A. EDITORIAL COMMITTEE S. A. Spongberg, Editor E. B. Schmidt, Managing Editor P. S. Ashton K. S. Bawa P. F. Stevens C. E. Wood, Jr. Printed at the Harvard University Printing Office, Boston, Massachusetts America, Europe, Asia Minor, and in eastern Asia where the center of species diversity occurs. The only North American representative of the genus, C. caroliniana grows along the edges of streams and in wet, rich soils in forested areas from Nova Scotia and Quebec southward to Florida and westward into Minnesota, Iowa, Missouri, eastern Texas, and Oklahoma. Disjunct popula- • - - Mexico and in Central The stems of Carpinus caroliniana, a small i bluish-gray, sinuate bark and an ; _ _ trees (Fagus spp.). The wood, which i Second-class postage paid at Boston, Massachusetts JOURNAL OF THE ARNOLD ARBORETUM REVISION OF MARGARITARIA (EUPIIORBIACEAE) GRADY L. WEBSTER AM ONC 1 THE SMALLER GENK aiphorbiaceae subfamily Phyllanthoi- deae 1Pax , Margaritaria has a larly broad distribution (MAP 1) in •he New and Old World (except for the Pacific islands). The name' (in alluding to the 'I e C\ n I.i< p i b white cndocarp of the fruit, v . used by Hermann (1689), and the plan. 3 illustrated by Pluke >92). Although the genus was estab- lished hv Linnaeus films in 1 early two centuries passed before it achie\ general acceptance taxonomic literature. Because the :ger Linnaeus included two different genera of plants in the original iption .1 taxonornh understanding ol the genu; \v;is dow in develop- Antoine de Jussieu (1789) and Adrien de Jussieu (1824) both re- ed Mariiaritaria as a gcnu> of uncertain position, allliough the latter- abed a West Indian form of M. nobilis as Cicca antillana A. Juss. on the basis of an observation by Choisy, reported a possible rela- hip of Mariiaritaria with Cicca L. Baillon (1858) treated Margari- in a rather con toed ni.mm i pi i« in< \ uious species in four different on , oi Cicca 3 \\( II i in lli ep it. t< < ( nu \ asp > m u B nllt n "w^.^X orld distribution oi \1< ai aria generalized). Base maps from t and Fellows of Harvard College, 1979. 404 JOURNAL OF THE ARNOLD ARBORETUM [VOL. 60 Mueller Argovicnsis (1866) continued this association between Cicca and Margaritaria but clarified the relationships to some extent by removing Cicca distkha (-- Phvllanthus acidus (L.) Skeels) to a separate subsec- tion of Phyllunthus sect. Cicca (L.) Mueller-Arg. Baillon and Mueller saw that the Old World taxa described under Pro- sorus Dalz. by Dalzell (1852) and Thwaites (1856) were related to the New World Margaritaria. However, they failed to realize that the taxa Margaritaria/Prosorus do not belong in close juxtaposition with Cicca (scn.su stricto). Bentham (1880), who rejected Margaritaria as being founded on. a mixtme of taxa., did point out that Prosorus was discordant with Cicca because of its tricarpellate fruit. Hooker (1887) judiciously maintained Cicca and Prosorus as separate sections of Indian Phvllanthus. However, this felicitous disposition of Hooker was ignored by Tax (1890) and Pax and Hoffmann (1931). who lumped Cicca and Margaritaria to- gether in Phyllanthus sect. Cicca. Although occasional writers of floras (e.g., Britton & Millspaugh, 1920; Gamble, 1924) maintained Margari- taria or Prosorus as distinct genera, most conservative botanists have fol- lowed the usage of Mueller and of Pax, at least until the rehabilitation of Margaritaria by Webster (1957) and Airy Shaw (1966). As I have pointed out earlier (Webster, 1957, 1968), Margaritaria is really quite distinct from Phvllanthus (including Cicca) and indeed is closer to Flucggca Willd. in Phyllantheae subtribe klueggeinae Mueller- Arg. (Webster, 1975). The unique seeds, with fleshy exotesta and thick, bony endotesta separati Ma ritar'u from II otiiei kinllantheae. The fruits of most species of Margaritaria. with a brittle, irregularly shatter- ing, papery endocarp (FIGURKS 5, 6), are equalh distinctive. Staminate material of Margaritaria max he distinguished from that of Flucggca by the lack of a pistillode, and from thai of Phyllanthus by the annular disc subtending a tetramerous androeciuni. The resemblance with Phyllanthus sect. Cicca that impressed 19lh century botanists is very superficial in- deed, sine.' Cicca (liiivtc markedly in its drupaceous fruits with thin- walled, dry seeds and in its very different vegetative structure (with typical phyllanthoid branching, the ultimate axes being short lived and deciduous, in contrast to the persistent mi specialized branches of Mar- Morphologically, the species of Margaritaria aie all rather similar, and the dioecious character of the plants makes it rather difficult to construct workable keys to imlividu I p aniens. Most taxa ire deciduous shrubs or small trees that grow in easonal for< t oi bu h .< fetation. The habit of Margaritaria agrees with that of many other unspeoialized Euphorbia- ceae; Halle ct al. (1978) cite M. discoidca as an example of the architec- tural model of Roux. in which the apline axi ha | u al phyllotaxy but the leaves on the branchlets are distichous. To judge from my own ob- servations of M. nobilis i well i from indi< it ion in the literature, all of the species in the genus would seem io belong to the model of Roux. The entire leaves, with diort petioles and distinct stipules, vary relatively slightly in morphology. Flowers, often produced at the proximal nodes VEBSTER. MARGARITARIA of leaves is mature. Opler. Frankie. and baker (1(>7(>) have reported that in the seasonal forest in Guanacaste. Costa Rica. .1/. nohilis flowers about two weeks after the first spring rain in April or May, and only remains in flower about three days. Other Maryjiritana populations in African and Asian monsoon forests may show similar flowering behavior. The tetra- meroii caly: i rathei imilai in ill no i, md th< annular floral disc is so constant that it provides a frond generic character. Tn all normal staminate flowers (FIGURE 12), the stamens are consistently 4, and the filaments are nearly always free. The tricolporate, semitectate pollen grains of M ar^aritaria are typical of many unspecialized Phyllantheae and in en imilai to tho ( oi in li <r in ra a h a (\ Lint l(<<> Kohler, 1965). The pistillate flowers of Margaritaria ( FIGURES 11. 13) have calyx and disc similar to that of the staminate. The main differential characters are found in the ovary, which has from 2 to 6 carpels. Mueller (1866) used carpel number as a diagnostic character in defining subsections of his sect. Circa, but its value in indicating phvlewenetic relationships is not very great Carpel number has keen independently reduced to 2 in the West Indies (.1/. hottcami). the Indian Ocean islands (.1/. aiwma'a). and .Australia (J/, dubiuni-traccyi) ; within the Arnencan .'!/. nobilis it fluc- Uiat iiom 1 to o in mli in tin ia!>l< tnann i thai it cannot be used to define subspecific taxa. However, despite some overlap, the pattern of carpel number distribution is distinctively different in M. nobilis and the closely related M. discoidea (TABLES 1,2). The fruit oi mo ! [U^M- ot Mar^aritaria an different from those of all other Phyllantheae. The endocarp is thin, papery, and hyaline, and easily shatters to expose the fleshy seeds (FIGURES 5, 6). The seeds are (exotesta), bony sclerotesta (endotesta I. and large hilum. These seeds would appear to be morphologically adapted for bird dispersal: van der Fiji (1969), in fact, discusses .10 nobilis as an example of "deceitful" ornithochorous seeds (although the sarcotesta is sufficiently thick that the notion of deceit is perhaps here misplaced). The scletole-m-i of the seed furnishes a useful systematic character, since it varies considerably in size and sculpturing. Tt is distinctively rugose in 1/. indka and M. rhom- boidalis, but smooth in most other taxa. Cytologically, Mur^aritaria appears to be diploid with a base number of Id (2// '() i hk' m,MO mi peciali/cd ta\a of subfamily Phyllan- thoideae (Webster & Ellis. 1962; Mangenot in Bancilhon, 1971). Al- though there are only two published counts (for ,1/. discoidea and M. no- bilis), the uniformity of the gemis makes it appear unlikrlv that nddi tional count would c\ ! n\ is niti< it a >l<> i< il h rsi Ecologically, species of Manjar/iaria might fairly be characterized as planl ol econdan ucce a'ons in tropical easonal climates, although they may also be found in < limax ot eve rgi '< n fore u Vdjanohoun (1964) reported that M. discoidea is common in secondary stages of moist forest JOURNAL OF THE ARNOLD ARBORETUM [VOL. 60 . 2. Carpel number variation in American and African taxa of Margaritaria. herbarium specimens, M. nobilis displays . Analysis of specimens from the great areas inhabited by M. nobilis, HI. discoidca, and .1/. ind'na. in America. Africa, and Asia, respectively, shows a considerable amount of ill-defined fluctuating morphological variability, as one might expect to find in extensive panmiclic populations. Only in Africa is there more than one distinct mainland continental species. Spe- ciation has occurred mainly in the Caribbean and Indian Ocean islands and thus appears to reflect a "peripheral'' effect; i.e., the definitive geo- 1979] WEBSTER, MARGARITARIA 407 graphic segregation of islands -reins to have been prerequisite to specia- tion, at least outside Africa.